revisiting soil resource limitation: resource predictors of tree growth

Transcription

REVISITING SOIL RESOURCE LIMITATION: RESOURCE PREDICTORS OF TREE
GROWTH AND FOREST PRODUCTIVITY CHANGE ACROSS ECOLOGICAL
GRADIENTS
By
Thomas Wyatt Baribault
A DISSERTATION
Submitted to
Michigan State University
in partial fulfillment of the requirements
for the degree of
DOCTOR OF PHILOSOPHY
Forestry
Ecology, Evolutionary Biology, and Behavior
2011
ABSTRACT
REVISITING SOIL RESOURCE LIMITATION: RESOURCE PREDICTORS OF TREE
GROWTH AND FOREST PRODUCTIVITY CHANGE ACROSS ECOLOGICAL
GRADIENTS
By
Thomas Wyatt Baribault
Multiple soil resources could limit aboveground net primary production (ANPP) and individual
tree growth, but most studies have focused on growth as a function of nitrogen (N) availability in
temperate forests and phosphorus (P) availability in tropical forests. In addition to soil resource
availability, individual tree growth is affected by competition among neighboring trees, but
relationships of local-scale soil resources to tree growth have not been assessed in conjunction
with competition. I hypothesized that temperate tree diameter growth and ANPP are correlated
with soil base cation availability as well as N. Similarly, tropical tree diameter growth and sitelevel basal area increment (BAI) should be correlated not only with soil P, but also with the base
cations. Furthermore, I hypothesized that the relative effect on diameter growth of competition
between neighboring trees strengthens if plant interactions are determined by a tolerancecompetitiveness tradeoff. Alternately, competition should be important at all sites if plant
interactions are governed by species‘ relative abilities to most efficiently exploit available soil
resources. Finally, I hypothesized that diameter growth, leaf nutrient content, and photosynthesis
should increase in response to fertilization of individual temperate forest trees with calcium (Ca)
and / or N. To address these hypotheses, I tested correlations between growth and production and
a suite of resources including water, multiple measures of inorganic N availability, phosphate
(PO4), total P, and extractable base cations (Ca, potassium (K), magnesium (Mg)), in northern
hardwood forests distributed across a multi-resource gradient and in lowland wet tropical forests
growing across a strong soil P gradient. I also developed individual-based growth models as
functions of local neighborhood index and fine-scale estimates of soil resources for single
species in both biomes and a variety of functional groups from the tropical sites.
In temperate forests, wood productivity was strongly correlated with soil Ca at lowfertility sites but with N at high fertility, whereas leaf production was correlated with N across all
sites. Diameter growth of individual temperate trees was unrelated to soil resource availability or
competition among species dominant at low-fertility sites; for species dominant at high-fertility
sites, diameter growth was related to several soil resources (Ca, K, N), and demonstrated strong
correlations with local neighborhood indices. Across tropical sites, individual diameter growth
and site mean BAI were correlated with both base cations and total P. The Fabaceae (legumes)
showed extremely weak correlation of individual growth to soil resources, and no relationship of
site mean BAI to resources. In contrast, the Arecaceae (palms) and species with low wood
density exhibited robust correlations of BAI to total soil P, while BAI was correlated with soil
base cation availability for non-legume dicots and species with higher wood density. My results
were consistent with limitation of temperate forest productivity by N or Ca; anthropogenic N
deposition may therefore have mixed effects, with increasing production in N-limited systems
but decreasing production in base cation-limited systems. Also for temperate forests, an increase
in the strength of competition in species common at high fertility suggests that communities may
be structured according to a tolerance-competition tradeoff. In lowland tropical forests, legumes
may partially escape soil resource limitation, while growth of other functional groups may be
limited by either soil P or base cations; consequently, base cations could be potentially limiting
resources globally.
This dissertation is dedicated to my parents, who inspired me by their example, and especially to
my wife, for whose love and support I am immeasurably grateful.
iv
ACKNOWLEDGEMENTS
I must thank Dr. Richard Kobe not only for his staid guidance throughout my time at Michigan
State University, but for his initial willingness to accept me as a student. Dr. Kobe is not
typically a gambling man, so I appreciate the risk that he assumed by accepting a chemist into an
ecology graduate program. I will always value the diversity of experiences Dr. Kobe made
available to me—course- and field-work in Costa Rica, abundant opportunity for fieldwork in
Michigan, and the environment of personal accountability fostered in his lab group. Perhaps
most importantly, I am grateful for his patience whilst teaching me how to write. I would also
like to thank my committee members, Drs. Stuart Grandy, David Rothstein, and Michael
Walters, for their mentoring. My fourth chapter rests on a practical foundation due to Dr.
Walters‘ recommendations, and Dr. Rothstein was always both helpful and encouraging while
my first chapter slowly took shape. One of the most important contributions of this dissertation,
fine-scale assessment of soil resource heterogeneity, was achieved at the encouragement of Dr.
Grandy. Additionally, I am indebted to Dr. Andrew Finley for his assistance with geospatial
modeling and GIS work, and for commiseration. From Dr. Kobe‘s lab group, I thank Dr. Sarah
McCarthy-Neumann for cheerfully helping me through numerous graduate school phases. Ellen
Holste, M.S., was a reliable colleague and a generous friend. Her work ethic made the fifth
chapter of this dissertation possible. I have watched David Minor progress from an excellent
undergraduate student to experienced field researcher, and I value both our past work and leisure.
Jesse Bramer provided invaluable assistance over many years. I thank Max Erickson for his
helpfulness, honesty, rationality, and remarkable capacity for technical innovation. Finally, I
thank Dr. Phu Nguyen and Carol Graysmith for their assistance, openness, and compassion.
v
TABLE OF CONTENTS
LIST OF TABLES ....................................................................................................................... viii
LIST OF FIGURES ..................................................................................................................... xiii
LIST OF ABBREVIATIONS ....................................................................................................... xv
CHAPTER I
INTRODUCTION .......................................................................................................................... 1
Soil resource limitation of forest productivity ............................................................................ 2
Competition and soil resource availability .................................................................................. 5
Dissertation objectives ................................................................................................................ 7
CHAPTER II
SOIL CALCIUM, NITROGEN, AND WATER ARE CORRELATED WITH
ABOVEGROUND NET PRIMARY PRODUCTION IN NORTHERN HARDWOOD
FORESTS ....................................................................................................................................... 9
Abstract ..................................................................................................................................... 10
Introduction ............................................................................................................................... 12
Methods ..................................................................................................................................... 16
Results ....................................................................................................................................... 22
Discussion ................................................................................................................................. 26
Acknowledgements ................................................................................................................... 34
CHAPTER III
NEIGHBOR INTERACTIONS STRENGTHEN WITH INCREASED SOIL RESOURCES IN A
NORTHERN HARDWOOD FOREST ........................................................................................ 43
Abstract ..................................................................................................................................... 44
Introduction ............................................................................................................................... 46
Methods ..................................................................................................................................... 51
Results ....................................................................................................................................... 60
Discussion ................................................................................................................................. 66
Acknowledgements ................................................................................................................... 43
CHAPTER IV
CALCIUM, NITROGEN, AND MICRONUTRIENT FERTILIZATION OF INDIVIDUAL
CANOPY TREES IN NORTHERN HARDWOOD FORESTS .................................................. 84
Abstract ..................................................................................................................................... 85
Introduction ............................................................................................................................... 87
Methods ..................................................................................................................................... 91
Results ....................................................................................................................................... 97
Discussion ............................................................................................................................... 101
vi
Acknowledgements ................................................................................................................. 108
CHAPTER V
TROPICAL TREE GROWTH IS CORRELATED WITH SOIL PHOSPHORUS, POTASSIUM,
AND CALCIUM, BUT LEGUMES MAY ESCAPE LIMITATION ........................................ 115
Abstract ................................................................................................................................... 116
Introduction ............................................................................................................................. 118
Methods ................................................................................................................................... 122
Results ..................................................................................................................................... 130
Discussion ............................................................................................................................... 136
Acknowledgements ................................................................................................................. 141
CHAPTER VI
SYNTHESIS ............................................................................................................................... 156
Summary ................................................................................................................................. 157
Competition and soil resource correlations transform across soil resource gradients ............ 157
Implications of heterogeneous resource limitation at multiple spatial scales ......................... 159
Conclusions and future directions ........................................................................................... 162
APPENDIX A ............................................................................................................................. 166
APPENDIX B ............................................................................................................................. 172
APPENDIX C ............................................................................................................................. 185
APPENDIX D ............................................................................................................................. 198
BIBLIOGRAPHY ....................................................................................................................... 222
vii
LIST OF TABLES
Table 2.1. Annual productivity increments were calculated between 1999 and 2007 for ANPPW,
ANPPL, and total ANPP ............................................................................................................... 35
Table 2.2. Model formulae, where RS indicates resource and A, B and/or C are estimated
parameters ..................................................................................................................................... 37
Table 2.3. AICc comparisons for top ANPPW models of each resource ...................................... 38
Table 2.4. AICc comparisons for top ANPPL models of each resource ....................................... 39
Table 3.1. Basic statistics, spatial distribution, and relative basal area for the ten most common
species across our sites.................................................................................................................. 73
Table 3.2. Principal results summarized by species. .................................................................... 74
Table 3.3. Supported models for realized growth framework ...................................................... 75
Table 3.4. Site-level correlations of mean individual basal area growth, mean diameter, stem
density, and mortality rate with the soil resource metrics ............................................................ 77
Table 4.1. Basic statistics—including sample size, diameter, annual diameter growth, and basal
area increment (± standard error)—for five treatment groups for each of four common species
..................................................................................................................................................... 109
Table 4.2. Fertilizer treatments and application rates ................................................................. 110
Table 4.3. Results from one-way ANOVA comparing mean values of response variables across
treatments for each species ......................................................................................................... 111
Table 4.4. Results from two-way ANOVA comparing mean values response variables across Ca,
N, Ca+N, and control treatments for each species ...................................................................... 112
Table 5.1. Summary of results for stem growth as a function of soil resources for all species
together, for functional groups, and for wood density categories............................................... 142
Table 5.2. Summary of results for stem growth as a function of soil resources for single species
..................................................................................................................................................... 143
Table 5.3. Results of fitting models of individual tree growth as a function of diameter, soil
resources, and neighborhood index for species for which more than 50 focal individuals were
available ...................................................................................................................................... 144
viii
resources and neighborhood indices for all species combined and categorized by functional
group and size class .................................................................................................................... 146
Table 5.5. Results of fitting models of individual tree growth for species grouped by wood
density as a function of diameter, soil resources, and neighborhood index ............................... 147
Table A.1 Site mean values, with standard error, for inorganic N pools (NO3 and NH4), soil N
dynamics (N-min rate and nitrification), and total C and N content .......................................... 166
-1
Table A.2. Site mean values for soil acidity and Al (cmol kg ), base cations (Ca, Mg, K), and P
-1
(all presented as cmolcharge kg soil) ........................................................................................ 167
Table A.3. Site median values for soil moisture and mean values for soil texture, with standard
deviation intervals where available ............................................................................................. 168
Table A.4. Soil resource correlation matrix, with correlation coefficients (r) for each pair-wise
comparison .................................................................................................................................. 169
Table A.5. AICc comparisons for significant or top non-significant models relating the fraction
of total ANPP comprising ANPPW to soil resources, species composition, and stand density . 170
Table B.2. Site occupancy by species ......................................................................................... 174
Table B.3. Soil resource availability by site ............................................................................... 175
Table B.4. Biological interpretation of neighborhood model parameters and model variations 176
Table B.5. Theoretical potential growth model selection and parameters .................................. 177
Table B.6. Summary statistics for USFS FIA plot data from northern Lower Michigan ........... 178
Table B.7. Model framework comparison: realized growth versus theoretical potential growth179
Table B.8. Parameter estimates and 95% confidence intervals for species whose growth was
either unrelated to neighborhood or related to a neighborhood that did not account for neighbor
species (terms (5) or (6)) ............................................................................................................. 180
Table B.9. Estimated model parameters for species-dependent neighborhoods ........................ 181
Table B.10. Mean growth (as basal area increment, cm2 year-1) for each species at each site;
standard deviation is in parentheses ............................................................................................ 182
Table B.11. Mean diameter (cm) for each species at each site; standard deviation is in
parentheses .................................................................................................................................. 183
ix
Table C.1. Site attributes, including mean soil Ca and inorganic N content, stem density,
dominant species, stand age, and location .................................................................................. 185
Table C.2. Soil resource content before and after fertilizer application, aggregated for all species
..................................................................................................................................................... 186
Table C.3. Soil pH before and after fertilizer addition, aggregated for all species .................... 187
Table C.4. Mean values per treatment for each species for chlorophyll fluorescence yield and
maximum .................................................................................................................................... 188
Table C.5. Leaf nutrient content of base cations and P .............................................................. 189
Table C.6. Content of N and C in leaves .................................................................................... 190
Table C.7. Percent canopy openness........................................................................................... 191
Table C.8. Diameter growth and BAI for each species, comparing treatments with and without
Ca, and with and without N ........................................................................................................ 192
-1
Table C.9.1. Parameters for simple linear regression of annual diameter growth (cm year ) as a
function of predictor variables in which the relationship between growth and the predictor
differed significantly in a treatment versus the control category ................................................ 193
Table C.9.2. Parameters for simple linear regression of annual basal area increment (BAI) (cm
2
-1
year ) as a function of predictor variables in which the relationship between growth and the
predictor differed significantly in a treatment versus the control category ................................ 194
-1
Table C.10.1. Parameters for multiple linear regression of diameter growth (cm year ) as a
function of initial diameter and another predictor variable for treatment categories with
significant slope parameters for both independent variables ...................................................... 195
Table C.10.2. Parameters for multiple linear regression of basal area increment (BAI) (cm
2
-1
year ) as a function of initial diameter and another predictor variable for treatment categories
with significant slope parameters for both independent variables .............................................. 196
Table D.1.1. Stand-level parameters, including mean diameter, diameter and basal area
increments with minimum and maximum values provided parenthetically, and density and
mortality rates ............................................................................................................................. 198
Table D.1.2. Stand-level mean resource measurements, with standard deviation ...................... 199
Table D.2. Model parameters, their biological interpretation, and the consequences of modifying
the full models by constraining or changing certain parameters ................................................ 200
x
2
Table D.3. Parameter estimates and R values for significant linear regressions of stand-level
2
-1
mean growth of all species (mean basal area increment (cm yr )) as a function of soil resources
..................................................................................................................................................... 201
Table D.4.1. Resource covariance for all species together and for functional groups ............... 202
Table D.4.2. Resource covariance for wood density groups ...................................................... 203
2
Table D.5. Parameter estimates and R values for linear regressions of stand-level mean growth
2
-1
(basal area increment (cm yr )) as a function of soil resources for functional groups ............ 204
Table D.6. Model parameters and 95% confidence intervals (parenthetically) for functional
groups .......................................................................................................................................... 205
2
2
-1
(basal area increment (cm yr )) as a function of soil resources for species with more than 50
focal individuals .......................................................................................................................... 206
Table D.8. Model parameters and 95% confidence intervals (parenthetically) for individual
species for which more than 50 focal individuals were available .............................................. 207
2
Table D.9.1. Parameter estimates and R values for linear regressions of stand-level mean
2
-1
growth (basal area increment (cm year )) as a function of soil resources for wood density
categories, excluding the Fabaceae ............................................................................................. 208
2
2
-1
categories, including the Fabaceae ............................................................................................. 209
2
Table D.9.3. Parameter estimates and R values for linear regressions of stand mean growth
2
-1
(basal area increment (cm year ) as a function of soil resources for wood density classes within
functional groups ........................................................................................................................ 210
Table D.10. Model parameters and 95% confidence intervals (parenthetically) for wood density
groups .......................................................................................................................................... 211
Table D.11. Model parameter mean estimates based on 101 sub-samples from the resource
posterior predictive distribution .................................................................................................. 212
Table D.12. Predicted growth across the range of resource estimates for functional groups, wood
density categories, and species for which growth was related to a soil resource ....................... 213
xi
Table D.13.1. Correlations (Pearson‘s r) between site-mean inorganic N and other resources for
all species and for the three single species—S. exorrhiza, I. deltoidea, and P. decurrens ......... 218
Table D.13.2. Sample sizes by site for each functional group .................................................... 218
Table D.14. Sample sizes by site for each wood density group ................................................. 219
Table D.15. Sample sizes by site for species with more than 50 focal individuals .................... 220
xii
LIST OF FIGURES
Figure 2.1. ANPPW expressed as functions of soil resources ...................................................... 40
Figure 2.2. ANPPL expressed as functions of soil resources........................................................ 41
Figure 2.3. Productivity partitioning to wood and leaves expressed as functions of ANPP and
several soil resources .................................................................................................................... 42
Figure 3.1. Normalized neighborhood index as a function of neighbor (a) displacement and (b)
diameter......................................................................................................................................... 78
Figure 3.2. Individual growth as linear (equation 3) or lognormal (equation 4) functions of
diameter......................................................................................................................................... 79
Figure 3.3. Predicted diameter growth as a function of soil resource, holding diameter constant at
the mean value for each species .................................................................................................... 80
Figure 3.4. Predicted diameter growth plotted as a function of soil water with neighborhood and
diameter held constant (a, b), and as a function of neighborhood with soil water and diameter
held constant (c, d) ........................................................................................................................ 81
Figure 3.5. Canopy diffuse non-interceptance as a function of each measured soil resources..... 82
Figure 3.6. Predicted diameter growth as a function of neighborhood ......................................... 83
Figure 4.1. Change in soil resource content aggregated across all species before (April 2008) and
after (October 2008) fertilizer application .................................................................................. 113
Figure 4.2. Diameter distributions for each species .................................................................... 114
Figure 5.1. Site mean basal area increment for all species as functions of soil resources .......... 148
Figure 5.2. Site mean basal area increment for six taxonomic groups as functions of soil
resources ..................................................................................................................................... 149
Figure 5.3. Site-mean basal area increment for individual species as a function of soil resources
..................................................................................................................................................... 152
Figure 5.4. Site mean stem area growth for wood density categories, excluding legume species,
plotted versus soil resources ....................................................................................................... 154
Figure D.12.2.i. Predicted individual diameter growth based on 101 estimated soil resource
datasets plotted against total soil P for the Arecaceae ................................................................ 214
xiii
Figure D.12.2.ii. Predicted diameter growth based on 101 estimated soil resource datasets plotted
as a function of total soil P for species with lowest wood density ............................................. 215
xiv
LIST OF ABBREVIATIONS
AIC: Akaike‘s Information Criterion, or AICc when corrected for small sample size
ANPP: Aboveground net primary productivity
ANPPL: Aboveground net primary productivity, leaf partition
ANPPW: Aboveground net primary productivity, wood partition
BAI: Basal area increment, annual
2+
Ca: Calcium, Ca
Dbh: Diameter at breast height (1.37m), alternately diam.
+
K: Potassium, K
2+
Mg: Magnesium, Mg
N: Nitrogen, either atmospheric or total plant-available soil inorganic pool
4+
NH : Ammonium
3-
NO : Nitrate
P: Phosphorus
PPD: Posterior predictive distribution
α: Exponent modifying diameter of neighbor trees
β: Exponent modifying displacement of neighbor trees
δ: Divisor of focal tree diameter
λ: Species-specific coefficient modifying separate neighbor trees
σ: Secondary divisor of focal tree diameter term
+
ΣN: Sum of inorganic N, or NH4 + NO3
-
xv
CHAPTER I
INTRODUCTION
1
Soil resource limitation of forest productivity
At a global scale, forest productivity is controlled by climate (Clark et al. 2010, Mund et al.
2010), soil type and resource availability (Vitousek et al. 2010), and species composition (Jacob
et al. 2010). Individual tree growth is ultimately limited by broad-scale factors like climate and
soils (Gradowski & Thomas 2008), but also by finer-scale processes, including competition
(Coates et al. 2009) and plant-soil feedbacks (McCarthy-Neumann & Kobe 2010). Temperate
forest production increases in response to supplemental nitrogen (N) (Will et al. 2001),
supporting the general consensus that temperate forests are N-limited. Similarly, tropical forest
productivity can increase in response to supplemental phosphorus (P) (Meason et al. 2009, Sayer
& Tanner 2010), leading to the conclusion that tropical forests are P-limited. Furthermore, Nlimitation of forest production may be globally distributed (LeBauer & Treseder 2008). Although
these general categorizations of resource limitation are often accurate (Finzi 2009), forest
ecosystems embody a diversity of species assemblages and varying degrees of soil resource
heterogeneity (Townsend et al. 2008) that complicate efforts to assess resource limitation.
Ecological and biogeochemical complexity of forest systems suggests that soil resource
limitation may not conform to simple dichotomous classifications (Vadeboncoeur 2010).
The expectation that a single soil resource limits production across large geographical
extents discounts local- to regional-scale variation in geological history, plant colonization
patterns, and extant species composition (Chesson 2003). Geologically recent soil development
in many temperate regions motivates the expectation that these soils contain low N levels, and N
limitation in these systems can be corroborated by N addition studies (LeBauer & Treseder
2008). By comparison, more ancient soil origins in many tropical regions support extensive loss
of soil P via leaching and immobilization, with P limitation in some tropical forests confirmed by
2
P addition studies (Vitousek et al. 2010). Nonetheless, several factors may change these expected
resource limitation patterns. Some temperate soils may exhibit P deficiencies, causing
productivity of forests growing on such soils to be limited by soil P availability (Gradowski &
Thomas 2006). Alternately, N limitation in temperate forests could be partially alleviated by
increased N use and resorption efficiency (Hidaka & Kitayama 2009) or by N absorption through
leaves (Vallano & Sparks 2008). In tropical systems, P limitation can be mitigated by
atmospheric deposition or as a consequence of parent material with relatively high P content
(Porder et al. 2006).
Although much research has focused on the potential for N or P limitation, many of the
same processes and conditions could contribute to limitation by base cations. As with P, nearly
all soil base cation content derives from parent material (Giesler et al. 1998, Blum et al. 2002,
Berger et al. 2006) and is thus inherently limited, with minimal contribution from atmospheric
deposition. Soil base cation availability is governed by the stoichiometry of weathering rates
(Blum et al. 2002), leaching and immobilization (Perakis et al. 2006, Singh & Agrawal 2008),
and cycling through plant tissues and microbial mass (McLaughlin 2009). The potential for base
cation limitation of forest productivity exists on soils that experience relatively low weathering
rates or slow nutrient cycling due to species composition (Fujinuma et al. 2005) or climate
(Hilley & Porder 2008, Yadav et al. 2008). Calcium (Ca) may be particularly important because
its ions are immobile in the phloem (Brown et al. 1999), which precludes resorption from
senescing leaves and fine roots (Park et al. 2008) thereby increasing demand. Additionally, base
cation limitation may be a consequence of repeated whole-tree harvest regimes (McLaughlin &
Phillips 2006).
3
Base cations fulfill physiological functions as fundamental and indispensable as either N
or P, supporting the idea that any of these soil nutrients may limit tree growth. As a structural
element, N is necessary for photosynthesis (chlorophyll and RuBisCO), proteins, and nucleic
acids (Masclaux-Daubresse et al. 2010). Phosphorus is a structural element, is required as
NADPH for electron transfer in photosynthesis, and is most important as a component of ATP
(Javot et al. 2007, Amtmann & Blatt 2009). Base cations are implicated in a variety of ecological
and physiological functions. Soil Ca is often correlated with stand- and individual-level growth,
and this relationship can be measured in forests that are leaching Ca over time (St Clair et al.
2008), sites distributed across Ca gradients (Bailey et al. 2004, Schaberg et al. 2006, Bigelow &
Canham 2007), or as a consequence of fertilizer addition (Kobe et al. 2002, Juice et al. 2006).
Indirectly, soil Ca can also influence production by stimulating nitrification (Hobbie et al. 2007,
Page & Mitchell 2008), N mineralization, and decomposition (Holzmueller et al. 2007,
Watmough 2008), with tree growth responding to elevated inorganic N availability. Productivity
is also related to availability of other base cations (Park et al. 2008, Rosberg et al. 2006,
Sverdrup et al. 2006, Wilmot et al. 1996). Higher magnesium (Mg) availability increases
photosynthesis (Vizcayno-Soto & Cote 2004), while greater soil potassium (K) availability
improves efficiency of water use (Bradbury & Malcolm 1977), nutrient uptake (Stevens et al.
1993), and leaf retention (Nowak et al. 1991, Hodson & Sangster 1998).
In numerous instances, fertilizer addition has shown that production in tropical forests
may not be limited by soil P (Hedin et al. 2009, Tripler et al. 2006), while production in
temperate forests may not be limited by inorganic N availability (Gradowski & Thomas 2006,
Wilmot et al. 1996). Moreover, many studies that have investigated resource limitation are
restricted to testing for effects of N versus P, neglecting to test whether other resources influence
4
growth (LeBauer & Treseder 2008). When many resources are tested (including N, P, and base
cations) for their effects on tree growth, each resource has been shown to increase the growth
response (Rosberg et al. 2006). Thus, a comprehensive first step to identifying the resource or
resources that limit plant productivity in an ecosystem is to assess the relationship of tree growth
to availability of a broad array of resources, not strictly N or P.
Competition and soil resource availability
Individual tree growth can be affected to similar degrees by soil resource limitation (Gradowski
& Thomas 2008) and by competition with neighboring trees (Canham et al. 2006). The
comparable magnitude of these effects necessitates measuring the influence on tree growth of
both soil resource availability and individual-level competition. Substantial evidence also
identifies changing intensity and importance of competition across soil resource gradients
(Craine 2005, Brooker & Callaway 2009); resource and competitive effects interact, so plant
growth must be compared to both factors. The simplest systems that demonstrate interaction
between resources and competition are monospecific plantations. Trees planted at high density
on soils with low resource availability grow more slowly than trees planted at lower density
(Erskine et al. 2005), while soils with higher resource availability can support faster growth rates
even at increased stem densities (Alves et al. 2010). Natural stands incorporate additional
complexity of species composition, which may shift significantly across resource gradients (Ares
& Fownes 2001, Pyke et al. 2001, Reich et al. 2010). Moreover, competitive interactions can
differ by species (Coates et al. 2009), meaning that tree growth is simultaneously influenced by
soil resource availability and local neighborhood configuration, including the distance between
individuals and neighbor species identity (Canham et al. 2006).
5
Relative effects of soil resource limitation and competition have been measured most
rigorously in plantations where stem density and spatial configuration are strictly controlled
(Binkley et al. 2010) or in herbaceous communities with rapid growth rates and easily
manipulated species composition (Wedin & Tilman 1993). Dual- and multi-species forest
plantations provide simplified models for studying both the spatial configuration of
neighborhoods and effects of neighbor species identity (Boyden et al. 2005); often, growth is
suppressed to a greater degree by intraspecific interactions than interspecific ones (Boyden et al.
2008), and displacement from competing neighbors is positively correlated with growth
(Canham et al. 2004). In contrast, scarce data exist to compare soil resource limitation to
competition in natural, uneven-aged forests comprising many species and spanning resource
gradients. Most studies concerning neighborhood effects (e.g. He & Duncan 2000, Uriarte et al.
2004) have not also explicitly tested effects of soil resource availability. In the few instances in
which tree growth has been modeled as a function of neighborhood and soil resources,
availability was weakly characterized (Canham et al. 2006) or not distinguished from
neighborhood effects (McPhee & Aarssen 2001, Coates et al. 2009). To appropriately measure
relative effects on tree growth of soil resource availability, competition, and neighbor species
identity, growth models must be informed by all three factors.
Across many lines of research concerning soil resource limitation of forest productivity,
resource availability is often measured at scales substantially coarser than individual trees or
local neighborhoods. Characterizing soil resource availability for individual trees in large
datasets becomes logistically challenging (Bigelow & Canham 2007, Finzi 2009), but high
resource heterogeneity at meter to tree scales (Reynolds et al. 2007, Townsend et al. 2008) may
necessitate greater sampling densities. If individual growth rates are strongly related to soil
6
resources within the diameter of the root system, measuring resources at coarser scales, as often
occurs (Lundholm 2009), would fail to identify the relationship. Similarly, if species colonize or
recruit into sites based on soil resource availability (Snyder & Chesson 2004), the colonization or
recruitment patterns may only be identifiable if resources have been measured at sufficiently fine
spatial scales. On the other hand, extensive root systems (Coll et al. 2008, Hertel et al. 2009) or
common mycorrhizal networks (Simard 2009) may effectively homogenize soil resources across
scales that include many individuals. It is ultimately important to measure soil resource
availability at scales commensurate with individual trees and their local neighborhoods, since the
scale at which resources relate to tree growth is poorly established.
Dissertation objectives
The purpose of this dissertation is to reevaluate potential soil resource limitation and competitive
effects across soil resource gradients in both northern temperate hardwood and lowland wet
tropical forests. Although N and / or P availability is typically implicated in limiting
productivity, other soil resources perform similarly vital physiological functions and are equally
likely to be in short supply. This research characterizes relationships of stand- and individuallevel tree growth to an array of potentially limiting soil resources, including N, P and base
cations. Any effects of soil resources on individual tree growth manifest within a neighborhood
comprising surrounding trees and their influence on availability of space, light, and soil
resources. The other principal facet of this dissertation research, therefore, is determining the
influence of neighborhood interactions on tree growth.
7
Objective i (Chapters II, III, V): Assess correlations and functional relationships of soil
resources, including N, P, Ca, K, Mg, and volumetric water, to stand- and individual-level
annual growth rates. Although it is expected that N will be correlated with growth in temperate
forests and P will be correlated with growth in tropical forests, it is hypothesized that growth will
also be related to soil base cation availability. Strong positive relationships of growth to a soil
resource are consistent with, but do not conclusively demonstrate limitation (see Objective ii).
Objective ii (Chapter IV): Test resource limitation in four common temperate tree species
growing across a gradient of multiple soil resources. It is hypothesized that both soil N and base
cation availability will be correlated with tree growth (see Objective i), such that addition of
either or both nutrients will increase annual basal area increment, photosynthetic rate, and / or
leaf nutrient content. Comparing the relative influence of N versus base cations is necessary to
establish whether soil N, as has been claimed by past research, truly limits production in northern
hardwood forests.
Objective iii (Chapters III, V): Measure effects of local neighborhood on individual tree
diameter growth, and compare relative influence of neighborhood versus resource effects across
soil resource gradients. Competition among individual trees is expected to exhibit a size- and
distance-dependent nature, with growth suppressed more by intraspecific than interspecific
interactions. It is hypothesized that soil resource availability influences the importance of
competition, potentially affecting individual-level interactions or emerging as a site-level effect.
8
CHAPTER II
SOIL CALCIUM, NITROGEN, AND WATER ARE CORRELATED WITH
ABOVEGROUND NET PRIMARY PRODUCTION IN NORTHERN HARDWOOD
FORESTS
Baribault, T.W., Kobe, R.K., Rothstein, D.E. 2010. 2010. Soil calcium, nitrogen, and water are
correlated with aboveground net primary production in northern hardwood forests. For. Ecol. and
Manage. 260, 723-733.
9
Abstract
Multiple soil resources could limit aboveground net primary production (ANPP) in temperate
forests, but most studies have focused on nitrogen (N) availability. We tested a suite of resources
including water, multiple measures of N, phosphate (PO4), and extractable ions (calcium (Ca),
potassium (K), magnesium (Mg)) as correlates of ANPP across a glacial landform gradient in
northwestern Lower Michigan. Our goals were to identify resource correlates of ANPP that
could potentially limit productivity, to characterize productivity partitioning to leaves (ANPP L)
and wood (ANPPW) across the gradient, and to test soil resource correlates of partitioning. We
measured soil resource availability and annual ANPP at 13 sites, and fit models of ANPPW and
ANPPL as functions of each resource. We used Akaike‘s Information Criterion to assess
empirical support for models with different functional forms. ANPPW was best correlated with
Ca, with a relationship that was asymptotic beyond the four lowest-productivity sites. Both Mg
and K covaried with Ca; Mg was also supported as an ANPPW correlate. Soil water and sum of
ammonium (NH4) and nitrate (NO3) (ΣN) explained substantial ANPPW variance, but with
weaker support than for Ca. ANPPL was best correlated with ΣN; NO3 received support, while
Ca received less support but still explained substantial ANPPL variance. These results suggest a
potential role for Ca in regulating ANPP, at least in low-productivity sites, while reaffirming the
importance of both N and water availability. Partitioning of annual production to ANPPW as a
fraction of total ANPP was related to Ca and soil water, but only weakly to ΣN. Production was
partitioned equally between ANPPL and ANPPW at low fertility, but ANPPW exceeded ANPPL
10
by 24% at maximum productivity. Given this shift in partitioning, measuring only ANPPW but
assuming that ANPPL scales proportionately would overestimate ANPP at higher fertility sites.
Although correlative, our results suggest that ANPP is constrained by some combination of N,
Ca, and water, and provide motivation for experimental manipulations of these resources to
better understand forest responses to human influences. If northern temperate forests are strictly
N-limited, as suggested by much of the literature, then anthropogenic N deposition should
enhance ANPP. In contrast, if ANPP is limited, or co-limited, by Ca or by soil water, then
anthropogenic change (altered precipitation, base cation leaching), could counteract production
gains from N deposition.
KEYWORDS: Forest productivity, resource correlation, productivity partitioning, calcium,
nitrogen, water
11
Introduction
Aboveground net primary productivity (ANPP) in temperate forests is often positively correlated
to soil nitrogen (N) availability (Pastor et al. 1984, Zak et al. 1989, Goodale & Aber 2001, Idol
et al. 2003, Hogberg et al. 2006, Joshi et al. 2003), and supplemental N can increase ANPP (Will
et al. 2001, Finzi 2009). Additional N from atmospheric deposition, in combination with
increased global atmospheric CO2 concentrations and temperatures, is projected to enhance
ANPP (Mohan et al. 2007). Nitrogen deposition may alter nutrient regimes in many ecosystems
by alleviating N limitation (Gradowski & Thomas 2008); however, N deposition could also
promote leaching of calcium (Ca) and other base cations (Izuta et al. 2004, Perakis et al. 2006),
and paradoxically of N itself (Brookshire et al. 2007). Since N limitation in temperate forests is
broadly assumed (Finzi & Canham 2000), tests of resource limitation often include only
correlations of production to N, even when other resource measures (e.g. soil texture) are
available (Pastor et al. 1984, Zak et al. 1989). In theory, productivity could be reduced by
deficiency of any resource, so it remains unknown whether N limits ANPP more than other
resources because other resources have rarely been evaluated.
Nitrogen cycling and productivity often co-vary (Hall et al. 2006, Johnson 2006, Mayer
2008), but production may influence N cycling as much as N cycling influences production. For
example, across an elevation–atmospheric N deposition gradient in the northeastern U.S., N
mineralization is strongly correlated with ANPP but ANPP is unrelated to mineral N availability,
which is controlled by both mineralization and atmospheric deposition (Joshi et al. 2003). In
Patagonian steppe systems that span a precipitation gradient, ANPP is related to soil moisture
and N, but N availability is controlled by soil moisture (Austin & Sala 2002). Furthermore, N
resorption from fine roots (Kunkle et al. 2009) and senescent leaves (Kobe et al. 2005), and
12
direct N absorption by leaves (Sparks et al. 2008), could partially decouple N metabolism from
soil N availability (Migita et al. 2007). In general, leaf production (ANPPL) may regulate soil N
if litter quality exerts strong control over decomposition, N mineralization, and nitrification rates
(Page & Mitchell 2008, Vesterdal et al. 2008). Recognizing that N availability is regulated by a
cycle of N-vegetation feedback, landscape gradients in soil N could be a function of—in addition
to being a reason for—variation in vegetation cover and ANPP (Hobbie 1996, Aerts & Chapin
2000).
Emerging evidence highlights the influence of Ca availability on productivity of
deciduous forests in northeastern North America (Bigelow & Canham 2007, Gradowski &
Thomas 2008, Park et al. 2008). Soil Ca availability is less likely than N to be a consequence of
production because Ca is weathered from parent material (Berger et al. 2006, Blum et al. 2002,
Giesler et al. 1998), but it is influenced, at least in late succession ecosystems, by mineralization
and decomposition rates (Holzmueller et al. 2007, Watmough 2008). Diminished soil Ca reduces
growth and disease resistance in Acer saccharum on acidic soils (St Clair et al. 2008).
Furthermore, Ca availability is strongly related to individual tree growth across natural (Bailey et
al. 2004, Schaberg et al. 2006, Sverdrup et al. 2006) and experimentally manipulated Ca
gradients (Bigelow & Canham 2007, Dauer et al. 2007, Kobe et al. 2002, Kulmatiski et al.
2007), and correlates with fine root production (Park et al. 2008, Reich et al. 1997). Fertilization
with Ca in the field can elicit positive growth responses in seedlings (Kobe et al. 2002), saplings
(Long et al. 1997, St Clair & Lynch 2005, Juice et al. 2006), and canopy adults (Gradowski &
Thomas 2008). Conversely, limited Ca availability contributes to low ANPP (Hallett &
Hornbeck 1997, McLaughlin & Wimmer 1999), while reduced productivity accompanies
leaching of soil Ca by acidic rain (Singh & Agrawal 2008). Moreover, Ca is immobile in the
13
phloem (Brown et al. 1999), so Ca uptake requirements likely exceed those of phloem-mobile
elements like K, Mg, P, or N.
Production also could be related to availability of other base cations (Park et al. 2008,
Rosberg et al. 2006, Sverdrup et al. 2006, Wilmot et al. 1996). Greater availability of Mg
increases photosynthesis (Vizcayno-Soto & Cote 2004), which is the primary physiological basis
for elevating ANPP. Increased K availability improves water use efficiency (Bradbury &
Malcolm 1977), nutrient uptake (Stevens et al. 1993), leaf retention (Nowak et al. 1991), and
leaf longevity (Hodson & Sangster 1998), while decreasing leaf mortality and senescence
(Covelo & Gallardo 2002, Ericsson & Kahr 1993). Higher K availability also is associated with
increased resistance to beetle infestation (Ylimartimo 1990), reduced tissue necrosis (Fluckiger
& Braun 1998), and reduced lateral root mass (Triboulot et al. 1997). Potassium and Mg are
readily re-translocated via phloem (Brown et al. 1999), and resorption rates for both elements are
high, further highlighting physiological importance. Soil texture can provide an integrated
measure of base cation availability (Knox et al. 1995), and predict wood production (ANPPW)
approximately as well as N (Pastor et al. 1984, Zak et al. 1989, Reich et al. 1997); soil moisture
often constrains production (Knox et al. 1995), so we included both volumetric soil water and
soil texture in our set of measured soil resources.
Species composition could further affect productivity, since some species (e.g. Acer
saccharum, Quercus rubra) grow faster than others (e.g. Quercus alba, Fagus grandifolia) under
equivalent resource conditions (Comas & Eissenstat 2004). As well as mediating the growth
response to resource availability, species composition could actively deplete or augment various
resources, whether through differences in resource uptake (Fujinuma et al. 2005) or litter cycling
rates (Dijkstra 2003).
14
In addition to total production, partitioning to leaf, stem, and root production would be
expected to vary across resource gradients (Vanninen & Makela 2000, Sternberg & Shoshany
2001, Vogel et al. 2008, Kobe et al. 2010). In many studies, productivity partitioning to stem
growth in seedlings responds idiosyncratically to increased soil N or base cation availability
(Berger & Glatzel 2001, Kobe et al. 2010), but partitioning to stem mass is consistently a
function of size for both seedlings and adults (Peichl & Arain 2007). It is unclear whether
partitioning of annual aboveground growth (as opposed to aboveground total mass) at the adult
stage is controlled primarily by tree size, nutrient availability, or an interaction of these factors.
We assume that growth and mass behave similarly, on the grounds that mass represents
accumulated growth, such that partitioning to annual stem growth increases with tree size.
Objectives
Our overall objective was to determine the strongest relationships of production to soil resources
in order to identify potential limiting resource candidates. We measured tree growth over an
eight-year interval at 13 sites that spanned a soil fertility and productivity gradient. A
comprehensive set of soil variables including N, base cations, texture, and moisture was tested
for correlation to ANPP. We expected a relationship of production—particularly ANPPL—to N,
but also suspected that production might be associated with base cation or water availability. Our
second objective was to test whether partitioning of production across the gradient was related to
soil resource availability. Based on allometric relationships, we expected wood production to
constitute a progressively larger fraction of total production as productivity increased; we also
expected that partitioning of production should be related to the soil resources identified from the
first objective.
15
Methods
Study sites
This study was conducted in mixed hardwood stands of the Manistee National Forest in the
lower peninsula of Michigan, USA (~44º12‘N, ~85º45‘W), a region with well-defined
associations among glacial landforms, soil fertility and forest community composition (Zak et al.
1989, Host & Pregitzer 1992). Net N-mineralization, nitrification rates, nitrate, and ammonium
pools increase from outwash through ice-contact and moraine landforms (Zak et al. 1989) (Table
A.1). Base cation concentrations also increase across the landform fertility gradient, although
phosphate follows no discernible pattern (Table A.2). From outwash through moraine, total
ANPP, ANPPW, and ANPPL increase approximately two-fold (Table 1). Soil properties (Table
A.3) that characterize landform are closely associated with characteristic species compositions
(Host & Pregitzer 1992, Host et al. 1988). Stands were 80 to 100 years old second-growth, and
have been undisturbed since the 1920‘s. Species composition shifts from dominance of Q. alba
and Quercus velutina on outwash plains through Q. rubra on intermediate sites, to A. saccharum
2
on moraines (Table 1). All sites were within an area of 960 km (maximum distance between
sites of 38.1 km, average distance 15.76 km, maximum elevation change 230 m) and are
expected to have very similar climate. Mapped stands were established, wherever possible, in
transects of dimensions 240m x 41m. Most sites conformed to this configuration, but
topographic variation, roads, or edaphic transitions at four sites required non-standard plot
dimensions. Two sites (12, 13) were adjacent halves of one standard plot, separated due to
different resource availability and species assemblage. Site (4) differed in shape (two 120m x
41m rectangles with ends overlapping at a 63° angle) and encompassed a slightly larger area.
16
Site (8) had a reduced area (40m x 21m) for tree diameter measurements. Leaf traps (n = 10, 1m
x 1m) were placed at 20 m intervals along a central transect within every mapped stand.
Productivity measurements
Stems of all mapped trees were measured using diameter tapes in 1999, 2005, and 2007. Growth
increment was calculated as the change in diameter measured at 1.37 m height over the eightyear interval from September 1999 to August 2007 for trees ≥ 10 cm diameter. Stem increment
for in-growth was calculated as the difference between measured diameter and the 10-cm
threshold. An intermediate census in 2005 enabled calculation of growth increment over the
1999-2005 interval for trees that died between 2005 and 2007. Of 6990 individuals, 57 (0.8%)
had diameter growth ≤ –1 mm during the eight-year interval, and were included in the
calculation of site-level productivity to minimize upward bias of ANPP due to measurement
error. Trees that grew after 1999 but died before 2005 or 2007 were included. Trees that died
during the interval but did not grow, or that lost diameter (due to bark loss or localized wood
-1
-1
decomposition), were excluded. Annual wood mass increment (ANPPW, Mg ha year ) (Table
1) was estimated using species-specific allometric biomass equations available in TerMikaelian
and Korzukhin (1997). Several allometric equations were available for each species, so where
possible we used equations developed for Lake States trees. For F. grandifolia and Tsuga
canadensis, which are relatively uncommon at these sites, it was necessary to use equations
developed for New Jersey trees. Equations developed for West Virginia were applied to Q. alba,
and for northern Kentucky to Q. velutina.
Leaf litter was collected from the 2006 growing season. All leaves had fallen at sites
dominated by A. saccharum or Q. rubra (sites 3 – 13) by early November 2006, and were
completely collected by 12.Nov.2006. Approximately 10% of leaves remained attached through
17
winter at sites dominated by Q. alba (sites 1 and 2), so leaf litter was collected on 12.Nov.2006
and again on 2.March.2007 for these sites. Collected litter was oven-dried at 70°C for ≥ 3 days
before weighing (following Pastor et al., 1984). Litter production for each site (ANPPL) was
-1
-1
calculated as mean leaf mass from the 10 litter traps scaled to Mg ha year (Table 1).
Resource measurements
Soil Moisture
Soil volumetric water was measured to 30 cm depth by time domain reflectometry
(Environmental Sensors Inc., Sidney, BC). Measurements were made in August 1998, April,
May, and September 1999, and August and October 2000 at 10 m intervals along a 200m central
transect at each site. Given the relatively small number of sampling dates, we used median soil
moisture for analysis at each site in order to avoid potential effects of extreme values that might
disproportionately influence mean soil moisture. Data for all sites were collected during rainless
periods over an interval of not more than two days per each sampling date.
Soil texture
Soil samples were collected in October 2005 at three points per site (40, 100, and 160m along
each central transect), and for three composite strata (0-50cm, 50-100cm, and 100-150cm).
Samples were dried for 24 hours at 105°C, sieved (2mm mesh), and 100g (±0.01g) subsamples
were suspended in 5% sodium hexametaphosphate to a final volume of 1L. Following Gee and
Bauder (1986), sedimentation of sand was measured by hydrometer after 40 seconds and of silt
after 2 hours; clay remained in suspension. Percentages of sand, silt, and clay were calculated for
each sample. We also used a composite sample of the three 100-150cm samples for each site to
measure the particle size profile. Analysis was conducted with a Mastersizer 2000E laser particle
size analyzer (Malvern Instruments Ltd., Worcestershire, UK) on chemically dispersed samples.
18
To correct for the consistent underestimation of the clay sized (<2 µm) fraction by these
instruments, the clay/silt division separating standard texture classes was set to 6 µm, which
most closely mimics particle size data using sieve-pipette methods, based on prior data (R.
Schaetzl, pers. comm.).
Soil nutrients
All soil nutrient analysis was conducted on cores (5 cm diameter, 10 cm depth) collected in
December 1999. Cores included the portion of the organic horizon in which tree roots were
observed, but excluded leaf litter. Samples were gathered at 10 m intervals along the central
transect at each site, with sets of 4 consecutive cores pooled into 5 composite soil samples per
site. Samples were transported to the lab in a cooler, sieved (2 mm mesh), and refrigerated for
two days prior to analysis.
Potential net N mineralization and nitrification rates were determined by aerobic
+
-
incubation. Initial concentrations of NH4 and NO3 were measured in a 10-g sub sample of each
+
composite. Soils were shaken for 30 minutes in 2 M KCl to extract inorganic N, and then [NH4 ]
-
and [NO3 ] were measured with an Alpkem Series 500 autoanalyzer (OI Analytical Alpkem,
College Station, TX). A second set of 10-g sub samples was incubated aerobically in darkness at
25°C for 28 days. Constant moisture content was maintained by periodic addition of deionized
+
-
water, and post-incubation concentrations of NH4 and NO3 were measured with the
autoanalyzer. Potential N mineralization was calculated as the difference between final and
+
-
initial ΣN [NH4 + NO3 ], and potential net nitrification was calculated as the difference
19
-
between final and initial [NO3 ]. Soils were sampled during an exceptionally warm December,
so microbial activity may not have attenuated for the season.
2+
+
-
Magnesium, Ca , K and phosphate (PO4 ) were extracted from 5-g sub samples with a
2+
2+
+
Mehlich III solution (Carter 1993). Concentrations of Mg , Ca , and K in extracts were
measured with Direct Current Plasma (DCP, SMI Corp., Glouchestershire, UK) atomic emission
-
spectrometry, and the Alpkem Series 500 autoanalyzer was used to measure PO4 . Total soil
carbon (C) and N content were measured from 5-g sub samples that were ground into powder by
a ball mill. Duplicate 25-mg secondary sub samples were compared against atropine standard by
combustion in a C-N analyzer (Carlo Erba NA 1500 Series 2, Lakewood, NJ). Extractable
+
3+
acidity ([H3O + Al ]) was measured by extraction in 1M KCl followed by titration against
NaOH.
Statistical analysis
Using the nonlinear routine in SYSTAT (v. 12, SPSS Corp., Chicago, IL USA), we fit models of
ANPPW and ANPPL as functions (Table 2) of single soil resources (Table 3, 4). A SYSTAT loss
function was defined as the negative natural log likelihood of a Gaussian probability density
function. For each resource, we evaluated four functional forms for the relationship between
ANPP and that resource. In the simplest case, ANPP was modeled as a linear function of
resource availability (Table 2), assuming that ANPP would increase proportionately with
resources across the measured domain. More complex models, including a Michaelis-Menten
function, a two-parameter sigmoid function, and a three-parameter sigmoid function, represented
scenarios in which ANPP could reach an upper bound within the measured resource domain
20
(Table 2). Production was negatively related to soil acidity and PO4, but these relationships were
so weak that it was unnecessary to test any further functional forms (e.g. exponential decay). We
also tested several paired resources using multiple linear regressions and a double MichaelisMenten function (Kobe 2006). None of these models achieved significant parameter estimates,
and many possible resource pairs were not justifiable due to multi-collinearity, so we report no
further on these. Models were compared using Akaike‘s Information Criteria (AIC) corrected for
small sample size (AICc) (Burnham & Anderson 2002). Models within two AICc units of the
minimum AICc are considered to have strong empirical support. We calculated coefficients of
2
determination (r ) as the square of the correlation coefficient between observed and predicted
productivity. To further explain variance in productivity, we fit monotonic (linear or MichaelisMenten) models of every soil resource to the residuals of each best-supported model for ANPPW
and ANPPL. To investigate potential stand history effects on productivity, we fit linear models of
stand age and mortality rate to the residuals of the three best models. Species composition effects
were explored using linear models comparing production fractions or basal area by species to
ANPP and resource availability. We quantified species composition by calculating the basal area
contributed by each species, and compared relationships of model residuals to each species‘
proportional basal area. Production partitioning patterns were characterized in two ways. First,
we compared the slope values of linear models of each partition as a function of total ANPP to
measure any change in partitioning with increasing ANPP. Second, using either log-normal or
linear models, we tested whether the fraction of total ANPP represented by ANPPW was related
to soil resources, species composition, and stand density.
21
Results
Summary of productivity and resource availability
ANPP increased nearly twofold from 5.4 Mg ha
-1
-1
-1
year at the least productive outwash site to
-1
9.2 Mg ha year at the most productive moraine site (Table 1). Production was lowest at
-1
-1
outwash sites and increased through ice-contact sites (5.4 - 7.0 Mg ha year ), while mesic icecontact through moraines, which constituted a majority of sites, had similar production (7.6 - 9.2
-1
-1
Mg ha year ) (Table 1). Production of different tissues followed similar rankings, with
-1
-1
-1
-1
maximum ANPPW (5.3 Mg ha year ) and ANPPL (4.1 Mg ha year ) measured on
-1
-1
-1
-1
moraines, whereas minimum ANPPW (2.6 Mg ha year ) and ANPPL (2.8 Mg ha year )
occurred at the outwash sites (Table 1). Productivity reached its maximum at high fertility, but
stem density peaked at intermediate fertility (Table 1). Soil resource availability increased along
the gradient from outwash through ice-contact, intermediate moraine, and moraine sites, with
some covariance among a subset of resources (Table A.4). Notably, base cations were correlated
with NO3, NH4, and nitrification rate, and covariance among the base cations themselves was
extensive (Table A.4).
ANPPW increased with availability of base cations, nitrogen, and water in soil
Among all measured resources, extractable Ca was the best correlate of ANPPW (ΔAICc = 0),
2
and explained much of the observed ANPPW variance (r = 0.72) (Table 3, Fig. 1a, b). There
-1
was a sharp productivity increase associated with Ca to 0.45 cmolcharge kg , while above that
threshold ANPPW approached an asymptote of 4.8 Mg ha
22
-1
-1
year with some minor scattering
2+
around that value (Fig. 1a). ANPPW was also related to Mg
2
(ΔAICc = 1.0, r = 0.70), but there
was strong covariance of Mg and Ca (Table 3, Table A.4). Soil moisture and ΣN were weakly
supported as ANPPW correlates (Table 3), but the positive relationships of ANPPW to soil
moisture (Fig. 1c) and N (Fig. 1d) nevertheless suggest a potential role for these resources. Loss
2
of standing mass due to tree mortality increased with ANPPW (r = 0.40) if site 4, with atypically
high density and mortality of mature Populus grandidentata (an early succession species), was
excluded from analysis. No residual variance from any model was related to stand age or
measured soil resources.
ANPPL increased with availability of N in soil
2
2
For ANPPL, ΣN (ΔAICc = 0, r = 0.61) was the best correlate, while NO3 (ΔAICc = 1.2, r =
0.51) was also a supported correlate (Table 4, Fig. 2a, b). Somewhat weaker support was found
2
2
for Ca (ΔAICc = 2.2, r = 0.42) and nitrification rate (ΔAICc = 3.7, r = 0.36) (Table 4), but there
was considerable scatter in these relationships (Fig. 2c, d). No soil resource was correlated with
residuals from the ΣN, NO3, or Ca models, nor was leaf production significantly related to stand
density, age, or mass lost through mortality.
Species composition and stand density affect productivity
Soil resources (Ca, ΣN, water) explain up to 72% of ANPPW variance and 61% of ANPPL
variance, leaving between 28% and 39% of productivity variance unexplained. Aside from soil
resources, perhaps the most striking difference among sites is species composition, so we tested
whether species proportional basal areas were correlated with residuals from production-resource
relationships. Residuals of the ANPPW-Ca model were unrelated to any species‘ basal area. In
23
contrast, the residuals of the ANPPL-ΣN model were negatively related to the basal area of Q.
2
2
rubra (r = 0.54) and Acer rubrum (r = 0.46). Stem density was positively correlated with
2
residuals from the ANPPW-Ca model (r = 0.29). Post hoc, we measured soil particle size
distributions (% fine sand, Table A.3), but this measure of soil texture was not correlated with
productivity or stand density.
With the goal of maximizing explained variance, we tested combinations of resourceproduction models and their residuals as functions of stand density and species basal areas,
where independent variables were not strong co-variates. Residuals of the ANPPW-Ca model
2
were related to stand density (r = 0.29), and Ca and density together explained 80.3% ANPPW
variance. For ANPPL, maximum variance (81.7%) was explained via a combination of the ΣN
model and its residuals as a function of Q. rubra basal area. Species composition itself, but not
stand density, was related to availability of some resources. Relative basal area of A. saccharum
2
2
(r = 0.73) and Tilia americana (r = 0.89) were strongly related to Ca availability. Nitrate was
2
+
related to A. saccharum basal area (r = 0.78) but not to any other species; NH4 was unrelated
to basal area of any species.
ANPPW increased disproportionately with total ANPP and soil resources
Wood and leaf production were roughly equal at both outwash sites and at one ice contact site,
whereas at maximum total ANPP, ANPPW exceeded ANPPL by 22% (Fig. 3a, Table 1). Both
productivity fractions increased as a function of total ANPP, but the proportion of ANPPW
2
2
increased faster (r = 0.95, slope = 0.71, 95% CI: (0.60, 0.82)) than the proportion of ANPPL (r
24
= 0.75, slope = 0.29, 95% CI: (0.18, 0.40)). As a fraction of total ANPP, ANPPW was most
2
strongly related to Ca (ΔAICc = 0, r = 0.59), was related with some support to soil water
2
2
(ΔAICc = 2.3, r = 0.51), and was weakly related to ΣN (ΔAICc = 9.4, r = 0.02) (Table A.5).
The relationship of ANPPW partition to both Ca and soil water was best described by a lognormal function because the fractional value reached a maximum within the observed domain of
ANPP (Fig. 3b, c). In contrast, only a monotonic function could justifiably characterize the
relationship of ANPPW partition to ΣN, and the linear model was most parsimonious (Fig. 3d).
Productivity partitioning was negatively related to aggregate relative basal area of species other
2
2
than Q. rubra (ΔAICc = 3.3, r = 0.39), positively related to stand density (ΔAICc = 3.6, r =
0.37), and not significantly related to mean tree size, but none of these variables provided a
better correlation than Ca or soil water (Table A.5). Resources, tree size, and species
composition were confounded, so we could not use our data to rigorously evaluate whether
productivity partitioning to leaves versus wood resulted simply from site differences in tree size
and species composition, or instead from resource effects.
25
Discussion
Extractable Ca correlated with ANPPW better than any other measured soil resource, supporting
our expectation that base cations should be productivity correlates. Production at low-fertility
sites was chiefly responsible for driving the relationship between ANPPW and Ca, however,
while production at higher-fertility sites was relatively invariant and unrelated to any measured
resource. In addition, ANPPW was positively related to soil moisture and to ΣN, each accounting
for a substantial amount of variance despite receiving weak support. This result was also
expected based on evidence in the literature supporting water and N as production correlates. For
ANPPL, ΣN was a better correlate than any other resource, supporting the expectation that N
should be particularly important for leaf production due to its central role in photosynthesis. Had
we tested only N and not a broad suite of soil resources, we would have found significant
relationships between N availability and both components of ANPP, but disregarded a stronger
relationship of ANPPW to Ca and a reasonably strong relationship of ANPPL to Ca. We
achieved our primary objective by identifying three soil resources that are the most likely
candidates for limiting productivity. Our second objective was partially achieved since we found
the expected pattern that ANPPW increased more rapidly than ANPPL as a function of ANPP.
This result has important implications for the use of wood production, or standing biomass, as an
approximation for total production.
Overall, our results suggest that it will be important to experimentally test whether Ca, N,
soil water, or some combination of resources is responsible for limiting production in these
forests. The lack of resource manipulations in our study, a characteristic shared by many studies
26
examining ANPP-soil resource relationships (Zak et al. 1989, Reich et al. 1997), precluded
distinguishing among Ca limitation, N, or water limitation, or the possibility of sequential
limitation, where Ca limits ANPPW at the four least productive sites but N or water limit
-1
production once a threshold of 0.5 cmolcharge Ca kg soil has been exceeded. The ANPPW-ΣN
and ANPPW-H2O models over-predicted ANPPW at the outwash sites (1 and 2) and one icecontact site (3) (plots not shown), suggesting either that the relationship between production and
water or N is strongly non-linear at low resource levels or that other factors, such as Ca
availability, may constrain production. Across sites with higher productivity, ANPPW was
roughly constant across an order of magnitude change in Ca availability (Fig. 1a, 3a), suggesting
that production at these sites may be controlled by water and N, both of which have been
identified as correlates of ANPP or tree growth at the same sites by previous studies (Zak et al.
1989, Kobe 2006).
ANPPW increased with availability of base cations, nitrogen, and water in soil
Wood production was best correlated with Ca, with Mg the next-best correlate, while soil water
and ΣN were weak correlates (Table 3). Availability of Ca at our lowest-fertility sites (Table
-1
A.2) was comparable to levels at Hubbard Brook, New Hampshire (0.1 cmolcharge kg ) (Nezat
et al. 2008), where Ca limitation is a concern. However, a direct comparison of production at
Hubbard Brook to production at our low-fertility sites was complicated because species
assemblages were entirely different. Calcium could stoichiometrically limit ANPPW if there
were insufficient Ca to incorporate into wood tissues. Direct limitation could also occur if
insufficient Ca availability reduced efficiency of photosynthesis, stomatal regulation, phloem
27
transfer of nutrients, or some other physiological process integral to growth. Although
correlative, the strong relationship between ANPPW and Ca would be consistent with direct
limitation of ANPPW by Ca across the four sites with lowest production. For sites where
-1
-1
-1
ANPPW exceeded 4 Mg ha year or where Ca exceeded ~0.5 cmolcharge kg , production was
essentially unrelated to Ca. At the higher-productivity sites, it is conceivable that Ca could
indirectly influence ANPP through stimulating nitrification (Hobbie et al. 2007, Page & Mitchell
2008), but we found no evidence that wood growth at those sites was related to increased NO3 or
ΣN. In fact, there is minimal variance in site level growth for these sites, and none of the soil
resources we tested were significant correlates of production for this subset.
ANPPW was also related to Mg, following a very similar relationship as Ca (i.e., rapid
approach to an asymptote). There is little physiological justification for stoichiometric limitation
of ANPPW by Mg, since wood tissue contains substantially less Mg than Ca (Fujinuma et al.
2005). If Mg availability limits ANPPW, the effect could be mediated by photosynthesis, and
would be restricted to the four lowest-productivity sites across which ANPP changes with
increasing Mg.
Higher soil water availability could increase Ca weathering (McLaughlin 2009) and
promote ANPPW, resulting in a spurious correlation between ANPPW and Ca. Sites with
rd
th
th
relatively low water availability (3 and 4 lowest), however, had relatively high ANPPW (7
rd
and 3 highest) (Fig. 1c). We offer the caveat that at all sites, TDR measurements of soil
moisture to 30cm depth and analysis of texture to 1.5m depth (Table 3, 4; Table A.3) did not
28
account for the possibility of deeper fine textured bands that retain water or nutrients (McFadden
et al. 1994).
ANPPL increased with availability of nitrogen in soil
Leaf production was best correlated with ΣN, in agreement with results from many forest
systems (Will et al. 2001, Xia & Wan 2008, Iversen & Norby 2008). In contrast to the
asymptotic relationship of ANPPW to Ca, the relationship of ANPPL to increasing ΣN was
dynamic across the entire environmental gradient that we sampled (Fig. 2a). If measured during
the growing season, pools of available N represent the difference between microbial Nmineralization activity and plant N uptake, not total mineral N availability. We measured N from
soils in December 1999 during an atypically warm autumn (air temperature on the sampling day
was ~16ºC). Soil microbes were likely still active, so ΣN probably represents the full capacity of
soil to convert organic matter to mineral N. While the data could support direct N limitation of
ANPPL, the possibility remains that soil N was a consequence of ANPPL, with higher N
availability at sites dominated by species that produce rapidly-decomposing litter. To determine
whether ANPPL was the cause or effect of soil N availability would require N fertilization and
measurement of microbial activity in the field. Species composition and stand development
could also influence the relationship between ANPPL and soil N.
ANPPW increased disproportionately with total ANPP
The proportion of ANPP represented by ANPPW becomes progressively larger across the
gradient (Fig. 3a), with ANPPW varying more than two-fold across sites whereas ANPPL varied
only 1.5-fold. Some studies (e.g. Zak et al. 1989) use ANPPW as a proxy for ANPP, but this
29
substitution would be problematic for systems—like those studied here—in which ANPPL does
not increase in proportion to ANPPW. Greater growth partitioning to wood with increasing
ANPP could be influenced by at least three factors: resource availability (Wikstrom & Ericsson
1995, Wu et al. 2008), species differences in productivity partitioning (TerMikaelian &
Korzukhin 1997, Jenkins 2004), and ―allometric effects‖ (Weiner 2004) where larger trees
partition more production to wood. Partitioning to wood production was related to Ca and soil
water (Fig. 3b, c), but only weakly to ΣN (Fig. 3d). The relationships of partitioning to Ca and
soil water were not monotonic, but were better supported than correlations between partitioning
and species composition, stand density, or tree size (Table A.5). Based on allometric equations
(TerMikaelian & Korzukhin 1997, Jenkins 2004), ANPPW increases more rapidly with diameter
than ANPPL for all species in this study; thus, partitioning to wood generally increases with tree
size. Several factors, however, were confounded in our data (and in similar datasets): as nutrient
availability increased, species composition changed (Host & Pregitzer 1992, Kobe 2006) and
trees tended to be larger (Table 1), so we cannot test the relative influence of these variables.
Species composition and stand density were correlated with productivity
Both species composition and stand density were significantly correlated with residuals of
production-resource models. Although we identified an effect of Q. rubra relative basal area on
ANPPL, this was probably an allometric effect rather than a consequence of species differences
in anatomy or physiology (Meerts 2002, Freer-Smith & Kennedy 2003). Sites with greater
2
proportional Q. rubra basal area contained smaller trees (r = 0.75 for Q. rubra basal area versus
mean diameter of other species), which produce fewer leaves and total mass than larger trees.
30
Smaller diameters of individuals of other species in the presence of Q. rubra could indicate
competitive or plant-soil feedback effects (McCarthy-Neumann & Kobe 2010).
Stand density may also partly explain why all models consistently under-predicted
productivity at two atypical sites (5 and 9) (Fig. 1b, 2b). The high productivity of these sites
arose from both higher stand densities (20% greater, Table 1) and higher individual-tree growth
rates. The other attribute most obviously shared by these two sites was their location at the
northwestern edge of our study area, which could be associated with unmeasured differences in
underlying parent material, establishment history, initial stocking density, or land use prior to
secondary forest re-growth.
Comparison with previous Great Lakes studies
The weak relationships between ANPPW and metrics of N availability differed from previous
research conducted at our sites (Zak et al. 1989) and at similar sites in the Great Lakes region
(Pastor et al. 1984, Reich et al. 1997). We estimated ANPPW based upon changes in tree size
over a relatively short interval, whereas Zak et al. (1989) estimated ANPPW as the quotient of
standing wood mass and stand age. With the latter approach, productivity is increasingly
underestimated with greater stand age because the growth of trees that die prior to census is
discounted (Clark et al. 2001). This underestimate is more severe at high-productivity sites
because mass lost to mortality is positively related to ANPP (data not shown). We re-calculated
wood production for our dataset as standing wood mass/stand age (ANPPW,A), which
2
substantially underestimated ANPPW (slope = 0.60, r = 0.68) across all sites.
31
Setting aside effects of how ANPP is measured, ANPPW,A was weakly related to N
2
2
mineralization rate (r = 0.18), which contrasts with the much stronger relationship (r = 0.85) in
Zak et al. (1989). This incongruous result is not likely related to lab methodology because
potential N mineralization was estimated in both studies from a single ex situ incubation and
resulting site ranks by N mineralization were identical. In contrast, Zak et al. (1989) regressed
mean values of N mineralization and ANPPW,A of nine forest ecosystem types, with each mean
value calculated from up to nine stands. Using means in the regression reduces site variation
within ecosystem type (e.g. Zak et al. (1989) reported standard errors up to ±18% for N-min and
±30% for ANPPW within ecosystem type) and could explain the discrepancy between studies in
the strength of the relationship between N availability and productivity.
Conclusions
In this study, Ca, N, and soil moisture were all important correlates of ANPP. Overall, Ca was
the best correlate of ANPPW, but this relationship reached an asymptote at relatively low soil Ca
content. In contrast, ANPPW was related to ΣN and soil moisture across the entire domain of
both resources, but these resources over-predicted ANPPW for low-productivity sites. Most
temperate forest ecosystems are thought to be N-limited (Finzi & Canham 2000), but we
hypothesize that Ca, N, and soil water could jointly control production, with variable importance
depending on soil type and species composition. We are more confident that ANPPL was related
to N availability, but cannot discern whether this was a causative or cyclical relationship. With
several resources possibly controlling ANPP, and with different potential limiting resources for
ANPPW versus ANPPL, edaphic conditions and biological responses could interact counter
32
intuitively. For instance, greater N availability from atmospheric deposition or fertilizer
application should elevate production overall by enhancing photosynthesis (Iversen & Norby
2008). If, however, N deposition causes Ca leaching (Izuta et al. 2004), then the greater
photosynthetic output expected from elevated N availability could fail to increase ANPPW due to
Ca limitation. Alternately, if ANPPW is regulated by soil moisture rather than Ca availability, N
deposition is unlikely to increase future production because most climate models (McMahon et
al. 2009, Rowell 2009) predict decreasing soil moisture levels for central North America. In such
situations, ecosystem models that forecast productivity based on soil N dynamics (e.g. Schimel et
al. 1996, Schimel et al. 1997) would mistakenly predict positive rather than negative ANPP.
Overall, our results show that multiple soil resources are correlated with ANPP, emphasizing that
a broad range of resources should be tested for their influences on ANPP, and that long-term
resource manipulation experiments are still necessary to determine resource limitation in these
forest systems.
33
Acknowledgements
We thank the following persons for assistance with field and lab work: M. Erickson, J. Bramer,
D. Minor, A. Edmonson, B. Bell, and M. Petrova. This manuscript was improved by helpful
suggestions from M. Walters, A. S. Grandy, S. McCarthy-Neumann, E. Holste, and two
anonymous reviewers. We appreciate the expertise of Dr. R. Schaetzl (Michigan State University
Department of Geography) and his help with soil texture measurement. The research was
supported by the Michigan Agricultural Experiment Station (NRSP-3, the National Atmospheric
Deposition Program), the Michigan State University Rackham Foundation, and the National
Science Foundation (DEB 0448058).
34
Table 2.1. Annual productivity increments were calculated between 1999 and 2007 for ANPPW, ANPPL, and total ANPP. Species
composition shifts predictably across landform, but outlying individuals can be found at some sites. Stand age should be interpreted as
elapsed time since significant, stand-replacing disturbance, with a mean of 87 years of minimal local anthropogenic disturbance.
Basal
ANPPW
ANPPL
ANPP
area
Density
Dominant
95%
Age
a
-1
-1
-1
2 -1
-1
b
dbh(cm)
(years)
Location
ID
(Mg ha ) (Mg ha ) (Mg ha ) (m ha ) (Stems ha )
Species
1, O
2.62
2.80
5.42
22.69
582
Qa, Qv, Pb
45.8
88
44°14‘N, 85°56‘W
2, O
2.73
2.78
5.51
23.07
510
Qa, Qv, Qr
49.7
80
44°14‘N, 85°57‘W
3, I
3.93
3.09
7.02
31.29
555
Qv, Qr, Ar
51.1
97
44°16‘N, 85°53‘W
4, I
3.19
3.34
6.53
30.23
486
Qr, Ar, Pg
51.8
111
44°17‘N, 85°53‘W
5, mI
5.09
3.52
8.61
38.10
725
Qr, Qv, Ar
44.3
83
44°19‘N, 85°52‘W
6, mI
4.34
3.16
7.50
36.46
623
Qr, Pg, Ar
58.4
90
44°12‘N, 85°48‘W
7, pM
4.40
3.73
8.13
33.10
568
As, Qr, Ar
48.3
79.0
44°11‘N, 85°45‘W
c
8, pM
4.88
2.99
7.87
636
As, Fg, Ar
37.9
77
44°20‘N, 85°29‘W
NA
9, iM
5.32
3.80
9.12
39.80
712
As, Fg, Qr
56.8
80
44°22‘N, 85°49‘W
10, M
4.62
3.77
8.39
39.97
642
As, Fg, Ta
59.6
84
44°15‘N, 85°45‘W
11, M
4.71
3.75
8.45
36.24
544
As, Fg, Ta
52.6
76
44°13‘N, 85°40‘W
12, M
4.31
3.19
7.49
36.10
612
As, Ta, Qr
55.9
92
44°13‘N, 85°45‘W
13, M
5.16
4.05
9.22
33.59
506
As, Ta, Qr
66.9
92
44°13‘N, 85°45‘W
a
Landform codes: O: outwash; I: ice contact; mI: mesic ice contact; pM: poor moraine; iM; intermediate moraine; M: Moraine.
b
Species codes: Ar: Acer rubrum; As: Acer saccharum; Fg: Fagus grandifolia; Fa: Fraxinus americana; Pb: Pinus banksiana; Ps:
Pinus strobus; Pg: Populus grandidentata; Qa: Quercus alba; Qr: Quercus rubra; Qv: Quercus velutina; Ta: Tilia americana. The
three most prevalent species are listed in order of abundance; other species occurred in small proportion.
35
c
Site area used for production calculations under-represented species prevalence and was excluded from basal area analysis.
36
Table 2.2. Model formulae, where RS indicates resource and A, B and/or C are estimated parameters. Mathematical effects of each
parameter are presented, with a corresponding biological interpretation. Models (1), (2), and (3) used two parameters, reducing their
AIC penalty in comparison to model 4.
Model
1.
Parameters
A: Slope
B: Minimum ANPP
Biological interpretation
Linear positive or negative effect of resource,
unlimited ANPP across the resource domain;
non-zero minimum ANPP
2.
A: ANPP upper bound
B: Low-resource ANPP
Non-linear positive effect of resource on ANPP;
ANPP may respond linearly to resource
increase or reach its upper bound within the
domain
3.
A: ANPP upper bound
B: Slope parameter governing where on the
domain the upper bound is reached (≈ zero
steep, >>zero shallow)
Positive saturating relationship of ANPP to
resource; rapid asymptotic behavior permits
ANPP to reach its upper bound at low resource
levels
A: Asymptotic ANPP (upper bound)
B: Median resource value; point on domain
corresponding to most rapid ANPP increase C:
Slope parameter (≈ zero steep, >>zero
shallow)
Positive saturating relationship of ANPP to
resource; rapid asymptotic behavior with
permits maximum ANPP a low resource
availability, but also restricts minimum ANPP
to greater than zero
ARSB
A RS
A  RS
B
AeB RS
4.
A
(RS
B)
C
1e
37
Table 2.3. AICc comparisons for top ANPPW models of each resource, with supported models presented in bold. Ca explained the
greatest percent variance and received strongest support. Models with soil water and ΣN received little support, but both explained
more than 40% ANPPW variance. Parameter estimates with 95% confidence intervals are presented as A, B, and/or C, which match
2
parameters listed in Table 1. Parameter number is k, with r and 95% confidence intervals around each parameter instead of p-values.
Independent
Variable
d
Soil H2O
Texture
+
Acidity, Al
Ca
2+
+
Model
2
r
k ΔAICc
A (95% C.I.)
B (95% C.I.)
1
1
1
0.464 3 4.215
0.478 (0.137, 0.820)
0.038 3 11.805 -0.069 (-0.296, 0.159)
0.235 3 8.831 -0.962 (-2.113, 0.189)
4
0.722 4
0.000
4.774 (4.297, 5.251)
0.196 (0.084, 0.309)
-1.131 (-4.996, 2.734)
4.930 (2.615, 7.244)
4.994 (3.976, 6.012)
1
0.109 3 10.813
14.49 (-12.99, 41.97)
2.431 (-1.068, 5.929)
2+
4
0.700 4
1.002
4.695 (4.236, 5.154)
0.110 (0.097, 0.123)
-
PO4
C
C:N
N
NH4
1
1
1
3
1
0.095
0.003
0.097
0.069
0.319
11.021 -2.603 (-7.942, 2.736)
12.271 0.119 (-1.234, 1.472)
10.993 -0.041 (-0.125, 0.043)
11.379 4.839 (3.215, 6.462)
7.122
0.759 (0.023, 1.495)
4.822 (3.535, 6.11)
4.021 (1.314, 6.728)
5.175 (3.228, 7.123)
0.011 (-0.016, 0.037)
2.700 (1.122, 4.279)
NO3
3
0.223 3
9.041
5.533 (3.757, 3.757)
0.182 (-0.036, 0.399)
ΣN
Nitrification
N-min rate
1
3
1
0.408 3 4.596
0.229 3 8.929
0.023 3 12.013
0.589 (0.118,1.060)
5.060 (3.907, 6.213)
0.949 (-3.16, 5.058)
2.543 (1.106, 3.979)
0.028 (-0.006, 0.062)
3.900 (2.269, 5.531)
K
Mg
d
3
3
3
3
3
C (95% C.I.)
d
0.144 (0.021, 0.267)
0.016 (0.001, 0.032)
Model received little support (ΔAICc > 4), but nonetheless explained at least 40% ANPPW variance
38
Table 2.4. AICc comparisons for top ANPPL models of each resource, with supported models
presented in bold. ΣN was the best predictor of ANPPL, although NO3 was also supported. Ca
and nitrification rate received less support, but still explained substantial ANPPL variance.
Parameter estimates with 95% confidence intervals are presented as A and/or B, which match
2
parameters listed in Table 1. The k value represents parameter number; r and 95% confidence
intervals around each parameter are presented in place of p-values.
Independent
Variable
Model
Soil H2O
Texture
+
Acidity, Al
2+e
Ca
+
2
k
ΔAICc
A (95% C.I.)
B (95% C.I.)
1
3
1
0.198 3
0.044 3
0.268 3
6.298
8.772
5.294
0.145 (-0.049, 0.339)
3.630 (2.769, 4.491)
-0.477 (-1.001, 0.046)
1.746 (-0.449, 3.941)
0.644 (-1.426, 2.714)
3.750 (3.287, 4.212)
3
0.422 3
2.222
3.687 (3.366, 4.007)
0.055 (0.010, 0.099)
r
3
0.234 3
5.894
4.902 (2.709, 7.096)
0.046 (-0.009, 0.101)
2+
3
0.294 3
4.834
3.781 (3.300, 4.262)
0.023 (-0.001, 0.047)
-
PO4
C
C:N
N
NH4
3
1
1
1
1
0.022
0.024
0.078
0.058
0.317
3
3
3
3
3
9.066
9.036
8.298
8.586
4.378
3.461 (3.009, 3.914)
0.148 (-0.474, 0.770)
-0.017 (-0.057, 0.022)
2.249 (-3.788, 8.285)
0.352 (0.009, 0.694)
0.004 (-0.013, 0.021)
3.092 (1.848, 4.337)
3.767 (2.853, 4.681)
3.160 (2.512, 3.809)
2.663 (1.929, 3.397)
NO3
ΣN
e
Nitrification
N-min rate
3
1
3
1
0.513
0.605
0.355
0.031
3
3
3
3
1.217
0.000
3.662
8.944
4.278 (3.634, 4.923)
0.333 (0.154, 0.512)
3.850 (3.367, 4.333)
0.515 (-1.387, 2.416)
0.162 (0.061, 0.264)
2.414 (1.869, 2.960)
0.021 (0.002, 0.039)
3.190 (2.436, 3.945)
K
Mg
e
Model received some support (2 < ΔAICc < 4); may be a significant correlate of ANPPL.
39
-1
-1
ANPPW (Mg ha yr )
-1
-1
-1
Predicted ANPPW (Mg ha yr )
-1
-1
ANPPW (Mg ha yr )
Ca (cmolc kg )
-1
Soil water (%)
NO3 + NH4 (mg kg )
Figure 2.1. ANPPW expressed as functions of soil resources. (a) ANPPW was best correlated
with Ca, with the best-fit model shown by the solid line and 95% confidence intervals by dashed
lines. (b) Observed ANPPW compared to predicted values from the Ca-ANPPW model.
Considerable ANPPW variance was explained by soil water (c) and by ΣN (d), but these
relationships received relatively weak support compared to base-cation models.
40
-1
-1
ANPPL (Mg ha yr )
-1
-1
-1
Predicted ANPPL (Mg ha yr )
-1
-1
ANPPL (Mg ha yr )
NO3 + NH4 (mg kg )
-1
-1
Ca (cmolc kg )
NO3 (mg kg )
Figure 2.2. ANPPL expressed as functions of soil resources. (a) ANPPL was best correlated
with ΣN, with the best-fit model shown by the solid line and 95% confidence intervals by dashed
lines. (b) Observed ANPPL compared to predicted values from the ΣN-ANPPL model.
Substantial ANPPL variance was explained by Ca (c) and by NO3 (d), but compared to the ΣNANPPL model these relationships received relatively little support.
41
ANPPW / total ANPP
-1
-1
ANPP partition (Mg ha yr )
-1
-1
-1
Total ANPP (Mg ha yr )
ANPPW / total ANPP
ANPPW / total ANPP
Ca (cmolc kg )
-1
Soil water (%)
NO3 + NH4 (mg kg )
Figure 2.3. Productivity partitioning to wood and leaves expressed as functions of ANPP
and several soil resources. (a) The solid line is the best-fit ANPPW line (open circles), while the
dash line is the best-fit ANPPL line (triangles). The fraction of total ANPP represented by
ANPPW was related to Ca (b) and soil water (c) by log-normal functions, and weakly to ΣN (d)
by a linear relationship.
42
CHAPTER III
NEIGHBOR INTERACTIONS STRENGTHEN WITH INCREASED SOIL RESOURCES
IN A NORTHERN HARDWOOD FOREST
43
Abstract
1. The relationship between soil resource availability and competitive interactions remains
unresolved, especially in forest ecosystems. If competition shifts from below- to above-ground
across soil resource gradients, then competitive interactions should constrain growth across all
soil fertility levels. Alternatively, competition may be less important under stressful conditions
associated with low soil resources.
2. We developed individual-based growth models as functions of local tree neighborhood and
direct measurements of soil resources for 10 common tree species from sites established across a
soil resource gradient in northwest Lower Michigan, USA. We hypothesized that tree growth
should increase with soil resource availability (rarely measured at a local scale), but decrease
with the density, size, and proximity of neighboring trees.
3. Correlations of growth to neighborhood effects were strongest for species occupying primarily
high-resource sites. Correlations of growth to soil resources were positive for species associated
with high and intermediate fertility. In contrast, correlations of growth to soil water were
negative for species associated with low fertility, suggesting that competitiveness of these
species decreased with higher soil resources and concomitant decreases in irradiance. Growth
was positively associated with focal tree diameter in all species.
4. For two species common at sites with high resources, interactions with conspecific neighbors
had the strongest influence on growth, and the overall effect of the neighborhood was
competitive.
5. Relationships between mean site-level tree growth and soil resources were much stronger than
individual growth-local resource relationships. Weaker species-specific, individual-level trends
likely arose from limited species distributions across each soil resource domain.
44
6. Synthesis. Neighborhood interactions were more prevalent in species associated with high soil
fertility sites, where canopy interceptance of irradiance was high. Furthermore, the high fertility
species that responded most strongly to neighborhood competition were shade intolerant. For
species dominant at low fertility, where canopy transmission of irradiance was relatively high,
neighborhood interactions were absent or negligible. The growth of low-fertility species was
negatively correlated with soil water, but decreasing site-level canopy openness with soil water
suggests that low-fertility species were out-competed for irradiance as soil resources increased.
Thus, irradiance likely mediated the stronger competitive interactions at high fertility sites.
KEYWORDS: Base cations, geological gradient, interspecific / intraspecific competition,
nitrogen, neighborhood models, resource heterogeneity, tree diameter growth
45
Introduction
Competition for resources between individual plants can structure communities (Davis et al.
1999, MacDougall & Turkington 2004), but how competitive interactions change across soil
resource gradients is unresolved (Brooker & Kikividze 2008, Craine 2005). Resource
competition could substantially influence plant growth at all fertility levels (Tilman 1988), with
allocation to resource acquisition structures decreasing in proportion to resource availability
(Aikio et al. 2009, Casper et al. 1998, Chen & Reynolds 1997). Alternately, resource
competition could be most significant when resources are abundant (Grime 1979, Grime 2001,
Coomes et al. 2009), while competition could decrease as plants are required to tolerate the
stressful conditions associated with low resource availability (Coomes & Grubb 2000).
Mitigation of competitive interactions with decreasing soil resource availability may lead to the
dominance of facilitative interactions in low-resource environments (Lortie & Callaway 2006,
Maestre et al. 2009).
Neighborhood models of individual tree growth typically have not explicitly included soil
resources and have used either purely phenomenological neighborhood indices (Wright et al.
1998, Uriarte et al. 2004, He & Duncan 2000) or have incorporated only crown-model estimates
of light attenuation by neighbors (Canham et al. 1999, Canham et al. 2004, Coates et al. 2009).
The rare studies that did consider soil resources did so indirectly by inferring fertility from
vegetation (Canham et al. 2006) or assuming that residual variance of aboveground interactions
could be ascribed to soil resources (Coates et al. 2009, McPhee & Aarssen 2001). Soil resource
availability for individual trees is difficult to characterize because of meter- to tree- scale (and
coarser) spatial heterogeneity in resources (Finzi et al. 1998, Bigelow & Canham 2002, Reynolds
et al. 2007, Townsend et al. 2008, Lundholm 2009). Measuring effects of soil resource
46
availability on individual tree growth requires local-scale resource measurements and
distinguishing soil resource effects from other factors that influence growth, including ontogeny
and neighborhood interactions.
In addition to soil resource effects, individual growth may be influenced by interactions
with neighboring trees (including competition for light), and by ontogenetic effects whereby
growth rate is a function of tree size (Stoll et al. 2002, Potvin & Dutilleul 2009). Ontogenetic and
neighborhood effects may be confounded in mature, relatively undisturbed forests. Small trees
grow slowly in closed-canopy forests because they are suppressed and shaded by canopy trees
(Coomes & Allen 2007, Vanhellemont et al. 2010). When growth is modeled as a function of
tree size and neighborhood effects, the slow growth of suppressed small trees may be attributed
to their size rather than to the neighborhood, especially if neighborhood indices have low
effective variability (Woods 2000, Thorpe et al. 2010). Separating ontogenetic from
neighborhood effects may be achieved by extrapolating a theoretical potential maximum growth
for an individual of a given size and introducing neighbor and resource effects to reduce this
maximum growth to the observed change in diameter (Canham et al. 2004, Uriarte et al. 2004);
this method requires broad effective neighborhood variability (Coates et al. 2009).
With increasing soil resource availability, competitive interactions between neighboring
trees could intensify due to depletion of irradiance (Canham et al. 1999, Kunstler et al. 2011).
Higher soil resource availability increases leaf thickness, size (Chen et al. 2010), and leaf area
index (Granier et al. 2000), thereby decreasing canopy openness and increasing competition for
photosynthetically active radiation (Dolle & Schmidt 2009). Light depletion due to reduced
canopy openness may also be a consequence of changing species assemblages (Simioni et al.
2004, Dent & Burslem 2009) or intraspecific, resource-dependent variation in canopy
47
characteristics (Henry & Aarssen 2001, Lefrancois et al. 2008). In the absence of direct
measurements of species-specific irradiance depletion across soil resource gradients, accurately
separating light effects from other effects of crowding may be impossible. Instead, we examined
the possibility that the interceptance of irradiance mediates changing competitive interactions
across a soil resource gradient by examining among-site variation in canopy openness.
From a theoretical perspective, individuals of all tree species could exert similar effects
on the growth of neighbors (i.e. ecological neutrality per Hubbell 2001), certain species could
compete more effectively (Caplat et al. 2008, Poorter & Arets 2003) and disproportionately
suppress neighbor growth, or a combination of neutral and asymmetric interactions could occur
(Adams et al. 2007, Vergnon et al. 2009). Neighborhood interactions have traditionally been
assessed as size- and distance-dependent functions (Berger & Hildenbrandt 2000, Stadt et al.
2007), assuming that closer and/or larger trees should influence focal tree growth more
substantially than smaller and/or farther-displaced trees. Simple size- and distance-dependent
neighborhood indices overlook the possibility of competitive asymmetry, effectively treating all
species as equivalent competitors. Abundant evidence for differences among species pair-wise
interactions and competitive capability (Stoll & Newbery 2005, Coates et al. 2009, Engel &
Weltzin 2008, McCarthy-Neumann & Kobe 2010) suggests that neighborhood analyses should
account for neighbor species identity(Coates et al. 2009, Uriarte et al. 2004). Differential
neighbor species influences can be accommodated by allowing neighbor species to contribute
disproportionately to a phenomenological index (Zhao et al. 2006) or by assigning speciesspecific canopy transmittance in quasi-mechanistic indices (Canham et al. 1999).
The major purpose of this study was to address how competitive interactions change with
soil resource availability by examining individual tree growth in relation to local soil resource
48
availability and interactions with neighboring trees. Debate persists over the ecological relevance
of the importance versus intensity of competition (Brooker & Kikividze 2008). For this study,
importance of competition was assessed relative to the amount of total variance in growth that
could be explained by tree diameter or by local soil resource availability; it was not our intention
to assess the absolute importance of competition (e.g. Freckleton et al. 2009). Any soil resource
required for plant physiological function could potentially limit individual growth if sufficiently
rare, but we focus on Ca, water, and N, which have been identified as correlates of stand-level
productivity at these (Zak et al. 1989, Baribault et al. 2010) and other sites (Finzi 2009, Hogberg
et al. 2006, Joshi et al. 2003) or established as limitations to growth through experimental
manipulations (Gradowski & Thomas 2008, Park et al. 2008, Finzi 2009, McDowell et al. 2009).
We build upon established methods for characterizing neighborhood interactions (Berger
& Hildenbrandt 2000, Canham & Uriarte 2006, Vettenranta 1999, Wimberly & Bare 1996, Zhao
et al. 2006) and extend these methods by including soil resource availability (calcium (Ca), soil
-
+
water, sum of nitrate (NO3 ) and ammonium (NH4 ) (ΣN)), and potential nitrogen (N)
mineralization) interpolated for each focal tree. As in similar studies, our phenomenological
index of neighborhood interactions implicitly measures net tree-tree interactions, including
competition for light (Canham et al. 2004), belowground competition (Coomes & Grubb 2000),
tree species effects on soil resource availability (Fujinuma et al. 2005), indirect effects of tree
species on soil biota and chemistry (McCarthy-Neumann & Kobe 2010), as well as contributions
from other unmeasured factors. Joint consideration of neighborhood interactions, individual-tree
soil resource availability, and site-level measures of light interceptance could help resolve how
competitive interactions change across broad variation in soil resources. We tested four
49
hypotheses using the 10 most common deciduous species from 13 sites across broad soil
resource gradients that vary with glacial landforms in northwest Lower Michigan:
Hypothesis 1: Individual tree growth is negatively related to interactions with neighboring trees
and positively related to local soil resource availability.
Hypothesis 2: Competitive interactions are more important for species common at high-resource
sites than for species common at low-resource sites.
Hypothesis 3: Intraspecific neighbor interactions are stronger than interspecific interactions.
Hypothesis 4: Site mean growth and individual-level growth are related to the same soil
resources, with similar strengths of correlation.
50
Methods
Study sites
Data for this study were collected in 13 mixed hardwood stands of the Manistee National Forest
in the lower peninsula of Michigan, USA (~44º12‘N, ~85º45‘W). These sites were specifically
selected to span the broadest possible ecological gradient, with two or three sites representing
each major forest ecosystem type (Zak et al. 1989) and its associated species assemblage (Host
& Pregitzer 1992) in the region. Stands were 80 to 100 year-old second growth, and experienced
minimal anthropogenic disturbance since the 1920‘s. All 13 sites were within an area of 960 km
2
(maximum distance between sites of 38.1 km, average distance 15.76 km, maximum elevation
difference 236 m) and are expected to have very similar climate. Mapped stands were
established, wherever possible, in plots of dimensions 240m x 41m, with all stems >10cm dbh
included in the original census. Most plots conformed to these dimensions, but a sharp change in
topography near site 4 necessitated a different plot shape (two 120 m x 41m rectangles connected
by their ends at a 63° angle), while a road at site 6 required the longer dimension to be extended
to 260 m. Sites 12 and 13 were adjacent halves of the one full-size plot that differed in species
composition (e.g. absence of Populus grandidentata at site 13 vs. 23% total P. grandidentata
stems at site 12 (see Table B.2 in Supporting Information)) and soil resources.
Forest community composition in this region has been associated with particular glacial
landforms and soil fertility levels (Zak et al. 1989, Host & Pregitzer 1992, Baribault et al. 2010),
with lowest resource conditions at outwash sites and maximum resource availability on moraines
(Table B.3). Nitrate and ammonium pools, along with potential N-mineralization rate, increased
from outwash through ice-contact and moraine landforms (Table B.3). Extractable Ca
concentrations and soil volumetric water content also increased across the landforms (Table B.3).
51
Species composition shifts from dominance of Quercus alba and Quercus velutina on outwash
plains through Quercus rubra on intermediate sites, to Acer saccharum on moraines (Table 1,
Table B.2). Some species occurred across as many as 11 sites (Acer rubrum, Q. rubra), whereas
other species were only present in three (Fagus grandifolia) or four (Tilia americana) sites
(Table B.2).
Individual growth measurements
Stems of all mapped trees were measured at 1.37 m height using diameter tapes in 1999, 2005,
2007, and 2009. Growth increment was calculated as the change in diameter over the decadelong measurement interval for trees ≥ 10 cm diameter. For trees that grew into the ≥ 10-cm class
during the census interval, growth was calculated as the difference between measured diameter
and the 10-cm threshold; these in-growth trees also were included as neighbors. Intermediate
census dates in 2005 and 2007 enabled calculation of growth increment over the 1999-2005 or
1999-2007 intervals for trees that died during 2005-2007 or 2007-2009; interval length was
adjusted as appropriate. Individuals that grew after 1999 but died before 2005, 2007, or 2009
were included.
We excluded from the growth analysis any trees that died during the interval but had not
grown. Such trees could have contributed to local interactions via shading and resource uptake,
so they were included as neighbors. Trees with final diameter < 10 cm were not measured and
thus are not included in the analysis. In our data, the 10-20cm class showed near-zero growth,
suggesting that strong competitive effects were experienced by most stems < 20 cm diameter.
Each species had multi-stemmed individuals. Treated as separate individuals, the close
proximity of neighboring stems could have disproportionately influenced their contributions to
the neighborhood index. Thus, we estimated two sets of growth models where multi-stemmed
52
individuals were treated as separate trees or as a single tree with basal area equal to the sum of
the individual stem basal areas. For nine species, <4% of the stems were members of multi-stem
aggregates and the two sets of models and final model selection were unaffected; for these
species, we report results based on treating multiple stems as separate individuals. For T.
americana, in contrast, 12% of stems belonged to aggregates, and both model performance and
selection changed substantially. Treating stems as individuals resulted in a positive growth
response to the neighborhood index, due to the strong influence of close-proximity stems on the
neighborhood index. Thus, for T. americana, multi-stem trees were treated as a single individual,
both as focal trees and as neighbors, in its growth model. As neighbors in heterospecific growth
models, results were indistinguishable for composite versus separate stems, and the latter were
used consistently.
Resource measurements
Soil volumetric water was measured to 30 cm depth by time domain reflectometry
(Environmental Sensors Inc., Sidney, BC). Measurements were made in August 2009 at 5-m
intervals along a central 200-m transect at each site. Data for all sites were collected during a
rainless period and over an interval of less than two days. Soil nutrient analyses were conducted
on composite samples each comprising three soil cores (2 cm diameter, 15 cm depth) collected in
June 2009. Cores included the A horizon, the AE horizon if present, and the portion of the
organic horizon in which tree roots were observed, but excluded leaf litter. Samples were
gathered at 10-m intervals along a central transect at each site, and at 20-m intervals along
longitudinal transects displaced 5 and 10 m from the central transect. Samples were air dried
prior to analysis.
53
Potential net N mineralization was determined by aerobic incubation. Initial
+
-
concentrations of NH4 and NO3 were measured in a 6-g subsample of each composite. Soils
+
-
were shaken for two hours in 2M KCl to extract inorganic N, and then [NH4 ] and [NO3 ] were
measured with a colorimetric assay using a fluorescent plate reader (ELx808 Absorbance
Microplate Reader, BioTek Instruments, Inc, Winooski, VT). A second set of 6-g subsamples
was incubated aerobically in darkness at 25°C for 28 days. Constant moisture content was
maintained by periodic addition of deionized water, and post-incubation concentrations of NH4
+
-
and NO3 were measured by the same method. Potential net N mineralization was calculated as
+
-
the difference between final and initial inorganic N [NH4 + NO3 ].
Base cations were extracted from 4-g subsamples shaken for 15 minutes in an acidic
Mehlich III solution (Carter 1993). Concentration of Ca
2+
in extracts was measured with
inductively coupled plasma atomic emission spectrometry (Optima 2100DV ICP Optical
Emission Spectrometer, Perkin-Elmer, Shelton, CT).
For Ca, soil water, ΣN, and N mineralization, we imputed the value of each at the
coordinates of each tree stem in the dataset using a distance weighted average from the R
package yaImpute (Crookston & Finley 2007). Resource availability for each tree was influenced
by the five nearest points at which the resource was measured, and including additional points
did not change the imputed resource values to at least the fifth decimal place. Uncertainty in
resource estimates due to the imputation was not ascertained. Irradiance was measured as diffuse
non-interceptance using a LAI 2000 (LI-COR Biosciences, Lincoln, NE) at 30cm above ground
54
level in 2-m intervals along the center of each site during August 2004. The mean and standard
error of diffuse non-interceptance (Table B.2) were used in site-level analysis.
Analysis
For each species, we modeled individual diameter growth (Gd) as a function of (i) tree diameter,
(ii) interactions with neighboring trees, and (iii) each of the soil resource metrics. The model
likelihood was based on a normal probability density for growth, which was confirmed by
normally distributed residuals. We predicted mean tree growth for a given soil resource level and
neighborhood using models that were additive,
Gd = Diameter effect – Neighborhood effect + Soil resource effect
(eq. 1)
and multiplicative,
Gd = Diameter effect  Neighborhood effect  Soil resource effect
(eq. 2)
Most investigators have adopted only multiplicative functions similar to equation 2 (Canham &
Uriarte 2006); we saw no a priori reason to exclude an additive model (equation 1) from testing.
The default was to test the full model (equations 1 or 2) for each species, but we also tested
reduced models for which we provide a biological interpretation in Table B.4. For the additive
model (equation 1), diameter effect was linear, requiring estimation of a single parameter:
Diameter effect =
A dbh
(eq. 3)
55
where A is the slope of the relationship between growth and diameter. For the multiplicative
model (equation 2), diameter effect was a lognormal function:
Diameter effect =

exp  0.5ln dbh    
2

(eq. 4)
where δ represents the diameter at which maximum growth rate is expected to occur based on the
data, and σ controls the rate the function achieves that maximum value.
All neighborhood indices were size- and distance-dependent functions of the number of
neighbors within a set or estimated radius. For additive models, the species-independent
neighborhood was:
 dbh exp  
n
Neighborhood effect = B
i 1

i
disti 

(eq. 5)
where B is a coefficient to control the relative contribution of the neighborhood overall, α is an
estimated exponent controlling the influence of neighbor diameter, and β is an estimated
exponent controlling the influence of distance to the focal tree. The sum is calculated for i =
1…n neighbors within either a fixed radius of 10 m or within a radius (≤ 10m) estimated from
the data. For multiplicative models, the neighborhood effect was itself an exponential:
n


 
Neighborhood effect = exp   B  dbhi exp
i 1

disti 



(eq. 6)
using the same definitions for B, α, and β. The final neighborhood indices, in which species
identity was parameterized, were extensions of equations 5 and 6:
56
s
n


B

dbh
exp

i
ij
Neighborhood-species effect =
i 1 j 1
 
distij


(eq. 7)
or
Neighborhood-species effect =
s
n


exp   B i dbhij dist ij 
i 1 j 1




(eq. 8)
where the index is calculated for j = 1…n neighbors of species i = 1…s. A species-specific
coefficient λi, where 0 < λi < 1, was estimated for each neighbor species. As with equations 5
and 6, B, α, β were assumed to be equal for all neighbor species in order to minimize the number
of estimated parameters. In equation 8, the exponential decay with neighbor distance was
simplified by placing the distance effect in the denominator and using β as an exponent. We
tested whether neighborhood effects disproportionately reduced growth of small trees by
replacing the coefficient B with a term that varied with focal tree size (Coates et al. 2009). This
term was never significant, however, so we report no further on size dependency of
neighborhood effects.
Soil resource effects were modeled both independently from the neighborhood index and
as an interactive term. For the additive model (equation 1), an independent soil resource effect
was included as another linear term:
Soil resource effect = C  resource
(eq. 9)
where C is an estimated scalar coefficient and the soil resource is expressed in units of ppm. For
the multiplicative model (equation 2), an independent soil resource effect was included as a
product,
57
C
Soil resource effect = resource
(eq. 10)
where C is an estimated exponent. Soil resources were also modeled as an interactive term with
all of the different neighborhood indices (5-8), where the soil resource value was included as a
quotient of the coefficient B on the neighborhood index.
To better separate neighborhood from size effects, we estimated a theoretical potential
growth for each species as a scalar coefficient on a model relating growth to tree size (following
Coates et al. 2009). Our data lacked instances of small trees growing rapidly in areas of low
neighborhood index, i.e. small trees released from competition, so we used USFS Forest
Inventory Analysis data (Bechtold & Patterson 2005) for the northern lower peninsula of
Michigan to estimate potential growth for smaller trees of each species. This potential growth
framework was not supported, so we report further details about methods and results in
Appendix B.1. To test for possible unmeasured effects of site, residuals from the best model for
each species were compared across sites (Figure S1). Residuals did not diverge by site; thus,
growth models did not include a separate term for site effects.
Parameter estimation and model comparison
Growth was estimated separately for each of the 10 deciduous species in the dataset for which
we measured more than 20 focal individuals. The simplest models required estimation of two
parameters and a variance term, while the most complex models required estimation of eight
parameters and variance. Model parameters were estimated with maximum likelihood using a
simulated annealing algorithm (Metropolis sampler) implemented in Delphi (Version 3.0,
Borland Corporation, Austin, TX), running each model for 50,000 steps. We calculated the 95%
58
confidence interval for each parameter through likelihood profiling. Whereas parameter
estimation was performed in Delphi due to its efficiency, subsequent analysis was performed in
R (The R Foundation for Statistical Computing, 2009, http://cran.r-project.org/). From the
correlation of model-predicted growth versus observed growth, we calculated the square of the
2
Pearson product moment correlation coefficient (r ), which we used to determine goodness of fit.
The slope values of these predicted versus observed relationships were used to check whether the
model over- (slope > 1) or under-predicted (slope < 1) individual growth. All models for a given
species were compared using Akaike‘s Information Criteria corrected for small sample size
(AICc) (Burnham & Anderson 2002). Models within two AICc units of the minimum AICc are
considered to have equivalent empirical support, whereas models with ΔAICc > 2 are not wellsupported by the data. The ΔAICc includes a penalty for additional model parameters.
Site level analysis
To test our fourth hypothesis—that individual growth-resource patterns scale up to the site
level—we used simple linear regression to assess the relationships of site-level properties,
including mean tree growth, mean tree size, stand density, and mortality rate to soil resources.
The goodness of fit for linear models of site-level variables as functions of soil resources was
2
measured by the coefficients of determination (R ) and 95% confidence intervals on slope and
intercept parameter estimates.
59
Results
Summary
Species associated with low versus high soil fertility had divergent growth responses to
neighborhood and local soil resource availability. Our division of species into associations with
low or high fertility sites was based on strong relationships between species composition and
glacial landform (Zak et al. 1989, Host & Pregitzer 1992). Neighborhood interactions were most
influential, and soil resources most important, among species dominant at sites with high
resource availability (Fig. 1, Table 2), supporting both our first and second hypotheses. There
were positive relationships of growth to soil resources in four species, and negative relationships
of growth to neighborhood index in five species (Table 2). The multiplicative model framework
exhibited the best performance for eight species, while the additive framework was best
supported for the remaining two species (Table 3). No obvious factor (e.g. species, site type,
taxonomy) distinguishes these species groups.
Among species common at low fertility sites, a negative correlation between individual
growth and soil water for three species suggests increasing competition (likely for irradiance)
with increased soil resources (Table 2). Individual growth was positively related to soil N for one
low-fertility species (P. grandidentata) (Table 2). For only one of these species (A. rubrum),
growth was weakly related to neighborhood (Fig. 1, Table 2), supporting our second hypothesis
that competitive interactions are weaker at low-resource sites. Prediction bias of the best model
for each species was typically low; slopes of 0-intercept regressions between predicted versus
observed growth were always greater than 0.9. Growth in all species was positively related to
focal tree diameter (Fig. 2).
60
We found competitive asymmetry in species-specific neighbor interactions in two of the
ten species (Table 2), partially supporting our third hypothesis. Consistent with our expectation,
intraspecific interactions were always of greatest magnitude, with extensive variation in the
importance of interspecific interactions (Table B.9).
Although soil resources are strongly correlated with stand productivity (Zak et al. 1989,
Baribault et al. 2010), they were more weakly correlated with individual tree growth, contrary to
our fourth hypothesis. Soil resources explained up to 8.6% of variation in growth whereas
neighborhood explained up to 19% and focal tree size explained 17% to 67%. In contrast, site
means of basal area increment and diameter growth were more strongly correlated with soil
2
resources (0.28 ≤ r ≤ 0.54). Because species composition changes across the resource gradient,
the correlations of mean basal area increment to resources may arise from both direct resource
effects on growth as well as indirect resource effects through changes in species composition.
Neighborhood generally was not correlated with growth of low-fertility species
Of the five species that were most prevalent on sites with lower soil fertility, growth of four
species, Q. alba, Q. rubra, Q. velutina, and P. grandidentata, was unrelated to neighborhood
(Fig. 3, Table 3, Table B.8). Growth of A. rubrum, typically associated with low to moderate
2
fertility, was negatively related to neighborhood (r increase of 2.2% above diameter effect),
-1
with expected growth approximately 1 mm year in the lowest-index neighborhood, but less
-1
than 0.1 mm year in a high-index neighborhood (Fig. 4, Table 3, Table B.8).
61
In general, growth of low-fertility species was negatively correlated with soil water, which
likely indicated growth responses to greater shading.
The only species for which a soil resource was positively correlated with growth was P.
grandidentata. Inclusion of ΣN explained 6.7% more growth variance than diameter alone
-1
(Table 3, Table B.8); growth was expected to increase from 1.8 to 3.2 mm year across the
domain of ΣN (Fig. 3). Growth of Q. velutina, typically found in greatest abundance at the
lowest fertility sites, was related neither to resources nor neighborhood (Fig. 2; Table 3, Table
B.8).
Growth was negatively correlated with soil water for Q. alba, Q. rubra, and A. rubrum
(Figs. 3 and 4, Table 3, Table B.8). Water explained 2.0-3.4% growth variance in these species
-1
(Table 3); growth from minimum to maximum water decreased from 1.3 to 0.9 mm year for A.
-1
rubrum, 3.4 to 2.2 mm year for Q. rubra, and 1.4 to 0.7 mm per year for Q. alba (Fig. 3). The
highest recorded volumetric water (21%) would not suppress growth in any of these species.
Consequently the apparently negative effect of water likely encapsulates shade-induced
decreases in growth: site-level canopy openness was negatively correlated with soil water (r = 0.69 for A. rubrum, r = -0.46 for Q. alba, r = -0.72 for Q. rubra). In contrast, neighborhood was
not correlated (r = -0.058) to soil water for A. rubrum, the sole low-fertility species for which
neighborhood was significant, indicating that shading effects across sites were not captured by
the neighborhood index. A tree of a given species could have similar neighborhood indices (sizes
of and distances to neighbors) at two sites, but experience different levels of shading due to interand intra-specific variation in light interception by neighbors (Montgomery 2004). Across sites,
mean canopy openness decreased from 12% to 1% with increased soil resource availability (Fig.
62
5). Within sites, there was extremely low variability in canopy openness and much greater
variability in soil resource availability (Fig. 5).
In high-fertility species, growth generally decreased with neighborhood and increased with
soil resources
Among the five species that predominated at sites of higher soil fertility, growth was negatively
related to neighborhood index for four species (F. americana, F. grandifolia, A. saccharum, and
P. serotina). The species most intolerant of shade, F. americana and P. serotina, showed the
strongest and third-strongest neighborhood effects, respectively (Table 3). While neighborhood
index formulations varied among species (Table 3), neighborhood effects explained from ~3.6%
(interacting with soil water) in A. saccharum to 19% total variance in F. americana (Fig. 6,
Table 3, Table B.8). Additionally, growth of F. grandifolia was positively related to soil water
(Fig. 4), P. serotina with soil water and Ca (Fig. 3), and A. saccharum with soil water through an
interaction with neighborhood index (Fig. 6, Table 3).
For F. americana, expected growth decreased substantially across the neighborhood
-1
domain, from 5.5 to 2.7 mm year (Fig. 6). For F. grandifolia, holding neighborhood constant at
its mean value, growth was expected to change across the soil water gradient from 0.5 to 1.2 mm
-1
-1
year (Fig. 4). Evaluated at mean water availability, a growth rate of 1.1 mm year could be
expected in the least suppressive neighborhood, compared to nearly zero growth in the most
suppressive neighborhood (Fig. 4). For P. serotina, mean growth increased from 1.4 to 3.4 mm
-1
year across the range of Ca; neighborhood index explained 4.6% variance in addition to Ca
-1
(Table 3), and reduced expected growth from 1.5 mm year in the lowest neighborhood to 0.6
-1
mm year in the highest (figure not shown).
63
In A. saccharum, the effects of neighboring trees on growth were diminished at higher
soil water. At minimum soil water availability, growth was expected to decrease from 1.6 to 0
-1
mm year across the range in neighborhood index, while at maximum soil water expected
-1
growth decreased from 1.6 to 1.2 mm year with increases in neighborhood (Fig. 6). Similarly,
growth in a minimum neighborhood configuration was expected to increase from 0 to 1.5 mm
-1
-1
year across the range in soil water, but only from 0 to 1.2 mm year in a maximum
neighborhood configuration (Fig. 6).
Conspecific neighbors had the strongest effects on focal tree growth
For A. saccharum and F. grandifolia, accounting for species of neighboring individuals
explained more growth variance than a species-neutral neighborhood model (Table 3, Table
B.8); for both species, conspecific neighbors had the strongest magnitude of effect. For A.
saccharum, interactions with T. americana suppressed growth most strongly after intraspecific
interactions (Fig. 6, Table B.9). For F. grandifolia, interactions with A. saccharum suppressed
growth most strongly after conspecific interactions; influence of other species was negligible
(Fig. 4, Table B.9).
Strong growth-resource relationships emerge at the site level
Site-level growth-resource relationships were stronger than individual growth-resource
relationships (Table 4). Whereas the strongest correlation of individual growth to any single soil
2
resource occurred for P. serotina and Ca (partial r = 0.086), all relationships of site mean basal
area to site mean soil resources were stronger. By simple linear regression, mean BAI was
2
2
2
significantly (p < 0.05) related to ΣN (R = 0.539), volumetric soil water (R = 0.500), Ca (R =
2
0.378), and potential N mineralization rate (R = 0.278). Mean basal area increment was
64
significantly related to all four soil resources, but mean diameter growth was related only to ΣN.
Neither stem density nor mortality rate was related to any resource (Table 4). The influence of
soil resources on site-level mean growth may be due to changes in species composition
associated with landform/ site-fertility. Outwash sites (11 and 12), characterized by lower
resource availability (Table B.3), were dominated by Q. alba and Q. velutina (Table B.2), both
species with relatively low mean growth rates (Table B.10) and diameters (Table B.11). In
contrast, at poor and intermediate moraines with higher resources (Table B.3) where Q. rubra
dominates (Table B.2), mean growth rates (Table B.10) and mean diameters (Table B.11) were
greater than for the outwash sites. Thus, the relationship of site-level growth to soil resources
was mediated by concurrent changes in species composition. Moreover, the individual-level
analyses were based on single species, and the majority of individuals of a given species
typically occupied less than half of the sites (Table B.2), representing a sharply restricted domain
of resource availability, which may have precluded detecting stronger growth-resource
correlations.
65
Discussion
Competition and soil resource effects increase in importance across the fertility gradient
Competitive interactions generally strengthened as soil resource availability increased (our
second hypothesis, and consistent with Grime 1979, Grime 2001), as supported by two lines of
evidence. First, neighborhood interactions were important for four of the five species dominant
at sites with high resource availability but were only weakly correlated with growth for one (A.
rubrum) of the five species dominant at lower-resource sites. For the high fertility species,
diameter growth was positively related to soil resources but negatively related to neighborhood,
supporting our first hypothesis. Evidence for increasing neighbor interactions was derived from
growth models for single species, but the individual trees used to calibrate these models were
sampled from the range of environments across which each species occurs. Species composition
changes across the sampled soil resource gradient (Baribault et al. 2010), however, so species
and resource effects were confounded. We cannot definitively determine whether stronger
neighborhood interactions arose from shifts in traits unique to the species that typically occupy
sites of high soil fertility (Kranabetter & Simard 2008) or direct effects of increasing resource
availability (Janse-ten Klooster et al. 2007, Liancourt et al. 2009). Nevertheless, correlations
between soil resource availability and interceptance of irradiance suggest that stronger
competitive interactions at higher soil fertility are mediated through greater shading. Greater
shading at high fertility likely arises as a result of changes in species composition (Canham et al.
1994) as well as mineral nutrient effects on leaf production within species (Chen et al. 2010).
Second, the negative correlation of growth to soil water in three species (Q. alba, Q.
rubra, and A. rubrum) common at lower-fertility sites was likely due to increased shading as soil
water increased and also supports competition increasing with resources. Additionally, these
66
three species are intolerant of shade (Kobe et al. 1995). Suppression of growth by increasing
water availability is unlikely across the range of water levels that we measured (McDowell et al.
2009); in fact, higher water availability is associated with increased growth of seedlings /
saplings (Kobe 2006, Schreeg et al. 2005). Across all sites at which these species occurred,
canopy openness was negatively correlated with soil water (and other soil resources, Fig. 5), but
neighborhood was unrelated to soil water for A. rubrum (the only low-fertility species for which
neighborhood was supported). A given neighborhood index, which is based on the diameter and
local density of trees, could be associated with a range of irradiance levels at different sites
because of inter- (Canham et al. 1999) and intra- (Dolle & Schmidt 2009) specific variation in
light transmittance through tree crowns. Neighborhood indices with explicit light attenuation
functions have not accounted for the potential influence of soil resources on species-specific
light transmittance (Lefrancois et al. 2008, Chen et al. 2010).
Species dominant at low soil fertility may have escaped light limitation because canopies
transmitted higher irradiance (Fig. 5) and stem densities were substantially lower (Table 1).
Slower growth of low-fertility species growing at sites with higher water and soil nutrient
availability (Schreeg et al. 2005) could have resulted because such sites were dominated by
species with denser canopies (Canham et al. 1994, Burton et al. 2009) that reduced light
transmittance. Weak or absent neighborhood effects in species dominant at low soil fertility may
have been principally a consequence of species distributions (Coates et al. 2009)—there were too
few individuals of low-fertility species growing at high-fertility sites to detect a correlation
between growth and neighborhood effects due to shading. When shade-intolerant A. rubrum and
Q. alba did occur under higher fertility conditions, they likely established when irradiance levels
were higher (Wang et al. 2010), as indicated by the larger diameters of individuals of these
67
species. These results are consistent with the interpretation that these species fail to effectively
compete at sites with more plentiful soil resources (Brooker & Callaway 2009, Craine 2005) and
supports that low fertility species are eliminated from high fertility sites due to growth sensitivity
to shading. Overall, our results add further evidence to the idea that the distribution of these
species across soil resource gradients arises from a species tradeoff between survival in low soil
resource conditions and growth in high-resource environments (Schreeg et al. 2005, Gaucherand
et al. 2006, Gravel et al. 2010), where low-fertility species are eliminated by shading.
The neighborhood effects that we detected for species associated with high soil fertility
also may have been mediated through light competition. Irradiance was uniformly lower at sites
with higher soil availability of water, Ca, and N (Fig. 5), suggesting increased competition for
light (Lintunen & Kaitaniemi 2010). It is important to note that the two most shade intolerant
species associated with high soil fertility, F. americana and P. serotina (Kobe et al. 1995),
exhibited some of the highest neighborhood effects (Table 3). Although our analysis likely
underestimates neighborhood effects for all species (see below), the stronger importance of
local-scale soil resource availability for species common at high soil fertility is consistent with
increasing competitiveness in the presence of abundant resources (Reiter et al. 2005). Thus, our
results provide evidence that apparent effects of soil resources on competitive interactions may
have been driven principally by changing levels of irradiance.
Intraspecific neighbor interactions were consistently more influential than interspecific
interactions in two of the five species for which growth was related to neighborhood index,
consistent with our third hypothesis. More closely-related individuals should theoretically be in
more intensive competition for the same resources (Uriarte et al. 2002, Canham et al. 2006, Zhao
et al. 2006) and thus unable to partition available resources with highest efficiency (Boyden et
68
al. 2008). In addition, high local conspecific density in intraspecific pairs could increase effects
of natural enemies or differential species-based chemical feedbacks (McCarthy-Neumann &
Kobe 2010). For A. saccharum, negative intraspecific interactions were nearly 30% more
important than interspecific interactions, and for F. grandifolia 14%. Even with evidence of
suppressive intraspecific interactions for only two species, competitive asymmetry has important
implications for silviculture. Both in multi-species plantations (Boyden et al. 2005, Forrester et
al. 2006) and in selection management of natural systems (Gronewold et al. 2010, Schwartz et
al. 2005), it would be useful to promote residual stand structures (Bauhus et al. 2009) consisting
of the least antagonistic neighbor interactions.
Despite limited variation in disturbance, neighborhood effects were detectable
We suspect that neighborhood effects were partially masked in our sites because of the limited
range of variation in neighborhood conditions; the growth of most trees, particularly of small
diameter, was probably suppressed by interactions with neighbors (Uriarte et al. 2004). Despite
limited variation in neighborhood densities at our minimally disturbed stands, neighborhood still
explained substantial growth variation for some species. For F. americana, neighborhood
explained 19% growth variance, likely because this shade intolerant species responded more
strongly to neighborhood variability than the shade tolerant species with which it co-occurs. Our
-1
-1
sites were relatively static with low mortality (~2 stems ha year ), limited gap formation, and
restricted opportunity for trees to be released from competition. Abundant evidence shows that
release from competition will dramatically increase growth (Stan & Daniels 2010, Noguchi &
Yoshida 2009). Greater variation in neighborhood conditions generally results in greater
measured effects of neighborhood on growth (Uriarte et al. 2004, Vanhellemont et al. 2010,
Zhao et al. 2006, Olano et al. 2009, Simard 2009, Thorpe et al. 2010). For example, in forests of
69
mixed northern temperate species encompassing unmanaged stands and 30-60% basal area
removal, neighborhood explains 25.8-77.3% total growth variance (Coates et al. 2009).
Some of the difference in growth variation associated with neighborhood can be
attributable to the statistics used to measure goodness of fit. We used the square of the Pearson
2
correlation coefficient (r ) to measure goodness of fit, but many other studies (e.g. Coates et al.
2
2009, Uriarte et al. 2004) use the coefficient of determination (R ) based on a simple linear
2
regression through the origin of predicted versus observed values. The interpretation of R for
models with the origin as the intercept is problematic (Myers 1986), however, and leads to
unrealistically high estimates of amount of total variance explained. For example, predicted
2
versus observed values for the supported model of Q. velutina generated r = 0.273 (Table 2),
2
but the origin-intercept regression of those same values produced R = 0.920. Thus, it is difficult
to directly compare our results with previous studies.
Individual level relationships did not scale up to site level patterns
Past research at our sites identified N (Zak et al. 1989), Ca, and soil water (Baribault et al. 2010)
as strong correlates of site-level productivity, but our current results do not support similarly
high correlations of individual growth with any of these resources for any species, contrary to
our fourth hypothesis. At the individual level, growth is determined by resource availability and
acquisition, which is related both to tree size and to local neighborhood (Canham et al. 1999). In
contrast, productivity or growth at the site level is controlled by the aggregate growth rate of all
the trees at a site and by stem density and mortality.
Significant relationships of growth to resource availability likely were more difficult to
detect for single species than for all species because most species had restricted distributions
70
across sites (Table B.2). For example, 65% of all Q. alba individuals occurred only at sites 11
and 12 (Table B.2), where the range in Ca availability was < 3% of the total range, soil water
56% of its range, and ΣN 8% of its range (Table B.3). Our results are consistent with other
studies that have found that individual growth is most strongly correlated with tree size
(Andreassen & Tomter 2003, Laubhann et al. 2009, Monserud & Sterba 1996), and that site
mean growth is often substantially correlated with soil resources (Bedison & McNeil 2009,
Devine & Harrington 2009, McDowell et al. 2009, Pastor et al. 1984, Reich et al. 1997, Wallace
et al. 2007, Zak et al. 1989)
Both stochastic variability in growth and measurement error introduced by diameter tapes
could contribute to variance in individual growth (Clark et al. 2007, Wyckoff & Clark 2005), but
this individual-scale noise exerts less influence at the site level, allowing a stronger correlation of
mean site growth to soil resources. In addition, the discrepancy between the individual and stand
also may be due to difference in spatial scales between measured soil resources and the area over
which a tree acquires resources (Das & Chaturvedi 2008, Kalliokoski et al. 2008, Yanai et al.
2008). Mycorrhizal networks also could connect trees across a broad area (Simard 2009),
decoupling local soil resources from plant available resources.
Conclusions
Two lines of evidence from this study support that the importance of competition increases
across resource gradients. First, competitive neighborhood interactions (H1) were important
primarily for species dominant at sites of high resource availability (H2). Second, increasing soil
water was associated with slower growth in three species dominant at low-resource sites,
signifying decreased competitiveness (H2) for light at higher-resource sites. The magnitude of
predicted neighborhood effects varied widely, but included complete suppression of growth at
71
some of the highest neighborhood indices. Intraspecific neighbor interactions were strongest in
two species (H3), though divergent species distributions likely precluded identifying other
species effects. Competitive neighborhood effects were prevalent in species associated with high
soil fertility, but largely absent for species associated with low soil fertility.
Acknowledgements
We thank S. Grandy, D. Rothstein, and M. Walters for insight about sampling, resource analysis,
and model design. D. Coomes, E. Holste, D. Minor, B. Bachelot, A. Maguire, D. Rozendaal, and
two anonymous reviewers provided helpful feedback in preparing the manuscript. Indispensable
field and laboratory assistance was provided by M. Erickson, D. Minor, J. Bramer, A. Maguire,
A. Pierce, and A. Stinson. In addition, we thank A. Finley and B. Walters for access to USFS
FIA data. The research was supported by the Michigan Agriculture Experiment Station (NRSP-3,
National Atmospheric Deposition Program) and NSF (DEB 0958943).
72
Table 3.1. Basic statistics, spatial distribution, and relative basal area for the ten most common
species across our sites. At least 90% of the stems for each species occurred across five or fewer
sites. All species had similar minimum diameter because our census intentionally excluded trees
smaller than ~10 cm. Sample size (n) refers to the number of focal trees for each species.
Species
Acer rubrum
Acer saccharum
Fagus grandifolia
Fraxinus americana
Populus grandidentata
Prunus serotina
Quercus alba
Quercus rubra
Quercus velutina
Tilia americana
code
acru
acsa
fagr
fram
pogr
prse
qual
quru
quve
tiam
n
554
676
110
30
84
27
486
589
184
183
Sites > 10%
total B.A.
3, 4, 9, 13
1, 2, 5, 8,10
1, 5, 8
1, 7, 8
1, 4, 7, 9, 13
2, 5, 6, 8
3, 9, 11, 12
2, 3, 4, 5, 13
3, 11, 12
1, 6, 7
73
mean
17.5
18.2
18.6
30.0
31.9
29.3
19.9
34.3
28.2
28.8
dbh
min
9.5
9.2
9.1
10
10
16
9.3
13.3
11.6
14
max
49.7
58.5
45.9
55.7
57.6
53.3
55.6
87.4
61.1
51
Rel.
B.A.
0.099
0.135
0.023
0.015
0.047
0.013
0.114
0.389
0.084
0.081
Table 3.2. Principal results summarized by species. Species are categorized by the soil fertility of sites at which each species is
dominant, ordered (top to bottom) from high-resource moraines through low- resource outwash plains (Table B.2), and by shade
tolerance. The existence and direction of significant correlations of growth to diameter, neighborhood, and/or soil resource effects are
indicated with ‗+‘ or ‗–‘ symbols; if neighbor species identity was important this is denoted with ‗  ‘.
Species
Acsa
Tiam
Fram
Prse
Fagr
Acru
Pogr
Quru
Quve
Qual
Predominant site
type
Moraine
Moraine
Moraine
Int. moraine
Poor moraine
Ice-contact
Ice-contact
Ice-contact
Outwash
Outwash
Shade
tolerance
v. tolerant
tolerant
intolerant
moderate
v. tolerant
moderate
intolerant
intolerant
intolerant
intolerant
Effect and direction of relationship
Neighbor
Diameter Neighborhood
H2O ΣN Ca
species
+
–
+

+
+
–
+
–*
+
+
–
+

+
–
–
+
+
+
–
+
+
–
2
* Relationship supported by strongest r and a support threshold of ΔAICc < 2.5.
74
Table 3.3. Supported models for realized growth framework. The amount of growth variance predicted by complex models (those that
included at least one term in addition to diameter) was typically higher than the variance explained by diameter alone. This
2
2
improvement (Add. var. expl.) represents the difference between diameter-only r and the r of the top supported model. If the
diameter-only model was supported, then it is not presented separately. Modifications (Table B.4) to full models, which are denoted
by the same equation numbering scheme found in the methods, are identified in superscript. General model type (additive or
multiplicative) is indicated in Model column with numbers corresponding to equation numbers in the methods. Similarly,
neighborhood type in the Ngb column corresponds to equation numbers.
2
Best and supported models
Diameter-only models
r
2
Model
Ngb.
Res.
k
n
ΔAICc
Model
Add. var. expl.
ΔAICc
r
acru
1
5
H2O
6
557
0.000
0.348
0.022
13.03
0.326
acsa
2
8
H2O
6
682
0.000
0.635
0.036
70.96
0.599
fagr
fram
fram
fram
fram
pogr
prse
prse
prse
qual
1
1*
2
1
2
2
2
2
2
2
7
5†‡
--5¶
6†
------6
---
H2O
--------ΣN
Ca
6
6
3
4
5
4
4
4
6
4
113
30
30
30
30
84
27
28
27
489
0.000
0.000
1.031
1.436
1.903
0.000
0.000
0.106
2.428
0.000
0.422
0.573
0.383
0.393
0.391
0.239
0.758
0.720
0.799
0.261
0.163
0.190
16.68
0.259
4.57
5.36
0.172
0.672
22.95
0.241
Species
H2O
Ca
H2O
75
0.010
0.008
0.067
0.086
0.048
0.127
0.020
Table 3.3 (cont‘d)
quru
2
--quve
2¶
--tiam
2
--tiam
2
--* Log-normal diameter effect
H2O
----Ca
4
4
3
4
591
184
180
180
0.000
0.000
0.000
0.706
0.623
0.273
0.388
0.392
†β=0
‡ Estimated radius
¶ Equivalent to potential growth
76
0.034
0.004
55.54
0.589
Table 3.4. Site-level correlations of mean individual basal area growth, mean diameter, stem
density, and mortality rate with the soil resource metrics. Sample size was 13 for all
2
relationships. Correlation strength presented as R , with parameter estimates for slope and
intercept, both with 95% confidence intervals.
Res.
Mean BAI
Mean Size
Density
Mortality
0.435
0.322
0.014
0.001
0.546
0.581
0.344
0.297
0.441
0.268
0.026
0.003
0.039
0.047
0.008
0.005
slope
2.18 (0.43, 3.93)
3.10 (-0.072, 6.28)
-24.3 (-166, 117.1)
-0.06 (-1.52, 1.40)
0.37 (0.13, 0.61)
0.10 (0.04, 0.16)
0.31 (0.01, 0.6)
0.46 (-0.038, 0.95)
0.15 (0.03, 0.26)
0.46 (-0.08, 1.00)
-4.97 (-26.6, 16.6)
0.46 (-5.29, 6.20)
-6.46 (-29.1, 16.2)
-0.07 (-0.29, 0.15)
0.01 (-0.05, 0.07)
-0.02 (-0.26, 0.21)
intercept
7.00 (6.10, 7.90)
22.51 (20.88, 24.15)
598 (525, 670)
2.41 (1.66, 3.17)
5.68 (4.20, 7.15)
6.28 (5.23, 7.34)
7.08 (6.10, 8.05)
21.06 (18.02, 24.09)
21.46 (19.44, 23.48)
22.60 (20.89, 24.31)
617 (485, 749)
582 (482, 682)
603 (531, 676)
2.78 (1.44, 4.13)
2.28 (1.26, 3.30)
2.44 (1.69, 3.20)
R
Ca
H2O
ΣN
N min.
Ca
H2O
ΣN
N min.
Ca
H2O
ΣN
N min.
2
77
Figure 3.1. Normalized neighborhood index as a function of neighbor (a) displacement and
(b) diameter. For F. americana, the neighborhood index was defined only across a radius of
7.86 m and was restricted to zero at larger distances.
78
-1
Diameter growth (cm year )
-1
Figure 3.2. Individual growth as linear (equation 3) or lognormal (equation 4) functions of
diameter.
79
-1
-1
-1
NO3 + NH4 (mg kg )
Figure 3.3. Predicted diameter growth as a function of soil resource, holding diameter
constant at the mean value for each species (solid lines with symbols). Diameter held constant
at minimum and maximum diameter values is presented as dashed lines.
80
-1
-1
Figure 3.4. Predicted diameter growth plotted as a function of soil water with
neighborhood and diameter held constant (a, b), and as a function of neighborhood with
soil water and diameter held constant (c, d). Mean values for constant terms are represented
by solid lines with symbols; maximum and minimum values for constant terms are represented
by dashed lines.
81
-1
-1
NO3 + NH4 (mg kg )
-1
Pot. N Min. (mg kg day )
Figure 3.5. Canopy diffuse non-interceptance as a function of each measured soil resources.
Uncertainty quantified as standard error bars for both vertical (diffuse non-interceptance) and
horizontal (soil resource) axes, with diffuse non-interceptance presented on a log scale. Sampling
intensity differed, with 80 – 100 samples for diffuse non-interceptance, 110 samples for soil
water, but less than 80 samples for soil nutrients.
82
-1
-1
Figure 3.6. Predicted diameter growth as a function of neighborhood. For A. saccharum (a),
predicted growth holding diameter constant at its mean value was calculated using minimum
(1%, dotted line), mean (7%, solid line with symbols), and maximum (16% dashed line)
measured soil water values. For F. americana (b), diameter held constant at its minimum and
maximum values is presented as dashed lines, the mean value as the solid line with symbols.
83
CHAPTER IV
CALCIUM, NITROGEN, AND MICRONUTRIENT FERTILIZATION OF INDIVIDUAL
CANOPY TREES IN NORTHERN HARDWOOD FORESTS
84
Abstract
Northern hardwood forest productivity and tree growth in the Manistee National Forest,
Michigan, USA, are correlated with soil calcium (Ca), inorganic nitrogen (N), and water
availability across simultaneous gradients of each resource, yet it remains uncertain which of
these resources is responsible for limiting aboveground net primary productivity (ANPP).
Despite correlations of growth with multiple resources, resource limitation in temperate forest
systems has typically been investigated in the context of N limitation. We initiated a full factorial
fertilization experiment to compare effects of additional soil Ca and N on growth of mature
canopy trees from the four most common species (Acer saccharum, Quercus rubra, Acer
rubrum, and Quercus alba) at these sites. Five fertilizer treatments—Ca, N, Ca + N, Ca + N + P
+ potassium (K) + micronutrients, and control—were applied to approximately equal numbers of
individuals from each species. For a subset of trees, soil resource availability was measured
before and after fertilization to assess effectiveness of the treatments. Diameter growth was
measured over three growing seasons, and leaf production was measured via canopy photographs
after the second growing season. Also at the end of the second season, canopy leaves were
removed to measure chlorophyll fluorescence and leaf nutrient content. As a consequence of
fertilization, content of Ca, magnesium (Mg), N, and P increased in leaves, though not
consistently across species or treatments. Neither diameter growth nor leaf production changed
significantly after fertilization, but non-significant changes in mean basal area increment were
consistent with the prediction that growth should increase in response to Ca at sites with low soil
fertility and in response to N at sites of high soil fertility. Non-significant changes in diameter
growth consistent with predictions suggest that the experiment could produce significant results
with some additional years of fertilization.
85
KEYWORDS: Calcium, fertilizer, leaf nutrients, nitrogen, northern hardwoods, photosynthesis,
resource limitation
86
Introduction
Tree growth and forest productivity in northern temperate hardwood forests are typically limited
by (Magill et al. 2004, van den Driessche et al. 2008) or correlated with (Reich et al. 1997, Joshi
et al. 2003) N availability, leading to the conclusion that most temperate plant ecosystems are
nitrogen (N)-limited (Vitousek & Farrington 1997). Growth is often correlated with other
resources, however, particularly base cations (Demidchik & Maathuis 2007, Fujinuma et al.
2005), suggesting that these elements could also potentially limit growth and productivity. Most
studies concerned with resource limitation in northern hardwood forests have been designed to
test for N limitation (Magill et al. 2004) or demonstrate that phosphorus (P) limitation is less
prevalent than N limitation (Will et al. 2006, Finzi 2009). A variety of studies have tested
whether base cations may limit tree growth (Rosberg et al. 2006, Gradowski & Thomas 2008),
yet relatively few studies (e.g. Bigelow & Canham 2007, Kulmatiski et al. 2007) explicitly test
whether growth is limited by N or by base cations.
A first step in identifying resource limitation is to demonstrate that growth is correlated
with the resource in question; in northern temperate hardwood forests, growth can be related to a
variety of resources (Newman et al. 2006, Mallik et al. 2008). Diameter and height growth are
strongly related to Ca availability across natural gradients (Bailey et al. 2004, Schaberg et al.
2006, Sverdrup et al. 2006), and diminished soil calcium (Ca) content reduces growth and
disease resistance in Acer saccharum on acidic soils (St Clair et al. 2008). Nitrogen availability
is related to aboveground net primary production (ANPP) (Austin & Sala 2002), N
mineralization rates and litter quality (Joshi et al. 2003, Hobbie 1996, Aerts & Chapin 2000), and
community composition and stand development (Hall et al. 2006, Johnson 2006, Mayer 2008).
Production also could be related to availability of other base cations (Wilmot et al. 1996,
87
Rosberg et al. 2006, Sverdrup et al. 2006, Park et al. 2008). Greater availability of magnesium
(Mg) increases photosynthesis (Vizcayno-Soto & Cote 2004), which is the primary physiological
basis for elevating ANPP. Leaf nutrient content may differ in trees suffering from resource
limitation; N:P less than ~14 can be symptomatic of N limitation (Blevins et al. 2005, Donaldson
et al. 2006, Mead et al. 2010), while N:P greater than ~16 can indicate P limitation (Verhoeven
et al. 1996, McGroddy et al. 2008, Soethe et al. 2008). Additionally, higher C:N can be expected
when N is limiting (Vilela et al. 2003).
Growth-soil resource correlations suggest resource limitation (Finzi & Canham 2000),
but growth must also respond to increased resource availability to conclusively demonstrate
resource limitation (Vilela et al. 2003). Fertilization with Ca in the field can elicit positive
growth responses in seedlings (Kobe et al. 2002), saplings (Long et al. 1997, St Clair & Lynch
2005, Juice et al. 2006), and adults in dominant canopy positions (Gradowski & Thomas 2008,
Rosberg et al. 2006). Similarly, fertilization of forest trees with N leads to increased specific leaf
area and nitrogen content, photosynthetic rates, and ultimately ANPP (Henderson & Jose 2005,
Will et al. 2006). Although additional N availability can increase tree growth, excessive N
deposition can actually reduce or even completely halt growth, and can increase mortality
(Magill et al. 2004). Tree growth increases in response to fertilization with other resources (P,
Mg, potassium (K), etc.) both with (Borders et al. 2004) and without simultaneous addition of N
(Will et al. 2006, Rosberg et al. 2006).
For stands in the current study, productivity is more strongly related to extractable soil Ca
than to inorganic N or volumetric water content (Baribault et al. 2010). However, the
relationship of wood production to soil Ca was asymptotic, with rapid increases in wood
production across a relatively small increase in soil Ca and then relatively constant production at
88
higher soil Ca availability (Baribault et al. 2010). An asymptotic relationship of ANPP to
inorganic N was documented across sites of which our sites are a subset (Zak et al. 1989). In
contrast, ANPP increased linearly as a function of soil inorganic N and water based on
measurements from 1999 to 2009 (Baribault et al. 2010). The growth response to resource
availability is additionally complicated by a complete shift in species composition across the
resource gradient, with Quercus alba (Qual) dominant at the lowest-productivity sites, replaced
by Quercus rubra (Quru) and Acer rubrum (Acru) at sites of intermediate fertility, in turn
replaced by A. saccharum (Acsa) on sites with highest resource availability (Host & Pregitzer
1992). Thus, growth of Q. alba at sites with low Ca availability was strongly correlated to soil
Ca, whereas growth of Q. rubra, A. rubrum, and A. saccharum at sites with higher Ca
availability was more strongly correlated with inorganic N. Furthermore, Q. rubra and A. rubrum
are relatively shade intolerant species that typically have higher growth rates than highly shadetolerant A. saccharum (Finzi & Canham 2000).
The change in species dominance may be controlled by a species tradeoff between
survival ability under low-resource conditions and effective competitive ability in high-resource
conditions (Grime 2001). Where soil resource availability is severely limited, species evolve a
resource-conservative, efficient physiology (Canham et al. 1996, Desilets & Houle 2005) that
precludes growth responses to high resource availability (Schreeg et al. 2005). A more
competitive strategy is promoted where resources are abundant (Liancourt et al. 2009), so
species adapted to low-resource conditions are excluded by virtue of their inferior competitive
ability (Sanchez-Gomez et al. 2008). An alternative model to explain community composition
compares species‘ abilities to draw limiting resources down to minimum levels (Tilman 1988).
Species that preempt the supply of a given resource dominate (Craine 2005). For example, A.
89
saccharum may acquire greater quantities of all soil nutrients due to higher fine root production
at sites where resources are abundant (Park et al. 2008).
The purpose of this study was to identify which soil nutrients limit tree growth across a
productivity and species composition gradient in northern hardwood forest in which ANPP has
been shown to correlate with both soil Ca and inorganic N content. We implement a factorial
resource addition of Ca and ammonium (NH4), with an additional treatment that also included P,
K, and micronutrients. Productivity is also correlated with soil volumetric water content in these
stands (Baribault et al. 2010), but we do not conduct water additions because of logistical
constraints. We test three hypotheses concerning how growth of the four most common species,
which constitute more than 80% of individuals at our sites, responds to fertilizer addition:
Hypothesis 1: Fertilization with Ca increases diameter growth and leaf production for Q. alba,
the species dominant at sites with low soil Ca, while growth of A. saccharum, dominant at sites
with high soil Ca, responds to N but not Ca fertilization. Species dominant at sites of
intermediate soil Ca (Q. rubra and A. rubrum) increase growth after fertilization with both Ca
and N.
Hypothesis 2a: The relative magnitude by which diameter growth and leaf production increase
following fertilizer treatment for Q. alba is lower (consistent with the ability to tolerate lowresource conditions) than for A. saccharum (consistent with superior competitive ability in highresource environments). Growth of Q. rubra and A. rubrum increases the most in response to
fertilizer addition, consistent with these species‘ shade intolerance and high potential growth
rates.
Hypothesis 2b: The relative magnitude by which diameter growth and leaf production increase
following fertilizer treatment for Q. alba is higher than for other species, consistent with the
90
interpretation that growth for this species is severely resource-limited and has the greatest
potential to increase.
Hypothesis 3: Diameter growth response to fertilizer addition is mediated by increased leaf
nutrient content (particularly N and Mg), which presumably enhances photosynthesis.
Methods
Study sites
Trees in this study were distributed across 11 mixed hardwood stands of the Manistee National
Forest in the lower peninsula of Michigan, USA (~44º12‘N, ~85º45‘W). Stands were minimallydisturbed, second-growth forests between 80 and 100 years old. Climate is expected to be very
similar across all sites (Schaetzl et al. 2005), with a maximum inter-site distance of 38.1 km and
maximum elevation change of 236 m. Glacial landforms are strongly associated with particular
soil fertility levels and community composition in this region (Table C.1). The range of sites
represents gradients of Ca and inorganic N that increased from outwash plains through icecontact and moraine landforms (Table C.1). Quercus alba was the most common species at
outwash plains, while Quercus rubra and Acer rubrum were common at ice-contact and poor
moraine sites; the highest-fertility moraines were occupied primarily by Acer saccharum
(Baribault & Kobe, unpublished manuscript).
Tree selection and fertilizer application
Trees were selected for fertilizer addition from the four most common species, A. saccharum, Q.
rubra, A. rubrum, and Q. alba, which together represented (~80%) of all individuals in mapped
stands adjacent to the study area (Baribault et al. 2010). Individuals with emergent canopies
exposed to irradiance were selected based on the rationale that they would be more likely to
91
respond to fertilizer addition than would smaller-statured (i.e. shaded) trees (Gradowski &
Thomas 2008). Canopy emergent trees varied in diameter from approximately 20cm to more
than 80cm (Table 1). Trees were selected across as many sites as possible, though relatively few
individuals were available at some sites where species were locally rare (Table C.1).
There were four fertilizer treatments (Ca, N, Ca and N (Ca+N), and Ca, N, and
micronutrients (Ca+N+µ)) and a control for each species (Table 2). Calcium alone was applied
as CaSO4·2H2O (gypsum), while N alone was applied as (NH4)2SO4 (Northstar Minerals,
Okemos, MI). Calcium and N were applied together at the same rates for the Ca+N treatment.
For the Ca+N+µ treatment, P, K, and micronutrients, including Mg, manganese (Mn),
molybdenum (Mo), copper (Cu), iron (Fe), and zinc (Zn)), were applied in addition to Ca and N.
In this treatment, Ca and N were applied at the same rates as in the other treatments, but with
slightly reduced proportion of Ca from gypsum and N from ammonium sulfate. The balance of
Ca was provided by the micronutrient mix (Micromax®, The Scotts Company), while the
balance of N was provided in the form of N-P-K blend fertilizer (The Scotts Company 12-12-12,
Table 2). For each species, control groups were established that contained the same number and
spatial distribution of trees in the treatment groups. Fertilizers were applied using hand-held
2
rotary broadcasters, evenly distributing the material in a ~4-m radius (covering 50 m )
surrounding each tree. Half of each treatment was applied in mid-May and the remaining half in
-1
mid-June of 2008, 2009, and 2010. Only light precipitation (< 1 cm day ) followed (within four
days) most treatments, although some heavier precipitation occurred during the June 2009
application.
92
Soil resource measurements
To demonstrate that the fertilizer addition had the intended effect of increasing the target soil
resources (Fig. 1, Table C.2) without altering pH (Table C.3), soil samples were collected in May
2008 (immediately prior to the first fertilizer addition) and again in October 2008, after the first
year of fertilization. Samples were taken from 33 A. saccharum, 31 Q. rubra, 26 A. rubrum, and
39 Q. alba trees, with approximately equal representation for each treatment. Three soil cores (5
cm diameter, 20 cm depth) were collected at equidistant points 2 m from the base of each tree at
each sampling date and thoroughly mixed into a composite sample. All nutrient analyses were
conducted on composite samples, which were air dried prior to analysis.
Potential net N mineralization was determined by aerobic incubation. Initial
+
-
concentrations of NH4 and NO3 were measured in a 6-g subsample of each composite. Soils
+
-
were shaken for two hours in 2M KCl to extract inorganic N, and then [NH4 ] and [NO3 ] were
measured with a colorimetric assay using a fluorescent plate reader (ELx808 Absorbance
Microplate Reader, BioTek Instruments, Inc, Winooski, VT). A second set of 6-g subsamples
was incubated aerobically in darkness at 25°C for 28 days. Constant moisture content was
maintained by periodic addition of deionized water, and post-incubation concentrations of NH4
-
and NO3 were measured by the same method. Potential net N mineralization was calculated as
+
-
the difference between final and initial inorganic N [NH4 + NO3 ].
Base cations were extracted from 4-g subsamples shaken for 15 minutes in an acidic
2+
2+
+
-
Mehlich III solution (Carter 1993). Concentrations of Ca , Mg , K , and PO4 in extracts
93
+
were measured with inductively coupled plasma atomic emission spectrometry (Optima 2100DV
ICP Optical Emission Spectrometer, Perkin-Elmer, Shelton, CT).
Growth measurements
Diameters of all trees were measured at 1.37 m height using diameter tapes in spring 2008
(before leaves emerged), fall 2009, and fall 2010 (both after leaves started to senesce). For 24
individuals, diameter could not be measured at 1.37 m due to obstructions, and was thus
measured at the lowest possible height above the 1.37 m point. Diameter growth increment was
calculated as the quotient of the change in diameter and the measurement interval—both the 2year and 3-year intervals were used for analysis. Several trees showed slightly negative growth
rates (as low as -0.2cm) due to bark loss or developing deformities; for these individuals, growth
was rounded to zero. Data from five trees were unusable due to mistaken fertilizer application or
mortality and were consequently excluded from the dataset.
Leaf nutrient content and chlorophyll fluorescence
Leaf tissue samples (up to three leaves per tree) were removed in early September 2009 before
annual senescence had started. Leaves as close as possible to the exposed top of the canopy were
selected, and brought to the ground using a shotgun; chlorophyll fluorescence yields (ΔF/Fm)
and maxima (Fv/Fm) were measured immediately in the field using a photosynthesis yield
analyzer (MINI-PAM, Walz, Effeltrich, Germany), with three readings taken at different points
on each leaf (Table C.4). Samples were then oven-dried at 60°C for three days and stored for up
to 4 months prior to analysis of nutrient concentrations. Dried leaves were ground into powder
using a ball-bearing mill (Kleco, Visalia, CA, USA). Sub-samples of approximately 0.5g were
dissolved in 5mL concentrated HNO3 for 24 hours, then heated in a block digester (AIM600
94
Block Digestion System, A.I. Scientific Pty Ltd, QLD, Australia) for 8 hours at temperatures up
to 180°C. Any liquid volume loss was adjusted back to 5mL with additional concentrated HNO3,
after which extracts were diluted to 100mL with ddH2O, and Ca, K, Mg, and P contents were
measured using ICP spectrometry (Table C.5). Separate sub-samples (0.1 - 0.2 mg) were used to
measure total C and N content (Table C.6). These samples were analyzed by dry combustion and
gas chromatography (Costech, ECS 4010 CHNSO analyzer, Valencia CA, USA), with elemental
content checked against an apple-leaf standard (NIST).
Canopy openness
Percent canopy openness was assessed from photographs of individual tree crowns. Photographs
were taken 1.5m above the ground and 2.5m from the north and south sides of each tree using a
EOS 5D Mark II (Canon, USA) and a diagonal fisheye lens (15mm F2.8 EX DG, Sigma,
Ronkonkoma, NY, USA), with the top of the lens oriented north for every photograph. The
Sidelook software package (http://www.appleco.ch/) was used to transform the color
photographs into black and white images (Nobis & Hunziker 2005). The software package GLA
(http://www.ecostudies.org/gla/) was then used to estimate canopy openness for the portion of
each photograph that corresponded to the tree crown. Two estimates of percent canopy openness
were calculated for each photograph to reduce the amount of error associated with manually
differentiating individual tree canopies from surrounding foliage. A single value of percent
canopy openness for each tree was calculated as the average of the two estimates from the northand south-side photographs of each individual (Table C.7).
Analysis
All data analysis was conducted in R (The R Foundation for Statistical Computing, 2010,
http://cran.r-project.org/). One-way analysis of variance (ANOVA) was used to identify whether
95
mean values of potential response variables differed between treatments, including addition of
-1
Ca; N; Ca+N; Ca+N+µ, and a control. Response variables included diameter growth (cm yr ),
2
-1
basal area increment (BAI, cm yr ), leaf nutrient content (Ca, K, Mg, P, N, C, C:N),
chlorophyll fluorescence (mean and maximum), and percent canopy openness. In the case of a
significantly different mean value (p < 0.05), Tukey‘s HSD test was applied to identify which
treatment group(s) differed. Data from the micronutrient treatment were excluded, and the
remaining dataset analyzed using two-way ANOVA to check for interaction effects between Ca
and N; the same response variables were tested in this factorial analysis. Again, Tukey‘s HSD
test was used to identify any differences among treatment means. Further tests for potential
differences in growth responses (diameter and basal area) among treatments were conducted
using linear regression. In simple linear regressions, BAI was modeled as functions of initial
diameter, leaf nutrient contents, chlorophyll fluorescence, and percent canopy openness. In
multiple linear regressions, diameter growth was modeled as functions of initial diameter and
one of the aforementioned predictor variables. A treatment was considered to be significantly
different from the control group if there was < 29% overlap in the 95% confidence intervals of
the slope parameter(s) of the treatment and control groups (Austin & Hux 2002, Kobe et al.
2010). Growth and / or BAI were significantly related to predictor variables in several instances,
but reasonable biological interpretations of these relationships were difficult to justify. Thus, we
present these results as supplementary material and do not discuss them further.
96
Results
Fertilizer addition changed content of added resources in soil
For the base cations and P, soil content of the resource in question typically increased following
the fertilizer application. Soil Ca content increased substantially with addition of Ca and Ca+N,
-1
increased modestly upon addition of Ca+N+µ and N, and decreased substantially (40.6 mg kg )
in the control groups (Fig. 1, Table C.2). Soil K content increased in the N, Ca+N, and Ca+N+µ
-1
treatments (between 1.6 and 2.6 mg kg ), but decreased by similar magnitude for the Ca
treatment and control groups (Fig. 1). Soil Mg content increased only following the Ca+N+µ
-1
-1
(6.8 mg kg ), with decreases between 1.3 and 12.0 mg kg in other treatments and the controls
-1
(Fig. 1). Similarly, soil P content increased only in Ca+N+µ treatment (1.99 mg kg , Fig. 1). In
contrast to P and the base cations, soil inorganic N content typically decreased after treatment
with N. For all treatments, including the control, soil NH4 decreased between 3.7 and 17.1 mg
-1
-1
kg ; soil NO3 decreased between 3.5 and 26.8 mg kg in all treatments except the control,
-1
where NO3 increased trivially (0.31 mg kg , Fig. 1). The fertilizer treatments did not affect soil
pH—the maximum pH increase was 0.26 and the maximum decrease 0.12 (Table C.3).
Some leaf nutrient contents responded to fertilizer addition
Several leaf nutrients differed between fertilizer treatments, though these differences were not
always intuitive or explicable. Based on one-way ANOVA, in A. saccharum, leaf Mg content
was significantly higher (p < 0.05) following addition of Ca, N, or both resources in comparison
to addition of Ca, N, and micronutrients (including Mg) or to the control (Table 3). For A.
saccharum, leaf P content increased significantly with addition of Ca (p < 0.05), while for Q.
97
alba, leaf P content decreased significantly (p < 0.01) with addition of Ca + N (Table 3). In A.
rubrum, leaf N:P was significantly greater than in the control (15.6 vs. 11.0, p < 0.01) with
addition of N, suggesting the onset of P limitation following N amendment. In contrast, leaf N:P
in Q. alba was significantly greater than the control (23.9 vs. 15.4, p < 0.001) after addition of
Ca + N (Table 3); both values were consistent with P limitation. Other leaf nutrient ratios (e.g.
C:N, N:Ca) were not significantly different after fertilizer treatment for any species based on
one-way ANOVA.
Based on two-way ANOVA, an interactive effect of Ca and N was significant (p < 0.01)
for leaf Ca content in Q. rubra (Table 4). Leaf N content was significantly higher in treatments
with N (p < 0.05) for A. rubrum. Leaf P content significantly increased with addition of Ca in A.
saccharum (p < 0.005) and N in A. rubrum (p < 0.05), while addition of N was associated with
significant decrease of leaf P content in Q. alba (p < 0.0005, Table 4). Consistent with the third
hypothesis, leaf C:N significantly decreased with addition of N (p < 0.05) for Q. alba (Table 4).
Consistent with the one-way ANOVA, two-way ANOVA showed that leaf N:P increased in both
A. rubrum (p < 0.001) and Q. alba (p < 0.00001) following N and Ca+N treatments, respectively
(Table 4).
Mean growth increased non-significantly following some treatments
For no species were there significant differences in diameter growth or basal area increment
following the fertilizer treatment. Similarly, no species showed significantly increased leaf
production, although a severe outbreak of forest tent caterpillar (Malacosma disstria) completely
defoliated all A. saccharum and most Q. rubra and A. rubrum trees in 2008 and 2009. A leaf
production response may have been impossible to detect given such an extreme event; indeed,
defoliation may have substantially reduced diameter growth for these species. Considering
98
strictly the mean diameter growth or BAI, without accounting for significance levels, results
were largely consistent with the prediction from hypothesis 1, which predicted that Ca should
elicit the largest effects in species dominant at low-Ca sites. For Q. alba, treatments that
contained Ca (Ca, Ca+N, Ca+N+µ) had non-significantly (NS) higher mean basal area increment
2
-1
(10.6, 10.4, and 11.0 cm year , respectively) than did treatments that did not contain Ca (N,
2
-1
control; 9.9, 9.3 cm year , respectively, Table 1). For Q. rubra, the mean of the Ca treatment
2
-1
2
-1
(28.1 cm year ) was higher (NS) than the control mean (27.0 cm year , Table 1). For A.
rubrum, treatment means for Ca, N, and Ca+N+µ were higher (NS) than the control (18.0, 14.8,
2
-1
15.2 versus 12.3 cm year , Table 1). For A. saccharum the Ca and Ca+N+µ treatment means
2
-1
2
-1
(5.0 and 4.5 cm year ) were lower (NS) than the control (5.4 cm year ), consistent with the
prediction that Ca should be less important at higher fertility (Table 1). The N and Ca+N
2
-1
treatment means (7.1 and 5.9 cm year ), but not the Ca+N+µ treatment mean, were higher (NS)
than the control. These results were mostly consistent with the prediction that N availability
should be of greater consequence at higher fertility (Table 1). Finally, there were no significant
differences in mean diameter growth or BAI comparing treatments with Ca (Ca, Ca+N,
Ca+N+µ) to those without Ca (N, control) or treatments with N (N, Ca+N, Ca+N+µ) to those
without (Ca, control, Table 8).
Non-significant growth changes associated with fertilizer addition were consistent with
tolerance-competition tradeoff
Although there were no significant differences in diameter growth or BAI among treatments, the
magnitude by which growth increased following fertilization was consistent with a tradeoff
99
between ability to tolerate low-resource conditions (but inability to respond to resource pulses)
versus superior competitive ability in higher-resource environments (hypothesis 2a). Growth of
Q. alba, which was dominant at low-fertility sites, increased (NS) 13% over the control after
fertilization with Ca, and after fertilization with Ca+N+µ, growth increased (NS) 18% (Table 1).
In contrast, growth of A. saccharum—dominant at high-fertility sites—increased (NS) 31% after
fertilization with N (Table 1). Of the species more common at sites of intermediate fertility, A.
rubrum demonstrated the largest increase (NS) in growth (47%) following fertilization with Ca,
while Q. rubra showed the smallest increase (NS) in growth (4%) after fertilization with Ca
(Table 1). These results were partially consistent with hypothesis 2a: Q. alba growth increased
(NS) proportionately less after fertilizer addition (Ca) than A. saccharum (N), but maximum
growth responses (NS) of Q. rubra (Ca) and A. rubrum (Ca) were highly dissimilar. It may be
premature, however, to discount support for our hypotheses based on measurements for Q. rubra
because of the substantial defoliation experienced by most individuals of these species. It is
possible that Q. rubra may respond strongly to fertilizer addition in future years when trees are
not defoliated. In contrast, results did not support hypothesis 2b, which predicted that Q. alba
should show the strongest proportional response to fertilizer addition. To test whether site
fertility level influenced growth in substantially different ways at low versus high fertility, we
partitioned the data into high- and low-fertility sites for each species and recalculated the 1-way
ANOVA. Neither among the low- nor high-fertility sites did mean growth rates or leaf nutrient
contents differ between treatments.
Leaf nutrient content, but not chlorophyll fluorescence, increased with fertilization
Our third hypothesis predicted that higher photosynthetic activity (measured by proxy as
chlorophyll fluorescence) resulting from increased leaf nutrient content should be the basis for
100
diameter growth response to fertilization. Results did not support this hypothesis—although leaf
nutrients increased in some instances following fertilizer application (Tables 3, 4) there was no
accompanying increase in chlorophyll fluorescence.
Discussion
Resource limitation across soil fertility gradients
Fertilizer addition increased leaf nutrient content in several treatments and changed soil resource
content across all species, yet we could conclude neither that the fertilizer treatments increased
soil resource availability nor that tree physiology responded to the treatments. Soil resource
availability was measured as standing pools, which experience relatively rapid turnover and are
depleted by plant and microbial uptake (Leduc & Rothstein 2010) and leaching (Bigelow &
Canham 2007). In addition, we did not account for potential secondary effects of base cation
addition, such as decreasing inorganic N availability (Groffman et al. 2006). Our hypotheses
were concerned with changes in diameter growth and leaf production, but significant differences
in these parameters were not observed, lending no support to the hypotheses. Hereafter, we
explore each hypothesis based on non-significant differences between treatments and controls,
showing that the results, though not statistically significant, were consistent with several
predictions.
Across plant ecosystems, the resources that limit productivity, including soil nutrients,
irradiance, and water are correlated with growth (Knox et al. 1995, Reich et al. 1997), yet
resources that do not limit growth are often correlated nonetheless (Joshi et al. 2003, Mayer
2008). In previous studies, we identified correlations of ANPP and diameter growth to several
soil resources, notably Ca, N, and soil water (Baribault et al. 2010). These results were consistent
101
with resource limitation, but because correlative evidence is not sufficient to demonstrate
limitation, it was necessary to manipulate resource availability (Will et al. 2006, Finzi 2009). To
reiterate, mean growth following fertilizer addition did not differ significantly from control
groups for any species, but non-significant differences were consistent with the potential for
resource limitation identified by correlative evidence, in accordance with our first hypothesis. At
sites with low soil Ca, where ANPP was strongly correlated with extractable Ca, Q. alba was the
dominant species. Following fertilizer addition, the largest non-significant increases in BAI for
Q. alba occurred in treatments that included Ca. In contrast, at sites with high availability of both
Ca and N where ANPP was correlated with N but not Ca, A. saccharum was the dominant
species. Only fertilization with N alone increased BAI substantially (not significantly) above the
control.
With our second hypothesis, we sought to address the nature of plant competition across
resource gradients by assessing predictions of either of two prominent theories (Tilman 1988,
Grime 2001). If competitive ability and growth rate plasticity must be compromised in order to
tolerate low-resource conditions (MacDougall & Turkington 2004), then plants growing at low
soil fertility should respond weakly to fertilizer addition (Schreeg et al. 2005). Alternately, if
plants occupy locations based on their ability to exploit low levels of available resources, the
magnitude by which growth increases in response to fertilizer addition should be greatest for
species occupying principally low-resource sites (Craine 2005). Although our results did not
show significant mean differences in diameter growth, basal area increment, or leaf production
following fertilizer treatment, the non-significant BAI changes were most consistent with the
first case—that is, a tradeoff between tolerance of low-resource conditions and ability to
effectively compete when resources are abundant (Weigelt & Jolliffe 2003, Gaucherand et al.
102
2006). The magnitude by which BAI increased in Q. alba following fertilization with Ca was
considerably lower than the increase observed in A. saccharum after fertilization with N. A.
saccharum, the species dominant at sites with the highest soil resource availability, demonstrated
a larger response to fertilization consistent with superior competitive ability and the capacity to
efficiently exploit resources (Davis et al. 1999). In contrast, Q. alba, dominant at sites with low
resources, failed to substantially increase growth rates even when resource limitation was
partially alleviated by fertilizer addition.
Despite observing significantly elevated content of several leaf nutrients after at least one
fertilizer treatment for each species, our third hypothesis was not supported because these
changes were not associated with significant increases in either chlorophyll fluorescence or
diameter growth. Two resources in particular, N and Mg, can lead to higher photosynthesis
because they are constituents of RuBisCO and chlorophyll, with more of both molecules required
to increase carbon fixation (Burselm et al. 1995, Walters & Reich 1997). Higher leaf N was
observed only in A. rubrum fertilized by ammonium sulfate, and higher leaf Mg occurred only in
A. saccharum, though after fertilization with Ca, not Ca+N+µ. In contrast to single nutrients,
ratios between different leaf nutrients, notably C:N and N:P, may reveal resource limitation in
the absence of significant growth responses (Blevins et al. 2005). A significant decrease in C:N
for Q. alba following N fertilization is consistent with reduced N limitation (Vilela et al. 2003).
Further supporting lessened N limitation, N:P in Q. alba increased to 23.6 following the Ca+N
treatment; N:P < 14 may correspond with N limitation, while N:P > 16 suggests P limitation
(Koerselman & Meuleman 1996, Romme et al. 2009). In A. rubrum N:P increased from 11.0 in
the control to 15.6 after fertilization with N—though not with Ca+N or Ca+N+µ—again
consistent with a shift away from N limitation (Blevins et al. 2005).
103
Overall, however, we cannot conclude that fertilizer application predictably increased
leaf nutrient content, and must also consider the possibility that fertilization could have
decreased N availability in all treatments and Mg availability in all treatments except Ca+N+µ
(Fig. 1). For example, adding ammonium nitrate to leaf litter can decrease inorganic N
availability because nitrification by litter microbes causes NO3 retention in the litter layer,
leading to net N immobilization and N2O production (Raat et al. 2010). To identical effect,
increased N immobilization is associated with litter from Quercus species (Piatek et al. 2009),
which were dominant from outwash through intermediate moraine sites. Extractable soil Ca, in
contrast, is positively associated with inorganic N availability (Page & Mitchell 2008),
suggesting that effects of simultaneous N and Ca addition could cancel one another. Moreover,
leaf nutrient content may be partially decoupled from soil resource availability for species with
low leaf nutrient content but high resource use efficiency (Rodenkirchen et al. 2009) and
resorption rates (Cote et al. 2002). Finally, nutrient availability could more immediately affect
fine root dynamics, including growth (Park et al. 2008) and senescence (Kunkle et al. 2009),
with changes in leaf nutrients arising in the longer term (Nilsen & Abrahamsen 2003).
Factors contributing to non-significant changes in diameter growth and leaf production
Among several potential reasons for lack of a significant growth response to fertilizer addition,
duration of the study likely had considerable consequences. Many fertilizer application studies
proceed for at least three years, even in fast-growing tropical plantations (Xu et al. 2002), and up
to a decade or longer in temperate silviculture (Neilsen & Lynch 1998, Bauer et al. 2004). For
example, Tsuga heterophylla and Thuja plicata plantations fertilized with N and P upon
establishment grow faster after a decade of fertilizer addition but growth response continues to
increase after 15 years (Blevins et al. 2006), demonstrating that fertilizer effects accrue over the
104
long term. In contrast, shorter-term response to fertilizer in mature stands can be restricted to
increased leaf, fine root, and seed production, with no measurable diameter growth (Davis et al.
2004).
Limited sample size (maximum 25 trees per treatment, minimum eight) and the diameter
of trees selected could have influenced the power to detect significant increases in growth and
the physiological capacity of trees to respond to fertilizer addition, respectively. Despite being
represented by trees with the smallest mean diameter and a size distribution skewed toward
smaller individuals (Fig. 2), A. rubrum showed BAI greater than A. saccharum and Q. alba, and
diameter growth nearly equivalent to Q. rubra, even given the much larger mean diameter of that
species (Table 1). Our rationale that larger trees are best situated to take advantage of increased
leaf nutrient content by virtue of their canopy position and irradiance interceptance may not be
justified (Gradowski & Thomas 2008, Jones et al. 2009). If smaller individuals, even those with
crowns mostly below the main canopy level, are nonetheless capable of responding to fertilizer
input (Rakhteenko 1981, Carlson et al. 2008), it may be necessary to add such individuals to the
study. On the other hand, A. rubrum typically grows more quickly across all size classes than do
more shade tolerant species (e.g. A. saccharum) or species characteristic of low-resource sites
(e.g. Q. alba) (TerMikaelian & Korzukhin 1997), so smaller trees of the other three species may
not be capable of achieving these higher growth rates, with or without the advantage of fertilizer.
Within species, differences in diameter distributions between treatments and
corresponding controls may have affected observed non-significant differences in BAI. For
example, the largest BAI in A. saccharum occurred in the N treatment, but mean diameter of
trees in this treatment was larger (39.4 cm) than in the control (36.1 cm), potentially accounting
for the faster growth rate (Table 1). Larger diameters in treatment groups versus the control may
105
also account for larger BAI following fertilizer addition for Q. alba (Table 1). In contrast, larger
BAI was observed in Q. rubra and A. rubrum for some treatments (e.g. N, Ca+N+µ for A.
rubrum, Ca for Q. rubra) despite smaller mean diameter in the treatment group than in the
control. Accounting for this effect of initial tree diameter may be accomplished either by analysis
of co-variance or by linear regression using diameter as a predictor; given the non-significant
differences observed in the current data, such analysis would be more useful for a future dataset
containing additional years of tree growth.
A final issue that could have affected the experiment is scale at which soil resources are
heterogeneous. In many forest systems, soil resources vary considerably over short distances,
such that trees separated by only a few meters can experience substantially different levels of e.g.
soil Ca or inorganic N (Holste et al. 2010). For example, at site 8 where mean soil Ca content in
-1
an adjacent mapped stand was 90 mg kg , one A. saccharum individual was growing on soils
-1
-1
with Ca content of 65 mg kg , while another grew at 135 mg kg . At site 10, mean adjacent soil
-1
Ca content was 1160 mg kg , but soils at individual A. saccharum trees varied from 945-1230
-1
-1
mg kg . Fertilization increased soil Ca content by an average of 13 mg kg , potentially a 20%
increase for A. saccharum at site 8 but a trivial 1.4% increase for trees at site 10. Thus, the
influence of fertilization could substantially differ both within and across sites, reducing the
likelihood of detecting effects in the aggregate, and necessitating a regression-based approach
that accounts for resource availability for every individual. A second possibility is that the soil
resource environment experienced by individual plants is unrelated to the resource content in soil
directly contacting the organism, as can occur with extensive common mycorrhizal networks
(Simard 2009). Both ectomycorrhizal and arbuscular mycorrhizal fungi are capable of integrating
106
into networks with large spatial extent that could effectively homogenize nutrient availability for
all trees connected to the network (van der Heijden & Horton 2009).
Conclusions
Application of Ca, N, P, K, and micronutrient fertilizers to individual forest trees increased the
content of most of these elements in surrounding soils, with the notable exception of N (NH4 and
NO3), availability of which decreased after nearly all treatments. Diameter growth and leaf
production did not significantly increase after three years of fertilizer addition. Nevertheless, the
non-significant changes in basal area increment followed our prediction that tree growth is
related to soil Ca at sites with low Ca availability but to inorganic N at sites with high soil Ca
and N content (H1). In addition, the non-significant basal area changes were consistent with
theoretical predictions that tree growth after fertilization should increase most in trees occupying
high-fertility sites, while trees tolerant of low-resource conditions should fail to respond as
vigorously to increased resource availability (H2). Results did not support our prediction that
diameter growth should increase with greater photosynthetic output made possible by higher leaf
nutrient content, despite elevated leaf nutrients following some treatments (H3). Non-significant
changes in growth were consistent with predictions from two of our hypotheses after only three
years of fertilizer application, portending significant changes in growth. Continuation of this
study, perhaps expanded to include smaller trees, will be an important step toward resolving
issues of nutrient limitation and plant competition across strong soil resource gradients.
107
Acknowledgements
Indispensable field assistance was provided by M. Erickson, D. Minor, E. Holste, J. Bramer, A.
Pierce, and A. Stinson. We thank M. Walters, D. Rothstein, and S. Grandy for their
recommendations during the planning of this experiment and for comments on the manuscript.
This research received support from the Michigan Agriculture Experiment Station (NRSP-3,
National Atmospheric Deposition Program) and NSF (DEB 0958943).
108
Table 4.1. Basic statistics—including sample size, diameter, annual diameter growth, and basal
area increment (± standard error)—for five treatment groups for each of four common species.
Treatment
Ca
N
Ca, N
Ca, N, µ
Control
A. saccharum
N
21
21
18
16
17
Q. rubra
A. rubrum
Ca
N
Ca, N
Ca, N, µ
Control
37.60 ± 2.10
39.41 ± 2.58
37.88 ± 1.98
35.75 ± 1.99
36.13 ± 2.35
24
17
18
16
22
15
19
8
20
8
Diameter (2007, cm)
49.33 ± 2.39 34.17 ± 1.99
48.98 ± 3.34 30.89 ± 1.66
45.83 ± 2.16 30.53 ± 1.43
45.74 ± 2.27 29.73 ± 2.00
49.96 ± 2.54 31.18 ± 2.75
0.09 ± 0.02
0.11 ± 0.02
0.10 ± 0.02
0.08 ± 0.01
0.09 ± 0.02
Diam. growth (cm year )
0.34 ± 0.03 0.30 ± 0.04
0.33 ± 0.02 0.29 ± 0.03
0.35 ± 0.01 0.23 ± 0.04
0.32 ± 0.03 0.30 ± 0.06
0.33 ± 0.02 0.25 ± 0.05
4.96 ± 1.24
7.06 ± 1.21
5.85 ± 1.17
4.52 ± 0.72
5.41 ± 1.06
BAI (cm
28.06 ± 3.66
26.20 ± 3.27
25.91 ± 1.85
24.77 ± 3.38
27.00 ± 2.61
Q. alba
24
25
23
18
21
32.42 ± 1.75
30.84 ± 1.73
32.99 ± 1.78
34.11 ± 2.44
31.80 ± 1.62
-1
Ca
N
Ca, N
Ca, N, µ
Control
Ca
N
Ca, N
Ca, N, µ
Control
2
0.20 ± 0.02
0.20 ± 0.01
0.20 ± 0.02
0.21 ± 0.02
0.18 ± 0.01
-1
year )
18.01 ± 3.47
14.81 ± 2.14
11.82 ± 2.89
15.22 ± 3.78
12.27 ± 2.47
109
10.6 ± 0.99
9.88 ± 0.82
10.4 ± 1.18
11.0 ± 1.00
9.32 ± 1.06
Table 4.2. Fertilizer treatments and application rates by fertilizer compound (top) and by
element (bottom).
-1
-1
Application rate (kg ha yr )
Ca, N,
Ca
N
Ca, N
µ
Fertilizer
compound
a
CaSO4
600
b
(NH4)2SO4
380
600
562
380
95
c
N-P-K
Micromax®,
d
(Scotts)
665
140
Element
Ca
139.6
0
139.6
139.2
N
0
106.4
106.4
106.4
a
CaSO4 ·2H2O (gypsum) contains 23.28% Ca by mass.
b
c
d
(NH4)2SO4 contains 21% N by mass.
N-P-K contained, by mass, 12% of each element.
Micromax (Scotts) contained, by mass, 0.1% Boron, 6.0% Ca, 1.0% Copper, 17% Iron, 3.0%
Mg, 2.5% Mn, 0.05% Molybdenum, 1.0% Zinc.
110
Table 4.3. Results from one-way ANOVA comparing mean values of response variables across treatments for each species. Mean
values reported only when at least one treatment category showed a difference from the others.
Species
Response
-1
Leaf Mg (mg g )
-1
Ca
A. saccharum 1.90 ± 0.18
b
Leaf P (mg g )
A. saccharum 1.74 ± 0.23
1.48 ± 0.06
Leaf N:P
Q. alba2
A. rubrum
Q. alba
15.3 ± 0.72
1
Treatment group mean
Ca, N
Ca, N, µ
N
b
b
11.9 ± 0.75
a
a
1.53 ± 0.11
a,b
1.11 ± 0.09
1.26 ± 0.09
a
a,b
15.6 ± 1.28
21.2 ± 2.06
1
b
a, b
Mean values with different letters are significantly different, p < 0.05
111
1.51 ± 0.06
1.52 ± 0.16
a,b
a,b
1.12 ± 0.09
15.0 ± 1.12
a
a, b
23.9 ± 2.37
b
Control
1.36 ± 0.06
1.66 ± 0.19
1.40 ± 0.10
a
a,b
a,b
a, b
12.0 ± 0.80
19.1 ± 2.17
a, b
a
1.37 ± 0.14
1.22 ± 0.12
a,b
1.47 ± 0.06
b
a
11.0 ± 1.02
15.4 ± 0.59
a
Table 4.4. Results from two-way ANOVA comparing mean values response variables across Ca, N, Ca+N, and control treatments for
each species. Mean values reported only when at least one treatment category or interaction effect showed a difference from the
others.
1
Response
-1
Leaf Ca (mg g )
Leaf N (%)
-1
Leaf P (mg g )
Leaf C:N
Leaf N:P
1
Species
Difference
Q. rubra
A. rubrum
Treatment group mean
Ca, N
Ca
N
Ca x N
N
5.86 ± 0.21
1.59 ± 0.05
5.98 ± 0.25
1.83 ± 0.11
A. saccharum
A. rubrum
Q. alba
Q. alba
A. rubrum
Ca
N
N
N
N
1.74 ± 0.23
1.40 ± 0.09
Q. alba
N
15.3 ± 0.72
b
b
1.48 ± 0.06
22.30 ± 0.5
a
12.0 ± 0.75
a
a,b
1.11 ± 0.09
1.24 ± 0.08
a,b
1.26 ± 0.09
20.98 ± 0.5
15.6 ± 1.28
21.2 ± 2.06
b
a,b
5.29 ± 0.22
1.76 ± 0.59
a,b
1.52 ± 0.16
1.27 ± 0.10
a
1.12 ± 0.09
21.27 ± 0.5
15.0 ± 1.12
a,b
23.9 ± 2.37
b
Mean values with different letters are significantly different based on Tukey‘s post-hoc test, p < 0.05
112
Control
Interaction
5.20 ± 0.30
1.63 ± 0.05
p < 0.05
NS
a,b
1.22 ± 0.12
1.54 ± 0.13
b
1.47 ± 0.06
22.04 ± 0.5
11.0 ± 1.02
15.4 ± 0.59
a
a
NS
NS
NS
NS
NS
NS
ΔK (mg kg )
-1
-1
ΔCa (mg kg )
-1
ΔNO3 (mg kg )
-1
ΔP (mg kg )
-1
ΔMg (mg kg )
-1
ΔNH4 (mg kg )
Ca
N
Ca+N
All
Ctr
Ca
Treatment
N
Ca+N
All
Ctr
Treatment
Figure 4.1. Change in soil resource content aggregated across all species before (April
2008) and after (October 2008) fertilizer application. Vertical bars represent standard error.
113
Density
Diameter (cm)
Figure 4.2. Diameter distributions for each species. A. rubrum (mean 31.6 cm ± standard
deviation 6.9 cm) was represented by substantially smaller individuals than A. saccharum (37.5 ±
9.6 cm) and Q. rubra (48.0 ± 11.4 cm), and by marginally smaller individuals than Q. alba (32.3
± 8.6 cm).
114
CHAPTER V
TROPICAL TREE GROWTH IS CORRELATED WITH SOIL PHOSPHORUS,
POTASSIUM, AND CALCIUM, BUT LEGUMES MAY ESCAPE LIMITATION
115
Abstract
1. Tropical forest productivity, a significant component of the terrestrial carbon cycle, is widely
assumed to be limited by soil phosphorus (P) availability. Biogeochemical processes that deplete
soil P also deplete calcium (Ca), potassium (K), and magnesium (Mg), yet it remains unclear
whether tropical tree growth is more strongly related to soil P or base cation availability.
2. We hypothesized that: A) tree growth is positively related to soil base cation and P availability
and negatively related to local competition; B) growth of Fabaceae is weakly correlated with soil
resources due to advantages of N2 fixation; C) growth of species with low wood density is more
strongly related to soil resource availability than species with high wood density.
3. Diameter growth and soil resource availability were measured in five mapped stands situated
across natural soil resource gradients in lowland wet tropical forest (La Selva Biological Station,
Costa Rica). Soil resource availability was estimated for each tree using a spatial hierarchical
Bayesian method. Individual tree growth was modeled as a function of diameter, local
neighborhood, and soil resources. Separately, site-level mean tree growth and resource
availability were modeled using linear regression.
4. Individual diameter growth and site mean basal area increment (BAI) correlated with soil base
cations and P, but rarely with N; individual growth correlated negatively with neighborhood
index for only three species. Growth of Fabaceae was unrelated to soil resources both at
individual and site levels. When species were categorized by wood density, growth was related
to soil P or base cation availability.
5. Synthesis. Several resources could potentially limit tropical forest productivity as tree growth
was similarly correlated with soil P and base cations, but tree functional groups responded
differently. Members of Fabaceae showed negligible growth relationships with soil resources,
116
suggesting that N fixation may indirectly alleviate mineral nutrient limitations. The growth of
(non-legume) species with low wood density was strongly correlated with soil P, whereas growth
of species with higher-density wood was correlated with base cations. Thus, tropical tree growth
may be limited by base cations and/or P, with the degree and type of resource limitation
dependent on tree functional group.
KEYWORDS: Base cations, legumes / Fabaceae, functional groups, neighborhood model, plantsoil interactions, phosphorus limitation, quantitative trait, resource limitation
117
Introduction
Tropical forest productivity is a significant component of terrestrial carbon cycling (Luyssaert et
al. 2007, Sitch et al. 2008), but accurately predicting tree growth in tropical forests is
complicated by fine-scale biogeochemical heterogeneity (Townsend et al. 2008, Holste et al.)
and species diversity (Clark et al. 1999). Productivity in tropical forests may be limited by soil
phosphorus (P) availability (Porder et al. 2007, Vitousek et al. 2010), nitrogen (N) (Vitousek &
Farrington 1997, Harrington et al. 2001, LeBauer & Treseder 2008), base cations (calcium (Ca),
potassium (K), magnesium (Mg)) (Andersen et al. 2010), water (Valdespino et al. 2009), or
irradiance (Graham et al. 2003). In wet tropical forests in particular, studies often focus on
whether soil P dynamics limit aboveground net primary productivity (Tanner et al. 1998), while
the possibility that other soil resources influence growth has not been eliminated (Schuur &
Matson 2001).
Evaluating a broad range of potentially limiting resources is important for understanding
resource-based productivity constraints, particularly because macronutrients critical to plant
physiology are equally likely to be lost from ecosystems (Asner et al. 2001, Corre et al. 2010,
Wood et al. 2005). Soil weathering and high rates of nutrient leaching can deplete phosphate
(PO4) in some tropical forest soils (Hedin et al. 2003) and have similar effects on soil base cation
and N content (Porder & Chadwick 2009, Porder et al. 2006). Physiologically, P participates in a
wide variety of processes (Plassard & Dell 2010, Comas et al. 2002), but the base cations also
perform indispensable functions (McLaughlin & Wimmer 1999), as does N (Millard & Grelet
2010). Since all soil macronutrients are subject to leaching and/or immobilization, a
comprehensive test for potential resource limitation of tree growth should investigate a broad
array of resources.
118
Individual tree growth is influenced both by intrinsic factors—ontogeny (Peichl & Arain
2007) and genotype (Boyden et al. 2005, Boyden et al. 2008)—and by extrinsic factors such as
climate (Clark et al. 2010), disturbance regime (Uriarte et al. 2004), resource availability
(Gradowski & Thomas 2008), interactions with surrounding trees (Coates et al. 2009), and
fungal (Hagerberg et al. 2003) or bacterial (Leblanc et al. 2005) mutualisms. Rarely, however,
do individual growth models incorporate soil resource data measured at appropriate scales
(Laubhann et al. 2009). A principal contribution of this study is to explicitly test how growth is
related to several soil resources measured at a scale commensurate with individual trees. These
estimates of individual-scale soil resource availability augment an established strategy for
modeling ontogenetic and neighborhood effects on tree growth (Canham et al. 1999, Canham et
al. 2006, Coates et al. 2009, Uriarte et al. 2004). Neighborhood effects characterize net local
tree-tree interactions, including light competition (Canham et al. 2004), belowground
competition (Coomes & Grubb 2000), plant-soil feedbacks (Fujinuma et al. 2005), (McCarthyNeumann & Kobe 2010), and other unmeasured size- and distance-dependent factors.
Soil resources are often measured at coarse spatial scales (Lundholm 2009), typically
sacrificing high spatial resolution for expedited processing or improved temporal representation.
Some recent work has attempted to measure resources at scales appropriate to individual trees
(Gradowski & Thomas 2008, Holste 2010), but it is nonetheless a considerable logistical
challenge to collect and analyze the thousands of soil samples required to properly assess
individual-level resource availability across stands. Evolving geostatistical methods, however,
facilitate interpolating resource availability at any scale from a relatively sparse sampling lattice
(Finley et al. 2007). Using increased computational capacity and new implementations of
Markov chain Monte Carlo algorithms in a Bayesian framework (Gelman 2006), it is possible to
119
estimate spatially dependent processes, like resource availability, and to measure the uncertainty
associated with the estimation procedure (Banerjee et al. 2004).
High alpha diversity, particularly in wet tropical forests, limits analysis of growth
responses at the level of individual species because many species are too rare to provide a
sufficient sample size (Gimaret-Carpentier et al. 1998, Lieberman & Lieberman 2007). The
ultimate simplifying assumption is ecological neutrality (Hubbell 2001); if most species respond
essentially the same way to resource availability and neighborhood interactions, then strong
relationships of growth to these variables should emerge when data are analyzed without
division. The opposing assumption, that every species differs ecologically in measurable ways,
motivates analysis of growth for individual species, as typically conducted for temperate
(Canham et al. 2006) or relatively low-diversity tropical systems (Uriarte et al. 2004). Phylogeny
(Swenson & Enquist 2009) or suites of functional traits (Swenson & Weiser 2010) also could
define groupings of species that may grow similarly in response to environmental factors.
As a distinct group, nearly all of the Fabaceae develop symbioses with N-fixing bacteria,
while endogenous N fixation is usually absent in other groups (Hedin et al. 2009). The N fixation
capability may alleviate P limitation if N-fixing species can utilize N for rhizosphere
phosphatase production that increases micro-scale PO4 availability by dissociating recalcitrant
forms of P (Treseder & Vitousek 2001). Thus the expectation is that growth of Fabaceae species
should be insensitive to N and P relative to other functional groups (Pearson & Vitousek 2002,
Menge et al. 2008, Cusack et al. 2009). Furthermore, species differ in terms of maximum
possible height / canopy position. Tree height governs physiological parameters such as
transpiration (Sperry et al. 2008) and the amount of carbon dedicated to support structures (King
120
2005, Buckley & Roberts 2006, Cole & Ewel 2006), and consequently carbon requirements for
respiration in non-photosynthetic tissues (MacFarlane & Kobe 2006, Elser et al. 2010).
Quantitative traits such as wood density (wood specific gravity) are strongly associated
with growth rate (Swenson & Enquist 2008), and species with different growth rate capabilities
may have different resource requirements (Ogle & Pacala 2009). In particular, species with lowdensity wood, and faster growth rates, and higher leaf turnover might require more soil resources
during rapid construction of foliar and other tissue (Oelmann et al. 2010). In contrast, species
with higher density wood, which are often shade tolerant (Poorter et al. 2010) and exhibit slower
growth rates (Fownes & Harrington 2004, Kelly et al. 2009), may tolerate lower nutrient levels
over short time scales (Easdale et al. 2007) due to greater nutrient use efficiency (Hidaka &
Kitayama 2009, Soolanayakanahally et al. 2009). Thus, growth of shade tolerant species with
high wood density could appear unrelated to soil resource availability (Baltzer et al. 2005).
The major purposes of this study are to test for correlations between a broad suite of soil
resources and individual and site-level tree growth and to compare relative strengths of these
correlations to identify potentially limiting soil resources. In particular, we hypothesized that:
Hypothesis 1: Tree growth is positively related to soil P and base cations (Ca, K, and Mg).
Hypothesis 2: Individual growth is negatively related to local neighborhoods defined by sizeand distance-dependent functions of neighboring trees.
Additionally, we also expected variation in growth-resource relationships among functional
groups. Specifically, we hypothesized that:
Hypothesis 3: Individual and stand-level growth of Fabaceae are unrelated to soil N and P (due
to N fixation and use in phosphatase) but strongly related to base cations.
121
Hypothesis 4: Across species, growth-soil resource relationships strengthen as wood density
decreases.
Methods
Field site
This study was conducted in five wet tropical forest sites at La Selva Biological Station in
northeastern Costa Rica (10°26‘N, 84°00‘W), an area that receives mean annual rainfall of 4306
mm (http:/81/www.ots.duke.edu/en/laselva/metereological.shtml) distributed relatively evenly
across the year. Sites were established across a gradient in soil P (Table D.1), with three sites
located on weathered volcanic ultisols and two sites located on entisols and inceptisols of recent
alluvial deposition (McDade & Hartshorn 1994). The original intention was to position sites in
undisturbed, primary forest (Vriesendorp 2003), but it was later realized that nearly 90% of one
alluvial site (4) was actually secondary forest less than 60 years old. In this site, both mean
growth and mortality rates in some taxa (e.g. Arecaceae) were slightly higher than at the other
alluvial site, but standard deviations from this site were well within the range of standard
deviations across all sites (Table D.1). In addition to representing a soil P gradient, the sites also
encompass a gradient in base cations and an opposing gradient in soil inorganic N availability
(Table D.1). Each site included an area 240m x 41m with the long dimension oriented at a
randomly selected azimuth. In 2000, all woody stems larger than 5 cm diameter at 1.3 m height
(excluding lianas) were measured (using diameter tapes) and their locations were mapped
(Vriesendorp 2003). In 2009, each site was expanded to encompass an area 280m x 81m, with
the original sites occupying the center of the new areas. All stems larger than 5 cm diameter at
1.3 m height were measured and mapped in this new area.
122
Growth measurements
Growth was measured three times following the original 2000 census, in 2005, 2007, and 2009,
with diameter increment calculated as the difference between the later and earlier measurements.
Encroaching buttress growth compromised the reliability of the original measurement, so growth
data used for this study were derived from the intervals between 2005-2009 or 2007-2009. For
dicot species with no buttress structures, diameter was measured at the same height above the
ground and at the same position on the stem in each census. Stem position was marked with a
nail (or multiple nails for trees of very large diameter), and the measurement was taken 10cm
above the nail to avoid any associated scarring. For individuals with extensive buttress
structures, measurements were taken above the termination of buttresses (Metcalf et al. 2009).
For 96 trees, or 1.8% of the total, buttress structures had grown above the 2005 measurement
point in later censuses, and it was necessary to infer the 2005 diameter. This inference was
accomplished by measuring diameter above the buttress at two heights separated by a distance of
1m, then projecting the 2005 diameter based on the assumption that the change in diameter was
constant across height. For some species in the Arecaceae, it was possible to measure diameter at
the same height above the ground, but several species had stilt root structures that progressively
lifted the measurement point. For these species, diameter was measured at the same location on
the stem regardless of how that point changed relative to ground level. In 2009, trees that had
been smaller than 5 cm diameter in 2005 (in-growth) were measured and mapped. Diameter
increment for in-growth was conservatively estimated as the difference between measured
diameter and 5 cm. Of the trees alive in 2005, 3.2% had died by 2009; for those trees, diameter
increment was calculated based on the 2005-2007 interval if the trees were alive in 2007. Trees
that were alive in 2005 but had died by 2007 (0.2%) were excluded from the analysis.
123
Soil sampling and resource measurements
At each site in June 2008, soil samples were taken from 254 locations. Along a central transect,
2
200 samples were taken from consecutive 1m quadrats. Outside of this transect, samples were
located at 5-m intervals on the long site axis and at alternating 10-m intervals on the short axis,
with 14 additional locations randomly selected in each site. Each composite sample consisted of
three cores (2 cm diameter, 20 cm depth) separated horizontally by 33 cm. Samples were air
dried for one week or until they achieved constant mass, then shipped from La Selva to Michigan
State University for processing. The samples typically formed into solid clay masses which were
coarsely ground so that roots, other biological debris, and stones could be removed. Samples
were then finely pulverized in an electric grinder prior to resource measurements.
+
-
Ammonium (NH4 ) and nitrate (NO3 ) were measured in a 2M potassium chloride
extract of 10g soil in 50 mL of solution. The NH4 pool in solution was measured following
Sinsabaugh et al. (2000), and the NO3 pool was measured according to Doane and Horwath
(2003). Both colorimetric assays were measured using an absorbance spectrometer (ELx808
Absorbance Microplate Reader, BioTek Instruments, Inc., Winooski, VT). A 28-day room
temperature incubation of soil at ~10% moisture content was conducted, but we do not include
potential N mineralization and nitrification rates here because of variability in moisture
conditions during the 28 day lab incubation.
2+
+
2+
The base cations (calcium (Ca ), potassium (K ), magnesium (Mg )), extractable
-
phosphorus (P), and phosphate (PO4 ), were measured in a weak acid Melich III extract (Mehlich
1984, Carter 1993, Loide et al. 2005) consisting of 6g soil in 42 mL of solution. Subsamples (15
124
mL) of the extract were used to measure Ca, K, Mg, and P with an Optima 2100DV ICP Optical
Emission Spectrometer (Perkin-Elmer, Shelton, CT). A colorimetric assay was used to measure
PO4 (Frank et al. 1998). Only extractable P was used in the analysis, however, because the soil
PO4 pool is extremely labile in these soils (Vincent et al. 2010) and extractable P more likely
captures the amount of P available to plants over longer time scales; nevertheless, PO4 and
extractable P were highly correlated (r = 0.920).
Local-scale soil resource estimation
Resource availability was directly measured at 254 points in each of the five sites, but these
locations were not always sufficiently close to focal trees to represent local availability. Thus, we
inferred resource values to each tree location using a spatial Bayesian model and a non-spatial
model in the package spBayes (Finley et al. 2007) implemented in R (The R Foundation for
Statistical Computing, 2009, http://cran.r-project.org/). For both spatial and non-spatial methods,
we tested effects of several geographic variables (elevation, slope, aspect, curvature, flow length,
flow direction) calculated in GRASS (GRASS development team, grass.osgeo.org) and QGIS
(Quantum GIS development team) from a LIDAR digital elevation map of La Selva (Kellner &
Asner 2009). Geographic factors were selected as independent variables based on a stepwise
regression starting with all factors and retaining only those that were related to each resource as
determined by Akaike‘s Information Criterion (AIC) (Burnham & Anderson 2002).
Multicollinear geographic factors were not permitted. For a given resource, all of the other
resources were additional covariates.
Each of the four resource estimation models—spatial framework with geographic
variables, spatial without geographic variables, non-spatial framework with geographic variables,
125
and non-spatial without geographic variables—was fitted using 75,000 iterations of the spBayes
Markov chain Monte Carlo (MCMC) algorithm. Model performance was assessed by comparing
G, P, and D values (Gelfand & Ghosh 1998), which measure goodness of fit, complexity penalty,
and dimensionality, respectively, from the spBayes routine spDIAG (Finley et al. 2007). For the
model with the best support for each site, resource availability was estimated as the mean value
of the posterior predictive distribution (PPD) at the coordinates of each tree using the spBayes
routine sp.predict (Finley et al. 2007). Each PPD was constructed by sampling 1000 regularly
spaced estimates from the converged MCMC chains, allowing 50,000 iterations for convergence.
Individual growth analysis
Data were analyzed at several levels, including all species together (representing ecological
neutrality), species with at least 50 focal trees (15 species), three functional groups further
divided into two size classes each (Fabaceae, canopy and smaller; Arecaceae, subcanopy and
smaller; non-legume dicots, canopy and smaller), and five wood density classes. For each of
these divisions, we evaluated individual diameter growth (Gd) as a function of (i) tree diameter,
(ii) a neighborhood index, and (iii) each of the soil resources (Ca, K, Mg, P, NH4, NO3), and a
species- or category-specific coefficient P:
Gd = P  Diameter effect  Neighborhood effect  Soil resource effect
(eq. 1)
where P was a scalar value representing theoretical maximum growth for each species (Canham
et al. 2006), and each remaining term was an exponential decay function constrained between 0
126
and 1 that could either reduce or have no effect upon P. This theoretical potential growth scalar
for each species was estimated simultaneously with other model parameters:

Gd = P  exp  0.5ln dbh    
2

(eq. 2)
where δ represents the diameter at which maximum growth occurs (consistent with the data), σ
controls the rate at which the function achieves the maximum. Neighborhood effect was
calculated as a size- and distance-dependent function (Berger & Hildenbrandt 2000, Canham &
Uriarte 2006, Vettenranta 1999, Wimberly & Bare 1996) of the number of neighbors:
 n

 
Neighborhood effect = exp  B dbhi exp
 i 1
disti 



(eq. 3)
where B is a coefficient to control the relative contribution of the neighborhood overall, α is an
estimated exponent controlling the influence of neighbor diameter, and β is an estimated
exponent controlling the influence of distance to the focal tree. The sum was calculated for i =
1…n neighbors within either a set radius of 20 m or within an estimated radius (≤ 20m)
determined by the data.
The soil resource effect was modeled as:
Soil resource effect = 1 – exp(C  resource)
(eq. 4)
where the exponential decay is subtracted from 1 so that lowest resource availability results in
the lowest growth. In addition to the full model (1), we tested a variety of simpler models,
127
including each of the effects (i), (ii), and (iii) alone, and every possible pair of effects (Table
D.2). Uncertainty based on the resource estimation procedure was quantified for individual-level
growth using 100 randomly selected datasets, calculating the range of parameter values and
predicted growth for the top model formulation for each category.
Parameter estimation and model comparison
Growth models were fit to all species, individual species with more than 50 focal trees, six
functional groups, and five wood density classes. Model parameters were estimated with
maximum likelihood using a simulated annealing algorithm (Metropolis sampler) implemented
in Delphi (Version 3.0, Borland Corporation, Austin, TX), running each model for 50,000 steps.
Growth was treated as a normal random variable in the likelihood function, with mean growth
specified by equation 1 and variance estimated as a parameter from the data. Residuals were
normally distributed, justifying the normal probability density function. The 95% confidence
interval for each parameter was calculated by likelihood profiling; the square of the Pearson
2
correlation coefficient (r ), used to determine goodness of fit, was calculated from the correlation
of model-predicted growth versus observed growth. For each model within a category, we
2
calculated the difference in r between it and the model that contained solely diameter effects;
this quantity represented the amount of total variance explained by non-diameter effects. The
models for each analysis group were compared using Akaike‘s Information Criterion corrected
for small sample size (AICc) (Burnham & Anderson 2002), which includes a penalty for model
parameters in order to select the most parsimonious model. Models with AICc values within 2
units of the minimum AICc are considered to have equivalent empirical support, so we report
only models that have ΔAICc < 2.
128
Site mean growth analysis
We also tested relationships of site-level mean basal area increment (BAI) to mean resource
availability. We used simple linear regressions because small sample size (n = 5 sites) precluded
fitting more complex functions. To account for differences in tree size among sites, site mean
growth was calculated as the average BAI of all of the focal individuals at each site (Table D.1).
Similarly, site mean resource values were calculated as the average of all of the estimated
resource values at each tree coordinate. Calculating distinct site-level mean growth and soil
resource values for each functional group was necessary because each tree was associated with
different soil resource conditions; different functional groups could experience different mean
resource levels even within a single site. In several instances where estimated intercepts were not
significantly different from zero, data were presented on the original scales, as opposed to
centered (Chatterjee & Hadi 2006), to facilitate comparison to other species / groups. To
propagate uncertainty from our resource estimation procedure, we calculated regression models
of BAI as a function of site-mean resource values for all 1000 datasets constituting the PPD for
each significantly-correlated resource.
129
Results
Overview
For many analytical categories, site mean BAI and individual tree growth were positively related
to soil P and base cations, results which generally supported our first, third, and fourth
hypotheses (Tables 1, 2; Figs. 1-4). In support of our first hypothesis, base cations (Ca, K, Mg)
and P were often correlated with growth (Tables 1, 2). Individual growth was rarely correlated
with local neighborhood, however, providing negligible support for our second hypothesis
(Tables 1, 2). At both the individual tree and site levels, the Fabaceae showed virtually no
relationship of growth to resources, whereas other functional groups had stronger growthresource correlations (Tables 1, 3, 4; Fig. 2). This result was consistent with our third hypothesis.
Among groups of species classified by wood density, only for the lowest-density species was
growth related to soil P, in accordance with our fourth hypothesis (Table 1, 5); for species in
higher wood density groups, growth generally was correlated with base cations (K or Mg).
Unexpectedly, focal tree diameter was not always the strongest correlate of individual tree
growth, and accounted for ≤ 19.3% total variance (Tables 3, 4, 5).
Stand-level growth correlated positively to soil P, Ca, and K, negatively to N and Mg
Grouping all species together, relationships of mean BAI to soil resources across sites supported
2
2
our first hypothesis, with growth related to P (R = 0.958), K (R = 0.790), and marginally
2
related to Ca (R = 0.744) (Figure 1, Table D.3). Sites located on the older volcanic soil types
had lower growth rates and considerably lower levels of P; the difference in Ca and K between
volcanic and alluvial sites was less pronounced (Table D.1). For all species together, the base
cations and P co-varied; many resources were also correlated in several other data classifications,
130
occasionally leading to multiple growth-resource correlations in some data categories (Table
D.4).
At the site level, dividing the dataset into functional groupings, results were consistent
with P, K, and / or Ca limitation, with P most important for the Arecaceae but base cations for
the non-legume dicot species. Mean BAI of subcanopy Arecaceae was related to extractable P
2
2
-1
(R = 0.868), with growth expected to vary from 2 to 12 cm yr across the range of soil P.
2
2
Mean BAI of smaller palms was related to P (R = 0.961) and to Ca (R = 0.843, Figure 2, Table
D.5, D.6); P and Ca were strongly correlated (r = 0.975, Table D.4). BAI of non-legume dicot
2
canopy species was related to K (R = 0.864), and was expected to increase from 12 to 21 cm
-1
2
2
yr across the K gradient. For subcanopy and smaller species, growth was related to K (R =
2
0.822) and Ca (R = 0.816, Table D.4), which themselves were correlated (r = 0.795, Table D.4);
2
-1
BAI was expected to vary from 3.5 - 7.5 cm yr (Figure 2, Table D.5).
Site-level mean BAI for single species was generally aligned with functional group
results. Among the Arecaceae, mean BAI in four species (S. exorrhiza, I. deltoidea, C.
warscewiczii, P. decurrens) was related to at least one resource, including N, P, and base cations
(Figure 3, Table D.7, 8). For the non-legume dicot species, mean BAI for three species (R.
deflexiflora, F. parvibractea, W. coccinea) was related to the base cations and P, but not to N
(Figure 3, Table D.7).
Growth of Fabaceae was unrelated to soil resources
Supporting our third hypothesis, site mean BAI of the Fabaceae was unrelated to soil resources,
either for canopy or smaller species (Figure 2, Table D.5). Site mean BAI for the single legume
131
species with sufficient sample size to warrant species-level analysis, Pentaclethra macroloba,
was not related to any soil resource (Table D.7).
Similarly, individual diameter growth for the Fabaceae was unrelated to P or inorganic N,
and only very weakly related to K. For canopy Fabaceae, K was significantly correlated with
growth but explained only 2.1% more variance than diameter alone (Table 4). For subcanopy and
smaller statured species, diameter alone was correlated with growth (Table 4). For P. macroloba,
despite having the largest sample size of any species (n = 368), growth was nominally correlated
2
with Ca (r = 0.026, Table 3).
Individual growth correlated positively but weakly to soil P, base cations, and N
Treating all species as one group, individual diameter growth was related only to initial diameter
nd
and soil K (Table 4, Table D.6, 2
lowest ΔAICc = 18.5, for Ca), and the model explained little
2
variance in growth (r = 0.081). Examining growth by functional group, the Arecaceae had
2
relatively strong correlations of individual growth to soil P (r = 0.105), while non-legume dicot
2
growth was correlated with K (r = 0.044). For subcanopy palm species, P explained 10.5% total
variance (Table 4). Growth of treelet and understory palms demonstrated a similarly robust
correlation to P (Table 4). For all other species, growth was weakly related to resources, with a
maximum of 4.4% additional total variance explained by soil K for species that could attain
canopy stature (Table 4, Table D.6).
For single species, diameter growth was often related to base cations or P, consistent with
our first hypothesis (Table 3, Table D.8). Although growth of the Arecaceae as a group was
related to P, growth of individual palm species was related to a variety of factors. Extractable P
2
was correlated with the growth of Iriartea deltoidea (r = 0.13, Table 3), one of the largest and
132
2
most common palm species in the study, and Prestoea decurrens (r = 0.059, Table 3). Growth
2
of the large palm species Socratea exorrhiza was correlated with K (r = 0.019), but for another
2
large palm, Euterpe precatoria, was positively correlated with NH4 (r = 0.116, Table 3).
Among non-legume dicot species, growth was unrelated to resource or neighborhood for
Dendropanax arboreus (Table 3); soil K was correlated with growth of Warszewiczia coccinea
2
2
(r = 0.107) and Rinorea deflexiflora (r = 0.065), while soil Ca was correlated with growth of
2
Capparis pittieri (r 0.109, Table 3).
Individual growth was rarely related to neighborhood index
Contrary to the prediction of hypothesis 2, neighborhood index was not correlated with growth
across all species or for any functional groups, and was only rarely correlated with growth of
single species. Welfia regia (a subcanopy palm) growth was related to the number of neighbors
2
within a radius of 6m (r = 0.100), and Cryosophila warscewiczii (a treelet palm) growth was
2
related to a diameter- and distance-dependent neighborhood with a set radius of 20m (r = 0.154,
Table 3; Table D.8). Growth of only one non-legume dicot species (Coussarea hondensis) was
2
related to the number of neighbors in a 7-m radius (r = 0.134, Table 3; Table D.8). For all three
species, growth was negatively correlated with neighborhood index.
Growth of species with low-density wood correlated more strongly with soil P
Among wood density categories, the base cations and P were all correlated with BAI. Extractable
P and K were correlated with BAI of species constituting the minimum wood density class, with
2
-1
growth expected to increase from 5.5-13.5 cm yr across the resource gradient (Figure 4).
133
Consistent with our fourth hypothesis, these relationships were slightly stronger (assessed by
goodness of fit for simple linear regression) for the minimum density class than for other
categories (Table D.9). For the higher density classes, site-mean BAI was related to Mg and Ca,
while for species of unknown wood density—likely representing many low-density species,
based on taxonomic groupings—BAI was again related to P (Figure 4).
Since growth of the Fabaceae was not correlated with soil resources, they were excluded
from the above wood density analyses (Figure 4). Including the Fabaceae still resulted in
2
significant correlations to the same soil resources and minor decreases in R values (Table D.9).
Furthermore, when functional groups were separated into wood density classes, there were no
significant BAI-resource relationships for the Fabaceae. In contrast, despite small sample sizes at
this level of analysis, both the Arecaceae and the non-legume dicots showed at least one
significant growth-resource relationship (Table D.9).
Considering individual diameter growth, the group of species with lowest wood density
showed considerably stronger correlation of diameter growth to P than species with higherdensity wood (supporting hypothesis 4), but in all cases resources explained little variation in
growth (Table 5, Table D.10). Extractable P explained 5.8% growth variance—440% more than
diameter alone—for the lowest density species (Table 5). By comparison, no more than 0.9% of
total variance could be attributed to soil resources for any other group of known density (Table
5). For species with unknown wood density, growth was related to Ca (Table 5).
Uncertainty propagated from resource estimation
The spatial Bayesian model provided estimates for both the mean resource level and uncertainty
in resources. To test the robustness of growth-resource relationships, we propagated resource
uncertainty through the models. At the site level, the envelope of possible predicted growth was
134
typically limited to a narrow range around the mean predictions (Figs. 1, 3, 4). Within a data
analysis class, the uncertainty resulting from estimating values of P was typically larger than that
of base cations. For example, with all species grouped together, the range of predicted BAI as a
2
-1
function of P at the center of the resource domain was approximately 4 cm year , while the
2
-1
range of predicted BAI as a function of K was only 2 cm year (Figure 1). These differences
were often even more pronounced for other data categories (Figs. 2, 3, 4). For the functional
groups, species, and wood density classes for which resources explained at least 5% total
variance in individual diameter growth, we calculated predicted growth given the uncertainty in
resource measurements (Table D.11, D.12). As expected, uncertainty in growth-resource
relationships due to uncertainty in resource estimates was generally greater in data categories
with lower sample sizes (Figure D.12).
135
Discussion
Growth and productivity in wet tropical forests are typically considered to be limited by P
availability (Vitousek et al. 2010), but results were consistent with limitation by P or base
cations, supporting our first hypothesis. Although individual tree growth was weakly related to
soil resources and neighborhood, there were very strong relationships between site-level mean
growth and soil resources. For all species together, site-level mean growth was strongly related
to soil P, K, and Ca. Deficiencies in P or the base cations are common on older soils subject to
high rates of leaching (Vitousek & Farrington 1997, Porder et al. 2007), so our results were
consistent with expected resource limitations on these types of soils. Extractable P represents a
longer-term pool more relevant to plant growth (Valdespino et al. 2009), likely contributing to
2
the very strong relationship of basal area growth to extractable P (r = 0.95). In contrast, soluble
K and Ca represent more dynamic resource pools (Tripler et al. 2006), potentially contributing to
the slightly weaker relationships of growth to these resources. Based on strength of relationship,
there may be more support for P limitation, but we cannot discount the possibility of base cation
limitation. Similar to all species as a group, individual tree growth was related to Ca and P for
two species each, and to K for four species, as well as inorganic N for two species, but none of
these relationships explained more than 13% of total growth variance.
Local neighborhood was negatively correlated with growth for only three species and
explained between 10 and 15% total growth variance, suggesting, contrary to our second
hypothesis, that competition was relatively unimportant (Uriarte et al. 2004). We believe,
however, that competition and neighborhood interactions likely exert much greater influence on
growth (Canham et al. 2006, Guariguata & Ostertag 2001, Hood et al. 2004) than detected here.
In forests with minimal disturbance, as with our data, growth of all trees, especially those in the
136
subcanopy, could be heavily suppressed by neighbors (Coates et al. 2009). Thus, lack of
variation in neighborhood conditions, particularly the absence of non-suppressed trees, may have
precluded detection of neighborhood effects.
Individual diameter growth and site mean BAI were related to soil resources for palm and
non-legume dicot species but not for legumes; a likely explanation for this apparent legume
escape from P limitation is diversion of fixed N2 to production of rhizosphere phosphatase
(Olander & Vitousek 2000, Wang et al. 2007) that dissociates plant-available PO4 from
recalcitrant sources (Adams et al. 2010). The extensive N fixation that occurs in tropical forests
appears paradoxical because N2 fixation should be down-regulated under conditions of excess N
availability (Hedin et al. 2009, Menge et al. 2010). Our results suggest that legume stem growth
was limited by neither N nor P, and that growth was correlated with soil P only for non-legume
species and functional groups.
Increased inorganic soil N content and plant-available P facilitated by legume N2 fixation
and phosphatase production also could contribute to an imbalance in base cation availability
(Treseder & Vitousek 2001) relative to N and P, particularly for legumes (Hogh-Jensen 2003).
Site mean legume growth was unrelated to soil base cations, however, and we cannot identify a
clear mechanism that explains how N2 fixation could alleviate base cation limitation (Smethurst
2010). In fact, N2 fixation in certain temperate forests contributes to base cation leaching at
ecosystem scales due to elevated soil NO3 (Compton et al. 2003); we identified negative
correlations of soil base cation content to inorganic N (Table D.13) consistent with leaching of
these mineral resources, potentially by increased soil N content or legume presence (Cahn et al.
137
1993, Emmett 1999). Furthermore, either possibility—alleviation of base cation limitation
related to N2 fixation or increased limitation as a consequence of leaching—may be difficult to
detect in our dataset because variability in site-level mean basal area increment for legumes was
limited (Table D.12). Although growth rates for canopy Fabaceae exceeded those of every other
functional group (Table D.12), it is unclear whether this resulted from alleviation of P / base
cation limitation or a different mechanism.
In contrast to the legumes, we found two lines of evidence consistent with N
accumulation and base cation limitation for other functional groups. The primary support for
base cation or P limitation was the positive correlations of BAI to these resources for the
Arecaceae and non-legume dicots, and for single species within these functional groups. Second,
soil resources whose abundance does not limit plant growth should be negatively correlated with
any limiting resources, as increasing levels of the limiting resource allow plants to further draw
down supplies of resources available in excess (LeBauer & Treseder 2008). For all species
grouped together, soil NH4 and NO3 were negatively correlated with all the base cations and
with extractable P (Table D.13). For three individual species (S. exorrhiza, I. deltoidea, and P.
decurrens) for which the strongest growth-resource relationships were negative correlations with
NH4 and NO3 (Table D.7), inorganic N was in turn negatively related to soil base cations and
extractable P (Table D.13), which is consistent with these latter resources limiting tree growth.
Evidence consistent with P and K limitation was strongest for species with lowest wood
density, in accordance with our fourth hypothesis which predicted that faster-growing species
should be more resource-limited (Stape et al. 2008). We also, contrary to prediction, found
strong correlations between growth and base cation availability for species with higher wood
138
density. Evidence consistent with P limitation at the site level existed only for presumably fastergrowing species of low wood density (Oelmann et al. 2010), while results for slower-growing
species with higher wood density (Kelly et al. 2009) were consistent with base cation limitation.
However, while species with low wood density may be fast growing from the perspective of
diameter or basal area increment, they may not be fast growing from the perspective of mass
increment. Furthermore, if stoichiometry of wood tissue is consistent across wood density
classes, then mass increment would more strongly influence demand (Easdale & Healey
2009)(Easdale & Healey 2009)for mineral nutrients than diameter or basal area increment
(Easdale & Healey 2009, Smethurst 2010). Thus, we calculated a proxy for mass increment as
the product of mean BAI and mean wood density for a 1-cm high cross-section of annual growth
at the point of diameter measurement. Based on this index, the fastest mass growth (10.79 g year
1
-
-3
) actually occurred in the highest wood density class (0.7-1.2 g cm ), followed by the 0.4-0.55
-3
-1
1
g cm class (2.28 g year ); slowest mass growth (1.59 g year ) occurred in the lowest density
-3
class of 0.2-0.4 g cm . We do not, however, have the information necessary to calculate whole
tree mass increment, which may differ from stem cross section mass increment due to variation
in tree height and branching patterns (Chave et al. 2005) and within-tree variation in wood
density. Nevertheless, a shift from P limitation in early-succession, low-density species to base
cation limitation in late-succession, shade-tolerant species with high wood density could indicate
differences in either stoichiometric resource limitation or resource use efficiency between earlyand late-succession species (Gleeson & Tilman 1994, van Kuijk & Anten 2009).
139
Conclusions
We identified strong correlations of site mean growth (BAI) to soil P and base cations,
particularly among the Arecaceae and non-legume dicot species and for several wood density
categories, but generally not for legume species. Whereas P has often been implicated in limiting
productivity of tropical forests, soil K was correlated with growth for more species and
functional groups, both at the site and individual levels (H1). In addition, competition appeared
to have limited influence on individual tree growth (H2), although data from a variety of
disturbance regimes could provide a different conclusion. For the legumes, both individual
diameter growth and site mean BAI were unrelated to soil resources (H3), consistent with
alleviation of resource limitation (for example, by rhizosphere phosphatase production supported
by N2 fixation). We emphasize that our results were based on natural variation in soil resources
and that resource limitation must ultimately be resolved using fertilizer addition experiments.
Many such fertilization studies, however, are designed to test for P limitation or are based on
factorial N and P additions. In contrast, although our results identify P as a possible limiting
resource, our strong correlative support for limitation by the base cations, particularly K, suggest
that base cations also should be more rigorously examined for their potential in limiting tree
growth and wood production in wet tropical forests (e.g. P x base cation fertilization
experiments). While we stress the potential importance of the base cations, our results also
suggest that P is the most probable limiting resource for growth of species with low wood
density (H4), which tend to be early succession species. Thus, P availability may determine the
rate of early successional dynamics on tropical soils. Overall, our results support that tropical
tree growth may be limited by base cations and/or P; furthermore, the degree and type of
140
resource limitation may depend on functional group differences in wood density and potential for
N fixation.
Acknowledgements
Many thanks to E. Holste for her extensive work on the soil resource dataset. We also thank S.
Grandy, D. Rothstein, and M. Walters for insight about resource analysis and model design. Two
anonymous reviewers provided useful feedback during preparation of the manuscript. Essential
field assistance was provided by A. Hurtado, M. Cascante, and R. Garcia. This research was
supported by NSF (DEB 0075472, 0640904, 0743609).
141
Table 5.1. Summary of results for stem growth as a function of soil resources for all species
together, for functional groups, and for wood density categories. Strongest relationships occurred
2
at the site level, for which goodness of fit is presented as R .
1
Analysis category
All species
FC
F<C
AS
A<S
NLDC
NLD<C
0.2-0.4 g cm
Mg
P
NH4 NO3
0.985
0.868
0.961
0.843
0.864
0.816 0.822
-3
0.55-0.7 g cm
0.815
-3
0.912
0.851
-3
-3
Unknown g cm
1
K
0.790
0.4-0.55 g cm
0.7-1.2 g cm
Ca
0.799
-3
0.780
Functional group and size class abbreviations: FC: Fabaceae, canopy; F<C: Fabaceae, smaller
than canopy; AS: Arecaceae, subcanopy; A<S: Arecaceae, smaller than subcanopy; NLDC: nonlegume dicot species, canopy; NLD<C: non-legume dicot species, smaller than canopy;
numerical categories and ‗Unknown‘ represent wood density classes.
142
Table 5.2. Summary of results for stem growth as a function of soil resources for single species.
All of the strongest relationships again occurred at the site level (with goodness of fit presented
2
as the R value), but for three species neighborhood effects (defined only at the individual level)
were significantly correlated with growth (Table 3).
1
Analysis category
were
pema
soex
ride
irde
capi
fapa
eupr
crwa
dear
caar
prde
coho
waco
gome
other
Ca
K
Mg
P
NH4
NO3
0.916 0.873
0.871
0.980
0.839 0.876
0.963 0.997 0.953 0.981
0.820
0.960
0.997
0.949
1
Species abbreviations: Welfia regia (were); Pentaclethra macroloba (pema); Socratea
exorrhiza (soex); Rinorea deflexiflora (ride); Iriartea deltoidea (irde); Capparis pittieri (capi);
Faramea parvibractea (fapa); Euterpe precatoria (eupr); Cryosophila warscewiczii (crwa);
Dendropanax arboreus (dear); Casearia arborea (caar); Prestoea decurrens (prde); Coussarea
hondensis (coho); Warszewiczia coccinea (waco); Goethalsia meiantha (gome).
143
Table 5.3. Results of fitting models of individual tree growth as a function of diameter, soil resources, and neighborhood index for
species for which more than 50 focal individuals were available. Total variance explained by supported models is compared to the
diameter-only model to identify any additional variance explained by neighborhood or resources (Add. Var.). If the best supported
model is the diameter-only model, then no additional variance can be explained by other factors.
Top model
Species
Funct.
group
Res.
Diameter-only
k
n
ΔAICc
r
2
Add. Var.
r
2
ΔAICc
6
168
0.000
0.124
0.100
0.024
13.490
1
AS
2
FC
Ca
5
368
0.000
0.126
0.026
0.100
0.822
2
AS
K
5
193
0.000
0.146
0.019
0.127
1.970
2
NLD<C
K
5
91
0.000
0.172
0.065
0.107
4.530
2
AS
P
5
360
0.000
0.277
0.130
0.147
56.985
2
NLD<C
Ca
5
72
0.000
0.149
0.109
0.040
6.334
3
NLD<C
4
83
0.000
0.011
2
AS
5
55
0.000
0.309
0.116
0.193
4.618
4
A<S
7
112
0.000
0.251
0.154
0.097
11.451
3
NLD<C
4
57
0.000
0.054
were
pema
soex
ride
irde
capi
fapa
eupr
crwa
dear
NH4
144
Table 5.3 (cont‘d)
2
NLD<C
NO3
5
207
0.000
0.076
0.024
0.052
3.031
2
AS
P
5
106
0.000
0.060
0.059
0.001
3.766
1
NLD<C
6
64
0.000
0.226
0.134
0.092
4.671
2
NLD<C
5
96
0.000
0.111
0.107
0.004
7.471
3
NLDC
4
51
0.000
0.066
5
2101
0.000
0.105
0.006
0.099
11.608
caar
prde
coho
waco
gome
2
other
K
K
1
Count of neighbors, estimated radius
2
Diameter and resource effects; no neighborhood
3
Diameter effect only; no resource, no neighborhood
4
Size and distance-dependent neighborhood, radius = 20m; no resource
145
resources and neighborhood indices for all species combined and categorized by functional
group and size class. Total variance explained by supported models is compared to the diameteronly model to identify any additional variance explained by neighborhood or resources (Add.
Var.). If the best supported model is the diameter-only model, then no additional variance can be
explained by other factors.
Top and supported models
Diameter-only
Analytical
category
All species
FC
F<C
AS
A<S
NLDC
NLD<C
Res.
k
n
ΔAICc
r
2
Add. Var.
r
2
ΔAICc
K
K
5
5
4
4184
446
151
0.000
0.000
0.000
0.091
0.111
0.227
0.010
0.021
0.081
0.090
79.56
7.728
NO3
P
P
K
Ca
K
5
5
5
5
5
5
151
887
119
726
726
1752
1.863
0.000
0.000
0.000
1.543
0.000
0.228
0.190
0.080
0.098
0.085
0.362
0.001
0.105
0.075
0.044
0.031
0.294
0.085
0.005
0.054
105.56
7.560
5.785
0.068
55.769
146
Table 5.5. Results of fitting models of individual tree growth for species grouped by wood
density as a function of diameter, soil resources, and neighborhood index. Total variance
explained by supported models is compared to the diameter-only model to identify any
additional variance explained by neighborhood or resources (Add. Var.). If the best supported
model is the diameter-only model, then no additional variance can be explained by other factors.
Top and supported models
Wood density
-3
(g cm )
0.2-0.4
0.4-0.55
0.55-0.7
0.7-1.2
Unknown
Diameter-only
Res.
k
n
ΔAICc
r
2
Add. Var.
r
2
ΔAICc
P
P
K
Ca
Ca
Ca
5
5
5
5
4
5
956
964
1111
1111
98
1043
0.000
0.000
0.000
1.256
0.000
0.000
0.071
0.037
0.006
0.004
0.182
0.151
0.058
0.009
0.002
0.000
0.013
0.028
0.004
55.942
10.062
8.234
0.016
0.135
17.037
147
2
R = 0.958
2
-1
Mean Δ basal area (cm yr )
p = 0.004
p = 0.044
2
R = 0.790
2
-1
P (ppm)
K (ppm)
Figure 5.1. Site mean basal area increment for all species as functions of soil resources.
Basal area increment was significantly related to soil P (a) and K (b). Solid lines represent simple
linear regression of growth as a function of mean soil resource with an estimated intercept,
2
presented with associated R and p values. Cross symbols represent the envelope of predicted
growth based on 1000 resource datasets sub-sampled from the posterior predictive distribution of
each resource.
148
0.003
pp ==0.508
p = 0.028
R =< 0.868
0.001
0.961
RR ==0.16
R = 0.843
2
22
2
2
-1
p = 0.021
0.972
P (ppm)
P (ppm)
Ca (ppm)
Figure 5.2. Site mean basal area increment for six taxonomic groups as functions of soil resources. Lines represent simple linear
2
regression of growth as a function of soil resource, with associated p and R values. Growth of subcanopy palm species was related to
P (a), while growth of treelet and understory (< subcanopy) palms was related to Ca and/or P (b, c). Among non-legume dicot species
(Other), mean diameter growth of canopy species was related to K (d), and of subcanopy, treelet, and understory (< canopy) species to
Ca or K (e, f). Notably, growth of Fabaceae was unrelated to resources regardless of size class (g – i), shown here as functions of the
soil resources most strongly correlated (none significantly) to individual growth. Cross symbols represent the envelope of predicted
growth based on 1000 resource datasets sub-sampled from the posterior predictive distribution of each resource. Envelopes were not
calculated for Fabaceae because regressions were not significant.
149
p = 0.972
p = 0.508
p = 0.036
R < 0.001
R = 0.16
R = 0.816
2
2
2
2
-1
Figure 5.2 (cont‘d)
K (ppm)
K (ppm)
150
Ca (ppm)
-1
p = 0.034
p = 0.740
R = 0.864
R = 0.822
R < 0.04
p = 0.022
2
2
2
K (ppm)
2
NO3 (ppm)
151
P (ppm)
2
-1
2
-1
K (ppm)
2
-1
2
-1
P (ppm)
NO3 (ppm)
NH4 (ppm)
Figure 5.3. Site-mean basal area increment for individual species as a function of soil
resources. Growth was positively related to P, Ca, and K, but negatively related to NO3, NH4,
and Mg. For species with growth related to more than one resource, the same symbol and line
type are used in all relevant panels. See Table D.6 for model parameters and goodness of fit
assessment.
152
2
-1
2
-1
Ca (mg kg )
-1
Mg (mg kg )
-1
153
-1
2
R = 0.912
0.2 – 0.4 g cm
(ppm)
p = 0.011
-3
0.2 – 0.4 g cm
(ppm)
p = 0.036
2
2
R = 0.815
P (ppm)
0.4 – 0.55 g cm
(ppm)
p = 0.026
K (ppm)
-3
0.7 – 1.2 g cm
(ppm)
p = 0.041
2
-3
2
R = 0.851
R = 0.799
2
-1
-3
Mg (ppm)
Ca (ppm)
Figure 5.4. Site mean stem area growth for wood density categories, excluding legume
species, plotted versus soil resources. Lines represent simple linear regression of growth as a
2
function of soil resource, with associated p and R values. For the two lightest density groups, P,
K, and Mg were correlated with growth (a, b, c). In the second-highest density group (d, e),
growth was related to K and Ca. Cross symbols represent the envelope of predicted growth based
154
on 1000 resource datasets sub-sampled from the posterior predictive distribution of each
resource.
p = 0.047
2
R = 0.780
2
-1
Unknown density
P (ppm)
155
CHAPTER VI
SYNTHESIS
156
Summary
In broadest terms, this dissertation reassessed the relative influence on tree growth of soil
resources, including nitrogen (N), phosphorus (P), calcium (Ca), potassium (K), magnesium
(Mg), and water. Whereas temperate forests are generally classified as N-limited systems and
tropical forests as P-limited, results from both biomes (Chapters II and V) identified strong
correlations of stand-level growth and productivity to soil base cation availability. The
possibility of base cation limitation of temperate forest productivity was investigated (Chapter
IV) by fertilizing individual forest trees with Ca and / or N and comparing resulting growth
responses. In addition to soil resource effects, individual tree growth is influenced by interactions
with neighboring trees. Neighborhood models often, however, discount soil resource
heterogeneity at both coarse and fine scales; simultaneous comparison of local-scale
neighborhood and soil resource effects (Chapter III, V) revealed that both may influence growth
to a similar extent, depending on species composition and site-level soil resource availability.
Competition and soil resource effects transform across soil resource gradients
Plant communities may be organized according to competitive hierarchies (Canham et al. 2006),
niche partitioning (Silvertown 2004), residual effects of colonization (Chesson 2003), or a
variety of tradeoffs and interactions among these processes. In north-temperate hardwood
forests, I found that aboveground net primary productivity of wood tissue (ANPPW) was
strongly correlated with soil Ca across sites with low resource availability overall, whereas
ANPPW was related to soil inorganic N content and volumetric water across sites with higher
overall soil resource availability (Ch. II).
157
The relative importance of potential soil resource effects at high fertility versus weak or
absent resource effects in species associated with low fertility suggests niche partitioning along
soil resource availability axes (Turner 2008). Simultaneously, changing competitive effects were
likely also important (Canham et al. 1999). Species distributions may be controlled by relative
ability to exploit available resources (Wedin & Tilman 1993) or by a tradeoff between ability to
tolerate scarce resources versus ability to compete given abundant resources (Grime 2001). In
temperate hardwood species most prevalent at low- or intermediate-fertility sites, I found that
individual diameter growth was essentially unrelated to competition as measured by local
neighborhood index (Ch. III). In contrast, diameter growth was, for the most part, negatively
related to competition among species dominant at high-fertility sites (Ch. III), consistent with a
tradeoff between tolerance of low-resource conditions and competitiveness at high-resource
sites. Tolerance of low resources and competitive exclusion of species common at low-resource
sites was not explicitly measured, so these results could not definitively reveal a tolerancecompetition tradeoff. Nonetheless, my results were consistent with seedling performance data
based on reciprocal transplants at the same sites that showed precisely this tradeoff (Schreeg et
al. 2005).
In tropical forest stands distributed across a soil P gradient, however, neighborhood
effects were not measurable for most species (Ch. V). Inability to detect correlations of diameter
growth to local neighborhood index does not necessarily mean that competition was absent from
the system (Coates et al. 2009). In fact, competition was probably intense (Guariguata &
Ostertag 2001), particularly for subcanopy trees, but nonetheless undetectable because all
neighborhood indices corresponded with suppressive conditions (Ch. III, V). Competitive effects
were likely obscured in a similar fashion in the temperate forests; in both biomes a more accurate
158
measure of competition could be assessed in stands encompassing a wider range of disturbance
regimes and developmental stages (Coates et al. 2009).
Implications of heterogeneous resource limitation at multiple spatial scales
Forest productivity is widely assumed to be limited by single soil resources across entire
bioregions (LeBauer & Treseder 2008), typically by P in tropical forests (Vitousek & Farrington
1997) and N in temperate forests (Finzi & Canham 2000). Results from my dissertation research
suggest that not only may tree growth and forest productivity be limited by soil base cation
availability in addition to N and / or P, but also that resource limitation may vary at fine spatial
scales (Ch II, III, V). In northern hardwood forests, ANPPW was most strongly correlated with
soil Ca in stands growing on soils low in both Ca and inorganic N, while ANPPW and leaf
production (ANPPL) were correlated with soil N in stands on soils with abundant Ca and N (Ch.
II). Across all sites, both ANPPW and ANPPL were correlated with soil volumetric water (Ch.
II). Thus, rather than clear support for N limitation across sites, as suggested by e.g. Zak et al.
(1989), results suggested that production could be sequentially limited by N and base cations, or
limited by soil water instead of either nutrient.
First, N-limitation could be restricted to high-resource sites, with Ca-limitation operating
at low-resource sites. I tested this scenario by factorial addition of N and Ca to individual trees
across sites. Preliminary results, including leaf nutrient content and non-significant increases in
diameter growth rates, suggest that Quercus alba, dominant at low-resource sites, responds to
addition of Ca, whereas Acer saccharum, dominant at high-resource sites, responds to N addition
(Ch. IV). I also tested correlations of individual tree diameter growth to local-scale soil resource
159
availability (Ch. III), finding that growth rates of species dominant at low fertility were largely
independent of soil resource availability, while growth of high-fertility species was correlated
with several resources, including Ca, K, N, and water.
Results from the fertilizer addition experiment will eventually resolve how resource
limitation changes across the fertility gradient; I expect those results to become significant as the
experiment proceeds (Neilsen & Lynch 1998, Bauer et al. 2004), especially since the first two
years of the experiment coincided with complete defoliation of A. saccharum by Malacosma
disstria caterpillars across sites. A second possibility is that growth and productivity are
ultimately controlled by soil water availability (Richards et al. 2010), and that correlations of
growth to soil Ca and N are spurious. Although we cannot easily manipulate soil water
conditions, failure to identify any significant growth response to Ca versus N addition could
indirectly support limitation of productivity by soil water.
If N is not, as my results suggest, the only limiting resource in north temperate hardwood
forests, there are several implications for forest ecology and management. Large-scale models of
forest productivity (e.g. Schimel et al. 1997) that incorporate soil N availability as their sole
resource predictor could be inappropriate. For example, high levels of atmospheric N deposition
(Gradowski & Thomas 2006, Gress et al. 2007, Finzi 2009) would increase productivity if
forests are N-limited, but Ca leaching caused by N deposition (Perakis et al. 2006) would
decrease productivity if forests are Ca-limited. Alternately, global climate change in which
precipitation decreases and temperatures increase across North America (Lutz et al. 2010) would
result in decreased forest productivity if tree growth is water-limited, yet models based on Nlimitation could fail to predict this outcome.
160
Reflecting my temperate system results, tropical tree diameter growth and stand-level
mean basal area increment (BAI) were not simply correlated with soil P (Ch. V), as should occur
if tropical production were P-limited (Vitousek et al. 2010). Global carbon cycling models that
assume P limitation in tropical forests (Lukac et al. 2010) may under-predict the importance of
tropical forest productivity relative to other biomes if production is in fact regulated by other soil
resources (e.g. base cations) or principally by climatic factors (Clark et al. 2010). Whereas
growth-resource correlations shifted across sites in my temperate results, potential resource
limitation patterns in tropical stands depended to a greater extent on species and functional group
identities (Ch. V), in agreement with other tropical research (Baker et al. 2003).
For the Fabaceae, growth was unrelated to soil resources at both the individual and site
levels and across wood density classes, evidence that this functional group escaped soil resource
limitation (Ch. V). Abundant N from symbiotic fixation could be allocated to production of
rhizosphere phosphatase, enhancing uptake of PO4 from recalcitrant mineral pools (Olander &
Vitousek 2000). Whatever the ultimate reason for weak relationships between legume growth
and soil resources (Menge et al. 2010), my results imply that tropical forests with a greater
proportion of legumes may exhibit weaker resource limitation or, conversely, higher levels of
productivity on low-resource soils (Stape et al. 2008). A management application of this result
could be to increase the relative proportion of legumes in species assemblages used for tropical
forest restoration (Siddique et al. 2008).
Other functional groups and trait (wood density) classes showed extremely strong
correlations of site-level BAI to mean soil resource availability (Ch. V), consistent with soil
resource limitation. For both the Arecaceae and species with low wood density (excluding
Fabaceae), BAI was related most strongly to total soil P; among the non-legume dicot species
161
and species with higher wood density (again excluding Fabaceae), BAI was related to soil Ca
and K (Ch. V). Early-succession forests typically contain a higher proportion of fast-growing
species with low wood densities (Doust et al. 2008); my results suggest that productivity in such
forests could be P-limited. In contrast, late-succession forests often contain a greater contingent
of slower-growing, shade-tolerant species with high-density wood (Gough et al. 2010), and
production in such forests could be more limited by soil base cation availability. Consequently, P
limitation in early succession second-growth tropical forests may be a significant impediment to
establishment and optimum productivity, but these problems could be alleviated by promoting
regeneration of legumes and species with higher wood densities (Binkley et al. 2003). In tropical
forests managed under selection systems (de Freitas & Pinard 2008) it could be advantageous to
maintain higher residual basal areas of species with low wood densities where soils contain
abundant total P, but to retain more basal area of species with high wood density on soils with
low total P.
Conclusions and future directions
This dissertation reinforces the growing recognition that a dichotomous classification of P
limitation in tropical forest ecosystems versus N limitation in temperate systems can be an
unrealistic simplification. I identified potential Ca and / or soil water limitation in addition to N
limitation in north temperate hardwood forests growing across a gradient of multiple resources.
In lowland wet tropical forests growing across a strong soil P gradient, I identified potential base
cation limitation in addition to P limitation, while also showing that soil resource limitation
likely does not affect the Fabaceae. This evidence, however, was conservatively interpreted as
potential resource limitation because it derived from correlations between tree growth and soil
162
resource availability. Actual resource limitation must be confirmed by documenting a growth
response to fertilizer application. My field experiment to test N versus Ca limitation in four
common temperate species is ongoing, but after only three years has shown that leaf nutrient
contents increase and that diameter growth increases, though not significantly, in response to
fertilization. The potential for base cation limitation that I identified in both temperate and
tropical forests needs to be confirmed by the continuation of the fertilization experiment that I
implemented at the temperate sites, and by a parallel manipulation in tropical forests.
Additionally, the nature of neighborhood effects across soil resource gradients deserves further
study, both to measure a wider array of neighborhoods and to clarify how plant competition
interacts with soil resource availability.
163
APPENDICES
164
APPENDIX A
165
APPENDIX A
Table A.1 Site mean values, with standard error, for inorganic N pools (NO3 and NH4), soil N dynamics (N-min rate and
nitrification), and total C and N content. Inorganic and total N content were measured on differing numbers of samples, which are
listed under n (N) and n (C,N).
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
n
(N)
4
5
5
5
5
5
5
5
5
5
5
5
5
-
NO3
-1
(mg kg )
0.52 ± 0.13
0.58 ± 0.12
0.59 ± 0.16
0.60 ± 0.10
0.57 ± 0.12
0.50 ± 0.16
0.54 ± 0.04
0.53 ± 0.12
0.89 ± 0.54
0.82 ± 0.32
1.99 ± 0.59
0.76 ± 0.22
2.28 ± 0.81
+
NH4
-1
(mg kg )
1.46 ± 0.46
1.41 ± 0.35
1.32 ± 0.73
1.57 ± 0.85
1.84 ± 0.88
1.07 ± 0.48
2.94 ± 0.44
2.43 ± 0.53
2.70 ± 1.78
2.66 ± 1.34
1.74 ± 0.44
2.92 ± NA
2.54 ± 0.98
N-min Rate
-1
-1
(mg kg day )
Nitrification
-1
-1
(mg kg day )
n
(C,N)
Total N
(%)
Total C
(%)
C:N
0.30 ± 0.07
0.42 ± 0.06
0.40 ± 0.02
0.33 ± 0.09
0.37 ± 0.10
0.29 ± 0.07
0.52 ± 0.16
0.37 ± 0.07
0.6 ± 0.15
0.57 ± 0.49
0.13 ± 0.03
0.25 ± 0.04
0.25 ± 0.08
0.11 ± 0.01
0.11 ± 0.01
0.11 ± 0.02
0.10 ± 0.01
0.11 ± 0.01
0.11 ± 0.01
0.17 ± 0.05
0.14 ± 0.02
0.14 ± 0.05
0.43 ± 0.20
0.58 ± 0.10
0.78 ± 0.75
0.79 ± 0.22
8
8
8
8
8
8
8
8
8
8
8
2
6
0.10 ± 0.02
0.08 ± 0.01
0.06 ± 0.01
0.05 ± 0.01
0.05 ± 0.02
0.06 ± 0.04
0.09 ± 0.02
0.12 ± 0.04
0.06 ± 0.01
0.12 ± 0.03
0.15 ± 0.03
0.17 ± 0.00
0.17 ± 0.04
2.11 ± 0.34
1.96 ± 0.34
1.76 ± 0.11
1.72 ± 0.25
1.45 ± 0.32
1.64 ± 0.73
1.77 ± 0.08
2.05 ± 0.32
1.74 ± 0.15
1.47 ± 0.84
2.12 ± 0.69
2.83 ± 0.02
2.81 ± 0.81
22.16
24.80
29.76
32.49
30.94
25.59
20.86
16.93
27.62
12.14
14.04
16.76
16.13
166
-1
-1
Table A.2. Site mean values for soil acidity and Al (cmol kg ), base cations (Ca, Mg, K), and P (all presented as cmolcharge kg
soil). Sample size (n) refers to the number of composite samples per site, each of which comprised four soil cores. Standard errors are
-
presented for the base cations and PO4 , but not for acidity and Al because one titration was conducted per sample.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
n (Al)
Acidity, Al
4
4
4
4
3
4
4
3
4
3
4
1
4
0.11
0.13
0.11
0.10
0.09
0.12
0.12
0.08
0.10
0.02
0.01
0.01
0.02
n
(base cations, PO4 )
5
5
5
5
5
5
5
5
6
5
5
5
5
-
Mg
Ca
K
PO4
0.12 ± 0.06
0.11 ± 0.05
0.17 ± 0.03
0.10 ± 0.05
0.16 ± 0.04
0.23 ± 0.07
0.17 ± 0.05
0.26 ± 0.09
0.15 ± 0.04
0.62 ± 0.19
1.03 ± 0.44
0.68 ± 0.36
0.83 ± 0.35
0.25 ± 0.12
0.20 ± 0.09
0.46 ± 0.11
0.43 ± 2.19
0.49 ± 1.30
0.69 ± 2.70
0.73 ± 3.55
1.98 ± 9.01
0.63 ± 2.34
4.51 ± 1.52
6.25 ± 1.77
4.66 ± 2.65
5.45 ± 2.76
0.10 ± 0.04
0.11 ± 0.04
0.10 ± 0.02
0.12 ± 0.02
0.12 ± 0.07
0.12 ± 0.02
0.10 ± 0.01
0.10 ± 0.01
0.13 ± 0.07
0.16 ± 0.03
0.16 ± 0.05
0.15 ± 0.06
0.13 ± 0.03
0.29 ± 0.22
0.14 ± 0.09
0.32 ± 0.27
0.32 ± 0.22
0.17 ± 0.13
0.18 ± 0.16
0.17 ± 0.14
0.05 ± 0.02
0.13 ± 0.11
0.41 ± 0.39
0.22 ± 0.13
0.34 ± 0.32
0.12 ± 0.06
167
Table A.3. Site median values for soil moisture and mean values for soil texture, with standard
deviation intervals where available. Soil volumetric moisture is expressed as volume water per
volume soil. Texture as (% silt + clay) was measured for several samples, but texture as (% fine
sand) was measured for only one sample per site, precluding calculation of standard deviation.
Texture as (% silt + clay) was used for the primary analysis (Table 3, 4), while texture as (% fine
sand) was used for a post-hoc test.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
n
160
172
172
176
174
156
126
126
176
151
128
79
72
Median vol.
3 -3
soil water (m m )
0.094
0.095
0.113
0.107
0.104
0.128
0.995
0.123
0.119
0.113
0.116
0.117
0.137
N
Texture
(% silt + clay)
3
3
3
3
3
3
3
1
3
3
3
2
2
9.3 ± 0.8
9.7 ± 0.8
11.2 ± 1.0
12.3 ± 2.4
9.2 ± 2.9
12.7 ± 2.1
8.8 ± 0.8
13.0 ± NA
9.8 ± 1.4
11.5 ± 1.8
14.2 ± 5.0
11.5 ± 5.0
11.5 ± 5.0
168
n
1
1
1
1
1
1
1
1
1
1
1
1
1
Texture
(% fine sand)
19.9
14.5
14.8
23.5
20.7
9.1
26.0
12.3
30.6
28.0
31.6
32.4
32.4
Table A.4. Soil resource correlation matrix, with correlation coefficients (r) for each pair-wise comparison.
Resource
2
Vol. H2O
2. Texture
-0.112
1.000
+ -0.493
0.027
3. Acidity, Al
2+
0.530
0.046
4. Ca
+
3
4
5
6
7
8
9
10
11
12
13
1.000
-0.959
1.000
0.395
0.473
-0.775
0.757
1.000
2+
0.521
0.151
-0.926
0.988
0.775
1.000
7. PO4
8. C
9. N
10. C:N
11. NO3
-
-0.222
0.477
-0.293
0.197
0.557
0.188
1.000
-0.049
-0.082
-0.005
0.546
0.275
0.251
-0.194
0.138
-0.066
-0.110
0.177
-0.683
0.266
0.209
-0.199
0.791
0.148
0.219
-0.305
0.551
0.336
0.287
-0.271
0.834
-0.067
0.071
-0.137
-0.133
12. NH4
13. ΣN
14. N-min rate
Nitrification
0.231
-0.519
-0.496
0.432
0.207
0.331
-0.044 0.070
0.485
-0.223
0.513
-0.273
-0.131
-0.084
-0.749
0.373
-0.899
0.770 0.472 0.727 -0.110 0.200 0.154 -0.136 0.741 0.811 1.000
-0.435 -0.209 -0.508 -0.029 -0.225 -0.236 0.213 -0.434 0.000 0.065
0.923 0.710 0.911 0.211 0.208 0.215 -0.284 0.751 0.637 0.817
5. K
6. Mg
14
169
1.000
0.813 1.000
-0.482 -0.879 1.000
0.253 0.200 -0.195 1.000
0.050
-0.028 0.208
1.000
1.000
-0.448
Table A.5. AICc comparisons for significant or top non-significant models relating the fraction of total ANPP comprising ANPPW to
soil resources, species composition, and stand density. For Ca and soil water, log-normal models were necessary to accommodate a
maximum reached within the domain. These log-normal models used the following function: (A*exp(-0.5*((Ln(R/B)/C)^2))), where
A, B, and C were estimated parameters and R was the soil resource. Species composition was tested as relative basal area of each
species, but the only proportionate basal area to which the ANPPW fraction was related was the relative basal area of species other
than Q. rubra. Parameter estimates with 95% confidence intervals are presented as A, B, and/or C for the log-normal function, and as
2
A or B for the linear function, with the parameters listed in Table 1. The k value represents parameter number; r and 95% confidence
intervals around each parameter are presented in place of p-values.
Independent variable/
Model
Ca
log-normal
Soil water
log-normal
linear
ΣN
linear
Species composition
linear
Stand density
linear
r
2
k
ΔAICc
A (95% C.I.)
B (95% C.I.)
C (95% C.I.)
0.592
4
0.000
0.598 (0.566, 0.631)
1.736 (1.320, 3.535)
3.184 (2.453, 5.556)
0.514
0.361
4
3
2.297
3.845
0.579 (0.554, 0.604)
0.021 (0.004, 0.037)
12.57 (11.48, 13.66)
0.318 (0.131, 0.504)
-0.511 (-0.687, -0.335)
0.018
3
9.435
0.013 (-0.01, 0.04)
0.514 (0.432, 0.595)
0.389
3
3.267
-0.09 (-0.16, -0.02)
0.608 (0.562, 0.654)
0.374
3
3.578
3.6e-4 (8.3e-5, 6.3e-4)
0.339 (0.017, 0.505)
170
APPENDIX B
171
APPENDIX B
Appendix B.1. Theoretical potential growth modeled using a composite of collected and USFS
FIA data
Theoretical potential growth model formulation—Methods
The FIA data included very few large trees. Combining the two datasets, we identified the
subset of trees with the top 5% of growth rates in each 10-cm diameter class (Gd,COMPOSITE)
and used these individuals to calculate the potential growth coefficient:
Gd,COMPOSITE

= P  exp  0.5ln dbh    
2

(11)
where δ and σ are the same as in equation 4, and P is an estimated coefficient for each focal
species. Theoretically, P may be interpreted as the growth rate of a tree free of neighbor
interactions, growing in optimum resource conditions, and of diameter at which maximum
growth occurs for the given species (Canham et al. 2006, Coates et al. 2009). To model realized
growth as a reduction to potential growth, effects of tree size, neighborhood, and soil resource
were constrained to values between 0 and 1 in a multiplicative equation:
Gd = P  Diameter effect  Neighborhood effect  Soil resource effect
(12)
with each of these terms defined in a similar fashion as above, except that each term was an
exponential decay. As above, we tested the full model (equation 12) for each species, but also
tested several reduced models (Table B.4). The diameter effect was identical to the lognormal
expression in equation 4. In contrast, we used only the neighborhood indices (equations 5 and 6),
both as exponents (using equation 5 as an example):
Neighborhood effect =
 n

exp  B dbhi exp 
 i 1
172
disti 



(13)
The same interpretations as in the main text apply for B, α, and β. For the species-dependent
neighborhood index (exponential decay with equation 6 as the exponent), the interpretation of λi
remains unchanged. The soil resource effect was separate from the neighborhood effect,
Soil resource effect = 1 – exp(C  resource)
(14)
where the exponential decay is subtracted from 1 so that increasing resource availability
decreases the contribution of the size effect (i.e. reduces potential growth less). Performance of
the potential growth framework was compared to that of the realized growth framework by
assessing overall differences in AICc (Table B.7).
Theoretical potential growth model formulation—Results
Models from the realized growth framework had stronger (A. saccharum, A. rubrum, F.
grandifolia, Q. alba, Q. rubra, and T. americana) or equivalent (F. americana, P. grandidentata,
and P. serotina) empirical support than models from the potential growth framework for all
species except Q. velutina (Table B.5). Under both frameworks, diameter was the only correlate
of Q. velutina growth (Figure 1; Table 3, B.5). Differences between realized growth and
extrapolated potential growth were most extreme for small diameters below the range of our
data, but driven by the smaller trees in FIA data (Table B.6). Thus, for the main text of this
paper, we focus on results from the realized growth modeling framework that does not rely on
the FIA data and uses only data that we collected.
173
Table B.2. Site occupancy by species. No single species was present across the entire landform
gradient, although Q. rubra and A. rubrum were present at 11 of 13 sites. Distributions of some
species were disjoint: A. saccharum and T. americana never co-occurred with Q. alba. Typically,
a few sites contained the majority of individuals for each species.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
Total stems
acru
acsa
Percentage stems of a species at a site
fagr fram pogr prse qual quru
---2.9
15.2
23.1
9.3
1.0
4.4
0.1
12.5
0.6
---1.7
29.1
1188
13.5
16.1
---0.1
15.6
6.6
7.7
20.7
---19.1
------0.6
1699
15.7
3.4
---1.7
48.7
1.7
0.8
26.3
0.8
0.4
------0.4
236
48.0
1.0
------5.0
5.0
11.0
24.0
---6.0
---------100
10.2
---2.3
21.5
------23.7
0.6
20.3
0.6
2.8
0.6
17.5
177
---10.0
------24.0
28.0
6.0
20.0
2.0
2.0
------8.0
50
174
------12.0
3.8
------------16.1
---35.0
30.3
2.9
1034
0.2
19.1
12.8
25.8
13.4
1.5
1.3
2.2
7.0
------2.1
14.6
1227
quve tiam
---0.2
16.9
3.9
------------1.7
---50.1
26.3
1.0
415
41.9
------------10.5
13.3
34.3
---------------420
Table B.3. Soil resource availability by site. Mean soil resource measurements for each site.
Values represent the mean of all imputed points used in the growth analysis. Standard deviations
(±) are presented instead of standard error because sample sizes differed substantially.
Site*
n
2+
Ca
(ppt)
Soil H2O (%)
-1
ΣN (mg kg )
-1
-1
N min. (mg kg day )
1, O
591 0.08 (±0.02)
3.78 (±2.56)
4.39 (±2.22)
0.91 (±0.38)
2, O
492 0.07 (±0.02)
2.88 (±2.32)
4.03 (±2.12)
0.61 (±0.34)
3, I
536 0.12 (±0.05)
4.48 (±1.64)
7.38 (±5.69)
1.19 (±0.61)
4, I
468 0.11 (±0.06)
3.28 (±1.62)
7.78 (±4.57)
1.23 (±0.60)
5, mI
690 0.13 (±0.04)
3.17 (±2.18)
14.9 (±8.68)
0.62 (±0.47)
6, mI
616 0.18 (±0.09)
5.42 (±4.16)
15.4 (±12.2)
2.15 (±0.98)
7, pM
563 0.15 (±0.07)
6.17 (±1.27)
10.4 (±6.48)
1.32 (±0.56)
8, pM
346 0.20 (±0.10)
6.36 (±1.72)
4.81 (±1.91)
1.78 (±0.66)
9, iM
678 0.09 (±0.09)
4.89 (±2.88)
8.28 (±5.76)
0.97 (±0.52)
10, M
622 0.84 (±0.47)
5.46 (±1.23)
28.0 (±10.2)
3.52 (±2.07)
11, M
553 1.16 (±0.39)
8.98 (±2.72)
31.3 (±8.10)
9.07 (±4.41)
12, M
318 0.71 (±0.43)
9.58 (±2.17)
20.9 (±12.5)
2.43 (±1.26)
13, M
215 0.71 (±0.43)
9.69 (±1.18)
25.3 (±16.2)
3.13 (±1.35)
*Landform codes: O: outwash; I: ice contact; mI: mesic ice contact; pM: poor moraine; iM;
intermediate moraine; M: Moraine.
175
Table B.4. Biological interpretation of neighborhood model parameters and model variations. Model parameters, their biological
interpretation, and the consequences of modifying the full models by constraining or changing certain parameters.
Models
1
2, 12
2, 12
1, 2, 12
1, 2, 12
1, 2, 12
Parameter
A
δ
σ
B
α
β
1, 2, 12
1, 2, 12
1, 2, 12
γ
λ
R
1, 2, 12
C
12
P
Interpretation
Influence of focal diameter on growth
Diameter expected for maximum growth rate
Rate at which growth increases with diameter
Influence of neighborhood index on growth
Effect of neighbor diameter on focal tree growth
Effect of neighbor distance on focal tree growth
Size-dependency of neighborhood effects
Species-specific neighbor coefficient
Estimated radius within which neighbors influence
growth
Effect of soil resource (coefficient (1) or exponent
(2, 12))
Potential growth extrapolated from dataset or
composite data
176
Modification
------B=0
α=0
β=0
Consequence
----R=x
------Growth is independent of local neighborhood
Neighborhood not influneced by neighbor size
Neighborhood not influneced by neighbor
distance
----Neighborhood radius ≠ 10m
C=0
---
Growth is independent of soil resources
---
Table B.5. Theoretical potential growth model selection and parameters. Supported models for potential growth framework (based on
equation 12 (Appendix B.1)) and comparison to realized growth framework (based on equations 1 and 2). The amount of growth
variance predicted by complex models (those that included at least one term in addition to diameter) was typically higher than the
2
variance explained by diameter alone. This improvement (Add. var. expl.) represents the difference between diameter-only r and the
2
r of the top supported model. If the diameter-only model was supported, then it is not presented separately. Modifications (Table B.4)
to full models, which are denoted by the same equation numbering scheme found in the methods, are identified in superscript.
Top realized growth models
Species
Top potential growth models
2
n
Ngb.
Res. k
AICc
r
Ngb.
Res.
k
acru
554
5
-2174.59
0.348
---
acsa
682
7
H2O 6
H2O 6
-2873.35
0.635
---
--H2O
fagr
fram
pogr
prse
110
30
84
7
6*
---
H2O 6
--- 6
ΣN 4
-485.72
-85.03
-288.53
0.422
0.573
0.239
-------
28
---
-84.42
0.758
---
qual
486
---
-2267.81
0.261
---
quru
quve
tiam
589
184
183
----7†
Ca 4
H2O 4
H2O 4
--- 4
--- 7
----ΣN
H2O
-2305.06
-677.33
-698.75
0.623
0.273
0.463
-------
*β = 0
Realized Potential
†α, β = 0
177
2
2
ΔAICc
Δr
3 -2163.23 0.324
-11.36
0.024
4 -2851.56 0.618
-21.79
0.017
3
3
4
-467.56
-81.46
-285.74
0.233
0.388
0.231
-18.16
-3.56
-2.79
0.189
0.185
0.008
4
-84.30
0.754
-0.12
0.004
---
3 -2252.35 0.240
-15.45
0.021
-------
7 -2300.91 0.622
3 -669.06 0.319
3 -687.89 0.407
-4.15
-8.27
-10.85
0.001
-0.046
0.056
AICc
r
Table B.6. Summary statistics for USFS FIA plot data from northern Lower Michigan.
Mean and maximum diameter and growth are presented with standard deviations. Uniformly, the
maximum diameter values in our dataset exceeded those in the FIA data, but maximum growth
rates in the FIA data typically exceeded those in our data.
Species
acru
acsa
fagr
fram
pogr
prse
qual
quru
quve
tiam
Mean
diameter
(cm)
7.2 ± 3.7
7.8 ± 3.9
7.6 ± 5.1
8.5 ± 4.7
7.3 ± 3.7
7.1 ± 4.1
9.1 ± 4.8
9.6 ± 4.9
10.2 ± 5.2
9.2 ± 4.1
Maximum
diameter
(cm)
29.4
33
36.7
29
27.2
27.7
41.7
38.4
35.9
46.5
Mean growth
2 -1
(cm yr )
Maximum
growth
2 -1
(cm yr )
1.12 ± 1.84
1.11 ± 1.61
0.98 ± 4.99
1.79 ± 2.32
1.64 ± 1.59
1.44 ± 2.24
1.13 ± 2.66
2.21 ± 4.62
2.09 ± 2.60
1.04 ± 10.8
26.99
23.86
9.91
32.32
17.85
22.31
13.71
46.75
18.72
20.72
178
Table B.7. Model framework comparison: realized growth versus theoretical potential growth. Comparison of realized framework
growth models in which neighbor species were treated as equivalent (NS, equations 5 or 6) or different (S, equations 7 or 8). Species
models occasionally fit better, but typically ΔAICc > 2 and the goodness of fit was worse. When the species model both fit better and
received more or equivalent support (ΔAICc < 2), the species model is highlighted in bold. Modifications (Table B.4) to full models,
denoted by the same equation numbering scheme found in the methods, are identified in superscript. General model type (additive or
multiplicative) is indicated in Model column with numbers corresponding to equation numbers in the methods. Similarly,
neighborhood type in the Ngb column corresponds to equation numbers.
Equivalent neighbors (NS)
Different neighbors (S)
2
k
S – NS
r
2
Species
n
Model
Ngb.
Res.
Model
Ngb.
Res.
NS
S
NS
S
Δr
ΔAICc
acru
557
1
5
H2O
1
7
H2O
6
9
0.348
0.329
-0.019
8.79
acsa
682
2
6, 12
H2O
2
8, 12
H 2O
6
10
0.607
0.635
0.028
-7.14
fagr
fram
pogr
prse
qual
113
30
84
27
486
1
1*
2
2
2
5
5†,‡
-------
H2O
--ΣN
Ca
1
1*
2
2
2
7
7†,‡
8
8
8
H 2O
---------
6
6
4
4
4
10
14
11
12
10
0.318
0.573
0.239
0.758
0.261
0.422
0.636
0.177
0.662
0.143
0.104
0.063
-0.062
-0.096
-0.118
-9.03
35.26
21.19
30.12
-3.52
quru
quve
tiam
589
184
183
2
2¶
2
----6‡,§
H2O
----†β = 0
2
2¶
2
8
8
8‡,§
-------
4
4
7
13
9
10
0.623
0.273
0.413
0.580
-0.043
0.295
0.022
0.463
0.050
§α, β = 0
27.77
15.70
-2.16
*Log-normal diameter effect
H2O
‡Estimated radius
179
¶Equivalent to potential growth
Table B.8. Parameter estimates and 95% confidence intervals for species whose growth was
either unrelated to neighborhood or related to a neighborhood that did not account for neighbor
species (terms (5) or (6)). Parameters are reported only for the best-supported model for each
species.
a. Parameter estimates and 95% confidence intervals for A. rubrum.
Species
acru
A
0.0099 (0.0095, 0.0103)
α
0.32 (0.26, 0.37)
B
0.0117 (0.010, 0.014)
β
1.59 (2.42, 1.12)
C
-0.002 (-0.003, -0.001)
Variance
0.085 (0.08, 0.091)
b. Parameter estimates and 95% confidence intervals for F. americana.
Species
fram
δ
225.8 (196, 267.12)
σ
2.02 (1.865, 2.162)
B
0.010 (0.0084, 0.012)
α
0.265 (0.211, 0.324)
radius
7.86 (7.52, 8.01)
Variance
0.124 (0.087, 0.151)
c. Parameter estimates and 95% confidence intervals for species for which growth was not
related to local neighborhood.
δ
σ
2727.7 (2452,
2.42 (2.36,
0.57
(0.52,
3088) 75.4)
2.48)
70.3 (66.1,
0.62)
2235.1 (2116,
2.26 (2.23,
2375)
2.28)
174.5
(171.1,
1.11
(1.09,
1.12)
* See Table 3 for178.2)
resources modified by
C
Specie
s
pogr
prse
qual
quru
C*
variance
0.14 (0.1, 0.17)
0.10 (0.09, 0.12)
0.23 (0.14, 0.36)
0.10 (0.08, 0.14)
-0.13 (-0.19, -0.09) 0.06 (0.058, 0.063)
-0.14 (-0.16, -0.13) 0.086 (0.08, 0.09)
d. Parameter estimates and 95% confidence intervals for Q. velutina and T. americana.
Species
quve
tiam
A
0.28 (0.26,
0.30)
----
δ
41.7 (38.8, 45.0)
161.2 (154.5,
168.3)
.3)
σ
0.70 (0.62, 0.78)
0.84 (0.81, 0.86)
180
variance
0.093 (0.085,
0.092
(0.082,
0.105)
0.102)
Table B.9. Estimated model parameters for species-dependent neighborhoods.
a. Parameter estimates and 95% confidence intervals of species for which growth was best
explained by models that included neighbor species identity
fagr
δ
159.2 (155.6, 164.5)
A
0.005 (0.004, 0.005)
σ
1.123 (1.103, 1.137)
B
0.004 (0.003, 0.004)
B
0.007 (0.006, 0.009)
C
0.006 (0.004, 0.008)
Species
acsa
fagr
α
1.319 (1.281, 1.376)
0.817 (0.740, 0.895)
β
0.625 (0.518, 0.712)
0.331 (0.403, 0.280)
variance
0.076 (0.070, 0.079)
0.064 (0.057, 0.074)
Species
acsa
b. Parameter estimates and 95% confidence intervals for neighbor species influence parameters
(λ). Strongest pair-wise interactions are presented in bold. To provide a quantitative comparison
of each species effect, we normalized the neighborhood parameters relative to the largest value
(normalized λ presented in first set of parentheses, confidence interval in second set). For
example, intraspecific interactions for A. saccharum have a value of 1.
Focal
Neighbor
Species acsa
Species
acsa
fagr
fagr
λ
0.878 (1.0)
0.898
(1.0)
(0.794,
(0.637,
0.999)
1.287)
Neighbor
Species
tiam
acsa
λ
Neighbor
0.627 (0.71) all others
0.78
(0.86) all others
(0.397,
(0.313,
0.935)
1.126)
181
λ
0.405 (0.46)
0.013
(0.01)
(-0.031,
(0,0.611)
0.102)
Table B.10. Mean growth (as basal area increment, cm2 year-1) for each species at each site;
standard deviation is in parentheses.
Site
acru
1
--2 2.44 (± 2.76)
3 3.35 (± 3.08)
4 3.55 (± 4.47)
5 7.01 (± 8.99)
6 10.39 (± 9.73)
7 9.85 (± 8.73)
8
0.38 (± 0)
9 3.71 (± 4.08)
10 14.06 (± 8.16)
11
--12 6.86 (± 4.46)
13 2.31 (± 2.57)
Table B.10, continued.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
prse
--7.35 (± 8.7)
----7.33 (± 8.07)
21.01 (± 18.64)
12.49 (± 0.47)
16.79 (± 12.11)
1.14 (± 0)
16.79 (± 0)
----2.83 (± 2.29)
acsa
10.08 (± 11.52)
3.21 (± 3.7)
--1.59 (± 1.1)
2.05 (± 3.06)
7.24 (± 10.01)
5.03 (± 7.84)
4.75 (± 6.22)
--3.75 (± 4.6)
----3.01 (± 3.4)
fagr
5.35 (± 5.9)
4.5 (± 5.81)
--6.17 (± 5.73)
3.07 (± 4.81)
1.09 (± 1.39)
0.58 (± 0.07)
5.49 (± 6.64)
3.59 (± 1.28)
0.64 (± 0)
----3.07 (± 0)
qual
----3.38 (± 3.96)
4.6 (± 4.43)
--------4.77 (± 4.54)
--2.94 (± 2.72)
3.32 (± 3.66)
5.83 (± 6.07)
quru
20.9 (± 30.56)
14.65 (± 8.89)
13.39 (± 11.73)
11.41 (± 8.56)
19.99 (± 13.95)
32.05 (± 11.6)
29.15 (± 17.84)
41.84 (± 13.41)
14.48 (± 10.06)
----16 (± 9.35)
17.21 (± 13.34)
182
fram
19.07 (± 17.53)
2.05 (± 0)
----22.01 (± 9.09)
16.39 (± 16.37)
8.42 (± 7.22)
11.75 (± 10.87)
--4.18 (± 6.6)
-------
quve
--2.6 (± 0)
13.03 (± 8.95)
15.91 (± 10.49)
--------14.08 (± 9.42)
--9.31 (± 7.07)
12.34 (± 8.7)
15.67 (± 3.35)
pogr
20.82 (± 11.19)
--34.26 (± 18.85)
15.19 (± 6.19)
----7.91 (± 3.68)
26.36 (± 0)
17.28 (± 12.1)
6.58 (± 0)
2.88 (± 1.82)
2.78 (± 0)
12.64 (± 6.05)
tiam
5.95 (± 6.98)
--------5.26 (± 5.15)
8.12 (± 7.28)
8.76 (± 8.67)
-----------
Table B.11. Mean diameter (cm) for each species at each site; standard deviation is in
parentheses.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
acru
acsa
fagr
--22.6 (± 10.2) 19.1 (± 8.3)
15.7 (± 4.5)
16.2 (± 6)
14 (± 5.2)
14.7 (± 5.2)
----15.7 (± 5.4) 10.3 (± 0.3) 15.7 (± 10.5)
20.2 (± 7.9) 16.2 (± 5.3) 18.2 (± 7.8)
24.4 (± 6.6) 19.9 (± 11.7)
12.6 (± 2)
25.2 (± 8.2) 18.6 (± 8.4) 12.2 (± 1.4)
12 (± 0)
19.4 (± 7.5) 23.7 (± 8.8)
17.9 (± 7.9)
--10.1 (± 0.2)
26.4 (± 7.4) 16.6 (± 4.9)
10 (± 0)
------14.6 (± 3.6)
----16.7 (± 5.7) 15.7 (± 5.2)
12.3 (± 0)
fram
31.3 (± 13)
15.9 (± 0)
----33.8 (± 5.9)
29.3 (± 11.7)
22.1 (± 7.5)
29.4 (± 7.8)
--27.8 (± 6.1)
-------
pogr
36.5 (± 9.5)
--39.2 (± 15.1)
32.2 (± 6.1)
----23.7 (± 4.2)
46.2 (± 0)
37.3 (± 8.3)
64.1 (± 0)
11.9 (± 2.7)
17.2 (± 0)
36.1 (± 5.7)
Table B.11, continued.
Site
1
2
3
4
5
6
7
8
9
10
11
12
13
prse
--25.6 (± 7.1)
----29.2 (± 7.4)
32.2 (± 8.7)
26.4 (± 1.6)
30.5 (± 7.8)
10.8 (± 0)
41.5 (± 0)
----25.3 (± 6.2)
qual
----22 (± 6.7)
27 (± 6.1)
--------26.9 (± 5.8)
--15.6 (± 5.4)
18.4 (± 7.8)
28.4 (± 8)
quru
quve
38.4 (± 42.4)
--33.4 (± 7.3)
27.3 (± 0)
29.1 (± 9.4) 33.4 (± 9.1)
27.3 (± 6.5)
33 (± 7.4)
38.8 (± 10.2)
--53.1 (± 11.4)
--45 (± 16.9)
--56.9 (± 8.1)
--34.3 (± 8.5) 33.6 (± 10.6)
------25.2 (± 9.8)
28.5 (± 9.1) 29.3 (± 9.7)
36.1 (± 12.5) 39.3 (± 4.5)
183
tiam
28.7 (± 5.5)
--------30 (± 8.2)
31 (± 7.6)
30.7 (± 7)
-----------
APPENDIX C
184
APPENDIX C
Table C.1. Site attributes, including mean soil Ca and inorganic N content, stem density,
dominant species, stand age, and location.
Site
a
ID
1, O
2, O
3, I
4, I
5, mI
6, mI
7, pM
8, iM
9, M
10, M
11, M
a
Ca
-1
(mg kg )
NO3 + NH4
(mg kg )
Density
-1
(Stems ha )
Dominant
b
Species
80
70
120
110
130
180
150
90
840
1160
710
4.4
4.0
7.4
7.8
14.9
15.4
10.4
8.3
28.0
31.3
22.5
582
510
555
486
725
623
568
712
642
544
506
Qa
Qa
Qr, Ar
Qr, Ar
Qr, Ar
Qr, Ar
As, Qr
As, Qr
As
As
As, Qr
-1
Stand
age
(years)
88
80
97
111
83
90
79.0
80
84
76
92
Location
44°14‘N, 85°56‘W
44°14‘N, 85°57‘W
44°16‘N, 85°53‘W
44°17‘N, 85°53‘W
44°19‘N, 85°52‘W
44°12‘N, 85°48‘W
44°11‘N, 85°45‘W
44°22‘N, 85°49‘W
44°15‘N, 85°45‘W
44°13‘N, 85°40‘W
44°13‘N, 85°45‘W
Landform codes: O: outwash; I: ice contact; mI: mesic ice contact; pM: poor moraine; iM;
intermediate moraine; M: Moraine.
b
Species codes: Ar: Acer rubrum; As: Acer saccharum; Qa: Quercus alba; Qr: Quercus rubra.
185
Table C.2. Soil resource content before and after fertilizer application, aggregated for all
species.
Treatment
Control
Ca
N
Ca, N
Ca, N, µ
Resource
Ca
K
Mg
P
PO4
NH4
NO3
Ca
K
Mg
P
PO4
NH4
NO3
Ca
K
Mg
P
PO4
NH4
NO3
Ca
K
Mg
P
PO4
NH4
NO3
Ca
K
Mg
P
PO4
NH4
NO3
Soil resource concentration (ppm)
Prior
Post
Change
227.84 ± 72.83
187.21 ± 62.52
-40.63
29.38 ± 6.3
27.14 ± 2.73
-2.24
26.01 ± 2.74
19.55 ± 2.75
-6.46
32.28 ± 2.53
25.21 ± 1.55
-7.07
71.31 ± 14.27
51.65 ± 15.18
-19.66
18.95 ± 1.53
5.38 ± 1.2
-13.57
15.54 ± 8.3
15.85 ± 11.05
0.31
245.18 ± 70.97
258.81 ± 79.84
13.63
26.7 ± 3.85
25.86 ± 2.03
-0.84
27.52 ± 4.17
26.21 ± 5.29
-1.31
27.74 ± 2.02
23.91 ± 1.75
-3.83
67.15 ± 16.36
56.8 ± 16.35
-10.35
18 ± 4.02
6.14 ± 0.78
-11.86
27.57 ± 11.14
24.12 ± 18.48
-3.45
233.97 ± 66.76
234.91 ± 80.03
0.94
26.58 ± 4.87
28.22 ± 3.67
1.63
34.75 ± 6.41
22.8 ± 3.35
-11.95
30.58 ± 2.12
27.45 ± 1.99
-3.13
105.28 ± 19.61
65.71 ± 13.56
-39.57
20.46 ± 3.11
12.41 ± 2
-8.06
41.37 ± 11.07
25.71 ± 15.37
-15.67
159.68 ± 44.76
164.25 ± 36.8
4.57
22.13 ± 2.08
24.77 ± 1.91
2.64
30.48 ± 4.66
20.67 ± 1.78
-9.81
41.07 ± 6.35
28.65 ± 2.41
-12.42
108.69 ± 17.44
68.02 ± 13.62
-40.67
27.34 ± 5.78
10.29 ± 1.49
-17.05
15.07 ± 8.5
5.57 ± 2.77
-9.49
239.97 ± 99.95 240.43 ± 106.13
0.46
30.19 ± 7.96
31.87 ± 1.91
1.68
24.67 ± 2.33
31.49 ± 10.84
6.82
31.5 ± 2.57
33.49 ± 2.96
1.99
73.81 ± 17.4
92.24 ± 9.27
18.43
15.19 ± 2.13
11.53 ± 2.27
-3.66
44.65 ± 17.07
17.82 ± 10.8
-26.82
186
Table C.3. Soil pH before and after fertilizer addition, aggregated for all species. Measured
changes indicated that the fertilizer treatment did not effectively alter pH.
Treatment
Control
Ca
N
Ca, N
Ca,N,µ
pH, prior
pH, post
pH, change
4.51 ± 0.13 4.63 ± 0.12
0.12
4.54 ± 0.19 4.59 ± 0.16
0.05
4.52 ± 0.13 4.53 ± 0.13
0.01
4.42 ± 0.13 4.3 ± 0.11
-0.12
4.35 ± 0.23 4.61 ± 0.2
0.26
187
Table C.4. Mean values per treatment for each species for chlorophyll fluorescence yield and
maximum.
Species
Treatment
A. saccharum Ca
N
Ca, N
Ca, N, µ
Control
Q. rubra
Ca
N
Ca, N
Ca, N, µ
Control
A. rubrum
Ca
N
Ca, N
Ca, N, µ
Control
Q. alba
Ca
N
Ca, N
Ca, N, µ
Control
Yield
(ΔF/Fm)
Max. (Fv/Fm)
715.3 ± 19.9
732.7 ± 11.6
735.1 ± 11.8
730 ± 22.5
719.2 ± 18.1
769.3 ± 3.9
768.3 ± 5.3
773 ± 3
753.6 ± 6.2
770.9 ± 3.9
771.2 ± 5.7
768.1 ± 5.6
772 ± 3.7
780.8 ± 3.1
762 ± 7.9
752.3 ± 9.3
763.3 ± 4
752.4 ± 5.9
764.3 ± 5.5
759.5 ± 3.2
2007 ± 104.4
2021.5 ± 72.3
1966.5 ± 60.7
2035.3 ± 114.3
1978.3 ± 91.1
1577.6 ± 51.8
1605.5 ± 82.7
1580.9 ± 51
1518.7 ± 59.7
1613.2 ± 64
2008.3 ± 53.1
1968.2 ± 74
2025.5 ± 55.7
2013.3 ± 55.4
1911.4 ± 110.7
1661.7 ± 46.8
1677.7 ± 40.5
1721 ± 60.8
1712.2 ± 57
1691 ± 51.8
188
Table C.5. Leaf nutrient content of base cations and P.
Species
Treatment
A. saccharum Ca
N
Ca, N
Ca, N, µ
Control
Q. rubra
Ca
N
Ca, N
Ca, N, µ
Control
A. rubrum
Ca
N
Ca, N
Ca, N, µ
Control
Q. alba
Ca
N
Ca, N
Ca, N, µ
Control
-1
-1
-1
-1
Ca (mg g )
K (mg g )
Mg (mg g )
P (mg g )
9.84 ± 0.89
8.99 ± 0.64
8.9 ± 0.31
7.94 ± 0.4
7.28 ± 0.58
5.86 ± 0.21
5.98 ± 0.24
5.29 ± 0.22
5.61 ± 0.22
5.19 ± 0.3
6.13 ± 0.43
6.29 ± 0.68
6.23 ± 0.38
6.13 ± 0.46
6.22 ± 0.46
6.6 ± 0.36
6.1 ± 0.27
6.45 ± 0.4
6.49 ± 0.3
5.52 ± 0.32
6.11 ± 0.57
5.78 ± 0.26
6.27 ± 0.46
6.75 ± 0.37
5.78 ± 0.31
7.32 ± 0.23
7.53 ± 0.47
7.25 ± 0.38
7.25 ± 0.43
6.91 ± 0.22
6.23 ± 0.37
6.48 ± 0.49
6.43 ± 0.35
7.59 ± 0.57
5.77 ± 0.33
7.19 ± 0.36
6.59 ± 0.26
6.84 ± 0.37
7.23 ± 0.3
7.18 ± 0.32
1.9 ± 0.15
1.53 ± 0.09
1.51 ± 0.05
1.36 ± 0.05
1.37 ± 0.12
1.35 ± 0.07
1.43 ± 0.08
1.33 ± 0.06
1.35 ± 0.07
1.39 ± 0.06
1.44 ± 0.07
1.5 ± 0.14
1.46 ± 0.06
1.5 ± 0.1
1.55 ± 0.12
1.25 ± 0.07
1.1 ± 0.06
1.15 ± 0.06
1.18 ± 0.07
1.08 ± 0.05
1.74 ± 0.19
1.11 ± 0.08
1.52 ± 0.13
1.66 ± 0.17
1.22 ± 0.1
1.39 ± 0.05
1.35 ± 0.06
1.33 ± 0.04
1.45 ± 0.06
1.35 ± 0.05
1.4 ± 0.09
1.24 ± 0.08
1.27 ± 0.1
1.43 ± 0.1
1.54 ± 0.12
1.48 ± 0.05
1.26 ± 0.09
1.12 ± 0.08
1.4 ± 0.09
1.47 ± 0.06
189
Table C.6. Content of N and C in leaves.
Species
Treatment
A. saccharum Ca
N
Ca, N
Ca, N, µ
Control
Q. rubra
Ca
N
Ca, N
Ca, N, µ
Control
A. rubrum
Ca
N
Ca, N
Ca, N, µ
Control
Q. alba
Ca
N
Ca, N
Ca, N, µ
Control
Leaf N (%)
1.54 ± 0.07
1.63 ± 0.08
1.61 ± 0.05
1.5 ± 0.08
1.51 ± 0.08
2.25 ± 0.04
2.28 ± 0.07
2.27 ± 0.05
2.29 ± 0.08
2.21 ± 0.07
1.59 ± 0.05
1.83 ± 0.11
1.76 ± 0.06
1.68 ± 0.05
1.63 ± 0.05
2.2 ± 0.05
2.31 ± 0.05
2.28 ± 0.04
2.35 ± 0.07
2.21 ± 0.05
Leaf C (%)
44.98 ± 1.33
42.03 ± 1.64
44.46 ± 1.13
44.75 ± 1.49
43.76 ± 1.61
48.06 ± 0.74
48.66 ± 1.65
48.63 ± 0.85
49.57 ± 0.69
50.69 ± 1.82
47.15 ± 0.82
48.23 ± 1.24
46.39 ± 0.91
49.41 ± 4.62
46.4 ± 1.57
48.49 ± 0.33
47.98 ± 0.65
48.13 ± 0.35
49.09 ± 1.35
48.25 ± 0.67
190
Leaf C : N
30.49 ± 1.59
26.54 ± 1.22
27.98 ± 0.91
30.87 ± 1.77
30.53 ± 2.05
21.64 ± 0.57
21.58 ± 0.74
21.76 ± 0.8
22.07 ± 0.82
23.44 ± 1.32
30.32 ± 1.31
27.83 ± 1.81
26.74 ± 0.95
29.5 ± 2.51
28.63 ± 1.16
22.3 ± 0.48
20.98 ± 0.48
21.27 ± 0.45
21.01 ± 0.44
22.04 ± 0.48
Table C.7. Percent canopy openness.
Species
Treatment
A. saccharum Ca
N
Ca, N
Ca, N, µ
Control
Q. rubra
Ca
N
Ca, N
Ca, N, µ
Control
A. rubrum
Ca
N
Ca, N
Ca, N, µ
Control
Q. alba
Ca
N
Ca, N
Ca, N, µ
Control
Canopy openness
(%)
16.64 ± 2.01
18.95 ± 2.99
13.17 ± 1.85
14.66 ± 3.03
15.4 ± 1.93
8.45 ± 2.02
6.93 ± 1.94
7.9 ± 0.91
12.04 ± 2.42
7.96 ± 1.15
3.85 ± 0.83
3.18 ± 0.4
4.59 ± 0.85
2.55 ± 0.73
4.69 ± 1.72
7.76 ± 0.92
5.88 ± 0.97
5.79 ± 0.95
7.28 ± 0.99
6.45 ± 0.79
191
Table C.8. Diameter growth and BAI for each species, comparing treatments with and without
Ca, and with and without N. Presented as mean ± standard error to compensate for differing
sample sizes.
Species
Ca
No Ca
N
No N
-1
Diam. growth (cm year )
A.
saccharum
Q. rubra
A. rubrum
Q. alba
0.091 ± 0.012
0.337 ± 0.015
0.275 ± 0.026
0.205 ± 0.01
0.102 ± 0.011
0.098 ± 0.01
0.33 ± 0.016
0.334 ± 0.013
0.277 ± 0.026
0.269 ± 0.024
0.192 ± 0.01
0.205 ± 0.01
2
-1
BAI (cm year )
0.091 ± 0.014
0.335 ± 0.02
0.286 ± 0.032
0.192 ± 0.011
A.
saccharum
Q. rubra
A. rubrum
Q. alba
5.121 ± 0.637
26.37 ± 1.769
15.131 ± 1.981
10.618 ± 0.612
6.322 ± 0.818
26.619 ± 2.043
13.961 ± 1.631
9.625 ± 0.65
5.158 ± 0.825
27.578 ± 2.301
16.173 ± 2.513
9.978 ± 0.72
192
5.925 ± 0.638
25.629 ± 1.602
13.743 ± 1.587
10.362 ± 0.577
-1
Table C.9.1. Parameters for simple linear regression of annual diameter growth (cm year ) as a
function of predictor variables in which the relationship between growth and the predictor
differed significantly in a treatment versus the control category.
Species
Predictor
Treatment
Q. rubra
Leaf K (mg g )
Q. alba
Leaf Mg (mg g )
-1
-1
R
2
Slope
Intercept
Ca*
Control
0.264
0.157
-0.26 (-0.45, -0.07)
0.15 (-0.02, 0.31)
0.69 (0.42, 0.95)
0.13 (-0.11, 0.37)
Ca, N*
Control
0.380
0.063
-0.03 (-0.05, -0.01)
0.01 (-0.009, 0.03)
0.43 (0.29, 0.57)
0.09 (-0.06, 0.25)
193
2
-1
Table C.9.2. Parameters for simple linear regression of annual basal area increment (BAI) (cm year ) as a function of predictor
variables in which the relationship between growth and the predictor differed significantly in a treatment versus the control category.
Species
Predictor
Treatment
R
A. rubrum
Diam. (cm), 2007
Ca, N, µ*
Control
Q. alba
Leaf K (mg g )
Q. rubra
Leaf Mg (mg g )
Q. alba
Leaf P (mg g )
2
Slope
Intercept
0.860
0.031
1.76 (1.05, 2.46)
0.16 (-0.73, 1.04)
-37.0 (-58.3, -15.6)
7.38 (-20.89, 35.65)
Ca, N†
Control
0.211
0.013
-1.5 (-2.85, -0.15)
0.38 (-1.26, 2.02)
20.83 (11.3, 30.36)
6.39 (-5.63, 18.42)
Ca‡
Control
0.176
0.058
-23.5 (-45.97, -1.03)
11.1 (-10.96, 33.17)
59.7 (28.63, 90.77)
11.52 (-19.72, 42.8)
Ca, N†
0.208
Control
0.001
Q. rubra
C:N
Ca, N, µ‡ 0.328
Control
0.014
Q. rubra
% Canopy open.
Ca†
0.175
N†
0.291
Control
0.001
*Treatment and control slopes had exclusive 95% CI.
-6.46 (-12.35, -0.58)
-0.66 (-9.06, 7.75)
-2.1 (-3.83, -0.38)
-0.23 (-1.36, 0.89)
-0.76 (-1.49, -0.03)
-0.93 (-1.73, -0.13)
0.06 (-1.06, 1.19)
17.81 (10.83, 24.79)
10.1 (-2.49, 22.68)
69.7 (31.13, 108.2)
34.99 (7.83, 62.15)
34.47 (25.11, 43.83)
32.21 (23.77, 40.65)
26.49 (15.88, 37.09)
-1
-1
-1
†Treatment slope differs from zero (95% CI excludes zero), but overlaps 95% CI of non-significant slope of control by > 29%
‡Treatment slope differs from zero (95% CI excludes) and significantly differs from the control (overlap of 95% CI is < 29%)
194
-1
Table C.10.1. Parameters for multiple linear regression of diameter growth (cm year ) as a function of initial diameter and another
predictor variable for treatment categories with significant slope parameters for both independent variables.
Species
Predictor
Treatment
Q. rubra
Leaf Mg (mg g )
-1
Ca*
Control
R
2
Slope (Iniditial dbh)
0.522 0.007 (0.0026, 0.011)
0.349 0.004 (0.0002, 0.007)
195
Slope (Predictor)
Intercept
-0.224 (-0.3826, -0.065)
0.16 (0.0059, 0.314)
0.29 (-0.03, 0.62)
-0.07 (-0.36, 0.22)
2
-1
Table C.10.2. Parameters for multiple linear regression of basal area increment (BAI) (cm year ) as a function of initial diameter
and another predictor variable for treatment categories with significant slope parameters for both independent variables.
Species
Q. rubra
Predictor
Treatment
-1
Leaf Ca (mg g )
-1
A. saccharum Leaf K (mg g )
Q. alba
Q. rubra
R
2
Slope (Initial dbh)
Slope (Predictor)
Intercept
Ca, N†
Control
0.801
0.594
0.67 (0.47, 0.87)
0.78 (0.44, 1.12)
1.96 (0.05, 3.86)
0.53 (-2.36, 3.42)
-15.21 (-26.24, -4.19)
-14.67 (-35.64, 6.29)
N†
Control
N†
Ca, N*
Control
0.534
0.001
0.486
0.454
0.706
0.3 (0.03, 0.58)
-0.001 (-0.75, 0.75)
0.3 (0.15, 0.45)
0.33 (0.09, 0.57)
0.54 (0.36, 0.71)
2.59 (1.01, 4.18)
0.09 (-1.98, 2.15)
1.16 (0.17, 2.15)
-1.77 (-2.94, -0.6)
0.6 (-0.33, 1.53)
-20.06 (-34.81, -5.31)
3.68 (-25.62, 32.97)
-6.91 (-15.38, 1.56)
11.81 (1.38, 22.23)
-12.21 (-21.41, -3.02)
-17.66 (-31.96, -3.35)
13.92 (0.63, 27.21)
-0.09 (-0.17, -0.01)
-0.03 (-0.13, 0.07)
-2.94 (-32.45, 26.58)
-33.06 (-57.9, -8.22)
65.15 (7.59, 122.71)
17.27 (-63.82, 98.36)
-1
Leaf Mg (mg g )
Ca*
0.690
1.11 (0.72, 1.5)
Control
0.682
0.81 (0.52, 1.11)
Q. alba
Chl. Fluor.
Ca, N†
0.344
0.33 (0.08, 0.59)
Control
0.652
0.50 (0.29, 0.71)
*Treatment and control predictor slopes had exclusive 95% CI.
†Predictor slope differs from zero (95% CI excludes zero), but overlaps 95% CI of non-significant predictor slope of control by >
29%.
196
APPENDIX D
197
APPENDIX D
Table D.1.1. Stand-level parameters, including mean diameter, diameter and basal area
increments with minimum and maximum values provided parenthetically, and density and
mortality rates.
Site
1
2
3
4
5
Diameter (cm)
14.98 (5, 224)
14.96 (5, 140.1)
14.73 (5, 88.5)
15.44 (5, 118.7)
16.30 (5, 168.3)
Diameter
increment
-1
(mm yr )
Basal area
increment
2 -1
(cm yr )
Density
-1
(stems ha )
Mortality
-1
(stems yr )
2.3 (0, 49.4)
2.5 (0, 35.9)
2.3 (0, 29.4)
4.1 (0, 33.7)
3.3 (0, 39.9)
7.92 (0, 365.5)
8.71 (0, 532.9)
7.83 (0, 221.3)
13.00 (0, 255)
11.31 (0, 209)
881
867
1021
824
663
18.2
16.9
14.8
27.6
20.6
198
Table D.1.2. Stand-level mean resource measurements, with standard deviation (units of ppm).
2+
Ca
Site
1
2
3
4
5
106.82 ± 50.5
89.75 ± 40.3
58.9 ± 3.5
153.61 ± 28.0
112.48 ± 2.1
Site
1
2
3
4
5
P
1.99 ± 0.3
1.84 ± 0.8
1.71 ± 0.7
5.08 ± 1.2
3.32 ± 0.1
+
K
82.17 ± 7.3
78.61 ± 8.0
67.56 ± 6.4
116.71 ± 14. 8
124.26 ± 1.6
2+
Mg
54.08 ± 17.2
45.64 ± 48.0
35.53 ± 6.9
66 ± 16.9
37.98 ± 3.0
NH4
NO3
159.57 ± 23.7
180.3 ± 48.0
120.93 ± 6.9
95.79 ± 16.9
112.51 ± 3.0
99.52 ± 21.8
110.6 ± 19.93
100.17 ± 14.4
74.26 ± 3.76
61.65 ± 1.68
199
Table D.2. Model parameters, their biological interpretation, and the consequences of modifying the full models by constraining or
changing certain parameters.
Parameter
P
δ
σ
B
α
β
γ
R
C
Interpretation
Theoretical potential maximum growth rate
Diameter expected for maximum growth rate
Rate at which growth increases with diameter
Influence of neighborhood index on growth
Effect of neighbor diameter on focal tree growth
Modification
------B=0
α=0
Effect of neighbor distance on focal tree growth
Size-dependency of neighborhood effects
Estimated radius within which neighbors influence growth
Effect of soil resource (coefficient (1) or exponent (2, 12))
200
β=0
--R=x
C=0
Consequence
------Growth is independent of local neighborhood
Neighborhood not influneced by neighbor size
Neighborhood not influneced by neighbor
distance
--Neighborhood radius ≠ 20m
Growth is independent of soil resources
2
Table D.3. Parameter estimates and R values for significant linear regressions of stand-level
2
-1
mean growth of all species (mean basal area increment (cm yr )) as a function of soil
resources. Mean basal area growth was related to P, K, and Ca, but no other resources.
Resource
P
K
Ca
2
R
0.958
0.790
0.744
Slope
1.533 (0.943, 2.123)
0.080 (0.004, 0.155)
0.056 (-0.004, 0.117)
Intercept
5.419 (3.608, 7.230)
2.203 (-5.09, 9.500)
3.841 (-2.74, 10.42)
201
Table D.4.1. Resource covariance for all species together and for functional groups.
Ca
Ca
K
Mg
P
NH4
NO3
1.000
0.798
0.824
0.898
-0.430
-0.619
Ca
K
Mg
P
NH4
NO3
1.000
0.820
0.843
0.920
-0.504
-0.669
Ca
K
Mg
P
NH4
NO3
1.000
0.788
0.792
0.898
-0.394
-0.635
Ca
K
Mg
P
NH4
NO3
1.000
0.795
0.827
0.887
-0.438
-0.635
K
Mg
P
All species
0.798 0.824 0.898
1.000 0.320 0.840
0.320 1.000 0.653
0.840 0.653 1.000
-0.625 -0.151 -0.744
-0.936 -0.115 -0.762
Fabaceae
0.820 0.843 0.920
1.000 0.387 0.868
0.387 1.000 0.698
0.868 0.698 1.000
-0.667 -0.237 -0.770
-0.951 -0.190 -0.776
Arecaceae
0.788 0.792 0.898
1.000 0.249 0.826
0.249 1.000 0.610
0.826 0.610 1.000
-0.588 -0.088 -0.702
-0.934 -0.093 -0.777
Non-legume dicot species
0.795 0.827 0.887
1.000 0.320 0.851
0.320 1.000 0.629
0.851 0.629 1.000
-0.630 -0.145 -0.741
-0.937 -0.131 -0.767
202
NH4
NO3
-0.430
-0.625
-0.151
-0.744
1.000
0.796
-0.619
-0.936
-0.115
-0.762
0.796
1.000
-0.504
-0.667
-0.237
-0.770
1.000
0.772
-0.669
-0.951
-0.190
-0.776
0.772
1.000
-0.394
-0.588
-0.088
-0.702
1.000
0.804
-0.635
-0.934
-0.093
-0.777
0.804
1.000
-0.438
-0.630
-0.145
-0.741
1.000
0.795
-0.635
-0.937
-0.131
-0.767
0.795
1.000
Table D.4.2. Resource covariance for wood density groups.
Ca
K
Mg
P
NH4
NO3
-0.384
-0.610
-0.081
-0.719
1.000
0.798
-0.621
-0.923
-0.123
-0.746
0.798
1.000
-0.400
-0.612
-0.159
-0.693
1.000
0.759
-0.618
-0.948
-0.144
-0.764
0.759
1.000
-0.440
-0.651
-0.165
-0.807
1.000
0.784
-0.610
-0.942
-0.111
-0.779
0.784
1.000
-0.547
-0.615
-0.309
-0.575
1.000
0.834
-0.671
-0.901
-0.233
-0.680
0.834
1.000
-3
Ca
K
Mg
P
NH4
NO3
1.000
0.779
0.824
0.864
-0.384
-0.621
Ca
K
Mg
P
NH4
NO3
1.000
0.812
0.849
0.925
-0.400
-0.618
Ca
K
Mg
P
NH4
NO3
1.000
0.794
0.828
0.881
-0.440
-0.610
Ca
K
Mg
P
NH4
NO3
1.000
0.873
0.856
0.980
-0.547
-0.671
0.2-0.4 g cm
0.779 0.824 0.864
1.000 0.290 0.833
0.290 1.000 0.600
0.833 0.600 1.000
-0.610 -0.081 -0.719
-0.923 -0.123 -0.746
-3
0.4-0.55 g cm
0.812 0.849 0.925
1.000 0.386 0.864
0.386 1.000 0.712
0.864 0.712 1.000
-0.612 -0.159 -0.693
-0.948 -0.144 -0.764
-3
0.55-0.7 g cm
0.794 0.828 0.881
1.000 0.322 0.842
0.322 1.000 0.639
0.842 0.639 1.000
-0.651 -0.165 -0.807
-0.942 -0.111 -0.779
-3
0.7-1.2 g cm
0.873 0.856 0.980
1.000 0.497 0.858
0.497 1.000 0.826
0.858 0.826 1.000
-0.615 -0.309 -0.575
-0.901 -0.233 -0.680
203
2
2
-1
(basal area increment (cm yr )) as a function of soil resources for functional groups. Compared
to other functional groups, the Fabaceae show the weakest relationships of growth to soil
resources—no model had significant parameters. Sample sizes presented in Table D.14.
Resource
R
2
Slope
FC—no significant parameter estimates
F<C—no significant parameter estimates
AS
P
0.868 2.699 (0.764, 4.634)
A<S
P
0.961 0.337 (0.212, 0.462)
Ca
0.843 0.013 (0.003, 0.023)
NLDC
K
0.864 0.121 (0.033, 0.210)
NLD<C
P
0.958 1.046 (0.644, 1.448)
K
0.822 0.057 (0.008, 0.107)
Ca
0.816 0.041 (0.005, 0.077)
Intercept
-2.455 (-8.261, 3.352)
-0.284 (-0.683, 0.115)
-0.687 (-1.824, 0.451)
4.957 (-3.580, 13.50)
2.150 (0.898, 3.402)
-0.317 (-5.073, 4.44)
0.797 (-3.118, 4.713)
204
Table D.6. Model parameters and 95% confidence intervals (parenthetically) for functional groups.
Analytical
Category
FC
F<C
F<C
AS
A<S
NLDC
NLDC
NLD<C
δ
29.7 (25.8, 34.0)
36.1 (29.9, 44.3)
34.9 (29.2, 43.2)
9.9 (9.5, 10.3)
11.7 (0.7, 49.0)
9.8 (8.4, 11.2)
10.0 (8.6, 11.4)
144 (131, 159)
σ
1.03 (0.90, 1.20)
0.96 (0.82, 1.10)
0.93 (0.80, 1.07)
0.38 (0.35, 0.43)
2.55 (0.86, 4.24)
2.99 (2.67, 3.31)
2.98 (2.67, 3.29)
1.98 (1.90, 2.04)
P
1.03 (0.94, 1.10)
0.73 (0.60, 0.86)
0.77 (0.64, 0.91)
3.61 (3.34, 3.94)
1.13 (0.82, 1.45)
0.58 (0.53, 0.65)
0.44 (0.39, 0.48)
2.48 (2.33, 2.62)
205
C
0.14 (0.12, 0.17)
NA
3.56 (1.73, 5.38)
0.05 (0.05, 0.06)
0.02 (0.02, 0.03)
0.11 (0.10, 0.14)
0.23 (0.17, 0.33)
0.03 (0.03, 0.03)
Variance
0.50 (0.46, 0.53)
0.40 (0.36, 0.45)
0.40 (0.36, 0.45)
0.40 (0.38, 0.42)
0.09 (0.08, 0.11)
0.53 (0.50, 0.55)
0.52 (0.50, 0.55)
0.29 (0.28, 0.30)
2
2
-1
(basal area increment (cm yr )) as a function of soil resources for species with more than 50
focal individuals. For most species, mean growth was positively related to soil resources, but
there were negative correlations of growth to resources for four species. Per-site sample sizes
presented in Table D.15.
Resource
soex
NH4
NO3
ride
P
irde
NO3
K
P
fapa
K
P
Ca
Mg
crwa
P
Ca
prde
NO3
waco
K
R
2
Slope
Intercept
0.916 -0.047 (-0.074, -0.021)
10.477 (6.983, 13.971)
0.873 -0.073 (-0.125, -0.022)
10.984 (6.245, 15.723)
0.980 0.937 (0.528, 1.345)
-1.055 (-1.995, -0.116)
0.876 -0.187 (-0.316, -0.058)
0.871 0.17 (0.05, 0.291)
0.839 3.014 (0.59, 5.438)
24.819 (12.728, 36.911)
-7.991 (-19.438, 3.457)
-0.808 (-8.317, 6.7)
0.997
0.981
0.963
0.953
-5.499 (-6.988, -4.011)
-9.846 (-15.793, -3.9)
-4.455 (-9.676, 0.767)
100.162 (34.866, 165.5)
0.109 (0.092, 0.125)
5.468 (3.15, 7.787)
0.11 (0.044, 0.175)
-2.428 (-4.075, -0.78)
0.960 0.329 (0.205, 0.454)
0.820 0.012 (0.002, 0.023)
-0.249 (-0.644, 0.146)
-0.645 (-1.849, 0.56)
0.997 -0.019 (-0.021, -0.017)
2.005 (1.838, 2.172)
0.949 0.047 (0.014, 0.08)
-1.158 (-4.265, 1.949)
206
Table D.8. Model parameters and 95% confidence intervals (parenthetically) for individual species for which more than 50 focal
individuals were available.
Funct.
Species Group
δ
σ
P
AS
4.50 (4.10, 5.00)
0.76 (0.70, 0.83)
0.80 (0.65, 1.01)
FC
28.7 (25.5, 33.5)
1.00 (0.85, 1.14)
0.88 (0.81, 0.95)
0.24 (0.18, 0.31)
0.47 (0.44, 0.51)
AS
6.70 (5.90, 7.50)
0.46 (0.37, 0.57)
1.83 (1.51, 2.15)
0.03 (0.03, 0.04)
0.4 (0.36, 0.44)
NLD<C
24.7 (20.0, 30.3)
0.99 (0.83, 1.21)
0.95 (0.75, 1.15)
0.02 (0.02, 0.03)
0.1 (0.08, 0.11)
AS
10.8 (10.3, 11.3)
0.35 (0.31, 0.39)
2.53 (2.28, 2.77)
0.12 (0.10, 0.13)
0.47 (0.44, 0.51)
NLD<C
4.70 (3.90, 5.80)
0.73 (0.56, 1.00)
0.80 (0.63, 0.96)
0.02 (0.02, 0.03)
0.09 (0.08, 0.11)
NLD<C
7.80 (4.8, 15.3)
0.86 (0.43, 1.28)
0.18 (0.13, 0.24)
AS
7.00 (6.60, 7.30)
0.14 (0.09, 0.19)
2.19 (1.57, 2.75)
A<S
5.80 (5.70, 5.90)
0.05 (0.04, 0.07)
1.63 (1.27, 2.01)
0.09 (0.08, 0.1)
NLD<C
13.0 (9.5, 19.7)
0.67 (0.37, 1.42)
0.42 (0.28, 0.55)
0.39 (0.32, 0.47)
NLD<C
77.7 (58.9, 107)
2.12 (1.82, 2.49)
1.56 (1.38, 1.72)
0.52 (0.45, 0.59)
0.3 (0.27, 0.33)
AS
1.50 (0.90, 2.50)
1.42 (1.05, 2.29)
0.26 (0.18, 0.35)
0.19 (0.11, 0.28)
0.11 (0.1, 0.13)
NLD<C
8.30 (7.40, 9.00)
0.28 (0.22, 0.45)
0.69 (0.58, 0.92)
NLD<C
3.30 (1.50, 5.10)
2.06 (1.22, 2.91)
2.16 (1.67, 2.64)
NLDC
23.8 (16.6, 35.6)
0.91 (0.58, 1.72)
0.97 (0.74, 1.20)
other
60.0 (55.6, 66.3)
1.40 (1.33, 1.47)
0.72 (0.67, 0.75)
0.17 (0.15, 0.20)
Funct.
Species Group
were
AS
crwa
A<S
coho
NLD<C
B
0.18 (0.14, 0.21)
0.05 (0.05, 0.06)
0.07 (0.05, 0.09)
α
β
Radius (m)
5.77 (5.51, 5.84)
1.08 (1.04, 1.12)
1.7 (1.64, 1.75)
1
were
2
pema
soex
2
2
ride
2
irde
capi
2
3
fapa
2
eupr
4
crwa
3
dear
2
caar
2
prde
coho
1
2
waco
3
gome
2
C
0.08 (0.07, 0.09)
0.22 (0.19, 0.26)
0.20 (0.13, 0.26)
0.23 (0.2, 0.29)
0.18 (0.16, 0.23)
0.01 (0.01, 0.02)
0.26 (0.23, 0.3)
0.67 (0.56, 0.83)
6.72 (6.09, 6.85)
207
Variance
0.39 (0.38, 0.4)
2
2
-1
categories, excluding the Fabaceae.
1
Resource
R
2
Slope
Intercept
-3
0.2-0.4 g cm
P
0.898 2.375 (0.900, 3.850) 1.948 (-2.56, 6.453)
K
0.792 0.130 (0.008, 0.253) -3.64 (-15.4, 8.131)
-3
0.4-0.55 g cm
Mg
0.806 0.155 (0.015, 0.294) -1.21 (-8.12, 5.695)
-3
0.55-0.7 g cm —no significant parameter estimates
-3
0.7-1.2 g cm
Ca
0.778 0.247 (0.004, 0.489)
Unknown density
P
0.790 0.907 (0.048, 1.767)
1
Per-site sample sizes presented in Table D.15.
208
-8.111 (-34.24, 18.0)
4.243 (1.611, 6.874)
2
2
-1
categories, including the Fabaceae.
1
Resource
R
2
Slope
Intercept
-3
0.2-0.4 g cm
P
0.898 2.375 (0.900, 3.850) 1.948 (-2.56, 6.453)
K
0.792 0.130 (0.008, 0.253) -3.64 (-15.4, 8.131)
-3
0.4-0.55 g cm
Mg
0.806 0.155 (0.015, 0.294) -1.21 (-8.12, 5.695)
-3
0.55-0.7 g cm —no significant parameter estimates
-3
0.7-1.2 g cm
K
Ca
Unknown density
P
Ca
0.779 0.371 (0.009, 0.734)
0.778 0.247 (0.004, 0.489)
-18.16 (-53.56, 17.2)
-8.111 (-34.24, 18.0)
0.790 0.907 (0.048, 1.767)
0.783 0.038 (0.001, 0.074)
4.243 (1.611, 6.874)
2.929 (-0.952, 6.81)
1
Per-site sample sizes presented in Table D.15.
209
2
2
-1
Table D.9.3. Parameter estimates and R values for linear regressions of stand mean growth (basal area increment (cm year ) as a
function of soil resources for wood density classes within functional groups. Growth was not related to soil resources for any density
classes in the Fabaceae.
Functional
Group
Wood Density
-3
(g cm )
FC
F<C
AS
no significant parameter estimates
no significant parameter estimates
0.2-0.4
P
0.899
Unknown
Mg
0.915
0.4-0.55
P
0.954
0.4-0.55
Ca
0.826
0.2-0.4
NO3
0.793
Unknown
NH4
0.803
0.4-0.55
P
0.962
0.55-0.7
P
0.927
0.55-0.7
Ca
0.906
Unknown
P
0.920
A<S
NLDC
NLD<C
Resource
R
2
Slope
Intercept
2.615 (1.003, 4.228)
0.048 (0.004, 0.093)
0.331 (0.197, 0.464)
0.012 (0.002, 0.023)
-0.646 (-1.252, -0.04)
0.166 (0.015, 0.317)
1.226 (0.778, 1.673)
1.135 (0.552, 1.718)
0.049 (0.02, 0.078)
0.958 (0.438, 1.478)
-1.679 (-6.512, 3.153)
-1.958 (-4.371, 0.454)
-0.257 (-0.684, 0.169)
-0.65 (-1.834, 0.534)
81.43 (26.6, 136.305)
-0.471 (-21.047, 20.11)
3.03 (1.652, 4.408)
1.965 (0.159, 3.771)
0.049 (-3.077, 3.176)
1.251 (-0.337, 2.84)
210
Table D.10. Model parameters and 95% confidence intervals (parenthetically) for wood density groups.
Wood density
-3
(g cm )
0.2-0.4
0.4-0.55
0.55-0.7
0.7-0.85
unknown
δ
σ
P
C
Variance
39.8 (32.7, 52.6)
49.9 (41.8, 62.6)
15.0 (14.0, 16.0)
15.0 (14.0, 15.9)
9.90 (4.30, 15.6)
81.8 (74.9, 90.8)
1.86 (1.60, 2.16)
1.65 (1.48, 1.90)
2.97 (2.41, 3.53)
2.98 (2.44, 3.52)
2.96 (1.90, 4.03)
1.28 (1.21, 1.34)
0.99 (0.90, 1.07)
0.43 (0.39, 0.48)
0.56 (0.52, 0.60)
0.43 (0.41, 0.47)
0.25 (0.19, 0.32)
0.95 (0.86, 1.02)
0.22 (0.20, 0.26)
0.64 (0.52, 0.85)
0.13 (0.12, 0.16)
0.28 (0.22, 0.41)
0.48 (0.46, 0.51)
0.34 (0.32, 0.35)
0.45 (0.43, 0.47)
0.45 (0.43, 0.47)
0.33 (0.29, 0.38)
0.33 (0.31, 0.34)
211
0.15 (0.13, 0.18)
Table D.11. Model parameter mean estimates based on 101 sub-samples from the resource posterior predictive distribution. Minimum
and maximum parameter estimates are presented parenthetically.
δ
Category
Functional group
Fabaceae, canopy
Parameter estimates: mean (minimum, maximum)
σ
P
29.37 (28.01, 30.48)
Fabaceae, < canopy
36.05 (20.98, 69.86)
Arecaceae, subcanopy
9.32 (5.07, 16.5)
Arecaceae, < subcanopy
15.27 (1.21, 47.02)
Other, canopy
19.96 (19.71, 20)
Other, < canopy
184.08 (122.63, 199.24)
R
1.04 (0.98, 1.07) 0.76 (0.73, 0.89) 3.1 (0.21, 16.27)
0.91 (0.6, 1.23)
0.53 (0.35, 3.52)
2.32 (0.06, 4.96)
1.74 (1.47, 3.06)
2.09 (1.91, 2.14)
1.06 (0.73, 2.92)
0.58 (0.28, 1.03)
0.13 (0.06, 1.68)
0.43 (0.39, 0.56)
0.64 (0.53, 1)
0.46 (0.02, 14.69)
9.48 (1.52, 17.32)
5.59 (0.15, 13.32)
2.68 (0.14, 13.37)
1.45 (0.09, 7.68)
Resource
K
NO3
P
P
K
K
-3
Wood density (g cm )
0.2-0.4
0.4-0.55
0.55-0.7
Unknown
Species
Pema
soex
ride
irde
capi
eupr
caar
prde
waco
74.51 (51.82, 79.95) 2.32 (2.11, 2.91) 0.5 (0.44, 0.69) 10.74 (0.42, 22.92)
158.95 (97.42, 199.47) 2.2 (1.94, 2.38) 0.44 (0.38, 0.49) 13.27 (5.96, 23.14)
39.12 (37.6, 41.17)
1.13 (1.09, 1.18) 0.63 (0.62, 0.69)
3 (0.3, 11.95)
112.42 (96.41, 154.43) 1.42 (1.35, 1.57) 0.77 (0.71, 0.86) 5.21 (0.63, 20.19)
27.94 (27.2, 28.55)
6.93 (4.9, 7.36)
36.87 (16.12, 44.86)
10.48 (10.28, 10.84)
4.73 (1.01, 7.03)
0.98 (0.95, 1.02)
0.42 (0.37, 0.69)
1.06 (0.69, 1.36)
0.34 (0.33, 0.37)
0.98 (0.54, 2.61)
0.77 (0.75, 0.78)
0.61 (0.49, 1.75)
0.38 (0.17, 1.48)
0.83 (0.71, 1.17)
0.12 (0.1, 0.16)
10.95 (4.24, 23.15)
3.74 (0.04, 24.73)
1.46 (0.01, 15.58)
8.51 (0.26, 27.73)
5.13 (1.44, 10.82)
P
P
K
Ca
Ca
K
K
P
Ca
0.18 (0.06, 0.43) 1.36 (0.31, 5.75) 4.76 (0.05, 31.47)
NH4
106.26 (37.75, 119.66) 2.22 (1.06, 4.41) 0.95 (0.58, 2.92) 1.38 (0.18, 5.07)
7.41 (1.26, 14.79)
1.14 (0.13, 2.5) 0.19 (0.06, 2.31) 4.15 (0.01, 14.56)
4.96 (0.68, 13.78)
1.97 (0.78, 4.29) 0.55 (0.22, 2.21) 1.72 (0.01, 16.56)
NO3
P
K
6.88 (5.7, 7.29)
212
Table D.12. Predicted growth across the range of resource estimates for functional groups, wood
density categories, and species for which growth was related to a soil resource.
-1
Category
Functional group
Fabaceae, canopy
Fabaceae, < canopy
Arecaceae, subcanopy
Arecaceae, < subcanopy
Other, canopy
Other, < canopy
-3
Wood density (g cm )
0.2-0.4
0.4-0.55
0.55-0.7
Unknown
Species
Pema
soex
ride
irde
capi
eupr
caar
prde
waco
Diameter growth (cm yr )
Mean
Min.
Max.
0.759
0.613
0.461
0.097
0.425
0.265
0.723
0.418
0.273
0.058
0.388
0.236
0.889
1.574
0.813
0.461
0.561
0.418
0.378
0.243
0.535
0.229
0.356
0.232
0.520
0.223
0.423
0.252
0.573
0.235
0.758
0.333
0.132
0.624
0.094
0.196
0.629
0.060
0.439
0.746
0.269
0.083
0.495
0.088
0.000
0.391
0.024
0.214
0.769
0.886
0.450
0.718
0.102
1.352
2.174
0.073
1.428
213
-1
Arecaceae, >
-1
-1
P (mg kg )
Arecaceae, <
-1
P (mg kg )
Figure D.12.2.i. Predicted individual diameter growth based on 101 estimated soil resource
datasets plotted against total soil P for the Arecaceae. Total soil P explained 7.9% total
growth variance for subcanopy (>) species, and 6.0% total growth variance for treelet and
understory species (<).
214
-1
0.2 – 0.4 g cm
-3
-1
P (mg kg )
Figure D.12.2.ii. Predicted diameter growth based on 101 estimated soil resource datasets
-3
plotted as a function of total soil P for species with lowest wood density (0.2 – 0.4 g cm ).
215
-1
irde
-1
-1
P (mg kg )
capi
-1
Ca (mg kg )
Figure D.12.2.iii. Predicted individual diameter growth based on 101 estimated soil
resource datasets plotted as a function of the most strongly related soil resource for single
species in which resources explained more than 4% of total growth variance.
216
-1
Figure D.12.12.iii (cont‘d)
eupr
-1
-1
NH4 (mg kg )
waco
-1
K (mg kg )
217
Table D.13.1. Correlations (Pearson‘s r) between site-mean inorganic N and other resources for
all species and for the three single species—S. exorrhiza, I. deltoidea, and P. decurrens—for
which growth was negatively related to soil N.
Species
N type
Ca
K
Mg
P
All species
NH4
-0.430
-0.625
-0.151
-0.744
All species
NO3
-0.619
-0.936
0.115
-0.762
soex
NH4
-0.487
-0.576
-0.141
-0.854
soex
NO3
-0.628
-0.899
0.037
-0.880
irde
NO3
-0.710
-0.934
-0.193
-0.723
prde
NO3
-0.438
-0.807
-0.115
-0.589
Table D.13.2. Sample sizes by site for each functional group.
Site
1
2
3
4
5
Fabaceae
Canopy Smaller
87
39
73
43
89
37
116
10
81
23
Arecaceae
Subcanopy Smaller
172
16
171
24
219
27
146
1
181
51
218
Other
Canopy Smaller
160
390
143
394
195
433
123
415
106
207
Unknown
3
5
5
0
3
Table D.14. Sample sizes by site for each wood density group.
-3
Site
1
2
3
4
5
0.2-0.4
187
169
212
232
158
Density (g cm )
0.4-0.55 0.55-0.7 0.7-0.85
233
174
26
224
202
23
206
228
44
110
342
5
191
165
12
219
Unknown
247
235
315
122
126
Table D.15. Sample sizes by site for species with more than 50 focal individuals.
Site
1
2
3
4
5
were
29
53
38
1
47
pema
64
57
76
106
65
soex
28
45
54
29
39
ride
18
10
28
0
35
irde
62
46
99
113
40
capi
15
13
28
0
16
fapa
8
18
55
0
2
eupr
16
17
20
1
1
1
2
3
4
5
crwa
15
21
24
1
51
dear
11
9
9
13
15
caar
4
14
10
176
3
prde
36
9
7
2
52
coho
28
24
6
5
1
waco
2
37
24
33
0
gome
2
9
1
30
9
other
529
471
526
301
276
220
BIBLIOGRAPHY
221
BIBLIOGRAPHY
Adams, M.A., Simon, J. & Pfautsch, S. (2010) Woody legumes: a (re)view from the South. Tree
Physiology, 30, 1072-1082.
Adams, T.P., Purves, D.W. & Pacala, S.W. (2007) Understanding height-structured competition
in forests: is there an R* for light? Proceedings of the Royal Society B-Biological
Sciences, 274, 3039-3047.
Aerts, R. & Chapin, F.S. (2000) The mineral nutrition of wild plants revisited: A re-evaluation of
processes and patterns. Advances in Ecological Research, Vol 30, pp. 1-67. Academic
Press Inc, San Diego.
Aikio, S., Ramo, K. & Manninen, S. (2009) Dynamics of biomass partitioning in two competing
meadow plant species. Plant Ecology, 205, 129-137.
Alves, L.F., Vieira, S.A., Scaranello, M.A., Camargo, P.B., Santos, F.A.M., Joly, C.A. &
Martinelli, L.A. (2010) Forest structure and live aboveground biomass variation along an
elevational gradient of tropical Atlantic moist forest (Brazil). Forest Ecology and
Management, 260, 679-691.
Amtmann, A. & Blatt, M.R. (2009) Regulation of macronutrient transport. New Phytologist, 181,
35-52.
Andersen, K.M., Turner, B.L. & Dalling, J.W. (2010) Soil-based habitat partitioning in
understorey palms in lower montane tropical forests. Journal of Biogeography, 37, 278292.
Andreassen, K. & Tomter, S.M. (2003) Basal area growth models for individual trees of Norway
spruce, Scots pine, birch and other broadleaves in Norway. Forest Ecology and
Ares, A. & Fownes, J.H. (2001) Productivity, resource use, and competitive interactions of
Fraxinus uhdei in Hawaii uplands. Canadian Journal of Forest Research-Revue
Canadienne De Recherche Forestiere, 31, 132-142.
Asner, G.P., Townsend, A.R., Riley, W.J., Matson, P.A., Neff, J.C. & Cleveland, C.C. (2001)
Physical and biogeochemical controls over terrestrial ecosystem responses to nitrogen
deposition. Biogeochemistry, 54, 1-39.
Austin, A.T. & Sala, O.E. (2002) Carbon and nitrogen dynamics across a natural precipitation
gradient in Patagonia, Argentina. Journal of Vegetation Science, 13, 351-360.
Austin, P.C. & Hux, J.E. (2002) A brief note on overlapping confidence intervals. Journal of
Vascular Surgery, 36, 194-195.
222
Bailey, S.W., Horsley, S.B., Long, R.P. & Hallett, R.A. (2004) Influence of edaphic factors on
sugar maple nutrition and health on the Allegheny Plateau. Soil Science Society of
America Journal, 68, 243-252.
Baker, T.R., Swaine, M.D. & Burslem, D. (2003) Variation in tropical forest growth rates:
combined effects of functional group composition and resource availability. Perspectives
in Plant Ecology Evolution and Systematics, 6, 21-36.
Baltzer, J.L., Thomas, S.C., Nilus, R. & Burslem, D. (2005) Edaphic specialization in tropical
trees: Physiological correlates and responses to reciprocal transplantation. Ecology, 86,
3063-3077.
Banerjee, S., Carlin, B.P. & Gelfand, A.E. (2004) Hierarchical Modeling and Analysis for
Spatial Data. Chapman and Hall / CRC, Boca Raton.
Baribault, T.W. & Kobe, R.K. Neighbour interactions strengthen with increased soil resources in
a northern hardwood forest.
Baribault, T.W., Kobe, R.K. & Rothstein, D.E. (2010) Soil calcium, nitrogen, and water are
correlated with aboveground net primary production in northern hardwood forests. Forest
Ecology and Management, 260, 723-733.
Bauer, G.A., Bazzaz, F.A., Minocha, R., Long, S., Magill, A., Aber, J. & Berntson, G.M. (2004)
Effects of chronic N additions on tissue chemistry, photosynthetic capacity, and carbon
sequestration potential of a red pine (Pinus resinosa Ait.) stand in the NE United States.
Forest Ecology and Management, 196, 173-186.
Bauhus, J., Puettmann, K. & Messier, C. (2009) Silviculture for old-growth attributes. Forest
Bechtold, W.A. & Patterson, P.L. (2005) The enhanced forest inventory and analysis program-national sampling design and estimation procedures. Gen. Tech. Rep. SRS 80, USDA
Forest Service, Asheville, N.C.
Bedison, J.E. & McNeil, B.E. (2009) Is the growth of temperate forest trees enhanced along an
ambient nitrogen deposition gradient? Ecology, 90, 1736-1742.
Berger, T.W. & Glatzel, G. (2001) Response of Quercus petraea seedlings to nitrogen
fertilization. Forest Ecology and Management, 149, 1-14.
Berger, T.W., Swoboda, S., Prohaska, T. & Glatzel, G. (2006) The role of calcium uptake from
deep soils for spruce (Picea abies) and beech (Fagus sylvatica). Forest Ecology and
Berger, U. & Hildenbrandt, H. (2000) A new approach to spatially explicit modelling of forest
dynamics: Spacing, ageing and neighbourhood competition of mangrove trees.
Ecological Modelling, 132, 287-302.
223
Bigelow, S.W. & Canham, C.D. (2002) Community organization of tree species along soil
gradients in a north-eastern USA forest. Journal of Ecology, 90, 188-200.
Bigelow, S.W. & Canham, C.D. (2007) Nutrient limitation of juvenile trees in a northern
hardwood forest: Calcium and nitrate are preeminent. Forest Ecology and Management,
243, 310-319.
Binkley, D., Senock, R., Bird, S. & Cole, T.G. (2003) Twenty years of stand development in
pure and mixed stands of Eucalyptus saligna and nitrogen-fixing Facaltaria moluccana.
Binkley, D., Stape, J.L., Bauerle, W.L. & Ryan, M.G. (2010) Explaining growth of individual
trees: Light interception and efficiency of light use by Eucalyptus at four sites in Brazil.
Blevins, D.P., Prescott, C.E., Allen, H.L. & Newsome, T.A. (2005) The effects of nutrition and
density on growth, foliage biomass, and growth efficiency of high-density fire-origin
lodgepole pine in central British Columbia. Canadian Journal of Forest Research-Revue
Canadienne De Recherche Forestiere, 35, 2851-2859.
Blevins, L.L., Prescott, C.E. & Van Niejenhuis, A. (2006) The roles of nitrogen and phosphorus
in increasing productivity of western hemlock and western redcedar plantations on
northern Vancouver Island. Forest Ecology and Management, 234, 116-122.
Blum, J.D., Klaue, A., Nezat, C.A., Driscoll, C.T., Johnson, C.E., Siccama, T.G., Eagar, C.,
Fahey, T.J. & Likens, G.E. (2002) Mycorrhizal weathering of apatite as an important
calcium source in base-poor forest ecosystems. Nature, 417, 729-731.
Borders, B.E., Will, R.E., Markewitz, D., Clark, A., Hendrick, R., Teskey, R.O. & Zhang, Y.
(2004) Effect of complete competition control and annual fertilization on stem growth
and canopy relations for a chronosequence of loblolly pine plantations in the lower
coastal plain of Georgia. Forest Ecology and Management, 192, 21-37.
Boyden, S., Binkley, D. & Senock, R. (2005) Competition and facilitation between Eucalyptus
and nitrogen-fixing Falcataria in relation to soil fertility. Ecology, 86, 992-1001.
Boyden, S., Binkley, D. & Stape, J.L. (2008) Competition among Eucalyptus trees depends on
genetic variation and resource supply. Ecology, 89, 2850-2859.
Bradbury, I.K. & Malcolm, D.C. (1977) Effect of phosphorus and potassium on transpiration,
leaf diffusive resistance and water-use efficiency in sitka spruce (Picea-Sitchensis)
seedlings. Journal of Applied Ecology, 14, 631-641.
Brooker, R.W. & Callaway, R.M. (2009) Facilitation in the conceptual melting pot. Journal of
Ecology, 97, 1117-1120.
Brooker, R.W. & Kikividze, Z. (2008) Importance: an overlooked concept in plant interaction
research. Journal of Ecology, 96, 703-708.
224
Brookshire, E.N.J., Valett, H.M., Thomas, S.A. & Webster, J.R. (2007) Atmospheric N
deposition increases organic N loss from temperate forests. Ecosystems, 10, 252-262.
Brown, P.H., Bellaloui, N., Hu, H.N. & Dandekar, A. (1999) Transgenically enhanced sorbitol
synthesis facilitates phloem boron transport and increases tolerance of tobacco to boron
deficiency. Plant Physiology, 119, 17-20.
Buckley, T.N. & Roberts, D.W. (2006) DESPOT, a process-based tree growth model that
allocates carbon to maximize carbon gain. Tree Physiology, 26, 129-144.
Burnham, K.P. & Anderson, D.R. (2002) Model Selection and Multimodel Inference: A Practical
Information-Theoretic Approach. Springer-Verlag, New York.
Burselm, D.F.R.P., Grubb, P.J. & Turner, I.M. (1995) Response to nutrient addition among
shade-tolerant tree seedlings of lowland tropical rain forest in Singapore. Journal of
Ecology, 83, 113-122.
Burton, J.I., Zenner, E.K., Frelich, L.E. & Cornett, M.W. (2009) Patterns of plant community
structure within and among primary and second-growth northern hardwood forest stands.
Cahn, M.D., Bouldin, D.R., Cravo, M.S. & Bowen, W.T. (1993) Cation and nitrate leaching in
an oxisol of the Brazilian Amazon. Agronomy Journal, 85, 334-340.
Canham, C.D., Berkowitz, A.R., Kelly, V.R., Lovett, G.M., Ollinger, S.V. & Schnurr, J. (1996)
Biomass allocation and multiple resource limitation in tree seedlings. Canadian Journal
of Forest Research-Revue Canadienne De Recherche Forestiere, 26, 1521-1530.
Canham, C.D., Coates, K.D., Bartemucci, P. & Quaglia, S. (1999) Measurement and modeling of
spatially explicit variation in light transmission through interior cedar-hemlock forests of
British Columbia. Canadian Journal of Forest Research-Revue Canadienne De
Recherche Forestiere, 29, 1775-1783.
Canham, C.D., Finzi, A.C., Pacala, S.W. & Burbank, D.H. (1994) Causes and consequences of
resource heterogeneity in forests - interspecific variation in light transmission by canopy
trees. Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere,
24, 337-349.
Canham, C.D., LePage, P.T. & Coates, K.D. (2004) A neighborhood analysis of canopy tree
competition: effects of shading versus crowding. Canadian Journal of Forest ResearchRevue Canadienne De Recherche Forestiere, 34, 778-787.
Canham, C.D., Papaik, M.J., Uriarte, M., McWilliams, W.H., Jenkins, J.C. & Twery, M.J. (2006)
Neighborhood analyses of canopy tree competition along environmental gradients in new
England forests. Ecological Applications, 16, 540-554.
Canham, C.D. & Uriarte, M. (2006) Analysis of neighborhood dynamics of forest ecosystems
using likelihood methods and modeling. Ecological Applications, 16, 62-73.
225
Caplat, P., Anand, M. & Bauch, C. (2008) Symmetric competition causes population oscillations
in an individual-based model of forest dynamics. Ecological Modelling, 211, 491-500.
Carlson, C.A., Burkhart, H.E., Allen, H.L. & Fox, T.R. (2008) Absolute and relative changes in
tree growth rates and changes to the stand diameter distribution of Pinus taeda as a result
of midrotation fertilizer applications. Canadian Journal of Forest Research, 38, 20632071.
Carter, M.R. (1993) Soil Sampling and Methods of Analysis. CRC Press, Boca Raton, FL.
Casper, B.B., Cahill, J.F. & Hyatt, L.A. (1998) Above-ground competition does not alter
biomass allocated to roots in Abutilon theophrasti. New Phytologist, 140, 231-238.
Chatterjee, S. & Hadi, A.S. (2006) Regression Analysis by Example. John Wiley & Sons, Inc.,
Hoboken, New Jersey.
Chave, J., Andalo, C., Brown, S., Cairns, M.A., Chambers, J.Q., Eamus, D., Folster, H.,
Fromard, F., Higuchi, N., Kira, T., Lescure, J.P., Nelson, B.W., Ogawa, H., Puig, H.,
Riera, B. & Yamakura, T. (2005) Tree allometry and improved estimation of carbon
stocks and balance in tropical forests. Oecologia, 145, 87-99.
Chen, F.S., Zeng, D.H., Fahey, T.J., Yao, C.Y. & Yu, Z.Y. (2010) Response of leaf anatomy of
Chenopodium acuminatum to soil resource availability in a semi-arid grassland. Plant
Ecology, 209, 375-382.
Chen, J.L. & Reynolds, J.F. (1997) A coordination model of whole-plant carbon allocation in
relation to water stress. Annals of Botany, 80, 45-55.
Chesson, P. (2003) Quantifying and testing coexistence mechanisms arising from recruitment
fluctuations. Theoretical Population Biology, 64, 345-357.
Clark, D.A., Brown, S., Kicklighter, D.W., Chambers, J.Q., Thomlinson, J.R., Ni, J. & Holland,
E.A. (2001) Net primary production in tropical forests: An evaluation and synthesis of
existing field data. Ecol. Appl., 11, 371-384.
Clark, D.B., Clark, D.A. & Oberbauer, S.F. (2010) Annual wood production in a tropical rain
forest in NE Costa Rica linked to climatic variation but not to increasing CO2. Global
Change Biology, 16, 747-759.
Clark, D.B., Palmer, M.W. & Clark, D.A. (1999) Edaphic factors and the landscape-scale
distributions of tropical rain forest trees. Ecology, 80, 2662-2675.
Clark, J.S., Wolosin, M., Dietze, M., Ibanez, I., LaDeau, S., Welsh, M. & Kloeppel, B. (2007)
Tree growth inference and prediction from diameter censuses and ring widths. Ecological
Applications, 17, 1942-1953.
226
Coates, K.D., Canham, C.D. & LePage, P.T. (2009) Above- versus below-ground competitive
effects and responses of a guild of temperate tree species. Journal of Ecology, 97, 118130.
Cole, T.G. & Ewel, J.J. (2006) Allometric equations for four valuable tropical tree species.
Coll, L., Potvin, C., Messier, C. & Delagrange, S. (2008) Root architecture and allocation
patterns of eight native tropical species with different successional status used in opengrown mixed plantations in Panama. Trees-Structure and Function, 22, 585-596.
Comas, L.H., Bouma, T.J. & Eissenstat, D.M. (2002) Linking root traits to potential growth rate
in six temperate tree species. Oecologia, 132, 34-43.
Comas, L.H. & Eissenstat, D.M. (2004) Linking fine root traits to maximum potential growth
rate among 11 mature temperate tree species. Functional Ecology, 18, 388-397.
Compton, J.E., Church, M.R., Larned, S.T. & Hogsett, W.E. (2003) Nitrogen export from
forested watersheds in the Oregon Coast Range: The role of N-2-fixing red alder.
Ecosystems, 6, 773-785.
Coomes, D.A. & Allen, R.B. (2007) Effects of size, competition and altitude on tree growth.
Journal of Ecology, 95, 1084-1097.
Coomes, D.A. & Grubb, P.J. (2000) Impacts of root competition in forests and woodlands: A
theoretical framework and review of experiments. Ecological Monographs, 70, 171-207.
Coomes, D.A., Kunstler, G., Canham, C.D. & Wright, E. (2009) A greater range of shadetolerance niches in nutrient-rich forests: an explanation for positive richness-productivity
relationships? Journal of Ecology, 97, 705-717.
Corre, M.D., Veldkamp, E., Arnold, J. & Wright, S.J. (2010) Impact of elevated N input on soil
N cycling and losses in old-growth lowland and montane forests in Panama. Ecology, 91,
1715-1729.
Cote, B., Fyles, J.W. & Djalilvand, H. (2002) Increasing N and P resorption efficiency and
proficiency in northern deciduous hardwoods with decreasing foliar N and P
concentrations. Annals of Forest Science, 59, 275-281.
Covelo, F. & Gallardo, A. (2002) Effect of pine harvesting on leaf nutrient dynamics in young
oak trees at NW Spain. Forest Ecology and Management, 167, 161-172.
Craine, J.M. (2005) Reconciling plant strategy theories of Grime and Tilman. Journal of
Ecology, 93, 1041-1052.
Crookston, N.L. & Finley, A.O. (2007) yaImpute: an R package for k-NN imputation. Journal of
Statistical Software, 23, 1-16.
227
Cusack, D.F., Silver, W. & McDowell, W.H. (2009) Biological Nitrogen Fixation in Two
Tropical Forests: Ecosystem-Level Patterns and Effects of Nitrogen Fertilization.
Ecosystems, 12, 1299-1315.
Das, D.K. & Chaturvedi, O.P. (2008) Root biomass and distribution of five agroforestry tree
species. Agroforestry Systems, 74, 223-230.
Dauer, J.M., Chorover, J., Chadwick, O.A., Oleksyn, J., Tjoelker, M.G., Hobbie, S.E., Reich,
P.B. & Eissenstat, D.M. (2007) Controls over leaf and litter calcium concentrations
among temperate trees. Biogeochemistry, 86, 175-187.
Davis, M.A., Wrage, K.J., Reich, P.B., Tjoelker, M.G., Schaeffer, T. & Muermann, C. (1999)
Survival, growth, and photosynthesis of tree seedlings competing with herbaceous
vegetation along a water-light-nitrogen gradient. Plant Ecology, 145, 341-350.
Davis, M.R., Allen, R.B. & Clinton, P.W. (2004) The influence of N addition on nutrient
content, leaf carbon isotope ratio, and productivity in a Nothofagus forest during stand
development. Canadian Journal of Forest Research-Revue Canadienne De Recherche
Forestiere, 34, 2037-2048.
de Freitas, J.V. & Pinard, M.A. (2008) Applying ecological knowledge to decisions about seed
tree retention in selective logging in tropical forests. Forest Ecology and Management,
256, 1434-1442.
Demidchik, V. & Maathuis, F.J.M. (2007) Physiological roles of nonselective cation channels in
plants: from salt stress to signalling and development. New Phytologist, 175, 387-404.
Dent, D.H. & Burslem, D.F.R.P. (2009) Performance trade-offs driven by morphological
plasticity contribute to habitat specialization of Bornean tree species. Biotropica, 41, 424434.
Desilets, P. & Houle, G. (2005) Effects of resource availability and heterogeneity on the slope of
the species-area curve along a floodplain-upland gradient. Journal of Vegetation Science,
16, 487-496.
Devine, W.D. & Harrington, C.A. (2009) Western redcedar response to precommercial thinning
and fertilization through 25 years posttreatment. Canadian Journal of Forest ResearchRevue Canadienne De Recherche Forestiere, 39, 619-628.
Dijkstra, F.A. (2003) Calcium mineralization in the forest floor and surface soil beneath different
tree species in the northeastern US. Forest Ecology and Management, 175, 185-194.
Doane, T.A. & Horwath, W.R. (2003) Spectrophotometric determination of nitrate with a single
reagent. Analytical Letters, 36, 2713-2722.
Dolle, M. & Schmidt, W. (2009) Impact of tree species on nutrient and light availability:
evidence from a permanent plot study of old-field succession. Plant Ecology, 203, 273287.
228
Donaldson, J.R., Kruger, E.L. & Lindroth, R.L. (2006) Competition- and resource-mediated
tradeoffs between growth and defensive chemistry in trembling aspen (Populus
tremuloides). New Phytologist, 169, 561-570.
Doust, S.J., Erskine, P.D. & Lamb, D. (2008) Restoring rainforest species by direct seeding: Tree
seedling establishment and growth performance on degraded land in the wet tropics of
Australia. Forest Ecology and Management, 256, 1178-1188.
Easdale, T.A., Gurvich, D.E., Sersic, A.N. & Healey, J.R. (2007) Tree morphology in seasonally
dry montane forest in Argentina: Relationships with shade tolerance and nutrient
shortage. Journal of Vegetation Science, 18, 313-326.
Easdale, T.A. & Healey, J.R. (2009) Resource-use-related traits correlate with population
turnover rates, but not stem diameter growth rates, in 29 subtropical montane tree species.
Perspectives in Plant Ecology Evolution and Systematics, 11, 203-218.
Elser, J.J., Fagan, W.F., Kerkhoff, A.J., Swenson, N.G. & Enquist, B.J. (2010) Biological
stoichiometry of plant production: metabolism, scaling and ecological response to global
change. New Phytologist, 186, 593-608.
Emmett, B.A. (1999) The impact of nitrogen on forest soils and feedbacks on tree growth. Water
Air and Soil Pollution, 116, 65-74.
Engel, E.C. & Weltzin, J.F. (2008) Can community composition be predicted from pairwise
species interactions? Plant Ecology, 195, 77-85.
Ericsson, T. & Kahr, M. (1993) Growth and nutrition of birch seedlings in relation to potassium
supply rate. Trees-Structure and Function, 7, 78-85.
Erskine, P.D., Lamb, D. & Borschmann, G. (2005) Growth performance and management of a
mixed rainforest tree plantation. New Forests, 29, 117-134.
Finley, A.O., Banerjee, S. & Carlin, B.P. (2007) spBayes: An R package for univariate and
multivariate hierarchical point-referenced spatial models. Journal of Statistical Software,
19.
Finzi, A.C. (2009) Decades of atmospheric deposition have not resulted in widespread
phosphorus limitation or saturation of tree demand for nitrogen in southern New England.
Biogeochemistry, 92, 217-229.
Finzi, A.C. & Canham, C.D. (2000) Sapling growth in response to light and nitrogen availability
in a southern New England forest. Forest Ecology and Management, 131, 153-165.
Finzi, A.C., Canham, C.D. & Van Breemen, N. (1998) Canopy tree soil interactions within
temperate forests: Species effects on pH and cations. Ecological Applications, 8, 447454.
229
Fluckiger, W. & Braun, S. (1998) Nitrogen deposition in Swiss forests and its possible relevance
for leaf nutrient status, parasite attacks and soil acidification. Environmental Pollution,
102, 69-76.
Forrester, D.I., BauhuS, J. & Cowie, A.L. (2006) Carbon allocation in a mixed-species plantation
of Eucalyptus globulus and Acacia mearnsii. Forest Ecology and Management, 233, 275284.
Fownes, J.H. & Harrington, R.A. (2004) Seedling response to gaps: separating effects of light
and nitrogen. Forest Ecology and Management, 203, 297-310.
Frank, K., Beegle, D. & Denning, J. (1998) Phosphorus. Recommended Chemical Soil Test
Procedures for the North Central Region (ed J. R. Brown). North Central Region
Research Publication No. 221, Missouri Agricultural Experiment Station, Missouri.
Freckleton, R.P., Watkinson, A.R. & Rees, M. (2009) Measuring the importance of competition
in plant communities. Journal of Ecology, 97, 379-384.
Freer-Smith, P.H. & Kennedy, F. (2003) Base cation removal in harvesting and biological limit
terms for use in the simple mass balance equation to calculate critical loads for forest
soils. Water Air and Soil Pollution, 145, 409-427.
Fujinuma, R., Bockheim, J. & Balster, N. (2005) Base-cation cycling by individual tree species
in old-growth forests of Upper Michigan, USA. Biogeochemistry, 74, 357-376.
Gaucherand, S., Liancourt, P. & Lavorel, S. (2006) Importance and intensity of competition
along a fertility gradient and across species. Journal of Vegetation Science, 17, 455-464.
Gee, G.W., and J.W. Bauder (1986) Particle-size analysis. Methods of soil analysis (ed A.
Klute), pp. 383–411. Agronomy Monographs, Madison, WI.
Gelfand, A.E. & Ghosh, S.K. (1998) Model choice: A minimum posterior predictive loss
approach. Biometrika, 85, 1-11.
Gelman, A. (2006) Prior distributions for variance parameters in hierarchical models(Comment
on an Article by Browne and Draper). Bayesian Analysis, 1, 515-533.
Giesler, R., Hogberg, M. & Hogberg, P. (1998) Soil chemistry and plants in Fennoscandian
boreal forest as exemplified by a local gradient. Ecology, 79, 119-137.
Gimaret-Carpentier, C., Pelissier, R., Pascal, J.P. & Houllier, F. (1998) Sampling strategies for
the assessment of tree species diversity. Journal of Vegetation Science, 9, 161-172.
Gleeson, S.K. & Tilman, D. (1994) Plant allocation, growth-rate and successional status.
Functional Ecology, 8, 543-550.
Goodale, C.L. & Aber, J.D. (2001) The long-term effects of land-use history on nitrogen cycling
in northern hardwood forests. Ecol. Appl., 11, 253-267.
230
Gough, C.M., Vogel, C.S., Hardiman, B. & Curtis, P.S. (2010) Wood net primary production
resilience in an unmanaged forest transitioning from early to middle succession. Forest
Gradowski, T. & Thomas, S.C. (2006) Phosphorus limitation of sugar maple growth in central
Ontario. Forest Ecology and Management, 226, 104-109.
Gradowski, T. & Thomas, S.C. (2008) Responses of Acer saccharum canopy trees and saplings
to P, K and lime additions under high N deposition. Tree Physiology, 28, 173-185.
Graham, E.A., Mulkey, S.S., Kitajima, K., Phillips, N.G. & Wright, S.J. (2003) Cloud cover
limits net CO2 uptake and growth of a rainforest tree during tropical rainy seasons.
Proceedings of the National Academy of Sciences of the United States of America, 100,
572-576.
Granier, A., Loustau, D. & Breda, N. (2000) A generic model of forest canopy conductance
dependent on climate, soil water availability and leaf area index. Annals of Forest
Science, 57, 755-765.
Gravel, D., Canham, C.D., Beaudet, M. & Messier, C. (2010) Shade tolerance, canopy gaps and
mechanisims of coexistence of forest trees. Oikos, 119, 475-484.
Gress, S.E., Nichols, T.D., Northcraft, C.C. & Peterjohn, W.T. (2007) Nutrient limitation in soils
exhibiting differing nitrogen availabilities: What lies beyond nitrogen saturation?
Ecology, 88, 119-130.
Grime, J.P. (1979) Plant Strategies and Vegetation Processes. John Wiley and Sons, Chichester.
Grime, J.P. (2001) Plant Strategies, Vegetation Processes, and Ecosystem Properties. John
Wiley and Sons, Ltd, Chichester.
Groffman, P.M., Fisk, M.C., Driscoll, C.T., Likens, G.E., Fahey, T.J., Eagar, C. & Pardo, L.H.
(2006) Calcium additions and microbial nitrogen cycle processes in a northern hardwood
forest. Ecosystems, 9, 1289-1305.
Gronewold, C.A., D'Amato, A.W. & Palik, B.J. (2010) The influence of cutting cycle and
stocking level on the structure and composition of managed old-growth northern
hardwoods. Forest Ecology and Management, 259, 1151-1160.
Guariguata, M.R. & Ostertag, R. (2001) Neotropical secondary forest succession: changes in
structural and functional characteristics. Forest Ecology and Management, 148, 185-206.
Hagerberg, D., Thelin, G. & Wallander, H. (2003) The production of ectomycorrhizal mycelium
in forests: Relation between forest nutrient status and local mineral sources. Plant and
Soil, 252, 279-290.
Hall, S.A., Burke, I.C. & Hobbs, N.T. (2006) Litter and dead wood dynamics in ponderosa pine
forests along a 160-year chronosequence. Ecol. Appl., 16, 2344-2355.
231
Hallett, R.A. & Hornbeck, J.W. (1997) Foliar and soil nutrient relationships in red oak and white
pine forests. Can. J. Forest Res., 27, 1233-1244.
Harrington, R.A., Fownes, J.H. & Vitousek, P.M. (2001) Production and resource use
efficiencies in N- and P-limited tropical forests: A comparison of responses to long-term
fertilization. Ecosystems, 4, 646-657.
He, F.L. & Duncan, R.P. (2000) Density-dependent effects on tree survival in an old-growth
Douglas fir forest. Journal of Ecology, 88, 676-688.
Hedin, L.O., Brookshire, E.N.J., Menge, D.N.L. & Barron, A.R. (2009) The Nitrogen Paradox in
Tropical Forest Ecosystems. Annual Review of Ecology Evolution and Systematics, 40,
613-635.
Hedin, L.O., Vitousek, P.M. & Matson, P.A. (2003) Nutrient losses over four million years of
tropical forest development. Ecology, 84, 2231-2255.
Henderson, D.E. & Jose, S.B. (2005) Production physiology of three fast-growing hardwood
species along a soil resource gradient. Tree Physiology, 25, 1487-1494.
Henry, H.A.L. & Aarssen, L.W. (2001) Inter- and intraspecific relationships between shade
tolerance and shade avoidance in temperate trees. Oikos, 93, 477-487.
Hertel, D., Moser, G., Culmsee, H., Erasmi, S., Horna, V., Schuldt, B. & Leuschner, C. (2009)
Below- and above-ground biomass and net primary production in a paleotropical natural
forest (Sulawesi, Indonesia) as compared to neotropical forests. Forest Ecology and
Management, 258, 1904-1912.
Hidaka, A. & Kitayama, K. (2009) Divergent patterns of photosynthetic phosphorus-use
efficiency versus nitrogen-use efficiency of tree leaves along nutrient-availability
gradients. Journal of Ecology, 97, 984-991.
Hilley, G.E. & Porder, S. (2008) A framework for predicting global silicate weathering and CO2
drawdown rates over geologic time-scales. Proceedings of the National Academy of
Sciences of the United States of America, 105, 16855-16859.
Hobbie, S.E. (1996) Temperature and plant species control over litter decomposition in Alaskan
tundra. Ecol. Monogr., 66, 503-522.
Hobbie, S.E., Ogdahl, M., Chorover, J., Chadwick, O.A., Oleksyn, J., Zytkowiak, R. & Reich,
P.B. (2007) Tree species effects on soil organic matter dynamics: The role of soil cation
composition. Ecosystems, 10, 999-1018.
Hodson, M.J. & Sangster, A.G. (1998) Mineral deposition in the needles of white spruce [Picea
glauca (Moench.) Voss]. Ann. Bot., 82, 375-385.
232
Hogberg, P., Fan, H.B., Quist, M., Binkley, D. & Tamm, C.O. (2006) Tree growth and soil
acidification in response to 30 years of experimental nitrogen loading on boreal forest.
Global Change Biology, 12, 489-499.
Hogh-Jensen, H. (2003) The effect of potassium deficiency on growth and N-2-fixation in
Trifolium repens. Physiologia Plantarum, 119, 440-449.
Holste, E.K. (2010) Spatial heterogeneity of soil resources in a wet tropical forest: influences of
substrate and plant-nutrient cycles. M.S., Michigan State University, East Lansing,
Michigan.
Holste, E.K., Kobe, R.K. & Finley, A.O. Fine-scale soil resource heterogeneity in a wet tropical
forest.
Holste, E.K., Kobe, R.K. & Finley, A.O. (2010) Spatial heterogeneity of soil resources in a wet
tropical forest: influences of substrate and plant-nutrient cycles. M.S., Michigan State
University, East Lansing, Michigan.
Holzmueller, E.J., Jose, S. & Jenkins, M.A. (2007) Relationship between Cornus florida L. and
calcium mineralization in two southern Appalachian forest types. Forest Ecology and
Hood, L.A., Swaine, M.D. & Mason, P.A. (2004) The influence of spatial patterns of dampingoff disease and arbuscular mycorrhizal colonization on tree seedling establishment in
Ghanaian tropical forest soil. Journal of Ecology, 92, 816-823.
Host, G.E. & Pregitzer, K.S. (1992) Geomorphic influences on ground-flora and overstory
composition in upland forests of northwestern lower Michigan. Can. J. Forest Res., 22,
1547-1555.
Host, G.E., Pregitzer, K.S., Ramm, C.W., Lusch, D.P. & Cleland, D.T. (1988) Variation in
overstory biomass among glacial landforms and ecological land units in northwestern
lower Michigan. Can. J. Forest Res., 18, 659-668.
Hubbell, S.P. (2001) The Unified Neutral Theory of Biodiversity and Biogeography. Princeton
University Press, Princeton, New Jersey.
Idol, T.W., Pope, P.E. & Ponder, F. (2003) N mineralization, nitrification, and N uptake across a
100-year chronosequence of upland hardwood forests. Forest Ecology and Management,
176, 509-518.
Iversen, C.M. & Norby, R.J. (2008) Nitrogen limitation in a sweetgum plantation: implications
for carbon allocation and storage. Can. J. Forest Res., 38, 1021-1032.
Izuta, T., Yamaoka, T., Nakaji, T., Yonekura, T., Yokoyama, M., Funada, R., Koike, T. &
Totsuka, T. (2004) Growth, net photosynthesis and leaf nutrient status of Fagus crenata
seedlings grown in brown forest soil acidified with H2SO4 or HNO3 solution. TreesStructure and Function, 18, 677-685.
233
Jacob, M., Leuschner, C. & Thomas, F.M. (2010) Productivity of temperate broad-leaved forest
stands differing in tree species diversity. Annals of Forest Science, 67.
Janse-ten Klooster, S.H., Thomas, E.J.P. & Sterck, F.J. (2007) Explaining interspecific
differences in sapling growth and shade tolerance in temperate forests. Journal of
Ecology, 95, 1250-1260.
Javot, H., Pumplin, N. & Harrison, M.J. (2007) Phosphate in the arbuscular mycorrhizal
symbiosis: transport properties and regulatory roles. Plant Cell and Environment, 30,
310-322.
Jenkins, J.C., Chojnacky, D. C., Heath, L. S., Birdsey, R. A. (2004) Comprehensive database of
diameter-based biomass regressions for North American tree species. pp. 1-48. United
States Department of Agriculture, Forest Service, Northeastern Research Station.
Johnson, D.W. (2006) Progressive N limitation in forests: Review and implications for long-term
responses to elevated CO2. Ecology, 87, 64-75.
Jones, T.A., Domke, G.M. & Thomas, S.C. (2009) Canopy tree growth responses following
selection harvest in seven species varying in shade tolerance. Canadian Journal of Forest
Research-Revue Canadienne De Recherche Forestiere, 39, 430-440.
Joshi, A.B., Vann, D.R., Johnson, A.H. & Miller, E.K. (2003) Nitrogen availability and forest
productivity along a climosequence on Whiteface Mountain, New York. Can. J. Forest
Res., 33, 1880-1891.
Juice, S.M., Fahey, T.J., Siccama, T.G., Driscoll, C.T., Denny, E.G., Eagar, C., Cleavitt, N.L.,
Minocha, R. & Richardson, A.D. (2006) Response of sugar maple to calcium addition to
northern hardwood forest. Ecology, 87, 1267-1280.
Kalliokoski, T., Nygren, P. & Sievanen, R. (2008) Coarse root architecture of three boreal tree
species growing in mixed stands. Silva Fennica, 42, 189-210.
Kellner, J.R. & Asner, G.P. (2009) Convergent structural responses of tropical forests to diverse
disturbance regimes. Ecology Letters, 12, 887-897.
Kelly, J., Jose, S., Nichols, J.D. & Bristow, M. (2009) Growth and physiological response of six
Australian rainforest tree species to a light gradient. Forest Ecology and Management,
257, 287-293.
King, D.A. (2005) Linking tree form, allocation and growth with an allometrically explicit
model. Ecological Modelling, 185, 77-91.
Knox, R.G., Harcombe, P.A. & Elsik, I.S. (1995) Contrasting patterns of resource limitation in
tree seedlings across a gradient in soil texture. Can. J. Forest Res., 25, 1583-1594.
Kobe, R.K. (2006) Sapling growth as a function of light and landscape-level variation in soil
water and foliar nitrogen in northern Michigan. Oecologia, 147, 119-133.
234
Kobe, R.K., Iyer, M. & Walters, M.B. (2010) Optimal partitioning theory revisited:
Nonstructural carbohydrates dominate root mass responses to nitrogen. Ecology, 91, 166179.
Kobe, R.K., Lepczyk, C.A. & Iyer, M. (2005) Resorption efficiency decreases with increasing
green leaf nutrients in a global data set. Ecology, 86, 2780-2792.
Kobe, R.K., Likens, G.E. & Eagar, C. (2002) Tree seedling growth and mortality responses to
manipulations of calcium and aluminum in a northern hardwood forest. Can. J. Forest
Res., 32, 954-966.
Kobe, R.K., Pacala, S.W., Silander, J.A. & Canham, C.D. (1995) Juvenile tree survivorship as a
component of shade tolerance. Ecological Applications, 5, 517-532.
Koerselman, W. & Meuleman, A.F.M. (1996) The vegetation N:P ratio: A new tool to detect the
nature of nutrient limitation. Journal of Applied Ecology, 33, 1441-1450.
Kranabetter, J.M. & Simard, S.W. (2008) Inverse relationship between understory light and
foliar nitrogen along productivity gradients of boreal forests. Canadian Journal of Forest
Kulmatiski, A., Vogt, K.A., Vogt, D.J., Wargo, P.M., Tilley, J.P., Siccama, T.G., Sigurdardottir,
R. & Ludwig, D. (2007) Nitrogen and calcium additions increase forest growth in
northeastern USA spruce-fir forests. Can. J. Forest Res., 37, 1574-1585.
Kunkle, J.M., Walters, M.B. & Kobe, R.K. (2009) Senescence-related changes in nitrogen in fine
roots: mass loss affects estimation. Tree Physiology, 29, 715-723.
Kunstler, G., Albert, C.H., Courbaud, B., Lavergne, S., Thuiller, W., Vieilledent, G.,
Zimmermann, N.E. & Coomes, D.A. (2011) Effects of competition on tree radial-growth
vary in importance but not in intensity along climatic gradients. Journal of Ecology, 99,
300-312.
Laubhann, D., Sterba, H., Reinds, G.J. & De Vries, W. (2009) The impact of atmospheric
deposition and climate on forest growth in European monitoring plots: An individual tree
growth model. Forest Ecology and Management, 258, 1751-1761.
LeBauer, D.S. & Treseder, K.K. (2008) Nitrogen limitation of net primary productivity in
terrestrial ecosystems is globally distributed. Ecology, 89, 371-379.
Leblanc, H.A., McGraw, R.L., Nygren, P. & Le Roux, C. (2005) Neotropical legume tree Inga
edulis forms N-2-fixing symbiosis with fast-growing Bradyrhizobium strains. Plant and
Soil, 275, 123-133.
Leduc, S.D. & Rothstein, D.E. (2010) Plant-available organic and mineral nitrogen shift in
dominance with forest stand age. Ecology, 91, 708-720.
235
Lefrancois, M.L., Beaudet, M. & Messier, C. (2008) Crown openness as influenced by tree and
site characteristics for yellow birch, sugar maple, and eastern hemlock. Canadian Journal
of Forest Research-Revue Canadienne De Recherche Forestiere, 38, 488-497.
Liancourt, P., Viard-Cretat, F. & Michalet, R. (2009) Contrasting community responses to
fertilization and the role of the competitive ability of dominant species. Journal of
Vegetation Science, 20, 138-147.
Lieberman, M. & Lieberman, D. (2007) Nearest-neighbor tree species combinations in tropical
forest: the role of chance, and some consequences of high diversity. Oikos, 116, 377-386.
Lintunen, A. & Kaitaniemi, P. (2010) Responses of crown architecture in Betula pendula to
competition are dependent on the species of neighbouring trees. Trees-Structure and
Function, 24, 411-424.
Loide, V., Moges, M. & Rebane, J. (2005) Assessment of agrochemical properties of the soil
using the extraction solution Mehlich 3 in Estonia. Agronomy Research, 73-80.
Long, R.P., Horsley, S.B. & Lilja, P.R. (1997) Impact of forest liming on growth and crown
vigor of sugar maple and associated hardwoods. Can. J. Forest Res., 27, 1560-1573.
Lortie, C.J. & Callaway, R.M. (2006) Re-analysis of meta-analysis: support for the stressgradient hypothesis. Journal of Ecology, 94, 7-16.
Lukac, M., Calfapietra, C., Lagomarsino, A. & Loreto, F. (2010) Global climate change and tree
nutrition: effects of elevated CO2 and temperature. Tree Physiology, 30, 1209-1220.
Lundholm, J.T. (2009) Plant species diversity and environmental heterogeneity: spatial scale and
competing hypotheses. Journal of Vegetation Science, 20, 377-391.
Lutz, J.A., van Wagtendonk, J.W. & Franklin, J.F. (2010) Climatic water deficit, tree species
ranges, and climate change in Yosemite National Park. Journal of Biogeography, 37,
936-950.
Luyssaert, S., Inglima, I., Jung, M., Richardson, A.D., Reichsteins, M., Papale, D., Piao, S.L.,
Schulzes, E.D., Wingate, L., Matteucci, G., Aragao, L., Aubinet, M., Beers, C.,
Bernhoffer, C., Black, K.G., Bonal, D., Bonnefond, J.M., Chambers, J., Ciais, P., Cook,
B., Davis, K.J., Dolman, A.J., Gielen, B., Goulden, M., Grace, J., Granier, A., Grelle, A.,
Griffis, T., Grunwald, T., Guidolotti, G., Hanson, P.J., Harding, R., Hollinger, D.Y.,
Hutyra, L.R., Kolar, P., Kruijt, B., Kutsch, W., Lagergren, F., Laurila, T., Law, B.E., Le
Maire, G., Lindroth, A., Loustau, D., Malhi, Y., Mateus, J., Migliavacca, M., Misson, L.,
Montagnani, L., Moncrieff, J., Moors, E., Munger, J.W., Nikinmaa, E., Ollinger, S.V.,
Pita, G., Rebmann, C., Roupsard, O., Saigusa, N., Sanz, M.J., Seufert, G., Sierra, C.,
Smith, M.L., Tang, J., Valentini, R., Vesala, T. & Janssens, I.A. (2007) CO2 balance of
boreal, temperate, and tropical forests derived from a global database. Global Change
Biology, 13, 2509-2537.
236
MacDougall, A.S. & Turkington, R. (2004) Relative importance of suppression-based and
tolerance-based competition in an invaded oak savanna. Journal of Ecology, 92, 422-434.
MacFarlane, D.W. & Kobe, R.K. (2006) Selecting models for capturing tree-size effects on
growth-resource relationships. Canadian Journal of Forest Research-Revue Canadienne
De Recherche Forestiere, 36, 1695-1704.
Maestre, F.T., Callaway, R.M., Valladares, F. & Lortie, C.J. (2009) Refining the stress-gradient
hypothesis for competition and facilitation in plant communities. Journal of Ecology, 97,
199-205.
Magill, A.H., Aber, J.D., Currie, W.S., Nadelhoffer, K.J., Martin, M.E., McDowell, W.H.,
Melillo, J.M. & Steudler, P. (2004) Ecosystem response to 15 years of chronic nitrogen
additions at the Harvard Forest LTER, Massachusetts, USA. Forest Ecology and
Mallik, A.U., Hossain, M.K. & Lamb, E.G. (2008) Species and spacing effects of northern
conifers on forest productivity and soil chemistry in a 50-year-old common garden
experiment. Journal of Forestry, 106, 83-90.
Masclaux-Daubresse, C., Daniel-Vedele, F., Dechorgnat, J., Chardon, F., Gaufichon, L. &
Suzuki, A. (2010) Nitrogen uptake, assimilation and remobilization in plants: challenges
for sustainable and productive agriculture. Annals of Botany, 105, 1141-1157.
Mayer, P.M. (2008) Ecosystem and decomposer effects on litter dynamics along an old field to
old-growth forest successnional gradient. Acta Oecologica-International Journal of
Ecology, 33, 222-230.
McCarthy-Neumann, S. & Kobe, R.K. (2010) Conspecific and heterospecific plant-soil
feedbacks influence survivorship and growth of temperate tree seedlings. J. Ecol., 98,
408-418.
McDade, L.A. & Hartshorn, G.S. (1994) La Selva Biological Station. La Selva: Ecology and
Natural History of Neotropical Rain Forest (eds L. A. McDade, K. S. Bawa, H. A.
Hespenheide & G. S. Hartshorn). University of Chicago Press, Chicago, Illinois.
McDowell, N.G., Allen, C.D. & Marshall, L. (2009) Growth, carbon-isotope discrimination, and
drought-associated mortality across a Pinus ponderosa elevational transect. Global
McFadden, J.P., Macdonald, N.W., Witter, J.A. & Zak, D.R. (1994) Fine-textured soil bands and
oak forest productivity in northwestern lower Michigan, USA. Can. J. Forest Res., 24,
928-933.
McGroddy, M.E., Daufresne, T. & Hedin, L.O. (2008) Scaling of C : N : P stoichiometry in
forests worldwide: implications of terrestrial Redfield-type ratios(vol 85, pg 2390, 2004).
Ecology, 89, 890-890.
237
McLaughlin, J. (2009) Boreal mixed-wood watershed riparian zone cation cycling during two
contrasting climatic years. Soil Science Society of America Journal, 73, 1408-1418.
McLaughlin, J.W. & Phillips, S.A. (2006) Soil carbon, nitrogen, and base cation cycling 17 years
after whole-tree harvesting in a low-elevation red spruce (Picea rubens)-balsam fir (Abies
balsainea) forested watershed in central Maine, USA. Forest Ecology and Management,
222, 234-253.
McLaughlin, S.B. & Wimmer, R. (1999) Tansley Review No. 104 - Calcium physiology and
terrestrial ecosystem processes. New Phytologist, 142, 373-417.
McMahon, S.M., Dietze, M.C., Hersh, M.H., Moran, E.V. & Clark, J.S. (2009) A predictive
framework to understand forest responses to global change. Year in Ecology and
Conservation Biology 2009, pp. 221-236. Blackwell Publishing, Oxford.
McPhee, C.S. & Aarssen, L.W. (2001) The separation of above- and below-ground competition
in plants - A review and critique of methodology. Plant Ecology, 152, 119-136.
Mead, D.J., Scott, J.T. & Chang, S.X. (2010) Using vector analysis to understand temporal
changes in understorey-tree competition in agroforestry systems. Forest Ecology and
Management, 259, 1200-1211.
Meerts, P. (2002) Mineral nutrient concentrations in sapwood and heartwood: a literature review.
Annals of Forest Science, 59, 713-722.
Mehlich, A. (1984) Mehlich-3 soil test extractant-a modification of Mehlich-2 extractant.
Communications in Soil Science and Plant Analysis, 15, 1409-1416.
Menge, D.N.L., DeNoyer, J.L. & Lichstein, J.W. (2010) Phylogenetic Constraints Do Not
Explain the Rarity of Nitrogen-Fixing Trees in Late-Successional Temperate Forests.
Plos One, 5.
Menge, D.N.L., Levin, S.A. & Hedin, L.O. (2008) Evolutionary tradeoffs can select against
nitrogen fixation and thereby maintain nitrogen limitation. Proceedings of the National
Academy of Sciences of the United States of America, 105, 1573-1578.
Metcalf, C.J.E., Clark, J.S. & Clark, D.A. (2009) Tree growth inference and prediction when the
point of measurement changes: modelling around buttresses in tropical forests. Journal of
Tropical Ecology, 25, 1-12.
Migita, C., Chiba, Y. & Tange, T. (2007) Seasonal and spatial variations in leaf nitrogen content
and resorption in a Quercus serrata canopy. Tree Physiology, 27, 63-70.
Millard, P. & Grelet, G.A. (2010) Nitrogen storage and remobilization by trees: ecophysiological
relevance in a changing world. Tree Physiology, 30, 1083-1095.
238
Mohan, J.E., Clark, J.S. & Schlesinger, W.H. (2007) Long-term CO2 enrichment of a forest
ecosystem: Implications for forest regeneration and succession. Ecol. Appl., 17, 11981212.
Monserud, R.A. & Sterba, H. (1996) A basal area increment model for individual trees growing
in even- and uneven-aged forest stands in Austria. Forest Ecology and Management, 80,
57-80.
Montgomery, R. (2004) Relative importance of photosynthetic physiology and biomass
allocation for tree seedling growth across a broad light gradient. Tree Physiology, 24,
155-167.
Myers, R.H. (1986) Classical and Modern Regression with Applications. Duxbury Press, Boston.
Neilsen, W.A. & Lynch, T. (1998) Implications of pre- and post-fertilizing changes in growth
and nitrogen pools following multiple applications of nitrogen fertilizer to a Pinus radiata
stand over 12 years. Plant and Soil, 202, 295-307.
Newman, G.S., Arthur, M.A. & Muller, R.N. (2006) Above- and belowground net primary
production in a temperate mixed deciduous forest. Ecosystems, 9, 317-329.
Nezat, C.A., Blum, J.D., Yanai, R.D. & Park, B.B. (2008) Mineral sources of calcium and
phosphorus in soils of the Northeastern United States. Soil Science Society of America
Journal, 72, 1786-1794.
Nilsen, P. & Abrahamsen, G. (2003) Scots pine and Norway spruce stands responses to annual
N, P and Mg fertilization. Forest Ecology and Management, 174, 221-232.
Nobis, M. & Hunziker, U. (2005) Automatic thresholding for hemispherical canopy-photographs
based on edge detection. Agricultural and Forest Meteorology, 128, 243-250.
Noguchi, M. & Yoshida, T. (2009) Individual-scale responses of five dominant tree species to
single-tree selection harvesting in a mixed forest in Hokkaido, northern Japan. Journal of
Forest Research, 14, 311-320.
Nowak, C.A., Downard, R.B. & White, E.H. (1991) Potassium trends in red pine plantations at
Pack-Forest, New-York. Soil Science Society of America Journal, 55, 847-850.
Oelmann, Y., Potvin, C., Mark, T., Werther, L., Tapernon, S. & Wilcke, W. (2010) Tree mixture
effects on aboveground nutrient pools of trees in an experimental plantation in Panama.
Plant and Soil, 326, 199-212.
Ogle, K. & Pacala, S.W. (2009) A modeling framework for inferring tree growth and allocation
from physiological, morphological and allometric traits. Tree Physiology, 29, 587-605.
Olander, L.P. & Vitousek, P.M. (2000) Regulation of soil phosphatase and chitinase activity by
N and P availability. Biogeochemistry, 49, 175-190.
239
Olano, J.M., Laskurain, N.A., Escudero, A. & De La Cruz, M. (2009) Why and where do adult
trees die in a young secondary temperate forest? The role of neighbourhood. Annals of
Forest Science, 66.
Page, B.D. & Mitchell, M.J. (2008) Influences of a calcium gradient on soil inorganic nitrogen in
the Adirondack Mountains, New York. Ecol. Appl., 18, 1604-1614.
Park, B.B., Yanai, R.D., Fahey, T.J., Bailey, S.W., Siccama, T.G., Shanley, J.B. & Cleavitt, N.L.
(2008) Fine root dynamics and forest production across a calcium gradient in northern
hardwood and conifer ecosystems. Ecosystems, 11, 325-341.
Pastor, J., Aber, J.D., McClaugherty, C.A. & Melillo, J.M. (1984) Above-ground production and
N and P cycling along a nitrogen mineralization gradient on Blackhawk Island,
Wisconsin. Ecology, 65, 256-268.
Pearson, H.L. & Vitousek, P.M. (2002) Soil phosphorus fractions and symbiotic nitrogen
fixation across a substrate-age gradient in Hawaii. Ecosystems, 5, 587-596.
Peichl, M. & Arain, M.A. (2007) Allometry and partitioning of above- and belowground tree
biomass in an age-sequence of white pine forests. Forest Ecology and Management, 253,
68-80.
Perakis, S., Maguire, D., Bullen, T., Cromack, K., Waring, R. & Boyle, J. (2006) Coupled
nitrogen and calcium cycles in forests of the Oregon Coast Range. Ecosystems, 9, 63-74.
Piatek, K.B., Munasinghe, P., Peterjohn, W.T., Adams, M.B. & Cumming, J.R. (2009) Oak
contribution to litter nutrient dynamics in an Appalachian forest receiving elevated
nitrogen and dolomite. Canadian Journal of Forest Research-Revue Canadienne De
Recherche Forestiere, 39, 936-944.
Plassard, C. & Dell, B. (2010) Phosphorus nutrition of mycorrhizal trees. Tree Physiology, 30,
1129-1139.
Poorter, L. & Arets, E. (2003) Light environment and tree strategies in a Bolivian tropical moist
forest: an evaluation of the light partitioning hypothesis. Plant Ecology, 166, 295-306.
Poorter, L., McDonald, I., Alarcon, A., Fichtler, E., Licona, J.C., Pena-Claros, M., Sterck, F.,
Villegas, Z. & Sass-Klaassen, U. (2010) The importance of wood traits and hydraulic
conductance for the performance and life history strategies of 42 rainforest tree species.
New Phytologist, 185, 481-492.
Porder, S. & Chadwick, O.A. (2009) Climate and soil-age constraints on nutrient uplift and
retention by plants. Ecology, 90, 623-636.
Porder, S., Clark, D.A. & Vitousek, P.M. (2006) Persistence of rock-derived nutrients in the wet
tropical forests of La Selva, Costa Rica. Ecology, 87, 594-602.
240
Porder, S., Vitousek, P.M., Chadwick, O.A., Chamberlain, C.P. & Hilley, G.E. (2007) Uplift,
erosion, and phosphorus limitation in terrestrial ecosystems. Ecosystems, 10, 158-170.
Potvin, C. & Dutilleul, P. (2009) Neighborhood effects and size-asymmetric competition in a
tree plantation varying in diversity. Ecology, 90, 321-327.
Pyke, C.R., Condit, R., Aguilar, S. & Lao, S. (2001) Floristic composition across a climatic
gradient in a neotropical lowland forest. Journal of Vegetation Science, 12, 553-566.
Raat, K.J., Tietema, A. & Verstraten, J.M. (2010) Nitrogen turnover in fresh Douglas fir litter
directly after additions of moisture and inorganic nitrogen. Plant and Soil, 330, 115-126.
Rakhteenko, L.I. (1981) Photosynthetic activity and growth of pine in uneven aged stands under
the effect of mineral fertilizers. Lesovedenie, 66-72.
Reich, P.B., Grigal, D.F., Aber, J.D. & Gower, S.T. (1997) Nitrogen mineralization and
productivity in 50 hardwood and conifer stands on diverse soils. Ecology, 78, 335-347.
Reich, R.M., Bonham, C.D., Aguirre-Bravo, C. & Chazaro-Basaneza, M. (2010) Patterns of tree
species richness in Jalisco, Mexico: relation to topography, climate and forest structure.
Plant Ecology, 210, 67-84.
Reiter, I.M., Haberle, K.H., Nunn, A.J., Heerdt, C., Reitmayer, H., Grote, R. & Matyssek, R.
(2005) Competitive strategies in adult beech and spruce: space-related foliar carbon
investment versus carbon gain. Oecologia, 146, 337-349.
Reynolds, H.L., Mittelbach, G.G., Darcy-Hall, T.L., Houseman, G.R. & Gross, K.L. (2007) No
effect of varying soil resource heterogeneity on plant species richness in a low fertility
grassland. Journal of Ecology, 95, 723-733.
Richards, A.E., Forrester, D.I., Bauhus, J. & Scherer-Lorenzen, M. (2010) The influence of
mixed tree plantations on the nutrition of individual species: a review. Tree Physiology,
30, 1192-1208.
Rodenkirchen, H., Gottlein, A., Kozovits, A.R., Matyssek, R. & Grams, T.E.E. (2009) Nutrient
contents and efficiencies of beech and spruce saplings as influenced by competition and
O-3/CO2 regime. European Journal of Forest Research, 128, 117-128.
Romme, W.H., Tinker, D.B., Stakes, G.K. & Turner, M.G. (2009) Does inorganic nitrogen limit
plant growth 3-5 years after fire in a Wyoming, USA, lodgepole pine forest? Forest
Rosberg, I., Frank, J. & Stuanes, A.O. (2006) Effects of liming and fertilization on tree growth
and nutrient cycling in a Scots pine ecosystem in Norway. Forest Ecology and
Rowell, D.P. (2009) Projected midlatitude continental summer drying: North America versus
Europe. Journal of Climate, 22, 2813-2833.
241
Sanchez-Gomez, D., Zavala, M.A., Van Schalkwijk, D.B., Urbieta, I.R. & Valladares, F. (2008)
Rank reversals in tree growth along tree size, competition and climatic gradients for four
forest canopy dominant species in Central Spain. Annals of Forest Science, 65.
Schaberg, P.G., Tilley, J.W., Hawley, G.J., DeHayes, D.H. & Bailey, S.W. (2006) Associations
of calcium and aluminum with the growth and health of sugar maple trees in Vermont.
Schaetzl, R.J., Knapp, B.D. & Isard, S.A. (2005) Modeling soil temperatures and the mesicfrigid boundary in the central great lakes region, 1951-2000. Soil Science Society of
America Journal, 69, 2033-2040.
Schimel, D.S., Braswell, B.H., McKeown, R., Ojima, D.S., Parton, W.J. & Pulliam, W. (1996)
Climate and nitrogen controls on the geography and timescales of terrestrial
biogeochemical cycling. Global Biogeochemical Cycles, 10, 677-692.
Schimel, D.S., Emanuel, W., Rizzo, B., Smith, T., Woodward, F.I., Fisher, H., Kittel, T.G.F.,
McKeown, R., Painter, T., Rosenbloom, N., Ojima, D.S., Parton, W.J., Kicklighter,
D.W., McGuire, A.D., Melillo, J.M., Pan, Y., Haxeltine, A., Prentice, C., Sitch, S.,
Hibbard, K., Nemani, R., Pierce, L., Running, S., Borchers, J., Chaney, J., Neilson, R. &
Braswell, B.H. (1997) Continental scale variability in ecosystem processes: Models, data,
and the role of disturbance. Ecol. Monogr., 67, 251-271.
Schreeg, L.A., Kobe, R.K. & Walters, M.B. (2005) Tree seedling growth, survival, and
morphology in response to landscape-level variation in soil resource availability in
northern Michigan. Can. J. Forest Res., 35, 263-273.
Schuur, E.A.G. & Matson, P.A. (2001) Net primary productivity and nutrient cycling across a
mesic to wet precipitation gradient in Hawaiian montane forest. Oecologia, 128, 431-442.
Schwartz, J.W., Nagel, L.M. & Webster, C.R. (2005) Effects of uneven-aged management on
diameter distribution and species composition of northern hardwoods in Upper Michigan.
Siddique, I., Engel, V.L., Parrotta, J.A., Lamb, D., Nardoto, G.B., Ometto, J., Martinelli, L.A. &
Schmidt, S. (2008) Dominance of legume trees alters nutrient relations in mixed species
forest restoration plantings within seven years. Biogeochemistry, 88, 89-101.
Silvertown, J. (2004) Plant coexistence and the niche. Trends in Ecology & Evolution, 19, 605611.
Simard, S.W. (2009) The foundational role of mycorrhizal networks in self-organization of
interior Douglas-fir forests. Forest Ecology and Management, 258, S95-S107.
Simioni, G., Gignoux, J., Le Roux, X., Appe, R. & Benest, D. (2004) Spatial and temporal
variations in leaf area index, specific leaf area and leaf nitrogen of two co-occurring
savanna tree species. Tree Physiology, 24, 205-216.
242
Singh, A. & Agrawal, M. (2008) Acid rain and its ecological consequences. J. Env. Biol., 29, 1524.
Sinsabaugh, R.L., Reynolds, H. & Long, T.M. (2000) Rapid assay for amidohydrolase (urease)
activity in environmental samples. Soil Biology & Biochemistry, 32, 2095-2097.
Sitch, S., Huntingford, C., Gedney, N., Levy, P.E., Lomas, M., Piao, S.L., Betts, R., Ciais, P.,
Cox, P., Friedlingstein, P., Jones, C.D., Prentice, I.C. & Woodward, F.I. (2008)
Evaluation of the terrestrial carbon cycle, future plant geography and climate-carbon
cycle feedbacks using five Dynamic Global Vegetation Models (DGVMs). Global
Smethurst, P.J. (2010) Forest fertilization: Trends in knowledge and practice compared to
agriculture. Plant and Soil, 335, 83-100.
Snyder, R.E. & Chesson, P. (2004) How the spatial scales of dispersal, competition, and
environmental heterogeneity interact to affect coexistence. American Naturalist, 164,
633-650.
Soethe, N., Lehmann, J. & Engels, C. (2008) Nutrient availability at different altitudes in a
tropical montane forest in Ecuador. Journal of Tropical Ecology, 24, 397-406.
Soolanayakanahally, R.Y., Guy, R.D., Silim, S.N., Drewes, E.C. & Schroeder, W.R. (2009)
Enhanced assimilation rate and water use efficiency with latitude through increased
photosynthetic capacity and internal conductance in balsam poplar (Populus balsamifera
L.). Plant Cell and Environment, 32, 1821-1832.
Sparks, J.P., Walker, J., Turnipseed, A. & Guenther, A. (2008) Dry nitrogen deposition estimates
over a forest experiencing free air CO2 enrichment. Global Change Biology, 14, 768-781.
Sperry, J.S., Meinzer, F.C. & McCulloh, K.A. (2008) Safety and efficiency conflicts in hydraulic
architecture: scaling from tissues to trees. Plant Cell and Environment, 31, 632-645.
St Clair, S.B. & Lynch, J.P. (2005) Base cation stimulation of mycorrhization and photosynthesis
of sugar maple on acid soils are coupled by foliar nutrient dynamics. New Phytologist,
165, 581-590.
St Clair, S.B., Sharpe, W.E. & Lynch, J.P. (2008) Key interactions between nutrient limitation
and climatic factors in temperate forests: a synthesis of the sugar maple literature. Can. J.
Forest Res., 38, 401-414.
Stadt, K.J., Huston, C., Coates, K.D., Feng, Z.L., Dale, M.R.T. & Lieffers, V.J. (2007)
Evaluation of competition and light estimation indices for predicting diameter growth in
mature boreal mixed forests. Annals of Forest Science, 64, 477-490.
Stan, A.B. & Daniels, L.D. (2010) Growth releases of three shade-tolerant species following
canopy gap formation in old-growth forests. Journal of Vegetation Science, 21, 74-87.
243
Stape, J.L., Binkley, D. & Ryan, M.G. (2008) Production and carbon allocation in a clonal
Eucalyptus plantation with water and nutrient manipulations. Forest Ecology and
Sternberg, M. & Shoshany, M. (2001) Aboveground biomass allocation and water content
relationships in Mediterranean trees and shrubs in two climatological regions in Israel.
Plant Ecology, 157, 171-179.
Stevens, P.A., Harrison, A.F., Jones, H.E., Williams, T.G. & Hughes, S. (1993) Nitrate leaching
from a sitka spruce plantation and the effect of fertilization with phosphorus and
potassium. Forest Ecology and Management, 58, 233-247.
Stoll, P. & Newbery, D.M. (2005) Evidence of species-specific neighborhood effects in the
dipterocarpaceae of a Bornean rain forest. Ecology, 86, 3048-3062.
Stoll, P., Weiner, J., Muller-Landau, H., Muller, E. & Hara, T. (2002) Size symmetry of
competition alters biomass-density relationships. Proceedings of the Royal Society of
London Series B-Biological Sciences, 269, 2191-2195.
Sverdrup, H., Thelin, G., Robles, M., Stjernquist, I. & Sorensen, J. (2006) Assesing nutrient
sustainability of forest production for different tree species considering Ca, Mg, K, N and
P at Bjornstorp Estate, Sweden. Biogeochemistry, 81, 219-238.
Swenson, N.G. & Enquist, B.J. (2008) The relationship between stem and branch wood specific
gravity and the ability of each measure to predict leaf area. American Journal of Botany,
95, 516-519.
Swenson, N.G. & Enquist, B.J. (2009) Opposing assembly mechanisms in a Neotropical dry
forest: implications for phylogenetic and functional community ecology. Ecology, 90,
2161-2170.
Swenson, N.G. & Weiser, M.D. (2010) Plant geography upon the basis of functional traits: an
example from eastern North American trees. Ecology, 91, 2234-2241.
Tanner, E.V.J., Vitousek, P.M. & Cuevas, E. (1998) Experimental investigation of nutrient
limitation of forest growth on wet tropical mountains. Ecology, 79, 10-22.
TerMikaelian, M.T. & Korzukhin, M.D. (1997) Biomass equations for sixty-five North
American tree species. Forest Ecology and Management, 97, 1-24.
Thorpe, H.C., Astrup, R., Trowbridge, A. & Coates, K.D. (2010) Competition and tree crowns:
A neighborhood analysis of three boreal tree species. Forest Ecology and Management,
259, 1586-1596.
Tilman, D. (1988) Plant Strategies and the Dynamics and Function of Plant Communities.
Princeton University Press, Princeton, NJ.
244
Townsend, A.R., Asner, G.P. & Cleveland, C.C. (2008) The biogeochemical heterogeneity of
tropical forests. Trends in Ecology & Evolution, 23, 424-431.
Treseder, K.K. & Vitousek, P.M. (2001) Effects of soil nutrient availability on investment in
acquisition of N and P in Hawaiian rain forests. Ecology, 82, 946-954.
Triboulot, M.B., Pritchard, J. & Levy, G. (1997) Effects of potassium deficiency on cell water
relations and elongation of tap and lateral roots of maritime pine seedlings. New
Phytologist, 135, 183-190.
Tripler, C.E., Kaushal, S.S., Likens, G.E. & Walter, M.T. (2006) Patterns in potassium dynamics
in forest ecosystems. Ecology Letters, 9, 451-466.
Turner, B.L. (2008) Resource partitioning for soil phosphorus: a hypothesis. Journal of Ecology,
96, 698-702.
Uriarte, M., Canham, C.D. & Root, R.B. (2002) A model of simultaneous evolution of
competitive ability and herbivore resistance in a perennial plant. Ecology, 83, 2649-2663.
Uriarte, M., Canham, C.D., Thompson, J. & Zimmerman, J.K. (2004) A neighborhood analysis
of tree growth and survival in a hurricane-driven tropical forest. Ecological Monographs,
74, 591-614.
Vadeboncoeur, M.A. (2010) Meta-analysis of fertilization experiments indicates multiple
limiting nutrients in northeastern deciduous forests. Canadian Journal of Forest
Valdespino, P., Romualdo, R., Cadenazzi, L. & Campo, J. (2009) Phosphorus cycling in primary
and secondary seasonally dry tropical forests in Mexico. Annals of Forest Science, 66.
Vallano, D.M. & Sparks, J.P. (2008) Quantifying foliar uptake of gaseous nitrogen dioxide using
enriched foliar delta N-15 values. New Phytologist, 177, 946-955.
van den Driessche, R., Thomas, B.R. & Kamelchuk, D.P. (2008) Effects of N, NP, and NPKS
fertilizers applied to four-year old hybrid poplar plantations. New Forests, 35, 221-233.
van der Heijden, M.G.A. & Horton, T.R. (2009) Socialism in soil? The importance of
mycorrhizal fungal networks for facilitation in natural ecosystems. Journal of Ecology,
97, 1139-1150.
van Kuijk, M. & Anten, N.P.R. (2009) Whole-canopy nitrogen-use efficiency of pioneer species
in early secondary forest succession in Vietnam. Ecological Research, 24, 811-820.
Vanhellemont, M., Verheyen, K., Staelens, J. & Hermy, M. (2010) Factors affecting radial
growth of the invasive Prunus serotina in pine plantations in Flanders. European Journal
of Forest Research, 129, 367-375.
245
Vanninen, P. & Makela, A. (2000) Needle and stem wood production in Scots pine (Pinus
sylvestris) trees of different age, size and competitive status. Tree Physiology, 20, 527533.
Vergnon, R., Dulvy, N.K. & Freckleton, R.P. (2009) Niches versus neutrality: uncovering the
drivers of diversity in a species-rich community. Ecology Letters, 12, 1079-1090.
Verhoeven, J.T.A., Koerselman, W. & Meuleman, A.F.M. (1996) Nitrogen- or phosphoruslimited growth in herbaceous, wet vegetation: Relations with atmospheric inputs and
management regimes. Trends in Ecology & Evolution, 11, 494-497.
Vesterdal, L., Schmidt, I.K., Callesen, I., Nilsson, L.O. & Gundersen, P. (2008) Carbon and
nitrogen in forest floor and mineral soil under six common European tree species. Forest
Vettenranta, J. (1999) Distance-dependent models for predicting the development of mixed
coniferous forests in Finland. Silva Fennica, 33, 51-72.
Vilela, A.E., Rennella, M.J. & Ravetta, D.A. (2003) Responses of tree-type and shrub-type
Prosopis (mimosaceae) taxa to water and nitrogen availabilities. Forest Ecology and
Vincent, A.G., Turner, B.L. & Tanner, E.V.J. (2010) Soil organic phosphorus dynamics
following perturbation of litter cycling in a tropical moist forest. European Journal of
Soil Science, 61, 48-57.
Vitousek, P.M. & Farrington, H. (1997) Nutrient limitation and soil development: Experimental
test of a biogeochemical theory. Biogeochemistry, 37, 63-75.
Vitousek, P.M., Porder, S., Houlton, B.Z. & Chadwick, O.A. (2010) Terrestrial phosphorus
limitation: mechanisms, implications, and nitrogen-phosphorus interactions. Ecological
Applications, 20, 5-15.
Vizcayno-Soto, G. & Cote, B. (2004) Boundary-line approach to determine standards of nutrition
for mature trees from spatial variation of growth and foliar nutrient concentrations in
natural environments. Communications in Soil Science and Plant Analysis, 35, 29652985.
Vogel, J.G., Bond-Lamberty, B.P., Schuur, E.A.G., Gower, S.T., Mack, M.C., O'Connell,
K.E.B., Valentine, D.W. & Ruess, R.W. (2008) Carbon allocation in boreal black spruce
forests across regions varying in soil temperature and precipitation. Global Change
Biology, 14, 1503-1516.
Vriesendorp, C.F. (2003) Maintenance of diversity in a neotropical seedling community. Ph.D.
dissertation, Michigan State University, East Lansing, Michigan.
246
Wallace, Z.P., Lovett, G.M., Hart, J.E. & Machona, B. (2007) Effects of nitrogen saturation on
tree growth and death in a mixed-oak forest. Forest Ecology and Management, 243, 210218.
Walters, M.B. & Reich, P.B. (1997) Growth of Acer saccharum seedlings in deeply shaded
understories of northern Wisconsin: Effects of nitrogen and water availability. Canadian
Journal of Forest Research-Revue Canadienne De Recherche Forestiere, 27, 237-247.
Wang, Y.C., Larsen, C.P.S. & Kronenfeld, B.J. (2010) Effects of clearance and fragmentation on
forest compositional change and recovery after 200 years in western New York. Plant
Ecology, 208, 245-258.
Wang, Y.P., Houlton, B.Z. & Field, C.B. (2007) A model of biogeochemical cycles of carbon,
nitrogen, and phosphorus including symbiotic nitrogen fixation and phosphatase
production. Global Biogeochemical Cycles, 21.
Watmough, S.A. (2008) Element mobility and partitioning along a soil acidity gradient in central
Ontario forests, Canada. Environmental Geochemistry and Health, 30, 431-444.
Wedin, D. & Tilman, D. (1993) Competition among Grasses Along a Nitrogen Gradient - Initial
Conditions and Mechanisms of Competition. Ecological Monographs, 63, 199-229.
Weigelt, A. & Jolliffe, P. (2003) Indices of plant competition. Journal of Ecology, 91, 707-720.
Weiner, J. (2004) Allocation, plasticity and allometry in plants. Perspectives in Plant Ecology
Evolution and Systematics, 6, 207-215.
Wikstrom, F. & Ericsson, T. (1995) Allocation of mass in trees subject to nitrogen and
magnesium limitation. Tree Physiology, 15, 339-344.
Will, R.E., Barron, G.A., Burkes, E.C., Shiver, B. & Teskey, R.O. (2001) Relationship between
intercepted radiation, net photosynthesis, respiration, and rate of stem volume growth of
Pinus taeda and Pinus elliottii stands of different densities. Forest Ecology and
Will, R.E., Markewitz, D., Hendrick, R.L., Meason, D.F., Crocker, T.R. & Borders, B.E. (2006)
Nitrogen and phosphorus dynamics for 13-year-old loblolly pine stands receiving
complete competition control and annual N fertilizer. Forest Ecology and Management,
227, 155-168.
Wilmot, T.R., Ellsworth, D.S. & Tyree, M.T. (1996) Base cation fertilization and liming effects
on nutrition and growth of Vermont sugar maple stands. Forest Ecology and
Wimberly, M.C. & Bare, B.B. (1996) Distance-dependent and distance-independent models of
Douglas-fir and western hemlock basal area growth following silvicultural treatment.
247
Wood, T.E., Lawrence, D. & Clark, D.A. (2005) Variation in leaf litter nutrients of a Costa
Rican rain forest is related to precipitation. Biogeochemistry, 73, 417-437.
Woods, K.D. (2000) Dynamics in late-successional hemlock-hardwood forests over three
decades. Ecology, 81, 110-126.
Wright, E.F., Coates, K.D., Canham, C.D. & Bartemucci, P. (1998) Species variability in growth
response to light across climatic regions in northwestern British Columbia. Canadian
Journal of Forest Research-Revue Canadienne De Recherche Forestiere, 28, 871-886.
Wu, F.Z., Bao, W.K., Li, F.L. & Wu, N. (2008) Effects of drought stress and N supply on the
growth, biomass partitioning and water-use efficiency of Sophora davidii seedlings.
Environmental and Experimental Botany, 63, 248-255.
Wyckoff, P.H. & Clark, J.S. (2005) Tree growth prediction using size and exposed crown area.
Canadian Journal of Forest Research-Revue Canadienne De Recherche Forestiere, 35,
13-20.
Xia, J.Y. & Wan, S.Q. (2008) Global response patterns of terrestrial plant species to nitrogen
addition. New Phytologist, 179, 428-439.
Xu, D., Dell, B., Malajczuk, N. & Gong, M. (2002) Effects of P fertilisation on productivity and
nutrient accumulation in a Eucalyptus grandis x E-urophylla plantation in southern China.
Yadav, S.K., Chakrapani, G.J. & Gupta, M.K. (2008) An experimental study of dissolution
kinetics of Calcite, Dolomite, Leucogranite and Gneiss in buffered solutions at
temperature 25 and 5 degrees C. Environmental Geology, 53, 1683-1694.
Yanai, R.D., Fisk, M.C., Fahey, T.J., Cleavitt, N.L. & Park, B.B. (2008) Identifying roots of
northern hardwood species: patterns with diameter and depth. Canadian Journal of
Forest Research-Revue Canadienne De Recherche Forestiere, 38, 2862-2869.
Ylimartimo, A. (1990) The effect of nitrogen and potassium availability on scleroderris canker of
scots pine-seedlings. Water Air and Soil Pollution, 54, 307-313.
Zak, D.R., Host, G.E. & Pregitzer, K.S. (1989) Regional variability in nitrogen mineralization,
nitrification, and overstory biomass in northern lower Michigan. Can. J. Forest Res., 19,
1521-1526.
Zhao, D.H., Borders, B., Wilson, M. & Rathbun, S.L. (2006) Modeling neighborhood effects on
the growth and survival of individual trees in a natural temperate species-rich forest.
Ecological Modelling, 196, 90-102.
248

revisiting soil resource limitation: resource predictors of tree growth

Transcription

Similar documents

Technical Sheet - Microbes Biosciences

Revel Casino News 2013

Angiosperm Latin Name: Callisia fragrans Common Name: Basket

AZOMITE granular - Down To Earth Fertilizer

Pro-Grow

NEW SOLUTION

MT55 | Mini Tracked Loaders

Bromeliaceae

- StopWaste.org

Angyo Star - Southern Living Plants