A holistic approach to Mima mound prairie restoration

Transcription

Eastern Washington University
EWU Digital Commons
EWU Masters Thesis Collection
Student Research and Creative Works
2013
A holistic approach to Mima mound prairie
restoration
Kristin R. Anicito
Follow this and additional works at: http://dc.ewu.edu/theses
Recommended Citation
Anicito, Kristin R., "A holistic approach to Mima mound prairie restoration" (2013). EWU Masters Thesis Collection. Paper 115.
This Thesis is brought to you for free and open access by the Student Research and Creative Works at EWU Digital Commons. It has been accepted for
inclusion in EWU Masters Thesis Collection by an authorized administrator of EWU Digital Commons. For more information, please contact
[email protected].
A Holistic Approach to Mima mound Prairie Restoration
________________________________________________________________________
A Thesis
Presented to
Cheney, Washington
________________________________________________________________________
In Partial Fulfillment of the Requirements
For the Degree
Master of Science
________________________________________________________________________
By
Kristin R. Anicito
Fall 2013
Thesis of Kristin Rae Anicito Approved by
______________________________________________
Dr. Rebecca Brown, Graduate study committee chair
Date_____________
______________________________________________
Dr. Robin O’Quinn, Graduate study committee member
Date_____________
______________________________________________
Dr. Suzanne Schwab, Graduate study committee member
Date_____________
______________________________________________
Dr. Stacy Warren, Graduate study committee member
Date_____________
ii
Master’s Thesis
In presenting this thesis in partial fulfillment of the requirements for a master’s degree at
Eastern Washington University, I agree that the JFK Library shall make copies freely
available for inspection. I further agree that copying of this project in whole or in part is
allowable only for scholarly purposes. It is understood, however, that any copying or
publication of this thesis for commercial purposes, or for financial gain, shall not be
allowed without my written permission.
Signature________________________
Kristin Anicito
Date____________________________
iii
Overall Abstract
Invasive winter annual grass (IWAGs) abundance has increased throughout much
of the western United States, with serious consequences for biodiversity and ecosystem
function. While a variety of tools have been used to control IWAGS, they often fail to
achieve the goal of creating a self-sustaining native plant community. Therefore, a major
goal of my thesis has been to explore and test restoration approaches that might help
realize such plant communities. These approaches include ecologically based invasive
plant management (EBIPM) and assisted succession. EBIPM is a holistic approach that
aims to restore ecological functions damaged by IWAGs with a combination of
restoration tools that remove the IWAGs and promote native growth. Assisted succession
can be used to further restore the native plant community by dispersing fast growing
native annual seed (instead of native perennial), forcing competition with invasive
annuals. It is thought that native perennials should be able to more easily succeed
following the native annuals that they evolved with rather than invasive annuals. Both
approaches seek to increase community resistance to invasion, which should promote
self-sustainability. Community resistance to invasion in western bunchgrass communities
is also enhanced by the presence of biological soil crust (biocrust), which is the moss,
lichen, algae, and cyanobacteria growing on the soil. Biocrust has been shown to prevent
seeds from establishing by creating a barrier between the seed and the soil but has only
been tested in arid areas. Much of the west, including Turnbull National Wildlife Refuge,
where my study is focused, is in semiarid habitat, with nearly twice the rainfall. It is not
known how biocrust affects IWAG invasion in such conditions, or even how biocrust
responds to invasive species control techniques such as herbicide and fire.
iv
I tested the effects of restoration techniques that control IWAGs, increase native
species abundance, and promote community resistance to invasion on plant community
and biocrust composition over a range of habitats in a Mima mound Prairie at Turnbull
National Wildlife Refuge (TNWR). Mima mounds are discrete mounds of soils,
surrounded by shallow intermound soils; mound size and intermound vegetation growth
is dependent on substrate type. The restoration techniques included prescribed fire (fall
burn), Outrider herbicide applied in late fall, and native annual and perennial seed
addition (seeds added in winter). I further tested the effect of boot trampling on biocrust.
Plants were sampled before and the first growing season after application. I found that 1)
V. dubia increased 15-fold on mounds from 2009 to 2013, and where V. dubia was
abundant, lichen cover was low but moss cover was high, 2) Outrider® herbicide was the
most effective technique at reducing invasive species and had minimal impact on natives,
on mounds; on intermounds native cover actually increased with herbicide, 3) Fire
effectively increased native species cover on mounds but had little effect on intermounds;
fire had little effect on invasive cover, 4) annual seed addition effectively increased
native cover while perennial seed addition had little effect the first year, and 5) in
intermounds areas where biocrust was abundant, treatments had little effect on the
biocrust community or on exotic plant species. My results can be used to help land
managers decide appropriate restoration approaches to increase community sustainability
in semiarid regions, however long-term monitoring is needed to test whether these results
are truly sustainable.
v
Preface
Invasive winter annual grasses (IWAGs) can reduce soil nutrients and water
availability, decrease native species and biological soil crust (biocrust; moss, lichens,
algae, and cyanobacteria) cover and diversity, and increase fire frequency and intensity,
all of which perpetuates further IWAG growth. When IWAG abundance is high, the
natural succession of semiarid prairies alters to a self-perpetuating, IWAG dominated
system with reduced ecological function than the natural prairie ecosystem.
In natural prairie succession, biocrust stabilizes the bare soil, and is then followed
by annuals, and finally perennial bunchgrasses with biocrust in the spaces between
individuals. When intact, biocrust communities provide important ecological functions
such as nitrogen fixation, erosion prevention, and limiting IWAG establishment by
creating a barrier between the seed and the soil. However, biocrust (especially in arid
regions) is extremely sensitive to disturbance and can take decades to recover, allowing
IWAGs to establish. Most biocrust research has been completed in arid regions, which
experience less rainfall and hotter temperatures than semiarid regions. It is not known
how biocrust responds to IWAGs or disturbance in semiarid regions.
It is important to understand the interactions between IWAGs, native plants,
biocrust, and restoration tools in semiarid regions if we hope to return to a more natural
and ecologically functional prairie ecosystem. As IWAGs expand their ranges and
population sizes throughout the west and into bunchgrass/biocrust communities,
innovative and thoughtful approaches are needed to control invasives, restore natives, and
increase resistance to invasion, thereby creating a self-sustaining ecosystem. Two such
approaches are ecologically based invasive plant management (EBIPM) and assisted
vi
succession, which I used for this study, and which are described in chapter 2. I applied
these approaches in the semiarid Mima mound prairie at Turnbull National Wildlife
Refuge.
The Mima mound prairie at Turnbull National Wildlife Refuge (TNWR) has
experienced an influx of IWAGs over the last decade. Mima mounds are discrete mounds
of soil over alluvial gravel or basalt bedrock substrate. The intermounds areas over basalt
bedrock have extremely shallow soils with well-developed biocrust communities. Brandy
Reynecke’s thesis (EWU 2012) and Jessica Bryant’s McNair study (Bryant et al. 2012)
documented an abundance of IWAGs and native species on basalt and alluvial Mima
mounds in 2008-2009. However, anecdotally, there appeared to have been a substantial
increase in the IWAG, Ventenata dubia, in the intermounds areas, which caused the
community to change from biocrust to dense IWAG cover. Refuge Managers,
particularly Mike Rule, would like to restore the Mima mound prairie to a native,
sustainable plant community, which has inspired (and funded) this research.
My primary research goals were to develop an approach to restore the Mima
Mound Prairie at Turnbull National Wildlife Refuge to a self-sustainable native plant
community. In doing so, I wanted to address the gap in biocrust research in semi-arid
regions, determine how disturbance caused by restoration affects biocrust, and determine
whether the rapid change in IWAG abundance over the past four years is related to
biocrust. To meet these goals, I conducted a three-part study, with each part represented
by one of the three chapters of my thesis. Each chapter was written as separate, standalone paper to facilitate publication in peer-review journals. The three chapters include:
1) Ventenata (Ventenata dubia) invasion in an eastern WA Mima mound prairie, 2)
vii
Prescribed fire, herbicide, and annual seed addition used to increase community
resistance to invasion, and 3) The effects of fire, herbicide, and human trampling on
semiarid biological soil crust and the invasive annual grass Ventenata dubia. In Chapter 1
I discuss the population explosion of Ventenata dubia at TNWR from 2009 to 2013 and
how its abundance compares to similar IWAGs and biocrust through out the prairie. This
chapter has been formatted for the journal, Weed Science, where it has been submitted for
publication. In Chapter 2, I analyzed the effects of fire, herbicide, and native annual and
perennial seed addition (assisted succession) on the native and invasive plant
communities growing on Mima mounds. In Chapter 3, I tested how basalt intermound
biocrust and vascular plant communities responded to fire, herbicide, and trampling.
It is my hope that my findings will assist Turnbull National Wildlife Refuge
managers in developing a long-term, integrated plan that will restore a self-sustaining
native dominated system.
viii
Acknowledgements
I would like to thank my advisor Dr. Rebecca Brown for all the help and guidance she
has given me throughout this process and for allowing me to create a project that I am
passionate about. I would also like to thank my committee members Dr. Suzanne
Schwab, Dr. Robin O’Quinn, and Dr. Stacy Warren for always allowing me to pester
them with questions. I am especially grateful for Mike Rule and Turnbull National
Wildlife Refuge for making my project happen. I would also like to extend special
gratitude to US Fish and Wildlife for funding my research. I would like to thank Turnbull
National Wildlife Refuge fire crew for setting fire to the prairie for me, Gil Cook for his
help apply herbicide, and my wonder field assistants Nita Rektor and Natalie Scott for
making field work a blast.
I would especially like to thank my fiancé Pat O’Brien, my parents, siblings, and friends
for all the moral support and love they have given me while my life became consumed
with invasive annual grasses.
ix
Table of Contents
Overall Abstract…………………………………………………………………………..iv
Preface …………………………………………………………………………………...vi
Acknowledgements……………………………………………………………………….ix
List of Figures…………………………………………………………………………….xi
List of Tables……………………………………………………………………………xiii
Chapter 1- Ventenata (Ventenata dubia) invasion in an eastern WA Mima mound prairie
Abstract……………………………………………………………………………………1
Introduction………………………………………………………………………………..3
Materials and Methods…………………………………………………………………….8
Results/Discussion……………………………………………………………………….11
Figures……………………………………………………………………………………17
Tables…………………………………………………………………………………….23
Chapter 2- Mound plant community responses to active restoration management
Abstract…………………………………………………………………………………..30
Introduction………………………………………………………………………………32
Materials and Methods…………………………………………………………………...40
Results……………………………………………………………………………………45
Discussion………………………………………………………………………………..48
Figures……………………………………………………………………………………55
Tables…………………………………………………………………………………….67
Chapter 3- Ventenata dubia and biological soil crust communities response to disturbance
Abstract…………………………………………………………………………………..91
Introduction………………………………………………………………………………93
Materials and Methods…………………………………………………………………...98
Results…………………………………………………………………………………..101
Discussion………………………………………………………………………………104
Figures…………………………………………………………………………………..110
Tables…………………………………………………………………………………...119
Conclusions and Management Implications……………………………………………137
Literature Cited…………………………………………………………………………140
x
List of Figures
Figure 1 Location of study area at Turnbull National Wildlife Refuge (TNWR) in
Cheney, WA and its location within Washington State…………………………….........17
Figure 2 Mean stem counts of IWAGs found in a 0.04 m2 corner of alluvial and basalt
mound plots from 2009, 2012, and 2013……………………………...............................18
Figure 3 Mean percent cover of ventenata, downy brome, and biocrust in all plot types in
2012……………………………..............................……………………………..............19
Figure 4 NMS, joint plot Ordination of the 2012 plant communities. Each triangle
represents an individual plot……..............................……………………………............20
Figure 5 NMS, joint plot Ordination of the 2012 biocrust communities on basalt
intermound plots……..............................……………………………..............................21
Figure 6 Relationship of (a) tall moss, (b) short moss, (c) total lichen cover, (d)
squamulose white lichen percent cover, and (e) biocrust diversity to ventenata percent
cover over 1 m2……..............................……………………………................................22
Figure 7 Aerial photograph of the Mima mound prairie at Turnbull National Wildlife
Refuge (TNWR)……..............................……………………………...............................55
Figure 8 Aerial photo of the Mima mound prairie; the colored polygons indicate the burn
areas……..............................…………………………….................................................56
Figure 9 Mean IWAG stem count changes from 2012 to 2013 after each treatment on
both basalt and alluvial mound plots found in 0.04 m2 corner the plot.............................57
Figure 10 Mean changes of native and invasive species richness per m2 from 2012 to
2013 on basalt and alluvial mounds after all treatments....................................................58
Figure 11 The effect of treatment on native and invasive species percent cover per m2
plots from 2012 to 2013 on basalt and alluvial mounds....................................................59
Figure 12 Mean change in Sisymbrium altissimum, Linaria dalmatica, and Vicia cracca
percent cover from 2012 to 2013 after treatment...............................................................60
xi
Figure 13 Mean change in native species cover from 2012 to 2013 on alluvial and basalt
burn plots that received seed or herbicide and seed addition.............................................61
Figure 14 Mean change of C. pulchella percent cover in 1 m2 plots from 2012 to 2013 on
basalt and alluvial mounds after treatment........................................................................62
Figure 15 Change of Madia gracilis abundance in 1 m2 plots from 2012 to 2013 on basalt
and alluvial mounds after each treatment..........................................................................63
Figure 16 Mean Bromus tectorum percent cover in P.f.-D7 rhizobacteria treated and
control plots from 2009, 2010, and 2012...........................................................................64
Figure 17 Mean total nitrogen in soil samples across all five sample sessions………….65
Figure 18 Mean 2012 nitrite/nitrate and ammonium levels near basalt and alluvial control
plots from each soil sample period....................................................................................66
Figure 19 Mean change in V. dubia stem counts after treatment from 2012 to 2013…..110
Figure 20 Mean change in V. duia percent cover and litter cover after treatment from
2012 to 2013....................................................................................................................111
Figure 21 The mean change in native and invasive species richness after each treatment
from 2012 to 2013............................................................................................................112
Figure 22 Mean change in native and invasive species percent cover after each treatment
from 2012 to 2013............................................................................................................113
Figure 23 NMS, joint plot Ordination of the 2012 and 2013 biocrust communities…...114
Figure 24 NMS ordination of the 2012 and 2013 biocrust communities………………115
Figure 25 Mean similarity of biocrust morphology composition per 1 m2 after each
treatment from 2012 to 2013............................................................................................116
Figure 26 Total monthly precipitation (mm) from 2007 to 2013 in Spokane, WA.........117
Figure 27 Mean monthly air temperature (C) from 2007 to 2013 in Spokane, WA......118
xii
List of Tables
Table 1 Summary of F-values from the repeated measures analyses to determine how
ventenata, downy brome and hairy chess respective stem counts varied with year,
substrate, and interaction effects of year and substrate………..........................................23
Table 2 Summary of F-values from a general linear model comparing change in
ventenata, downy brome and hairy chess stem count from 2009-2013 and how they
varied among substrate type…………………………………….......................................24
Table 3 Summary of F-values from a general linear model analyzing ventenata, downy
brome and hairy chess 2012 percent cover and how they varied among substrate type...25
Table 4 Regression of tall moss, short moss, total lichen, and squamulose white lichen
percent cover and biocrust diversity compared to ventenata cover in 1 m2 plots.………26
Table 5 All vascular plant species, lichen morphology groups, and moss morphology
groups recorded in 2009, 2012, and 2013 and the plot types within which each was
found..……………………………………...…………...………………………………..27
Table 6 Eight treatment combinations of fire, Outrider herbicide, and native annual and
perennial seed that were applied in fall 2012 to winter 2013..…………………………..67
Table 7 Summary of F and p-values from a mixed linear model analysis to determine the
effect of treatment and substrate type on total IWAG stem counts...……………………68
Table 8 Summary of t and p values from a Tukey post-hoc analysis that compares the
effect of treatment on total IWAG stem counts...………………………………………..69
Table 9 Summary of F and p values for two mixed linear model analyses comparing the
effects of substrate type, fire (burned), herbicide/seed application (HAP) and the
interaction of fire and herbicide/seed application on total IWAG stem counts..………..70
xiii
effects of substrate type, herbicide and seed addition type (annual seed only or annual and
perennial seed together) in burned plots on total IWAG stem counts.…………………..71
Table 11 Summary of F and p-values from two mixed linear model analyses to determine
the effect of treatment and substrate type on invasive and native species richness...……72
Table 12 Summary of t and p values from Tukey post-hoc analyses of two mixed linear
model analyses that compares the effect of treatment on native and invasive species
richness....………………………………………………………………………………..73
Table 13 Summary of F and p values from a mixed linear model comparing the effects of
substrate type, treatment, to total species richness.....…………...………………………74
Table 14 Summary of t and p-values from a Tukey post-hoc analysis of a mixed linear
model analyses to determine if treatments affected the change of native species richness
differently than invasive species richness.……...…………………………...…………..75
effects of substrate type, fire (burned), and herbicide/seed treatment (HAP) on invasive
and native species richness.………………...……………………………………………76
Table 16 Summary of F and p values from two mixed model analyses that compared the
effects of substrate type, herbicide, and seed dispersal after fire on invasive and native
species richness..……...…………………………...……………………………………..77
Table 17 Summary of F and p values from two mixed linear model analyses to determine
if treatment or substrate type influenced invasive and native species percent cover.……78
xiv
Table 18 Summary of t and p values from Tukey post-hoc analyses of two mixed linear
model analyses comparing the effect of treatments on native and invasive species percent
cover..………………...…….………………...…….………………...…………………..79
effects of substrate type, fire (burned), herbicide/seed treatment (HAP), and the
interaction of burning, herbicide, and seed addition on invasive and native species percent
cover.………………...…………………………………………………………………...81
species percent cover.………...……………………..…………………………………...82
Table 21 Summary of t and p-values from a Tukey post-hoc analysis a mixed linear
model analyses to determine if post burn treatments (herbicide and native annual and
native perennial seed) affected native species percent cover differently than invasive
species percent cover. Significant values are indicated in bold.……………...………….83
Table 22 Summary of F and p values from two mixed linear model analyses, which
compared the effects of substrate type and treatment to determine the success of annual
seed addition.…………...…………………………...…………………………………...84
Table 23 Summary of t and p-values from Tukey post-hoc analyses of two mixed linear
model analysis to determine how each treatments changed the percent cover of M.
gracilis and C. pulchella compared to the control……………………………………….85
cover…………...…………………………...………………………...…………………..86
xv
effects of substrate type, herbicide and seed addition type on M. gracilis and C. pulchella
percent cover in burned plots only from 2012 to 2013. …………...…………………….87
Table 26 Summary of F and p values of a mixed linear model comparing the effect of
year and P.f.-D7 rhizobacteria treatment on B. tectorum stem counts…………...……...88
Table 27 Summary of F and p values from a mixed linear model analysis comparing
substrate type, time of soil collection, and levels of available nitrogen………...……….89
Table 28 Summary of t and p value from a Tukey post-hoc analysis of a mixed linear
model analysis comparing total nitrogen availability of each sample session…………...90
Table 29 The basalt intermound treatment combinations of burning, herbicide, and
trampling………………………………………………………………………………..119
Table 30 Summary of F and p values from two mixed linear model analyses comparing
V. dubia stem counts changes and percent cover changes with treatment.……………..120
Table 31 Summary of t and p values from a Tukey post-hoc analysis of a mixed linear
model, comparing the effect of treatment litter cover change from 2012 to 2013……..121
V. dubia stem counts changes and percent cover changes with post fire treatments…...122
V. dubia stem counts and percent cover changes with herbicide applied after fire…….123
the change in native and invasive species richness with treatment……...……………..124
xvi
Table 35 Summary of t and p values from a Tukey post-hoc analysis of two mixed linear
model analyses comparing the change in native and invasive richness per 1 m2 after
treatment……...………………….……...………………….……...…………………...125
the change in native and invasive species richness with fire and post fire treatments…126
the change in native and invasive species richness with herbicide applied after fire…..127
the change in native and invasive species percent cover with treatment……………….128
Table 39 Summary of t and p values from a Tukey post-hoc analysis of two mixed linear
model analyses comparing the change in native and invasive percent cover per 1 m2 after
treatment………………………….……...………………….……...…………………..129
the change in native/invasive species cover with fire and post fire treatments...............130
Table 41 Summary of F and p values from a Tukey post-hoc analysis of a mixed linear
model analysis comparing the change in native species percent cover with fire and post
fire treatments………………….……...………………….……...……………………..131
the change in native and invasive species percent cover with herbicide applied after
fire………………………………………………………..…………………..…………132
Table 43 Comparing distance matrix output, analysis the effect of treatment on the
similarity values of biocrust community composition………………………………….133
xvii
Table 44 Summary of F and p values from three mixed linear model analyses comparing
the changes of short moss, tall moss, and squamulose lichen cover with treatment…...134
the changes of short moss, tall moss, and squamulose lichen cover with post fire
treatments………………………………………………………….……………………135
the changes of short moss, tall moss, and squamulose lichen cover with burning and
burn/herbicide treatments………………………………………….……………………136
xviii
Ventenata (Ventenata dubia) invasion in an eastern WA Mima mound prairie
Abstract
Invasive winter annual grasses (IWAGs), such as ventenata and downy brome, threaten
western United States ecosystems and reduce forage. While downy brome has been
extensively studied, ventenata is less known although its population appears to be rapidly
expanding into previously IWAG resistant habitat. In arid habitats, intact biological soil
crust (biocrust) can increase IWAG resistance. However, biocrust patches also exist in
semiarid habitats, such as the Inland Northwest, USA, where their effects on IWAGs are
less understood. Biocrusts’ effects on ventenata are unknown and there has been little
formal documentation of ventenata’s spread. Consequently, my goals were to: 1)
document changes in ventenata’s population size (and other IWAGs), 2) determine how
IWAG invasion relates to biocrust abundance and composition, and 3) determine habitat
characteristics related to IWAG susceptibility in a semiarid Mima mound prairie at
Turnbull National Wildlife Refuge (TNWR) in Cheney, WA. Mima mounds are soil
aggregations underlain by basalt bedrock or alluvial gravel interspersed by intermounds
with shallow soil and biocrust. I recorded changes in vascular plant species composition
and IWAG stem counts in 1 m2 plots in 2009, 2012, and 2013. Plots were randomly
stratified among basalt mounds, alluvial mounds, and basalt intermounds. Biocrust
composition and cover were recorded in 2012. From 2009 to 2013 ventenata abundance
increased 15-fold on mounds, while other IWAGs did not change. The highest cover of
ventenata and biocrust was on basalt intermounds, where other IWAGs were rare. On
intermounds, biocrust diversity and lichen cover were negatively related to ventenata
cover, but not moss cover. Before ventenata invasion, basalt intermounds were dominated
1
by biocrust with little vascular plant growth, despite the presence of downy brome and
other IWAGs. But it appears that ventenata has been able to enter this niche and could
potentially replace most biocrust species in the semiarid west.
2
Introduction
Invasive winter annual grasses (IWAGs) are a major threat to biodiversity in the
western United States; they replace native plants, limit biological soil crust (biocrust)
growth, and negatively affect wildlife, thus altering ecosystem structure and function
(D'Antonio and Vitousek 1992, Beckstead and Augspurger 2004, Davies and Sheley
2007). IWAGs can keep prairie ecosystems in an early successional stage by altering
factors like fire frequency and intensity (Kerns et al. 2006), water (White et al. 2006) and
soil nitrogen availability (James 2008a), and herbivory (Hill et al. 2005), thereby
promoting IWAG growth while suppressing native species less tolerant to the new
environmental conditions (Brown et al. 2008). For example, downy brome (Bromus
tectorum) is well known for forming vast monotypic stands that increase fire
frequency/intensity and reduce soil nitrogen levels (Link et al. 2006, Vasquez et al.
2008a). Other threatening IWAGs include ventenata (Ventenata dubia), field brome
(Bromus arvensis), hairy chess (Bromus commutatus), Australian brome (Bromus
arenarius), and medusa head (Taeniatherum caput-medusae), among others. The goals of
this study were to document changes in IWAG abundance, particularly ventenata, and
determine how IWAG invasion relates to biocrust distribution, soil depth and substrate
type in a semiarid prairie habitat.
Efforts to manage IWAG invasion in natural systems, particularly in arid
sagebrush steppe habitat, have focused in part on preventing IWAG establishment in the
first place, by maintaining healthy biocrust (Prasse and Bornkamm 2000, Root and
McCune 2011) and perennial bunchgrass communities (Beckstead and Augspurger
2004). Maintaining a native dominated system can slow IWAG invasion, particularly for
3
downy brome (Ponzetti et al. 2007, Vasquez et al. 2008a), but their effect on ventenata is
unknown. Intact biocrust communities, which are comprised of cyanobacteria, algae,
mosses, and lichens that grow on soil, are common in the semiarid/arid regions of the
western United States. Biocrust communities can fix nitrogen, prevent seedling
establishment, retain soil moisture, reduce wind and water erosion (Maestre et al. 2006,
Ponzetti et al. 2007), reduce wildfire risk (Belnap 2003, Link et al. 2006), and indicate
environmental conditions, such as pollution and disturbance intensity (Ponzetti et al.,
2007, Eldridge 2000). Biocrust community structure is highly influenced by vascular
plant presence, the intensity and frequency of disturbances, precipitation, temperature,
and soil quality (Ponzetti et al. 2007, Chamizo et al. 2011). Vascular plants alone can
have a significant effect on microhabitats and consequently the biocrust species present
(Belnap 2003, Root et al. 2011). For instance, severe shading from dense vascular plant
growth can reduce the abundance of lichen (Cornelissen et al. 2001) and bryophyte
communities (Ponzetti et al. 2007).
Numerous studies have focused on the interactions between IWAGs and biocrust
in arid sagebrush steppe habitats (Prasse and Bornkamm 2000, Belnap 2003, Ponzetti et
al. 2007, O’Bryan et al. 2009), however there have been relatively few studies in semiarid
regions (Root et al. 2011, Root and McCune 2011) that experience higher annual rainfall
than arid habitats (Schwinning et al. 2004). In the semiarid Channeled Scablands of
eastern Washington, which receive around 430 mm of rain per year, IWAGs are
abundant, even in areas with extensive biocrust cover. This is a concern because subtle
variations in rainfall and biocrust abundance can be driving factors in IWAG invasion
4
and thereby influence the management strategy needed. One IWAG that appears to be
rapidly increasing its population size and distribution is ventenata.
Ventenata is a winter annual grass with similarities to other infamous IWAGs,
such as downy brome and medusahead. Ventenata is native to Mediterranean Europe and
Africa, and prefers shallow rocky soils that flood in spring. Since it was first documented
in 1956, ventenata has spread from Kootenai County, ID (Northam and Callihan 1994) to
most western states and several northeastern states of the United States and southern
Canada (USDA 2008). Like other IWAGs, ventenata senesces in early summer,
providing no foraging or grazing benefits beyond this time. Although its presence has
been noted for decades (Northam et al. 1993, Gray and Lichthardt 2003, James 2008a,
Vasquez et al. 2008b), to date very little has been published on its phenology,
morphology, or effects on neighboring species and wildlife.
Along with downy brome, ventenata has been compared to the IWAG,
medusahead, which initially established in the late 1800’s in heavy clay soils where few
native species could grow (Young and Evans 1970). Since then medusahead population
has expanded and has proven capable of outcompeting the aggressive downy brome when
soil moisture is abundant (Young 1992). Along with their shared ability to thrive in
seemingly undesirable areas, another suggested similarity between ventenata and
medusahead is silica content (Prather and Steele 2009); medusahead has a silica content
that is about 10% higher than other IWAGs (Young 1992), which contributes to its slow
decomposition and deterrence to grazers (Young and Mangold 2008). If ventenata also
has higher silica content, it could explain why it provides little forage for wildlife and
5
livestock, results in quick litter accumulation, and effectively perpetuates its own growth
by altering the trophic dynamics of the ecosystem.
One area in particular that is experiencing a rapid increase of ventenata is the
Channeled Scablands of eastern Washington (personal observation). The Channeled
Scablands were formed by the Lake Missoula floods about 20,000 years ago, creating
four main tracts, the largest of which is the Cheney-Palouse River tract, about 20 miles
southwest of Spokane, Washington (Bretz 1928). This area is generally comprised of
shallow soils intermixed with patches of deeper soil underlain by basalt bedrock. Typical
Channeled Scabland features include Mima mounds (Berg 1990), also referred to as
pimple mounds, biscuit swale formations, and hog wallows (Washburn 1988). Mima
mounds are aggregations of fine loess up to 2 m tall and 30 m in diameter surrounded by
intermound areas of shallow soil. Soil depths of both the mounds and intermounds are
strongly influenced by the underlying substrate (Del Moral and Deardorff 1976), which
in eastern Washington tends to be either basalt bedrock or alluvial gravels (Bretz 1928).
Mounds underlain by basalt consist of fine loess on basalt bedrock with shallow rocky
intermound areas that often form vernal wetlands or biocrust. Mounds underlain by
alluvial are comprised of deeper re-deposited sediments and fine loess over basalt
bedrock. These mounds are often smaller than basalt underlain mounds due to deeper
intermound soils (Washburn 1988). The variations in soil depth and substrate allow for a
highly diverse native prairie flora intermixed with biocrust communities in these semiarid
Mima mound habitats. In this article, I refer to mounds and intermounds underlain by
basalt as basalt mounds and basalt intermounds and mounds and intermounds underlain
by alluvial gravels over basalt bedrock as alluvial mounds and alluvial intermounds.
6
In many places, Channeled Scablands habitat has been altered by grazing,
agriculture, or development, but some of the best remaining examples can be found at
Turnbull National Wildlife Refuge (TNWR; Figure 1), 5 miles south of Cheney,
Washington. Grazing was banned on the refuge in 1993, and it has since experienced
minimal human disturbance, creating an ideal location to study natural plant
communities. The area surrounding TNWR includes farmland, rangeland, and remnant
Palouse Prairie; all, except agriculture fields, contain Mima mound prairies with the
native flora and biocrust.
Ventenata was first reported at TNWR in 1990 and despite refuge managers’
efforts to eliminate human disturbance there has been a noticeable increase in ventenata
abundance at TNWR and throughout semiarid regions of the Inland Northwest, even on
biocrust dominated areas that had previously remained impervious to invasion by
IWAGS. The conversion of these native communities and biocrust to ventenata
dominated monocultures will likely negatively impact plant and wildlife biodiversity at
the refuge and in the surrounding areas (James 2008a, Vasquez et al. 2008a).
Understanding how, where and why ventenata is increasing, and why biocrust is no
longer offering any resistance would enable land managers at TNWR and areas
throughout the semiarid regions of the west to better manage and control this invasion.
Consequently, the three goals of this study were to 1) document changes in IWAG
abundance, particularly ventenata, on the Mima mound prairie at TNWR, 2) identify the
characteristics of sites that are most susceptible to invasion, and 3) determine how IWAG
abundance relates to biocrust abundance and distribution. I predicted an increase in
ventenata abundance between 2009 to 2013, while the abundance of other IWAGs
7
remained relatively consistent. I also predicted that ventenata and biocrust abundance
would be higher on the basalt intermounds than on mounds, but biocrust abundance will
be negatively correlated with ventenata abundance.
Materials and Methods
Study Site
TNWR covers approximately 18,217 acres encompassing several distinct habitats
including Mima mound prairie, ponderosa pine forest, wetlands, and riparian areas and is
surrounded by rangeland and agricultural fields. The Mima mound prairie remnants cover
about 1,270 acres and include both alluvial gravel and basalt substrates. The climate is
semiarid with an average annual rainfall of 432 mm (17 in) and an average annual
temperature of 48° F. My study site was in a 380-acre portion of the Mima mound prairie
located in the Stubble Field Tract and Public Use Area (47.40N, -117.52W; Figure 1).
Study Design
To assess changes in IWAG abundance from 2009 to 2013, 22 1-m2 plots were
established in 2009 (6 on alluvial mounds, and 16 on basalt mounds) and resurveyed in
2012 and 2013. For each plot, I recorded underlying substrate type (alluvial or basalt),
vascular plant species composition and percent cover, IWAG stem count for a 0.04 m2
area on the southeast corner of the plot, and soil depth (at 2 corners of the plot). Vascular
plants were identified to species using the Flora of the Pacific Northwest (Hitchcock and
Cronquist 1973) and nomenclature was updated where necessary using the Integrated
Taxonomic Information System (www.ITIS.gov).
To better determine how IWAG abundance relates to plant community
composition and biocrust distribution, I added an additional 177, 1 m2 plots in 2012 for a
8
total of 199 plots, this time including intermound areas (57 on basalt mounds, 90 on
basalt intermounds and 52 on alluvial mounds). Alluvial intermounds were not surveyed.
Along with the previously described vascular plant assessment, I also recorded biocrust
composition and percent cover, as well as bare soil and litter percent cover. Biocrust
composition classes included six lichen morphology categories and two moss
morphology categories. Lichen morphology categories were crustose, areolate,
gelatinous, squamulose, foliose, and fruticose. Each lichen morphology category was
further assessed by color (white, light and dark green, red, black, light and dark brown,
grey, orange, and yellow). The two moss morphology categories were characterized as
tall or short (Table 5; Rosentreter et al. 2007).
The survey period was June and July for all three years. To control for
phenological affects over the survey period (during which time plants were gradually
desiccating), plots in different substrate types and locations were randomly sampled over
the survey duration.
Data Analysis
Data analyses were performed only for IWAG species (including ventenata,
downy brome, and hairy chess) present in greater than 45% of the plots. Other IWAG
species were either very infrequent or had very low cover. To test whether stem counts of
these species increased from 2009 to 2013 across the different substrates, I performed
repeated measures analyses (one for each species) using a spatial power law covariance
structure (SP(POW); PROC MIXED, SAS 9.3 software; SAS Institute Inc. 2011). This
covariance structure was selected because the sampling years were unequally spaced
(Littell et al., 2002). To test whether ventenata increased more than the other two species
9
over time, I used a general linear model (PROC GLM in SAS) comparing change in stem
count from 2009-2013 among the three species. To test whether percent cover of
ventenata, downy brome, and biocrust varied across the three plot types (basalt mound,
alluvial mound, basalt intermound) based on the 199 plots from 2012 in which biocrust
was assessed, I used a general linear model (PROC GLM in SAS). I present percent
cover data for the plant species because it was assessed at the same scale (1 m2) as
biocrust; however stem count data, which was assessed at a smaller scale, yielded the
same results.
Data were tested for normality and homoscedasticity and transformed as
necessary using Box-Cox transformations for stem count data and arcsine-square root
transformations for percent cover variables.
I assessed patterns in vascular plant species composition related to substrate, soil
depth, and percent cover of litter, bare soil, rock, total biocrust, ventenata, total moss, and
total lichen with nonmetric multidimensional scaling (NMS) ordination using PC-Ord
software (MjM Software Design 2006). A second NMS ordination was used to analyze
the patterns in biocrust morphology composition in the basalt intermound areas;
ordination axes were then related to the variables stated above (except substrate), as well
as biocrust diversity (calculated using the Shannon diversity index) and total cover of tall
moss, short moss, and squamulose white lichen (the three most abundant biocrust
morphology categories). Other than ventenata, IWAGs were not compared to biocrust
composition because of their extremely low abundance in the basalt intermounds. For
both NMS ordinations, I used a Relative Sorensen distance measurement, Varimax
10
rotation with 50 runs on the real data, 15 iterations to evaluate stability, a stability criteria
of 0.00001, an initial step length of 0.2 and a maximum of 250 iterations.
To further analyze the relationship between ventenata, biocrust diversity, and the
three most prominent biocrust morphological classes (tall moss, short moss, and
squamulose white lichen), I used a series of regressions (PROC REG in SAS) to test the
effects of biocrust diversity and the percent cover of each biocrust morphology type on
ventenata percent cover. All data, except biocrust diversity, were arcsine square root
transformed to improve normality.
Results and Discussion
Despite its broad geographic range, the invasive tendencies of ventenata have
received relatively little attention, likely because its abundance has remained low
compared to other IWAGs. However, this is no longer the case at TNWR, where from
2009 to 2013, ventenata stem count increased about 15-fold from 3.25 to 47.4 stems per
0.04 m2 on alluvial and basalt mound plots, while downy brome and hairy chess, the other
most common IWAGs, stem counts, remained consistent (Figure 2; Table 1-2). Final
lambda values for Box-Cox transformations were 0.25 for both ventenata and downy
brome and 0 for hairy chess. The transformations improved distributions, however due to
the presence of large numbers of zero values, some of the data were still not normally
distributed. Ventenata was present across all plot types in 2012, however it was almost
five times more abundant on basalt intermound plots, whereas both downy brome and
hairy chess were nearly absent (Figure 3; Table 3). Ventenata makes up on average
95.2% of all IWAG cover in basalt intermound plots. This is of particular concern
because it suggests that ventenata has managed to invade a habitat that was once
11
relatively free of IWAGs and is where biocrust tends to be most abundant (Figure 3;
Table 3).
Basalt intermound areas, with their shallow soils and extremes of wet and dry
conditions, provide a unique habitat that is well suited for biocrust and highly adapted
native plants. Biocrust was 12 times more abundant on basalt intermounds (Table 3;
Figure 4) than both basalt and alluvial mounds with a mean percent cover of 73.05% per
plot, with 49.68% moss cover and 7.99% lichen cover. In 2012, forty-four vascular plant
species were identified in basalt intermound plots with an average percent cover (totaled
over all species) of 81.37% per plot, of which ventenata comprised 66.66%. It is not clear
how ventenata presence is affecting other IWAGs and biocrust, but it appears to have
been able to fill a niche that was not filled by other IWAGs in the past.
Species composition in basalt intermound plots was relatively consistent
compared to basalt and alluvial mound plots as revealed by NMS ordination. Species
composition, total biocrust cover (axis 1 r2=0.336; axis 2 r2=0.419; Figure 4), and
ventenata cover (axis 1 r2=0.479; axis 2 r2=0.597; Figure 4) were positively related on the
basalt intermounds. From this, I infer that biocrust and ventenata have similar habitat
preferences of shallow soils and the capacity to withstand the moisture extremes of basalt
intermound areas. Species composition on basalt and alluvial mound plots was
overlapping, and on both was strongly correlated with soil depth (axis 1 r2=0.360; axis 2
r2=0.585; 2-Dimensional; final stress=19.581; final instability=0.00179; Figure 4). In
deeper soils, prickly lettuce (Lactuca serriola; axis 2 r2=0.221), common yarrow
(Achillea millefolium; axis 1 r2=0.239), and downy brome (axis 2 r2=0.193) tended to
dominate, while there was more ventenata (axis 2 r2=0.597) in shallow soils. In general it
12
appeared that plant communities on basalt mounds tended to be less uniform than on the
alluvial mounds, but species composition on both mound types is different than basalt
intermounds species composition, which was dominated by ventenata.
Basalt intermound biocrust community composition (n=90; biocrust groups=38)
was correlated with total cover of ventenata (axis 1 r2=0.268), litter (axis 1 r2=0.381), tall
moss (axis 1 r2=0.493), short moss (axis 2 r2=0.222), squamulose white lichen (axis 1
r2=0.605), and biocrust diversity (axis 1 r2=0.317) per the NMS biocrust ordination (2Dimensional; final stress=13.430; final instability=0.00003; Figure 5). Both litter cover
and ventenata cover have a strong and similar relationship with species composition,
implying that tall moss is currently able to grow with ventenata. However, I did not find a
positive relationship between tall moss and ventenata when compared independently with
regression (Figure 6a; Table 4). Currently, the tall moss is most likely benefiting from the
shade and litter produced by living ventenata, which may serve to increase soil moisture
retention, but it is unclear how tall moss abundance changed after ventenata invaded or
how the bryophyte community will respond as ventenata abundance increases. I suspect
that increasing litter depth will overwhelm the bryophyte community causing it to decline
like other biocrust species (Ponzetti et al. 2007).
Conversely short moss cover shares a strong relationship with ordination axis 2,
compared to ventenata’s relationship with axis 1, suggesting that short moss and
ventenata are driven by different factors (Figure 5). When compared independently with
regression, there is no relationship between short moss cover and ventenata cover (Figure
6b; Table 4). The plots farthest from those with an association with high ventenata cover,
share a relationship with high biocrust diversity and squamulose white lichen cover
13
(Figure 5). From this I can conclude that with high ventenata cover there is lower biocrust
diversity, as well as low squamulose white lichen cover (the most abundant lichen
morphology type). When analyzed independently with regression, total lichen cover,
squamulose white cover, and biocrust diversity are all negatively correlated with
ventenata cover (Figure 6c, d, and e; Table 4).
The current negative relationship between ventenata and most of the biocrust
community is alarming; anecdotal observations from 2009 suggest the basalt intermound
communities then were primarily biocrust despite the presence of several aggressive and
highly competitive IWAGs, (i.e. downy brome) on the adjacent mounds. If ventenata
cover continues to increase, it may lead to loss of the biocrust community, and alter the
ecosystem as a whole through changes to nutrient, water, and space availability (Rimer
and Evans 2006), soil composition (DiTomaso 2000, Brown et al. 2008), increased
biomass and litter (Facelli and Pickett 1991). Many native vascular and non-vascular
plant species are susceptible to the altered environmental factors, which may explain the
reduction in lichens and native plant species where high ventenata cover is predominant,
which is consistent with other IWAG studies (Young et al. 1987, Ridenour and Callaway
2001, Cox and Anderson 2004, Maestre et al. 2005, Martínez et al. 2006). In addition to
the effects of direct competition between living ventenata and biocrust, I suspect that the
litter produced by high ventenata cover is drastically changing microhabitats by altering
soil temperature and shade (Facelli and Pickett 1991), which may result in further
declines of biocrust and native plants.
It is unclear why ventenata has been so successful at TNWR, or why its
population exploded when it did, especially given the presence of intact biocrust
14
communities, which typically slow establishment of other IWAG species (Prasse and
Bornkamm 2000, Ponzetti et al. 2007, Root and McCune 2011). Possible factors that
could have affected the relationship between ventenata and intact biocrust include: 1)
ventenata has a relatively high silica content (Prather and Steele 2009), thereby reducing
forage and decomposition (Massey and Hartley 2006), giving it a competitive advantage
over species favored by herbivores, 2) with low decomposition rates the increased litter
layer may increase shade, thereby reducing biocrust abundance and diversity, by
increasing soil moisture and temperature, potentially favoring its own growth (Evans and
Young 1970; Facelli and Pickett 1991), 3) biocrust may not be able to compete against
ventenata in particular due to an unknown allelopathy (Ridenour and Callaway 2001;
Levine et al., 2003), 4) the higher rainfall levels of semiarid regions may render biocrust
less effective at inhibiting invasion (Belnap et al., 2001), 5) other IWAG species have
larger root systems (e.g. 91 cm for downy brome; Hironaka 1961) that the shallow basalt
intermound soils (0-19.5 cm) cannot support. Ventenata has smaller root length and
biomass (James 2008a), and is not limited by low water availability when it is most active
due to the dependable vernal flooding of the basalt intermounds, or 6) some combination
of these factors. Future research should seek to determine why ventenata populations
have expanded so quickly, and evaluate the relationship between ventenata and biocrust
species individually and as a community.
IWAG growth has been increasing throughout western United States for decades
but the rate of ventenata increase at TNWR, where human impact is minimal, is alarming
and will not be contained to refuge boundaries; surrounding areas with extensive human
disturbance may see an even larger expansion of ventenata, considering it appears that
15
ventenata has found its preferred habitat on basalt intermounds: shallow, rocky, spring
inundated soils, which is a common habitat in the semiarid west. Potential areas impacted
by the spread of ventenata are economically important (farm and rangeland) and rare
(Palouse Prairie and Channeled Scablands). If ventenata expands into these areas their
economic and ecological value will continue to diminish.
16
Chapter 1 Figures
Figure 1 Location of study area at Turnbull National Wildlife Refuge (TNWR) in
Cheney, WA and its location within Washington State.
17
80
Mean Stem Count per 0.04 m
2
Hairy Chess
Downy Brome
Ventenata
60
40
20
0
AL
BA
2009
AL
BA
2012
AL
BA
2013
Figure 2 Mean stem counts of IWAGs found in a 0.04 m2 corner of alluvial and basalt
mound plots from 2009, 2012, and 2013. Statistical inferences are based on Box-Cox
transformed data. Error bars represent + 1 standard error.
18
Mean Percent Cover per 1 m
2
100
80
Biocrust
Ventenata
Downy Brome
60
40
20
0
AL
BA
IN
Plot Type
Figure 3 Mean percent cover of ventenata, downy brome, and biocrust in all plot types in
2012: alluvial mound plots (AL), basalt mound plots (BA), and basalt intermound plots
(IN). All statistical inferences are based on arcsine-square root transformed data. Error
bars represent + 1 standard error. Repeated letters across bars indicate no significant
difference.
19
Figure 4 NMS, joint plot Ordination of the 2012 plant communities. Each triangle
represents an individual plot; distance between triangles represents similarity of species
composition among plots. The ordination axes were correlated to substrate, soil depth,
and percent cover of litter, bare soil, rock, total biocrust, ventenata, total moss, and total
lichen. Those with Pearson correlations greater than 0.2 are shown as vectors; the length
of the vector lines indicates the strength of the correlation. Alluvial mound plots (AL),
basalt mound plots (BA), and basalt intermound plots (IN) are indicated by light gray
circles, gray squares, and black triangles respectively.
20
Figure 5 NMS, joint plot Ordination of the 2012 biocrust communities on basalt
intermound plots. The ordination scores were then correlated to soil depth and percent
cover of litter, bare soil, rock, total biocrust, ventenata, total moss, total lichen, tall moss,
short moss and squamulose white lichen. Those with Pearson correlations greater than
0.2 are shown as vectors; the length of the vector lines indicates the strength of the
correlation.
21
2
2
R =0.083
A
R =0.0280
B
2
R =0.3314
2
R =0.4698
2
R =0.4698
C
E
C
D
2
R =0.1378
E
Figure 6 Relationship of (a) tall moss, (b) short moss, (c) total lichen cover, (d)
squamulose white lichen percent cover, and (e) biocrust diversity to ventenata percent
cover over 1 m2. All data, except biocrust diversity, were arcsine square root transformed.
22
Chapter 1 Tables
Table 1 Summary of F-values from the repeated measures analyses to determine how
ventenata, downy brome and hairy chess respective stem counts varied with year,
substrate, and interaction effects of year and substrate. Significant values are indicated in
bold.
Year
Ventenata
Downy
Brome
Hairy
Chess
F (numDF, denDF)
10.82(2,40)
Substrate
Year*Substrate
P
F (numDF, denDF)
0.69(1,23.1)
0.0002
P
0.4161
F (numDF, denDF)
3.13(2,40)
P
0.0544
0.29(2,24)
0.7524
0.43(1,20)
0.5189
1.25(2,24)
0.3048
1.49(2,28)
0.2429
0.43(1,20)
0.5196
1.35(2,28)
0.2748
23
Table 2 Summary of F-values from a general linear model comparing change in
ventenata, downy brome and hairy chess stem count from 2009-2013 and how they
varied among substrate type. Significant values are indicated in bold.
Stem Count
Change
Species
F (DF)
P
4.64(2)
0.0133
Substrate
F (DF)
P
1.99(1)
0.1634
24
Species*Substrate
F (DF)
P
1.85(2)
0.1655
Table 3 Summary of F-values from a general linear model analyzing ventenata, downy
brome and hairy chess 2012 percent cover and how they varied among substrate type.
Significant values are indicated in bold.
Species
Ventenata
Downy Brome
Biocrust
F (DF)
125.46(2)
40.86(2)
223.43(2)
25
P
<0.0001
<0.0001
<0.0001
Table 4 Regression of tall moss, short moss, total lichen, and squamulose white lichen
percent cover and biocrust diversity compared to ventenata cover in 1 m2 plots.
Independent
Variable
Tall Moss
Short Moss
Total Lichen
Squamulose
White
Biocrust
Diversity
DF
Dependent Variable=Ventenata Cover
SS
F
P
R2
1
1
1
1
0.0023
0.0017
5.55
4.21
3.51
2.53
77.10
43.62
0.0644
0.1152
<0.0001
<0.0001
3.83%
2.80%
46.98%
33.14%
1
1.75
14.06
0.0003
13.78%
26
Table 5 All vascular plant species, lichen morphology groups, and moss morphology
groups recorded in 2009, 2012, and 2013 and the plot types (alluvial mound, basalt
mound, and basalt intermound) within which each was found.
Vascular
Species
Achillea millefolium
Acmispon americanus
Acmispon denticulatus
Agoseris grandiflora
Agoseris heterophylla
Allium columbianum
Amsinkia menzesii
Apera interrupta
Balsamorhiza sagittata
Blepharipappus scaber
Bromus arenarius
Bromus arvensis
Bromus commutatus
Bromus secalinus
Bromus tectorum
Buglossoides arvensis
Camassia quamash
Centaurea solstitialis
Cirsium sp
Clarkia pulchella
Claytonia exigua
Collinsia parviflora
Collomia linearis
Convolvulus arvensis
Danthonia intermedia
Descurainia incisa
Delphinium nuttallianum
Draba verna
Elymus elymoides
Epilobium brachycarpum
Eriogonum heracleoides
Festuca idahoensis
Frittillaria pudica
Gaillardia aristata
Galium aparine
Alluvial
Mound
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Basalt
Mound
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Basalt
Intermound
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
Geranium viscosissimum
Geum triflorum
Holosteum umbellatum
Hypericum perforatum
Idahoa scapigera
Lactuca pulchella
Lactuca serriola
Lagophylla ramosissima
Lewisia rediviva
Linaria dalmatica
Linum usitatissimum
Lomatium ambiguum
Lomatium macrocarpum
Lomatium triternatum
Lupinus sericeus
Madia gracilis
Microsteris gracilis
Myosotis stricta
Perideridia gairdneri
Poa bulbosa
Poa Protensis
Poa secunda
Polemonium micranthum
Polygonum douglasii
Polygonum polygaloides
Potentilla argentea
Pseudoroegneria spicata
Rosa woodsii
Sedum lanceolatum
Sisymbrium atlissimum
Tragopogon dubius
Turritis glabra
Ventenata dubia
Veronica peregrina
Vicia cracca
Lichen Morphology
Crustose
Areolate
Gelatinous
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
X
28
X
X
X
X
X
X
X
Squamulose
Foliose
Fruticose
Moss Morphology
Tall
Short
X
X
29
X
X
X
X
X
X
X
X
X
Prescribed fire, herbicide, and annual seed addition used to increase community
resistance to invasion
Abstract
The threat of invasive winter annual grasses (IWAGs) in the west has forced land
managers to use aggressive restoration tools (i.e. prescribed fire, herbicide, and
biocontrols), but they can both help or harm native species. To further increase native
cover, native annual seed can be dispersed instead of native perennial seed to assist native
succession to a later stage. My objective was to determine the most effective approach to
restore the native plant communities to the Mima mound prairie at Turnbull National
Wildlife Refuge (TNWR; see description in chapter 1). In 2012, 107 1 m2 plots on both
basalt and alluvial mounds, were treated with prescribed fire (October/November),
Outrider® herbicide (late November), and/or native annual seed alone or mixed with
native perennial seed (assisted succession; February) in a multifactorial experimental
design, with 8 treatment combinations per substrate with 5 replicates each, except the
controls and burn only which had 10 plots each. All plots were surveyed in summer 2012
and 2013. In 2009, 14 P.f.-D7 rhizobacteria plots (7 inoculated, 7 control) and 29 native
perennial seed addition plots (7 treated, 22 control) were surveyed and treated, then
resurveyed in 2010 and 2012. For all plots I recorded vascular plant species composition
and cover and IWAG stem counts. I analyzed the effect of treatments and the interaction
of restoration tools on IWAG abundance, invasive and native species richness/cover, and
cover of the seeded species using mixed linear models. While herbicide reduced IWAGs,
invasive species cover/richness, and native species richness, fire increased native cover.
Of the seed species, fire increased Madia gracilis cover, even if herbicide was applied
30
after, but neither native perennial seed addition nor the rhizobacteria were successful.
Long-term monitoring must be done to fully understand the effects of these treatments.
31
Introduction
Land managers have long understood the biological and economic threats posed
by invasive plants, because they can change ecosystem function, reduce biodiversity, and
limit forage for wildlife and livestock (D'Antonio and Vitousek 1992, Beckstead and
Augspurger 2004). However, many invasive species management plans focus on
controlling invasive species without considering why they were able to establish and
thrive in the first place. When a system is healthy (i.e., native vegetation and the natural
disturbance regime to which the vegetation is adapted), it often has self-sustaining factors
that prevent invasive plant establishment. For example, prairies with abundant
bunchgrass populations have been associated with lower invasive winter annual grass
(IWAG) abundance (Beckstead and Augspurger 2004, Ponzetti et al. 2007) and abundant
biological soil crust (biocrust; which are the mosses, lichens, and algae that grow on soil)
that acts as a natural barrier to prevent seedling establishment (Maestre et al. 2006).
However, when the natural defenses of the system are disturbed, or the natural
disturbance regime is altered, invasive plants can establish and potentially form
monotypic stands. Thus, a major goal of restoration should not only be to control target
invasive species but also to increase community resistance to invasion, resulting in a selfsustaining system.
In the western United States, IWAGs are a major threat to prairies and rangelands,
as they can establish with or without disturbance (DiTomaso 2000), and can quickly
replace native species (D'Antonio and Vitousek 1992, Davies and Sheley 2007) by
changing the environmental conditions in which native species have evolved. When
IWAGs are prevalent they can alter factors like fire frequency and intensity (Kerns et al.
32
2006), water availability (White et al. 2006), soil nitrogen availability (James 2008a), and
herbivory (Hill et al. 2005). By altering these factors, IWAGs can keep the prairie in a
stage that promotes their own growth while suppressing native species less tolerant to the
new environmental conditions (Brown et al. 2008). A key to successful IWAG invasion
is their growth habit as they are active when most native species are not; IWAGs
germinate in fall, grow roots throughout winter, and have prolific shoot growth in early
spring (Wainwright et al. 2011). Once they set seed in early summer, IWAGs senesce,
and are no longer palatable to grazers (DiTomaso 2000). One of the more common
IWAGs is Bromus tectorum (downy brome), which as of 2000 it infested over 40 million
ha of rangeland and agriculture fields and though one of the more common IWAGs, B.
tectorum is not the only problematic species; other IWAGs include Ventenata dubia
(ventenata), field brome (Bromus arvensis), hairy chess (Bromus commutatus), Australian
brome (Bromus arenarius), and Taeniatherum caput-medusae (medusahead), among
others. Due to their profound effects on western habitats, IWAG species have become the
targets of prairie restoration management.
Through human disturbances, intense grazing, and frequent wildfires, IWAG
population sizes have increased, reducing native perennial bunchgrass/biocrust systems.
Native perennial bunchgrass species like Pseudoroegneria spicata (Bluebunch
wheatgrass), Poa secunda (Sandberg bluegrass), and Festuca idahoensis (Idaho fescue),
provide quality habitat for other native plant species, biocrust, animals, and insects (Klaas
et al. 1998). When a system is dominated by native bunchgrasses, often with space
between individuals occupied by biocrust (Root et al. 2011), IWAGs have difficulty
establishing due to lack of space and reduced resource availability (Corbin and D'Antonio
33
2004a, Beckstead and Augspurger 2004, Ponzetti et al. 2007, Vasquez et al. 2008a, Root
et al. 2011). However, native bunchgrasses, which are highly palatable to wildlife and
livestock (USDA NRCS Aberdeen Plant Materials Center 2011), do not have much
seedling vigor (especially compared to IWAGs) or recover quickly from grazing
(DiTomaso 2000). When disturbances result in competition between IWAG and
bunchgrass seedlings, IWAGs are generally able to outcompete bunchgrass seedlings
through species priority (Ray-Mukherjee et al. 2011, Corbin and D'Antonio 2004a,
Keeley and McGinnis 2007, James et al. 2011) meaning IWAG seedlings use space and
nutrients before native seedlings establish, enabling them to produce more seeds sooner
and potentially usurping the native population (James et al. 2011). In semiarid habitats
soil nitrogen can be limited so if IWAGs can utilize it first, bunchgrass seedlings cannot
capture enough nitrogen to successfully establish (McLendon and Redente 1992, James
2008a, James et al. 2011, Blumenthal et al. 2003). If soil nitrogen availability increases
before or during fall IWAG germination, then IWAG abundance could increase
dramatically, further increasing it competitive advantage over native seedlings.
Promoting bunchgrass seedling establishment and early growth is an essential
management technique for prairie restoration (Antos and Halpern 1997).
There are many management approaches that reduce IWAG population density
and size (Kennedy et al. 2001, Frost and Launchbaugh 2003, Cox and Anderson 2004,
Keeley 2006, DiTomaso et al. 2009, Nyamai et al. 2011), but most aim exclusively on
controlling the target invasive and not on restoring damaged ecological processes. One
example that aims to restore ecological processes damaged by invasive species is the
ecologically based invasive plant management approach (EBIPM; James et al. 2010).
34
EBIPM is a holistic method that requires land managers to develop a long-term, multistep plan that removes target invasives while promoting natural resistance to invasion
with the use of multiple management techniques (James et al. 2010). For example, by
applying herbicide when IWAGs are susceptible followed by dispersing native seed, all
while preventing further disturbance and damage to the system’s natural defenses, should
reduce IWAG establishment, increase the native seed bank, and increase community
resistance to future invasion.
All active invasive plant management plans require the use of at least one
management technique to physically impede growth or reproduction of the target
invasive species. These techniques can include biocontrols (Kremer and Kennedy 1996),
fire (Pollak and Kan 1998), and herbicide (Mittelhauser et al. 2011). One of the most
common IWAG control methods is herbicide, due to its relatively easy application and
effectiveness at eliminating plant growth. The sulfosulfuron herbicide, Outrider®, which
can control annual and perennial weeds (DPR 2008), but can have minimal effect on
established native bunchgrasses like P. spicata and Elymus multisetus (Sand Hollow Big
Squirreltail; Stevens 2010, Monaco and Creech 2004).
In fire-adapted systems, such as many semiarid habitats of the west, prescribed
fires can also control IWAGs, however this technique entails a delicate balance between
too much and not enough, as many fire-adapted species do best under an intermediate
disturbance regime (Grime 1973, Miller et al. 2012). When fires occur too frequently or
at too high an intensity some IWAG populations will increase, resulting in an overall
decrease in biodiversity (Keeley et al. 2005). Yet fire suppression can also lead to an
increase in IWAG abundance because as time goes on more IWAGs can establish but the
35
native perennial seedlings are not able to replace lost individuals (Keeley 2001). As
IWAG abundance increases, so does the litter layer, providing fuel for wildfires (Collins
et al. 2008) and facilitating IWAG establishment by increasing winter soil moisture and
temperature, yet in spring and summer the insulation increases soil temperature but
decrease moisture availability (Facelli and Pickett 1991). So while winter extremes are
ameliorated, summer conditions can be more extreme, potentially impacting native
species. When fire has been suppressed past the intermediate stage, prescribed burns are
an option (Miller et al. 2012) if the fires target IWAG germination, which is often when
biocrust is relatively moist (Belnap and Eldridge 2001) and native species are not actively
growing (James et al. 2011).
When the problematic invasive is primarily one species, a host specific biocontrol
can be used. The rhizobacterium, Pseudomonas fluorescens D7 (P.f.-D7), which targets
B. tectorum (Kennedy et al. 1991, Kremer and Kennedy 1996), is a non-parasitic
bacterium that colonizes the rhizosphere of B. tectorum, potentially suppressing early
growth (Kremer et al. 1990). P.f.-D7 releases a phytotoxin (Tranel et al. 1993) that
inhibits root growth and seed germination (Kennedy et al. 2001). However, results may
not be seen until up to five years after application. Use of rhizobacteria is still in the
experimental stage and is not yet widely available, nor are the long-term results of this
approach well researched.
A common management technique for restoring native communities to vast areas
is to disperse seed of the desired late successional stage species (usually a perennial), but
this does not always yield the desired establishment rates when IWAG abundance is high
(Cox and Anderson 2004). A relatively new management approach called assisted
36
succession, attempts to increase native perennial abundance by using the system’s natural
successional processes to force a more balanced competition between IWAGs and native
annuals, instead of native perennials. Eventually, when the native annual species are
dominant, native perennial species should be able to establish and outcompete their
natural native predecessors. As the native population increases, IWAG abundance should
decrease with the reduction of space and resources (Eldridge 2000, Belnap 2003, Corbin
and D'Antonio 2004, Cox and Anderson 2004, Link et al. 2006, Maestre et al. 2006,
Ponzetti et al. 2007, Brown et al. 2008, Chamizo et al. 2011). Assisted succession may be
a viable approach in areas where IWAGs are promoting an early successional stage and
tilling or drilling of native perennial seeds are not options.
One large region that has been profoundly altered by IWAGs is the Channeled
Scablands of eastern Washington. This unique geological region was formed by
successive rounds of intense flooding from periodic breaks in the ice dam that held back
glacial Lake Missoula about 20,000 years ago (Bretz 1928). The flooding created four
main tracts, the greatest of which is the Cheney-Palouse River tract (75 miles long and
about 20 miles wide), about 20 miles southwest of Spokane, Washington (Bretz 1928).
Much of this land has been converted to rangeland, agriculture fields, or development,
but intact remnants often contain Mima mounds (Berg 1990), also referred to as pimple
mounds, prairie mounds, and hog wallows (Washburn 1988). They have been described
from regions of North American, Africa, and South America (Cox 1990). The mounds
are aggregations of fine loess about 2 m tall and 30 m in diameter, but soil depth of
mounds and intermounds is strongly influenced by the underlying substrate type (Del
Moral and Deardorff 1976). In eastern Washington, Mima mounds are underlain by one
37
of two substrates: basalt bedrock or alluvial gravel (Bretz 1928, Bryant et al. 2013).
Basalt underlain mounds (basalt mounds) consist of fine loess on basalt bedrock with
shallow rocky intermounds that often flood in spring, forming ephemeral wetlands. The
alluvial underlain mounds (alluvial mounds) are redeposited sediments and fine loess on
the basalt bedrock (Washburn 1988). Alluvial mounds are often smaller (in height and
diameter) than basalt mounds due to deeper intermound soils.
Native species common to the Channeled Scablands include the annuals
Amsinckia menziesii (tarweed fiddleneck), Clarkia pulchella (pinkfairies) and Madia
gracilis (grassy tarweed), and the perennials Eriogonum heracleoides (parsnipflower
buckwheat), F. idahoensis, P. secunda and P. spicata (Bryant et al. 2013, Reynecke,
2012). Over the last few decades these native populations across the west have been
greatly reduced by IWAGs (D'Antonio and Vitousek 1992, DiTomaso 2000b), including
V. dubia, B. tectorum, B. arvensis (field brome), B. commutatus (bald brome), and T.
caput-medusae, as well as invasive perennial species including Poa bulbosa (bulbous
bluegrass), and Vicia cracca (bird vetch).
Objectives
The overall objective of my study was to experimentally determine the most
effective approach to sustainably restoring native Mima mound prairie plant communities
by restoring a more natural pattern of succession. I did this by testing combinations of
techniques that aim to: 1) reduce invasives, 2) assist succession through annual seed
addition, and 3) increase community resistance to invasion. Specifically, I wanted to
determine: 1) whether burning, herbicide, or a combination of the two was most effective
at reducing invasives and promoting natives, and 2) whether annual seed alone or in
38
combination with perennial seed effectively limits invasives and promotes native growth.
I addressed these questions in a three part study.
For the first part of my study, I experimentally tested the effects of fire, herbicide,
and native annual and perennial seed addition on species composition on Mima mounds
underlain by two different geologic substrates (basalt and alluvial). I predicted that fire
and herbicide together would be most effective at reducing IWAG abundance.
Experiment 1 also allowed me to assess the benefits of assisted succession by comparing
establishment success of native annual to perennial seed. I predicted that initially more
native annuals would establish than native perennials but that native perennial abundance
will increase after 1 year.
For the second part of my study, I experimentally tested whether addition of P.f.D7 rhizobacteria reduced IWAG cover over 3 years by comparing control and treated
plots in 2009 (before treatment), 2010 (after treatment), and 2012. I predicted that the
2012 abundance of B. tectorum and other IWAG would be lower than in 2009 and 2010,
because rhizobacteria may take up to five years to influence growth.
Finally in part three of my study, I experimentally tested the effect of adding
native perennial seed to plots over three years by comparing seeded species abundances
in control and treated plots in 2009, 2010, and 2012. I predicted that seeded perennial
species, which take a long time to germinate, would have increased in abundance since
before treatment and one year after treatment; with a concurrent reduction in IWAG
abundance.
In addition, I assessed how soil nitrogen availability changed from August to
October to establish a base line of ammonium and nitrite/nitrate levels available during
39
fall germination. I predicted that ammonium and nitrite/nitrate levels would increase from
August to October, in conjunction with fall IWAG germination.
Study Site
Turnbull National Wildlife Refuge (TNWR) is located about 5 miles south of
Cheney, Washington and is about 18,217 acres, encompassing several habitats including
Mima mound prairie, ponderosa pine forest, wetlands, and riparian areas surrounded by
Palouse Prairie, rangeland, and agricultural fields. The Mima mound prairie remnants
encompass about 1,270 acres (47.40N, -117.52W). The climate is semiarid with an
average annual rainfall of 43 cm (17 in) and an average annual temperature of 48° F. My
study site was located on a 154-hectare portion of the Mima mound prairie known as the
Stubble Field Tract and Public Use Area (Figure 7). Mounds in this area are underlain by
either basalt bedrock or alluvial gravel substrates, which for the purpose of this paper will
be hereafter referred to as basalt mounds and alluvial mounds.
Study Design
Within the Mima mound prairie, three experiments were carried out. Experiment
1 was replicated on both alluvial and basalt mounds and tested the effect of prescribed
fire, herbicide, addition of native annual seed only, and addition of native annual and
perennial seed combined on species composition. The multi-factorial combination of 8
treatments, including a control is listed in Table 6. The experiment was not fully factorial
due to limited time and resources; treatments were selected to include combinations most
likely to be used by refuge managers. There were 5 replicates per treatment, except for
basalt and alluvial burn plots, for which there were 10 replicates per substrate type. There
40
were 17 basalt control plots established in 2009 (seven extra plots were surveyed in case
plots were lost before treatment application) and 10 alluvial control plots, 5 of which
were established in 2009 and 5 were new. All plots that received treatments were
established in 2012. Plots were sampled prior to treatment in 2012, and after treatment in
2013, except for 3 plots that could not be relocated. Treatments were applied
sequentially, with burning in fall (October/early November), followed by herbicide 3
weeks after the last burn (late November), and finally seed addition in winter (early
February).
Low intensity burning was administered to five different large areas of the prairie
(Figure 8) to provide replication. Areas 1, 2, (basalt) and 4 (alluvial) were burned on
October 18, 2012, and areas 3 (basalt) and 5 (alluvial) were burned on November 8, 2012
(Figure 8). At the time of the first burns, IWAGs had begun to germinate and weather
conditions were favorable; unfortunately two weeks of rain delayed burning of the final
two areas. Because the weather window was small, the last two burns were only of
individual mounds with plots rather than the entire areas to better control burn intensity
(the wet conditions would not have allowed intensity levels similar to the previous fires).
To prevent late IWAG establishment, Outrider® herbicide was applied on
November 27 and 28, 2012 to each designated plot, at 1 oz/acre by backpack sprayer. On
February 5, native annual seed (C. pulchella and M. gracilis; collected in August of 2012
from TNWR outside the study area) and perennial seed (F. idahoensis, P. secunda, and
P. spicata; supplied by TNWR) were dispersed either together or annual seed only. All
seed was hand dispersed, with annual seed at a rate of 0.16 g/m2 and perennial seed at a
rate of 2.11 g/m2. Both rates were adjusted to account for dispersal type (hand broadcast)
41
using the BFI native seed calculator (www.bfinativeseeds.com).
To determine the effect of rhizobacteria on IWAG abundance, I resurveyed 14
fenced plots, of which seven had been inoculated with rhizobacteria in 2009; the other
seven were fenced controls. Plots were surveyed before inoculation (2009), and one
(2010) and three years (2012) post inoculation. All 14 plots are on alluvial substrate
mounds that were initially selected for their high abundance of B. tectorum in June and
July of 2009. In November 2009, P.f.D7 was applied to 7 plots on the soil surface at rate
of 108 cells/m2 with a backpack sprayer in water equivalent to 10 gallons per acre. All 14
plots were fenced before application and remain so.
To assess the effects of native perennial seed addition over three years, 22 plots
that were established and surveyed in 2009 were relocated and resurveyed in 2012. There
were 22 control plots (5 alluvial and 17 basalt substrate) and 7 treated plots (3 alluvial
and 4 basalt substrate). The seed of two native perennial bunchgrass species (P. spicata
and F. idahoensis) were hand broadcast in late November 2009 at a rate of 1.36 g/m2.
Finally, soil samples were taken from 10 basalt and 10 alluvial control plots every
two weeks from August 10, 2012 to October 12, 2012 to determine how soil nitrogen
levels changed over time. I removed the top layer of soil and collected a 50 g sample
from the side of the hole, about 10-20 cm deep. Soil nitrogen samples were put in paper
bags in direct sunlight until dry and all rocks and organic matter were removed with a
three-tiered sieve. These samples were sent to University of Idaho Analytical Sciences
Laboratory for analysis of nitrogen concentrations (ammonium, nitrite, and nitrate) using
the 2N KCl soil extraction for nitrate, nitrite, and ammonium.
Vegetation Survey Details
42
To account for plant phenology changes over the survey period, the different
treatments and locations were sampled randomly. For all 128 plots surveyed in 2012, the
following information was recorded: substrate type (alluvial and basalt), species
composition and percent cover, IWAG stem count in a 0.04 m2 corner of the plot, and
soil depth (at 2 corners of the plot). Soil samples were collected from selected plots as
described previously.
Data Analysis
To determine which combination of treatments most effectively reduced invasives
and increased natives, I used mixed linear models to compare the effect of treatments on
IWAG stem counts, native and invasive species richness and percent cover, M. gracillis
percent cover and C. pulchella percent cover (PROC MIXED; SAS 9.3 software). I used
mixed linear models because I had both random (burn area nested within substrate) and
fixed (treatment) variables. These analyses included two factors: underlying substrate
(alluvial or basalt), and treatment (with 8 levels per substrate type: burn (B), burn/annual
seed (BA), burn/annual and perennial seed (BAP), burn/herbicide/annual seed (BHA),
burn/herbicide/annual and perennial seed (BHAP), herbicide/annual seed (HA),
herbicide/annual and perennial seed (HAP), and control (C)). Native species richness and
percent cover of M. gracilis and C. pulchella were not normally distributed but
transformations did not improve normality. For all significant factors a Tukey post hoc
analysis was done.
To determine how burning and herbicide (followed by seed addition) interacted to
effect IWAG stem counts, native/invasive species richness/percent cover, or percent
cover of seeded annual plants (M. gracillis and C. pulchella) I used five mixed linear
43
models (one for each variable) that accounted for post burn treatments (PROC MIXED;
SAS 9.3 software). I used mixed linear models because I had both random (burn area
nested within substrate) and fixed (treatment) variables. The three factors were substrate
type (basalt or alluvial), burning (2 levels: burned or not) and herbicide followed by seed
addition (3 levels: no herbicide or seed, herbicide with annuals only, and herbicide with
annuals and perennials), for a total of 6 treatments each on both alluvial and basalt
substrates (B, BHA, BHAP, HA, HAP, and C). The change in IWAG stem counts and M.
gracillis and C. pulchella percent cover were Box-Cox transformed (λ=1.25 for all).
Native and invasive species richness were not normally distributed and distribution could
not be improved with transformations. For all significant factors a Tukey post hoc
analysis was done.
On burned plots, to determine how seed addition type interacted with herbicide
and substrate to affect IWAG stem counts, native/invasive species richness/percent cover,
and M. gracillis and C. pulchella percent cover, I used mixed linear models (PROC
MIXED; SAS 9.3 software). I used mixed linear models because I had both random (burn
area nested within substrate) and fixed (treatment) variables. The models included three
factors: substrate type (basalt or alluvial), herbicide (used/not used), and seed addition
type (annual seed only/annual and perennial seed), which resulted in the following 4
treatments compared over basalt and alluvial substrates: BA, BAP, BHA, and BHAP. M.
gracillis and C. pulchella percent covers were Box-Cox transformed with a final λ of
0.75. For all significant factors a Tukey post hoc analysis was done.
To determine if there was a long-term affect of P.f.-D7 on stem counts and
percent cover of B. tectorum and total IWAGs I used a repeated measures ANOVA
44
(PROC MIXED; SAS 9.3 software) comparing 2009, 2010, and 2012 stem counts and
percent cover with and without treatment. For all significant factors a Tukey post hoc
analysis was done.
All data were tested for normality and homoscedasticity and transformed as
necessary using Box-Cox transformations.
Results
Effects of Fire, Herbicide, and Seed Addition on Plant Communities
The combination herbicide/seed addition treatments reduced IWAG stem counts
(Tables 7 and 8; Figure 9) irrespective of burning, substrate, or type of seed added
(annual alone or annual and perennial). Burning also reduced IWAG stem counts across
both substrates except for the burning/annual and perennial seed (BAP) treatment on
basalt, which I suspect is due seed contamination or that the soil seed bank was primarily
IWAGs (Table 8; Figure 9). In control plots, IWAG stem counts increased (Table 8).
When analyzed as a fully factorial model (e.g. removing nonfactorial variables) to assess
interactions between burning and herbicide, there was no interaction between burning,
herbicide/seed applications, or substrate type in terms of their effect on IWAG stem
counts (Table 9). When analyzed as a fully factorial model to assess the interaction
between substrate, herbicide, or type of seed added (e.g. with plots that were not burned
removed), IWAG stem counts were not influenced by an interaction between substrate,
herbicide, or type of seed addition (annual or annual/perennial; Table 10).
Similar to IWAG stem counts, invasive and native species richness decreased
when herbicide was applied (irrespective of seed addition type and burning), yet both
invasive and native species richness increased in control, burn, and burn/seed treatments
45
(Tables 11 and 12; Figure 10). However, herbicide decreased invasive richness more than
native richness (Table 13), particularly in the burn/herbicide/seed treatments (Table 14).
Invasive and native species richness was not influenced by the interaction between fire
and herbicide (Table 15; Figure 10). On burned plots, herbicide and seed addition type
interacted; herbicide reduced invasive species less when perennial seeds were added to
the mix (Table 16; Figure 10). This interaction effect was marginal and could be due to
lack of normally distributed data. Herbicide, irrespective of seed addition type, reduced
native and invasive species richness (Table 16; Figure 10).
From 2012 to 2013, the most aggressive restoration tools used
(burn/herbicide/seed and herbicide/seed) reduced mean invasive species percent cover by
38% and 57% per m2, respectively. Meanwhile, mean native species percent cover
increased across all treatments, not including the controls, from 35% to 57% per m2.
Invasive species percent cover was reduced by all herbicide treatments, but increased in
the control, burn, and burn/seed treatments (Tables 17 and 18; Figure 11). The three most
common invasive species that increased cover with burn and burn/seed treatments were
Sisymbrium altissimum, Linaria dalmatica, and Vicia cracca, yet all decreased when
herbicide was applied (Figure 12). There was no significant effect of any treatments on
native cover (Table 17; Figure 11), however burning alone appeared to increase native
cover more than herbicide with annual and perennial seed (per Tukey tests, though the
overall model was not significant; Table 18). All herbicide treatments reduced invasive
species percent cover more than native cover (Table 21; Figure 11) on both basalt
(p=0.0157, t=-2.50) and alluvial mound (p=0.0015, t=-3.34). Invasive and native cover
were not affected by the interaction between burning and herbicide (Table 19; Figure 11).
46
On burned plots, the interaction between seed treatments and herbicide had no effect on
invasive cover (Table 20; Figure 11). On burned plots, an interaction between substrate
type and herbicide affected native cover; on basalt plots native cover increased most in
the presence of herbicide; while on alluvial plots, native cover only increased in the
absence of herbicide (Table 20; Figure 13).
C. pulchella percent cover was not influenced by any treatment (Table 22) except
herbicide/annual and perennial seed which reduced its cover (Table 23; Figure 14). C.
pulchella cover was not influenced by an interaction between burning and herbicide
(Table 24) or seed addition type and herbicide in burned plots (Table 25).
Unlike C. pulchella, burning and seeding increased M. gracilis percent cover
(Tables 22 and 23; Figure 15). In burn/herbicide/seed treatments M. gracilis cover
increased, however herbicide/seed treatments reduced M. gracilis cover (Table 23).
Burning with seed addition increased M. gracilis cover the most, but the interaction
between fire and herbicide did not affect M. gracilis cover (Table 24), nor did an
interaction between herbicide and seed addition type in burned plots (Table 25).
There was no effect of any treatments on perennial seeded species; only 4
identifiable bunchgrass individuals (2 F. idahoensis, 1 P. spicata, and 1 P. secunda) were
found, and all except the P. secunda individual were present the year before treatments
were applied. There were 6 unidentifiable bunch grass individuals found in burned plots,
but only two of those had received native perennial seed. Anecdotally, I observed that
mature P. spicata and F. idahoensis responded well to burning on alluvial intermounds.
Rhizobacteria Effects
47
B. tectorum stem counts and percent cover decreased in both control and
rhizobacteria inoculated plots from 2009 to 2012 (Table 26), with no significant effect of
inoculation (p=0.0139, t=2.59; Figure 16). Likewise, total IWAG percent cover was not
reduced by the rhizobacteria treatment (p=0.5646; F=0.35). In both treated and control
plots, total IWAG percent cover changed little from 2009 to 2012 (p=0.0731; t=1.85).
2009 Native Perennial Seed
Of the 14 plots seeded in 2009, only 7 were surveyed in 2012; of these, only one
plot had 0.5% cover of F. idahoensis; otherwise the seeded perennial grasses did not
establish.
Soil Samples
Sample time and substrate type did not influence total levels of available soil
nitrogen (Table 27; Figure 16), and there was no difference in soil ammonium or
nitrite/nitrate levels among sample times or substrate type (Tables 27 and 28; Figure 17).
Soil nitrite/nitrate levels where higher than ammonium levels (Table 27; Figure 17).
Discussion
My study found that while the most effective restoration tool for reducing
invasive species was herbicide, irrespective of burning, prescribed fire was most effective
at increasing native species cover. I also found that annual seed addition can increase M.
gracilis cover over short time periods. By following the EBIPM restoration guidelines I
not only developed an understanding of common prairie restoration approaches but also
learned from previous restoration projects that were completed in my study area and
similar regions. This knowledge led me to choose restoration tools that were not used
previously at TNWR (i.e. fire and native annual seed addition). I also changed the type of
48
herbicide and rate application, as well as the rate and time of native perennial seed
application based on findings of Reynecke (2012).
Herbicide is one of the most common IWAG management tools as it consistently
reduces IWAG abundance (Steers et al. 2009, DiTomaso et al. 1997, Monaco and Creech
2004, Corbin and D'Antonio 2004b, Huddleston and Young 2005, Simmons et al. 2007,
Geisel et al. 2008, Sheley et al. 2012). There are a wide variety of herbicides available
and each can affect plants differently. However, as herbicide use increases so will the
number of resistant populations. While having populations of native species resistant to
herbicide can be beneficial for restoration, invasive species resistance will only
exacerbate the issues associated with high invasive cover. Populations of the invasive
species, Lactuca serriola (Prickly Lettuce), are resistant to certain sulfonylurea,
imidazolinone, imazapyr, and imazethypyr herbicides (Mallory-Smith et al. 1990) and as
of 1990 at least 40 eudicot and 15 grass (including B. tectorum) populations were
documented to be resistant to triazine herbicides (Holt and LeBaron 1990). Because
herbicide resistant populations will increase as use increases, herbicide may not be a
viable long-term solution; instead restoration projects should generate and promote selfsustaining ecological communities.
In the short-term, however, herbicides are able to reduce the cover of IWAGs and
other invasives species, and when applied in combination with fire, native cover can
increase (Davies et al. 2011, DiTomaso et al. 1997, Simmons et al. 2007). This works
partially because fire triggers germination of some invasive species, such as Bromus
hordeaceus (soft brome), thus a post fire herbicide application could remove those
individuals (Sheley et al. 2007, Boyd and Rafael 1996). With the litter layer removed by
49
fire, herbicide efficiency should increase (Rhoades et al. 2002, DiTomaso et al. 2009).
Most IWAG species have relatively short seed dormancy periods, like ventenata and
downy brome, which have seeds that are only viable for about 2-3 years (USDA 2008,
Young et al. 1987) therefore, long-term use of aggressive restoration tools could reduce
future IWAG cover. The rate and timing of my herbicide application successfully
targeted IWAG fall growth and although native richness decreased, native cover
increased. This implies that while many TNWR native species are not tolerant to
Outrider® herbicide, those that are can benefit from IWAG removal. Other studies have
found similar trends where native species richness is lower shortly after fire and/or
herbicide application but gradually increases while invasive cover decreases over time
(DiTomaso et al. 1997, Keeley et al. 2005, Nyamai et al. 2011, Sheley et al. 2012, Rinella
et al. 2012).
Another aggressive restoration tool is prescribed fire, which reduces IWAG cover,
removes litter and potentially seed banks, and promotes native growth. The effect of fire
on IWAG abundance is highly dependent on fire return intervals; IWAG abundance has
been shown to increase if too frequent fires kill native species (DiTomaso 2000, Young
and Allen 1997, McIver and Starr 2001, Keeley et al. 2005, Keeley 2006). However, as
IWAG abundance increases so does the fire reoccurrence interval (Davies et al. 2007,
Miller et al. 2012), further promoting IWAG establishment and a more-frequent fire cycle
(Whisenant and Uresk 1990).
There are several seasons where periodic prescribed fires can control IWAGs;
early spring fires can prevent established IWAGs from setting seed (Whisenant and
Uresk 1990), while early summer fires that occur after seed production but before seeds
50
are released can reduce IWAG seed bank and cover (Davies et al. 2013, Pollak and Kan
1998). Fall burns can reduce the number of established IWAG individuals and thereby
limit competition with native species in spring (Nyamai et al. 2011). Plus, when
prescribed fires mimic the historic fire cycle of the area, native growth has been shown to
increase (Davies et al. 2013, Davies and Sheley 2007). Long-term monitoring is required
after prescribed fires, even if the appropriate fire cycle has been mimicked, because
IWAG abundance can be high the year following the fire, while native species may
require several years to become dominant (Keeley et al. 2005, Davies and Sheley 2007).
Along with killing IWAG individuals, prescribed fires remove dense IWAG
generated litter that favors IWAG growth (Keeley and McGinnis 2007) and can
potentially reduce the seed bank in the litter or top cm of soil (DiTomaso et al. 1999).
Dense litter shades out native seedlings and biocrust communities while trapping seeds
close to the soil (Evans and Young 1970) and increasing winter soil temperature and
moisture favorable to IWAG germination and growth (Facelli and Pickett 1991). Fires
that are hot enough to penetrate the top few cm of soil and remove the litter layer can
potentially destroy a large portion of the seed bank (Boudell et al. 2002), but a single cool
burn may be insufficient to dramatically reduce the soil seed bank (DiTomaso et al.
1999). Removing IWAG seeds from the seed bank is important as future competition will
primarily depend on what species are present (Keeley et al. 2005, Keeley 2006). After a
cool fire, if the viable seed bank is comprised of mainly IWAG species, the next year’s
growth will still be dominated by IWAGs. This may explain my finding that invasive
species richness/cover increased when areas were burned but herbicide was not used. A
51
low intensity burn may still confer benefit by reducing the amount of IWAG seeds being
added to the soil, and the competitive advantage allotted by dense litter cover.
Along with aggressive IWAG restoration tools, sowing seed of desired perennial
species is common, but establishment is not always successful (Cox and Anderson 2004,
Nyamai et al. 2011) as I found. Most native perennial establishment success is highest
with tilling or drilling rather than broadcast seed dispersal (Davies et al. 2013, Cox and
Anderson 2004, Nyamai et al. 2011). But tilling and drilling are not viable options for
certain habitats, such as Mima mound prairies. The limited native perennial establishment
I observed may be due to low seedling vigor especially when competing with IWAG
seedlings (Ray-Mukherjee et al. 2011, James et al. 2011). However, when competition is
between IWAG seedlings and an established native perennial, the perennial tends to
outcompete the IWAG (Beckstead and Augspurger 2004, Ponzetti et al. 2007). Another
possible explanation for low perennial recruitment is that mature native bunchgrass
populations may not do well on the tops of Mima mounds at TNWR. Anecdotal
observations from 2013 suggest that more perennial grasses are on the alluvial
intermounds where mature bunchgrasses responded well to fire; however I did not survey
alluvial intermounds.
The low native perennial establishment rates may also be due to IWAGs having
altered the soil chemistry and microbiota (Belnap et al. 2005, Busby et al. 2012), which
can prevent native perennial establishment (Mangla and Callaway 2007). In this case,
assisted succession with annual seed could potentially benefit native growth by providing
a rapid growing annual species to compete with the IWAGs, and by potentially restoring
the soil chemistry and microbiota to a state where native perennial seed can germinate.
52
This assumes that because native perennials evolved to establish after the native annuals,
they share the same microbiota and soil chemistry preferences.
I tested the effect of assisted succession at TNWR by seeding with two native
annuals; M. gracilis responded well to fire, though it was not associated with lower
invasive species numbers. C. pulchella did not increase with seeding, which could be due
to M. gracilis being a more aggressive establisher, seed viability, or plant size, as M.
gracilis tends to be larger than C. pulchella. Of the two annuals used here, M. gracilis
would be a better choice for rapidly increasing native cover, but the desired end result of
restoration is important when choosing which annual species to disperse. C. pulchella and
other Clarkia sp. are highly attractive to people and pollinators and are used in pollinator
seed mixes. So while C. pulchella may not be an aggressive establisher, its use should be
further assessed with different seed densities and possibly in different years when the
weather differs. Other potential native annuals can include, Amsinkia menzesii, Collomia
lineraris, and Polygonum douglasii, as all appear to colonize quickly after disturbance.
More research needs to be done to find more/better native annual species that can
increase native cover quickly and long-term monitoring is required to determine if these
responses persist; my results are based on the first year (Rinella et al. 2012).
After three years, the biocontrol P.f.-D7 rhizobacteria, which was applied in 2009,
had no effect on IWAG stem counts or cover. There are a variety of possible reasons that
this application did not work despite success in other projects (Kennedy et al. 1991,
Kremer and Kennedy 1996, Kennedy et al. 2001). P.f.-D7 is a cold bacterium, but on the
day of application, the ambient temperature was below freezing, most likely resulting in
the bacteria freezing before they could enter the soil. It is also possible that the bacteria
53
successfully entered the soil but did not survive the winter. A third possible explanation
is that TNWR has a wetter climate than some of the other study areas and the excess
water availability may have allowed B. tectorum to overcome the negative impacts of the
rhizobacteria.
My research goal was to restore a native plant community on the Turnbull
National Wildlife Refuge Mima mound prairie that is able to sustain itself with minimal
long-term intervention. I found that while herbicide was effective at reducing invasive
species and may be an excellent short-term solution, fire appears to more strongly benefit
native species, the success of which could help limit further invasion, increase
community resistance and sustain the restoration. Further, seeding with local native
annuals, such as M. gracilis, may help promote succession to a sustainable native
community by taking up space and resources that would otherwise be available for
invasive annuals. I would suggest that managers at TNWR and similar regions develop a
prescribed burn schedule based on the historic fire regime, and only use an herbicide that
targets invasive species (but minimizes effects to natives) in areas with dense IWAG
growth. Herbicide and fire application will need to be timed carefully to maximize
damage to invasives while minimizing harm to natives. After these have been applied, I
recommend dispersing seed of M. gracilis and other native annual species, followed by
native perennial species once the native annual abundance has increased. This plan
should increase native cover, while reducing IWAG abundance, and promote a selfsustaining system.
54
Chapter 2 Figures
Figure 7 Aerial photograph of the Mima mound prairie at Turnbull National Wildlife
Refuge (TNWR). The area of TNWR is in the upper right hand corner and the pink
polygon is the general study area. The white points indicate individual plots.
55
Figure 8 Aerial photo of the Mima mound prairie; the colored polygons indicate the burn
areas. Areas 1, 2, and 3 are underlain by basalt substrate and areas 4 and 5 are underlain
by alluvial gravel.
56
Total Stem Count Change
per 0.04m2
100
50
0
-50
-100
-150
-200
-250
-300
AL
BA
Treatment
Figure 9 Mean IWAG stem count changes from 2012 to 2013 after each treatment on
both basalt and alluvial mound plots. Error bars represent + 1 standard error. Statistical
inferences are based on Box-Cox transformed data.
57
Control
Herb/Ann/Per
Herb/Ann
Burn/Herb/Ann/Per
Burn/Herb/Ann
Burn/Ann/Per
Burn/Ann
Burn
Control
Herb/Ann/Per
Native
Herb/Ann
Burn/Herb/Ann/Per
Burn/Herb/Ann
Burn/Ann/Per
Burn/Ann
Burn
Species Richness Change per m2
Invasive
6
AL
4
BA
2
0
-2
-4
-6
-8
Treatment
Figure 10 Mean changes in native and invasive species richness per m2 from 2012 to
2013 on basalt and alluvial mounds after all treatments. Error bars represent + 1 standard
error.
58
Control
Herb./Ann./Per.
Herb./Ann.
Burn/Herb./Ann./Per.
Burn/Herb./Ann.
Burn/Ann./Per
Burn/Ann.
Burn
Control
Herb./Ann./Per.
Native
Herb./Ann.
Burn/Herb./Ann./Per.
Burn/Herb./Ann.
Burn/Ann./Per
Burn/Ann.
Burn
Percent Cover Change per m2
Invasive
80
60
Alluvial
40
Basalt
20
0
-20
-40
-60
-80
-100
-120
Figure 11 The effect of treatment on native and invasive species percent cover per m2
plots from 2012 to 2013 on basalt and alluvial mounds. Error bars represent + 1 standard
error. All statistical inferences are based on Box-Cox transformed data.
59
30
20
10
0
-10
-20
-30
-40
Treatment
Linaria
dalmatica
Sisymbrium
altissimum
Vicia cracca
Figure 12 Mean change in Sisymbrium altissimum, Linaria dalmatica, and Vicia cracca
percent cover from 2012 to 2013 after treatment. Error bars represent + 1 standard error.
60
Native Species Percent Cover
Change per m2
70
60
50
40
30
20
10
0
-10
-20
-30
AL
BA
Treatment
Figure 13 Mean change in native species cover from 2012 to 2013 on alluvial and basalt
burn plots that received seed or herbicide and seed addition. Error bars represent + 1
standard error.
61
10
AL
5
BA
0
-5
-10
-15
-20
Treatment
Figure 14 Mean change in C. pulchella percent cover in 1 m2 plots from 2012 to 2013 on
basalt and alluvial mounds after treatment. Error bars represent + 1 standard error.
62
40
30
20
10
0
-10
-20
-30
AL
BA
Treatment
Figure 15 Change in Madia gracilis abundance in 1 m2 plots from 2012 to 2013 on basalt
and alluvial mounds after each treatment. Error bars represent + 1 standard error.
Repeated letters across bars indicate no significant difference between treatments.
63
A
A
AB
AB
B
B
Figure 16 Mean Bromus tectorum percent cover in P.f.-D7 rhizobacteria treated and
control plots from 2009, 2010, and 2012. Error bars represent + 1 standard error.
Repeated letters indicate no significant difference.
64
12
A
Mean total N ug/g
10
8
A
A
A
A
6
4
2
0
1
2
3
Sampling Time
4
5
Figure 17 Mean total nitrogen in soil samples across all five sampling times. 1=August
10, 2012, 2=August 20, 2012, 3=September 13, 2012, 4=September 28, 2012, and
5=October 12, 2012. Letters indicate no statistical difference. Error bars represent + 1
standard error.
65
12
Nitrogen amount (ug/g)
Alluvial
10
Basalt
8
6
4
2
0
1
2
3
4
5
Ammonium
1
2
3
4
5
Nitrite/Nitrate
Figure 18 Mean 2012 nitrite/nitrate and ammonium levels near basalt and alluvial control
plots from each soil sample period. 1=August 10, 2=August 20, 3=September 13,
4=September 28, and 5=October 12. Error bars represent + 1 standard error.
66
Chapter 2 Tables
Table 6 Eight treatment combinations of fire, Outrider® herbicide, and native annual and
perennial seed that were applied fall 2012 to winter 2013. Each treatment was replicated
on basalt and alluvial mounds.
Treatments
Burn
(B)
1
2
3
4
5
6
7
Control
X
X
X
X
X
Herbicide
(H)
X
X
X
X
Native
Annual
Seed (A)
X
X
X
X
X
X
67
Native
Number
Perennial
of
Seed (P)
Basalt
Plots
11
5
X
5
5
X
4
4
X
3
17
Number
of
Alluvial
Plots
11
3
5
5
5
5
5
10
Table 7 Summary of F and p-values from a mixed linear model analysis to determine the
effect of treatment and substrate type on total IWAG stem counts. Significant values are
in bold.
Type 3 Tests of Fixed Effects on
Total IWAG Stem Counts
Effect
F Value (DF) p Value
Substrate
1.06 (86)
0.31
Treatment
5.86 (86)
<0.0001
Substrate*Treatment
1.19 (86)
0.31
68
Table 8 Summary of t and p values from Tukey post-hoc analysis of a mixed linear model
analyses that compare the effect of treatment on total IWAG stem counts. B=Burn,
BA=Burn/Annual Seed, BAP=Burn/Annual and Perennial Seed, BHA=Burn/Herbicide/
Annual Seed, BHAP= Burn/Herbicide/Annual and Perennial Seed,
HA=Herbicide/Annual Seed, HAP= Herbicide/Annual and Perennial Seed, C=Control.
Treatment
B
B
B
B
B
B
B
BA
BAP
BHA
BHAP
HA
HAP
BA
BA
BA
BA
BA
BHA
BHA
BHA
BHA
BHAP
BHAP
BHAP
BAP
BAP
HA
Total IWAG Stem
Count
t Value(DF)
p Value
0.18(86)
0.86
1.18(86)
0.24
1.94(86)
0.056
-1.83(86)
0.071
1.85(86)
0.068
1.64(86)
0.11
-3.26(86)
0.0016
-2.45(86)
0.016
-0.69(86)
0.49
-3.72(86)
0.0004
-4.22(86)
<0.0001
4.13(86)
<0.0001
3.92(86)
0.0002
0.77(86)
0.44
1.43(86)
0.16
-1.59(86)
0.12
1.36(86)
0.18
1.19(86)
0.24
0.76(86)
0.45
-2.55(86)
0.013
0.68(86)
0.50
0.5(86)
0.62
-3.12(86)
0.0025
-0.07(86)
0.94
-0.24(86)
0.81
3.04(86)
0.0031
2.86(86)
0.0053
-0.17(86)
0.86
Treatment
BA
BHA
BHAP
BAP
HA
HAP
C
C
C
C
C
C
C
BHA
BHAP
BAP
HA
HAP
BHAP
BAP
HA
HAP
BAP
HA
HAP
HA
HAP
HAP
69
Table 9 Summary of F and p values for a mixed linear model analysis comparing the
effects of substrate type, fire (burned), herbicide/seed application (HAP) and the
interaction of fire and herbicide/seed application on total IWAG stem counts. Significant
values are indicated in bold.
Total IWAG Stem Count
Effect
F Value(DF) p Value
Substrate
1.34(72)
0.25
Burned
0.86(72)
0.36
Substrate*Burned
0.01(72)
0.92
HAP
11.53(72)
<0.0001
Substrate*HAP
2.64(72)
0.079
Burned*HAP
2.64(72)
0.078
Substrate*Burned*HAP
0.73(72)
0.48
70
Table 10 Summary of F and p values for a mixed linear model analysis comparing the
effects of substrate type, herbicide and seed addition type (annual seed only or annual and
perennial seed together) in burned plots on total IWAG stem counts. Significant values
are indicated in bold.
Type 3 Tests of Fixed Effects
IWAG Stem Counts
Effect
F Value(DF) Pr > F
Substrate
0.33(2.51)
0.61
Herbicide
10.20(24.3) 0.0039
Substrate*Herbicide
0.32(24.3)
0.58
Seed Addition
0.50(24.2)
0.48
Substrate*Seed
0.11(24.2)
0.74
Herbicide*Seed
2.21(25.1)
0.15
Substrate*Herbicide*Seed
0.08(25.1)
0.79
71
Table 11 Summary of F and p-values from two mixed linear model analyses to determine
the effect of treatment and substrate type on invasive and native species richness.
Significant values in bold.
Effect
Substrate
Treatment
Substrate*Treatment
Invasive Species
Native Species
Richness
Richness
F Value(DF)
p Value
F Value(DF) p Value
0.06(2.95)
0.82
1.32(86)
0.25
19.11(24.4)
3.92(86)
<0.0001
0.0009
0.75(24.4)
0.63
1.33(86)
0.25
72
Table 12 Summary of t and p values from Tukey post hoc analysis of two mixed linear
model that compare the effect of treatment on native and invasive species richness.
B=Burn, BA=Burn/Annual Seed, BAP=Burn/Annual and Perennial Seed,
BHA=Burn/Herbicide/ Annual Seed, BHAP= Burn/Herbicide/Annual and Perennial
Seed, HA=Herbicide/Annual Seed, HAP= Herbicide/Annual and Perennial Seed,
C=Control. Significant values are indicated in bold.
Treatment
B
B
B
B
B
B
B
BA
BAP
BHA
BHAP
HA
HAP
BA
BA
BA
BA
BA
BHA
BHA
BHA
BHA
BHAP
BHAP
BHAP
BAP
BAP
HA
Treatment
BA
BHA
BHAP
BAP
HA
HAP
C
C
C
C
C
C
C
BHA
BHAP
BAP
HA
HAP
BHAP
BAP
HA
HAP
BAP
HA
HAP
HA
HAP
HAP
Invasive Species
Richness
t Value(DF) p value
-1.05(84.3)
0.30
7.84(84.6) <0.0001
5.31(83.9) <0.0001
0.46(85.9)
0.64
2.23(6.05)
0.067
2.67(6.05)
0.037
-0.87(3.52)
0.44
-0.05(5.93)
0.96
-1.1(5.01)
0.32
-5.88(4.96)
0.0021
-4.38(5.96)
0.0047
4.48(83.2) <0.0001
5.15(83.2) <0.0001
7.13(83.5) <0.0001
5.17(84)
<0.0001
1.28(84.5)
0.20
2.65(9.03)
0.026
3.05(9.03)
0.014
-1.56(84.5)
0.12
-6.28(83.9) <0.0001
-2.43(7.86)
0.042
-2.02(7.86)
0.079
-4.17(85.4) <0.0001
-1.24(9.05)
0.25
-0.84(9.05)
0.42
1.8(7.9)
0.11
2.22(7.9)
0.058
0.54(83.2)
0.59
73
Native Species
Richness
t Value
p value
-0.92(86)
0.36
2.43(86)
0.017
2.45(86)
0.016
0.98(86)
0.33
1.55(86)
0.12
3.25(86)
0.0016
-0.43(86)
0.67
0.63(86)
0.53
-1.33(86)
0.19
-2.79(86)
0.0065
-2.78(86)
0.0067
1.87(86)
0.065
3.59(86)
0.0006
2.7(86)
0.0084
2.74(86)
0.0074
1.57(86)
0.12
2.01(86)
0.048
3.39(86)
0.0010
0.23(86)
0.82
-1.22(86)
0.22
-0.55(86)
0.58
0.93(86)
0.36
-1.36(86)
0.18
-0.73(86)
0.47
0.65(86)
0.52
0.58(86)
0.56
2.06(86)
0.042
1.38(86)
0.17
Table 13 Summary of F and p values from a mixed linear model comparing the effects of
substrate type, treatment, to total species richness. Significant values indicated in bold.
Type 3 Tests of Fixed Effects on Total Species
Richness
Effect
F Value(DF)
p Value
Substrate
0.38(2.85)
0.58
Treatment
17.56(28.5)
<0.0001
Substrate*Treatment
1.12(28.5)
0.38
Native vs Invasive (NvI)
26.19(169)
<0.0001
Substrate*NvI
0.31(169)
0.58
Treatment*NvI
2.56(169)
0.016
Substrate*Treatment*NvI
0.50(169)
0.84
74
Table 14 Summary of t and p-values from Tukey post-hoc analysis of a mixed linear
model to determine if treatments affected the change of native species richness differently
than invasive species richness. Significant values are indicated in bold. B=Burn,
Treatment
B
BA
BAP
BHA
BHAP
HA
HAP
C
t Value (DF)
-1.60(169)
-1.23(169)
-0.41(169)
-4.87(169)
-2.63(169)
-1.97(169)
-0.93(169)
-0.77(169)
75
P Value
0.11
0.22
0.68
<0.0001
0.0093
0.050
0.36
0.44
effects of substrate type, fire (burned), and herbicide/seed treatment (HAP) on invasive
and native species richness. Significant values are indicated in bold.
Invasive Species
Richness
Effect
F Value(DF) p Value
Substrate
0.00(2.35)
0.97
Burned
2.81(2.35)
0.22
Substrate*Burned
0.12(2.35)
0.76
HAP
49.36(70)
<0.0001
Substrate*HAP
1.03(70)
0.36
Burned*HAP
2.08(70)
0.13
1.12(70)
0.33
76
Native Species
Richness
F Value(DF)
p Value
0.15(1.66)
0.75
0.00(1.66)
0.95
3.13(1.66)
0.24
10.97(70.7)
<0.0001
0.76(70.7)
0.47
0.42(70.7)
0.66
0.33(70.7)
0.72
species richness. Significant values are indicated in bold.
Effect
Substrate
Herbicide
Substrate*Herbicide
Seed
Substrate*Seed
Herbicide*Seed
Invasive Species
Richness
F Value(DF) p Value
0.55(2.84)
0.52
75.68(25.6)
<0.0001
0.06(25.6)
0.82
0.03(25.6)
0.86
2.11(25.6)
0.16
4.56(26.5)
0.042
1.64(26.5)
77
0.21
Native Species
Richness
F Value(DF)
p Value
4.22(28)
0.049
6.14(28)
0.020
0.13(28)
0.73
1.21(28)
0.28
0.15(28)
0.70
0.66(28)
0.42
0.78(28)
0.38
Table 17 Summary of F and p values from two mixed linear model analyses to determine
if treatment or substrate type influenced invasive and native species percent cover.
Effect
Substrate
Treatment
Invasive Species
Native Species
Percent Cover
Percent Cover
p Value
F Value p Value
F Value
0.19(2.04)
0.70
0.18(2.23)
0.71
13.86(23.3)
1.31(23.1)
0.29
<0.0001
Substrate*Treatment
0.56(23.3)
0.78
78
1.86(23.1)
0.12
Table 18 Summary of t and p values from Tukey post-hoc analysis of two mixed linear
models comparing the effect of treatments on native and invasive species percent cover.
B=Burn, BA=Burn/Annual Seed, BAP=Burn/Annual and Perennial Seed,
BHA=Burn/Herbicide/ Annual Seed, BHAP= Burn/Herbicide/Annual and Perennial
Seed, HA=Herbicide/Annual Seed, HAP= Herbicide/Annual and Perennial Seed,
C=Control. Significant values are indicated in bold.
Treatment Treatment
Invasive Species
Percent Cover
t ValueDF
p Value
Native Species
Percent Cover
t ValueDF
p Value
B
BA
0.2884.8
0.78
0.1585.7
0.88
B
BHA
4.6885.1
<0.0001
1.885.8
0.076
B
BHAP
4.8683.9
<0.0001
1.1285
0.27
B
BAP
-0.3285.4
0.75
1.582.5
0.14
B
HA
5.057.02
0.0015
1.113.4
0.29
B
HAP
4.847.02
0.0019
2.5413.4
0.024
B
C
0.472.82
0.67
1.773.72
0.16
BA
C
0.146.85
0.90
1.1413
0.27
BAP
C
0.665.11
0.54
-0.038.55
0.97
BHA
C
-3.455.16
0.017
-0.319.13
0.76
BHAP
C
-3.726.8
0.0078
0.212.6
0.84
HA
C
5.6782.2
<0.0001
-0.1882.8
0.86
HAP
C
5.4282.2
<0.0001
1.3582.8
0.18
BA
BHA
3.4583.1
0.0009
1.2884
0.20
BA
BHAP
3.7284.2
0.0004
0.7985.5
0.43
BA
BAP
-0.4985.2
0.62
1.0586
0.30
BA
HA
4.1412.6
0.0012
0.825.9
0.43
BA
HAP
3.9612.6
0.0017
1.9925.9
0.058
BHA
BHAP
0.5485.2
0.59
-0.4386
0.66
BHA
BAP
-4.2583.9
<0.0001
-0.2585
0.81
BHA
HA
1.3610.4
0.20
-0.3921.4
0.70
BHA
HAP
1.1710.4
0.27
0.8721.4
0.39
BHAP
BAP
-4.4386
<0.0001
0.2185.3
0.84
BHAP
HA
0.8412.5
0.42
0.0325.2
0.98
79
BHAP
HAP
0.6612.5
0.52
1.2225.2
0.24
BAP
HA
4.7810.2
0.0007
-0.1720.1
0.86
BAP
HAP
4.5910.2
0.0009
1.0920.1
0.29
HA
HAP
-0.2182.2
0.84
1.2382.8
0.22
80
cover. Significant values are indicated in bold.
Effect
Substrate
Burned
Substrate*Burned
HAP
Substrate*HAP
Invasive Species
Percent Cover
F Value(DF) p Value
0.08(1.79)
0.81
1.32(1.79)
0.38
0.46(1.79)
0.57
38.92(70.4)
<0.0001
1.11(70.4)
0.34
Native Species Percent
Cover
F Value(DF)
p Value
2.05(2.18)
0.28
0.85(2.18)
0.45
0.04(2.18)
0.85
1.68(70.7)
0.19
1.88(70.7)
0.16
Burned*HAP
0.51(70.4)
0.60
1.01(70.7)
0.37
0.56(70.4)
0.57
0.46(70.7)
0.63
81
species percent cover. Significant values are indicated in bold.
Effect
Substrate
Herbicide
Substrate*Herbicide
Seed
Substrate*Seed
Herbicide*Seed
Invasive Species
Percent Cover
F Value(DF) p Value
0.63(2.56)
0.49
57.25(25.7)
<0.0001
0.01(25.7)
0.92
0.01(25.8)
0.92
0.79(25.8)
0.38
1.2(27)
0.28
1.2(27)
0.28
82
Native Species
Percent Cover
F Value(DF)
p Value
0.04(28)
0.84
0.49(28)
0.49
7.03(28)
0.013
0.13(28)
0.72
0.29(28)
0.60
0.77(28)
0.39
0.51(28)
0.48
Table 21 Summary of t and p-values from Tukey post-hoc analysis of a mixed linear
model to determine if post burn treatments (herbicide and native annual and native
perennial seed) affected native species percent cover differently than invasive species
percent cover. Significant values are indicated in bold. B=Burn, BHA=Burn/Herbicide/
Treatment
B
BHA
BHAP
HA
HAP
C
t Value (DF)
-1.04(141)
-3.47(141)
-3.94(141)
-4.51(141)
-3.29(141)
-0.17(141)
83
P Value
0. 11
0.0007
0.0001
<0.0001
0.0013
0.87
Table 22 Summary of F and p values from two mixed linear model analyses, which
examined the effects of substrate type and treatment on the success of annual seed
addition. Significant values are indicated in bold.
Effect
Substrate
Treatment
Substrate*Treatment
Madia gracilis
Clarkia pulchella
Percent Cover
Percent Cover
F Value(DF) Pr > F
F Value(DF) Pr > F
0.28(87)
0.60
0.82(2.8)
0.44
1.97(87)
0.068
1.44(21.2)
0.24
0.6(87)
0.76
1.02(21.2)
0.45
84
Table 23 Summary of t and p-values from Tukey post-hoc analysis of two mixed linear
model to determine how each treatment changed the percent cover of M. gracilis and C.
pulchella compared to the control. Significant values are indicated in bold.
HA=Herbicide/Annual Seed, HAP= Herbicide/Annual and Perennial Seed. Significant
values are indicated in bold.
Treatment
BA
BAP
BHA
BHAP
HA
HAP
Madia gracilis
Percent Cover
t Value (DF)
p Value
2.56 (87)
0.012
2.02 (87)
0.046
1.01 (87)
0.31
0.77 (87)
0.44
0.62(87)
0.54
0.46 (87)
0.64
85
Clarkia pulchella
Percent Cover
t Value (DF)
p Value
-0.39 (5.22)
0.71
-0.17 (4.49)
0.87
-0.72 (4.44)
0.51
-0.93 (4.74)
0.40
1.06(84)
0.29
2.45 (84)
0.016
cover. Significant values are indicated in bold.
Effect
Substrate
Burn
Substrate*Burn
Herbicide/Seed (HAP)
Substrate*HAP
Burn*HAP
Substrate*Burn*HAP
Madia gracilis Percent
Clarkia pulchella Percent
Cover
Cover
F Value(DF)
Pr > F
F Value(DF)
Pr > F
1.79(73)
0.19
0.41(2.52)
0.58
4.62(73)
0.05(2.52)
0.85
0.035
0.69(73)
0.41
0.17(2.52)
0.72
0.35(73)
0.71
3.95(70.7)
0.024
0.11(73)
0.90
1.47(70.7)
0.24
0.03(73)
0.97
0.24(70.7)
0.79
0.17(73)
0.85
1.09(70.7)
0.34
86
effects of substrate type, herbicide and seed addition type on Madia gracilis and Clarkia
pulchella percent cover in burned plots only from 2012 to 2013. Significant values are
indicated in bold.
Madia gracilis
Clarkia pulchella
Percent Cover
Percent Cover
Effect
F Value(DF) Pr > F F Value(DF) Pr > F
Substrate
0.24(2.58)
0.66
2.04(2.61)
0.26
Herbicide
3.3(27.3)
0.080
2.58(26.6)
0.12
Substrate*Herbicide
1.11(27.3)
0.30
0.02(26.6)
0.88
Seed Addition
0.13(27)
0.72
0(26.3)
0.10
Substrate*Seed
0.02(27)
0.89
3.31(26.3)
0.080
Herbicide*Seed
0.24(28.8)
0.63
1.06(27.9)
0.31
0.63(28.8)
0.43
0.14(27.9)
0.71
87
Table 26 Summary of F and p values from a mixed linear model comparing the effect of
year and P.f.-D7 rhizobacteria treatment on B. tectorum stem counts. Significant values
indicated in bold.
Effect
F Value(DF)
Pr > F
Treatment
0.04(14.9)
0.85
Year
Year*Treatment
17.38(25.4) <0.0001
0.10(25.4)
88
0.90
Table 27 Summary of F and p values from a mixed linear model analysis comparing
effect of substrate type and time of soil collection on available nitrogen. Session indicates
time of soil collection. Significant values are indicated in bold.
Effect
F Value P Value
Session
1.02(56)
0.40
Substrate
1.87(19)
0.19
Ammonium VS Nitrite/Nitrate
(N Form)
15.32(19)
0.0009
Session*Substrate
1.41(56)
0.24
Substrate*N Form
0.01(19)
0.91
Session*N Form
0.56(56)
0.69
Session*Substrate*N Form
0.54(56)
0.70
89
Table 28 Summary of t and p values for a Tukey post-hoc test of a mixed linear model
comparing total nitrogen availability of each session (1=August 10, 2=August 20,
3=September 13, 4=September 28, and 5=October 12).
Session
1
1
1
1
2
2
2
3
3
4
Session
2
3
4
5
3
4
5
4
5
5
t Value
-1.46(56)
-0.21(56)
-0.85(56)
-0.8(56)
1.81(56)
0.9(56)
0.95(56)
-0.93(56)
-0.85(56)
0.06(56)
90
p Value
0.15
0.84
0.40
0.43
0.075
0.37
0.35
0.36
0.40
0.95
The effects of fire, herbicide, and human trampling on semiarid biological soil crust
and the invasive annual grass Ventenata dubia
Abstract
Aggressive restoration tools (i.e. prescribed fire and herbicide) are increasingly being
used to combat the spread of invasive winter annual grasses (IWAGs; i.e. Ventenata
dubia), but it is not known how these tools affect semiarid biological soil crust (biocrust;
includes moss, lichen, cyanobacteria, and algae). It is important to understand the impacts
of these treatments on biocrust communities, as they can prevent IWAG establishment.
At Turnbull National Wildlife Refuge (TNWR), areas with biocrust (often the basalt
intermounds areas, previously described in chapter 1) were formerly relatively free of
IWAGs but the impact by V. dubia is increasing, and it is not clear how the restoration
approaches used for communities with deeper soil conditions will work in these sites. The
goal of this study was to determine how semiarid biocrust, IWAGs, and native plant
communities respond to fire, herbicide, and trampling in basalt intermound areas. In
2012, 90 1 m2 plots were surveyed in June/July for vegetation and biocrust and were then
treated with fire (October 18 and November 8), Outrider® herbicide (November 27-28),
and trampling (December 3) in a multifactorial experimental design, with seven treatment
combinations with ten replicates each (controls had 30). Plots were resurveyed June/July
of 2013.
From 2012 to 2013, V. dubia stem counts decreased by 60% in all treatments
including control, which could be from low 2013 March rainfall, resulting in drought-like
stress in the intermounds. Invasive species cover/richness and native species richness
decreased in all treatments including the controls, but native cover increased, especially
when herbicide was used, suggesting that not all native species were harmed by the
91
drought or herbicide. Treatment type also had no effect on biocrust morphology
composition, but long-term monitoring is needed to ensure this is a common response and
to document long-term changes from treatment applications.
92
Introduction
As invasive winter annual grass (IWAG) populations expand, they impact
ecosystem function, native biodiversity, forage quality (D'Antonio and Vitousek 1992,
Beckstead and Augspurger 2004) and biological soil crust (biocrust), which consists of
moss, cyanobacteria, algae, and lichens (Eldridge 1993). Biocrust community
composition is dependent on vascular plant presence (Ponzetti et al. 2007), disturbance,
precipitation, temperature, and soil textures, pH, and chemistry (Root and McCune 2011).
Intact biocrust communities stabilize the soil, prevent wind and water erosion, cycle
nitrogen, carbon, and other nutrients, retain soil moisture, and prevent seedling
establishment by creating a barrier between the seeds and soil (Prasse and Bornkamm
2000, Bowker et al. 2006, Maestre et al. 2006, Ponzetti et al. 2007). These biocrust
communities can cover extensive areas in semiarid and arid grasslands (Root and
McCune 2011). Both habitats are common in western United States, with arid receiving
less rainfall than semiarid (Schwinning et al. 2004), and although common in both
habitats, more biocrust research has been completed in arid (Prasse and Bornkamm 2000,
Belnap 2003, Ponzetti et al. 2007) than semiarid habitats (Johansen et al. 1993, O’Bryan
et al. 2009).
In semiarid sagebrush steppe, biocrust is commonly found growing between
native bunchgrass individuals (Ponzetti et al. 2007, Root and McCune 2011). When
undisturbed the community is primarily lichens (Clair et al. 1993), with higher moss
cover near vascular plants (Clair et al. 1993, Maestre et al. 2011). The early successional
cyanobacteria and algae species colonize bare soil after a disturbance (Belnap and
Eldridge 2001, Maestre et al. 2011). Depending on environmental factors, regrowth of
93
mosses and lichens can take decades, if not longer (Maestre et al. 2006, Lass and Prather
2007). An intact biocrust barrier is essential for reducing IWAG establishment, as 8090% of the seed bank can be in the litter and top 2 cm of soil in arid regions (Boudell et
al. 2002). Boudell et al. (2002) found that 60% of all seeds trapped above intact biocrust
were Bromus tectorum (downy brome), despite not being a dominant species above
ground, while the majority of seeds under the crust were perennials. Disturbances to the
biocrust or crevices caused by drying can facilitate IWAG establishment (Boudell et al.
2002) and with the higher nutrient and water availability provided by the previously
intact biocrust (Harper and Pendleton 1993, Zaady et al. 2004, Maestre et al. 2005,
Godínez-Alvarez et al. 2011, Maestre et al. 2011) IWAG individuals will do well
(Boudell et al. 2002, Godínez-Alvarez et al. 2011).
In arid regions, biocrust can be extremely sensitive to disturbance, so as IWAG
cover increases in biocrust patches, they can transition from minimally to severely
impacted due to altered soil water (White et al. 2006), nitrogen availability (James
2008a), shade, litter cover (Ponzetti et al. 2007), fire frequency and fire intensity (Kerns
et al. 2006). All of these can significantly alter microhabitats and thereby biocrust
community composition (Root and McCune 2011). The threat of IWAG species,
including B. tectorum (downy brome), Taeniatherum caput-medusae (medusahead), and
Ventenata dubia (ventenata), and others to western prairies has caused them to become
the target of countless restoration projects (Kennedy et al. 2001, Frost and Launchbaugh
2003, Cox and Anderson 2004, Keeley 2006, DiTomaso et al. 2009, Nyamai et al. 2011),
however it is unclear how semiarid biocrust communities respond to IWAGs and IWAG
removal techniques.
94
Prescribed fire is a commonly used tool to reduce IWAG cover in habitats with
fire-adapted vegetation, but fire management can have a range of effects on vascular
plant and biocrust communities. Evidence from arid habitats suggests that the absence of
fire can increase biocrust diversity if vascular plant cover remains low (Evans and Belnap
1999), while with high fire frequency and intensity, biocrust recolonization can take
decades which often results in lower biocrust diversity in arid regions (Belnap and
Eldridge 2001). In contrast, in semiarid areas with higher rainfall, biocrust abundance can
significantly increase post fire (Johansen et al. 1993, O’Bryan et al. 2009). Similar to fire,
herbicides are commonly used to reduce IWAG cover (Young et al. 1987, Monaco and
Creech 2004, Huddleston and Young 2005, Simmons et al. 2007); and some can have
minimal impact on desired native species if used properly (Stevens, 2010, Monaco and
Creech 2004). However, there has been very little research on biocrust community
response to herbicide; what there is suggests herbicide reduces biocrust cover and
richness (Zaady et al. 2004, Maestre et al. 2011). It is disconcerting that there is so little
research on how semiarid biocrust communities respond the two most common IWAG
controls (prescribed fires and herbicide; Pollak and Kan 1998, Brown et al. 2008,
Mittelhauser et al. 2011, Nyamai et al. 2011).
Other common biocrust disturbances include trampling by large animals and
vehicles (Prasse and Bornkamm 2000, Lass and Prather 2007, Chamizo et al. 2011, Root
et al. 2011), especially when the biocrust is dry or frozen (Belnap and Eldridge 2001,
Rosentreter et al. 2001). Grazing at low intensity and frequency generally has minimal
impact on biocrust communities as only compression force is applied. However, if
livestock grazing is intense, biocrust diversity decreases (Belnap and Eldridge 2001,
95
Martínez et al. 2006) as patches are killed by being covered or pulled up (Belnap and
Eldridge 2001, Ponzetti et al. 2007). In contrast to animal trampling, any form of vehicle
disturbance can devastate all biocrust that is driven over. The tires apply compression and
shear force, creating bare soil, (Belnap and Eldridge 2001), seed traps, and channels for
water (Prasse and Bornkamm 2000), effectively providing ideal establishment conditions
for IWAG individuals that may not have been able to get through the once intact biocrust.
An example of a semiarid region where biocrust and IWAGs are common is the
Channeled Scablands of eastern Washington. This unique geological feature was formed
by intense flooding from glacial Lake Missoula about 20,000 years ago (Bretz 1928). The
flooding created four main tracts, the greatest of which is the Cheney-Palouse River tract,
about 20 miles southwest of Spokane, Washington. The tract is 75 miles long and about
20 miles wide (Bretz 1928). Within the sagebrush steppe of the Cheney-Palouse River
tract lie Mima mounds (Berg 1990), which are aggregations of fine loess up to 2 m tall
and 30 m in diameter, with soil depth of the mounds and intermounds strongly influenced
by the substrate type (Del Moral and Deardorff 1976). In eastern Washington, Mima
mound prairies are commonly underlain by basalt (Bretz 1928), with rocky, typically
biocrust containing intermounds that often form ephemeral wetlands or vernal pools.
Vernal pools which fill with precipitation and lose water through evaporation or
evapotranspiration (Marty 2005), are formed when deeper soils (mounds) are surrounded
by shallow soils (intermounds; Crowe et al. 1994). Inundation tolerance influences the
species present in these areas and the wetter zones are dominated by water logged
tolerant species, which have high internal root aeration (Mcdonald et al. 2001, Gerhardt
and Collinge 2007).
96
One species that seems to prefer vernal pool-like habitat is the IWAG, V. dubia
(Scheinost, USDA), which, although not new to the Channeled Scablands (native to
Mediterranean Europe and Africa), has recently become problematic (VanVleet, 2009).
Similar to other IWAGs, V. dubia germinates in fall, grows roots in winter, and uses all
available resources for shoot growth and seed production in spring and early summer.
Once senesced, V. dubia provides no foraging or grazing benefits (Lass and Prather
2007). V. dubia has a smaller root system than other IWAG species (James 2008b),
suggesting that it may be more sensitive to low water availability than other IWAGs and
perennial species with larger root systems (Volaire et al. 1998). Even though V. dubia has
been noted in many articles as a problematic invasive (Gray and Lichthardt 2003, James
2008a, Vasquez et al. 2008b), very little has been published on species phenology,
morphology, and its effects on neighboring species.
Objectives
The goals of this study were to increase community resistance to IWAG invasion
by restoring intermound native plant communities and to determine how semiarid
biocrust responds to common prairie disturbances and restoration techniques at the Mima
mound prairie at Turnbull National Wildlife Refuge (TNWR). I addressed these goals in
an experiment in which I tested how invasive (V. dubia in particular) and native vascular
plant species and biocrust respond to fire, herbicide, and human trampling. I predicted
that fire and herbicide together would reduce invasive species cover the most, but native
cover would increase with fire. I further predicted that the combination of all three
disturbances would reduce biocrust cover the most.
97
Study Site
Turnbull National Wildlife Refuge (TNWR) is located about 5 miles south of
Cheney, Washington and is about 7,372 hectares, encompassing several habitats
including Mima mound prairie, ponderosa pine forest, wetlands, and riparian areas
surrounded by rangeland and Palouse Prairie agricultural fields. The Mima mound prairie
remnants are about 514 hectare (47.40N, -117.52W). The climate is semiarid with an
average annual rainfall of 43 cm (17 in) and average annual temperature of 48° F. Our
study site was in a 154 hectare portion of the Mima mound prairie located in the Stubble
Field Tract and Public Use Area (Figure 7).
Study Design
Fire, herbicide, and trampling, were applied to the basalt intermound areas in a
multi-factorial combination of seven treatments (including controls; Table 29), with ten
replicates each, for an initial total of 70 plots. Some plots could not be relocated in 2013,
with replicates ranging between 8 to 10 plots per treatment. Seven control transects were
also established, with four plots per transect, for a total of 28 transect plots. Transects
were located across a gradient from predominantly biocrust cover to predominantly V.
dubia cover. No treatments were applied to any transect plot. All plots were established
in 2012 and resurveyed in 2013; eleven plots (across all treatments) could not be
relocated. Only one randomly selected plot per transect was used as a control for
statistical analyses.
Low intensity fires were set by the TNWR fire crew in three areas to provide
replication (Figure 8). Areas 1 and 2 (Figure 8) were burned on October 18, 2012, and
98
area 3 was burned on November 8, 2012. At the time of the first fires, weather conditions
were favorable for a low intensity fire, but heavy rain delayed burning the final area for
about two weeks. The ground was still not dry enough for a large-scale fire, but the
weather window was small, so individual plots were burned to ensure similar intensity as
the first fires. To prevent late IWAG establishment, Outrider herbicide was applied
November 27 and 28, 2012 by backpack sprayer at 1 oz/acre. On December 3, 2012 plots
were trampled on by myself for one minute; I began in one corner and stomped along the
inside edge of the plot, continuing in parallel lines until the entire plot was covered.
Vegetation Survey Details
Plots were randomly surveyed over time across locations and treatments to avoid
bias due to phenological changes in vascular plants. All 1 m2 plots were surveyed in June
and July of 2012 and 2013 and the following information was recorded: vascular plant
species and biocrust morphological composition and percent cover, IWAG stem count in
0.04 m2 in the southeast corner of the plot, percent cover of litter, rock, and bare soil, and
soil depth (at 2 corners of the plot).
Biocrust morphology was split into two main categories: lichen and moss. Lichen
morphology was divided into six structure categories including crutose, areolate,
gelatinous, squamulose, foliose, and fruticose, and further categorized by color (white,
light and dark green, red, black, light and dark brown, grey, orange, and yellow). Moss
morphology was separated into two categories: tall and short (Rosentreter et al. 2007).
Data analysis
To determine which combination of treatments had the most effect on V. dubia
stem counts and percent cover, native and invasive percent cover and richness, and cover
99
of litter, short moss, tall moss, and squamulose white. I used 10 mixed linear model
analyses (one per dependent variable; PROC MIXED; SAS 9.3 software). Mixed linear
models were used because I had both random (transect nested within burn areas) and
fixed (treatment) variables. All treatments were included in these analyses: burn (B; n=9),
burn/herbicide (BH; n=9), burn/herbicide/trample (BHT; n=8), herbicide/trample (HT;
n=10), herbicide (H; n=9), trample (T; n=9), and control (C; n=12). For these analyses,
one representative plot from each control transect was randomly chosen and added to the
control group.
To analyze the interaction between fire and herbicide/trampling (when applied
together) on V. dubia stem counts and percent cover, native and invasive percent cover
and richness, and cover of short moss, tall moss, and squamulose white lichen, I used
nine mixed linear model analyses (one per dependent variable; PROC MIXED; SAS 9.3
software). I used mixed linear models because I had both random (transect nested within
burn areas) and fixed (treatment) variables. The following treatments were included; B,
BHT, HT, and C. For these analyses, one representative plot from each control transect
was randomly chosen and added to the control group.
I then compared the effects of burning alone to burning and herbicide together on
V. dubia stem counts and percent cover, native and invasive percent cover and richness,
and short moss, tall moss, and squamulose white using nine mixed linear model analyses
(one per dependent variable; PROC MIXED; SAS 9.3 software). I again used mixed
linear models because I had both random (transect nested within burn areas) and fixed
(treatment) variables. Only B and BH treatments were analyzed.
100
All data were tested for normality and homoscedasticity and transformed as
necessary using Box-Cox transformations.
I assessed changes in biocrust morphology composition after treatment and how
biocrust composition related to soil depth and percent cover of litter, rock, bare soil, total
biocrust, total lichen, total moss, and V. dubia, using nonmetric multidimensional scaling
(NMS) ordination with PC-Ord software (MjM Software Design 2006). For the NMS
ordination, I used a Relative Sorensen distance measurement, Varimax rotation with 50
runs on the real data, 15 iterations to evaluate stability, a stability criteria of 0.00001, an
initial step length of 0.2, and a maximum of 250 iterations.
From the NMS ordination, a distance matrix was created, where Relative
Sorensen similarity rate of change values were assigned to each plot. These values
indicate how much biocrust morphology has changed in each plot from 2012 to 2013.
With the Relative Sorenson values I determined the average similarity in biocrust
community composition for each treatment. I then compared the effect of treatment on
the change in biocrust community composition with a mixed linear model analysis
(PROC MIXED; SAS 9.3 software).
Results
Effects of Burning, Herbicide, and Trampling on Basalt Intermound Plant Communities
In 2012, V. dubia made up 95.2% of all IWAG cover (all other IWAG species
found had under 1.0% cover) in the basalt intermounds and averaged 66.7% cover per
plot, but in 2013 V. dubia’s cover decreased to 8.6% per plot despite all other IWAG
species cover remaining under 1.0%. On the intermounds, V. dubia stem counts and
percent cover where not affected by treatment type (Table 30; Figures 19 and 20
101
respectively), however, all fire treatments reduced litter cover (p=0.0004, F=4.88; Table
31; Figure 20). V. dubia change in stem counts and percent cover were Box-Cox
transformed with a final lambda of 1.5 and 0.5 respectively, however litter cover was not
normally distributed and normality was not improved by transformation. V. dubia stem
counts were not influenced by an interaction between burning and herbicide/trampling
(together), however herbicide/trample, irrespective of burning, reduced V. dubia stem
counts (Table 32; Figure 19) but not cover (Table 32; Figure 20). V. dubia percent cover
was Box-Cox transformed with a final lambda of 0.5. In burn plots, herbicide and fire had
similar effects as burning alone on V. dubia stem counts and cover (Table 33).
All treatments had similar effects on invasive and native species richness (Table
34; Figure 21), but native cover decreased more in burn/herbicide plots than
herbicide/trample plots (Tables 34 and 35; Figure 21). Invasive species richness was
Box-Cox transformed with a final lambda of 1.25. Native species richness was not
normally distributed but transformations did not achieve or improve normality. The
interaction between burning and herbicide/trampling did not affect invasive or native
species richness (Table 36; Figure 21). Burning and herbicide had similar effects on
invasive and native species richness as burning alone (Table 37; Figure 21).
From 2012 to 2013, invasive species cover decreased from 73% per plot to 19%,
in all treatments including the control (Figure 22); all treatments had similar affects on
invasive species cover (Tables 38 and 39). Native species cover increased from 9.1% per
plot in 2012 to 21.6% per plot in 2013, in all treatments including the controls. The three
most common native species in 2013 were present in over 82% of plots surveyed and all
increased their cover over the year, including Polygonum douglasii (Douglas’ knotweed;
102
increased from 1.87% to 4.18% per 1 m2), Poa secunda (Sandberg’s bluegrass; increased
from 2.31% to 5.88% per 1 m2), and Epilobium brachycarpum (tall annual willowherb;
increased from 1.21% to 5.16% per m2). Native species percent cover increased in in all
treatments, particularly those with herbicide applications (Tables 38 and 39; Figure 22).
The interaction between burning and herbicide/trampling did not affect invasive or native
species percent cover (Table 40; Figure 22), but herbicide and trampling, irrespective of
burning, increased native cover (Tables 40 and 41; Figure 22). Burning alone had similar
effects as burning and herbicide on invasive and native cover (Table 42; Figure 22).
Effect of Burning, Herbicide, and Trampling on Biocrust Community Composition
Biocrust morphology composition shares a strong relationship with litter (axis 1
r2=0.347) and rock cover (axis 1 r2=0.334; Figure 36), however V. dubia cover, bare soil
cover and soil depth do not strongly relate to biocrust morphology composition. Tall
moss is the dominant morphology type in plots in the top left corner, which shares a
strong relationship with litter cover (axis 1 r=-0.659, axis 2 r=0.688; Figure 36). The
plots in the center of the ordination tend to have more short moss cover, which shares a
strong relationship with rock cover (axis1 r=0.247; Figure 36), while plots on the right
tend to be squamulose lichen dominated (axis 2 r=0.869; Figure 36). When comparing
treatment type to the change in biocrust community composition there no noticeable trend
(Figure 37). Biocrust composition changed the least in control plots and the most with
trampling (Figure 38), but there was no significant difference between treatments (alpha
= 0.05, Table 43). Short moss, tall moss, and squamulose lichen cover were affected
similarly by treatments (Table 44). Short moss and squamulose lichen percent cover were
normally distributed and normality could not be achieved, but tall moss cover was Box-
103
Cox transformed with a final lambda of 0.75. The three dominate biocrust morphology
types (short moss, tall moss, and squamulose white) were not affected by an interaction
between burning and herbicide/trampling (when applied together; Table 45). Tall moss
cover was normally distributed but normality of short moss and squamulose lichen cover
were not achieved with transformation. Burning alone had the same affect on all
morphology types as burning with herbicide (Table 46). Short moss, tall moss, and
squamulose lichen cover were Box-Cox transformed with final lambdas of 0.75, 0.5, and
2.25, respectively.
Weather
In March 2012, there was 113.6 mm of precipitation (the highest March
precipitation in the last six year), however in 2013 there was only 24.8 mm (Figure 39).
March air temperature in 2012 was 3.4C compared to 3.7C in 2013 (Figure 40), which
may have influenced these study results.
Discussion
Biocrust research is fairly limited in comparison to the vascular plant
communities they are often associated with, but as IWAGs spread throughout the arid and
semiarid west, more aggressive restoration tools (i.e. fire and herbicide) will be required.
While others have documented the interactions between biocrust, IWAGs, and native
species (Ponzetti et al. 2007), as well as how to increase native cover with IWAG
removal techniques (DiTomaso et al. 1997, Simmons et al. 2007, Nyamai et al. 2011),
few have looked at how the biocrust responds to these tools (Belnap et al. 2001, Root et
al. 2011, O’Bryan et al. 2009), especially herbicide. The loss of biocrust to IWAGs, and
most likely IWAG removal techniques in arid habitats, may result in an alternate climax
104
species composition, which may not be as resistant to invasion, requiring more active
restoration management. The goal of my project was to increase community resistance to
invasion by understanding how biocrust relates to vascular plants, removing IWAGs,
increasing native cover, and determining how semiarid biocrust will respond to common
restoration techniques.
The relationship between vascular plants and the three main biocrust communities
found (tall moss, short moss, and squamulose lichen), appear to be consistent with what
others have described; moss species have higher cover closer to high litter and vegetation
cover (Clair et al. 1993, Maestre et al. 2011), while lichens are more prominent in open
areas (Clair et al. 1993). However, if vegetation is too dense then moss and lichen cover
will decrease (Eldridge 1993), which is problematic as IWAG species can form dense
monotypic stands, eventually shading out biocrust species. IWAGs can also alter the soil
biota (Belnap et al. 2005), nitrogen availability (Sperry et al. 2006), the fire cycle (Link et
al. 2006), and presence of native species (James et al. 2011), all of which can adversely
affect biocrust. Without biocrust, semiarid bunchgrass communities would lose a major
piece that makes the community as a whole resistant to further invasion. The gap in V.
dubia and biocrust literature has led to many questions, in particular why V. dubia’s
population exploded in the prairie at TNWR (especially in intermound areas), as well as
how semiarid biocrust would respond to this relatively new invader and to common
IWAG removal techniques.
V. dubia dominated the basalt intermounds in 2012, however by 2013 its cover
decreased in all treatments including the controls, which may be due to early spring
precipitation levels; while the March mean temperature was similar in 2012 and 2013,
105
there was about a fifth more rain in 2012 than 2013 (www.NOAA.gov Updated Nov.
2013). In 2012, the excessive rainfall may have triggered, or at least supported, the
population explosion. Yet March 2013 had the lowest average rainfall in the last seven
years, so the shallow soil’s inability to retain moisture may have increased drought stress
on V. dubia seeds and seedlings attempting to establish, but there is no research
indicating how this species would react to low spring precipitation. Other species adapted
to vernal pools can cope with inundation stress (Gerhardt and Collinge 2007), but are
more susceptible to drought (Volaire et al. 1998) and are generally not good competitors
(Gerhardt and Collinge 2007). Though vernal pools tend to have high abiotic stress
(Gerhardt and Collinge 2007), which generally lowers invasive cover (D'Antonio 1993,
King and Grace 2000, Gerhardt and Collinge 2003), there is often low native cover, so an
invasive species well adapted to inundation will thrive with the lack of competition
(Gerhardt and Collinge 2007). This could be why V. dubia did so well in previous years
and why native cover increased once competition with V. dubia decreased.
While most of the invasive species cover at in my study area at TNWR consisted
of V. dubia (even in 2013), invasive cover and richness declined across the board, but
were not affected differently by treatments, which could be due to the reduced
precipitation. However, native cover increased in every treatment, including the control,
suggesting there are native species that were not negatively affected by the lower spring
precipitation, fire, herbicide, or trampling, and they were able to take advantage of the
lack of invasive competition. The timing of my herbicide application targeted fall IWAG
germination but the native species were not actively growing, and similar to what others
have found, invasive cover decreased while native cover increased (Cox and Anderson
106
2004, Huddleston and Young 2005, Mittelhauser et al. 2011, Nyamai et al. 2011).
However, invasive cover decreased in the controls as well, suggesting that the drought
may have had more of an affect on invasive cover than treatment types. The timing of
application and the growth stage of native species is extremely important, as native
perennial seedlings are susceptible to herbicides (Monaco and Creech 2004), and cannot
outcompete IWAG seedlings (James 2008a). So with the reduction of IWAGs and
minimal herbicide exposure native species should be able to acquire more resources,
increasing their size and competitive advantage over future IWAG generations. However,
only a few studies have analyzed the effect of herbicide on biocrust communities and in
these arid habitats biocrust cover was reduced (Zaady et al. 2004, Maestre et al. 2011),
which is inconsistent with my findings. The herbicide treatments had similar effects on
biocrust morphology composition as the controls, implying that either the herbicide I
used specifically did not have an effect or it did but the higher rainfall and cooler winter
temperatures (compared to arid regions) reduced abiotic stress, allowing biocrust to heal.
Similar to the effects of a well timed herbicide application, prescribed fires at low
rates and intensities can increase native cover (Pollak and Kan 1998, Keeley et al. 2005),
as I found at TNWR. Prescribed fires that occur when IWAGs are growing but natives
are not (e.g. in fall), would reduce next season’s IWAG abundance and seed bank, litter
cover, future wildfire risk and increase nutrient availability for native species (Collins et
al. 2008). However, prescribed fires cannot be used often in the same areas, as native
species in this region are not adapted to yearly fires, so a combination of tools should be
used, with periodic prescribed fires that mimic the historic fire cycle (McIver and Starr
2001, Keeley 2006, James et al. 2010). In comparison, only a few studies have
107
determined the effect of fire on arid (Belnap et al. 2001, Root et al. 2011) and semiarid
biocrust (Ponzetti et al. 2007, O’Bryan et al. 2009), but only those done in arid habitats
found that fire reduced biocrust cover. In a higher rainfall arid habitat, the plow line was
more destructive than the fire itself (Ponzetti et al. 2007), and biocrust cover increased
after a fire in semiarid Australia, which is consistent with my findings (O’Bryan et al.
2009). My control plots had the least amount of change but as there was no difference
between any treatments, implying that the semiarid biocrust community at TNWR may
not be as susceptible to disturbance as arid biocrust communities. It is important to note
that there has not been a fire in this area of the prairie in over twenty years, and excessive
fires may increase stress on the remaining biocrust species, resulting in decreased cover.
Also, the extreme dry conditions may have swamped out treatment effects that could
have occurred in more typical years.
Although, at a relatively smaller scale than fire and herbicide, livestock grazing
can have a wide variety of effects on vascular plants and biocrust communities. If grazing
is left unattended then native perennial species abundance decreases and invasive cover
increases (DiTomaso 2000), but if prescribed grazing is applied, then target invasives can
be reduced while preserving the native species (Frost and Launchbaugh 2003). However,
grazing can also have detrimental effects on biocrust (Bowker et al. 2006, Chamizo et al.
2011), especially if heavy grazing occurs when the biocrust is dry or frozen (Clair et al.
1993, Belnap and Eldridge 2001). Grazers hooves can pull up or bury the biocrust, but if
crushed and not moved then the biocrust can still function as a seed barrier (Belnap and
Eldridge 2001). My trampling disturbance analyzed the effects of compression on the
biocrust community, which were relatively un-phased by human trampling. However,
108
trampling by humans (which are smaller in size) may not be directly comparable to
livestock trampling studies, as livestock are much heavier. However I can conclude that
the moist biocrust was not affected by human foot traffic in this short-term study. A longterm analysis is needed to evaluate long-term effects of these treatments.
Overall, native species percent cover increased in all treatments, the biocrust
community was not significantly affected by any treatment, and invasive cover was
reduced, though this may be due to reduced precipitation. Herbicide alone had little effect
on the biocrust morphology composition and increased native cover, however to remove
the extensive V. dubia litter layer, which is promoting IWAG establishment and
hindering biocrust and native seedling growth, I recommend that prescribed fires be used.
After a fire, herbicide can be used to control dense IWAG stands as they establish in fall.
This restoration plan, in conjunction with my recommendations for mounds (Chapter 2),
should promote native vascular growth and preserve biocrust cover, thereby increasing
community resistance as a whole, to future invasion. However, my findings represent
only the first year following treatment, and the community will undoubtedly change over
the long-term (Rinella et al. 2012, Rinella et al. 2009, Morris et al. 2009, Wilson and
Pinno et al. 2013). More monitoring should be completed to determine long-term effects
of fire and herbicide on semiarid biocrust and vascular plant community composition.
109
Chapter 3 Figures
0
Stem Count Change per m2
-50
-100
-150
A
-200
-250
A
A
-300
A
-350
-400
-450
A
A
A
Treatment
Figure 19 Mean change in V. dubia stem counts after treatment from 2012 to 2013.
Repeated letters indicate no significant difference. Error bars represent + 1 standard error.
110
60
40
20
Ventenata
0
Litter
-20
-40
-60
-80
-100
Treatment
Figure 20 Mean change in V. duia percent cover and litter cover after treatment from
2012 to 2013. Error bars represent + 1 standard error.
111
0.5
0
-0.5
-1
-1.5
-2
-2.5
-3
-3.5
-4
-4.5
Control
Herb/Tram
Trample
Herbicide
Burn/Herb/Tram
Burn/Herb
Burn
Control
Herb/Tram
N
Trample
Herbicide
Burn/Herb/Tram
Burn/Herb
Burn
Species Richness Change per m2
I
Treatment
Figure 21 The mean change in native and invasive species richness after each treatment
from 2012 to 2013. Error bars represent + 1 standard error. I is invasive and N is native.
112
Control
Herb/Tram
Trample
Herbicide
Burn/Herb/Tram
Burn/Herb
Burn
Control
Herb/Tram
N
Trample
Herbicide
Burn/Herb/Tram
Burn/Herb
Burn
I
Percent Cover Change per 1 m2
40
20
0
-20
-40
-60
-80
-100
Treatment
Figure 22 Mean change in native and invasive species percent cover after each treatment
from 2012 to 2013. Error bars represent + 1 standard error. I is invasive and N is native.
113
Tall Moss
Short Moss
SQ Lichen
Figure 23 NMS, joint plot Ordination of the 2012 and 2013 biocrust communities. Each
triangle represents an individual plot; distance between triangles represents similarity of
species composition among plots. The ordination axes were correlated to 2012 and 2013
soil depth, and percent cover of litter, bare soil, rock, total biocrust, and ventenata. Those
with Pearson correlations greater than 0.2 are shown as vectors; the length of the vector
lines indicates the strength of the correlation. Plots are grouped by treated or control and
year surveyed. Plots dominated by tall moss are within the red circle, short moss within
the blue circle, and squamulose (SQ) lichens within the black circle.
114
Figure 24 NMS ordination of the 2012 and 2013 biocrust communities. Black vectors
indicate the change community composition from 2012 to 2013. Plots are grouped by
treatment type and year surveyed.
115
Relative Sorenson Similarity
0.6
0.5
0.4
0.3
0.2
0.1
0
Treatment
Figure 25 Mean Sorenson similarity of biocrust morphology composition per 1 m2 after
each treatment from 2012 to 2013. Error bars represent + 1 standard error.
116
180
Total Precipitation (mm)
160
140
2007
120
2008
100
2009
80
2010
60
2011
40
2012
2013
20
0
Jan
Feb Mar April May June July Aug Sept Oct
Month
Nov Dec
Figure 26 Total monthly precipitation (mm) from 2007 to 2013 at the NOAA Spokane
Weather Forecast Office.
117
Mean Air Temperature (C)
25
20
15
2007
10
2008
2009
5
2010
0
2011
-5
2012
-10
2013
Month
Figure 27 Mean monthly air temperature (C) from 2007 to 2013 at the NOAA Spokane
Weather Forecast Office.
118
Table 29 The basalt intermound treatment combinations of burning, herbicide, and
trampling.
Treatments
1
2
3
4
5
6
Control
Burn
X
X
Herbicide
X
X
X
X
X
119
Trampling
X
X
X
ventenata stem counts changes and percent cover changes with treatment. Significant
values indicated in bold.
Effect
Treatment
Ventenata
Ventenata
Stem Counts
Percent Cover
F Value(DF)
Pr > F
F Value(DF)
Pr > F
2.19(78)
0.27
3.48(78)
0.78
120
Table 31 Summary of t and p values from Tukey post-hoc tests for a mixed linear model,
comparing the effect of treatment on the change in litter cover from 2012 to 2013.
Significant values indicated in bold.
Litter cover
Treatment
Treatment
t Value(DF)
Pr > |t|
B
B
B
B
B
B
BH
BH
BH
BH
BH
C
C
C
C
H
H
H
T
T
TBH
BH
C
H
T
TBH
TH
C
H
T
TBH
TH
H
T
TBH
TH
T
TBH
TH
TBH
TH
TH
0.05(56)
-2.31(56)
-2.03(56)
-1.99(56)
1.82(56)
-2.46(56)
-2.28(56)
-2.02(56)
-1.98(56)
1.73(56)
-2.43(56)
0.14(56)
0.12(56)
4.07(56)
-0.26(56)
-0.02(56)
3.72(56)
-0.37(56)
3.64(56)
-0.34(56)
-4.15(56)
0.96
0.024
0.047
0.052
0.074
0.017
0.026
0.048
0.053
0.090
0.018
0.89
0.91
0.0001
0.80
0.99
0.0005
0.71
0.0006
0.73
0.0001
121
ventenata stem counts changes and percent cover changes with post fire treatments.
Effect
Burn
Herbicide/Trample
Burn*Herbicide/Trample
Ventenata
Ventenata
Stem Counts
Percent Cover
F
F
Pr > F
Pr > F
Value(DF)
Value(DF)
3.32(35)
0.077
0.27(35)
0.61
4.21(35)
0.42(35)
0.52
0.048
0.44(35)
0.51
0.49(35)
0.49
122
ventenata stem counts changes and percent cover changes with herbicide applied after
fire. Significant values indicated in bold.
Effect
Burn*Burn/Herbicide
Ventenata
Ventenata
Stem Counts
Percent Cover
F Value(DF)
Pr > F
F Value(DF)
Pr > F
1.69(16)
0.21
0.65(16)
0.43
123
the change in native and invasive species richness with treatment. Significant values
indicated in bold.
Effect
Treatment
Invasive Species
Native Species
Richness
Richness
F Value(DF) Pr > F
F Value(DF)
Pr > F
0.96(59)
0.46
0.93(59)
0.48
124
Table 35 Summary of t and p values from Tukey post-hoc tests for two mixed linear
model analyses comparing the change in native and invasive richness per 1 m2 after
treatment. Significant values indicated in bold.
Invasive Species
Richness
Native Species
Richness
Treatment
Treatment
t Value(DF)
Pr > |t|
t Value(DF)
Pr > |t|
B
B
B
B
B
B
BH
BH
BH
BH
BH
C
C
C
C
H
H
H
T
T
TBH
BH
C
H
T
TBH
TH
C
H
T
TBH
TH
H
T
TBH
TH
T
TBH
TH
TBH
TH
TH
1.03(59)
1.51(59)
-0.20(59)
0.11(59)
-0.33(59)
0.26(59)
0.41(59)
-1.23(59)
-0.93(59)
-1.33(59)
-0.80(59)
-1.72(59)
-1.39(59)
-1.81(59)
-1.28(59)
0.31(59)
-0.14(59)
0.46(59)
-0.43(59)
0.15(59)
0.59(59)
0.31
0.14
0.84
0.92
0.74
0.80
0.69
0.22
0.36
0.19
0.43
0.090
0.17
0.076
0.21
0.76
0.89
0.65
0.67
0.88
0.56
0.80(59)
0.16(59)
0.50(59)
-0.50(59)
0.40(59)
-1.17(59)
-0.70(59)
-0.30(59)
-1.30(59)
-0.38(59)
-2.00(59)
0.38(59)
-0.70(59)
0.27(59)
-1.42(59)
-1.00(59)
-0.09(59)
-1.69(59)
0.89(59)
-0.66(59)
-1.55(59)
0.43
0.87
0.62
0.62
0.69
0.25
0.49
0.76
0.20
0.71
0.050
0.71
0.49
0.79
0.16
0.32
0.93
0.097
0.38
0.51
0.13
125
the change in native and invasive species richness with fire and post fire treatments.
Invasive Species
Native Species
Richness
Richness
Effect
Burn
1.86(35)
0.18
1.16(35)
0.29
Herbicide/Trample
1.06(35)
0.31
0.45(35)
0.51
0.38(35)
0.54
1.71(35)
0.20
126
the change in native and invasive species richness with herbicide applied after fire.
Effect
Burn*Burn/Herbicide
Invasive Species
Native Species
Richness
Richness
F Value(DF)
Pr > F
F Value(DF)
Pr > F
1.71(16)
0.21
0.69(16)
0.42
127
the change in native and invasive species percent cover with treatment. Significant values
indicated in bold.
Effect
Treatment
Invasive Species
Native Species
Percent Cover
Percent Cover
F Value(DF) Pr > F
F Value(DF)
Pr > F
0.88(59)
0.52
2.93(59)
0.014
128
Table 39 Summary of t and p values from Tukey post-hoc tests for two mixed linear
model analyses comparing the change in native and invasive percent cover per 1 m2 after
treatment. Significant values indicated in bold.
Invasive Species
Percent Cover
Native Species
Percent Cover
Treatment
Treatment
t Value(DF)
Pr > |t|
t Value
Pr > |t|
B
B
B
B
B
B
BH
BH
BH
BH
BH
C
C
C
C
H
H
H
T
T
TBH
BH
C
H
T
TBH
TH
C
H
T
TBH
TH
H
T
TBH
TH
T
TBH
TH
TBH
TH
TH
0.83(59)
0.52(59)
1.41(59)
0.02(59)
0.95(59)
1.72(59)
-0.38(59)
0.58(59)
-0.81(59)
0.14(59)
0.87(59)
1.00(59)
-0.49(59)
0.51(59)
1.32(59)
-1.39(59)
-0.43(59)
0.27(59)
0.92(59)
1.70(59)
0.70(59)
0.41
0.61
0.16
0.98
0.35
0.090
0.71
0.56
0.42
0.89
0.39
0.32
0.63
0.61
0.19
0.17
0.67
0.79
0.36
0.095
0.49
-2.51(59)
-0.59(59)
-2.34(59)
-1.38(59)
-2.55(59)
-3.20(59)
2.10(59)
0.18(59)
1.14(59)
-0.11(59)
-0.62(59)
-1.91(59)
-0.88(59)
-2.14(59)
-2.82(59)
0.96(59)
-0.28(59)
-0.80(59)
-1.21(59)
-1.79(59)
-0.49(59)
0.015
0.56
0.023
0.17
0.014
0.0022
0.040
0.86
0.26
0.92
0.54
0.061
0.38
0.036
0.0065
0.34
0.78
0.43
0.23
0.079
0.63
129
the change in native and invasive species percent cover with fire and post fire treatments.
Effect
Burn
Herbicide/Trample
Invasive Species
Native Species
Percent Cover
Percent Cover
F Value(DF) Pr > F
F Value(DF)
Pr > F
1.07(35)
0.31
0.58(35)
0.45
3.59(35)
0.067
14.27(35)
0.0006
0.04(35)
0.85
0.00(35)
0.97
130
Table 41 Summary of F and p values from a mixed linear model analysis comparing the
change in native species percent cover with fire and post fire treatments. Significant
values indicated in bold.
Native Species
Percent Cover
Treatment Treatment t Value(DF) Pr > |t|
C
HT
-2.82(35) 0.0078
C
B
0.59(35)
0.56
C
BHT
-2.14(35)
0.039
HT
B
3.20(35)
0.0029
HT
BHT
0.49(35)
0.63
B
BHT
-2.55(35)
131
0.015
the change in native and invasive species percent cover with herbicide applied after fire.
Effect
Burn*Burn/Herbicide
Invasive Species
Native Species
Percent Cover
Percent Cover
F Value(DF)
Pr > F
F Value(DF)
Pr > F
0.51(16)
0.49
3.43(16)
0.083
132
Table 43 F and p values of a mixed linear model analysis the effect of treatment on
Relative Sorenson distance matrix values, which were created by a nonmetric
multidimensional scaling ordination. The values indicate how much biocrust morphology
composition changes in each plot from 2012 to 2013.
Effect
Treatment
F Value(DF) Pr > F
0.42(78)
133
0.89
Table 44 Summary of F and p values from three mixed linear model analyses comparing
the changes of short moss, tall moss, and squamulose lichen cover with treatment.
Effect
Treatment
Short Moss
Tall Moss
0.73(19.7)
0.63
0.56(9.53)
0.75
134
Squamulose Lichen
F Value(DF) Pr > F
1.19(59)
0.32
the changes of short moss, tall moss, and squamulose lichen cover with post fire
treatments. Significant values indicated in bold.
Effect
Burn
Herbicide/Trample
(HT)
Burn*HT
Short Moss
Tall Moss
0.27(35)
0.61
0.4(35)
0.53
Squamulose Lichen
F Value(DF) Pr > F
0.01(35)
0.94
0.44(35)
0.51
0.27(35)
0.60
0.04(35)
0.84
0.49(35)
0.49
0.52(35)
0.48
2.57(35)
0.12
135
the changes of short moss, tall moss, and squamulose lichen cover with burning and
burn/herbicide treatments. Significant values indicated in bold.
Effect
Burn*Burn/Herbicide
Short Moss
Tall Moss
Squamulose Lichen
F Value(DF) Pr > F F Value(DF) Pr > F F Value(DF) Pr > F
0.08(16)
0.78
3.57(16)
0.077
1.11(16)
0.31
136
Conclusions and Management Recommendations
The main findings of my study at Turnbull National Wildlife Refuge were that 1)
Ventenata dubia increased 15-fold on mounds from 2009 to 2013, but in 2012 it was
most abundant in the basalt intermounds, along with biological soil crust (biocrust), 2)
herbicide was the only tool that reduced invasives, while fire increased natives, in
particular the seeded annual Madia gracilis, and 3) on the basalt intermounds V. dubia
abundance decreased about 60%, irrespective of treatment, native cover increased with
herbicide, and biocrust was not affected by treatments. These results can be used to
inform local land and refuge managers about the most effect ways to reduce invasive
winter annual grasses (IWAGs), increase native cover, and promote a self-sustainable
native plant community.
My long-term monitoring (Chapter 1) of IWAG abundance documented a 15-fold
increase in V. dubia stem counts control mound plots, while all other IWAG species
populations remained relatively constant. Despite the dramatic increase on mounds, V.
dubia was most abundant on basalt intermounds, as was biocrust. However, biocrust
diversity and lichen cover were highest in areas with the lowest V. dubia and litter cover.
In contrast, moss was most abundant when V. dubia and litter cover were high.
I found that on mounds (Chapter 2) herbicide was the only restoration tool that
effectively reduced IWAG stem count/cover and invasive richness/cover; prescribed
burning had no effect relative to the control. Native richness responded similarly to
invasive richness, but burning increased native cover, but this was similar to the controls.
Of the seeded species, M. gracilis cover increased most with burning, even when burning
137
was followed by herbicide, but Clarkia pulchella and the seeded native perennials did not
establish well in the first year.
On the basalt intermounds (Chapter 3), V. dubia cover decreased about 60%,
irrespective of treatment, which could be due to low March 2013 rainfall. Similarly,
invasive species richness/cover and native species richness decreased in all treatments,
including the controls. However, native cover increased in all treatments, especially those
with herbicide. The three species associated with this increase were Polygonum
douglasii, Epilobium brachycarpum, and Poa secunda. I also found there were three
main biocrust communities: Tall moss, which was associated with litter cover,
squamulose lichen, which was associated with open areas, and short moss was the
intermediate community. Herbicide, burning, and trampling did not significantly affect
biocrust community composition or cover of tall moss, short moss or squamulose lichens
relative to the control.
My findings can help land managers of similar habitat develop an integrated
management plan that would remove invasives, promote native growth (through assisted
succession), and increase community resistance to invasion. If prescribed burns mimic
the natural fire regime and Outrider® herbicide is used within reason the biocrust may
not be harmed. While these results are promising more research is needed to ensure these
results in fact sustained over long time periods, and until then treatments should be used
sparingly as overexposure could harm the biocrust community. Based on my findings, I
suggest that TWNR and similar areas develop a prescribed burn schedule based on the
historic fire regime, and only use herbicide in areas with dense IWAG growth. After
these have been applied, I recommend dispersing seed of M. gracilis and other aggressive
138
native annual species, followed by native perennial species once the native annual
abundance has increased. This plan should increase native cover, while reducing IWAG
abundance, and maintaining biocrust, which promotes a self-sustaining system.
139
Literature Cited
Antos, J. and C. Halpern. 1997. Root system differences among species: Implications for
early successional changes in forests of Western Oregon. Am Midl Nat 138:97–108.
Beckstead, J. and C. K. Augspurger. 2004. An Experimental Test of Resistance to
Cheatgrass Invasion: Limiting Resources at Different Life Stages. Biol Invasions
6:417–432.
Belnap, J. 2003. The world at your feet: desert biological soil crusts. Frontiers in Ecology
and the Environment 1:181–189.
Belnap, J. and D. Eldridge. 2001. Disturbance and recovery of biological soil crusts.
Biological soil crusts: structure, function, and management:363–383.
Belnap, J, R Rosentreter, S Leonard, JK Kaltenecker, J Williams, and D Eldridge (2001)
Biological Soil Crusts: Ecology and Management. USDI Bureau of Land
Management Technical Reference 1730-2, BLM/ID/ST-01/001+1730, Denver.
Belnap, J., S. L. Phillips, S. K. Sherrod, and A. Moldenke. 2005. Soil biota can change
after exotic plant invasion: Does this affect ecosystem processes? Ecology 86:3007–
3017.
Berg, A. W. 1990. Formation of Mima mounds: a seismic hypothesis. Geology 18:281–
284.
Blumenthal, Dana M., Nicholas R. Jordan, and Michael P. Russelle. 2003. "Soil Carbon
Addition Controls Weeds and Facilitates Prairie Restoration." Ecological
Applications 13: 605-615.
Boudell, J. A., S. O. Link, and J. R. Johansen. 2002. Effect of soil microtopography on
seed bank distribution in the shrub-steppe. Western North American Naturalist
62:14–24.
Bowker, M. A., J. Belnap, and M. E. Miller. 2006. Spatial modeling of biological soil
crusts to support rangeland assessment and monitoring. Rangeland Ecol Manag
59:519–529.
Boyd, D. and S. Rafael. 1996. Use of fire to control French broom 2:9–12.
Bretz, J. H. 1928. The channeled scabland of eastern Washington. Geographical Review
18:446–477.
Brown, C. S., V. J. Anderson, V. P. Claassen, M. E. Stannard, L. M. Wilson, S. Y.
Atkinson, J. E. Bromberg, T. A. Grant III, and M. D. Munis. 2008. Restoration
ecology and invasive plants in the semiarid west. Invasive Plant Science and
Management 1:399–413.
Bryant, J. A. M., Reynecke, B. K., and Brown, R.L. 2013. Geology and topography
effects on exotic plant distribution in a semi-arid Mima mound prairie in eastern
Washington. Northwest Science, 87: 12-23. Busby, R. R., M. E. Stromberger, G.
Rodriguez, D. L. Gebhart, and M. W. Paschke. 2012. Arbuscular mycorrhizal fungal
community differs between a coexisting native shrub and introduced annual grass.
Mycorrhiza 23:129–141.
Chamizo, S., A. Stevens, Y. Cantón, I. Miralles, F. Domingo, and B. Van Wesemael.
2011. Discriminating soil crust type, development stage and degree of disturbance in
semiarid environments from their spectral characteristics. European Journal of Soil
Science 63:42–53.
Clair, L. L. S., J. R. Johansen, and S. R. Rushforth. 1993. Lichens of soil crust
communities in the intermountain area of the western United States. Western North
140
American Naturalist 53:5–12.
Collins, B. M., J. D. Miller, A. E. Thode, M. Kelly, J. W. Wagtendonk, and S. L.
Stephens. 2008. Interactions Among Wildland Fires in a Long-Established Sierra
Nevada Natural Fire Area. Ecosystems 12:114–128.
Corbin, J. D. and C. M. D'Antonio. 2004a. Competition between native perennial and
exotic annual grasses: implications for an historical invasion. Ecology 85:1273–1283.
Corbin, J. D. and C. M. D'Antonio. 2004b. Can Carbon Addition Increase
Competitiveness of Native Grasses? A Case Study from California. Res Ecol 12:36–
43.
Cornelissen, JHC, TV Callaghan, JM Alatalo, A Michelsen, E Graglia, AE Hartley, DS
Hik, SE Hobbie, MC Press, and CH Robinson. 2001. Global change and arctic
ecosystems: is lichen decline a function of increases in vascular plant biomass?
Journal of Ecology 89:984–994.
Cox, G. W. 1990. Soil mining by pocket gophers along topographic gradients in a Mima
moundfield. Ecology 71:837–843.
Cox, R. D. and V. J. Anderson. 2004. Increasing native diversity of cheatgrass-dominated
rangeland through assisted succession. Rangeland Ecol Manag 57:203–210.
Crowe, E. A., A. J. Busacca, J. P. Reganold, and B. A. Zamora. 1994. Vegetation zones
and soil characteristics in vernal pools in the Channeled Scabland of eastern
Washington. Western North American Naturalist 54:234–247.
D'Antonio, C. M. 1993. Mechanisms controlling invasion of coastal plant communities
by the alien succulent Carpobrotus edulis. Ecology 74:83–95.
D'Antonio, C. M. and P. M. Vitousek. 1992. Biological invasions by exotic grasses, the
grass/fire cycle, and global change. Annual review of ecology and systematics
23:63–87.
Davies, K. W. and R. L. Sheley. 2007. A Conceptual Framework for Preventing the
Spatial Dispersal of Invasive Plants. Weed Science 55:178–184.
Davies, K. W. and R. L. Sheley. 2011. Promoting Native Vegetation and Diversity in
Exotic Annual Grass Infestations. Res Ecol 19:159–165.
Davies, K. W., J. D. Bates, and R. F. Miller. 2007. Short-term effects of burning
Wyoming big sagebrush steppe in southeast Oregon. Rangeland Ecol Manag 60:515–
522.
Davies, K. W., A. M. Nafus, and D. D. Johnson. 2013. Are Early Summer Wildfires an
Opportunity to Revegetate Exotic Annual Grass–Invaded Plant Communities?
Rangeland Ecol Manag 66:234–240.
Del Moral, R. and D. C. Deardorff. 1976. Vegetation of the Mima Mounds, Washington
State. Ecology 57:520–530.
DiTomaso, J. M. 2000. Invasive weeds in rangelands: species, impacts, and management.
Weed Science 48:255–265.
DiTomaso, J. M., G. B. Kyser, and M. S. Hastings. 1999. Prescribed burning for control
of yellow starthistle (Centaurea solstitialis) and enhanced native plant diversity.
Weed Science 20:233–242.
DiTomaso, J. M., M. L. Brooks, E. B. Allen, R. Minnich, P. M. Rice, and G. B. Kyser.
2009. Control of Invasive Weeds with Prescribed Burning. Weed Technology
20:535-548.
DiTomaso, J., E. Healy, D. Marcum, G. Kyser, and M. Rasmussen. 1997. Post-fire
141
herbicide sprays enhance native plant diversity. California Agriculture 51:6–11.
[DPR] California Department of Pesticide Regulation. 2008. Sulfosulfuron:1–8.
Eldridge, D. 2000. Ecology and management of biological soil crusts: recent
developments and future challenges. The Bryologist 103:742–747.
Eldridge, D. J. 1993. Cryptogams, vascular plants, and soil hydrological relations: some
preliminary results from the semiarid woodlands of eastern Australia. Western North
American Naturalist 53:48–58.
Evans, R. and J. Belnap. 1999. Long-term consequences of disturbance on nitrogen
dynamics in an arid ecosystem. Ecology 80:150–160.
Evans, R. A. and J. A. Young. 1970. Plant litter and establishment of alien annual weed
species in rangeland communities. Weed Science 18:697–703.
Facelli, J. M. and S. T. Pickett. 1991. Plant litter: its dynamics and effects on plant
community structure. The Botanical Review 57:1–32.
Facelli, J. M., P. Chesson, and N. Barnes. 2005. Differences in seed biology of annual
plants in arid lands: a key ingredient of the storage effect. Ecology 86:2998–3006.
Frost, R. A. and K. L. Launchbaugh. 2003. Prescription grazing for rangeland weed
management: A new look at an old tool to control weeds on rangelands. Rangelands
25:43–47.
Geisel, B. G. L., J. J. Schoenau, F. A. Holm, and E. N. Johnson. 2008. Interactions of
ALS-inhibiting herbicide residues in three prairie soils. Weed Science 56:624–627.
Gerhardt, F. and S. K. Collinge. 2003. Exotic plant invasions of vernal pools in the
Central Valley of California, USA. Journal of Biogeography 30:1043–1052.
Gerhardt, F. and S. K. Collinge. 2007. Abiotic constraints eclipse biotic resistance in
determining invasibility along experimental vernal pool gradients. Ecol Appl 17:922–
933.
Godínez-Alvarez, H., C. Morín, and V. Rivera-Aguilar. 2011. Germination, survival and
growth of three vascular plants on biological soil crusts from a Mexican tropical
desert. Plant Biology 14:157–162.
Gray, K. and J. Lichthardt. 2003. Field surveys for Silene spaldingii (Spalding's catchfly)
on the Lower Salmon River and Eagle Creek, Idaho:1–13.
Harper, K. T. and R. L. Pendleton. 1993. Cyanobacteria and cyanolichens: can they
enhance availability of essential minerals for higher plants? Western North American
Naturalist 53:59–72. 14 p.
Hill, J. P., C. J. Willson, and W. K. Smith. 2005. Enhanced photosynthesis and flower
production in a sagebrush morphotype associated with animal burrows. Plant Ecol
177:1–12.
Hironaka, M. 1961. The relative rate of root development of cheatgrass and medusahead.
J Range Manage 14:263–267.
Holt, J. S. and H. M. LeBaron. 1990. Significance and distribution of herbicide
resistance. Weed Technology:141–149.
Huddleston, R. T. and T. P. Young. 2005. Weed control and soil amendment effects on
restoration plantings in an Oregon grassland. Western North American Naturalist
65:507–515.
James, J. J. 2008a. Effect of soil nitrogen stress on the relative growth rate of annual and
perennial grasses in the Intermountain West. Plant Soil 310:201–210. 10 p.
James, J. J. 2008b. Leaf nitrogen productivity as a mechanism driving the success of
142
invasive annual grasses under low and high nitrogen supply. Journal of Arid
Environments 72:1775–1784. 10 p.
James, J. J., B. S. Smith, E. A. Vasquez, and R. L. Sheley. 2010. Principles for
Ecologically Based Invasive Plant Management. Invasive Plant Science and
Management 3:229–239.
James, J. J., T. J. Svejcar, and M. J. Rinella. 2011. Demographic processes limiting
seedling recruitment in arid grassland restoration. Journal of Applied Ecology
48:961–969.
Johansen, J. R., J. Ashley, and W. R. Rayburn. 1993. Effects of range fire on soil algal
crusts in semiarid shrub-steppe of the lower Columbia Basin and their subsequent
recovery. Western North American Naturalist 53:73–88.
Keeley, J. E. 2001. Fire and invasive species in Mediterranean-climate ecosystems of
California:81–94.
Keeley, J. E. 2006. Fire Management Impacts on Invasive Plants in the Western United
States. Conserv Biol 20:375–384.
Keeley, J. E. and T. W. McGinnis. 2007. Impact of prescribed fire and other factors on
cheatgrass persistence in a Sierra Nevada ponderosa pine forest. Int. J. Wildland Fire
16:96.
Keeley, J. E., M. Baer-Keeley, and C. J. Fotheringham. 2005. Alien plant dynamics
following fire in Mediterranean-climate California shrublands. Ecol Appl 15:2109–
2125.
Kennedy, A. C., B. N. Johnson, and T. L. Stubbs. 2001. Host Range of a Deleterious
Rhizobacterium for Biological Control of Downy Brome. Weed Science 49:792–797.
Kennedy, A., L. Elliott, F. Young, and C. Douglas. 1991. Rhizobacteria suppressive to
the weed downy brome. Soil Sci. Soc. Am. J 55:722–727.
Kerns, B. K., W. G. Thies, and C. G. Niwa. 2006. Season and severity of prescribed burn
in ponderosa pine forests: implications for understory native and exotic plants.
Ecoscience 13:44–55.
King, S. E. and J. B. Grace. 2000. The effects of soil flooding on the establishment of
cogongrass (Imperata cylindrica), a nonindigenous invader of the southeastern United
States. Wetlands 20:300–306.
Klaas, B. A., B. J. Danielson, and K. A. Moloney. 1998. Influence of pocket gophers on
meadow voles in a tallgrass prairie. Journal of mammalogy:942–952. 11 p.
Kremer, R. and A. Kennedy. 1996. Rhizobacteria as biocontrol agents of weeds 10:601–
609.
Kremer, R. J., M. F. Begonia, L. Stanley, and E. T. Lanham. 1990. Characterization of
rhizobacteria associated with weed seedlings. Appl. Environ. Microbiol. 56:1649–
1655.
Lass, L. and T. Prather. 2007. A scientific evaluation for noxious and invasive
weeds of the highway 95 construction project between the Uniontown cutoff
and Moscow. QuilaVision Inc., Missoula, MT. Online. Available:
http://northwestmedia.net/us95/pdf-final/Final-Weed- Report.pdf.
Levine JM, M Vilà, CM D Antonio, JS Dukes, K Grigulis and S Lavorel. 2003.
Mechanisms underlying the impacts of exotic plant invasions. Proceedings of the
Royal Society London B 270: 775-781.
Link, S. O., C. W. Keeler, R. W. Hill, and E. Hagen. 2006. Bromus tectorum cover
143
mapping and fire risk. Int. J. Wildland Fire 15:113–119.
Littell, RC, WW Stroup, and RJ Freund (2002) SAS for Linear Models. 4th ed. Cary, NC,
USA: SAS Institute.
Maestre, F. T., A. Escudero, I. Martínez, C. Guerrero, and A. Rubio. 2005. Does spatial
pattern matter to ecosystem functioning? Insights from biological soil crusts. Funct
Ecology 19:566–573.
Maestre, F. T., M. A. Bowker, Y. Cantón, A. P. Castillo-Monroy, J. Cortina, C. Escolar,
A. Escudero, R. Lázaro, and I. Martínez. 2011. Ecology and functional roles of
biological soil crusts in semi-arid ecosystems of Spain. Journal of Arid Environments
75:1282–1291.
Maestre, F. T., N. Martín, B. Díez, R. López-Poma, F. Santos, I. Luque, and J. Cortina.
2006. Watering, Fertilization, and Slurry Inoculation Promote Recovery of Biological
Crust Function in Degraded Soils. Microb Ecol 52:365–377.
Martínez, I., A. Escudero, F. T. Maestre, A. de la Cruz, C. Guerrero, and A. Rubio. 2006.
Small-scale patterns of abundance of mosses and lichens forming biological soil
crusts in two semi-arid gypsum environments. Aust. J. Bot. 54:339–348.
Mallory-Smith, C. A., D. C. Thill, and M. J. Dial. 1990. Identification of sulfonylurea
herbicide-resistant prickly lettuce (Lactuca serriola). Weed Technology:163–168.
Mangla, S. and R. M. Callaway. 2007. Exotic invasive plant accumulates native soil
pathogens which inhibit native plants. J Ecology 96: 58-67.
Marty, J. T. 2005. Effects of Cattle Grazing on Diversity in Ephemeral Wetlands.
Conserv Biol 19:1626–1632. 7 p.
Massey, FP, and SE Hartley. 2006. Experimental demonstration of the antiherbivore
effects of silica in grasses: impacts on foliage digestibility and vole growth rates.
Proceedings of the Royal Society B 273: 2299-2304.
McDonald, M. P., N. W. Galwey, and T. D. Colmer. 2001. Waterlogging tolerance in the
tribe Triticeae: the adventitious roots of Critesion marinum have a relatively high
porosity and a barrier to radial oxygen loss. Plant Cell Environ 24:585–596.
McIver, J. and L. Starr. 2001. Restoration of degraded lands in the interior Columbia
River basin: passive vs. active approaches. Forest Ecology and Management 153:15–
28.
McLendon, T. and E. F. Redente. 1992. Effects of nitrogen limitation on species
replacement dynamics during early secondary succession on a semiarid sagebrush
site. Oecologia 91:312–317. 6 p.
Miller, J. D., C. N. Skinner, H. D. Safford, E. E. Knapp, and C. M. Ramirez. 2012.
Trends and causes of severity, size, and number of fires in northwestern California,
USA. Ecol Appl 22:184–203.
Mittelhauser, J. R., P. W. Barnes, and T. G. Barnes. 2011. The Effect of Herbicide on the
Re-Establishment of Native Grasses in the Blackland Prairie. Natural Areas Journal
31:226–233.
Monaco, T. and J. Creech. 2004. Sulfosulfuron effects on growth and photosynthesis of
15 range grasses. J Range Manage 57:490–496.
Northam, F and R Callihan.1994. New weedy grasses associated with downy brome.
Monsen, SB and SG Kitchen, compilers. Proc. Ecology and management of annual
rangelands, Boise, ID. Ogden, UT: USDA-Forest Service, Intermountain Research
Station. GTR-INT-313:211–212.
144
Northam, FE, RR Old, and RH Callihan. 1993. Little lovegrass (Eragrostis minor)
distribution in Idaho and Washington. Weed Technology:771–775.
Nyamai, P. A., T. S. Prather, and J. M. Wallace. 2011. Evaluating Restoration Methods
across a Range of Plant Communities Dominated by Invasive Annual Grasses to
Native Perennial Grasses. Invasive Plant Science and Management 4:306–316.
O’Bryan, K. E., S. M. Prober, I. D. Lunt, and D. J. Eldridge. 2009. Frequent fire
promotes diversity and cover of biological soil crusts in a derived temperate
grassland. Oecologia 159:827–838.
Pollak, O. and T. Kan. 1998. The use of prescribed fire to control invasive exotic weeds
at Jepson Prairie Preserve:241–249.
Ponzetti, J. M., B. McCune, and D. A. Pyke. 2007. Biotic soil crusts in relation to
topography, cheatgrass and fire in the Columbia Basin, Washington. Bryologist
110:706–722.
Prasse, R. and R. Bornkamm. 2000. Effect of microbiotic soil surface crusts on
emergence of vascular plants. Plant Ecol 150:65–75. 11 p.
Prather, TS and V Steele. 2009. Ventenata control strategies found for forage producers.
University of Idaho Extensions: Idaho Impact Statements. 1-2.
Ray-Mukherjee, J., T. A. Jones, P. B. Adler, and T. A. Monaco. 2011. Immature Seedling
Growth of Two North American Native Perennial Bunchgrasses and the Invasive
Grass Bromus tectorum. Rangeland Ecol Manag 64:358–365.
Reddy, MN. 2010. Damping-Off of Festuca arizonica Caused by Fusarium. AJPS
01:104–105.
Reynecke, B. K. 2012. Plant Community Restoration on Mima Mounds at Turnbull
National Wildlife Refuge, WA. Master's Thesis, Eastern Washington University.
Ridenour, WM and RM Callaway. 2001. The relative importance of allelopathy in
interference: the effects of an invasive weed on a native bunchgrass. Oecologia
126:444–450.
Rimer, R and R Evans. 2006. Invasion of downy brome (Bromus tectorum L.) causes
rapid changes in the nitrogen cycle. Am Midl Nat 156:252–258.
Rinella, M. J., J. M. Mangold, E. K. Espeland, R. L. Sheley, and J. S. Jacobs. 2012.
Long-term population dynamics of seeded plants in invaded grasslands. Ecol Appl
22:1320–1329.
Rhoades, C., T. Barnes, and B. Washburn. 2002. Prescribed fire and herbicide effects on
soil processes during barrens restoration. Res Ecol 10:656–664.
Root, H. T. and B. McCune. 2011. Regional patterns of biological soil crust lichen
species composition related to vegetation, soils, and climate in Oregon, USA. Journal
of Arid Environments:1–8.
Root, H. T., J. E. D. Miller, and B. McCune. 2011. Biotic soil crust lichen diversity and
conservation in shrub-steppe habitats of Oregon and Washington. The Bryologist
114:796–812.
Rosentreter, R., D. Eldridge, and J. Kaltenecker. 2001. Monitoring and management of
biological soil crusts. Biological soil crusts: structure, function, and
management:457–468. 12 p.
Rosentreter, R., M. Bowker, and J. Belnap. 2007. A Field Guide to Biological Soil Crusts
of Western US Drylands: Common Lichens and Bryophytes.
Sheley, R. L., E. A. Vasquez, A.-M. Chamberlain, and B. S. Smith. 2012. Landscape-
145
Scale Rehabilitation of Medusahead (Taeniatherum caput-medusae)-Dominated
Sagebrush Steppe. Invasive Plant Science and Management 5:436–442.
Schwinning, S., O. Sala, M. Loik, and J. Ehleringer. 2004. Thresholds, memory, and
seasonality: understanding pulse dynamics in arid/semi-arid ecosystems. Oecologia
141.
Sheley, R. L., M. F. Carpinelli, and K. J. R. Morghan. 2007. Effects of Imazapic on
Target and Nontarget Vegetation during Revegetation. Weed Technology 21:1071–
1081.
Simmons, M. T., S. Windhager, P. Power, J. Lott, R. K. Lyons, and C. Schwope. 2007.
Selective and Non‐ Selective Control of Invasive Plants: The Short‐ Term Effects of
Growing‐ Season Prescribed Fire, Herbicide, and Mowing in Two Texas Prairies.
Res Ecol 15:662–669.
Steers, R. J. and E. B. Allen. 2009. Post-Fire Control of Invasive Plants Promotes Native
Recovery in a Burned Desert Shrubland. Res Ecol 18:334–343.
Simmons, M. T., S. Windhager, P. Power, J. Lott, R. K. Lyons, and C. Schwope. 2007.
Selective and Non‐ Selective Control of Invasive Plants: The Short‐ Term Effects of
Growing‐ Season Prescribed Fire, Herbicide, and Mowing in Two Texas Prairies.
Res Ecol 15:662–669.
Sperry, L. J., J. Belnap, and R. D. Evans. 2006. Bromus tectorum invasion alters nitrogen
dynamics in an undisturbed arid grassland ecosystem. Ecology 87:603–615.
Stevens, R. E. 2010. The use of weed technology in Palouse prairie remnants for
management and restoration. Master's Thesis Washington State University.
Tranel, P. J., D. R. Gealy, and A. C. Kennedy. 1993. Inhibition of downy brome (Bromus
tectorum) root growth by a phytotoxin from Pseudomonas fluorescens strain D7.
Weed Technology:134–139.
[USDA] U.S. Department of Agriculture. 2008. Ventenata Ventenata dubia (Leers) Coss.
USDA NRCS Aberdeen Plant Materials Center. 2011. Bluebunch Wheatgrass
(Pseudoroegneria spicata) Plant Guide:1–5.
Vasquez, E., R. Sheley, and T. Svejcar. 2008a. Nitrogen Enhances the Competitive
Ability of Cheatgrass (Bromus tectorum) Relative to Native Grasses. Invasive Plant
Science and Management 1:287–295.
Vasquez, E, R Sheley, and T Svejcar. 2008b. Creating Invasion Resistant Soils via
Nitrogen Management. Invasive Plant Science and Management 1:304–314.
Wainwright, C. E., E. M. Wolkovich, and E. E. Cleland. 2011. Seasonal priority
effects: implications for invasion and restoration in a semi-arid system. Journal of
Applied Ecology 49:234–241.
Volaire, F., H. Thomas, and F. Lelievre. 1998. Survival and recovery of perennial forage
grasses under prolonged Mediterranean drought: I. Growth, death, water relations
and solute content in herbage and stubble. New Phytologist 140:439–449.
Washburn, A. 1988. Mima Mounds. Report of Investigations 29:1–53.
Whisenant, S. G. and D. W. Uresk. 1990. Spring burning Japanese brome in a western
wheatgrass community. J Range Manage 43:205–208.
White, E. M., J. C. Wilson, and A. R. Clarke. 2006. Biotic indirect effects: a neglected
concept in invasion biology. Divers Distrib 12:443–455.
Young, J. A. and F. L. Allen. 1997. Cheatgrass and range science: 1930-1950. J Range
Manage 50:530–535.
146
Young, JA. 1992). Ecology and management of medusahead (Taeniatherum caputmedusae ssp. asperum Melderis). Great Basin Naturalist 52:245–252.
Young, JA and RA Evans. 1970. Invasion of Medusahead into the Great Basin. Weed
Science 18: 89-97.
Young, J. A., R. A. Evans, R. E. Eckert Jr, and B. L. Kay. 1987. Cheatgrass. Rangelands
9:266–270.
Young, K and J Mangold. 2008. Medusahead (Taeniatherum Caput-medusae)
Outperforms Squirreltail (Elymus Elymoides) Through Interference and Growth Rate.
Invasive Plant Science and Management 1:73–81.
Zaady, E., R. Levacov, and M. Shachak. 2004. Application of the Herbicide, Simazine,
and its Effect on Soil Surface Parameters and Vegetation in a Patchy Desert
Landscape. Arid Land Research and Management 18:397–410.
147
VITA
EDUCATION
Eastern Washington University, GPA 3.80
Expected Fall 2013
MS in Biology, Cheney, WA
Thesis: A Holistic Approach to Mima mound Prairie Restoration
Certificate in Geographic Information Systems (GIS)
Hofstra University, GPA 3.35
May 2011
BS in Biology concentration in Ecology and Evolution, Hempstead, NY
TEACHING
2012- 2013
Graduate teaching assistant, Cheney, WA
 Instructed laboratory courses in botany, introductory biology, field ecology,
plant physiology, and mycology
Guest Lecturer
Introduction biology
Oct 2013
Hofstra University
Teaching assistant, Hempstead, NY
 Instructed introductory biology lab
Jan-May 2011
PUBLICATIONS Anicito, K., B. Reynecke, and R. Brown In review. Ventenata (Ventenata
dubia) invasion in an eastern WA Mima mound prairie. Weed Science
PRESENTATIONS
Inland Empire Society of American Foresters Annual Meeting Nov 2013
Cheney, WA
Oral presentation: A Holistic Approach to Mima mound Prairie Restoration
98th Ecological Society of America Conference
August 2013
Minneapolis, MN
Oral presentation: Ventenata dubia invasion within a Mima mound prairie in
Eastern Washington
16th Annual Creative Works Symposium
May 2013
Eastern Washington University, Cheney WA
Oral presentation: Ventenata dubia invasion within a Mima mound prairie in
Eastern Washington
15th Annual Creative Works Symposium
Dec 2012
Eastern Washington University, Cheney, WA
Poster: Relationships between invasion, plant community composition, and
biological soil crust on and between Mima mounds at Turnbull National
Wildlife Refuge
Colonial Academic Alliance Undergrad. Research Conference April 2011
Hofstra University, Cheney, WA
Poster: Ixodes scapularis host preference and feeding sites on lizards
and small mammals
148
POSITIONS
Control of Mile-A-Minute Using Weevils
Jan 2011-May 2011
Field Technician, Hempstead, NY
 Designed and implemented study of how native plant communities
responded to the removal of Mile-A-Minute (Persicaria perfoliata) by the
weevil Rhinoncomimus latipes
Ecology of Lyme Disease Research Project
May 2010-Sept 2011
Field Technician, Aiken, SC and Pine Barrens, NJ
 Assisted with wildlife trapping and tagging (small and medium mammals
and reptiles)
 Performed vegetation surveys
 Collected ticks from wildlife and vegetation (by dragging and flagging)
Badlands Herbivore Diet Analysis
March 2011-May 2011
Lab Technician, Hempstead, NY
 Previously collected and dried fecal matter from several Badland
herbivores were broken, potted, and placed in the greenhouse.
Diamond Back Turtle Gender Identification Study
Jan-May 2010
 Prepared slides of Diamondback Terrapin Turtle gonads to sex
individuals in the hopes of finding identifying features to sex
individuals in the field
Shadowed Wildlife Manager
Field Technician, JFK Airport, Jamaica, NY
 Assisted with general bird surveys of airport grounds
 Performed laugh gull nest surveys
June 2010
Beach Grass Study
Nov 2009
 Measured growth by the surface area of each plant using a leaf area
meter. All plants were then dried and weighed to find the effect of
various treatments.
SKILLS
Vegetation:
Thesis research, Lyme Disease research, Mile-A-Minute research
 Surveys to estimate frequency and cover of vegetation within sample plots
 Monitored invasive species
 Tree measurements
 Pressing and keying plants
 Surveys to estimate frequency and cover of biological soil crust with
sample plots
 Designed and implemented invasive annual grass removal project using
fire, herbicide, and seed addition
 Designed and implemented invasive vine (Mile-A-Minute) project using a
weevil biocontrol
149
Small and mammals:
Ecology of Lyme Disease Research
 Extensive experience with Sherman, Tomahawk, and pitfall traps
 Mark-recapture using tagging method
 Species identification of live specimens
Lizard:
Ecology of Lyme Disease Research
 Extensive experience with pitfall and cover board traps
 Mark-recapture using branding method
 Species identification of live specimens
Data management:
Thesis research, Lyme Disease research
 Experience recording data by hand and with a Trimble
 Data entry
 Statistical analysis SAS 9.3 and PC-Ord
 Extensive mapping in ArcGIS
Other experience:
Lyme Disease research, Thesis research
 Operation of motion sensor-triggered cameras
 Mist-netting birds
 Supervised and instructed field technicians
 Soil sample collection
AWARDS
 EWU Graduate Assistantship – 2012-2013
 EWU High Demand Tuition Reduction- 2012-2013
 Academic Presidential Scholarship recipient (Hofstra U)-2007-2011
 Plumber’s Local 24 Scholarship-2007
 OSI pharmaceutical Summer Fellowship-2011
 EWU Biology Student Organization, Treasurer–2012-2013
150

A holistic approach to Mima mound prairie restoration

Transcription

Similar documents

Photographs showing Herbicide - Agent Orange Storage and the

www.ninsar.com

Kalustyan-Essential Oils.indd

Heart to Heart: Seed of Life Mandalas

Nuevas técnicas para el aseguramiento fitosanitario de plantas de

Product Sheet

IDENTIFICATION OF MINNESOTA INVASIVE

Common Burdock (Arctium minus)

spring news - Starke Ayres

MOBILE WASHING UNIT We come to you