The Makgadikgadi basin, a dynamic hydrological input and aeolian

The Makgadikgadi basin, a dynamic hydrological input and aeolian
output system.
Frank D. Eckardt a,* , Robert G. Br yant b, Warren W. Wood c, Kevin White d, Baruch Spiro e, Graham
McCulloch f
a Environmental and Geographical Scienc es, U niversity of Cape T own, P Bag X3, Rondebosch, 7701, South Africa
b Department of Geography, Uni versity of Sheffiel d, Dept. of Geography, Wi nter Street, Sheffiel d, S10 2TN, U K
c Department of Geol ogical Scienc es, Michigan State Uni versity, Eas t Lansi ng, Michigan 48824, USA
d Department of Geography, The Uni versity of R eading, Whiteknights, R eading, UK
e Department of Miner alogy, The Natural Histor y Mus eum, Cromwell Road, London SW7 5BD, UK
f Department of Z ool ogy, Uni versity of Dublin, Trinity College, Dublin 2, Irel and
* Correspondi ng author: frank.ec kardt@ .uct.ac.z a
This study focuses on modern day processes surrounding the Makgadikgadi pan basin system in NE Botswana. We
present results on the chemistry and dynamics of fluvial surface water inputs, examine the nature and chemistry of
various surface and subsurface evaporation products, provide data on the nature and chemistry of aeolian outputs
and demonstrate the impact of aerosol deposition in the downwind sector of the pans. During the wet season surface
water enters the saline pans and precipitates mostly calcite and halite, as well as dolomite and traces of other salts
associated with the desiccation of the lake. The hypersaline subsurface brine (up to TDS 190,000 mg/l) is
homogenous with minor variations due to pumping by BotAsh mine (Botswana Ash (Pty) Ltd), which extracts 2400 m 3
of brine/h from a depth of 38m. Average Pan surface sediments yielded 15 % Na and 22 % Cl which has been
identified in the downwind dust and could be traced as far as Johannesburg during a single dust episode in 2003.
Data for the 1980–2000 period suggest that dust loadings are intermittently influenced by the extent and frequency of
lake inundation, fluvial sediment influx, and surface wind speed variability. In addition, a significant proportion of the
observed variability in the dust and hydrological cycle of this source could be attributed to El Nino-Southern
Oscillation (ENSO) and Indian Ocean sea surface temperature anomalies. Detailed soil chemistry data suggests that
the pans produce over three million metric tons of chloride, sodium, and bicarbonate each year. In addition it was
noted that significant masses of naturally-occurring radioactive 226-radium and 238-uranium are present in the
aerosol load.
Keywords: Makgadikgadi, hydrology, brine, evaporites, aerosols
The Makgadikgadi Basin
A Dynamic
Hydrological input
and Aeolian Output System
Presented By
Frank Eckardt
EGS Department
University of Cape Town
Rondebosch South Africa
945 m
1500
Elevation (m)
2000
Gamsberg
Southern African Topography
Kalahari Basin
Windhoek
Escarpment
Ghanzi
1000
500
0
Kuiseb
Limpopo
600
M
ap
ut
o
Jo
ha
nn
es
b
ur
g
G
ab
or
on
e
Fr
an
ci
st
ow
n
S
er
ow
e
R
ak
op
s
200
G
ha
nz
i
G
ob
ab
is
W
itv
le
i
W
in
dh
o
ek
K
a
rib
ib
S
w
ak
op
m
un
d
Rainfall (mm)
Southern African Rainfall
800
Tropical
400
Semi-arid
0
Hyper-arid
A Dynamic System
June 2005
Florida Air Boat
April 2000
April 2008
August 2003
Regatta on Sua Pan
Bigg G - Department of Geography, University of Sheffield, Sheffield, UK
Bryant RG - Department of Geography, University of Sheffield, Sheffield, UK
Coetzee S - University of Botswana, Gaborone Botswana
Eng K - U. S. Geological Survey, Reston, Virginia , USA
Helmlinger M – NASA JPL, Pasadena, California USA
Irvine K - Department of Zoology, University of Dublin, Trinity College, Dublin, Ireland
Kraemer T U. S. Geological Survey, Reston, Virginia, USA
Nkala G - Botash Pty. Ltd, Sua Town, Botswana
Mahowald N - National Centre for Atmospheric Research, Boulder, Colorado, USA
McCulloch G - Department of Zoology, University of Dublin, Dublin , Ireland
Resane T - University of the Witwatersrand, Johannesburg, South Africa
Ross S - Department of Geography, University of Sheffield, Sheffield, UK
Spiro B Department of Mineralogy, The Natural History Museum, London
Vickery K - Environmental and Geographical Science, Uni of Cape Town, South Africa
Washington R – School of Geography, Oxford, UK
White K - Department of Geography, The University of Reading, Reading, UK
Wood WW - Department of Geological Sciences, Michigan State University,USA
Wood WW, Eckardt FD, Kraemer TF, Eng K.(in press) Quantitative eolian transport of
evaporate salts from the Makgadikgadi Depression (Ntwetwe and Sua Pans) in
northeastern Botswana. Implications for regional ground-water quality. In Sabka
Ecosystems Volume 3, Africa, Barth H-J, Böer B (eds) Kluwer: Dordrecht.
Eckardt FD, R.G. Bryant, G. McCulloch, B. Spiro, W. W. Wood (2008) The
hydrogeochemistry of a semi-arid pan basin case study: Makgadikgadi, Botswana
Applied Geochemistry, Volume 23, Issue 6, Pages 1563-1580
Bryant R.G. Bigg G., Mahowald N, Eckardt FD, and Ross S. (2007). Climate controls on
dust emissions from S African sources. Journal of Geophysical Research Atmospheres. Volume 23, Issue 6, June 2008, Pages 1563-1580
McCulloch, G., Irvine, K., Eckardt, FD. and Bryant, R. (2007).Hydrochemical fluctuations
and crustacean community composition in an ephemeral saline lake (Sua Pan,
Makgadikgadi Botswana). Hydrobiologia, vol. 596, no. 1, pp. 31-46
White K, Eckardt FD (2006) Geochemical mapping of the Makgadikgadi basin, Botswana
using moderate and high resolution remote sensing instruments, Earth Surface
Processes and Landforms, 31: 665-681
SAFARI 2000 Aerosols in Southern Africa
Drought
2000 Lake
Bryant R.G. Bigg G., Mahowald N, Eckardt FD, and
Ross S. (2007). Climate controls on dust emissions
from S African sources. Journal of Geophysical
Research - Atmospheres. Volume 23, Issue 6, June
2008, Pages 1563-1580.
Input
The Nata River is the most important source of
lake water in the Makgadikgadi
Bryant R.G. et al
Input
Flamingo Tracking
Brine Shrimp
& Water
Feed
Breed
Data via McCulloch
Input
Surface Water Samples
Inflow water 87 mg/l to 1,711 mg/l
• River Inputs and
• Lake Brine Evolution
Dec Lake samples 258 mg/l to 17,788 mg/l
April Lake samples 709 mg/l to 31,172 mg/l
North Lake (NL)
Middle Lake (ML)
Flood Water (FW)
South Lake (SL)
Eckardt FD et al.
From Input to Storage
• The soils in the catchment add much of the Ca, HCO3, Mg and K
• Inflow into Sua generally Ca–HCO3
• Transformation from calcium and bicarbonate rich water to sodium and chloride domination
• Mi xing of surface waters and recycling of surface salts
• Production wells and lake waters are tightly grouped in the Na and Cl sectors
• Na and Cl are most likely provided by marine atmospheric aerosols
Cations
Eckardt FD, R.G. Bryant, G. McCulloch, B. Spiro, W. W. Wood (2008) The hydrogeochemistry of a semi-arid
pan basin case study: Makgadikgadi, Botswana Applied Geochemistry, Volume 23, Issue 6, Pages 1563-1580
Anions
Storage
Subsurface Water Samples
• Subsurface Brine Wellfield
• Production and Monitoring Wells 38m depth
.
Wellfield (WF)
Ion log concentration (mmol/L) plotted against Cl
From Input to Storage
Average Flood Water (FW)
North Lake (NL)
Middle Lake (ML)
South Lake (SL)
Wellfield (WF)
.
Eckardt FD et al.
Na and K and
remain in solution
From Input to Storage
Ca and Mg are lost from solution as evaporation
products (calcite and dolomite formation)
Eckardt FD et al.
From Input to Storage
SO2 remains in solution
Eckardt FD et al.
The Earth Observatory
Storage &
Brine Product
• 98 wells operational since 1991
• brine 90,000-190,000 mg/l
• rate of 2400m3/h
• deposit 1 billion m3
• semi-confined clay aquifer
Evaporation Products
Sodium Carbonate Na2CO3
Sodium Sulphate Na2SO4
Sodium Chloride NaCl
Eckardt FD et al.
Storage
δ 34S average 34.35 (Sulphur Isotopes)
Makgadikadi
Namibia
• It appears that the major control on the variability of d34S is its source.
• Schroder et al. (2004) studied the d34S of sulfates in Oman across the Neoproterozoic Cambrian boundary
• Global marine value for that period is in the range 35–40‰.
• Further high values of d34S are reported from sulfates of Cambrian age elsewhere
• Claypool et al., (1980), Irkutsk Basin, 26.7–34.2‰, Amadeus Basin, 26.8–32.0‰,
• Peryt et al. (2005), East Siberian Basin, 22.6–34.5‰.
Eckardt FD et al.
87Sr/86Sr
Storage
(Strontium Isotopes)
Nata River
North Lake
Subsurface Brine
0.730000
0.728000
0.726000
0.724000
Min
Max
0.722000
0.720000
0.718000
e
Su
bs
ur
fa
c
So
ut
h
M
id
dl e
No
rth
M
os
et
se
Se
m
ow
an
e
Na
ta
La
ke
Fl
oo
d
0.716000
•Sr isotope for surface water from the Nata River and Lake are decoupled from the underlying brine
•The variability indicator of the heterogeneity of the sources at Sua Pan
•Subsurface water has a limited response to changes in surface processes.
Eckardt FD et al.
Southern Africa
Output
Burning Biomass
Peak Aerosol Emission in September
JAS TOMS Mean 1980-1999
Worlds Dustiest Place (Top11)
TOMS AI
Rainfall m m
Bodele D epressi on of s outh c entral Sahara
>30
17
West Sahar a in Mali and Mauritania
>24
5–100
Arabia (souther n Oman/Saudi border)
>21
<100
Eastern Sahara (Li bya)
>15
22
Southwes t Asi a (Makran c oas t)
>12
98
Takl amakan/Tarim Basin
>11
<25
Etos ha Pan (Namibia)
>11
435–530
Lake Eyre Basi n (Aus tralia)
>11
150–200
9 Makg adikg adi Basin (Botswan a)
>8
460
Salar de U yuni (Bolivia)
>7
178
Great Basi n (United States)
>5
400
6000 km2
10
9
Etosha
Makadikadi
Mea n Monthly TO MS A AI
8
7
6
5
4
3
2
1
0
jan
feb
mar
apr
m ay
jun
jul
aug
s ep
oct
nov
dec
Via Richard Washington
Output
• Historic stockpile – Loess soils (China), Bodele (Transport Ltd.)
Global Dust • Pan – Seasonal replenishment (Amadeus, Etosha) (Supply Ltd.)
Point Sources • Desertification (Dust Bowl 1934/Sahel 60-70‘s)
• Palaeolake margin – recent exposure (Aral Sea, Owens L.)
Meteosat Daily Dust 2005-2008
Temperature Brightness Product
“Pink Dust Method”
Kathryn Vickery - EGS MSc
MSG Brightness Temperature
MODIS True Colour
Output
Dust Samples
Dust
Quartz
Output
Si is the
prevailing
element
Sticky SEM Stub
Halite
NaCl is a
pan/evaporative
indicator
Clay
Calcium
Also found
Al, Fe, Ca as
well as K and
Mg
Eckardt FD et al.
Output
Dust Samples
Na
Cl
• Upwind (East)
• Downwind (West)
Ca
Pan surface samples yielded 15 % (Na) and 22 % (Cl)
Output
Dust Samples
25.0
Downwind concentrations range from
12 – 7 % (Na) and 14 to 9 % (Cl)
Gaborone dust is at 13 % Na and 19 % Cl.
Traces of NaCl in Joburg and Okavango Delta
20.0
15.0
Na
Cl
10.0
5.0
ne
sb
ha
n
Jo
O
ka
va
ng
o
ur
g
r
st
e
Ko
d
or
on
e
G
ab
in
nw
ow
D
ar
ch
pl
um
e
ce
M
S
ur
fa
an
P
P
an
n
S
id
e
Up
w
wi
nd
in
d
0.0
Pa
% concentration
Eckardt FD et al.
Long Range Transport
Output
Output
Too much water produces no dust
Bryant R.G. et al
Cyclone Eline
February 2000
The Earth Observatory
• Variations in rainfall entering the Nata catchment are consistent with SST controls.
• Close relationships between extreme regional rainfall events, flow in the Nata River
catchment, and subsequent flooding events on the surface of Sua Pan.
• Importance of rare landfall of tropical storms and depression
• Relationship between ENSO and dust
Bryant R.G. et al
Double Lake
Texas High Plains
Wood, W. W., and W. E. Sanford (1995)
Eolian Transport, Saline Lake Basins, and Groundwater Solutes, Water Resour. Res., 31(12), 3121–3129.
Deposition
Deposition
Soil Samples
• Soil Contribution to Floodwaters
• Pan Aerosol Deposition (Trace)
Terra MODIS Mixture Map
•Surface Carbonate Concentration
0%
100%
White K, Eckardt FD (2006) Geochemical
mapping of the Makgadikgadi basin, Botswana
using moderate and high resolution remote
sensing instruments, Earth Surface Processes
and Landforms, 31: 665-681
Wood WW, Eckardt FD, Kraemer TF, Eng K.(in press) Quantitative eolian
transport of evaporate salts from the Makgadikgadi Depression (Ntwetwe and
Sua Pans) in northeastern Botswana. Implications for regional ground-water
quality. In Sabka Ecosystems Volume 3, Africa, Barth H-J, Böer B (eds)
Kluwer: Dordrecht.
Deposition
Logarithmic concentration of soil leaches as a
function of logarithm of distance from the source
Solute footprint
140 km wide and 125 km long (17,500 km2)in size
Wood WW et al
Metric Tons Per Annum
Sodium
309,905
Bicarbonate
2,737,266
Chloride
176,615
Deposition
Naturally occurring
226-radium
0.0154
238-uranium
52,159
(Metric Tons Per Annum)
Wood WW et al
The Makgadikgadi Is Not A Complicated System
• Inflow into Sua generally Ca–HCO3 (Soils)
• Mixing of surface runoff and pan surface salts (NaCl)
• ENSO and cyclones influence lake hydrology
•Surface evaporation produces calcite, dolomite and halite
• Subsurface brine is similar but has bedrock contributions
• Dust depends on surface water recharge but limited by lake cover
• Dust rich in clays, salt and calcium carbonate
• Chemical footprint 17,500 km2
• Regional Dust Dispersal Possible
Pans – Soils, Duricrust, Groundwater
Chivas, A.R., Andrew, A.S., Lyons, W.B., Bird, M.I., Donnelly,
T.H., 1991. Isotopic constraints on the origin of salts
in Australian plays, 1. Sulphur. Palaeogeogr., Palaeoclimatol.,
Palaeoecol. 84, 309–332
Drake NA, Eckardt FD, White KH (2004) Sources of sulphur in
gypsiferous sediments and crusts and pathways of gypsum
redistribution in southern Tunisia. Earth Surface Processes and
Landforms 29, 1459-1471.
Risacher, F., Alonso, H., Salazar, C., 2003. The origin of brines
and salts in Chilean salars: a hydrochemical review. Earth Sci.
Rev. 45, 249–293.
4mm/yr or 4 L/m2yr of
Re charge
Solute
a
b
R2
RMSE
Area Under
Background
Average Conc.
Bulk Amount
Curve
Conc.
Over 125km
Deposited Metric Tons/yr
Sodium
341.8
-1.1305
0.816
0.285
790.902
1.9
4.427
310,000
Bicarbonate
1064.8
-0.6687
0.571
0.309
10825.476
47.5
39.104
2,740,000
Chloride
39.956
-0.2042
0.582
0.089
2190.206
15
2.522
177,000
226-Radium
2.3491
-0.3895
0.434
1.757
98.452
0.3
0.488
0.0154
238-Uranium
2.4142
-0.3939
0.347
1.848
104.230
0.275
0.559
52,000
y = a xb
Regression:
recharge flux of 4 mm/yr
y is the concentration
x is the distance from source (km)
a is the intercept
b is the slope
• Establishing a linear regression between the sample concentrations and the distance from the source
•The distance where the sampled concentrations reached a background value was identified as the upper limit
•The area beneath the curves was determined by integration from the closest site at 4 km to approximately 125 km
downwind. .
•Average concentrations was obtained by dividing the area underneath the curve by the effecti ve distance of transport
125 k
•The background concentrations were then subtracted from these average concentrations before any bulk estimates
were calculated.
Wood WW et al
The Makgadikgadi basin, a dynamic hydrological input and aeolian
output system.
Frank D. Eckardt a,* , Robert G. Br yant b, Warren W. Wood c, Kevin White d, Baruch Spiro e, Graham
McCulloch f
a Environmental and Geographical Scienc es, U niversity of Cape T own, P Bag X3, Rondebosch, 7701, South Africa
b Department of Geography, Uni versity of Sheffiel d, Dept. of Geography, Wi nter Street, Sheffiel d, S10 2TN, U K
c Department of Geol ogical Scienc es, Michigan State Uni versity, Eas t Lansi ng, Michigan 48824, USA
d Department of Geography, The Uni versity of R eading, Whiteknights, R eading, UK
e Department of Miner alogy, The Natural Histor y Mus eum, Cromwell Road, London SW7 5BD, UK
f Department of Z ool ogy, Uni versity of Dublin, Trinity College, Dublin 2, Irel and
* Correspondi ng author: frank.ec kardt@ .uct.ac.z a
This study focuses on modern day processes surrounding the Makgadikgadi pan basin system in NE Botswana. We
present results on the chemistry and dynamics of fluvial surface water inputs, examine the nature and chemistry of
various surface and subsurface evaporation products, provide data on the nature and chemistry of aeolian outputs
and demonstrate the impact of aerosol deposition in the downwind sector of the pans. During the wet season surface
water enters the saline pans and precipitates mostly calcite and halite, as well as dolomite and traces of other salts
associated with the desiccation of the lake. The hypersaline subsurface brine (up to TDS 190,000 mg/l) is
homogenous with minor variations due to pumping by BotAsh mine (Botswana Ash (Pty) Ltd), which extracts 2400 m 3
of brine/h from a depth of 38m. Average Pan surface sediments yielded 15 % Na and 22 % Cl which has been
identified in the downwind dust and could be traced as far as Johannesburg during a single dust episode in 2003.
Data for the 1980–2000 period suggest that dust loadings are intermittently influenced by the extent and frequency of
lake inundation, fluvial sediment influx, and surface wind speed variability. In addition, a significant proportion of the
observed variability in the dust and hydrological cycle of this source could be attributed to El Nino-Southern
Oscillation (ENSO) and Indian Ocean sea surface temperature anomalies. Detailed soil chemistry data suggests that
the pans produce over three million metric tons of chloride, sodium, and bicarbonate each year. In addition it was
noted that significant masses of naturally-occurring radioactive 226-radium and 238-uranium are present in the
aerosol load.
Keywords: Makgadikgadi, hydrology, brine, evaporites, aerosols