Content uploaded by Gary E Belovsky
Author content
All content in this area was uploaded by Gary E Belovsky on Jan 15, 2015
Content may be subject to copyright.
Available via license: CC BY 3.0
Content may be subject to copyright.
The Great Salt Lake Ecosystem (Utah, USA):
long term data and a structural equation approach
GARY E. BELOVSKY,
1,
DOYLE STEPHENS,
2,5
CLAY PERSCHON,
3,6
PAUL BIRDSEY,
3
DON PAUL,
3,7
DAVI D NAFTZ,
2
ROBERT BASKIN,
2
CHAD LARSON,
1,8
CHAD MELLISON,
4
JOHN LUFT,
3
RYAN MOSLEY,
3
HEIDI MAHON,
1
JAMES VAN LEEUWEN,
3
AND DAVI D V. ALLEN
2
1
Environmental Research Center, University of Notre Dame, Notre Dame, Indiana 46556 USA
2
United States Geological Survey, Salt Lake City, Utah 84119 USA
3
Utah Division of Wildlife Resources, Salt Lake City, Utah 84114 USA
4
United States Fish and Wildlife Service, Reno, Nevada 89502 USA
Abstract. Great Salt Lake (Utah, USA) is one of the world’s largest hypersaline lakes, supporting many
of the western U.S.’s migratory waterbirds. This unique ecosystem is threatened, but it and other large
hypersaline lakes are not well understood. The ecosystem consists of two weakly linked food webs: one
phytoplankton-based, the other organic particle/benthic algae-based.
Seventeen years of data on the phytoplankton-based food web are presented: abundances of nutrients (N
and P), phytoplankton (Chlorophyta, Bacillariophyta, Cyanophyta), brine shrimp (Artemia franciscana),
corixids (Trichocorixa verticalis), and Eared Grebes (Podiceps nigricollis). Abundances of less common species,
as well as brine fly larvae (Ephydra cinerea and hians) from the organic particle/benthic algae-based food web
are also presented. Abiotic parameters were monitored: lake elevation, temperature, salinity, PAR, light
penetration, and DO. We use these data to test hypotheses about the phytoplankton-based food web and its
weak linkage with the organic particle/benthic algae-based food web via structural equation modeling.
Counter to common perceptions, the phytoplankton-based food web is not limited by high salinity, but
principally through phytoplankton production, which is limited by N and grazing by brine shrimp.
Annual N abundance is highly variable and depends on lake volume, complex mixing given thermo- and
chemo-clines, and recycling by brine shrimp. Brine shrimp are food-limited, and predation by corixids and
Eared Grebes does not depress their numbers. Eared Grebe numbers appear to be limited by brine shrimp
abundance. Finally, there is little interaction of brine fly larvae with brine shrimp through competition, or
with corixids or grebes through predation, indicating that the lake’s two food webs are weakly connected.
Results are used to examine some general concepts regarding food web structure and dynamics, as well
as the lake’s future given expected anthropogenic impacts.
Key words: Artemia franciscana; brine shrimp; food webs; Great Salt Lake, Utah; hypersaline; phytoplankton; Podiceps
nigricollis; terminal lake; waterbirds.
Received 15 September 2010; revised 14 January 2011; accepted 17 January 2011; final version received 23 February 2011;
published 16 March 2011. Corresponding Editor: D. P. C. Peters.
Citation: Belovsky, G. E., D. Stephens, C. Perschon, P. Birdsey, D. Paul, D. Naftz, R. Baskin, C. Larson, C. Mellison, J. Luft,
R. Mosley, H. Mahon, J. Van Leeuwen, and D. V. Allen. 2011. The Great Salt Lake Ecosystem (Utah, USA): long term data
and a structural equation approach. Ecosphere 2(3):art33. doi:10.1890/ES10-00091.1
Copyright: Ó2011 Belovsky et al. This is an open-access article distributed under the terms of the Creative Commons
Attribution License, which permits restricted use, distribution, and reproduction in any medium, provided the original
author and sources are credited.
5
Deceased.
6
Present address: Riverdale, Utah 84405 USA.
7
Present address: AvianWest, Inc., Salt Lake City, Utah 84050 USA.
8
Present address: University of Texas, Arlington, Arlington, Texas 76010 USA.
E-mail: belovsky.1@nd.edu
vwww.esajournals.org 1March 2011 vVolume 2(3) vArticle 33
INTRODUCTION
We summarize results from 13 years (1994–
2006) of our ongoing study on the ecosystem of
the Great Salt Lake (Utah), a hypersaline terminal
lake, along with four years found in the literature
(1970–1994 ). Data collection focused on physical
characteristics of the lake, nutrients (nitrogen and
phosphorus), phytoplankton (Chlorophyta, Ba-
cillariophyta, Cyanophyta), brine shrimp (Arte-
mia franciscana) populations, and Eared Grebe
(Podiceps nigricollis) populations. Incidental pop-
ulation data on other species, such as brine fly
larvae (Ephydra cinerea and hians), corixids
(Trichocorixa verticalis), protozoans, dinoflagel-
lates, and copepods, are presented. These data
were employed to address hypotheses regarding
intra- and inter-annual variability in abiotic and
biotic effects on the food web involving nutrients,
phytoplankton, brine shrimp and Eared Grebes,
as well as incidental interactions with other
species. We hypothesize that this simple highly
productive food web is driven by the bottom-up
process of nutrient availability, not the stressful
conditions posed by extreme salinity and season-
al temperatures. Each trophic level is examined
using a structural equation approach (regression
models constructed a priori based on hypothe-
sized relationships).
Lack of long term and extensive data on the
Great Salt Lake became a concern due to the
Great Salt Lake’s potential environmental degra-
dation with urbanization along its shores, in-
creased agricultural runoff, decreased freshwater
inputs, mining of its waters for minerals, and
commercial harvesting of brine shrimp. In
addition, the commercial harvesting of brine
shrimp became a conservation issue, because
the lake’s large waterbird populations (one-third
of all western U.S. waterbirds nest or migrate
through Great Salt Lake: Paul and Manning 2002)
rely upon this abundant and relatively large
zooplankton, as well as brine fly larvae (Ephyra
spp.), to supply energy for spring and fall
migrations. Therefore, our study was initiated
in 1994 to monitor the brine shrimp and
waterbird populations, as well as the lake’s
general limnological dynamics.
The Great Salt Lake is the fourth or fifth largest
terminal saline lake in the world (depending on
metric): only the Caspian Sea (Russia and Iran),
Aral Sea (Kazakhstan and Uzbekistan), Lake
Balkash (Kazakhstan), which is partially fresh-
water, and Lake Urmia (Iran) are larger. Great
Salt Lake is the largest terminal lake in the
western hemisphere. Hypersaline lakes are inter-
esting because their harsh environmental condi-
tions, especially high salinity, lead to relatively
simple food webs and ecosystems, which allow
basic ecological concepts (e.g., top-down vs.
bottom-up control) to be more easily investigat-
ed. Yet the food webs and ecosystems are not so
simple (Boetius and Joye 2009) that they do not
reflect the dynamics shared by more complex
ecosystems. Therefore, hypersaline systems may
provide a model that combines some of the
simplicity of laboratory experiments and the
complexity of natural systems.
Saline, especially hypersaline, lakes and inland
seas have not received much attention from
ecologists for a number of reasons (Collins
1977, Hammer 1986). First, these ecosystems are
not considered common, which is true for North
America and Europe, but this is not the case
worldwide where saline lakes and inland seas
encompass nearly as much water volume as
freshwater lakes (1.04 310
5
versus 1.25 310
5
km
3
: Horne and Goldman 1994). Second, saline
lakes are often considered uninteresting because
of their simplicity due to the small number of
species that are able to cope with the stresses of
high salinity (Williams 1978, Williams et al.
1990). Finally, saline lakes are often considered
unimportant to people because they provide few
economic benefits, especially such as not having
potable waters, which makes their environmental
degradation of less concern to societies (e.g.,
destruction of Aral Sea: Pala 2006).
This is not to say that hypersaline environ-
ments have gone unstudied: recent microbial
work in hypersaline ponds (Yannarell and Paerl
2007, Yannarell et al. 2007, Paerl and Yannarell
2010) and concern over the conservation of saline
lakes (Williams 1993a,b, 1998, 2002) are notable.
However, after an extensive bibliographic com-
pilation of saline lake and inland sea literature
(.1200 papers dated prior to 1994, Larson and
Belovsky 1999), we found few studies available
vwww.esajournals.org 2March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
on the Great Salt Lake ecosystem and its biota
(e.g., Hayes 1971, Wirick 1972, Porcella and
Holman 1972, Stephens 1974, Brock 1975, Ste-
phens 1990, 1998, Post 1975, 1980, Stephens and
Gillespie 1976, Cuellar 1990, Collins 1977, Felix
and Rushforth 1977, 1979, 1980, Rushforth and
Felix 1982, Stephens and Birdsey 2002, Wurts-
baugh 1988, Wurtsbaugh and Berry 1990), and
these were short term and very limited in scope.
In addition, most of the extensive and long term
studies of other hypersaline lakes were for much
smaller lakes with very different environmental
patterns (e.g., Mono Lake, USA, a deep lake
basin: Dana et al. 1990, 1993, 1995; Lake Grass-
mere, NZ, a seaside lagoon lake: Wear and
Haslett 1986, 1987, Wear et al. 1986).
STUDY SYSTEM
Gwynn (1980, 2002) provides a detailed
descriptionoftheGreatSaltLake’shistory,
natural history and geology, as well as its
economic values. Great Salt Lake (Fig. 1) is a
hypersaline terminal lake that is the remnant of
Pleistocene Lake Bonneville. In historical times
the lake’s watershed has encompassed .89,000
km
2
. The lake’s salinity has ranged between 50
and 280 ppt (;5–28%) as the lake’s surface area
varied between 5490 and 2470 km
2
. At its highest
surface elevation (1284 m), the lake’s maximum
depth was 13.7 m; at its lowest surface elevation
(1277 m) the lake’s maximum depth was 7.6 m.
At the historical mean lake elevation of 1280 m,
the mean depth is only 5.5 m. Because it has a
large watershed, is shallow, and is a terminal
lake, Great Salt Lake is hypereutrophic.
The lake is composed of two major arms that
are now separated by a railroad causeway
through which some exchange occurs via cul-
verts and a more recently constructed breach
(Fig. 1). The South Arm has a lower salinity,
because 95%of the lake’s surface inflows are
located here (Bear, Weber, Ogden and Jordan
Rivers: Fig. 1), and contains a much more diverse
biota. The North Arm, because of its current
separation from most surface inflow by the
railroad causeway constructed in 1959 (Fig. 1),
has a salinity that is usually near halite saturation
(;26–28%) so that most of the biota is bacteria
and cyanophytes, and Dunaliella salina can be
present when salinities decline below saturation-
levels. The same railroad causeway crosses Bear
River Bay and maintains a much lower salinity
there due to the Bear River’s inflow so that the
biota in places contains fish. Another more recent
causeway partially isolates Farmington Bay in
the South Arm (Fig. 1) and maintains lower
salinity (;1–9%) there due to the Jordan River’s
inflow and higher nutrient concentrations due to
treated sewage inflows from Salt Lake, Davis and
Tooele Counties. A series of dikes in the Stans-
bury Basin in Carrington Bay (Fig. 1) were
constructed to enhance evaporation for salt
extraction from the lake. Therefore, the South
Arm today is more representative of the original
lake prior to separation of three of its bays and
the North Arm by causeways and dikes.
The South Arm of the lake’s biota has been
typically characterized as brine shrimp (Artemia
franciscana), two species of phytoplankton (Du-
naliella viridis and salina: Chlorophyta), two
species of brine fly (Ephydra cinerea and hians), a
corixid (Trichocorixa verticalis), and numerous
water birds. However, as our investigations of
the lake have progressed, the number of identi-
fied phytoplankton species (Chlorophyta, Bacil-
lariophyta, Cyanophyta, Dinophyta) and benthic
algae has increased to .60 species and several
species of rotifers, nematodes, ciliates, and
crustacean zooplankton have been found resi-
dent (Belovsky et al. 2000, Belovsky and Larson
2001, 2002, Larson 2004).
More than a third of all western US water birds
pass through the Great Salt Lake in their spring
and fall migrations, many of which are species of
conservation concern (e.g., Snowy Plover, Char-
adrius alexandrinus), and a number nest along the
lakeshore (Aldrich and Paul 2002); this makes the
lake of high conservation value. Most abundant
are the Eared Grebe (Podiceps nigricollis), three
common gulls (Franklin’s: Larus pipixcan, Cali-
fornia: L. californicus and Ring-billed: L. delawar-
ensis), and two common phalaropes (Red-necked:
Phalaropus lobatus and Wilson’s: P. tricolor) (Post
1975, 1980). The Eared Grebe is particularly
important not only given its abundance and
reliance on the lake as a migratory staging
location (.70%of all individuals of the species),
but because this species forages intensely on
brine shrimp (;90%of diet) to continue its
spring and fall migration (Caudell 2001, Conover
and Caudell 2009, Conover et al. 2009). Studies
vwww.esajournals.org 3March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
reporting lower consumption of brine shrimp
(Gafney 2008) were based on Eared Grebes
collected from areas of very low salinity in
Farmington Bay where brine shrimp were absent
(collected by J. Luft, personal communication).
However, the single species that most typifies
the ecosystem is the brine shrimp (Artemia
franciscana), because of its abundance and
uniqueness for saline lakes. The Great Salt Lake
brine shrimp is an Anastrocan crustacean zoo-
plankton that reaches up to 12 mm in length and
reproduces sexually via two modes: ovoviparity
which produces eggs that hatch in the ovisac and
the nauplii are released to the lake, and oviparity
which produces diapausing cysts. Ovoviparity is
the primary way that the population grows
during the summer. Cysts (oviparity) are the
means by which brine shrimp overwinter when
water temperatures are too cold for nauplii,
juvenile or adult survival, and they hatch in
spring when water temperatures rise and a
freshwater lens (less dense freshwater floating
on dense hypersaline water) appears on the lake.
The cysts of brine shrimp in the Great Salt Lake
float and at times accumulate in wind rows
called ‘‘streaks’’ . Cysts can remain in the lake
over the winter or can be deposited on beaches,
where they may or may not be washed or blown
back into the lake. Cysts deposited on the beach
can also serve as a ‘‘refuge’’ for brine shrimp
populations in areas that may dry up.
Brine shrimp cysts are commercially harvested
for the aquaculture industry, because they can be
hatched to provide a highly nutritious food for
larval fish and other crustaceans that are com-
mercially cultured. Floating and beach-deposited
cysts are easy to gather, and with minimal
treatment the cysts can be packaged, transported
and made ready for hatching. The commercial
harvest of cysts in the Great Salt Lake began in
Fig. 1. Satellite image of Great Salt Lake, Utah (USGS Earth Shots) identifies major areas and aspects of the
region. South Arm sampling sites in our study are indicated by blue dots.
vwww.esajournals.org 4March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
1952 by the Sanders Brine Shrimp Company; by
2000, the number of companies harvesting cysts
increased to 32, and since then the number of
companies has declined as a few large companies
bought up smaller ones (Sturm et al. 1980, Kuehn
2002). The annual economic value of this
industry has been estimated between US $50 –
100 million (Isaacson et al. 2002).
Based on available natural history, a simplified
food web diagram for the Great Salt Lake
ecosystem can be constructed (Fig. 2). We
hypothesize that the food web is actually
comprised of two weakly linked food chains:
one based on direct consumption of phytoplank-
ton (autotrophs), and the other based on con-
sumption of particulate detritus (detritivores)
and benthic algae. The data reported here is
restricted to the food web based on direct
consumption of phytoplankton (right of dashed
line in Fig. 2), because the commercial harvesting
of brine shrimp cysts was the initial concern of
Utah Division of Wildlife Resources (UDWR),
which funded the study.
METHODS
We report here on our sampling conducted
between 1994 and 2006. Preliminary nonrandom
sampling was conducted from June 1994 through
June 1995 by Wurtsbaugh (1995) and Gliwicz et
al. (1995) for UDWR; these data, which were
published by Wurtsbaugh and Gliwicz (2001),
are included in our analysis to increase the time
series. Furthermore, when available, earlier ob-
servations found in the literature were added to
the database to assess whether our observations
are consistent with earlier observations and to
assess whether major changes in the lake have
taken place.
Sampling sites
Fig. 1 presents our 1995–2006 sampling sites in
the South Arm of Great Salt Lake that were
located by GPS. Initial sampling in 1994 em-
ployed 10 sites that were more or less uniformly
distributed over the South Arm of the lake with
four sites in the shallow littoral zone, four sites in
the deepest areas (.7.9 m), and two sites at
intermediate depths (Gliwicz et al. 1995, Wurts-
baugh 1995, Wurtsbaugh and Gliwicz 2001).
With this initial sampling, a power analysis
indicated that 10 sites were adequate to ensure
standard errors no greater than 25%of the mean
at most times (Gliwicz et al. 1995).
We decided to further reduce standard errors
by sampling more than 10 sites starting in fall of
1995, which still met logistical limits to sampling.
In addition, the initial sampling had indicated
that standard errors could be reduced even more
by stratifying the lake into two regions: ,and 4
m depth, even though this constrained analyses
based on depth. Consequently, we randomly
selected 14 sites within the two strata using a
square kilometer grid system: seven sites were at
depths ,4 m; seven sites were at depths 4m.In
addition to the 14 random sites, one site north of
Fremont Island was selected nonrandomly to be
near the major inflow of freshwater (Bear River),
a region exhibiting high variability. In 1997, two
additional sites were nonrandomly added: one
east of Fremont Island to better sample the Bear
River estuary and one in Farmington Bay to
examine the inflow of freshwater (Jordan River)
and urban sewage there. In 1998, four additional
random sites were added (DWR 1–4), because
four original sampling sites were discontinued
when decreasing lake levels prohibited access to
shallow areas (Sites 2433, 2636, 3641, 3954): this
maintained the sampling of 17 random sites since
1997.
Sampling frequency
In 1994, 2–4 week sampling was employed
depending on weather to develop the power
analysis. From 1995–1996, each site was sampled
monthly. After 1996, each site was sampled
monthly from February–March, bimonthly from
April–September, and sometimes weekly from
October–January (monitoring for regulation dur-
ing the commercial harvest season for cysts). This
sampling regimen was not always achieved due
to inclement weather, which made it unsafe to be
on the lake; however, all sites were sampled
within three days of each other, weather permit-
ting. All measurements were averaged for all
sites and days within a weekly sampling period
to provide an average for that week. All weekly
averages within a month of a year were averaged
to provide a monthly average value for that year.
Measurements
At each site the following measures were
vwww.esajournals.org 5March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
made.
1. Lake elevation was obtained from the USGS
gage at Saltair Marina on Great Salt Lake.
2. Physical measures were collected at various
depths at each site. Salinity was measured at 1 m
below the surface and 1 m above the bottom
using an optical refractometer specific for NaCl
at 208C. Salinity measures were corrected for
water temperature using a regression developed
in the laboratory using a hydrometer. Photosyn-
thetically active radiation (PAR) was measured
using a LiCor LI-193SA spherical quantum
sensor in the air at the water surface and at 1-m
intervals from 1995–2000. Depth profiles of water
temperature and dissolved oxygen were mea-
sured at 1-m intervals using a YSI Model 59.
Secchi disc measures were taken at each site.
3. Nutrient concentrations were measured from
water samples taken at 5–11 sites on each
sampling date using a pump at 1 m below the
surface and 1 m above the bottom. At each depth,
two 500 ml samples were taken, filtered through
0.45lm filters and frozen. This provides 1 m
values for all sites sampled and .4 m values only
for sites deeper than 4 m. The water was
analyzed for ammonia, nitrite, nitrate (DIN)
and total dissolved phosphorus (TDP) at the
USGS National Water Quality Laboratory using
methods modified for high salinity (Fishman and
Friedman 1989), because hypersaline conditions
Fig. 2. Our hypothesized simple food web diagrams for the Great Salt Lake South Arm are presented. Two food
webs are proposed: a phytoplankton-based web (right of dashed line) and an organic particle/benthic algae web
(left of dashed line). Arrow thickness represents suspected relative interaction strengths. The phytoplankton-
based web was examined here.
vwww.esajournals.org 6March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
can distort chemical analyses (Fishman and
Friedman 1989).
4. Phytoplankton abundance were made using
two 500 ml water samples collected at 1 m below
the surface and passed through 153 lm mesh to
remove brine shrimp. One bottle was used to
measure chlorophyll concentration by fluorome-
try and the other to measure phytoplankton
species composition by cell count. These samples
were collected from 4–9 sites on each date.
To measure chlorophyll with fluorometry
(Turner Designs, Model 10AU), each 500 ml
water sample was filtered through a 0.8 lm glass
fiber filter (Whatman GF/F) and extraction was
conducted following Wetzel and Likens (1979)
with methanol in 1994–1995 (Gliwicz et al. 1995,
Wurtsbaugh 1995) and thereafter, with acetone.
Prior to 1999, concentrations of chlorophyll-a
(Chla) were corrected for phaeophytin-autilizing
the acidification procedure (Wetzel and Likens
1979), and after 1999, chlorophyll-awas directly
measured (Welschmeyer 1994). Chlorophyll
measures were unavailable for large portions of
2001–2002 and 2003.
To enumerate and identify phytoplankton
taxa, the second sample was immediately pre-
served using Lugol solution in 1994 (Gliwicz et
al. 1995); however, concern that Lugol solution
was less effective at the high salinities of Great
Salt Lake led Wurtsbaugh (1995) to employ 5%
sugar-formalin as a preservative in 1995. After
1995, Stephens and others kept the sample
refrigerated and in the dark, and enumeration
occurred within several days. Samples were
filtered through a 1.2 lm Millipore filter and
algal cells on the filter were re-suspended in 5 ml
of distilled water. If the sample was not
immediately enumerated, the distilled water re-
suspension was preserved in Lugol solution.
Aliquots of the sample were removed and placed
in a Palmer counting chamber for enumeration
(Palmer and Maloney 1954). Algal cells in the
Palmer cell were identified and counted along
one or two transects at 4003, rather than 1603,
magnification under a microscope with interfer-
ence phase contrast. Counting at higher magni-
fication (4003) facilitates species identification,
especially for Bacillariophytes, and provides a
more representative suite of the phytoplankton
from the pumped water sample than from
plankton net samples. The mean counts per
transect were converted to cells per liter using
the appropriate Palmer cell multiplication factor.
Cell counts were available for 1995–2002 and
2005–2006, as conducted by the laboratories of
Dr. S. Rushforth at the Botany Department of
Brigham Young University (1995–1999), of Dr. G.
Belovsky at the Department of Fisheries and
Wildlife, Utah State University or Department of
Biology, University of Notre Dame (1999–2002),
or Great Salt Lake Ecosystem Project Laboratory
(GSLEP), UDWR (2005–2006). Cell counts were
converted to phytoplankton biovolume based on
cell shape and dimensions.
5. Brine shrimp were sampled using a 0.5 m
diameter 32 m plankton net (153 lm mesh). This
was the largest net size that could be employed
by hand; a large net is needed given times of the
year when brine shrimp are not very abundant
and a large net reduces sample variation (Wiebe
and Holland 1968). Previous studies of brine
shrimp in Great Salt Lake and other hypersaline
lakes have used 80–250 lm mesh (larger number
is coarser mesh: Dana et al. 1990, Wear and
Haslett 1987, Wurtsbaugh and Berry 1990,
Wurtsbaugh and Marcarelli 2004). Quality con-
trol results indicated that a 153 lm mesh did not
lose any nauplii or cysts, while a 183 lm mesh
lost up to 6%.
The net was lowered to the lake bottom and
raised vertically to the surface at a rate of
approximately 1 m/s. The depth of the lake
bottom was recorded and was used to compute
the volume of water sampled (depth 3net mouth
area). Three tows were made at each site. The
computed volume was calibrated in 1995 using a
non-reversing TSK Model 901 flow meter, indi-
cating a sampling efficiency of 87%by volume
and this did not change between the first and
third tow. Therefore, as expected using a net with
a 4:1 length to mouth ratio and relatively large
mesh size, little sampling loss due to back
pressure was indicated.
Each tow’s brine shrimp sample (500ml) was
preserved in buffered formalin in the field and
then transferred to 70%ETOH until they could
be counted. Numbers of brine shrimp were
counted in 10 ml subsamples under a dissecting
microscope. Subsamples were examined until a
minimum of 50 adults and juveniles were
counted. Cysts (unhatched and hatched) and
nauplii were counted in a single 5 ml subsample.
vwww.esajournals.org 7March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
Nauplii were defined as 1.5 mm; juveniles were
.1.5 mm that were not sexually mature (females
without ovisac and males without claspers);
adults were sexually mature (.6 thorocopods).
Adult females were examined to determine
whether they had eggs or cysts in their ovisac.
Counts were extrapolated to numbers per liter
baseduponeachtow’s volume. This was
repeated three times for each tow and the values
were averaged. In addition, brine shrimp bio-
mass was computed in terms of adult equivalents
(nauplii ¼0.13 adults, juveniles ¼0.63 based on
live-mass).
6. Other taxa were identified incidentally in the
phytoplankton (protozoans, dinoflagellates,
nematodes, rotifers, copepods and cladocerans)
and plankton net samples (brine fly larvae and
corixids) when encountered.
7. Eared Grebe populations were estimated in the
fall of all years (except 2002) using a photo
survey of transects of the Great Salt Lake to
project the population for the entire lake (Paul et
al. 1999a,b). Transects were flown with a Cessna
185 using in-plane GPS and altimeter to gauge
area size covered by photos. Grebes were
identified in photos and enumerated to account
for density.
8. Brine shrimp cyst commercial harvesting
(metric tons per year) was estimated by UDWR
using random surveys of harvest boats returning
to harbor and legally mandated reports filed by
the harvest companies.
Statistics
All analyses were conducted using the month-
ly average values for each year using SYSTAT
(version 13), Systat Software, Inc., Chicago,
Illinois, USA. Proportions were arcsine-square
root transformed to obtain normality and hetero-
scedascity was minimized using log
10
-trans-
forms when necessary. Temporal variations in
survey data were analyzed using GLM to assess
whether observed inter- (among year) and intra-
(monthly) annual fluctuations were statistically
significant, where month and year were treated
as blocks. Specific hypotheses were tested using
t-test and regression.
Dynamics of the phytoplankton-based food
web were investigated using structural equation
modeling with our survey data (Grace 2006,
Hampton and Schindler 2006, Shipley 1999).
Structural equation modeling is more powerful
than standard multiple regression or path anal-
ysis approaches that attempt to fit unspecified
linear models to observations (Shipley 2000).
Structural equation modeling is accomplished by
first hypothesizing relationships, then depicting
the relationships in an equation, and finally using
regression methods to fit the equations to data to
determine whether the proposed hypotheses
may be operating. Inclusion of an independent
variable required its partial correlation to have a
p,0.15 tolerance, and the Akaike Information
Criteria index for the model including the
variable must be less than the model without it.
Both linear and non-linear equations were
employed, depending on the specific hypothe-
sized relationship. Least square and robust (least
median square) functions were employed to
minimize error, and both were examined for
consistency. Robust regressions were used, be-
cause least square regressions are sometimes
strongly impacted by outliers (Gotelli and Ellison
2004), and outliers were often evident in our
data.
RESULTS
Here we summarize our Great Salt Lake
dataset (Supplement) and examine inter- and
intra-annual patterns. This database is the foun-
dation necessary for testing specific hypotheses
regarding the phytoplankton-based food chain as
developed in the Discussion.
Abiotic measures
Depth profiles in temperature and salinity at the
same time in two years are presented for a deep
and shallow site as an example (Fig. 3). These
data were selected because they represent typical
patterns observed across all sites. The deep site
(Site 3510) in both years demonstrated tempera-
ture stratification (i.e., thermocline) and demon-
strated salinity stratification (i.e., chemocline) in
2005, but not in 2006. The shallow site (Site 2935)
did not demonstrate a chemocline or thermocline
in either year (2005, 2006); however, a weak
thermocline can be seen to be developing in 2006,
which may be a diel condition due to calm
waters. Average PAR with depth is presented for
March (peak phytoplankton abundance) and
June (lowest phytoplankton abundance) in 1998
vwww.esajournals.org 8March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
(Fig. 4A, B). PAR at both times was reduced to
less than 200 lEm
2
s
1
, the light intensity
required for good algal growth (Horne and
Goldman 1994), at a depth between 3–5 m;
therefore, the photic zone may be generally
defined as the first 4 m of the water column.
Within year (Fig. 5A, B, and C) and among year
patterns (1994–2006: Fig. 5D, E, and F) of lake
elevation, water temperature (1 m), and salinity for
Great Salt Lake are presented. While lake elevation
varied both within (GLM: F ¼18.20, df ¼11, 132,
p,0.001) and among years (GLM: F ¼252.01, df
¼12, 132, p ,0.001), among year variation was
far more variable than within year variation.
Annually, lake elevation was greatest from April
through June. While salinity varies both within
(GLM: F ¼6.85, df ¼11, 117, p ,0.001) and
among years (GLM: F ¼137.39, df ¼12, 117, p ,
0.001), among year variation was far more
variable than within year variation. Mean annual
salinity during the study varied from 8.8–16.3%.
Annually, salinity is lowest from May–July
during spring runoff. While water temperature
at 1 m varied both within (GLM: F ¼269.76, df ¼
11, 118, p ,0.001) and among years (GLM: F ¼
3.74, df ¼12, 118, p ,0.001), within year
variation was far more variable than among year
variation. Mean surface water temperatures
varied by 7.68C among years. Annual mean
water temperature was greatest in July–August
(25.88C) and lowest in January–February (1.58C).
PAR measurements were only available for 4
full years (1996–1999). There was no significant
annual variation in PAR at the surface (GLM: F ¼
1.34, df ¼3, 28, p ,0.28); however, there was
significant variation among months (GLM: F
3.11, df ¼11, 28, p ,0.008) with October through
January receiving the least PAR. Annually, 73%
Fig. 3. Plots of water temperature (red) and salinity (blue) with depth in the Great Salt Lake South Arm are
presented for a deep and shallow site in April/May for two years (2005, 2006). In 2005, the deep site exhibits the
deep brine layer (chemocline), but not in 2006. Both years exhibit a thermocline at the deep site, but never at the
shallow site.
vwww.esajournals.org 9March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
of maximum PAR is received and during October
through January, 65%is received. Light avail-
ability/turbidity as measured by Secchi depth
transparencies (Fig. 6A, D) varied both within
(GLM: F ¼14.73, df ¼11, 103, p ,0.001) and
among years (GLM: F ¼4.50, df ¼12, 103, p ,
0.0001). Secchi transparencies were greatest in
July (averaging 3.4 m) and least in January
(averaging 1.5 m).
Nutrient measures
Photosynthesis occurs year-round in Great Salt
Lake, in fact, peak phytoplankton abundances
occur in late fall to early spring when brine
shrimp grazing is absent. When phytoplankton
are at their lowest abundances (summer), nutri-
ents are tied up in the shrimp population.
Therefore, measures of lake nutrient concentra-
tions do not represent nutrient pools, but unused
nutrients available to phytoplankton and benthic
algae.
Within year patterns of lake nutrient concen-
trations (dissolved inorganic N: DIN and total
dissolved P: TDP) at shallow (1 m) and deep (.4
m) depths are presented in Fig. 6B and C. Among
year patterns (1994–2006) of lake nutrient levels
(DIN and TDP) at shallow (1 m) and deep (.4m)
depths are presented in Fig. 6E and F. Nitrate and
nitrite values of N were always below detection
values and therefore, ammonia was the principal
form of inorganic N. Shallow DIN did not vary
significantly within years (GLM: F ¼1.36, df ¼11,
58, p ,0.21), but varied among years (GLM: F ¼
7.38, df ¼9, 58, p ,0.001). Deep DIN varied
within years (GLM: F ¼3.27, df ¼11, 47, p ,
0.011) and among years (GLM: F ¼82.69, df ¼9,
47, p ,0.001). Deep DIN was lowest from
February through May. Shallow TDP did not
vary within years (GLM: F ¼0.75, df ¼11, 54, p ,
0.69), but varied among years (GLM: F ¼4.72, df
¼8, 54, p ,0.001). Deep TDP did not vary within
years (GLM: F ¼1.15, df ¼11, 43, p ,0.35) or
among years (GLM: F ¼3.12, df ¼8, 43, p ,
0.007). Nutrient measures under hypersaline
conditions can be distorted and require special
techniques (Fishman and Friedman 1989), which
were employed; nonetheless, we still view the
values as relative, rather than absolute, changes
over time.
Dissolved oxygen
Dissolved oxygen measures are difficult to
make in hypersaline water and DO correction
tables for salinity are only available up to 67,000
lS/cm (Lewis 2006). Therefore, sensor measures
were viewed as relative rather than absolute
values and were not as regularly taken as other
parameters (Fig. 7). For the same sites and dates
reported for temperature and salinity depth
profiles, DO (%saturation: Fig. 7A, C) declined
with depth as expected and approached anoxia
only at greatest depths, especially in 2005 when a
chemocline re-emerged at Site 3510. Average DO
measures for all sites on a given date (Fig. 7B, D)
varied within years (GLM: F ¼3.36, df ¼10, 12, p
Fig. 4. Plots of average photosynthetically active
radiation (PAR: lEm
2
s
1
) with depth in 1998 for
March (peak phytoplankton abundance) and June
(lowest phytoplankton abundance).
vwww.esajournals.org 10 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
,0.025) and among years (1994 –2004, GLM: F ¼
7.41, df ¼5, 12, p ,0.002).
Biotic measures
Phytoplankton.—Abundance of Chla (Fig. 8A,
C) varied approximately to an equal degree
within years (GLM: F ¼3.77, df ¼11, 93, p ,
0.001) and among years (GLM: F ¼2.54, df ¼12,
93, p ,0.006). Chla was greatest between
November and April, with January–February
providing peak levels. Peak Chla levels were
greatest in 1998 (228 lg/l), more than seven times
Fig. 5. Basic physical measurements of the Great Salt Lake South Arm are presented for variations within year
(intra-annual) by month and among years (inter-annual). Average lake elevation (A, D), salinity (B, E), and water
temperature (C, F) observations are presented for the period of our study (1994–2006) along with standard errors.
vwww.esajournals.org 11 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
greater than in 1995, 2000, 2002 and 2003. The
next highest Chla levels occurred in 1997 (114 lg/
l).
Phytoplankton composition for all years except
2003–2004, when enumeration was discontinued
by UDWR, (Chlorophyta, Bacillariophyta, Cya-
nophyta: Fig. 8B, D) did not vary within years
(GLM respectively: F ¼1.09, df ¼11, 49, p ,0.39;
F¼1.41, df ¼11, 49, p ,0.20; F ¼0.70, df ¼11, 49,
p,0.74), but varied dramatically among years
Fig. 6. Measures related to phytoplankton resource requirements in the Great Salt Lake South Arm are
presented for variation within year (intra-annual) by month and among years (inter-annual). Average
transparency (Secchi Depth: A, D), photic zone (1 m) DIN (red) and TDP (blue) concentrations (B, E ) and
deep (.5 m) DIN (red) and TDP (blue) concentrations (C, F) are presented for the period of our study (1994–
2006) along with standard errors.
vwww.esajournals.org 12 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
(GLM respectively: F ¼5.88, df ¼10, 49, p ,
0.000009; F ¼1.41, df ¼10, 49, p ,0.000002; F ¼
4.58, df ¼10, 49, p ,0.0001). Lowest Chlor-
ophyte abundances occur from July through
November, lowest Bacillariophyte abundances
occur from March through June, and Cyanophyte
abundances were lowest from November
through March. Chlorophytes dominated in
2001, 2002, and 2006, Bacillariophyta dominated
in 1996, 1997 and 2005, and Cyanophyta domi-
nated in 1998–2000. Chlorophyta and Bacillar-
iophyta were equally dominant in 1995. All three
taxa were approximately equal in 1994. There-
fore, the phytoplankton was far more diverse and
variable than previously reported (Wirick 1972,
Felix and Rushforth 1977, 1979, 1980, Rushforth
and Felix 1982) and more than 60 species were
identified during our study as compared to
previous reports 20 species (Felix and Rush-
forth 1977, 1979, 1980, Rushforth and Felix 1982).
Brine shrimp.—Brine shrimp generally hatch
from overwintering cysts in March/April and
largely disappear during December, a period
marked with water temperatures ,48C (Fig. 9A,
B and C). Nauplii and juvenile densities varied
within years (Fig. 9A, B, GLM respectively: F ¼
10.39, df ¼11, 120, p ,0.001; F ¼4.57, df ¼11,
120, p ,0.009) and among years (Fig. 9D, E,
GLM respectively: F ¼2.43, df ¼12, 120, p ,
0.007; F ¼3.02, df ¼12, 120, p ,0.001). Nauplii
and juvenile densities peaked in May and then
declined over the summer and fall with nauplii
diminishing more rapidly than juveniles. The
highest nauplii densities were observed in 2000
(56.4 nauplii/l) and the lowest densities (25 times
lower) occurred in 1994. The highest juvenile
densities were observed in 1997 (9.4 juveniles/l)
and the lowest densities (three times lower)
Fig. 7. Dissolved oxygen (DO: %of saturation) measures in the Great Salt Lake South Arm are presented. Plots
of DO with depth at two sampling sites for April in two years (red, 2005; blue, 2006) are presented in (A) for a
deep site and (C) for a shallow site. DO values for variations are presented within year by month (intra-annual)
in (B) and among years (inter-annual) in (D) for the period of our study (1994–2006) along with standard errors.
vwww.esajournals.org 13 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
occurred in 1995. Adult densities varied within
years (Fig. 9C, GLM: F ¼11.32, df ¼11, 120, p ,
0.001), but did not vary among years (Fig. 9F,
GLM: F ¼0.73, df ¼12, 120, p ,0.72). Adult
numbers peak in May or June and then slowly
decline. Lack of significant differences among
years arises from the great variability in adult
numbers within a year. However, adults densities
were observed to be highest in 2004 (5.82 adults/
l) and the lowest densities (2.07 adults/l or 2.8
times lower) occurred in 1995. Given the large
numbers of nauplii relative to juveniles in May,
nauplii survival is low. However, the survival of
juveniles to adults is much greater.
Densities of cysts, the diapausing life stage
which overwinters and initiates each year’s
population, (Fig. 10A, E) approached significant
variation within years (GLM: F ¼1.62, df ¼11,
109, p ,0.10), but did vary significantly among
years (GLM: F ¼3.18, df ¼11, 109, p ,0.001).
One would expect significant within-year varia-
tion, because cysts are primarily produced prior
to winter as the diapausing life stage. However,
cysts were found at relatively high abundances
throughout the year, especially in May. This
occurs for two reasons that particularly apply to
May counts: 1) counting of broken cysts after
they hatch (up to 90%broken on occasion); 2)
appearance of cysts re-entering the lake with
rain, wind or rising water levels from shoreline
deposits. To compensate for this, May was
excluded from analyses and within year varia-
tion was then observed to be significant (GLM: F
¼1.93, df ¼10, 109, p ,0.04). Peak densities of
cysts were observed from September–November,
as expected for a diapausing life stage. The
highest cyst densities were observed in 1995 (249
cysts/l) and the lowest (.eight times lower)
occurred in 1999, neither of these values occurred
in May. Because adult density did not vary
among years to the extent that cyst densities did,
production of the overwintering cysts was not
likely due to variation in adult densities as much
as to variation in per capita reproductive output.
Fig. 8. Phytoplankton measures in the Great Salt Lake South Arm are presented for variations within year
(intra-annual) by month and among years (inter-annual). Average chlorophyll-aconcentrations (A, C: no
measures for large portions of 2000–2001, 2003), and phytoplankton relative composition by taxa (Chlorophytes,
black; Bacillariophytes, gray; Cyanophytes, the difference from 100%; B, D: no counts for 2003–2004) are
presented for the period of our study (1994–2006) along with standard errors.
vwww.esajournals.org 14 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
There are a number of ways that per capita
reproductive output might vary. First, the adult
sex ratio might be skewed in favor of males (Fig.
10B, F). Sex ratios were observed to be skewed
for some months within years (GLM: F ¼1.90, df
¼10, 85, p ,0.05), but not among years (GLM: F
¼1.44, df ¼12, 85, p ,0.16). Sex ratios were
skewed in favor of males only at hatching time
(April) which suggests that cysts containing
males hatch earlier, while from September–
December, females predominated which suggest-
ed that females survive better than males.
Fig. 9. Brine shrimp developmental stage abundances in the Great Salt Lake South Arm are presented for
variations within year (intra-annual) by month and among years (inter-annual). Average nauplii (A, D), juvenile
(B, E) and adult (C, F) densities are presented for the period of our study (1994–2006 ) along with standard errors.
vwww.esajournals.org 15 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
Second, the proportion of females that are
reproductive might vary. Proportion of repro-
ductive females (Fig. 10C, G) varied within years
(GLM: F ¼3.53, df ¼10, 85, p ,0.001), and
among years (F ¼11.62, df ¼12, 85, p ,0.001).
The proportion of reproductive females was
lowest in April, June and July. However, vari-
ability in this proportion was greater among
Fig. 10. Brine shrimp population attributes in the Great Salt Lake South Arm are presented for variations
within year (intra-annual) by month and among years (inter-annual). Average cyst density (A, E), adult sex ratio
(B, F), proportion of adult females that are reproductive (C, G), and density of females that reproduce via
ovoviparity (black) and oviparity (gray) (D, H) are presented for the period of our study (1994 –2006) along with
standard errors.
vwww.esajournals.org 16 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
years than within years, with the greatest value
(0.90) in 1995 and the lowest value (0.02) in 1997,
a 45-fold difference.
Per capita reproductive output by females is
complicated further because brine shrimp can
reproduce in two ways: ovoviparity (eggs that
hatch in the ovisac) and oviparity (diapausing
cysts). Density of females reproducing by ovovi-
parity and oviparity (Fig. 10D, H) differed within
years (GLM respectively: F ¼6.62, df ¼11, 116, p
,0.001; F ¼8.79, df ¼11, 116, p ,0.001) and
among years (GLM respectively: F ¼2.12, df ¼12,
116, p ,0.02; F ¼6.22, df ¼12, 116, p ,0.001).
Ovoviparity peaked in May and then declined,
while oviparity peaked in September–October,
which indicated reproductive mode shifted after
the period of low reproduction in June and July.
Some years had more females reproducing via
ovoviparity (1995, 2002, 2006), but in most years
oviparity dominated (1994, 1999–2001, 2003,
2004). Therefore, reproductive output depends
on the proportion of females that are reproduc-
tive, and the shift from ovoviparity to oviparity.
Predators of brine shrimp.—There are three
potentially important predators of brine shrimp:
corixids that consume brine shrimp, grebes that
consume brine shrimp and their cysts, and
humans that commercially harvest brine shrimp
cysts. Corixid densities in plankton tows (Fig.
11D, H) were never great and did not vary within
years (GLM: F ¼1.52, df ¼11, 45, p ,0.16 ), but
did vary among years (GLM: F ¼2.37, df ¼6, 45,
p,0.05). Corixids were most abundant at the
same times of the year as brine shrimp (April–
December), and appeared to exhibit two gener-
ations each year (April, July). Over the six years
of measurements, corixid densities varied from
an annual mean approaching 0/m
3
to 1.46/m
3
.
Grebe numbers in the fall varied dramatically
among years (Fig. 12A). The commercial harvest
of brine shrimp cysts (Fig. 12B) averaged 4597 6
892 metric tons/yr and varied considerably
among years. Grebe and harvest numbers could
not be statistically examined for within and
among year variability, because they were based
on point samples by UDWR (respectively, single
lake survey and sum of commercial harvest
records).
Other taxa.—Several taxa were occasionally
encountered (protozoans, dinoflagellates, cope-
pods and brine fly larvae), all but brine fly larvae
were exceedingly rare at most times of the year.
Protozoans (primarily ciliates, but also amoebae)
and dinoflagellates (primarily Glenodinium sp.,
but occasionally Ceratium sp.) were only abun-
dantfromDecember–March(Fig.11A,E).
Copepods (Fig. 11B, F) were only abundant at
the time of spring inflows to Great Salt Lake
(March–July) and varied by more than 2800-fold
over six years; however, neither within year
(GLM: F ¼1.35, df ¼11, 45, p ,0.23) nor among
year (GLM: F ¼1.43, df ¼6, 45, p ,0.23)
variation were significant. Also, rotifers and
nematodes were occasionally encountered.
Brine fly larvae (Fig. 11C and G) in the water
column were abundant, but not as abundant as
brine shrimp, and exhibited both within year
(GLM: F ¼2.85, df ¼11, 33, p ,0.01) and among
year variation (GLM: F ¼5.55, df ¼3, 33, p ,
0.003). Brine fly larvae were sampled starting in
2003. They were most abundant at the same
times of year as brine shrimp (April–December),
but their populations exhibit two-annual peaks:
April–May and August–October, which may
reflect two distinct annual generations. Over four
years, brine fly larvae densities varied by more
than 2.7-fold.
DISCUSSION
Using our Great Salt Lake survey data, the
dynamics of the phytoplankton-based food web
can be investigated (right of dashed line in Fig. 2)
using structural equation modeling. Our initial
premise for the structural equation modeling is
that there are two distinct times of the year that
need to be examined separately: without brine
shrimp present (December–February) and with
brine shrimp present (March–November). The
ecosystem’s inter-annual variation is most strong-
ly exhibited when brine shrimp are absent (e.g.,
annual peak phytoplankton abundance), which
suggests that the strong effect of brine shrimp
can mask annual differences in abiotic influences.
On the other hand, intra-annual variation is most
strongly exhibited when brine shrimp are pre-
sent, because brine shrimp abundance is highly
variable among months.
Our initial hypotheses about the food web
included the following:
vwww.esajournals.org 17 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
1. In the absence of brine shrimp, maximum
annual phytoplankton abundance is primarily
determined by nutrient availability. Nutrient
availability should be strongly affected by
abiotic factors, but these relationships were
unknown. Two alternative starting points
were considered: variation in watershed in-
flows of nutrients or variation in lake volume
changing nutrient concentration of a fixed
nutrient quantity already in the lake.
Fig. 11. Abundances of other biota in the Great Salt Lake South Arm are presented. Average densities for
protozoans (A) and dinoflagellates (E ) are presented by month, because they were only measured for two years
(1994, 2000). Copepod density (B, F), brine fly density (C, G), and corixid density (D, H) for variations within year
(intra-annual) by month and among years (inter-annual) are presented along with standard errors, because 4–5
years had been sampled.
vwww.esajournals.org 18 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
2. With brine shrimp, phytoplankton abundance
should depend on its annual maximum in the
absence of brine shrimp, intensity of brine
shrimp grazing, and its ability to recover after
grazing. Therefore, brine shrimp abundance is
critical and we hypothesized that brine shrimp
populations were ultimately limited by phy-
toplankton availability.
3. Brine shrimp, we hypothesized, strongly
affected the annual abundances of Eared
Grebes, but these birds did not significantly
impact brine shrimp numbers.
4. We suspected that other species in Great Salt
Lake as depicted in the food web (Fig. 2) exert
little influence on phytoplankton, brine
shrimp or Eared Grebes, and vice a versa.
Each of these initial predictions will be
addressed.
Nutrient availability
We established that most nutrients did not
come from watershed inflows. First, annual
watershed inflow to the lake was not significant-
ly correlated with nutrient concentrations (DIN
or TDP at 1 m: respectively, p ,0.72 and p ,
0.36). Second, nutrient concentrations at 1 m for
sites near watershed inflow sources (Sites 2267,
2336, 2583, and 2963) were not greater than at
sites farther from inflows (Sites 2935, 3641, 3510
and 3954) ( paired t-test: DIN, t ¼0.19, df ¼16, p
¼0.25; TDP, t ¼0.56, df ¼7, p ,0.60). Finally,
nutrient concentrations at 1 m actually increased
with salinity, and salinity increases as inflow
diminishes. This is not to say that watershed
inputs of nutrients are unimportant, but this
change occurs at decadal time scales (recent
measures indicate only 8%of annual inputs come
from watershed inflows: D. Naftz, unpublished
data). Finally, DIN dynamics were more strongly
affected by salinity than TDP dynamics.
Another source of nutrients might be atmo-
spheric deposition. Maximum annual atmospher-
ic input can be estimated using the maximum
surface area of Great Salt Lake and maximum
DIN deposition values for the region (NADP
2000). This indicates that atmospheric deposition
accounts for less than 1.5%of DIN, again, a
minor annual input.
The above observations indicate that inter-
annual variation in nutrients is primarily a
dilution of an endogenous pool. However, we
originally hypothesized that nutrients would be
well mixed in Great Salt Lake, because the lake’s
shallow depth and high surface to depth ratio
would permit winds to easily mix the waters
(Baskin 2005, 2006, Baskin and Allen 2005,
Baskin and Turner 2006). However, mixing and
the effect on nutrient concentrations were more
complicated. For example, much greater concen-
trations of DIN ( principally ammonia) at greater
depths were observed (Fig. 13A, paired t-test: t ¼
8.90, df ¼66, p ,0.001). Although TDP
concentrations were far less different with depth,
they also increased with depth (Fig. 13A, paired
t-test: t ¼7.59, df ¼61, p ,0.001). Therefore,
nutrients appear to accumulate at greater depths.
Greater nutrient concentrations with depth,
Fig. 12. Annual fall Eared Grebe population num-
bers (A) and commercial cyst harvest (B) in the Great
Salt Lake are presented for the period of our study
(1994–2006).
vwww.esajournals.org 19 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
especially N in ammonia, are typical of eutrophic
lakes. Terminal lake systems with large water-
sheds, like Great Salt Lake, generally are hyper-
eutrophic, as they accumulate nutrients. This
explanation is based on high biotic activity in
eutrophic lakes which exhausts shallow nutrients
and produces products (e.g., excrement, carcass-
es, etc.) that sink and accumulate in deep layers
until mixing occurs. However, several other
nutrient patterns emerge with depth that com-
plicates this simple explanation.
As expected, when shallow DIN or TDP
increase, then their deep pools respectively
increase, but shallow and deep pools are weakly
coupled, especially for DIN (regression: DIN, r
2
¼
0.06, n ¼67, p ,0.049; TDP, r
2
¼0.23, n ¼62, p ,
0.001). Also, as expected, when shallow or deep
DIN increase, then shallow or deep TDP respec-
tively increase, but DIN and TDP pools are
weakly coupled, especially at shallow depths
(regression: shallow, r
2
¼0.05, n ¼72, p ,0.06;
deep, r
2
¼0.38, n ¼61, p ,0.001). Weak linkages
Fig. 13. Relationships between shallow and deep DIN and TDP in the Great Salt Lake South Arm are
presented. (A) Average concentrations of DIN (red) and TDP (blue) at shallow (1 m) and deep (.5 m) depths. (B)
Relationship between shallow salinity and DIN (red) and TDP (blue) concentrations. (C) Monthly DIN (red) and
TDP (blue) at deep sites in relation to the number of months since average lake salinity was below 12%.(D)
Relationship between monthly shallow and deep DIN in months when salinity is greater than 12%.(E)
Relationship between monthly shallow and deep TDP in months when salinity is greater than 12%.
vwww.esajournals.org 20 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
between shallow and deep nutrient pools, and
DIN and TDP pools at the same depth indicate
that they are partially distinct. This would occur
if Great Salt Lake does not mix as thoroughly as
we originally hypothesized.
Frequent mixing is indicated for shallow
portions (,4 m) of the lake, because they never
exhibited a chemocline or a thermocline, except
possibly a diel thermocline with calm conditions.
However, deep portions (.4 m) always exhibited
a thermocline between five and 7 m, and
exhibited a chemocline in some years. The
chemocline is termed the ‘‘deep brine layer’’
and it is always present at the deepest sites, but
expands into shallower areas as salinities decline
(;,12%). It expanded for a 7–8 year period and
again for two years at the end of our 13-year
study. When a deep brine layer is present at a
site, it was always below the thermocline. This
creates complex patterns of mixing in the lake
with the waters above the thermocline being well
mixed by wind/wave action, the waters between
the thermocline and chemocline periodically
mixing with spring and fall turnover, and the
waters below the chemocline mixing only in
some years when salinities are high and wind/
wave action is great.
The more complicated pattern of mixing may
explain the accumulation of nutrients at greater
depth and the weak linkage between shallow and
deep nutrient pools, as oligomixis reflects the
expansion and contraction of a ‘‘ deep brine’’
layer (Lin 1976a,b, Stephens 1976, Stephens and
Gillespie 1976). Oligomixis, rather than mero-
mixis, occurs because at greatest depths, a deep
brine layer never disappears, a deep brine layer
may appear periodically at intermediate depths,
and a deep brine layer may never appear in the
shallowest areas (.74%of volume, Baskin 2005,
2006, Baskin and Allen 2005, Baskin and Turner
2006). Therefore, explaining the pattern of
increasing N and P with depth is complicated
and dynamic, and we propose the following
scenario.
Deep (.5 m) nutrients.—We start with deep
nutrients, because we hypothesize that: 1) the
deep pool sequesters nutrients from the shallows
as oligomixis intensifies and releases nutrients as
oligomixis diminishes, and 2) being below the
photic zone, biotic influences will be weak.
A key factor is identifying when nutrients are
sequestered in the deep layer of the lake versus
when nutrients are released. To do this, we
employed piecewise regression to identify a
breakpoint in salinity when deep DIN or TDP
were correlated with shallow DIN or TDP at
higher salinities (mixing) or independent at lower
salinities (very limited mixing). A breakpoint was
observed at 12%salinity and no significant
correlations were ever observed with the previ-
ous or current months’abundance of phyto-
plankton or brine shrimp. At salinities ,12%,
deep nutrients were uncorrelated with shallow
nutrients (DIN: r
2
¼0.023, n ¼40, p ,0.99; TDP:
r
2
¼0.016, n ¼30, p ,0.57), and deep nutrients
increased as the number of months since salinity
had been ,12%increased (Fig. 13C, N: r
2
¼0.62,
n¼40, p ,0.001; TDP: r
2
¼0.27, n ¼30, p ,
0.002). At salinities 12%, deep nutrients were
correlated with shallow nutrients (Fig. 13D, E, N:
r
2
¼0.21, n ¼27, p ,0.016; TDP: r
2
¼0.46, n ¼16,
p,0.008).
Shallow nutrients.—We proposed three hypoth-
eses for nutrient pools in shallow regions: 1) they
will increase as lake volume decreases (salinity
rises) and concentrates the endogenous nutrient
pool; 2) they will increase as the presence of a
‘‘deep brine’’ layer diminishes (salinity rises) and
releases nutrients, or decrease as the ‘‘deep
brine’’ layer increases in extent (salinity declines)
and sequesters nutrients, and 3) they will
increase as biotic activity ( phytoplankton and
brine shrimp abundance) decreases and fewer
nutrients are used.
Hypothesis 2 was already supported above in
discussion of deep nutrients. Hypothesis 1 was
supported, as nutrient concentrations were pos-
itively correlated with the current salinity, but
hypothesis 3 was not supported, because nutrient
concentrations were not negatively correlated
with the current or previous months’phyto-
plankton abundance and the previous month’s
brine shrimp abundance was positively, not
negatively, correlated with nutrients (DIN: r
2
¼
0.74, n ¼71, p ,0.001, Salinity, p ,0.001,
Shrimp, p ,0.013; TDP: r
2
¼0.55, n ¼55, p ,
0.001, Salinity, p ,0.001, Shrimp, p ,0.007). The
biotic relationships will be discussed further
below, but it appears that when brine shrimp
are present they can influence nutrients through
their consumption of phytoplankton and releas-
ing the consumed nutrients through excrement,
vwww.esajournals.org 21 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
especially by production of ammonia. At times of
the year when brine shrimp were absent (brine
shrimp biomass ,0.1 adult equivalent/l), hy-
pothesis 1 dominates and phytoplankton still
exhibit no influence (Fig. 13B, DIN: r
2
¼0.92, n ¼
7, p ,0.001; TDP: r
2
¼0.95, n ¼5, p ,0.005).
In summary (Fig. 14), DIN and TDP dynamics
in Great Salt Lake appear to reflect two pools: the
shallow or photic zone and a deep zone. The
deep zone increases and decreases as a nutrient
sink due to oligomixis with nutrients settling into
the deep brine layer. When the deep brine layer
diminishes as salinity increases, mixing of the
photic and deep zones occurs with wind/wave
action and results in a single nutrient pool.
During portions of the year when brine shrimp
are present, they make nutrients available in the
photic zone through their excrement.
Phytoplankton
Even though phytoplankton abundance (Chla)
was negatively correlated with Secchi depth (Fig.
15A: r
2
¼0.42, n ¼100, p ,0.001), there is no
indication that light limits phytoplankton pro-
duction, because PAR was very high, averaging
73%of maximum, and the photic zone was
always at least 4 m deep. We also know from
laboratory studies with Great Salt Lake phyto-
plankton that reduced production is exhibited at
higher salinities and lower temperatures (Herbst
and Bradley 1989, Larson 2004). Therefore, we
hypothesize that phytoplankton abundance is
primarily limited by nutrients and brine shrimp
grazing with smaller effects of temperature and
salinity.
Phytoplankton growth appeared to be limited
by N availability, as reflected by the Redfield
Ratio. The ratio of N:P in phytoplankton based
on particulate organic N and P in many aquatic
ecosystems (Redfield Ratio) tends to approach a
value of 16. Consequently, optimal phytoplank-
ton growth should occur when inorganic sources
of N and P approach a ratio of 16 in the
ecosystem (Redfield 1934, 1942, 1958, Redfield
et al. 1963, but see Sterner et al. 2008 for
exceptions). Inorganic N:P ratios less than 16
indicate N limitation for phytoplankton and
values greater than 16 indicate P limitation.
Observed inorganic N:P ratios were always less
than 16 for shallow and deep nutrient pools,
indicating N limitation (Fig. 15B and C ). These
ratios did not vary within years (GLM: shallow, F
¼1.28, df ¼11, 52, p ,0.27; deep, F ¼1.04, df ¼
11, 41, p ,0.43), but did vary among years
(GLM: shallow, F ¼43.49, df ¼8, 52, p ,0.001;
deep, F ¼3.31, df ¼8, 41, p ,0.005). N limitation
is supported by previous laboratory studies of
phytoplankton reared in water from the South
Arm of Great Salt Lake (Stephens and Gillespie
1976, Wurtsbaugh 1988).
If N is limiting, we would expect to observe
annual peak phytoplankton abundances (Chla
peak
)
to be greater in years when DIN concentrations
are greater. These peak abundances occur in
months when brine shrimp are absent, because
brine shrimp grazing is very intense. Therefore,
we hypothesize the following relationship:
Chlapeak ¼ðDINphoÞðk1k2Salinity
þk3TemperatureÞ
where k
1
,k
2
and k
3
are constants. Temperature
had no significant effect, because observed winter
water temperatures when brine shrimp were
absent only varied by a few degrees among
years. As expected, phytoplankton abundance
(Chla
peak
) diminished as salinity increased (p ,
0.005). However, DIN in the photic zone (DIN
pho
)
had a slightly larger effect (53%of variance) as
hypothesized (Fig. 15D: nonlinear regression: r
2
¼
0.88, n ¼9, p ,0.001).
When brine shrimp are present, we hypothe-
size that phytoplankton abundance by month
(Chla
mo
) depends on the preceding winter’s peak
phytoplankton abundance (Chla
peak
), which re-
flects that year’s nutrient limitation, its reduction
by brine shrimp grazing and observed effects of
temperature and salinity from the laboratory:
Chlamo ¼ðChlapeak Þðk4k5Salinity
þk6Temperature k7Shrimpt1Þ
where k
4
–k
7
are constants, and Shrimp
t1
is the
brine shrimp biomass in the previous month
(nonlinear regression using robust LMS to
achieve convergence: r
2
¼0.76, n ¼85, p ,
0.001). Brine shrimp biomass dominated the
decrease in Chla
mo
(p ,0.001) from that year’s
Chla
peak
, increasing salinity had a smaller effect in
decreasing Chla
mo
(p ,0.001), and temperature
had no effect (p ,0.17). Therefore, we conclude
that the annual availability of DIN sets a
maximum phytoplankton abundance and this is
strongly reduced by brine shrimp grazing with
vwww.esajournals.org 22 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
little or no effect of salinity or temperature on the
phytoplankton.
Phytoplankton dynamics are more than
changes in overall abundance, but also include
considerable annual and monthly variation in
taxa composition (Fig. 8B, D: Chlorophyte,
Bacillariophyte, and Cyanophyte). We hypothe-
size that taxa relative abundances in the lake
should respond similarly to observations in
laboratory experiments with Great Salt Lake
Fig. 14. The phytoplankton-based food web in the Great Salt Lake South Arm is summarized for periods when
brine shrimp are present (March–November) and absent (December–February). Arrow size reflects interaction
strength, with blue arrows indicating indirect interactions and black arrows indicating direct interactions. Beside
the arrow is its partial correlation coefficient and each factor’sr
2
is in parentheses.
vwww.esajournals.org 23 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
phytoplankton, as salinity, temperature, overall
phytoplankton abundance (surrogate for nutri-
ent availability) and abundances of other taxa
are changed (Larson 2004). Using backward
stepwise regression, the annual and intra-annual
relative abundances of each phytoplankton taxa
were examined with salinity, temperature,
Chla
peak
(surrogate for nutrients), and absolute
abundances of other taxa (Table 1) to assess
whether laboratory and field observations agree.
Chlorophyte relative abundance in the labora-
tory increases as salinity increases, as tempera-
ture declines, as nutrients increase, and as
Bacillariophyte absolute abundance declines
(competition) (Larson 2004). In the lake, annual
variation in Chlorophyte relative abundance only
was negatively correlated with Bacillariophyte
abundance (r
2
¼0.58, n ¼10, p ,0.01), while
intra-annual variation in relative abundance was
positively correlated with salinity, and negatively
correlated with temperature, Chla
peak
(nutrient
abundance), and Bacillariophyte abundance (r
2
¼
0.52, n ¼51, p ,0.001). These results were
consistent with laboratory observations, except
for the negative relationship with Chla
peak
.We
suggest that the unanticipated Chla
peak
result may
reflect the concurrent Bacillariophyte increased
abundance (competition) as described below,
which prevents Chlorophytes from responding
in the field.
Bacillariophyte relative abundance in the lab-
oratory increases as Chlorophyte and Cyano-
phyte absolute abundances decrease
(competition), but it was not very sensitive to
salinity, temperature or nutrient availability
(Larson 2004). In the lake, annual variation in
Bacillariophyte relative abundance was negative-
ly correlated with salinity, temperature and
Chlorophyte abundance (r
2
¼0.87, n ¼10, p ,
0.01), while intra-annual variation in relative
abundance was positively correlated with salin-
ity and Chla
peak
, but negatively correlated with
Chlorophyte and Cyanophyte abundances (r
2
¼
0.49, n ¼51, p ,0.001). The consistently strong
negative responses to Chlorophyte and Cyano-
phyteabundanceswereexpectedfromthe
laboratory. The mixed annual and intra-annual
results for temperature and salinity might be
expected, given their weak effects in the labora-
tory. Finally, the positive response to Chla
peak
cannot be explained from laboratory results, but
Fig. 15. The relationship between phytoplankton
abundance (Chla) and Secchi Depth in the Great Salt
Lake South Arm is presented (A). Monthly (B) and
yearly (C) inorganic N:P ratios (6SE) are presented for
the photic zone (1 m) (gray) and deeper levels (.5m)
(black). The dashed line represents the 16:1 N:P ratio
(Redfield Ratio). The relationship between photic zone
DIN and phytoplankton abundance (Chla) is present-
ed (D) when brine shrimp are absent.
vwww.esajournals.org 24 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
variation in laboratory nutrients far exceeded
field variation and Bacillariophytes may be
particularly sensitive to low nutrient availability.
Cyanophyte relative abundance in the labora-
tory increases as salinity declines and as temper-
ature increases (Larson 2004). In the lake, annual
variation in Cyanophyte relative abundance was
negatively correlated with salinity (r
2
¼0.51, n ¼
10, p ,0.02), while intra-annual variation in
relative abundance was negatively correlated
with salinity, temperature, Chla
peak
, and Chlor-
ophyte and Bacillariophyte abundance (r
2
¼0.60,
n¼51, p ,0.001). The very strong negative
response to salinity is expected from laboratory
observations. We were surprised that tempera-
ture did not produce a strong positive effect in
the field. A negative competitive effect of
Chlorophyte and Bacillariophyte abundances
was unexpected from laboratory studies. This,
however, may be consistent with the observed
negative effect of Chla
peak
(nutrients), because
many of the Cyanophytes can fix nitrogen, unlike
Chlorophytes and Bacillariophytes; therefore,
Cyanophytes do better when nutrients are less
abundant because they can fix their own nutri-
ents and low nutrients reduce competition from
other taxa.
In summary (Fig. 14),amongyears,peak
phytoplankton abundance (Chla
peak
) occurred
when brine shrimp were absent (December–
February) and was primarily limited by photic
zone DIN concentrations. Within a year (month)
when brine shrimp were present, phytoplankton
abundance primarily depended on the year’s
Chla
peak
and brine shrimp grazing. The relative
abundances of phytoplankton taxa varied with
Chla
peak
, salinity, temperature and competition
with other taxa. Phytoplankton abundance and
composition were highly variable among and
within years as DIN and biotic interactions
among phytoplankton taxa varied.
Brine shrimp
Brine shrimp did not appear to be restricted by
dissolved oxygen (%saturation) except at depths
well below the photic zone (.5–7 m, Fig. 7A),
even though DO declined with temperature ( p ,
0.001) and salinity (p ,0.001) as expected
(regression: r
2
¼0.68, n ¼28, p ,0.001). Studies
in other lakes indicate that different Artemia
species do better at higher values within the
range of salinities and temperatures observed in
our study (e.g., Dana and Lenz 1986, Dana et al.
1993, 1995, Wear and Haslett 1986, Wear et al.
1986). The positive effect of temperature is
expected because brine shrimp are more produc-
tive in warmer water and consume more
phytoplankton. From our results presented
above, we also know that brine shrimp increase
photic zone DIN through their excretion (ammo-
nia), which increases phytoplankton abundance,
but decrease phytoplankton abundance through
grazing. Therefore, we hypothesized that brine
shrimp abundance is primarily limited by phy-
toplankton abundance. We also examined wheth-
er the availability of different phytoplankton taxa
differentially affected brine shrimp abundance.
This was accomplished using backward stepwise
regression to compare brine shrimp abundance
with phytoplankton abundance, salinity, temper-
ature and the abundances of phytoplankton taxa.
Assessing whether phytoplankton abundance
limits brine shrimp abundance is not straightfor-
ward, because phytoplankton abundance peaks
Table 1. Significant partial correlation signs for inter- and intra-annual relative abundances of phytoplankton
taxa.
Period Salinity Temperature Peak Chla
Absolute Abundance
r
2
NChlorophyte Bacillariophyte Cyanophyte
Inter-annual
Chlorophyte 0.58 10
Bacillariophyte 0.87 10
Cyanophyte 0.51 10
Intra-annual
Chlorophyte þ 0.58 10
Bacillariophyte þþ 0.87 10
Cyanophyte 0.51 10
vwww.esajournals.org 25 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
at times of the year when it is too cold for brine
shrimp to persist in the Great Salt Lake (winter/
spring) and phytoplankton recover (increase)
after grazing as the brine shrimp population
declines due to cooler water temperatures in the
fall (Fig. 16). Consequently, the role of phyto-
plankton abundance on brine shrimp abundance
(#/l) was examined in two ways: inter-annually
using each year’s average shrimp abundance (#/l)
from April–October, and intra-annually using
monthly brine shrimp abundances of each
developmental stage (#/l) from April–October.
Inter-annual average brine shrimp abundance
is hypothesized to follow the relationship:
Shrimp Abundance ¼k8Chlapeak þk9Temperature
þk10Shrimp þk11
where k
8
–k
11
are constants and Shrimp Abundance
is the average #/l from April–October during a
year, Temperature is the maximum water temper-
ature during a year, and Salinity is the maximum
salinity during a year. Shrimp Abundance was
positively correlated with annual maximum
phytoplankton abundance (Chla
peak
:lg/ml, p ,
0.021) and maximum annual water temperature
(p ,0.043) (Fig. 17A: r
2
¼0.71, n ¼14, p ,0.013).
Chla
peak
accounted for 80%of the explained
variance. There was no significant effect of the
maximum annual salinity. The analysis included
data from years prior to our study (1970, 1971,
1973, 1985: Stephens and Gillespie 1976, Wirick
1972, Wurtsbaugh and Berry 1990, Wurtsbaugh
1992, Wurtsbaugh and Gliwicz 2001), which
indicates that current and past brine shrimp
dynamics may be similar. Therefore, we conclude
that brine shrimp are primarily food-limited
among years.
Absence of a salinity effect was surprising
Fig. 16. Comparison between average monthly
phytoplankton (gray) and brine shrimp (black) densi-
ties in the Great Salt Lake South Arm are presented
along with standard errors.
Fig. 17. The relationship between peak annual
phytoplankton abundance (Chla) and brine shrimp
density is presented with the year for the Great Salt
Lake South Arm (A). The relationship between last
year’s brine shrimp density per grebe and the per
capita change in grebe numbers between the previous
and current year is presented (B).
vwww.esajournals.org 26 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
given results for other Artemia species (e.g., Wear
and Haslett 1986, Wear et al. 1986, Dana and
Lenz 1986, Dana et al. 1993, 1995, Gliwicz et al.
1995, Barata et al. 1996a,b, Williams 1998,
Browne and Wanigasekera 2000). We suggest
that the correlations with salinity observed in
these studies may reflect the negative effects of
salinity on phytoplankton production reported
above (Discussion: Phytoplankton), not necessarily
the direct effect of salinity on brine shrimp.
Furthermore, from laboratory studies with Great
Salt Lake brine shrimp (Belovsky and Larson
2002), salinity only exhibited negative population
effects at very low (4.5%) and very high (12%)
salinities, which were seldom observed in our
study (8.8–17.4%).
One can question our conclusion that brine
shrimp are food-limited given predation by
corixids (Wurtsbaugh and Berry 1990, Wurts-
baugh 1992) and the abundant Eared Grebes
(Caudell and Conover 2006) might limit brine
shrimp. However, when corixid and Eared Grebe
numbers were correlated with the residuals of
the above relationships or were included in the
relationships, predator impacts were not nega-
tive as expected if it is limiting, but exhibited a
weak positive effect for corixids (p ,0.02) and
no effect for Eared Grebes (p ,0.24). Previous
claims that corixids might limit brine shrimp
arise, because the studies were conducted in
littoral areas, not open water, and in years and
regions of the lake (e.g., Farmington Bay with its
causeway and inflows) where salinities were
very low (;5%) so that corixid numbers were
50 times higher than observed in our whole lake
samples. Mellison’s (2000) study of corixids in
littoral areas of Farmington Bay when salinities
were between 3–9%found that they did not limit
brine shrimp even though their densities were
141–241 times greater than our average value,
which supports our conclusion. Finally, we
suspect that Eared Grebe numbers, although
large, are not large enough to decrease the high
abundance and reproduction of brine shrimp.
For the smaller number of years (8 vs. 14 )
when the abundances of the three phytoplankton
taxa (Chla
peak
multiplied by taxa average relative
abundances) are known, the backward stepwise
regression was repeated with these values
substituted for Chla
peak
, which provides a better
fit (r
2
¼0.95, n ¼8, p ,0.005). Again maximum
summer water temperature was positively corre-
lated with average brine shrimp abundance (p ,
0.005), maximum annual salinity had no effect,
and phytoplankton abundances accounted for
the majority of variation explained (82%). Chlor-
ophyte abundance had no significant effect ( p ,
0.77), Bacillariophyte abundance had a signifi-
cant positive effect (p ,0.002) and Cyanophyte
abundance had a significant negative effect ( p ,
0.06). The absence of a Chlorophyte effect was
surprising because this is assumed to be the
principal food for brine shrimp; however, Chlor-
ophyte abundance did not vary appreciably
among years, which weakens its influence in
the regression, while Bacillariophyte and Cyano-
phyte abundances vary much more among years.
Intra-annual (monthly) brine shrimp abun-
dance (#/l) changes (DShrimp) with survival and
reproduction during the period after cysts hatch
in the spring (April–November). Because inter-
annual brine shrimp abundance was food-limited
and increased with water temperature, we expect
that survival and reproduction, and thereby
intra-annual DShrimp, also should be food-
limited and increase with temperature. Because
nauplii, juveniles and adult brine shrimp require
different amounts of food, differ in survival, and
only adults reproduce, each stage must be
accounted for separately in assessing DShrimp.
Therefore, we hypothesize the following rela-
tionship which can be examined via backward
stepwise regression:
DShrimp ¼k12Chlat1k13 Naupliit1
k14Juvenilest1k15 Adultst1
þk16Temperaturet1þk17 Salinityt1
þk18
where k
12
–k
18
are constants, DShrimp is the
change in brine shrimp density between months,
and t1refers to the value from the previous
month. The function reflects intraspecific exploit-
ative competition for food (i.e., density depen-
dence) so that as phytoplankton abundance in
the previous month (Chla
t1
) increases relative to
brine shrimp energetic requirements then
DShrimp will become positive, and as Chla
t1
decreases relative to brine shrimp energetic
requirements then DShrimp will become negative
(Schoener 1973).
As hypothesized, DShrimp was positively
correlated with the previous month’s phyto-
vwww.esajournals.org 27 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
plankton abundance (Chla
t1
)(p,0.007) and
temperature (p ,0.003), but negatively correlat-
ed with the previous month’s density of nauplii
(p ,0.001) and juveniles (p ,0.018) (r
2
¼0.78, n
¼79, p ,0.001). Density of adults did not have a
significant effect (p ,0.29). Significant influence
of nauplii and juvenile densities on DShrimp is
due to their large numbers, while the insignifi-
cant influence of adult density is due to their
relatively small numbers. This is not to say that
adult densities are unimportant, because adults
produce nauplii. As with inter-annual brine
shrimp abundance, salinity had no effect and
any limitation by corixid or Eared Grebe preda-
tion was not indicated. Finally, a more detailed
examination using phytoplankton abundance by
taxa did not contribute to the correlation.
Therefore, monthly changes in shrimp numbers
arise from competition for food and increased
survival at warmer temperatures, and explain the
observed annual pattern of phytoplankton abun-
dance and brine shrimp (Fig. 16).
Demography’simportance is emphasized by the
intra-annual function, which reflects characteris-
tics of a complex stage-dependent matrix popu-
lation model, where monthly changes depend on
food and temperature. Brine shrimp demogra-
phy is critical to the population’s production and
can be tracked through the annual life cycle
(Table 2). Understanding biotic and abiotic
factors influencing crustacean populations is not
well understood (Twombly et al. 2007 ), and our
brine shrimp data can contribute to a better
understanding.
1. Spring (March–April) nauplii, as they hatch
from cysts, constitute the onset of the annual
cycle. The peak density of spring nauplii is highly
correlated with the density of cysts present in the
spring (average of January–March), as expected
because cysts are the stage that permits brine
shrimp to survive winter’s cold (Table 2: r
2
¼0.75,
n¼11, p ,0.001).
2. Changes in nauplii density after the March–
April hatching are important because these are
the source of adults. Nauplii density should
increase primarily from adult females producing
live young via ovoviparity, but some nauplii may
emerge from cysts, which is the change in cyst
density (DCysts). Nauplii density should decrease
as some die or transition into juveniles, and
lower water temperatures and salinity should
decrease survival and reproduction. Finally, the
abundance of phytoplankton per shrimp should
increase nauplii through greater reproduction,
but decrease it through increased transition to
juveniles; the net effect should be positive,
because reproduction exceeds transition. We
hypothesize the following relationship which
can be examined via backward stepwise regres-
sion:
DNauplii ¼k19Reproductive Femalest1
þk20DCystst1k21 Naupliit1
þk22Temperaturet1þk23 Salinity
þk24
Chlat1
Shrimpt1
þk25
where k
19
–k
25
are constants. The hypothesized
relationship is supported (Table 2: r
2
¼0.76, n ¼
19, p ,0.001), except salinity again is found to
have no effect. Food is the dominant limiting
factor and examining it by taxa did not improve
correlations.
3. Density of reproductive adult females is a
critical aspect of DNauplii and depends on
density of adult females and the proportion that
are reproductive. Density of reproductive fe-
males should be a function similar to DShrimp
presented above with Adults
t1
replaced by Adult
Females
t1
:
Adult femalest1¼k26Naupliit1
þk27Juvenilest1
þk28Adult femalest1
þk29Chlat1
þk30Temperaturet1
þk31Salinity þk32
where k
26
–k
32
are constants. As for DShrimp,
salinity has no effect, but all other hypothesized
relationships are observed (Table 2: r
2
¼0.65, n ¼
27, p ,0.001), and again food is the principal
limiting factor and examining it by taxa did not
improve correlations.
The proportion of adult females that are
reproductive (producing eggs via ovoviparity
or cysts via oviparity) should increase with food
availability, decline due to intraspecific competi-
tion for food (brine shrimp biomass), and salinity
was not considered important given the above
results. Given our earlier inter-annual results, we
hypothesize that food abundance should primar-
ily be defined by Chlorophyte and Bacillario-
vwww.esajournals.org 28 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
phyte abundances, while Cyanophyte abundance
should have a negative effect:
Proportion Females Reproductive
¼k33Chlorophytest1þk34 Bacillariophytest1
k35Cyanophytest1
k36Brine Shrimp Biomasst1
þk37Temperaturet1þk38
where k
33
–k
38
are constants. The relationship is
supported with food again the dominant limiting
factor (considering taxa instead of Chla
t1
im-
proved the correlation), except that Bacillario-
phyte abundance and temperature were not
significant (Table 2: r
2
¼0.91, n ¼27, P ,0.001).
4. Brine shrimp reproduction is more complex
than the proportion of females that are repro-
ductive, because reproduction can occur through
ovoviparity (eggs hatch in the female’s egg sac)
and oviparity (diapausing cysts). While repro-
duction occurs via both modes throughout the
annual cycle, there is a strong seasonal pattern
with ovoviparity dominating early in the cycle
and oviparity dominating late. From laboratory
studies, we know that the switch from ovovi-
parity to oviparity is triggered by low food
abundance and increased temperature (Gliwicz
et al. 1995, Belovsky and Larson 2002). We
hypothesize the following relationship:
Proportion Ovoviparity ¼k39 Chlorophytest1
þk40Bacillariophytest1
k41Cyanophytest1
þk42Temperaturet1
þk43
where k
39
–k
43
are constants. The function is
supported with food being the dominant limiting
factor, but again Bacillariophytes are not signif-
icant (Table 2: r
2
¼0.83, n ¼27, p ,0.001).
Proportion oviparity is 1 minus proportion
ovoviparity.
5. Change in cyst density (DCysts) is important
because it reflects accumulation of cysts (dia-
pausing eggs) that permit the population to
survive winter (December–February) and initiate
the next year’s population. This should increase
as the density of adult reproductive females
producing cysts increases, as food increases and
as temperature increases, but it should decrease
if cysts hatch. We hypothesize the following
relationship:
DCysts ¼k44Chlorophytest1
þk45Bacillariophytest1
k46Cyanophytest1þk47 Temperaturet1
þk48
where k
44
–k
48
are constants. The function is
supported, but again Bacillariophytes are not
Table 2. Significant multiple regressions obtained for brine shrimp demography (partial correlations in
parentheses).
Life Stage Independent Variables r
2
NP
1) Peak spring nauplii density (Mar or Apr) Spring cyst density 0.75 11 ,0.001
2) DNauplii (ovoviparity production: May–Jun) Adult females w/eggs
t1
(þ0.28) 0.76 19 ,0.001
Per capita phytoplankton
t1
(þ0.08)
DCysts (þ0.07)
Nauplii
t1
(0.39)
Temperature
t1
(0.05)
3) Adult female density (May–Nov) Nauplii
t1
þJuveniles
t1
(þ0.68) 0.65 27 ,0.001
Adult females
t1
(þ0.09)
Per capita phytoplankton
t1
(þ0.02)
Temperature
t1
(þ0.02)
Proportion of females reproducing (May–Nov) Chlorophyte abundance
t1
(þ0.40) 0.91 27 ,0.001
Cyanophyte abundance
t1
(0.28)
Brine shrimp biomass
t1
(0.28)
4) Proportion of reproducing females producing Chlorophyte abundance
t1
(þ0.43) 0.83 27 ,0.001
nauplii (May–Nov) Cyanophyte abundance
t1
(0.28)
Temperature
t1
(0.20)
5) DCysts (oviparity production: Jul–Nov) Adult females w/cysts
t1
(þ0.31) 0.53 27 ,0.001
Per capita chlorophyes
t1
(þ0.28)
Per capita cyanophytes
t1
(0.01)
Nauplii
t1
(0.08)
Temperature
t1
(þ0.03)
6) Spring cyst density (Dec–Mar) Peak fall cyst density (þ0.87) 0.77 11 ,0.001
Winter salinity (60.09)
vwww.esajournals.org 29 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
significant (Table 2: r
2
¼0.53, n ¼27, p ,0.001).
6. Spring cyst density (average of January–
March) completes the annual cycle (see above:
Discussion: Brine Shrimp:Demography:1.Spring
(March–April) nauplii ). Obviously, spring cysts
should increase with the peak fall density. Lower
overwinter salinities should decrease the spring
cysts, because at lower salinities cysts can break
diapause and hatch during brief warm periods,
but these nauplii die as it is too cold to feed.
Furthermore, at lower salinities cysts can sink,
temporarily being lost in the deep brine layer and
sediments. However, at higher overwinter salin-
ities, spring cysts also decline as greater cyst
buoyancy allows winds and waves to deposit
more of them on beaches, where they are
temporarily lost until washed back into the lake.
Warmer temperatures might decrease spring
cysts if diapause breaks at lower salinities as
discussed above. We hypothesize the following
relationship:
Spring Cysts ¼k49 Peak Fall Cysts þk50 Salinity
k51Salinity2
k52Average Winter Temperature
k53
where k
49
–k
53
are constants. This is supported,
but temperature is not significant because it may
not vary enough among years (Table 2: r
2
¼0.77,
n¼11, p ,0.001).
In summary (Fig. 14), brine shrimp density and
demography are primarily food-limited as ex-
pected from our bottom-up hypothesis for this
food web. Temperature has a less important
effect, while salinity and predation exert no
effect. Our observation of food-limitation is
consistent with Wurtsbaugh and Gliwicz’s
(2001) suggestion based on a single year (1994 )
from our dataset. Our more comprehensive
dataset also indicates that phytoplankton taxa
differ in suitability as foods for brine shrimp.
Predators
Even though we found that corixids and Eared
Grebes did not exert a negative effect on brine
shrimp populations, we still expected corixid and
Eared Grebe numbers to increase with brine
shrimp density, given our bottom-up hypothesis
for this food web.
Corixid density should not be a simple function
of brine shrimp density, because we know that
corixids exhibit no preference for nauplii due to
their small size, avoid adults due to their large
size, but prefer juveniles (Belovsky and Mellison
1998). Salinity should exhibit a negative effect on
corixids, because they cannot tolerate salinities
approaching 9%, and prefer salinities between 2–
6%(Mellison 2000). Temperature may exhibit a
positive effect, because corixids are restricted to
shallows where water temperatures are highest
(Hayes 1971, Mellison 2000). We hypothesize the
following relationship:
Corixids ¼k54Juvenilest1k55 Adultst1
k56Salinityt1þk57 Temperaturet1
k58
where k
54
–k
58
are constants. Corixid density was
poorly explained by this relationship (r
2
¼0.12, n
¼62, p ,0.03). As expected from corixid feeding
behavior, density is not correlated with nauplii
(p ,0.57), is negatively correlated with adults ( p
,0.02), and is positively correlated with juve-
niles (p ,0.01). Temperature is positively
correlated with density as predicted (p ,0.06).
However, salinity unexpectedly exhibits a posi-
tive effect (p ,0.04) for which we have no
explanation except that there may be some other
environmental variable important to corixids that
is positively correlated with salinity. A weak
relationship between brine shrimp and corixid
densities is not surprising, as corixids are not
restricted to Great Salt Lake, but can fly between
the lake and surrounding brackish waters where
they have abundant alternate prey (Hayes 1971).
Also, we suspect that substrate in the shallows
may exert strong limits to corixids.
Grebe numbers should exhibit a per capita
change between years that is positively correlat-
ed with the abundance of brine shrimp (all
developmental stages) in the previous year,
because more food should lead to greater
survival and reproduction. Grebes consume large
amounts of brine shrimp and their cysts in the
South Arm (as much as 95%of diet: Caudell
2001, Conover and Caudell 2009, Conover et al.
2009). This relationship emerges (Fig. 17B: r
2
¼
0.81, n ¼9, p ,0.001), indicating the dependence
of grebes on brine shrimp. Because Great Salt
Lake is primarily a staging area for spring and
fall grebe migration (most grebes reproduce and
all winter elsewhere), this indicates that migra-
tory success is critical to grebe populations.
vwww.esajournals.org 30 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
A brine shrimp density of 5.85/l is predicted by
the above regression to provide no net change in
grebe numbers. Jehl (1988, 2007) concluded that
Eared Grebe migration from Mono Lake in
California was influenced by comparable brine
shrimp densities. In contrast, a smaller value of
0.04–1.0/l was estimated to provide basal and
foraging energetic needs for Eared Grebes in
Great Salt Lake (Caudell 2001, Caudell and
Conover 2006, Conover and Caudell 2009), but
this does not include the greater demands of
migration, survival in the wild, and reproduc-
tion. Finally, our estimated minimum brine
shrimp density for maintenance of grebe popu-
lations (5.85/l) is not much smaller than average
observed brine shrimp density during our study
(6.97/l), which suggests that grebe numbers may
be sensitive to brine shrimp populations.
In summary (Fig. 14), corixids were not closely
coupled with the nutrient/phytoplankton/brine
shrimp food web in the Great Salt Lake’s South
Arm, while grebe populations were.
Anthropogenic impacts: brine shrimp harvesting
Commercial harvests of brine shrimp cysts are
consistently large (4597 6892 metric tons/yr
(SE)), averaging 61 611%(SE ) of peak fall cyst
densities and with some years approaching
.90%. This might influence the nutrient/phyto-
plankton/brine shrimp food web, because spring
cysts, which initiate the brine shrimp population,
are in part the fall peak cyst density less the
harvest. However, no significant correlation
could be detected between annual harvest and
the following year’s brine shrimp population (r
2
¼
0.01, n ¼11, p ,0.70). First, we might not expect
to observe an impact, because starting in 1997,
UDWR limited harvests to levels projected to
prevent a negative impact on brine shrimp
population production. Second, to detect harvest
effects, one needs a better understanding of
overwinter cyst survival and how spring hatch-
ing numbers subtly impact brine shrimp age
structure and reproduction. This is the subject of
additional papers on an experimental study of
overwinter cyst survival and the State of Utah’s
developmentofcystharvestregulations(G.
Belovsky et al., unpublished manuscript; G. Belov-
sky and C. Perschon, unpublished manuscript).
Other Great Salt Lake biota
A number of other species were incidentally
encountered in our phytoplankton and brine
shrimp sampling. Brine fly larvae were frequent-
ly encountered and abundant, but most species
were rarely encountered and at low densities
compared to other aquatic systems (protozoans,
dinoflagellates, copepods, cladocerans, nema-
todes and rotifers). For example, cladocerans
were encountered only twice in 13 years and
probably reflect high freshwater inflows to Great
Salt Lake. However, some of the less abundant
species (protozoans, dinoflagellates and cope-
pods) were encountered frequently enough to
have their densities analyzed. Many of these
species are not part of the phytoplankton-based
food web discussed in this paper (right of dashed
line in Fig. 2), but the organic particle/benthic
algae-based food web (left of dashed line).
Brine fly larval densities were positively corre-
lated with brine shrimp densities (r
2
¼0.24, n ¼
48, p ,0.001), which indicates that both species
similarly respond to the environment. Therefore,
we hypothesized that brine fly larvae should
respond, like brine shrimp, positively with
phytoplankton abundance and water tempera-
ture, and negatively with salinity:
Brine Fly Larvae ¼k59 Chlat1þk60Temperaturet1
k61Salinityt1þk62
where k
59
–k
62
are constants. The hypothesized
relationship was observed (r
2
¼0.70, n ¼41, p ,
0.001) with water temperature dominating (p ,
0.001). We suspect that Chla is not a better
predictor of brine fly larval density (p ,0.04),
because the larvae primarily consume organic
particles (detritus) and benthic algae, not phyto-
plankton like brine shrimp, and this correlation
reflects benthic algae and phytoplankton both
increasing with greater nutrients (DIN). Brine
shrimp density is not correlated ( p ,0.20) with
the residuals of this relationship, indicating the
absence of strong interspecific competition as
also found in the laboratory (Belovsky and
Mellison 1997). Grebe numbers were positively
correlated with the residuals (p ,0.001),
indicating no negative effect of grebe predation,
and corixid predation could not be examined
because they were sampled in different years.
Results support the ecosystem depiction in Fig. 2
where two weakly linked food webs comprise
vwww.esajournals.org 31 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
the Great Salt Lake’s South Arm.
Protozoan and dinoflagellate densities only re-
sponded to abiotic conditions. Protozoans, prin-
cipally ciliates, were negatively correlated with
water temperature and salinity (r
2
¼0.59, n ¼10,
p,0.05). Dinoflagellates were negatively corre-
lated with temperature (r
2
¼0.51, n ¼7, p ,0.07).
Increased density with lower temperature re-
flects both taxa’s presence only in winter.
Copepod densities are positively correlated with
phytoplankton abundance (p ,0.09) and tem-
perature (p ,0.002), and negatively correlated
with salinity (p ,0.12) and brine shrimp
biomass (p ,0.004) (r
2
¼0.41, n ¼42, p ,
0.009). Copepods are most abundant when
salinity is low, but their ability to respond to
low salinities may be curtailed by competition
with brine shrimp. However, the residuals of the
relationship for brine shrimp density were not
correlated with copepods (p ,0.76), indicating
that copepods have no competitive effect on
brine shrimp.
CONCLUSION
Our 17 years of data through 2006 for the Great
Salt Lake’s South Arm and continued monitoring
represents one of the most extensive and long
term studies of a large hypersaline lake, since
most long term studies are from much smaller
shallow (e.g., Lake Grassmere, NZ: Wear and
Haslett 1986, 1987, Wear et al. 1986) or deeper
(e.g., Mono Lake, USA: Dana et al. 1990, 1993,
1995) hypersaline lakes that present very differ-
ent environmental patterns. In general, there are
few long term aquatic data sets (e.g., Lake
Washington, USA: Hampton and Schindler
2006, Hampton et al. 2006; Peter and Paul Lakes,
USA: Carpenter et al. 2001, Carpenter 2003; Lake
Tahoe, USA: Goldman et al. 1993; Mirror Lake:
Likens 1985; North American Great Lakes:
numerous researchers).
The Great Salt Lake ecosystem is not complex,
but it is not as simple as often portrayed (e.g.,
Jordan 1889, Pack 1919, Wirick 1972, Post 1975,
1980, Stephens 1974, Stephens and Gillespie
1976). While containing fewer species than most
lakes, Great Salt Lake has more species (;100)
than expected for salinity levels of 10–20%(;8
species expected: Williams 1978, Williams et al.
1990, see Boetius and Joye 2009 for a counter
view). Yet, Great Salt Lake’s ecosystem is simple
enough that its food web dynamics may be
examined in entirety, as we present for one of the
lake’s two food webs: nutrient-phytoplankton-
zooplankton-avian consumers (Fig. 2 and 14).
For this food web, we are able to quantify how
different abiotic and biotic factors operate (Fig.
14), and use these findings to examine several
ecological concepts and provide some manage-
ment projections for Great Salt Lake (Menge et al.
2009).
General ecological insights
Extreme environments.—Ecologists have long
debated the relative importance of abiotic factors
in controlling populations (Andrewartha and
Birch 1954, 1984). The importance of abiotic
factors has been particularly emphasized for
extreme environments and there is no question
that the hypersaline conditions of Great Salt Lake
are extreme (Williams 1998). Nonetheless, Great
Salt Lake is highly productive with dissolved
inorganic nitrogen and salinity about equally
limiting phytoplankton production when brine
shrimp are not present, but the effect of salinity
diminishes considerably when brine shrimp are
grazing. A major impact of salinity on phyto-
plankton production does appear through its
effect on phytoplankton taxa composition. We
found little or no direct effect of salinity on brine
shrimp density, but the above effect of salinity on
phytoplankton abundance and taxonomic com-
position constitutes an indirect effect, because
brine shrimp are food limited. The predators of
brine shrimp were either unaffected (Eared
Grebes) or weakly affected (corixids) by salinity.
Therefore, salinity did not appear to dominate
this food web’s dynamics, supporting the idea
that extreme environments are not extreme to
species adapted to live there (Sanders 1969,
Slobodkin and Sanders 1969).
Food web dynamics.—Ecologists for many years
have debated whether food webs are primarily
controlled by productivity (bottom-up control) or
by consumption (top-down control) (Lindeman
1942, Hairston et al. 1960, Slobodkin et al. 1967,
Fretwell 1977, Oksanen et al. 1981, Carpenter et
al. 1985, Schmitz 1992, Carpenter and Kitchell
1993, Hairston 1989, Hairston and Hairston
1993). The above distinctions have been built
upon simple food chain/web models that closely
vwww.esajournals.org 32 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
resemble the Great Salt Lake food web examined
by us. Using these simple models, some argue
that top-down control should become stronger as
primary production increases (e.g., Oksanen et
al. 1981), while others argue that top-down
control diminishes as primary production in-
creases (e.g., Schmitz 1992). Our Great Salt Lake
study supports the latter hypothesis, as bottom-
up control is observed in this highly productive
ecosystem.
Another possibility is that food webs are
neither top-down nor bottom-up controlled,
rather they are a mix of the two, and the relative
importance of each can vary (Strong 1992, Chase
2000, Hunter and Price 1992, Schmitz 1992, Vanni
and De Ruiter 1996, Schmitz et al. 2000). It has
been suggested that top-down control decreases
in relative importance at higher trophic levels,
because predators are more likely to be food-
limited, and this should result in greater reduc-
tion in food populations (next lower trophic
level) as one moves up the food web (Menge and
Sutherland 1987). The first part of this hypothesis
is supported by our study as the partial
correlation between food and consumer abun-
dances increases from nutrients and phytoplank-
ton (þ0.22) to phytoplankton and brine shrimp
(þ0.82) to brine shrimp and grebes (þ0.91).
However, the second part is not supported,
because Eared Grebe and corixid predation
exerted no limits on brine shrimp density. The
absence of predators reducing brine shrimp
densities means that trophic cascades cannot
emerge (Polis and Winemiller 1996).
Food web structure.—Ecologists using mathe-
matical models have hypothesized that ecosys-
tems should be composed of weakly linked food
chains, making the ecosystem more stable,
resilient and resistant (Pimm 2002, Pimm et al.
1991, Polis and Winemiller 1996, Teng and
McCann 2004). While we only examined in detail
the South Arm of Great Salt Lake’s nutrient/
phytoplankton/brine shrimp/Eared Grebe food
web (right of dashed line in Fig. 2), there is
another simple food web for which we have
some information (organic particles/benthic al-
gae/brine fly larvae/gulls: left of dashed line in
Fig. 2). The two food webs are potentially cross
linked in three ways: 1) phytoplankton and
benthic algae competing for common nutrients,
2) brine shrimp and brine fly larvae competing
for phytoplankton and benthic algae, and 3)
corixids and Eared Grebes preying on both brine
shrimp and brine fly larvae. We measured the
last two potential cross links and found them to
be either very weak or nonexistent, supporting
the idea that ecosystems may be composed of
weakly cross linked food webs.
Nutrient dynamics.—Some ecologists consider
that nutrient availability to autotrophs in an
ecosystem is constant and independent of con-
sumers, which simplifies food web dynamics
(Hairston et al. 1960, Slobodkin et al. 1967,
Fretwell 1977, Oksanen et al. 1981, Hairston
1989, Hairston and Hairston 1993). Others argue
that nutrient availability is not constant, but
modified by consumers, which complicates food
web dynamics (e.g., Porter 1976, Porter et al.
1996, DeAngelis 1992, Pace 1993, Wetzel 1983,
Wardle 2002, Weisser and Siemann 2004). The
latter perspective is supported by our study, as
brine shrimp increase nutrient availability to
phytoplankton through their consumption of
phytoplankton and re-release of nutrients
through excrement. Similar observations have
been reported for the simple ecosystems found in
the harsh environment of desert streams (Grimm
1987, Grimm and Fisher 1989).
Highly variable annual nutrient availability to
autotrophs also emerges in Great Salt Lake with
the expansion and contraction of a deep brine
layer creating oligomixis. Over years as the deep
brine layer expands and oligomixis increases
with lower salinities, more nutrients are lost to
greater depths and the recycling of nutrients by
consumers gains in importance until lake mixing
increases and the deep brine layer contracts with
higher salinities. Similar complexities have been
reported for Mono Lake, another hypersaline
lake (Melack and Jellison 1998, Carini and Joye
2008, MacIntyre et al. 2009), and lakes with
higher salinity due to mine runoff (Pieters and
Lawrence 2009).
Management implications
The Great Salt Lake is a unique ecosystem,
especially for North America, but it is increas-
ingly being threatened as the surrounding region
(Salt Lake, Davis and Tooele Counties, Utah: Fig.
1) becomes more impacted by anthropogenic
activities. Our study helps to identify potential
anthropogenic threats, many of which have not
vwww.esajournals.org 33 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
previously been fully recognized.
Commercial harvesting of brine shrimp cysts
affecting waterbird populations by reducing their
brine shrimp food base was the original concern
motivating UDWR to fund our study. Even with
the harvest removing annually an average of 61%
(some years .90%) of cysts, our database did not
identify a negative effect of harvesting on shrimp
numbers. However, since 1997 (all but three
years of our 13 year study), UDWR regulated
harvests based on experimental and modeling
studies that were funded by this project. The
criteria for regulation requires that the harvest be
stopped when cyst densities are reduced to a
level where maximum brine shrimp production,
as a waterbird food base, is assured, and
harvesting is not allowed if cyst densities are
below this value at the start of the harvest season.
The basis and development of these regulations
will be presented elsewhere (G. Belovsky and C.
Perschon, unpublished manuscript). However, im-
plementation of these regulations by UDWR
requires careful and frequent monitoring of cyst
densities to terminate harvesting at the appro-
priate level.
Water diversion of Great Salt Lake inflows by
agriculture, industry and urbanization is increas-
ing, because continued economic growth de-
pends on water. The sum of current proposed
projects to impound and divert freshwater that
otherwise would flow into Great Salt Lake
already exceeds the annual inflow (Great Salt
Lake Planning Team 1999, 2000). Mining of
minerals in Great Salt Lake water would further
reduce water. Obviously, less water decreases.
Our study indicates that this will increase water
temperatures and salinity, which will strongly
affect nutrient availability, phytoplankton abun-
dance and composition, brine shrimp and Eared
Grebes (Fig. 14). Given Great Salt Lake’s oligo-
mixis, salinity increases will enhance mixing,
which releases nutrients from depths to the
photic zone. This will increase phytoplankton
production, but this will be short term, because
phytoplankton abundance will diminish as sa-
linity continues to increase. Changes in phyto-
plankton taxa composition are very complex and
not as easily predicted. Nonetheless, the outcome
will be lower brine shrimp abundances, a critical
food base for many waterbirds.
Nutrient additions from agricultural fertilizers
and sewage are expected to increase and these
will accumulate in this terminal lake. We already
know that N concentrations are much greater in
Farmington Bay, where the inflow of Salt Lake
and Davis County sewage is located, and this
leads to dramatically higher phytoplankton
productivity and a different taxa composition
than the South Arm (Marcarelli et al. 2006,
Wurtsbaugh and Marcarelli 2004, 2006). Our
study indicates that effects of nutrient additions
may be masked by the lake’s oligomixis during
periods when the deep brine layer is extensive
and long lasting, which can lead to complacency
about pollutants. During these periods of lower
salinity, nutrients can be sequestered at depths
having little effect on phytoplankton production
and composition; then as the deep brine layer is
reduced in periods of higher salinity, nutrients
will be released from the depths and phyto-
plankton production and composition will
change dramatically. This means that the lake
can appear to be unaffected by nutrient additions
for long periods and then their effects can
suddenly be manifested; water diversions (see
above) changing salinity could help to suddenly
release nutrients and other substances by dimin-
ishing the deep brine layer. Finally, cumulative N
additions may lead to P replacing N as the limit
to phytoplankton production.
Anthropogenic climate change with its increased
temperatures and reduced precipitation and
runoff in the lake’s drainage (Wagner 2003,
Knowles et al. 2006, Stewart et al. 2005, Wagner
and Adrian 2009) could further exacerbate the
effects of water diversion and nutrient addition.
Climate change will reduce lake volume, increase
salinity and temperature, and make oligomixis
less important so that all of the concomitant
effects described above could emerge sooner and
be exacerbated. Finally, the more frequent mixing
will increase the incidence of gas releases (e.g.,
sulfides) from lake sediments that are annoying
to people today, making public complaints more
frequent and intense (Reese and Anderson 2009).
Great Salt Lake is a unique ecosystem in North
America with high biological production, wild-
life and commercial value that may be threatened
by anthropogenic pressures. This should be a
conservation concern, and our study begins to
provide ecological understanding of this unique
system and how to protect it. This simple system
vwww.esajournals.org 34 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
also informs us about ecological concepts that
maybemoredifficulttoaddressinmore
complex ecosystems. But it provides only a start,
because it illustrates how little we know about
nutrient availability and phytoplankton respons-
es, dynamics in another major food web (organic
particles, bacteria, benthic algae, brine fly larvae
and their avian consumers), and how dynamics
in specific areas of the lake may differ with
depth, current, wind, inflows and proximity to
anthropogenic influences. Our program is ad-
dressing these additional questions.
ACKNOWLEDGMENTS
We wish to thank the Great Salt Lake Ecosystem
Program, Utah Division of Wildlife Resources (to GEB,
DS, DN, RB), National Science Foundation (to GEB,
DEB-9322576), and U.S. Geological Survey (to DS, DN,
RB) for funding. The Great Salt Lake Technical
Advisory Group provided feedback. S. Kilham, B.
Marden and J. Butler provided discussion on data and
methods. We thank T. Crowl, J. Slade, J. Neill and W.
Wurtsbaugh for comments.
LITERATURE CITED
Aldrich, T. W., and D. S. Paul. 2002. Avian ecology of
Great Salt Lake. Pages 343–374 in J. W. Gwynn,
editor. Great Salt Lake: An overview of change.
DNR Special Publication, Utah Geological Survey,
Salt Lake City, Utah, USA.
Andrewartha, H. G. and L. C. Birch. 1954. The
distribution and abundance of animals. University
of Chicago Press, Chicago, Illinois, USA.
Andrewartha, H. G., and L. C. Birch. 1984. The
ecological web. University of Chicago Press,
Chicago, Illinois, USA.
Barata, C., F. Hontoria, F. Amat, and R. Browne. 1996a.
Competition between sexual and parthenogenetic
Artemia: temperature and strain effects. Journal of
Experimental Marine Biology and Ecology 196:313–
328.
Barata, C., F. Hontoria, F. Amat, and R. Browne. 1996b.
Demographic parameters of sexual and partheno-
genetic Artemia: temperature and strain effects.
Journal of Experimental Marine Biology and
Ecology 196:329–340.
Baskin, R. L. 2005. Calculation of area and volume for
the south part of Great Salt Lake, Utah. USGS
Open–File Report 2005-1327. US Geological Survey,
Salt Lake City, Utah, USA.
Baskin, R. L. 2006. Calculation of area and volume for
the north part of Great Salt Lake, Utah; USGS
Open–File Report 2006-1359. US Geological Survey,
Salt Lake City, Utah, USA. hhttp://pubs.usgs.gov/
of/2006/1359/i
Baskin, R. L., and D. V. Allen. 2005. Bathymetric map
of the south part of Great Salt Lake, Utah, 2005.
USGS Scientific Investigations Map 2005-2894. US
Geological Survey, Salt Lake City, Utah, USA.
hhttp://pubs.usgs.gov/sim/2005/2894i
Baskin, R. L., and J. Turner. 2006. Bathymetric map of
the north part of Great Salt Lake, Utah, 2006. USGS
Scientific Investigations Map 2954. US Geological
Survey, Salt Lake City, Utah, USA. hhttp://pubs.
usgs.gov/sim/2006/2954/i
Belovsky, G. E., and C. Larson. 2001. Brine shrimp
population dynamics and sustainable harvesting in
the Great Salt Lake, Utah. 2000 Progress Report to
the Utah Division of Wildlife Resources. Salt Lake
City, Utah, USA.
Belovsky, G. E., and C. Larson. 2002. Brine shrimp
population dynamics and sustainable harvesting in
the Great Salt Lake, Utah. 2001 Progress Report to
the Utah Division of Wildlife Resources. Salt Lake
City, Utah, USA.
Belovsky, G. E., C. Larson, and C. Mellison. 2000. Brine
shrimp population dynamics and sustainable har-
vesting in the Great Salt Lake, Utah. 2000 Progress
Report to the Utah Division of Wildlife Resources.
Salt Lake City, Utah, USA.
Belovsky, G. E., and C. Mellison. 1997. Brine shrimp
population dynamics and sustainable harvesting in
the Great Salt Lake. 1997 Progress Report to the
Utah Division of Wildlife Resources. Salt Lake City,
Utah, USA.
Belovsky, G. E., and C. Mellison. 1998. Brine Shrimp
Population Dynamics and Sustainable Harvesting
in the Great Salt Lake. 1998 Progress Report to the
Utah Division of Wildlife Resources. Salt Lake City,
Utah, USA.
Boetius, A., and S. Joye. 2009. Thriving in salt. Science
324:1523–1525.
Brock, T. D. 1975. Salinity and the ecology of Dunaliella
from Great Salt Lake. Journal of General Microbi-
ology 89:285–292.
Browne, A. A., and G. Wanigasekera. 2000. Combined
effects of salinity and temperature on survival and
reproduction of five species of Artemia. Journal of
Experimental Marine Biology and Ecology 244:29–
44.
Carini, S. A., and S. B. Joye. 2008. Nitrification in Mono
Lake, California: Activity and community compo-
sition during contrasting hydrological regimes.
Limnology and Oceanography 53:2546–2557.
Carpenter, S. R. 2003. Regime shifts in lake ecosystems:
pattern and variation. Volume 15 in the Excellence
in Ecology Series. Ecology Institute: Oldendorf/
Luhe, Germany.
Carpenter, S. R., J. J. Cole, J. R. Hodgson, J. F. Kitchell,
M. L. Pace, D. Bade, K. L. Cottingham, T. E.
vwww.esajournals.org 35 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
Essington, J. N. Houser, and D. E. Schindler. 2001.
Trophic cascades, nutrients and lake productivity:
whole-lake experiments. Ecological Monographs
71:163– 186.
Carpenter, S. R., and J. F. Kitchell, editors. 1993. The
trophic cascade in lakes. Cambridge University
Press, Cambridge, UK.
Carpenter, S. R., J. F. Kitchell, and J. R. Hodgson. 1985.
Cascading trophic interactions and lake productiv-
ity. BioScience 35:634 –639.
Caudell, J. N. 2001. Biology and management of eared
grebes (Podiceps nigricollis) on the Great Salt Lake,
Utah. Dissertation. Utah State University, Logan,
Utah, USA.
Caudell, J. N., and M. R. Conover. 2006. Behavioral
and physiological responses of Eared Grebes
(Podiceps nigricollis) to variations in brine shrimp
(Artemia franciscana) densities. Western North
American Naturalist 66:12–22.
Chase, J. M. 2000. Are there real differences among
aquatic and terrestrial food webs? TREE 15:408–
412.
Collins, N. C. 1977. Ecological studies of terminal
lakes: their relevance to problems in limnology.
Pages 411–420 in D. C. Greer, editor. Desertic
terminal lakes: Proceedings from the International
Conference on Desertic Terminal Lakes. Utah
Water Research Laboratory, Logan, Utah, USA.
Conover, M. R., and J. N. Caudell. 2009. Energy
budgets for Eared Grebes on the Great Salt Lake
and implications for harvest of brine Shrimp.
Journal of Wildlife Management 73:1134–1139.
Conover, M. R., J. Luft, and C. Perschon. 2009.
Concentrations of selenium in Eared Grebes from
the Great Salt Lake, Utah. Report to Utah Depart-
ment of Water Quality. Salt Lake City, Utah, USA.
Cuellar, O. 1990. Ecology of brine shrimp from Great
Salt Lake, Utah, U.S.A. (Branchiopoda, Anostraca).
Crustaceana 59:26–34.
Dana, G. L., R. Jellison, and J. M. Melack. 1990. Artemia
monica cyst production and recruitment in Mono
Lake, California, USA. Hydrobiologia 197:233–243.
Dana, G. L., R. Jellison, and J. M. Melack. 1995. Effects
of different natural regimes of temperature and
food on survival, growth and development of
Artemia monica Verrill. Journal of Plankton Re-
search 17:2117–2130.
Dana, G. L., R. Jellison, J. M. Melack, and G. L. Starrett.
1993. Relationships between Artemia monica life
history characteristics and salinity. Hydrobiologia
263:129–143.
Dana, G. L., and P. H. Lenz. 1986. Effects of increasing
salinity on an Artemia population from Mono Lake,
California. Oecologia 68:428–436.
DeAngelis, D. L. 1992. Dynamics of nutrient cycling
and food webs. Chapman and Hall, London, UK.
Felix, E. A., and S. R. Rushforth. 1977. The algal flora of
the Great Salt Lake: A preliminary report. Pages
385–392 in D. C. Greer, editor. Desertic terminal
lakes: Proceedings from the International Confer-
ence on Desertic Terminal Lakes. Utah Water
Research Laboratory, Logan, Utah, USA.
Felix, E. A., and S. R. Rushforth. 1979. The algal flora of
the Great Salt Lake, Utah, U.S.A. Nova Hedwigia
31:163–195.
Felix, E. A., and S. R. Rushforth. 1980. Biology of the
South Arm of the Great Salt Lake, Utah. Pages 305–
312 in J. W. Gwynn, editor. Great Salt Lake: a
scientific, historical, and economic overview. Utah
Geological and Mineral Survey Bulletin 116, Salt
Lake City, Utah, USA.
Fishman, M. J., and L. C. Friedman. 1989. Methods for
determination of inorganic substances in water and
fluvial sediments. USGS Techniques of Water-
Resources Investigations. TWRI Book. 5-A1:Reston,
Virginia, USA.
Fretwell, S. D. 1977. The regulation of plant commu-
nities by the food chains exploiting them. Perspec-
tives in Biology and Medicine 20:169–185.
Gafney, J. M. 2008. Tracking eared grebe (Podiceps
nigricollis) migration using fatty acid signature
analysis. Thesis, University of San Diego, San
Diego, California, USA.
Gliwicz, Z. M., W. Wurtsbaugh, and A. Ward. 1995.
Brine shrimp ecology in the Great Salt Lake. Report
to the Utah Division of Wildlife Resources, Salt
Lake City, Utah, USA.
Goldman, C. R., A. D. Jassby, and S. H. Hackley. 1993.
Decadal, interannual, and seasonal variability in
enrichment bioassays at Lake Tahoe, California–
Nevada, USA. Canadian Journal of Fisheries and
Aquatic Sciences 50:1489–1496.
Gotelli, N. J., and A. M. Ellison. 2004. A primer of
ecological statistics. Sinauer Associates, Inc., Sun-
derland, Massachusetts, USA.
Grace, J. B. 2006. Structural equation modeling and
natural systems. Cambridge University Press,
Cambridge, UK.
Great Salt Lake Planning Team. 1999. Great Salt Lake
Draft Comprehensive Management Plan, Novem-
ber 3, 1999. Utah Department of Natural Resources,
Salt Lake City, Utah, USA.
Great Salt Lake Planning Team. 2000. Great Salt Lake
Comprehensive Management Plan and Decision
Document, March 1, 2000. Utah Department of
Natural Resources, Salt Lake City, Utah, USA.
Grimm, N. B. 1987. Nitrogen dynamics during
succession in a desert stream. Ecology 68:1157–
1170.
Grimm, N. B., and S. G. Fisher. 1989. Stability of
periphyton and macroinvertebrates to disturbance
by flash floods in a desert stream. Journal of the
North American Benthological Society 8:293–307.
Gwynn, J. W., editor. 1980. Great Salt Lake: A scientific,
vwww.esajournals.org 36 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
historical, and economic overview. Utah Geological
and Mineral Survey Bulletin, 116, Salt Lake City,
Utah, USA.
Gwynn, J. W., editor. 2002. Great Salt Lake: An
overview of change. DNR Special Publication.
Utah Geological Survey, Salt Lake City, Utah, USA.
Hairston, N. G., Sr. 1989. Ecological experiments:
Purpose, design, and execution. Cambridge Uni-
versity Press, Cambridge, UK.
Hairston, N. G., F. E. Smith, and L. B. Slobodkin. 1960.
Community structure, population control, and
competition. American Naturalist 94:421–425.
Hairston, N. G., Jr., and N. G. Hairston, Sr. 1993.
Cause-effect relationships in energy flow, trophic
structure and interspecific interactions. American
Naturalist 142:379–411.
Hammer, U. T. 1986. Saline lake ecosystems of the
world. Monographiae Biologicae 59, Dr. W. Junk,
Publishers, Dordrecht, The Netherlands.
Hampton, S. E., M. D. Scheuerell, and D. E. Schindler.
2006. Coalescence in the Lake Washington story:
Interaction strengths in a planktonic food web.
Limnology and Oceanography 51:2042–2051.
Hampton, S. E., and D. E. Schindler. 2006. Empirical
evaluation of observation scale effects in commu-
nity time series. Oikos 113:424–439.
Hayes, C. R. 1971. Distribution, populations, and
species diversity of phytoplankton and zooplank-
ton of Farmington Bay. Pages E2–E21 in C. K.
Carter, editor. Some ecological considerations of
Farmington Bay estuary and adjacent Great Salt
Lake State Park. University of Utah, Salt Lake City,
Utah, USA.
Herbst, D. B., and T. J. Bradley. 1989. Salinity and
nutrient limitations on growth of benthic algae
from two alkaline salt lakes of the western great
basin (USA). Journal of Phycology 25:673–678.
Horne, A. J., and C. R. Goldman. 1994. Limnology.
Second Edition. McGraw-Hill, Inc., New York,
New York, USA.
Hunter, M. D., and P. W. Price. 1992. Playing chutes
and ladders: heterogeneity and the relative roles of
bottom-up and top-down forces in natural com-
munities. Ecology 73:724–732.
Isaacson, A. E., F. C. Hachman, and R. T. Robson. 2002.
The economics of Great Salt Lake. Pages 187–200 in
J. W. Gwynn, editor. Great Salt Lake: An overview
of change. DNR Special Publication. Utah Geolog-
ical Survey, Salt Lake City, Utah, USA.
Jehl, J. R., Jr. 1988. Biology of the Eared Grebe and
Wilson’s Phalarope in the nonbreeding season: A
study of adaptations to saline lakes. Studies in
Avian Biology No. 12:1–74.
Jehl, J. R., Jr. 2007. Why do Eared Grebes leave
hypersaline lakes in Autumn? Waterbirds 30:112–
115.
Jordan, D. S. 1889. Report of exploration of Colorado
and Utah during the summer of 1889. Bulletin of
the U. S. Fish Commission 9:31–68.
Knowles, N., M. D. Dettinger, and D. R. Cayan. 2006.
Trends in snowfall versus rainfall for the western
United States, 1949–2004. Journal of Climate
19:4545–4559.
Kuehn, D. 2002. The brine shrimp industry in Utah.
Pages 259–264 in J. W. Gwynn, editor. Great Salt
Lake: An overview of change. DNR Special
Publication. Utah Geological Survey, Salt Lake
City, Utah, USA.
Larson, C. A. 2004. Experimental examination of the
factors affecting growth and species of composition
of phytoplankton from Great Salt Lake, Utah.
Thesis, Utah State University, Logan, Utah, USA.
Larson, C. A., and G. E. Belovsky. 1999. Database for
Utah DNR—Great Salt Lake Literature. UDWR
Publication, Salt Lake City, Utah, USA.
Lewis, M. E. 2006. Dissolved Oxygen. National field
manual for the collection of water-quality data.
USGS Techniques of Water-Resources Investiga-
tions Book 9-A1–9. hhttp://pubs.water.usgs.gov/
twri9Ai
Likens, G. E., editor. 1985. An ecosystem approach to
aquatic ecology Mirror Lake and its environment.
Springer-Verlag, New York, New York, USA.
Lin, A. 1976a. A survey of physical limnology of the
Great Salt Lake. Utah Division of Water Resources
Technical Report.
Lin, A. 1976b. The Meromictic Great Salt Lake. Journal
of Great Lakes Research 2:374–383.
Lindeman, R. L. 1942. The trophic-dynamic aspect of
ecology. Ecology 23:399–418.
MacIntyre, S., J. F. Clark, R. Jellison, and J. P. Fram.
2009. Turbulent mixing induced by nonlinear
internal waves in Mono Lake, California. Limnol-
ogy and Oceanography 54:2255–2272.
Melack, J. M., and R. Jellison. 1998. Limnological
conditions in Mono Lake: contrasting monomixis
and meromixis in the 1990s. Hydrobiologia 384:21–
39.
Marcarelli, A. M., W. A. Wurtsbaugh, and O. Griset.
2006. Salinity controls phytoplankton response to
nutrient enrichment in the Great Salt Lake, Utah,
USA. Canadian Journal of Fisheries and Aquatic
Sciences 63:2236–2248.
Mellison, S. C. 2000. Functional response of a water-
boatman (Trichocorixa verticalis) and environmental
conditions that affect its distribution in the Great
Salt Lake, Utah, USA. Thesis, Utah State University,
Logan, Utah, USA.
Menge, B. A., F. Chan, S. Dudas, D. Eerkes-Medrano,
K. Grorud-Colvert, K. Heiman, M. Hessing-Lewis,
A. Iles, R. Milston-Clements, M. Noble, K. Page-
Albins, E. Richmond, G. Rilov, J. Rose, J. Tyburczy,
L. Vinueza, and P. Zarnetske. 2009. Terrestrial
ecologists ignore aquatic literature: Asymmetry in
vwww.esajournals.org 37 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
citation breadth in ecological publications and
implications for generality and progress in ecology.
Journal of Experimental Marine Biology and
Ecology 377:93–100.
Menge, B. A., and J. P. Sutherland. 1987. Community
regulation: variation in disturbance, competition,
and predation in relation to environmental stress
and recruitment. American Naturalist 130:730–757.
NADP. 2000. Nitrogen in the nation’s rain. Illinois State
Water Survey, Champaign, Illinois, USA.
Oksanen, L., S. D. Fretwell, J. Arruda, and P. Niemela.
1981. Exploitation ecosystems in gradients of
primary productivity. American Naturalist
118:240–261.
Pace, M. L. 1993. Heterotrophic microbial processes.
Pages 252–277 in S. R. Carpenter and J. F. Kitchell,
editors. The trophic cascade in lake ecosystems.
Cambridge University Press, Cambridge, UK.
Pack, D. A. 1919. Two ciliata of Great Salt Lake.
Biological Bulletin 36:273–282.
Paerl, H. W., and A. C. Yannarell. 2010. Environmental
dynamics, community structure and function in a
hypersaline microbial mat. Pages 421–442 in J.
Seckbach and A. Oren, editors. Microbial mats.
Springer, Berlin.
Pala, C. 2006. Once a terminal case, the north Aral Sea
shows new signs of life. Science 312:183–183.
Palmer, C. M., and T. E. Maloney. 1954. A new
counting slide for nannoplankton. American Soci-
ety of Limnology and Oceanography Special
Publication 21.
Paul, D. S., E. M. Annand, and J. Flory. 1999a. Great
Salt Lake Waterbird Survey: 1997 and 1998
Seasons. Great Salt Lake Ecosystem Project, Utah
Division of Wildlife Resources and Waterbird
Survey Cooperators, Publication No. 99-19. Salt
Lake City, Utah, USA.
Paul, D. S., J. Flory, and E. M. Annand. 1999b. 1997
Great Salt Lake Eared Grebe Photo Survey. Great
Salt Lake Ecosystem Project, Utah Division of
Wildlife Resources and Waterbird Survey Cooper-
ators, Publication No. 99-20. Salt Lake City, Utah,
USA.
Paul, D. S., and A. E. Manning. 2002. Great Salt Lake
Waterbird Survey Five-Year Report (1997–2001).
Publication Number 08-38. Utah Division of
Wildlife Resources, Salt Lake City, Utah, USA.
Pieters, R., and G. A. Lawrence. 2009. Effect of salt
exclusion from lake ice on seasonal circulation.
Limnology and Oceanography 54:401–412.
Pimm, S. L. 2002. Food webs. The University of
Chicago Press, Chicago, Illinois, USA.
Pimm, S. L., J. H. Lawton, and J. E. Cohen. 1991. Food
web patterns and their consequences. Nature
350:669–674.
Polis, G. A., and K. O. Winemiller, editors. 1996. Food
webs: Integration of patterns and dynamics.
Chapman and Hall, New York, New York, USA.
Porcella, D. B., and J. A. Holman. 1972. Nutrients, algal
growth, and culture of brine shrimp in the southern
Great Salt Lake. Pages 142–155 in J. P. Riley, editor.
The Great Salt Lake and Utah’s water resources.
Utah Water Research Laboratory, Logan, Utah,
USA.
Porter, K. G. 1976. Enhancement of algal growth and
productivity by grazing zooplankton. Science
192:1332–1334.
Porter, K. G., P. A. Saunders, K. A. Haberyan, A. E.
Macubbin, T. R. Jacobsen, and R. E. Hodson. 1996.
Annual cycle of autotrophic and heterotrophic
production in a small, monomictic Piedmont lake
(Lake Oglethorpe): Analog for the effects of
climatic warming on dimictic lakes. Limnology
and Oceanography 41:1041–1051.
Post, F. J. 1975. Life in the Great Salt Lake. Utah Science
36:43–47.
Post, F. J. 1980. Biology of the North Arm. Pages 313–
321 in J. W. Gwynn, editor. Great Salt Lake: a
scientific, historical, and economic overview. Utah
Geological and Mineral Survey Bulletin 116, Utah
DNR, Salt Lake City, Utah, USA.
Redfield, A. C. 1934. On the proportions of organic
derivatives in sea water and their relation to the
composition of plankton. Pages 176–192 in R. J.
Daniel, editor. James Johnstone Memorial Volume.
Liverpool University Press, Liverpool, UK.
Redfield, A. C. 1942. The processes determining the
concentration of oxygen, phosphate and other
organic derivatives within the depths of the
Atlantic Ocean. Papers in Physical Oceanography
and Meteorology 9:1–22.
Redfield, A. C. 1958. The biological control of chemical
factors in the environment. American Scientist
46:205–221.
Redfield, A. C., B. H. Ketchum, and F. A. Richards.
1963. The influence of organisms on the composi-
tion of seawater. Pages 26 –77 in M. N. Hill, editor.
Comparative and Descriptive Oceanography. Wi-
ley, New York, New York, USA.
Reese, B. K., and M. A. Anderson. 2009. Dimethyl
sulfide production in a saline eutrophic lake, Salton
Sea, California. Limnology and Oceanography
54:250–261.
Rushforth, S. R., and E. A. Felix. 1982. Biotic
adjustments to changing salinities in the Great Salt
Lake, Utah, USA. Microbial Ecology 8:157–161.
Sanders, H. L. 1969. Benthic marine diversity and the
stability-time hypothesis. Brookhaven Symposium
in Biology 22:71–80.
Schmitz, O. J. 1992. Exploitation in model food chains
with mechanistic consumer-resource dynamics.
Theoretical Population Biology 41:161–183.
Schmitz, O. J., P. Hamba¨ck, and A. P. Beckerman. 2000.
Trophic cascades in terrestrial systems: a review of
vwww.esajournals.org 38 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
the effect of top predator removals on plants.
American Naturalist 155:141–153.
Schoener, T. W. 1973. Population growth regulated by
intraspecific competition for energy or time: some
simple representations. Theoretical Population Bi-
ology 4:56–84.
Shipley, B. 1999. Testing causal explanations in
organismal biology: causation, correlation and
structural equation modelling. Oikos 86:374–382.
Shipley, B. 2000. Cause and correlation in biology: a
user’s guide to path analysis, structural equations
and causal inference. Cambridge University Press,
Cambridge, UK.
Slobodkin, L. B., and H. L. Sanders. 1969. On the
contribution of environmental predictability to
species diversity. Brookhaven Symposium in Biol-
ogy 22:82–93.
Slobodkin, L. B., F. E. Smith, and N. G. Hairston. 1967.
Regulation in terrestrial ecosystems, and the
implied balance of nature. American Naturalist
101:109–124.
Stephens, D. W. 1974. A summary of biological
investigations concerning the Great Salt Lake, Utah
(1861 – 1973). Great Basin Naturalist 34:221–229.
Stephens, D. W. 1976. Meromictic condition of the
Great Salt Lake, Utah. Abstract, 39
th
Annual
Meeting of American Society of Limnology and
Oceanography, June 21, 1976, Savannah, Georgia,
USA.
Stephens, D. W. 1990. Changes in lake levels, salinity,
and the biological community of Great Salt Lake
(Utah, USA), 1847–1987. Hydrobiologia 197:139–
146.
Stephens, D. W. 1998. Salinity-induced changes in the
aquatic ecosystem of Great Salt Lake, Utah. Pages
1–7 in J. Pitman and A. Carroll, editors. Modern
and ancient lake systems. Utah Geological Survey
Guidebook 26, Salt Lake City, Utah, USA.
Stephens, D. W., and P. W. Birdsey, 2002. Population
dynamics of the brine shrimp, Artemia franciscana,
in Great Salt Lake and regulation of commercial
shrimp harvest. Pages 327–336 in J. W. Gwynn,
editor. Great Salt Lake: An overview of change.
DNR Special Publication. Utah Geological Survey,
Salt Lake City, Utah, USA.
Stephens, D. W., and D. M. Gillespie. 1976. Phyto-
plankton production in the Great Salt Lake, Utah,
and a laboratory study of algal response to
enrichment. Limnology and Oceanography 21:74 –
87.
Sterner, R. W., T. Anderson, J. J. Elser, D. O. Hessen,
J. M. Hood, E. McCaughley, and J. Urabe. 2008.
Scale-dependent carbon: nitrogen: phosphorus
seston stoichiometry in marine and freshwaters.
Limnology and Oceanography 53:1169–1180.
Stewart, I. T., D. R. Cayan, and D. M. Dettinger. 2005.
Changes toward earlier streamflow timing across
the western North America. Journal of Climate
18:1136–1155.
Strong, D. R. 1992. Are trophic cascades all wet?
Differentiation and donor-control in speciose eco-
systems. Ecology 73:747–754.
Sturm, P. A., G. C. Sanders, and K. A. Allen. 1980. The
brine shrimp industry on the Great Salt Lake. Pages
243–248 in J. W. Gwynn, editor. Great Salt Lake: A
scientific, historical and economic overview. Utah
Geological and Mineralogical Survey Bulletin 116.
Utah DNR, Salt Lake City, Utah, USA.
Teng, J., and K. S. McCann. 2004. Dynamics of
compartmented and reticulate food webs in rela-
tion to energetic flows. American Naturalist
164:85–100.
Twombly, S., G. Wang, and T. Hobbs. 2007. Composite
forces shape population dynamics of copepod
crustaceans. Ecology 88:658–670.
Vanni, M. J., and P. C. De Ruiter. 1996. Detritus and
Nutrients in Food Webs. Pages 25–29 in G. A. Polis
and K. O. Winemiller, editors. Food webs, integra-
tion of patterns & dynamics. Chapman and Hall,
New York, New York, USA.
Wagner, C., and R. Adrian. 2009. Exploring lake
ecosystems: hierarchy responses to long-term
change. Global Change Biology 15:1104–1115.
Wagner, F., editor. 2003. Preparing for a changing
climate: the potential consequences of climate
variability and change. Rocky Mountain/Great
Basin regional climate-change assessment team.
U.S. Global Change Research Program. Utah State
University Press, Logan, Utah, USA.
Wardle, D. A. 2002. Communities and ecosystems:
Linking the aboveground and belowground Com-
ponents. Princeton University Press, Princeton,
New Jersey, USA.
Wear,R.G.,andS.J.Haslett.1986.Effectsof
temperature and salinity on the biology of Artemia
fransiscana Kellogg from Lake Grassmere, New
Zealand. 1. Growth and mortality. Journal of
Experimental Marine Biology and Ecology 98:153–
166.
Wear, R. G., and S. J. Haslett. 1987. Studies on the
biology and ecology of Artemia from Lake Grass-
mere, New Zealand. Pages 102–126 in P. Sorgeloos,
D. A. Bengtson, W. Decleir, and E. Jaspers, editors.
Artemia research and its applications. Volume 3:
Ecology, culturing, use in aquaculture. Universa
Press, Wetteren, Belgium.
Wear, R. G., S. J. Haslett, and N. Alexander. 1986.
Effects of temperature and salinity on the biology
of Artemia franciscana Kellogg from Lake Grass-
mere, New Zealand. 2. Maturation, fecundity, and
generation times. Journal of Experimental Marine
Biology and Ecology 98:167–183.
Weisser, W. W., and E. Siemann, editors. 2004. Insects
and ecosystem function. Ecological Studies. Spring-
vwww.esajournals.org 39 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
er-Verlag, Berlin, Germany.
Welschmeyer, N. A. 1994. Fluorometric analysis of
chlorophyll a in the presence of chlorophyll b and
phaeopigments. Limnology and Oceanography
39:1985–1992.
Wiebe, P. H., and W. R. Holland. 1968. Plankton
patchiness: effects on repeated net tows. Limnolo-
gy and Oceanography 13:315–321.
Wetzel,R.G.1983.Limnology.W.B.Saunders,
Philadelphia, Pennsylvania, USA.
Wetzel, R. G., and G. E. Likens. 1979. Limnological
analyses. W. B. Saunders Company, Philadelphia,
Pennsylvania, USA.
Williams, W. D. 1978. Limnology of Victorian Salt
Lakes, Australia. Verhandlungen des Internationa-
len Verein Limnologie 20:1165–1174.
Williams, W. D., A. J. Boulton, and R. G. Taaffe. 1990.
Salinity as a determinant of salt lake fauna: a
question of scale. Hydrobiologia 197:257–266.
Williams, W. D. 1993a. Conservation of salt lakes.
Hydrobiologia 267:291–306.
Williams, W. D. 1993b. The conservation of salt lakes—
important aquatic habitats of semiarid regions.
Aquatic Conservation—Marine and Freshwater
Ecosystems 3:71–72.
Williams, W. D. 1998. Salinity as a determinant of the
structure of biological communities in salt lakes.
Hydrobiologia 381:191–201.
Williams, W. D. 2002. Environmental threats to salt
lakes and the likely status of inland saline
ecosystems in 2025. Environmental Conservation
29:154–167.
Wirick, C. D. 1972. DunaliellaArtemia plankton com-
munity of the Great Salt Lake, Utah. Thesis,
University of Utah, Salt Lake City, Utah, USA.
Wurtsbaugh, W. A. 1988. Iron, molybdenum and
phosphorus limitation of N2 fixation maintains
nitrogen deficiency of plankton in the Great Salt
Lake drainage (Utah, USA). Verhandlungen des
Internationalen Verein Limnologie 23:121–130.
Wurtsbaugh, W. A. 1992. Food-web modification by an
invertebrate predator in the Great Salt Lake (USA).
Oecologia 89:168–175.
Wurtsbaugh, W. A. 1995. Brine shrimp ecology in the
Great Salt Lake, Utah. Report to Utah Division of
Wildlife Resources, December. The Utah Division
of Wildlife Resources, Salt Lake City, Utah, USA.
Wurtsbaugh, W. A., and T. S. Berry. 1990. Cascading
effects of decreased salinity on the plankton,
chemistry, and physics of the Great Salt Lake
(Utah). Canadian Journal of Fisheries and Aquatic
Sciences 47:100–109.
Wurtsbaugh, W. A., and Z. M. Gliwicz. 2001. Limno-
logical control of brine shrimp population dynam-
ics and cyst production in the Great Salt Lake,
Utah. Hydrobiologia 466:119–132.
Wurtsbaugh, W. A., and A. M. Marcarelli. 2004.
Analysis of phytoplankton nutrient limitation in
Farmington Bay and the Great Salt Lake. Report to
the Central Davis Sewer Improvement District,
Kaysville, Utah, USA.
Wurtsbaugh, W. A., and A. M. Marcarelli. 2006.
Eutrophication in Farmington Bay, Great Salt Lake,
Utah 2005 annual report. Report to the Central
Davis Sewer Improvement District, Kaysville,
Utah, USA.
Yannarell, A. C., and H. W. Paerl. 2007. Effects of
salinity and light on organic carbon and nitrogen
uptake in a hypersaline microbial mat. FEMS
Microbial Ecology 62:345–353.
Yannarell, A. C., T. F. Steppe, and H. W. Paerl. 2007.
Disturbance and recovery of microbial community
structure and function following Hurricane Fran-
ces. Environmental Microbiology 9:576–583.
SUPPLEMENT
Monthly data for the 13 years of our study (Ecological Archives C002-005-S1).
vwww.esajournals.org 40 March 2011 vVolume 2(3) vArticle 33
BELOVSKY ET AL.
