Potential Influence of Climate Change on Offshore Primary Production in Lake
S. Brooks and John C. Zastrow
Department of Biological Sciences and
Center for Great Lakes Studies
Great Lakes Water Institute
University of Wisconsin-Milwaukee
P.O. Box 413 Milwaukee, Wisconsin 53201
Change and Primary Production in Lake Michigan
ABSTRACT. This paper examines
the potential influence of climate change on the primary productivity of Lake
Michigan. Two general circulation models (GCMs) provided physical information
on projected regional climate for the years 2030, 2050 and 2090. A 30-year
record of meteorological data, centered on 1975, was used to define present,
BASE conditions for the lake. GCM output was used to develop scenarios of
future thermal characteristics, mixing patterns and surface irradiance, which
were then used to drive primary production calculations. Mean annual primary
production for the base period was 116 g C m-2. Under base conditions
thermal stratification of the lake occurred on June 13 and extended 135 days
until October 26. Conditions projected for 2090 showed the mean date of stratification
beginning by April 5 and remaining for 225 days until November 20. Estimated
mean annual primary production under these conditions, totaled 112 g C m-2,
a decrease of 3% from the mean base value. Under the most extreme conditions
of maximum projected cloud cover, primary production in 2090 could fall to
101 g C m-2,a decrease of 13% from the base mean. The projected
decrease was principally caused by physical/chemical constraints imposed on
spring primary production. Early stratification shortened the period of winter-spring
mixing, during which time nutrients from the sediment are transported to the
productive euphotic zone. The spring bloom was diminished when early stratification
capped the nutrient supply and increased cloud cover reduced light input for
photosynthesis. To a lesser extent fall production was also reduced by the
extension of the stratified period. Reduced primary production in the face
of climate change will be an important factor to consider in assessing the
food web dynamics of the lake and the future productivity of the fishery.
climate change, primary
production, Lake Michigan
The Laurentian Great
Lakes of North America are unique in limnological and ecological character
as well as their vast size. The lakes represent a diversity of habitats,
from small wetland communities to vast ocean-like expanses of deep water inhabited
by pelagic organisms. Under present conditions, the temperatures to which
the organisms of the lakes are exposed range from the freezing point of water
to upwards of 30° C in protected, nearshore areas. Offshore
surface dwellers may experience temperatures between 2 and 25° C, while inhabitants of deep basins may only experience an
annual change between 2 and 4° C.
A climatic warming with higher temperatures and altered
meteorological conditions could result in changes in the physical and chemical
nature of the lakes as well as the species composition in the ecosystem.
It has been suggested that species ranges could be altered by higher temperatures
and competition from northward-moving native and exotic warmwater species
intolerant of the present temperature regime (Magnuson et al 1990, Meisner
et al 1987).
Alterations of the annual thermal and mixing cycles
that are driven by solar heating and the wind could change the nature of the
physical and chemical environment of the lakes (Lehman 2001 this issue).
Extended periods of thermal stratification and reduced vertical mixing could
also alter nutrient flux from the bottom sediments and contribute to the degradation
of water quality in the hypolimnion (McCormick 1990). These factors could
all contribute to a change in the food web, the fishery it supports and the
overall Great Lakes ecosystem, as it exists today.
The principal primary producers in the open waters
of the Great Lakes are the photosynthetic phytoplanktonic algae. It is these
primary producers upon which consumer organisms depend for nourishment.
Primary production in the Great Lakes
is influenced by water temperature, sunlight, mixing, and nutrients such as
nitrogen, phosphorus and silicon.
Observations made over the past 3 decades indicate
that in winter the offshore areas of the larger lakes remain ice-free and
vertically mixed from top to bottom at temperatures at or below 4ºC. The
mixed water column is in contact with the bottom sediments that can supply
the nutrients needed for growth (Brooks and Edgington 1994). However, the low winter sun angle and the short day length reduce the
amount of sunlight reaching the lakes and limit photosynthesis and the rate
of primary production. The few algal cells that are in the water during winter
are mixed to depths greater than that to which the sunlight can penetrate,
so little primary production occurs under these conditions even though nutrients
As spring approaches, sunlight increases and penetrates
to greater depths. When light of high enough intensity reaches a “critical
depth” below the surface (Sverdrup, 1953), more carbon is fixed by photosynthesis than is consumed by respiration
and algal biomass increase rapidly during the spring bloom. As long as the
water column remains mixed to the bottom so phosphorus released from the sediments
can be mixed upward into the euphotic zone, positive, net primary production
will increase the biomass of the algae (Scavia et al 1986, Brooks and Edgington
1994). As soon as the surface waters warm above 4º C and thermal stratification
begins to set up, full mixing to the bottom ceases. Under these conditions,
the spring bloom ends due to a lack of new phosphorus entering the euphotic
zone from below, even though light intensity is still high enough to support
photosynthesis. In the autumn as surface waters cool, the mixed layer deepens
and nutrients are again mixed back to the surface. Now, however, light intensity
is on the wane as winter approaches and only a slight pulse of production
Regional climate acts as a master force on the physical/chemical
variables discussed above and, hence, primary production. Although the mechanisms
by which climate acts on these physical variables are complex, there is enough
known about critical linkages to assess the potential influence of climate
change on the food web of the lake. One way is to evaluate basic biological
processes, such as primary production. The composite processes that influence
primary production integrate the effects of physical, chemical and biological
changes that can be expected from climate change. As such, primary production
was selected as the biological process to be examined in this study with respect
to the effects of future climate change on the ecology of the lake.
Assessments of the impact of climate change on primary
production have been published for marine waters (Woods and Barkmann 1993,
Rowe and Baldauf 1995, Smith 1995), but there have been few, if any, specific
assessments of the effect of climate change on the primary productivity of
the Great Lakes. Studies on the Great Lakes have reported the influence of
seasonal and interannual variability in the physical and chemical factors
that influence primary production (Brooks and Torke 1977, Scavia et al 1986,
Brooks and Edgington 1994), but no fully integrated assessment is known to
Previous studies have used output from 2 X CO2
climate change scenarios to drive temperature, mixing and nutrient models
(McCormick 1990, Lehman, 2001 this issue, Blumburg and Di Toro 1990). General
results have predicted increased water temperatures, longer periods of warm
surface stratification, deeper depth of warming, and more extensive depletion
of oxygen from deep waters. McCormick (1990) estimated that under a warmer
climate scenario Lake Michigan could remain thermally stratified up to 2 months
longer than present and might not mix thoroughly during the winter. Such
conditions could lead to the development of a permanently isolated deep zone
with degraded water quality conditions.
Other studies have addressed the potential implications
for thermal habitats of Great Lakes fish. Magnuson et al (1990) concluded
that the size of the habitat favorable for cold-, cool- and warmwater fish
would increase in Lake Michigan, but habitats suitable only for cool-and warmwater
fish would increase in Lake Erie. Fish yields, estimated from empirical models
relating thermal habitat to sustained yields, remained about the same for
lake trout and lake whitefish, but increased for walleye.
Hill and Magnuson (1990) examined growth of lake trout,
yellow perch, and largemouth bass (cold, cool, and warm-water fish respectively)
at three nearshore sites in Lake Erie, Lake Michigan, and Lake Superior. Their
findings indicated that growth of yearling fish would increase with climate
warming if prey consumption also increased, but would decrease if prey consumption
remained constant. They noted that changes in growth would be most pronounced
in spring and fall due to the projected lengthening of the period of thermal
stratification, during which time habitats of differing temperatures are available
for fishes that can move to an area with appropriate temperatures for optimal
Estimated ratios of primary production, zooplankton
abundance and fishery yields developed for thermal conditions under 2 X CO2
to 1 X CO2 climate change scenarios ranged from an increase of
1.6 to 2.7 for phytoplankton production, from 1.3 to 2.3 for zooplankton biomass,
and from 1.4 to 2.2 for fishery yields (Hill and Magnuson, 1990). They note,
however, that the actual rates of primary and secondary production will depend
on a myriad of food web interactions. They further state that the dynamics
of Great Lakes must be considered in detail to answer the question of whether
increases in primary and secondary production will be sufficient to meet the
increased predatory demands of fishes.
The present study attempts to assess the influence
of potential climate change on the primary producers at the base of the Lake
Michigan food web. Producers that must be present in great enough abundance
to support prey species and any projected increase of fishery yield.
The intent of this assessment
was to compare primary production under “BASE” conditions against calculated
production for a future time under the influence of projected climate change
scenarios. In order to accomplish this, recent biological and physical data
were assembled to establish a baseline against which the influence of projected
climate changes could be compared. Figure 1 presents a flowchart illustrating
the sources of data used in this study.
used for the BASE reference was derived from meteorological data for the Great
Lakes region spanning the years 1961-1990 (Lofgren et al 2001 this issue).
From these data monthly maximum, minimum and mean values for cloud cover and
stratification dates were used to define the base conditions for Lake Michigan.
Output from two global
climate models (GCMs) was prepared by the NOAA Great Lakes Environmental Research
Laboratory (Lofgren et al 2001, this issue). Each model produced a set of
physical forcing functions for three time periods centered about the years
2030, 2050 and 2090. The two models, described by Sousounis, (2001 this issue),
were the Canadian Global Coupled Mode 1 (CGCM1) and the Hadley Centre Coupled
Model v2, (HadCM2). The latter utilized more input variables specific to
the Great Lakes region than the CGCM1. The resolution of both GCMs was such
that only two grid cells covered the whole of Lake Michigan. Data generated
for these cells that projected future climate conditions were averaged to
generate a composite result for the entire basin.
We recognize that differences
exist between nearshore and offshore areas of the lake and over longitudinal
gradients, however, these model projections were the best available at the
time of this study. The physical modeling provided a single date of stratification
for the entire lake, which was a necessary simplification of the known, complex
seasonal thermal structure. For example, northern sections of Lake Michigan
are known to stratify later in the season than more southerly locations (Bolgrien
and Brooks 1993) and nearshore areas warm and cool more rapidly than offshore
regions (Brooks et al 1990, Brooks and Sandgren 1995).
Data from the climate models were used by
Lehman (2001 this issue) to derive input variables needed for primary production
calculations. The physical variables derived from the GCMs used for the calculations
were the initial date and duration of whole-lake stratification, mixing depth
and monthly average photosynthetically available radiation (PAR) at ground
solar irradiance data were generated by scaling half-hour cloudless irradiance
curves from Fee (1990) for latitude N 43°, to match the monthly average value
predicted by the GCM-forced calculations. The derivation of total incident
short wave radiation data is presented by Lehman (2001, this issue). These
monthly values were then linearly interpolated at two-week intervals. A comparison
of irradiance values calculated using BASE data and data recorded hourly at
the Center for Great Lakes Studies in Milwaukee from April 1996 to June 2000
show good agreement (Figure 2).
[Figure 2] [Caption]
The sub-surface light extinction data and the biological variables
required as input for the calculation of integrated primary production using
the Fee application (Fee, 1990,1998) were derived from numerous studies at
a 100m deep station in Lake Michigan, referenced as Fox Point (W87.65, N43.22)
(Table 1). These studies spanned a period from 1985 through 1999 (Brooks et
al 1990, Brooks and Sandgren1995 and Center for Great Lakes Studies monitoring
cruises R. Cuhel, personal communication). Sub-surface light intensity data
observed at Fox Point were entered for depths of 5, 10, 20 and 30 m and interpolated programmatically
over time and depth for the indicated segments of the water column using the
methods described by Fee (1990, 1998).
Although there are few values for the biological
input variables reported in the literature for fall through early spring,
published values for summer conditions (Fahnenstiel and Scavia 1987) are comparable
with those used here for the stratified season.
Aerial primary production for a one square meter column of water was calculated
using the computer applications developed by Fee (1990) and updated on the
WEB (Fee 1998). Daily integral photosynthesis was determined with program
default settings (e.g. resolution of interpolation) using values describing
the vertical light gradient, chlorophyll a and photosynthesis versus
The calculations of primary production for projected climate change
scenarios were made using two different procedures. Initial calculations
were run using the GCM-based projected physical conditions for the lake, but
with mean BASE alpha and Pmax input values of 6.5 mg C « mg chl a –1
« Ein m–2 and
2.8 mg C hr-1 respectively, that did not change seasonally. Chlorophyll
a input values ranged from 0.5 to 3 mg m-3 and followed
seasonal BASE data. The second method utilized seasonal BASE P vs I inputs
adjusted for projected changes in physical conditions in the lake. For example,
under BASE conditions just prior to summer stratification a relatively high
concentration of chlorophyll is vertically homogeneous throughout the water
column, while after stratification surface chlorophyll values decrease and
a sub-thermocline chlorophyll peak begins to develop (Brooks and Torke 1977).
If the date of stratification was projected to occur earlier than the mean
BASE date, post-stratification BASE biological conditions were imposed on
the production calculation as of the projected date of stratification.
Primary production calculations determined using the
mean observational BASE data and biological data from the Fox Point station
showed that the spring bloom commenced in March when the lake was fully mixed,
increased rapidly through April and declined following stratification on or
about June 13 (Figure 3). The rising limb of the curve during spring mixing
reflected the increase in PAR as the season progressed. During this period,
PAR levels were under saturating for optimal photosynthesis (<Ik)
at depths >10m, while most of the algal biomass (measured as chlorophyll
a) was below that depth. Following thermal stratification and the
cessation of full vertical mixing, production declined. A slight pulse in
production occurred just after stratification that may be attributed to increased
chlorophyll in the sub-thermocline strata and continuing increases in PAR
input through the summer solstice. Another pulse was observed in late October
coincident with fall overturn. This pulse was short lived, however, due to
diminishing PAR values as winter approached. Low light inhibited production
throughout the winter months even though the water column was fully mixed
and nutrients were abundant.
Under these conditions, mean base annual primary production
was estimated to be 116 g C m-2. The range of production values
as influenced by BASE minimum and maximum cloud cover, which results in maximal
and minimal PAR, were shown to be 129 and 104 g C m-2 respectively
3] [Figure 4] [Caption]
Primary production calculated using mean annual P vs. I input
factors that were not changed seasonally, and the mean stratification and
cloud cover conditions projected for the year 2090 are superimposed on the
base production data illustrated in Figure 3. These data show the effect
of the earliest projected date of spring stratification and the latest date
of fall overturn (Table 2). Under these conditions annual primary production
was projected to decline to 100 g C m-2 yr-1, a decrease
of 14% from BASE values. These production calculations illustrate the hypothesized
influence of changes in the duration of thermal stratification in Lake Michigan,
independent of biological variability. The greatest loss of production occurred
during spring with the early truncation of the spring bloom, while the later
overturn in fall delayed the upward mixing of nutrients to a date when insufficient
light was present to support the fall pulse shown under BASE conditions.
Primary production calculated using mean annual biological
input factors that were changed seasonally according to the mean stratification
and cloud cover conditions projected for the years 2030, 2050 and 2090 are
illustrated in Figure 4. Both daily production estimates and cumulative values
over the year are plotted. In comparison to the BASE conditions with a mean
annual production of 116 g C m-2, projected annual production calculated
with mean outputs from the HADCM2 model show a decrease in primary production
of about 2% in 2030, 2% in 2050 and 3% by 2090 (Table 2). The CCM1 model projections
resulted in decreases of approximately 3% in 2030, 2050 and 2090. Using the
maximum cloud cover projected for 2090, annual production could decrease by
approximately 13% to 101 g C m-2 yr-1, whereas, with
minimum clouds, production was projected to increase 7% above the BASE value.
The anticipated changes in the physical characteristics
of Lake Michigan may impact primary production in two ways, both of which
are related to incoming solar radiation. First, altered light intensity, due
to an increase or decrease in cloud cover, would directly influence rates
of photosynthesis. Second, changes in incoming solar radiation could alter
surface warming and the thermal structure of the lake by extending or retarding
the onset and ending dates of stratification (McCormick 1990). Other meteorological
variables, such as wind and precipitation would also be important in determining
the physical characteristics of the lake (Lehman 2001 this issue) and/or the
overall hydrology and runoff inputs to the lake (Lofgren et al 2001 this issue).
Both GCMs used here suggest a warming of the lake and longer periods of stratification,
starting earlier in the spring and extending later into the fall.
The biological implications of the physical changes
predicted by the climate models suggest that for Lake Michigan the extended
duration of thermal stratification will reduce the duration of winter-spring
mixing. Given that most algal biomass is produced during the spring bloom
under well lit, vertically mixed, nutrient-replete conditions, any diminution
of the mixed period would be expected to reduce the amount of primary biomass
produced. Scavia et al. (1986) noted inter-annual decreases in algal biomass
caused by a shortened mixing period resulting from extensive ice cover. Brooks
and Torke (1977) reported that chlorophyll a was reduced by as much
as 40% at the end of the spring bloom in 1974 when stratification occurred
6 weeks earlier than that observed in the previous year. Such changes may
be indicative of future conditions under projected climatic change scenarios.
The BASE scenario that was determined from recent observations
of the lake represent the coolest conditions and the shortest period of
thermal stratification considered in this study. Under this scenario, the
mean date for the onset of thermal stratification was June 13, just prior
to the summer solstice, and extended for135 days through October 26. The mean
BASE annual production of 116 g C m-2, calculated for these conditions,
is comparable to published values for Lake Michigan (Fahnenstiel et al 1989,
Fahnenstiel and Scavia 1986).
The projected extension of stratification into the
spring bloom period was hypothesized to limit the bloom by placing a cap on
the reservoir of nutrients in the sediments that are carried into the euphotic
zone during spring mixing (Brooks and Edgington 1994).
A similar phenomenon was assumed to occur in fall when
stratification was projected to extend beyond the date at which light could
support production in the fully mixed water column. In other words in the
fall, if mixing and the release of nutrients from the bottom waters did not
commence until after the “critical depth” for light (Sverdrup 1953) rose above
the depth of mixing, there would be no net production realized at the time
of fall overturn. These reductions of the spring and to a lesser extent the
fall algal blooms, coupled with projected changes in cloud cover, appear to
be the principal causes of the climate-induced changes in primary production.
The significance of light variance alone, to an otherwise
identical annual production calculation, indicated that BASE production varied
±10% when the surface irradiance was modified using
maximum and minimum cloud conditions. Similar ranges were also seen about
the means of projected production under altered climate scenarios (Table 2).
The influence of the most extreme climate projections
on primary production are shown in data plotted for the year 2090 using projections
from CGCM1 (Figure 3, 4 and Table 2). Under this scenario stratification
was projected to occur on April 5 and extend for 225 days through November
20. The extended period of stratification, as compared with BASE conditions,
reduced the spring and fall production pulses when both the annual mean biological
input variables were used (Figure 3) and when they were varied seasonally
(Figure 4). In spring, early stratification was assumed to have imposed limiting
conditions on the phytoplankton. This was not as apparent as initially expected,
however, as production continued to climb after the projected date of stratification.
The continued increase in PAR beyond the early projected
date of thermal stratification appears to have strongly influenced the primary
production calculations independent of any limitation imposed by the application
of post-stratification P vs. I input variables indicative of nutrient limitation,
such as a lower Pbm. This is not unexpected given that mean light levels
at depths >10 m were estimated to still be less than saturating values
at the projected date of early stratification in early April. Any increase
in PAR beyond that date would be expected to result in increased production.
As PAR data input to the production calculation continued to increase, the
production estimates followed until light began to diminish in late June.
For example, daily average scalar irradiance at 10 m for January 1 under BASE
conditions was 32 mmol photons m-2 sec-1 and saturating
light intensity at the same depth (Ik) was 69. By June 24, that value increased
to an annual maximum of 110, and the June Ik was 117 mmol photons m-2 sec-1. These
numbers simply illustrate that for most of the year, irradiance at about 10
m and below is sub-saturating and phytoplankton production would be sensitive
to increases or decreases in light at or below about 10 m. For reference,
the BASE annual average Ik at 10 m in this study was 116 (annual
SD = 49). Annual Ik calculated from Fahnenstiel for Lake Michigan
(1989) were 108 (annual SD = 21).
Although the availability of nutrients does not directly
influence the production calculations presented here, with such an early date
of stratification projected the reservoir of nutrients remaining in the surface
waters after stratification would be expected to be greater than if the bloom
had been allowed to continue later into the spring. Normally, under BASE
conditions with stratification occurring in mid-June after an extended period
of mixing and production, nitrogen and silicon would be diminished leaving
little reserve for further production in the surface waters (Brooks and Edgington
1994). With the earlier projected date of stratification, this reserve would
be greater and could continue to fuel the spring bloom for a time after stratification
had occurred as PAR continued to increase. In fall the extended period of
stratification would delay the upward mixing of nutrients that occurs at overturn
to a date beyond which light could support any new production in the fully
mixed water column.
The results of this research suggest that primary production
in Lake Michigan will decline as climate warms. This decline will occur principally
as a result of increased duration of thermal stratification that will limit
the availability of nutrients in the lighted euphotic zone of the lake. When
these primary production results are coupled with the estimates of zooplankton
abundance and fishery yield developed by Hill and Magnuson (1990), they suggest
if the productivity of the lower food web is diminished, then fishery production
will also decline. The magnitude of the fishery decline will require more
detailed study of the intermediate links in the food web to better understand
the complexities of the system.
Compounding these predictions are unknowns, such as
changes in primary producer metabolism resulting from warmer conditions, possible
algal species shifts, changes in tributary runoff and nutrient inputs and
the invasion or introduction of new exotic species that could completely change
the structure of the food web as we know it today. Changes brought about
by exotic species have been well documented in the past with the invasion
of the alewife, sea lamprey, gobies, zebra mussels, Bythotrephes and
the stocking of exotic salmon. The effects of climate change alone on the
biological productivity of the Great Lakes would appear to be the easiest
to predict in the face of unknown invaders and to changes related to politically-driven
fishery management decisions.
The scientific community must respond to climate change
with an observational program that can detect the strengths and weaknesses
of these and other predictions. Studies will be required that integrate the
results of research conducted in many disciplines. Critical to our understanding
of the food web in the lakes will be knowledge of changing cloud cover and
wind patterns that may alter irradiance, nutrient dynamics, thermal cycles
and mixing patterns in the lakes. Links in the food web between the primary
producers and the top, economically important fish in the system must also
be examined in greater detail.
While there is much to be gained by effectively monitoring
present conditions, much knowledge can be gained by looking back at extant
historical records. The collection of good quality data at frequent intervals
will aid in addressing the research needs outlined above, while the analysis
of past conditions may elucidate inter-annual variance and extreme events
that will strengthen the validity of longer-term climate-coupled projections.
Blumburg, A.F. and D.M. Di Toro. 1990. Effects of climate
warming on dissolved oxygen concentrations in Lake Erie. Trans. Am. Fish.
D.W. and A.S. Brooks. 1992. Analysis of thermal features of Lake Michigan
from AVHRR satellite images. J. Great Lakes Res. 18:259-266.
Brooks, A.S. and C.D. Sandgren 1995. Nearshore Hydrodynamics Affecting
the Lake Michigan Coastal Waters of Wisconsin: Biological Productivity Along
Thermal Fronts in Coastal Waters. Final report to. NOAA Great Lakes Environmental
Research Laboratory, Ann Arbor, MI.
A.S. and D.N. Edgington. 1994. Biogeochemical control of phosphorus cycling
and primary production in Lake Michigan. Limnol. Oceanogr 39: 961-968
A.S. et al. 1990. Algal productivity measurements at Great Lakes water intakes.
Final report to Great Lakes National Program Office, U.S. Environmental Protection
Agency, Chicago, IL
Brooks, A.S. and B. G. Torke.
1977. Vertical and seasonal distribution of chlorophyll a in Lake
Michigan. J. Fish Res. Bd. Can. 34(12): 2281-2287
Fahnenstiel, G. L. J. F. Chandler,
H. J. Carrick and D. Scavia, 1989: Photosynthetic characteristics of phytoplankton
communities in Lakes Huron and Michigan: P-I parameters and end-products.
J. Great Lakes Res., 15: 394-407.
Fahnenstiel, G.L. and D. Scavia, 1986: Dynamics of Lake Michigan phytoplankton:
primary production and growth. Can. J. Fish. Aquat. Sci. 44: 499-508.
Fee, E.J. 1990. Computer Programs for Calculating
In Situ Phytoplankton Photosynthesis. Can. Tech. Rept. Fisheries and Aquatic
-------- 1998 Revision of 1990 Computer Program
for Calculating in-situ Phytoplankton Photosynthesis. At: http://www.umanitoba.ca/institutes/fisheries/PSpgms.html
Hill, D. and J.J. Magnuson,
1990: Potential effects of global climate warming on the growth and prey consumption
of Great Lakes fish. Trans. Am. Fish. Soc.119: 265-275.
Lehman, J.T. 2001. Mixing
dynamics and plankton biomass of the St. Lawrence Great Lakes under climate
change scenarios. J. Great Lakes Res. (This issue)
Lofgren, B.M. F.H. Quinn,
A. H. Clites and R. A. Assel. 2001. Climate change impacts on Great Lakes
basin water resources. J. Great Lakes Res. (This issue)
McCormick, M.J., 1990: Potential
changes in thermal structure and cycle of Lake Michigan due to global warming.
Trans. Am. Fish. Soc.119:183-194.
Magnuson, J.J., J.D. Meisner
and D.K. Hill, 1990: Potential changes in the thermal habitat of Great Lakes
fish after global climate warming. Trans. Am. Fish. Soc. 119: 254-264.
Meisner, D.J., J. L. Goodier, H. A. Regier, B.J. Shutter and J. Christie.
1987. An assessment of the effects of climate warming on Great Lakes basin
fishes. J. Great Lakes Res. 13:340-352.
Rowe, G.T. and J. G. Baldauf,
1995: Biofeedback in the ocean in response to climate change. 233-245. Biotic
feedbacks in the global climatic system. G.M. Woodwell and F. T. Makcenzie
eds. Oxford Univ. Press 416 p.
Scavia, D, G.L. Fahnenstiel,
M.S. Evans, D.J. Jude and J.T. Lehman, 1986: Influence of salmonine predation
and weather on long-term water quality trends in Lake Michigan. Can. J.
Fish. Aquat. Sci. 43: 435-443.
Smith, S.V., 1995. Net carbon metabolism of oceanic margins and estuaries.
246-250. Biotic feedbacks in the global climatic system. G.M. Woodwell
and F.T. Makcenzie eds. Oxford Univ. Press 416 p.
Sverdrup, H. U. 1953. On conditions
for the vernal blooming of phytoplankton.
Cons. Explor. Mer, 18, 287-295.
Woods, J. and W. Barkmann,
1993: The plankton multiplier—positive feedback in the greenhouse. J. Plankton
Res. 15: 1053-1074.