Full text of DOI:10.1016/S0380-1330(02)70608-4

J. Great Lakes Res. 28(4):597–607
Internat. Assoc. Great Lakes Res., 2002
The Potential Influence of Climate Change on Offshore Primary
Production in Lake Michigan
Arthur 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
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 and limno-
logical observations, from 1961 to through1990, was used to define present, baseline 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/m2. Under base conditions thermal stratification of the lake
occurred on 13 June and extended 135 days until 26 October. Conditions projected for 2090 showed the
mean date of stratification beginning as early as 5 April and remaining for 225 days until 20 November.
Estimated mean annual primary production under these conditions totaled 113 g C/m2, a decrease of 3%
from the mean base value. Under the most extreme conditions of maximum projected cloud cover for
2090, primary production in that year could fall to 101 g C/m2, a decrease of 13% from the base mean,
or down 22% from maximum base production calculated under minimum base cloud cover conditions.
The projected decrease may be attributed to physical/chemical constraints imposed on spring primary
production by altered climate conditions. Early stratification would shorten the period of winter-spring
mixing, during which time nutrients from the sediment are transported to the productive euphotic zone.
The spring bloom was projected to diminish if early stratification capped the nutrient supply, and
increased cloud cover reduced light input for photosynthesis. To a lesser extent fall production could also
be reduced by the extension of the stratified period. Altered physical/chemical conditions influenced by a
changing climate will be an important factor to consider in assessing future water quality conditions, pri-
mary production and the food web dynamics of the Great Lakes.
INDEX WORDS: Climate change, primary production, Lake Michigan.
INTRODUCTION
tures between 2 and 25°C, while inhabitants of deep
basins may only experience an annual change be-
tween 2 and 4°C.
The Laurentian Great Lakes of North America
are unique in limnological and ecological character
as well as their vast size. The lakes represent a di-
versity 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 tempera-
Observations made over the past 3 decades indi-
cate that in winter the offshore areas of the larger
lakes remain relatively free of ice 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 nutri-
ents needed for phytoplankton growth. However,
the low winter sun angle and the short day length
reduce the amount of sunlight reaching the lakes,
which limits photosynthesis and the rate of primary
*Corresponding author. E-mail: abrooks@uwm.edu
597

598
Brooks and Zastrow
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 are abundant.
nisms by which climate acts on these physical vari-
ables are complex, there is enough known about
critical linkages to assess the potential influence of
climate change on the food web of the lakes. One
way is to evaluate basic biological processes, such
as primary production. The composite processes
that influence primary production integrate the ef-
fects 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 Michigan.
The 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. Pri-
mary production in the Great Lakes is influenced by
water temperature, sunlight, mixing, and nutrients,
such as nitrogen, phosphorus, and silicon.
Assessments of the impact of climate change on
primary production have been published for marine
waters (Woods and Barkmann 1993, Rowe and Bal-
dauf 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 pri-
mary production (Brooks and Torke 1977, Scavia et
al. 1986, Brooks and Edgington 1994), but no fully
integrated assessment is known to exist.
Previous studies have used output from 2 × CO2
climate change scenarios to drive temperature, mix-
ing, and nutrient models (McCormick 1990,
Lehman 2002, Blumburg and Di Toro 1990). Gen-
eral results have predicted increased water tempera-
tures, longer periods of warm surface stratification,
a deeper depth of warming, and more extensive de-
pletion of oxygen from deep waters. McCormick
(1990) estimated that under a warmer climate sce-
nario Lake Michigan could remain thermally strati-
fied up to 2 months longer than present and might
not mix thoroughly during the winter. Such condi-
tions could lead to the development of a perma-
nently isolated deep zone with degraded water
quality conditions.
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 increases rapidly during the
spring bloom. As long as the water column remains
mixed to the bottom, phosphorus released from the
sediments will be mixed upward into the euphotic
zone and primary production will increase the bio-
mass of the algae (Scavia et al. 1986, Brooks and
Edgington 1994). As soon as the surface waters
warm above 4ºC and thermal stratification is estab-
lished, full mixing to the bottom ceases. Under
these conditions, the spring bloom diminishes due
to a lack of new phosphorus entering the euphotic
zone from below, even though light intensity is still
high enough to support continued primary produc-
tion. 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 occurs.
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 the range of
species could be altered by higher temperatures and
competition from northward-moving, native and ex-
otic warm-water species intolerant of the present
temperature regime (Magnuson et al. 1990, Meisner
et al. 1987).
Alterations of the annual thermal and mixing cy-
cles that are driven by solar heating and the wind,
could change the nature of the physical and chemi-
cal environment of the lakes (Lehman 2002). Ex-
tended periods of thermal stratification and reduced
vertical mixing could lead to the degradation of
water quality in the hypolimnion (McCormick
1990) and reduce nutrient flux from the bottom sed-
iments (Brooks and Edgington 1994). These factors
could all contribute to a change in the food web, the
fishery it supports and the Great Lakes ecosystem,
as it exists today.
Other studies have addressed the potential impli-
cations for thermal habitats of Great Lakes fish.
Magnuson et al. (1990) concluded that the size of
the habitats favorable for cold, cool, and warm-
water fish would increase in Lake Michigan. Fish
yields, estimated from empirical models relating
Regional climate acts as a master force on the
physical/chemical variables discussed above and,
hence, primary production. Although the mecha-

Climate Change and Primary Production in Lake Michigan
599
thermal habitat to sustained yields, remained about
the same for lake trout and lake whitefish, but in-
creased for walleye.
was derived from meteorological data for the Great
Lakes region spanning 30 years from 1 January
1961 through 31 December 1990 (Lofgren et al.
2002). From these data monthly maximum, mini-
mum and mean values for cloud cover and stratifi-
cation dates were used to define the base conditions
for Lake Michigan.
Projections from two global climate models
(GCMs) were prepared by the NOAA Great Lakes
Environmental Research Laboratory (Lofgren et al.
2002). Each model produced a set of physical forc-
ing functions for three time periods centered about
the years 2030, 2050, and 2090. The two models,
described by Sousounis (2002), were the Canadian
Global Coupled Model 1 (CGCM1) and the Hadley
Centre Coupled Model v2, (HadCM2). The resolu-
tion of both GCMs was such that only two grid
cells covered the whole of Lake Michigan. Data
generated from these cells were averaged to pro-
duce a composite result for Lake Michigan basin in-
dicative of future climate conditions.
Although differences exist between nearshore
and offshore areas of the lake and over longitudinal
gradients, 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 simplifica-
tion of the complex seasonal thermal structure
known to exist. For example, northern sections of
Lake Michigan stratify later in the season than more
southerly locations and nearshore areas warm and
cool more rapidly than offshore regions (Bolgrien
and Brooks 1992).
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 warm-
ing 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 tempera-
tures are available for fishes that can move to an
area with appropriate temperatures for optimal
growth.
Estimated ratios of primary production, zoo-
plankton abundance, and fishery yields developed
for thermal conditions under 2 × CO2 to 1 × 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
environmental 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 suffi-
cient to meet the increased predatory demands of
fishes.
Aerial primary production was estimated using
computer programs that were originally developed
for the calculation of primary production in Lake
Michigan (Fee 1969). The most recent published
versions of the Fee application were used for this
study (Fee 1990), as updated on the WEB (Fee
1998). The Fee method has been successfully used
in numerous studies of Great Lakes primary pro-
duction (Fahnenstiel and Scavia 1987, Fahnenstiel
et al. 1988, Brooks et al. 1990, Brooks and Sand-
gren 1995). Fahnenstiel et al. (1989) used similar
methods in their study of photosynthetic character-
istics of phytoplankton in Lakes Michigan and
Huron. Lean et al. (1987) estimated primary pro-
duction in Lake Ontario using the method of Lean
and Burnison (1979), and found agreement with
data independently derived by Cuhel and Lean
(1987) using the Fee calculation.
The present study attempts to assess the influence
of potential climate change on the primary produc-
ers at the base of the Lake Michigan food web; pro-
ducers that must be present in great enough
abundance to support prey species and any pro-
jected increase of fishery yield.
METHODS
The intent of this assessment was to compare pri-
mary production under recent conditions against
calculated production for a future time under the in-
fluence 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.
Daily depth-integrated primary production was
estimated using input variables of light extinction
Physical information used for the base reference

600
Brooks and Zastrow
FIG. 1. Data flowchart followed to determine projected primary production in Lake
Michigan. a) This project was supplied with refined climate data for Lake Michigan,
the processing of which is detailed in Loftgren et al. (2002). b) Thermal structure
and surface irradiance was determined by the methods detailed in Lehman (2002). c)
The biological inputs were averaged from multiple studies conducted at the Univer-
sity of Wisconsin – Milwaukee, Center for Great Lakes Studies. d) The application of
Fee (1990, 1998) was used to calculate rates of primary production.
coefficients to determine the vertical light gradient,
chlorophyll a concentrations as an estimate of phy-
toplankton biomass, and photosynthesis versus irra-
diance relationships (P vs. I plots) to assess the rate
of photosynthesis over a range of light intensities.
Daily solar irradiance data were generated by
scaling half-hour cloudless irradiance curves for
latitude N 43° from Fee (1990), 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
(2002). These monthly values were then linearly in-
terpolated at 2-week intervals. A comparison of ir-
radiance values calculated using baseline data and
data recorded hourly at the Center for Great Lakes
Studies in Milwaukee from April 1996 to June 2000
show good agreement (Fig. 2).
FIG. 2. Monthly average irradiance at ground
level calculated for baseline mean, minimum and
maximum cloud cover conditions over Lake
Michigan. Maximum and minimum cloud cover
translates to minimum and maximum irradiance
values, respectively. Daily average observed irradi-
ance was obtained from hourly data (April 1996
through June 2000) at the Center for Great Lakes
Studies, Milwaukee, WI (N 43°).
The sub-surface light extinction data and the bio-
logical variables required as input for the calcula-
tion of integrated primary production using the Fee

Climate Change and Primary Production in Lake Michigan
601
application (Fee 1990,1998) were derived from
studies at a 100-m 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 Sand-
gren 1995; R. Cuhel, Center for Great Lakes Stud-
ies, 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 indi-
cated segments of the water column using the meth-
ods described by Fee (1990, 1998).
The required biological input variables derived
from observations at the Fox Point station (Table 1)
are comparable to those reported in the literature
for the stratified season (Fahnenstiel and Scavia
1987). This lends confidence to the methods applied
to derive the values used here, although there are
few published values available for direct compari-
son during the winter-spring mixed period.
from March through May, and declined following
stratification on or about 13 June (Fig. 3). The ris-
ing limb of the curve during spring mixing was
strongly influenced by increasing photosyntheti-
cally active radiation, ((400–700 nm (PAR)), as
the season progressed. During this period, PAR
levels were under-saturating for optimal photosyn-
thesis at depths > 10 m. Following thermal stratifi-
cation and the cessation of full vertical mixing,
production declined. Production stabilized in July
as chlorophyll increased in the sub-thermocline
strata while PAR values were still near seasonal
maximum. A second pulse of production was ob-
served in late October coincident with the begin-
ning of fall mixing. 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
adequate to support further growth.
The calculations of primary production for the
baseline and projected climate change scenarios
were run using the baseline physical and biologi-
cal data, as well as with GCM-based physical con-
ditions projected for the years 2030, 2050, and
2090. Chlorophyll a input values ranged from 0.5
to 3 mg/m3 and followed the baseline data season-
ally. For the projected climate change scenarios,
seasonal baseline alpha and Pmax inputs were ad-
justed for projected changes in physical conditions
in the lake. For example, under baseline condi-
tions just prior to summer stratification relatively
high concentrations of chlorophyll are vertically
homogeneous throughout the water column, while
after stratification surface and deep chlorophyll
concentrations decrease as a sub-thermocline
chlorophyll peak begins to develop (Brooks and
Torke 1977). For a given climate change scenario,
if the date of stratification was projected to occur
earlier than the mean baseline date, post-stratifica-
tion baseline biological conditions were imposed
on the production calculation as of the projected
date of stratification.
Under these conditions, mean base annual pri-
mary production was estimated to be 116 g C/m2.
The range of production values, as influenced by
baseline minimum and maximum cloud cover,
which results in maximal and minimal PAR input,
were shown to be 129 and 104 g C/m2 respectively
(Fig. 3, Table 2).
Projected Production
Primary production calculated using mean sea-
sonal biological input factors that were changed ac-
cording to the stratification and cloud cover
conditions projected for the years 2030, 2050, and
2090 are illustrated in Figure 4. Both daily produc-
tion estimates and cumulative values over the year
are plotted. In comparison to the baseline condi-
tions, with a mean annual production of 116 g
C/m2, projected annual production calculated with
mean outputs from the HADCM2 model show a de-
crease in primary production of 2% in 2030, 2% in
2050, and 3% by 2090 (Table 2). The CGM1 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 from the base mean by approxi-
mately 13% to 101 g C/m2/yr, or down 22% in
comparison with the values calculated for minimum
base cloud cover. On the other hand, under the min-
imum cloud cover projected for 2090, production
was estimated to increase 7% above the mean base-
line value.
RESULTS
Base-Year Production
Primary production calculations determined
using the mean observational baseline physical
data and biological data from the Fox Point station
showed that the spring bloom commenced in Feb-
ruary when the lake was fully mixed at tempera-
tures < 4°C. The production increased rapidly

602
Brooks and Zastrow
TABLE 1. Average values from the 100-m-deep Fox Point station (W87.65, N43.22) in Lake Michigan
used for baseline calculations of primary production. Within the parentheses in the body of the table, the
first value is number of samples and the second is the standard deviation about the mean. NA refers to
values that were estimated.
Depth Interval
(m)
Chl-a
AlphaB
PBm
Depth
(m)
Light
(% Surface)
Month
Jan
(mg/m3)
(mg C/mg Chl a/Ein/m2) (mg C/mg Chl a/hr)
0–5
5–15
> 30
0.63 (2, 0.04)
0.85 (4, 0.11)
0.74 (6, 0.29)
6.78(NA)
6.72 (2,1.29)
8.47 (2,1.62)
1.7 (NA)
1.25 (2, 0.22)
1.34 (2, 0.23)
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
5
10
20
30
31.3 (3,13.3)
13.8 (3, 9.3)
6.9 (3,5.5)
1.5 (1)
34.2 (NA)
15.1 (NA)
7.0 (NA)
Feb
Mar
Apr
May
Jun
0–5
5–15
> 30
0.80 (6, 0.20)
0.82 (10, 0.31)
0.72 (11, 0.30)
7.38 (2, 2.81)
7.38 (NA)
7.38 (NA)
2.11 (2,0.09)
1.88 (6,0.56)
1.51 (NA)
1.1 (NA)
0–5
5–15
> 30
1.20 (8, 0.58)
1.27 (15, 0.47)
1.35 (14, 0.46)
7.18 (NA)
7.18 (NA)
7.18 (NA)
5.59 (2, 4.94)
3.26 (5, 3.34)
1.67 (NA)
37.1 (8, 6)
16.4 (8, 3.6)
7.1 (8, 2.1)
0.7 (5, 0.2)
34.5 (12, 12.1)
12.9 (12, 6.5)
5.2 (12, 3.7)
0.7 (6, 0.4)
26 (17, 13.8)
9.4 (17, 7.2)
3.5 (17, 2.4)
0.3 (17, 0.3)
30.3 (17, 10.2)
13.7 (17, 5.9)
6.5 (17, 3.1)
0.5 (16, 0.2)
29.6 (11, 9.3)
9.3 (10, 5.2)
3.2 (8, 2.4)
0.3 (2, 0.3)
21.4 (NA)
9.0 (NA)
3.2 (NA)
.2 (NA)
13.3 (4, 11.6)
8.6 (3, 7.3)
3.2 (3, 2.3)
.1 (2, 0)
18 (1)
7.6 (1)
3.2 (1)
0.3 (1)
0–5
5–15
> 30
2.32 (13, 0.85)
1.83 (14, 0.44)
2.14 (17, 0.49)
7.29 (2, 1.30)
4.73 (4, 3.8)
4.73 (NA)
1.26 (6, 1.10)
3.26 (8, 1.59)
1.84 (1)
0–5
5–15
> 30
2.26 (29, 0.48)
1.86 (16, 0.93)
1.61 (25, 0.53)
3.90 (24, 1.82)
7.51 (2, 0.95)
4.73 (NA)
2.15 (35, 1.54)
1.84 (2, 1.84)
2.99 (2, 1.38)
0–5
5–15
> 30
1.46 (18, 0.61)
1.72 (20, 0.78)
1.25 (25, 0.52)
6.10 (7, 4.26)
4.72 (4, 1.55)
4.72 (NA)
2.57 (14, 0.89)
2.12 (13, 0.87)
2.54 (1)
Jul
0–5
5–15
> 30
1.30 (15, 1.14)
1.37 (10, 0.64)
0.86 (16, 0.86)
2.68 (5, 1.14)
4.72 (NA)
4.72 (NA)
2 .00 (13, 0.98)
2.64 (1)
2.34 (NA)
Aug
Sep
Oct
Nov
Dec
0–5
5–15
> 30
1.00 (4, 0.12)
1.36 (4, 0.59)
0.76 (8, 0.37)
6.1(NA)
4.72 (NA)
4.72 (NA)
2.51 (NA)
3.42 (NA)
2.14 (NA)
0–5
5–15
> 30
1.28 (4, 0.48)
1.17 (4, 0.55)
0.73 (4, 0.06)
6.1(NA)
4.72 (NA)
4.72 (NA)
3.02 (NA)
4.21 (NA)
1.94 (NA)
0–5
5–15
> 30
0.68 (2, 0.18)
0.51 (3, 0.02)
0.07 (4, 0.06)
6.1(NA)
4.72 (NA)
4.72 (NA)
3.53 (NA)
4.99 (NA)
1.74 (NA)
0–5
5–15
> 30
0.99 (4, 0.14)
0.81 (6, 0.26)
0.67 (8, 0.54)
6.78(NA)
6.78(NA)
6.78(NA)
4.04 (2, 1.08)
3.29 (NA)
1.54 (NA)
27.6 (5, 13.5)
17.4 (5, 15.4)
8.0 (5, 9.5)
1.0 (3, 1.1)
12.1 (2, 7.7)
4.0 (2, 4.9)
1.5 (2, 2.1 )
0.2 (1)
0–5
5–15
> 30
1.06 (1)
1.16 (3, 0.05)
1.05 (4, 0.30)
6.78(NA)
6.78(NA)
6.78(NA)
2.48 (NA)
2.48 (NA)
2.48 (NA)

Climate Change and Primary Production in Lake Michigan
DISCUSSION
603
The anticipated changes in the physical charac-
teristics of Lake Michigan may impact primary pro-
duction in two ways, both of which are related to
incoming solar radiation. First, altered light inten-
sity, due to an increase or decrease in cloud cover,
were shown to directly influence rates of photosyn-
thesis. Second, changes in incoming solar radiation
altered the rate of 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 were important in determining the
physical characteristics of the lake (Lehman 2002)
and/or the overall hydrology and runoff inputs to
the lake (Lofgren et al. 2002). Both GCMs used
here suggest a warming of the lake and longer peri-
ods of thermal stratification, starting earlier in the
spring and extending later into the fall.
FIG. 3. Daily (stepped plot) and cumulative
annual (smooth curve) primary production for
Lake Michigan calculated using baseline light and
mean observed photosynthetic parameters. The
gray lines surrounding the dark lines illustrate the
production estimated under high and low cloud
cover observed in the baseline data. The area
between the vertical dashed lines represents the
period of thermal stratification under baseline
conditions.
The biological implications of the physical
changes predicted by the climate models suggest
that for Lake Michigan the extended duration of
TABLE 2. Results of primary production calculations determined for baseline conditions and for the
years 2030, 2050, and 2090 using climate change data from two global climate models.
Calculation
Scenario
Cloud
Cover
Annual Primary
Production (g/m2)
% BASE
MEAN
Stratification
Start Date
Stratification
End Date
Duration
(days)
BASE
Mean
Max
Min
Mean
Max
Min
Mean
Max
Min
Mean
Max
Min
Mean
Max
Min
Mean
Max
Min
Mean
Max
Min
116
104
129
113
101
125
112
101
125
113
102
124
114
102
126
113
102
126
112
101
125
100.0
90
111
97
87
108
97
87
107
97
88
106
98
88
108
98
88
108
97
13-Jun
13-May
1-May
5-Apr
26-Oct
7-Nov
7-Nov
20-Nov
3-Nov
6-Nov
10-Nov
135
177
190
225
155
162
189
CGCM1 2030
CGCM1 2050
CGCM1 2090
HADCM2 2030
HADCM2 2050
HADCM2 2090
31-May
28-May
5-May
87
107

604
Brooks and Zastrow
FIG 4. Daily (stepped plot) and cumulative annual (smooth curve) primary production for Lake
Michigan calculated using projected physical conditions derived from the Canadian Climate Cen-
tre (CGCM1) and Hadley (HadCM2) global climate models, and baseline biological variables
adjusted for the projected climate change for the years 2030, 2050, and 2090. The area between
the vertical dashed lines represents the projected period of thermal stratification. Baseline produc-
tion (gray line) is superimposed over conditions projected for 2090 by the CGCM1 model.

Climate Change and Primary Production in Lake Michigan
605
thermal stratification will reduce the duration of
winter-spring mixing. Given that most algal bio-
mass is produced during the spring bloom under
well-lit, vertically mixed, nutrient-replete condi-
tions, 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
during the winter of 1977. 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 ob-
served in the previous year. Such changes brought
about by extreme meteorological variability may be
indicative of potential future conditions under pro-
jected climatic change scenarios.
The baseline scenario that was determined from
recent observations of the lake represent the coolest
conditions and the shortest period of thermal strati-
fication considered in this study. Under this sce-
nario, the mean date for the onset of thermal
stratification was 13 June, just prior to the summer
solstice. The stratified period extended for 135 days
through 26 October. The mean baseline annual pro-
duction of 116 g C/m2, calculated for these condi-
tions, is comparable to published values for Lake
Michigan (Fahnenstiel et al. 1989, Fahnenstiel and
Scavia 1987).
Under the climate change scenarios, the projected
extension of stratified conditions into the spring
bloom period was hypothesized to limit production
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).
In fall, stratification was projected to extend be-
yond the date at which light could support produc-
tion in the fully mixed water column. In other
words in the fall, if mixing and the release of nutri-
ents 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.
projected production under the cloud cover pro-
jected by the climate change scenarios (Table 2).
The influence of the most extreme climate pro-
jections on primary production are shown in data
plotted for the year 2090 using projections from
CGCM1 (Fig. 4 and Table 2). Under this scenario
stratification was projected to occur on 5 April and
extend for 225 days through 20 November. The ex-
tended period of stratification, as compared with
baseline conditions, reduced the spring production
pulse, however, calculated production continued to
climb after the projected date for summer stratifica-
tion. The seasonal increase in PAR, which was not
diminished by the onset of stratification, appears to
have strongly influenced the primary production
calculations independent of any limitation imposed
by the application of post-stratification biological
input variables indicative of nutrient limitation,
such as a lower Pmax. This is not unexpected given
that mean light levels at depths > 10 m were esti-
mated to be less than saturating values at the pro-
jected date of early stratification in early April. Any
increase in PAR beyond that date would be ex-
pected 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 1 Janu-
ary under baseline conditions was 32 µmol pho-
tons/m2/sec, while saturating light intensity (Ik) at
the same depth was estimated to be 69 µmol-pho-
tons/m2/sec. By 24 June, daily average scalar irradi-
ance at 10 m increased to an annual maximum of
110 µmol photons/m2/sec, and Ik was 117 µmol
photons/m2/sec. These numbers simply illustrate
that for the period of spring phytoplankton bloom,
irradiance at 10 m and below is sub-saturating and
phytoplankton production would be highly sensitive
to increases or decreases in PAR.
Although the availability of nutrients does not di-
rectly influence the production calculations pre-
sented here, with such an early date of stratification
projected, the reservoir of nutrients remaining in
the surface waters after stratification would be ex-
pected to be greater than if the bloom had been al-
lowed to continue later into the spring. Normally,
under baseline conditions with stratification occur-
ring in mid-June after an extended period of mixing
and production, nitrogen and silicon would be di-
minished leaving little reserve for further produc-
tion in the surface waters (Brooks and Edgington
1994). With the earlier projected date of stratifica-
tion, this reserve would be greater and could con-
Light variance alone changed baseline production
±10% when the surface irradiance was modified
using maximum and minimum cloud conditions.
Similar ranges were also seen about the means of

606
Brooks and Zastrow
tinue to fuel the spring bloom for a time after strati-
fication had occurred and PAR continued to in-
crease.
elucidate inter-annual variance and extreme events
that will strengthen the validity of longer-term cli-
mate-coupled projections.
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 euphotic
zone of the lake. When these primary production re-
sults 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 pro-
duction will also decline. The magnitude of the fish-
ery decline will require more detailed study of the
intermediate links in the food web to better under-
stand the complexities of the system.
Compounding these predictions are unknowns,
such as changes in primary producer metabolism re-
sulting from warmer conditions, possible algal
species shifts, changes in tributary runoff and nutri-
ent inputs, and the invasion or introduction of new
exotic species that could completely change the
structure of the food web as it is known 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
(Mills et al. 1993). 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 cli-
mate 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 an understanding of the food
web in the lakes will be knowledge of changing
cloud cover and wind patterns that may alter irradi-
ance, 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.
ACKNOWLEDGMENTS
We thank J. Lehman for his collaboration in de-
veloping the physical input variables for the pro-
duction calculations and for his helpful comments
over the life of this project. R. Cuhel provided the
results of ongoing primary production research in
the lake. The comments of P. Bertram in review
were most helpful. P. Sousounis and J. Bisanz were
extremely helpful with the administrative aspects of
the project and assembling the final manuscript. E.
Trendel and J. Flores provided superb office logisti-
cal support. This study was sponsored by the U.S.
Environmental Protection Agency as part of the Na-
tional Climate Assessment Program. John Zastrow
was supported for his MS research by the Paul
Frederick Memorial Fellowship of the Great Lakes
Cruising Club, Great Lakes Foundation.
REFERENCES
Blumburg, A.F., and Di Toro, D.M. 1990. Effects of cli-
mate warming on dissolved oxygen concentrations in
Lake Erie. Trans. Am. Fish. Soc. 119:210–213
Bolgrien, D.W., and Brooks, A.S. 1992. Analysis of
thermal features of Lake Michigan from AVHRR
satellite images. J. Great Lakes Res. 18:259–266.
Brooks, A.S., and Edgington, D.N. 1994. Biogeochemi-
cal control of phosphorus cycling and primary pro-
duction in Lake Michigan. Limnol. Oceanogr 39:
961–968
, and Sandgren, C.D. 1995. Nearshore Hydrody-
———
namics 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.
, and Torke, B.G. 1977. Vertical and seasonal dis-
tribution of chlorophyll a in Lake Michigan. J. Fish
Res. Board Can. 34(12):2281–2287
———
, Edgington, D.N., Warne, E.L., Back, R.C., Bol-
———
grien, D.W., and Galarneau, S.G. 1990. Algal produc-
tivity measurements at Great Lakes water intakes.
Final report R005880-01 to Great Lakes National Pro-
gram Office, U.S. Environmental Protection Agency,
Chicago, IL
There is much to be gained by monitoring present
conditions in order to maintain databases that can
be used to detect changes as they occur, while
building upon the extant historical records. The col-
lection of good quality data at frequent intervals
will aid in addressing the research needs outlined
above, while the analysis of past conditions may
Cuhel, R.L., and Lean, D.R.S. 1987. Influence of light
intensity, light quality, temperature and daylength on
uptake and assimilation of carbon dioxide and sulfate
by lake plankton. Can. J. Fish. Aquat. Sci. 44:
2118–2132.

Climate Change and Primary Production in Lake Michigan
607
Fahnenstiel, G.L., and Scavia, D. 1987: Dynamics of
Lake Michigan phytoplankton: primary production
and growth. Can. J. Fish. Aquat. Sci. 44:499–508.
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., Meisner, J.D., and Hill, D.K. 1990.
Potential changes in the thermal habitat of Great
Lakes fish after global climate warming. Trans. Am.
Fish. Soc. 119:254–264.
, Scavia, D., Lange, G.A., Saylor, J.H., Miller,
———
G.S., and Schwab, D.J. 1988. Impact of inertial period
internal waves on fixed-depth primary production
estimates. J. Plankton Res. 10:77–87.
, Chandler, J.F., Carrick, H.J., and Scavia, D.
Meisner, D.J., Goodier, J.L., Regier, H.A., Shutter, B.J.,
and Christie, J. 1987. An assessment of the effects of
climate warming on Great Lakes basin fishes. J. Great
Lakes Res. 13:340–352.
Mills, E.L., Leach, J.H., Carlton, J.T., and Secor, C.L.
1993. Exotic species in the Great Lakes: a history of
biotic crises and anthropogenic introductions. J. Great
Lakes Res. 19:1–54.
Rowe, G.T., and Baldauf, J.G. 1995: Biofeedback in the
ocean in response to climate change. In Biotic feedbacks
in the global climatic system, G.M. Woodwell and F. T.
Makcenzie, eds., pp. 233–245. Oxford Univ. Press.
Scavia, D, Fahnenstiel, G.L., Evans, M.S., Jude, D.J.,
and Lehman, J.T. 1986: Influence of salmonine preda-
tion 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. In Biotic feedbacks in the
global climatic system, G.M. Woodwell and
F.T. Makcenzie eds., pp. 246–250. Oxford Univ.
Press.
Sousounis, P.J., and Grover, E.K. 2002. Potential future
weather patterns over the Great Lakes region. J. Great
Lakes Res. 28:496–520.
Sverdrup, H. U. 1953. On conditions for the vernal
blooming of phytoplankton. J. Cons. Explor. Mer 18,
287–295.
Woods, J., and Barkmann, W. 1993. The plankton multi-
plier—positive feedback in the greenhouse. J. Plank-
ton Res. 15:1053-1074.
———
1989: Photosynthetic characteristics of phytoplankton
communities in Lakes Huron and Michigan: P-I para-
meters and end-products. J. Great Lakes Res. 15:
394–407.
Fee, E.J. 1969. A numerical model for the estimation of
integral photosynthetic production, integrated over
time and depth in natural waters. Limnol. Oceanogr.
14:906–911
. 1990. Computer Programs for Calculating In
Situ Phytoplankton Photosynthesis. Can. Tech. Rept.
Fisheries and Aquatic Sci. 1740
———
. 1998. Revision of 1990 Computer Program for
———
Calculating in-situ Phytoplankton Photosynthesis. At:
http://www.umanitoba.ca/institutes/fisheries/PSpg
ms.html
Hill, D., and Magnuson, J.J. 1990: Potential effects of
global climate warming on the growth and prey con-
sumption of Great Lakes fish. Trans. Am. Fish.
Soc.119:265–275.
Lean, D.R.S., and Burnison, B.K. 1979. An evaluation of
errors in the 14C method of primary production mea-
surement. Limnol. Oceanogr. 24:918–928.
, Fricker, J.H., Charlton, M.N., Cuhel, R.L., and
Pick, E.R. 1987. The Lake Ontario life support sys-
tem. Can. J. Fish. Aquat. Sci. 44:2230–2240.
———
Lehman, J.T. 2002. Mixing patterns and plankton bio-
mass of the St. Lawrence Great Lakes under climate
change scenarios. J. Great Lakes Res. 28:583–596.
Lofgren, B.M. Quinn, F.H., Clites, A.H., Assel, R.A.,
Eberhardt, A.J., and Luukkonen, C.L. 2002. Evalua-
tion of potential impacts on Great Lakes water
resources based on climate scenarios of two GCMs. J.
Great Lakes Res. 28:537–554.
Submitted: 19 October 2000
Accepted: 9 March 2002