&EPA
United States
Environmental Protection
Agency
Policy, Planning,
And Evaluation
(PM-221)
EPA-230-05-89-055
June 1989
The Potential Effects
Of Global Climate Change
On The United States
Appendix E
Aquatic Resources
-------
THE POTENTIAL EFFECTS OF GLOBAL CLIMATE CHANGE
VN ON THE UNITED STATES:
APPENDIX E • AQUATIC RESOURCES
Editors: Joel B. Smith and Dennis A. Tirpak
OFFICE OF POLICY, PLANNING AND EVALUATION
US. ENVIRONMENTAL PROTECTION AGENCY
WASHINGTON, DC 20460
MAY 1989
Text Printed on Recycled Paper
-------
TABLE OF CONTENTS
APPENDIX E: AQUATIC RESOURCES
PREFACE iii
THE EFFECTS OF GLOBAL CLIMATE CHANGE ON THE WATER QUALITY
OF MOUNTAIN LAKES AND STREAMS 1-1
Earl R. Byron, Alan Jassby, and Charles R. Goldman
POTENTIAL RESPONSES OF GREAT LAKES FISHES AND THEIR HABITAT
TO GLOBAL CLIMATE WARMING 2-1
John J. Magnuson, David K. Hill, Henry A. Regier, John A. Holmes,
J. Donald Meisner, and Brian J. Shuter
ECOLOGICAL EFFECTS OF GLOBAL CLIMATE CHANGE: WETLAND
RESOURCES OF SAN FRANCISCO BAY 3-1
Michael Josselyn and John Callaway
PROJECTED CHANGES IN ESTUARINE CONDITIONS BASED ON MODELS OF
LONG-TERM ATMOSPHERIC ALTERATION 4-1
Robert J. Livingston
-------
PREFACE
The ecological and economic implications of the greenhouse effect have been the subject of discussion within
the scientific community for the past three decades. In recent years, members of Congress have held hearings
on the greenhouse effect and have begun to examine its implications for public policy. This interest was
accentuated during a series of hearings held in June 1986 by the Subcommittee on Pollution of the Senate
Environment and Public Works Committee. Following the hearings, committee members sent a formal request
to the EPA Administrator, asking the Agency to undertake two studies on climate change due to the greenhouse
effect.
One of the studies we are requesting should examine the potential health and environmental
effects of climate change. This study should include, but not be limited to, the potential impacts
on agriculture, forests, wetlands, human health, rivers, lakes, and estuaries, as well as other
ecosystems and societal impacts. This study should be designed to include original analyses, to
identify and fill in where important research gaps exist, and to solicit the opinions of
knowledgeable people throughout the country through a process of public hearings and
meetings.
To meet this request, EPA produced the report entitled The Potential Effects of Global Climate Change on the
United States. For that report, EPA commissioned fifty-five studies by academic and government scientists on
the potential effects of global climate change. Each study was reviewed by at least two peer reviewers. The
Effects Report summarizes the results of all of those studies. The complete results of each study are contained
in Appendices A through J.
Appendix Subject
A Water Resources
B Sea Level Rise
C Agriculture
D Forests
E Aquatic Resources
F Air Quality
G Health
H Infrastructure
I Variability
J Policy
GOAL
The goal of the Effects Report was to try to give a sense of the possible direction of changes from a global
warming as well as a sense of the magnitude. Specifically, we examined the following issues:
o sensitivities of systems to changes in climate (since we cannot predict regional climate change, we
can only identify sensitivities to changes in climate factors)
o the range of effects under different warming scenarios
o regional differences among effects
o interactions among effects on a regional level
iii
-------
o national effects
o uncertainties
o policy implications
o research needs
The four regions chosen for the studies were California, the Great Lakes, the Southeast, and the Great Plains.
Many studies focused on impacts in a single region, while others examined potential impacts on a national scale.
SCENARIOS USED FOR THE EFFECTS REPORT STUDIES
The Effects Report studies used several scenarios to examine the sensitivities of various systems to changes in
climate. The scenarios used are plausible sets of circumstances although none of them should be considered to
be predictions of regional climate change. The most common scenario used was the doubled CO2 scenario
(2XCO2), which examined the effects of climate under a doubling of atmospheric carbon dioxide concentrations.
This doubling is estimated to raise average global temperatures by L5 to 4-5°C by the latter half of the 21st
century. Transient scenarios, which estimate how climate may change over time in response to a steady increase
in greenhouse gases, were also used. In addition, analog scenarios of past warm periods, such as the 1930s, were
used.
The scenarios combined average monthly climate change estimates for regional grid boxes from General
Circulation Models (GCMs) with 1951-80 climate observations from sites in the respective grid boxes. GCMs
are dynamic models that simulate the physical processes of the atmosphere and oceans to estimate global climate
under different conditions, such as increasing concentrations of greenhouse gases (e.g., 2XCO2).
The scenarios and GCMs used in the studies have certain limitations. The scenarios used for the studies assume
that temporal and spatial variability do not change from current conditions. The first of two major limitations
related to the GCMs is their low spatial resolution. GCMs use rather large grid boxes where climate is averaged
for the whole grid box, while in fact climate may be quite variable within a grid box. The second limitation is
the simplified way that GCMs treat physical factors such as clouds, oceans, albedo, and land surface hydrology.
Because of these limitations, GCMs often disagree with each other on estimates of regional climate change (as
well as the magnitude of global changes) and should not be considered to be predictions.
To obtain a range of scenarios, EPA asked the researchers to use output from the following GCMs:
o Goddard Institute for Space Studies (GISS)
o Geophysical Fluid Dynamics Laboratory (GFDL)
o Oregon State University (OSU)
Figure 1 shows the temperature change from current climate to a climate with a doubling of CO2 levels, as
modeled by the three GCMs. The figure includes the GCM estimates for the four regions. Precipitation
changes are shown in Figure 2. Note the disagreement in the GCM estimates concerning the direction of
change of regional and seasonal precipitation and the agreement concerning increasing temperatures.
Two transient scenarios from the GISS model were also used, and the average decadal temperature changes
are shown in Figure 3.
iv
-------
FIGURE 1. TEMPERATURE SCENARIOS
GCM Estimated Change in Temperature from 1xCO2 to 2xCO2
8
8
Great Southeast Great California United
Lakes Plains States'
Great Southeast Great California United
Lakes Plains States*
4-
2
SUMMER
Great Southeast Great California United
Lakes Plains States'
GISS
GFDL
OSU
* Lower 48 States
-------
FIGURE 2. PRECIPITATION SCENARIOS
GCM Estimated Change in Precipitation from 1xCC>2 to 2xCO2
<
Q
0)
QC
HI
I.U
0.8
0.6
0.4
0.2
.0
-0.2-
•0.4-
-fl.fi.
ANNUAL
^nln l^ftfl
lP
Great Southeast Great California United
Lakes Plains States*
1.0
0.8
0.6
0.4
0.2
0.0
-0.2
-0.4
-0.6
WINTER
I
Great Southeast Great California United
Lakes Plains States*
1.0
0.8-
0.6-
0.4-
0.2-
0.0
-0.2-
-0.4-
-0.6
No
Change
SUMMER
Great Southeast Great California Unite
Lakes Plains State
GISS
GFDL
OSU
* Lower 48 States
-------
4
3.5
O
g 2.5
UJ
UJ
1.5
1
0.5
0
o
3.5
3
2.5
flC
UJ _
a. 1.5
UJ
I- 1
0.5
0
3.72
2.99
2.47
1.72
1.36
0.70
0.88
0.30
1980s 1990s 2000s 2010s 2020s 2030s 2040s 2050s
TRANSIENT SCENARIO A
1.26
1.(
0.59
0.18
0.35
\/ / / A
Y77/
1980s 1990s 2000s 2010s
TRANSIENT SCENARIO B
2020s
FIGURE 3
GISS TRANSIENTS "A" AND "B" AVERAGE
TEMPERATURE CHANGE FOR LOWER 48 STATES
GRID POINTS.
vu
-------
EPA specified that researchers were to use three doubled CO, scenarios, two transient scenarios, and an analog
scenario in their studies. Many researchers, however, did not nave sufficient time or resources to use all of the
scenarios. EPA asked the researchers to run the scenarios in the following order, going as far through the list
as time and resources allowed:
1. GISS doubled CO2
2. GFDL doubled CO2
3. GISS transient A
4. OSU doubled CO2
5. Analog (1930 to 1939)
6. GISS transient B
ABOUT THESE APPENDICES
The studies contained in these appendices appear in the form that the researchers submitted them to EPA.
These reports do not necessarily reflect the official position of the ILS. Environmental Protection Agency.
Mention of trade names does not constitute an endorsement.
vm
-------
THE EFFECTS OF GLOBAL CLIMATE CHANGE ON THE WATER QUALITY
OF MOUNTAIN LAKES AND STREAMS
by
Earl R. Byron
Alan Jassby
Charles R. Goldman
Tahoe Research Group
Division of Environmental Studies
University of California, Davis
Davis, CA 95616
Contract No. CR-814581-01-0
-------
CONTENTS
Page
FINDINGS 1-1
CHAPTER 1: INTRODUCTION 1-2
PERSPECTIVE 1-2
SITE DESCRIPTIONS 1-3
CHAPTER 2: METHODS ' 1-4
THE EMPIRICAL EFFECTS MODELS 1-4
Castle Lake 1-4
Snowfall Estimates 1-5
Model Using Estimated Snowfall 1-5
Ward Creek, Lake Tahoe 1-6
Limitations of the Empirical Models 1-6
THE CLIMATE CHANGE SCENARIOS 1-18
The Scenarios Used 1-18
Issues Resulting From the Scenarios 1-24
CHAPTER 3: RESULTS 1-25
CASTLE LAKE 1-25
WARD CREEK, LAKE TAHOE 1-25
CHAPTER 4: INTERPRETATION 1-33
CHAPTER 5: IMPLICATIONS OF RESULTS 1-34
ENVIRONMENTAL 1-34
SOCIOECONOMIC 1-34
CHAPTER 6: POLICY IMPLICATIONS 1-35
CHAPTER 7: FUTURE RESEARCH 1-36
REFERENCES 1-37
-------
Byron
FINDINGS1
We researched the effects of global climate change on the water quality of subalpine lakes and streams in
California Using our long-term data bases of Castle Lake and Ward Creek, Lake Tahoe, both in the mountains
of northern California, we were able to relate our water quality records to climatic variables. We then applied
several scenarios of global warming to the empirical models to predict the potential effects of climatic warming
on water quality in these systems.
Our results indicated that the changing precipitation patterns and global warming predicted by various
climate change scenarios would have a detrimental effect on water quality for our systems. Algal growth rates
(primary productivity) in Castle Lake increased under all scenario projections. Average increases ranged from
16% to 87% over control values. The enhanced algal productivity under scenario conditions resulted from the
combined effects of an increase in the length of the growing season and a change in hydraulic loading. Our
preliminary analysis of Ward Creek did not reveal an increase in runoff and nutrient concentrations in the
scenario conditions. However, since we were not able to incorporate temperature change in our Ward Creek
model, the results are much less certain than for Castle Lake.
'Although the information in this report has been funded wholly or partly by the U.S. Environmental
Protection Agency under Contract No. CR-814581-01-0 with the University of California, Davis, it does not
necessarily reflect the Agency's views, and no official endorsement should be inferred from it.
1-1
-------
Byron
CHAPTER 1
INTRODUCTION
PERSPECTIVE
The emphasis of our research on the ecological effects of climate change has been directed toward
anticipated changes in the water quality of mountain lakes and streams. The long-term limnological research
of the Lake Tahoe and Castle Lake research groups of University of California, Davis has provided a unique
opportunity to examine an extensive limnological data base that spans several decades. In our previous research,
we have established strong, quantifiable relationships between climatic variables and water quality, and it is
logical that further research can be applied to the EPA effort in modeling the ecological effects of changing
climate.
Our work interprets the long-term limnological data bases of a northern California mountain lake and
stream. Our first data set, from subalpine Castle Lake, spans the longest period of time and has yielded the
strongest relationships. In this analysis we have taken advantage of 29 years of limnological and precipitation
data to establish relationships between annual phytoplankton productivity and climatic variables. Extensive
limnological research over the years has been focused on understanding the mechanisms concerned with
nutrient dynamics and system productivity (e.g., Goldman, in press; Goldman et al., submitted; Goldman and
de Amezaga, 1984; Axler et al., 1981). The lake is clearly sensitive to climatic change, with most of the
interannual variability in heat budget (Strub et al., 1985) and primary productivity accountable to variation in
climate (Goldman et al., submitted). One of the key variables responding to climate change is the extent and
timing of ice cover. Long-term records have established clear responses between ice cover and climate
(Semtner, 1984; Rannie, 1983; Kuusisto, 1987). Long-term records have been used empirically to project lake
ice changes, but not subsequent water quality effects in changing climate scenarios (Kuusisto, 1987).
For our analysis of mountain stream water quality, we have integrated our 12 years of data on Ward Creek,
a tributary of Lake Tahoe, to yield precipitation-streamflow and streamflow-nutrient concentration relationships
useful in predicting the water quality changes associated with changing climate. Ward Creek has also shown
large water quality fluctuations in response to drought or high-precipitation years (Leonard et al., 1979; Byron
and Goldman, 1986).
The study of the causes of variability in the water quality of lakes and streams is of both theoretical and
practical interest. Most previous research has focused on the causes of long-term changes, but interannual
variability is receiving increasing attention as well. For example, long-term records have revealed water quality
changes in lakes due to eutrophication (Edmondson and Lehman, 1981; Goldman, 1986, 1981, 1974) and the
effects of climatic fluctuations (Wetzel, 1981; Strub et al., 1985; Goldman et al., submitted). In addition,
streams may experience long-term changes in water quality due to anthropogenic disturbances of watersheds
(Leonard et al., 1979) and changing patterns of atmospheric deposition (Smith et al., 1987).
The realization that unprecedented global climate changes brought about by projected CO2 and trace gas
increases will occur, has spawned a new body of research concerned with potential secondary effects (e.g.
Ramanathan, 1988). Hydrologic research has focused on projections of changes in water supply and subsequent
societal impacts (Cohen, 1986,1987; Kuusisto, 1987). Studies in California have projected significant hydrologic
changes under doubled atmospheric CO2 scenarios (Gleick, 1987). In this study, we propose to establish
empirical models of water quality and climatic variables from our long-term data sets. These models will allow
us to add water quality projections to the ongoing hydrologic and water supply modeling of other researchers.
1-2
-------
Byron
SITE DESCRIPTIONS
Castle Lake is typical of subalpine lakes in the Western United States. It is located at an elevation of 1657
m in a protected cirque basin in the Klamath mountains of northern California The lake has a surface area
of 0.20 km2, a maximum depth of 35 m, and a mean depth of 11.4 m. It is meso-oligotrophic, dimictic, and
ice-covered in winter. Ward Valley, Lake Tahoe, is also a rather typical subalpine watershed of the central
Sierra Nevada Ward Creek runs shallow, cold, and clear for the majority of the year. The relief of the
watershed ranges from 2700 m on the surrounding peaks to 1898 m at the level of its drainage into Lake Tahoe.
The total area is 2510 ha, contributing approximately 6% of the annual surface runoff to the lake. The Ward
Valley watershed has been glaciated, and soil parent material is primarily andesitic and basaltic volcanic rock
(Leonard et al., 1979).
The remaining portion of our report includes two major sections. First, we describe the development of the
empirical water quality models for Castle Lake and Ward Creek, Lake Tahoe, from our two long-term data sets.
Second, we apply our empirical models to the climatic variable outputs provided by EPA to examine the effects
of climate change scenarios on the water qualitv of the lake and stream. In each case, the results are discussed
with regard to model limitations and to the environmental and socioeconomic implications of the predicted
water quality changes.
1-3
-------
Byron
CHAPTER 2
METHODS
THE EMPIRICAL EFFECTS MODELS
Castle Lake
The 29-year data record at Castle Lake has given us an excellent opportunity to understand the importance
of long-term variability in the water quality of a subalpine lake. Climatic events appear to dominate limnological
variation in Castle Lake, at least as indicated by the primary productivity of the phytoplankton community.
Primary production makes an excellent water quality indicator since it simultaneously reflects local weather
conditions, nutrient availability, and the size of the producer populations. Further, it indirectly reflects the
activities of other trophic levels, such as graying and nutrient regeneration. In addition, it is of direct importance
to our investigation of oligotroptiic, subalpine lakes, since primary productivity is causally related to water clarity,
an important measure of water quality in these systems (Goldman, in press). Although results based on primary
productivity are not immediately transferable to other lakes, in part because of the rarity of long-term
productivity records, the variable gives us an immediate measure of eutrophication, perhaps the most important
overall water quality indicator. Our empirical water quality model for Castle Lake gives us the ability to forecast
primary production from a knowledge of climatic variation and previous production.
Our long-term record at Castle Lake results from samples collected over the entire water column at a single
deep station every 5 days during the summer months and irregularly during other months since 1959. Our
model uses summer primary productivity estimates for the period of June 1 to September 30 of each year
(measured to be 75% of total annual production). Temperature and precipitation estimates were provided by
the nearby Mt. Shasta City weather station (NWS #5983).
Model development
Annual primary production at Castle Lake is clearly related to climate. In particular, Castle Lake has
exhibited anomalous (unusually high or low) primary production during ENSO (El Nino/Southern Oscillation)
years. In order to investigate the dependence of annual primary production on specific climate variables, the
1960-1984 series of primary productivity measurements was first prewhitened with a Markov process to remove
an autocorrelation at lag 1 year. The prewhitened series was created as follows:
WPt = (Pt-a0)-a1*(Pt.1-a0) (1)
where WPt is the series of prewhitened production at year t, Pt (mg C m"2d ) is the series of actual primary
production at year t, and the a, are constants; ag = 355 ± 29 and a1 = 0.385 ± 0.189.
A significant cross-correlation at lag zero was detected between prewhitened primary production and both
winter snowfall and annual hydraulic load. Although snowfall and hydraulic load are associated, the relation
between prewhitened primary production and each of these variables is still significant, even when adjusted for
the other variable. Snowfall is important because of its effect on ice breakup time; hydraulic load because of its
effect on outwash rates in the spring. The autocorrelation in primary production may arise from persistence of
certain phytoplankton populations from one year to the next (Vincent and Goldman, 1980).
Based on these associations with the prewhitened series, a time series regression model was built relating
measured annual primary production to snowfall and precipitation. Primary production was transformed with
logarithms to stabilize variance; the autocorrelation was addressed by including lag 1 primary production; climate
1-4
-------
Byron
effects were represented by the product of January-April snowfall and January-May total precipitation; and a
moving-average model for the residuals was included. All coefficients in the model were significantly different
from zero at the p <0.05 level. The initial model accounted for 81% of the variability in the logarithm of
primary production:
(1- S *B) In Pt = WQ + Wl * St * Ht + i\ (2)
where B is the backshift operator (i.e. B x. = x.,), Pt (mg C m d~1) is primary production during year t, S» is
January-April snowfall (actual snow depth in cm, not water equivalent) during year t, Hf (cm) is January-May
hydraulic loading during year t, ru is the residual series, and 5, WQ, and w1 are constants. P. is measured summer
productivity (June-September). In 1968, when measurements were made year-round, these summer months
accounted for 75% of the annual production.
Since some autocorrelation was found in the residual series at lag 2, we selected a model for the residuals
as:
r^ = aj + 0*^.2 (3)
where a^ is a series of independent white noise N(0, aa) and 6 is constant. The overall initial model using inputs
of measured productivity, snowfall, and total precipitation became:
(1- 5*B) In Pt = WQ + w1 * St * Ht + (1 + 0*B2)at (4)
Snowfall Estimates
Output from the global climate models does not include snowfall, which is required for the final Castle
Lake production model (Eq. 4). We therefore attempted to develop a forecasting model for monthly snowfall,
based on monthly values for the climate model outputs of total precipitation and temperature. Using the actual
monthly data for the years 1960-1984, the same years for which Eq. 1 was developed, a variety of approaches
was tried, including time-series regression models for individual months, and regression models developed using
data for all winter months. A second-degree polynomial model applicable for all winter months was chosen;
it performed almost as well as the more complicated time-series regression models developed for individual
months:
S = b,* H + b2 * H2 + b3 * H * T (5)
where S is monthly snowfall (cm snow depth), H is monthly total precipitation (cm), and T is mean monthly
temperature (°K). The constants are b1 = 191 ± 27, b2 = -0.100 ± 0.015, and b^ = -0.672 ± 0.097. The variance
accounted for (r) is 56%, and the residuals exhibit no significant autocorrelation or other internal structure.
Model Using Estimated Snowfall
An attempt was made to develop a new forecasting model for annual production that excluded snowfall and
used temperature (monthly maximum, minimum, and average). However, no improvement could be made over
the form of Eq. 4, even when St was estimated by summing values determined from Eq. 5 for January through
1-5
-------
Byron
April. As a result, the parameters in Eq. 4 were estimated again using these predicted values of S. The
resulting model explained much less of the variance (r2 = 58%), but the parameters remained significantly
nonzero (p <0.01) and a moving average term was not needed to remove residual structure (i.e., 9=0). The
new parameter values were S = 0.571a ± 0.142, WQ = 2.85 ± 0.81, and w1 = -(3.78 ± 0.79)*10'5 The structure
of the model is given in Table 1.
Ward Creek. Lake Tahoe
Our data records for precipitation and streamflow for Ward Valley encompass 12 years of daily streamflow
and nutrient load estimates and storm-event based precipitation sampling. In developing the water quality
models, we limited the record to our longest unbroken string of data: 1980-1986. The period of modeling is
further limited by the weather records for this study, which end in 1984. Approximately 100-150 stream samples
were taken each year to generate gauge height-concentration curves that allowed us to use continuously recorded
gauge height to estimate average daily concentration values. Precipitation data were provided
by the VS. Weather Bureau, Tahoe City (NWS #8758). While this is not a particularly long data set compared
to our lake data, it has nevertheless yielded significant regressions relating water quality to climatic variables.
We had to examine several hydrologic parameters in order to establish the water quality regressions. Our
basic two-part approach was First to establish a relationship between precipitation and streamflow and then to
describe the relationship between flow and nutrient concentrations. Our data record is long enough to include
a drought year (1981) as well as very high water years (1982, 1983). Thus we feel that the significant
relationships we have established encompass most of the present range of climatic variation.
The precipitation-flow model for Ward Creek used precipitation at lag 3 and 6 months and a moving
average model for the residuals (1 month) to predict monthly total streamflow (Figure 1, Table 2). The model
was well behaved, i.e., all coefficients were significant and the residuals had all structure removed. However,
only 55% of the variation was accounted for, and it was not possible to incorporate temperature into the model
in a statistically acceptable manner. Neither present month nor lagged temperature terms added to the
explanation of variance. The 3- and 6-month lag terms for precipitation were necessary to incorporate snowmelt
accumulation and melt effects into the model (and, thus, incorporate some potential temperature effects).
The second step of our model estimates various aspects of annual average stream nutrient concentrations
as predicted from stream runoff. In this case we were not limited by the weather data, and our period of
analysis was from January 1980 through September 1986 (81 observations). Stream discharge was significantly
associated with both soluble phosphorus and nitrate concentrations for this period (Figures 2, 3; Table 3). As
might be expected, stream discharge was also significantly associated with loading of both soluble phosphorus
and nitrate (Figures 4,5; Table 4). The increase in predictive capability with loadings is due to the fact that load
is the product of stream discharge and nutrient concentration.
Limitations of the Empirical Models
In the case of Castle Lake, the model is limited to the prediction of annual total'values of algal production.
Seasonal effects are lost. Still, the model was created empirically, using enough years of data to validate the
relationships for a wide range of climatic variation. It offers a good unbiased estimate of direction and overall
magnitude of changes when using scenario data For Ward Creek, the seasonal effects are preserved by the use
of monthly data However, the relative lack of change in cumulative precipitation in the climate change
scenarios and inability to incorporate temperature in the models precluded the use of the monthly data in any
attempts to model water quality changes.
The models make use of effects of precipitation and temperature change but do not incorporate the direct
effects of a doubling of CO2 in the atmosphere into their predictions. Atmospheric CO2 increase is only
potentially important in the case of the Castle Lake primary productivity model, but even in that case the effects
of carbon limitation on algal growth are unlikely to be significant. Subalpine lakes like Castle are relatively rich
1-6
-------
Byron
Table 1. Castle Lake Primary Production Model
PPR = "Annual" primary production (June September)
LPPR = Ln(PPR)
TMPOit = mean temperature (°K) for year t and month i (i = 1,...,4)
PRCOit = total precipitation (cm) for year t and month i (i = 1.....5)
SNWOit = MAX(191*PRCOit-0.0996*PRCOit2-0.672*PRCOit*TMPOit, 0)
PRCt = PRC01t+PRC02t+PRC03t+PRC04t+PRC05t
SNWt = SNW01t+SNW02t+SNW03t+SNW04t
LPPRt = MAX(2.85+0.571*LPPRt_r3.78*10-5*SNWt*PRCt, 0)
PPRt = EXP(LPPRt)
1-7
-------
Byron
P.
so
82
84
Measured
Date
Model
Figure 1. Prediction of water discharge from precipitation, Ward Creek, Lake Tahoe.
1-8
-------
Byron
Table 2. Ward Creek Streamflow Prediction
Dependent Variable is Water Discharge
1980.07 1984.12
Number of observations: 54
Convergence achieved after 5 iterations
VARIABLE
Precipitation (-3)
Precipitation (-6)
Residuals:
Moving Average (1)
R-squared
S.E. of regression
Durbin-Watson stat
F-statistic
COEFFICIENT STD. ERROR T-STAT. 2-TAIL SIG.
403 85.1 4.73 0.000
376 75.3 5.00 0.000
0.725 0.140 5.17 0.000
0.545
2600
1.98
30.6
1-9
-------
Byron
26
a
a
o
a.
(Monthly Averages)
0
1 !
82
i
84
i i
86
Date
a Measured
+ Model
Rgure 2. Prediction of soluble phosphorus concentrations from water discharge, Ward Creek, Lake Tahoe.
1-10
-------
Byron
Table 3a. Ward Creek Soluble Phosphorus Concentration Prediction
Dependent Variable is Soluble Phosphorus Concentration
1980.02 - 1986.09
Number of observations: 80
Convergence achieved after 3 iterations
VARIABLE
Constant
Water Discharge (-1)
Residuals:
Moving Average (1)
R- squared
S.E. of regression
Durbin-Vlatson stat
F-statistic
COEFFICIENT STD. ERROR T-STAT. 2-TAIL SIG.
10.1 0.350 28.8 0.000
-0.000436 7.81D-05 -5.59 0.000
0.503 0.114 4.41 0.000
0.422
2.39
1.95
28.1
1-12
-------
Byron
3
84
86
M*a»ur«d
Dot*
Model
Figure 3. Prediction of nitrate concentrations from water discharge, Ward Creek, Lake Tahoe.
1-11
-------
Byron
Table 3b. Ward Creek Nitrate Concentration Prediction
Dependent Variable is Nitrate Concentration
1980.04 - 1986.09
Number of observations: 78
Convergence achieved after 5 iterations
VARIABLE
Constant
Water Discharge
Discharge (-1)
Discharge (-3)
Residuals:
Moving Average (1)
Moving Average (2)
R-squared
S.E. of regression
Durbin-Watson stat
F-stat1stic
COEFFICIENT STD. ERROR T-STAT. 2-TAIL SIG.
11.2 1.07 10.5 0.000
0.000782 0.000236 3.31 0.002
-0.000939 0.000235 -4.00 0.000
-0.000461 0.000193 -2.39 0.020
0.462 0.119 3.90 0.000
0.298 0.118 2.52 0.014
0.410
5.73
2.00
10.0
1-13
-------
Byron
u.
i
80
D Measured
Figure 4. Prediction of soluble phosphorus loading from water discharge, Ward Creek, Lake Tahoe.
1-14
-------
Byron
280
o
z
n
o
Q M««ur»d
Figure 5. Prediction of nitrate loading from water discharge, Ward Creek, Lake Tahoe.
1-15
-------
Byron
Table 4a. Ward Creek Soluble Phosphorus Loading Prediction
Dependent Variable is Soluble Phosphorus Loading
1980.02 - 1986.09
Number of observations: 80
Convergence achieved after 3 iterations
VARIABLE
Water Discharge
Discharge (-1)
Residuals:
Moving Average (1)
R-squared
S.E. of regression
Durbin-Uatson stat
F-statistic
COEFFICIENT STD. ERROR T-STAT. 2-TAIL SIG.
0.00733 0.000244 30.0 0.000
-0.000717 0.000239 -2.99 0.004
0.455 0.117 3.88 0.000
0.933
6.23
1.88
539
1-16
-------
Byron
Table 4b. Ward Creek Nitrate Loading Prediction
Dependent Variable is Nitrate Loading,
1980.02 - 1986.09
Number of observations: 80
Convergence achieved after 3 iterations
VARIABLE
Water Discharge
Discharge (-1)
Residuals:
Moving Average (1)
R-squared
S.E. of regression
Durbin-Watson stat
F-statistic
COEFFICIENT STD. ERROR T-STAT. 2-TAIL SIG.
0.014 0.000918 15.3 0.000
-0.00426 0.000898 -4.75 0.000
0.264 0.118 2.25 0.028
0.758
23.3
1.96
121
1-17
-------
Byron
in free CO2 and low in bicarbonate and carbonates, and experience algal production limited by nutrient
availability, not carbon (Wetzel, 1983). Increased carbon in the system could be dealt with in a more complex
model. The effects of increased carbonic acid in precipitation are not incorporated into our models either.
However, precipitation acidity is presently dominated by the effects of strong acids. Presumably, future increases
in nitric and sulfuric acid (not estimated as part of the scenarios) would be more important in causing future
ecological changes.
The Castle Lake algal productivity predictions that we have produced using the scenario results are only
realistic when productivity estimates fall within or close to the historic range. As production estimates rise, the
limiting factors on total production at some point will undoubtedly shift from factors associated with hydraulic
washout and the length of the growing season (as the present model was developed) to nutrient input, speed
of recycling, or other limitations. Nutrient inputs, for example, may change dramatically with forest vegetation
and runoff changes under a doubled CO2 regime. The present model cannot account for the possibility of
changing limitations to growth associated with changing climate. Microbial processes in the soil and sediment
and watershed plant communities may be expected to respond to climatic change and deserve consideration in
terms of their effect on nutrient flux. A series of predictive ecological models should be incorporated into an
integrative ecosystem level model in order to most accurately predict water quality changes.
THE CLIMATE CHANGE SCENARIOS
The Scenarios Used
We used the GFDL, GISS, and OSU control and 2xCO~ and the GISS transient A scenarios to apply to
our empirical water quality models. Surface temperature ana total precipitation were used from each. Although
this does not represent the complete list of possible scenarios, it gives a strong consensus in direction and
magnitude suggested by most of the climate change models. For Castle Lake models we used weather station
data from Mt. Shasta City, California For Ward Creek, we used Tahoe City, California, data
The projected temperature and precipitation changes resulting from the scenarios are shown in Figures 6
through 10. In all cases, ratios of change from present conditions for each month were multiplied by historical
weather data to yield scenario climatic conditions. Note that temperature in the 2xCO, scenarios is higher than
historical values (Figure 6), whereas precipitation changes are much less discernible (Figures 7, 8). The lack of
change hi precipitation can also be easily seen for the GISS transient A scenario (Figure 9). For our empirical
model of Castle Lake, it is of greater significance that snowfall estimates show a strong negative trend for the
transient scenario (Figure 10).
It was necessary to change the application of the scenarios for Castle Lake. As described above, our
empirical model requires that productivity estimates begin in a year for which the previous summer's productivity
was known. In our case, the model output was restricted to beginning in 1960, the second year of Castle Lake
data collection. Thus, the measured Castle Lake productivity for 1957 was used as input to both the control and
2xCO2 conditions. Since the years 1951-1959 were removed from the series, we included the four additional
years Tor which we had data (1981-1984) at the end of the data series. Thus, the series for Castle Lake runs
from 1960 through 1984 for control and 2xCO2 data The GISS transient A starts by using the measured 1980
Castle Lake productivity to begin the model run in 1970 for the series through 2060.
The Ward Creek models were run using the series 1980-1984 to establish precipitation and temperature
conditions because our empirical models were developed for that period. The scenarios were not used to predict
water quality changes for the 1951-1980 period, since the only scenario to show changing precipitation did not
cause a consequent change in stream water discharge (see results for GISS 2xCO2 below), and temperature was
not incorporated into the model. There is, therefore, no basis for examining predictions of Ward Creek water
quality based on the scenario outputs. The modeling of average concentrations of stream solutes as predicted
from precipitation would require more detailed information on rain, snow, and snowmelt than was available as
climate model output. Instead, we present our empirical models of nutrient concentrations and load over time
1-18
-------
Byron
180
I
170 -
160 -
150 -
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
50 -
40 -
30 -
20 -
10 -
Mt. Shasta City, CA.
2XCO2 Scenario*. Coetle Lake
Total Praelp. (Jon-May)
58
i i
62
Measured
66
70
Year
OFDL
i
74
i
78
82
86
GISS
OSU
Figure 7. Measured and 2xCO2 scenario, January-May precipitation, Castle Lake.
1-20
-------
12
I
I
11 -
10 -
9 -
8 -
7 -
6 -
5 -
4 -
3 -
2 -
1 -
Byron
Mt. Shasta City, CA.
2XC02 Scenarios. Cart* laks
Avg. T«mp. (Jan-Apr)
98
62
66
70
74
78
82
86
Year
Figure 6. Measured and 2xCO2 scenario, January-April average temperatures, Castle Lake.
GISS gridpoint: 120W, 39.13N;
GFDL gridpoint 120W, 40N;
OSU gridpoint: 120W, 40N.
1-19
-------
Byron
80
Measured
Date
GFDL
GISS
OSU
Figure 8. Measured and 2xCO2 scenario, monthly precipitation, Ward Creek, Lake Tahoe.
GFDL gridpoint: 120W, 35.55N;
GISS gridpoint: 120W, 39.13N;
OSU gridpoint: 120W, 36N.
1-21
-------
Byron
180 - —
150 -
140 -
130 -
120 -
110 -
100 -
90 -
80 -
70 -
60 -
SO -
40-
30 -
20 -
10 -
0 -j-
70
Prtdp. (Jon-May)
90
10
30
90
70
Year
Figure 9. GISS transient A, monthly precipitation, Castle Lake.
1-22
-------
Byron
350
300 -
230 -
200 H
I 150-
100 -
50 -
Snow Depth (Jen-Apr)
70
r
90
10
30
50
70
Year
Figure 10. Monthly snow estimates calculated using precipitation and temperature results from GISS
transient A, Castle Lake.
1-23
-------
Byron
for the years 1980-1984 as predicted from the relationships between monthly stream water discharge and nutrient
concentrations.
Issues Resulting From the Scenarios
Our empirical models require a knowledge of snow accumulation and melt to adequately predict water
quality parameters. The present models would be greatly aided by refinements in the scenario outputs that
would allow more accurate predictions of precipitation, temperature, and snowfall on a fine-grained scale. The
coarse nature of the large global climate model grids yields ratios of change in parameters that are of dubious
value in predicting change in mountainous subregions of the larger grids.
The years 1960-1984 are adequate for establishing a full range of baseline conditions. In terms of hydrology
and water quality parameters, they encompass an extensive range of drought, "normal," and high-water years.
Although these represent a deviation from the basic 1951-1980 data set, they are still representative of the same
hydrologic range.
1-24
-------
Byron
CHAPTERS
RESULTS
CASTLE LAKE
Annual primary production was forecast for five cases using the equations shown in Table 1: (1) actual
conditions (Table 5, Figure 11); (2) climate output from the the three 2xCO2 models (Table 6, Figure 12); and
(3) climate output from the GISS transient scenario A (Figure 13). Table 1 contains the actual equations used
in the simulation. Note that for all simulations, snowfall was set to 0 when Eq. 2 predicted negative snowfall.
The results for 2xCO2 are summarized in Figure 14.
The simulation reproduced the actual primary production values quite accurately (Figure 11), although the
model had difficulty recreating the large swings experienced during the period 1969-1977, during which time
four major ENSO events occurred (1969,1972,1976,1977) (Strub et al., 1985). It should be noted that Figure
11 represents a dynamic simulation that propagates errors in estimated primary production through the lag
1-year term.
The three 2xCO2 simulations behaved in a similar fashion, although quite variable in magnitude (Figure
12). The increase in temperature caused by the 2xCO, conditions resulted in an increase in mean annual
primary production of 69.0% (GFDL), 87.0% (GISS), and 163% (OSU) over estimated "control" conditions for
the 2xCO2 scenarios. These changes are close to our measured historical range, but show a consistent direction
of change toward higher production under conditions of higher atmospheric CO, concentrations. It is possible
to substitute an average 2xCO2 value for our measured 19S9 value to start the 2xCO2 runs. The mean 2xCO2
primary production values increase slightly under those conditions, from 03% (OSU) to 3% (GISS) greater than
the averages shown in Table 6.
The GISS transient A scenario produced an increasing but variable primary productivity estimate for the
period 1970-2059 for Castle Lake, showing the expected strong 30-year scale of variation inherent in the
1951-1980 data base (Figure 13). In all scenarios, therefore, we project increasing productivity. Whereas the
increases may be small at Castle Lake, the variability between the lakes is unknown. The direction of change
is clear. The implications of eutrophication are discussed below.
WARD CREEK, LAKE TAHOE
Since our empirical models for stream water quality did not use temperature, we were limited in our ability
to apply our results to the 2xCO- and transient scenarios. Of the climate conditions we examined for the Tahoe
City weather station, the GISS £*CO>> precipitation estimates stood apart as representing consistently wetter
conditions. We tested whether the GISS-altered precipitation regime would affect stream discharge by using
the empirical regression equation for precipitation-discharge described above.
The results show isolated periods of deviation of GISS discharge from measured streamflow but no
consistent trends in magnitude or direction (Figure 15). Apparently, the slightly elevated precipitation regime
of the GISS scenario is not enough to significantly alter runoff conditions in our models, perhaps reflecting the
relatively small watersheds of mountain lakes. If we could incorporate temperature as weft as precipitation into
our precipitation-runoff model, there would be a greater likelihood of demonstrating some effect of the scenario
conditions. With our present models we have no further basis for examining water quality parameters under
the scenario conditions.
1-25
-------
Byron
Table 5. Modeled vs. Measured Primary Production, Castle Lake
1960 - 1984
Number of observations: 25
SERIES
MEAN
S.D. MAXIMUM MINIMUM
Measured Production
Modeled Production
349
340
99.6
68.9
569
468
136
216
PPR,PPR1XO
Covariance Correlation
427
0.648
1-26
-------
Byron
I
r
600
500 H
400 H
300 H
200 H
100
htooaured
58 62
i
66
i i
70 74 78 82 86
Year
Figure 11. Measured and modeled annual primary production, Castle Lake, 1960-1984.
1-27
-------
Byron
Table 6. 2xCO2 Scenario Results, Castle Lake: Production Estimates
1960 - 1984
Number of observations: 25
SERIES MEAN S.D. MAXIMUM MINIMUM
GFDL 576 65.5 677 422
GISS 636 59.0 720 446
OSU 396 72.7 493 250
1-28
-------
Byron
I
I
800
700 -
600 -
SOO -
400 -
300 -
200
Model
58
I
62
I
66
i i i^ i i
70 74 78
Year
I I I
82
86
Figure 12. Control and 2xCO, scenario, annual primary production estimates, Castle Lake.
1-29
-------
Byron
x
04
O
I
800
500 -
400
300 -
200 -
100 -
Castle Lake
Annual Production EitimotM
1981-89 1990-99 2000-09 2010-19 2020-29 2030-39 2040-49 2050-99
QBS Traratwit A
Mean UTft Standard Dwfatfon
Figure 13. GISS transient A, annual primary production estimates, Castle Lake, plus model "control" values.
Transient values are decadal means. In all cases, one standard deviation is shown as added to the
mean.
1-30
-------
Byron
TOO -i
N
O
600 -
900 -
400 -
300 -
200 -
100 -
Castle Lake
Annual Production EitfmatM
M«Mured Model GFDL
[771 Mean U77\ Standard delation
OJSS
osu
Figure 14. Summary of production estimates: all scenarios plus measured and model "control" values. One
Standard deviation is shown as added to the mean.
1-31
-------
Byron
so
84
Control
Dot*
GISS
Figure 15. Modeled "control" and GISS 2xCO2 water discharge estimates, Ward Creek, Lake Tahoe.
1-32
-------
Byron
CHAPTER 4
INTERPRETATION
The changes in Castle Lake associated with the climatic change scenarios were driven primarily by
temperature. The increased temperature under the 2xCO, scenario conditions contributed to relatively less
snow accumulation and consequently an early melting of the lake ice. The trend in snowfall can clearly be seen
in the transient results (Figure 10). Early melting and a longer growing season resulted in higher annual primary
production estimates. The relative lack of change of precipitation as compared to temperature in the scenarios
lessened the impact of total precipitation on the production estimates.
The 2xCO2 conditions produced clearly altered productivity estimates (Figure 12) and the transient
provided a definite trend (Figure 13) for Castle Lake. Short-term changes are less clear in the transient,
however, owing to the masking effect of strong interannual variation. As can be seen from the summary figure,
the first two 30-year periods of the transient do not produce mean production estimates more than one standard
deviation from our measured historical values (Figure 14). Short-term trends occurring at the rate of the GISS
transient A would be difficult to differentiate from interannual variability at a scale of 50 years or less.
In the case of Ward Creek, Lake Tahoe, water quality changes associated with climate change projections
were undetectable in our analysis.
1-33
-------
Byron
CHAPTERS
IMPLICATIONS OF RESULTS
ENVIRONMENTAL
The implications of the projected direction of change in water quality center on the effects of enhanced
primary productivity on oligotrophic mountain lakes. Our results suggest that algal production in Castle Lake
would be stimulated through increasing hydraulic inputs and a longer growing season. The work on Ward
Creek, although inconclusive, indicates that enhanced precipitation (as demonstrated in the GISS scenario)
combined with a changing rain-snow balance (unknown in our model) could easily increase the in-stream
concentrations of particulates and soluble nutrients. There is thus the possibility of increased watershed nutrient
contributions as well as the direct enhancement of production through an increase in length of the growing
season. It seems certain that the projected global changes associated with the increasing atmospheric CO,
conditions would result in enhanced primary production for Castle Lake and possibly for all similar high
mountain lakes.
Enhanced primary production (eutrophication) would cascade to all elements of the food web and possibly
lake and downstream water quality through changing nutrient dynamics, storage, and release. As eutrophication
proceeds, the dynamics of algal and bacterial blooms begin to dominate the changes in water quality.
Zooplankton, macroinvertebrates, and fish all increase in biomass but are subject to large seasonal changes.
Fish in the Great Lakes have been projected to change dramatically in response to climate change (Meisner
et al., 1987). During stratification, decaying biota deplete oxygen in deep waters and increase the solubility of
nutrients and potentially toxic compounds (e.g., metals, HjS) which could be added to the surface water and
downstream releases during mixing events. Although increased system productivity might affect such aspects
of community dynamics as species diversity or the size, age structure, and dynamics of populations, the
relationships are complex and require additional empirical and mechanistic models before reasonable predictions
can be made.
SOCIOECONOMIC
The socioeconomic implications of mountain lake eutrophication and altered water quality are in terms
of recreational uses, fisheries, and the quality of water supplies. All are adversely affected by increasing algal
productivity with the exception of fisheries, which, at least in these unproductive lakes, might be enhanced by
a greater system productivity. The enhancement of system productivity could result in major shifts in fish
community structure as well. Lakes which are so warm and productive as to be marginal for trout production
might exceed the threshold of oxygen and temperature requirements and revert to warm water fisheries
conditions, as has been projected for the Great Lakes area (Meisner et al., 1987). As eutrophication proceeds,
deep lakes will show oxygen depletion in the hypolimnion and lose deep fish habitat during the summer
stagnation and winter ice cover. The eutrophication of mountain lakes and streams could affect local and
downstream water supplies and recreation in a multiplicative fashion. Oxygen depletion, nutrient enrichment,
and algal blooms may contribute to locally decreased water quality for domestic or industrial use. Subsequently,
the eutrophication of upstream areas will have a negative impact on downstream receiving lakes and streams
independent of climate-induced changes. The negative impact on water supplies will also be immediately
evident in recreational use. Although limited eutrophication could enhance fisheries, the deterioration of water
quality will negatively affect other recreational uses.
1-34
-------
Byron
CHAPTER 6
POLICY IMPLICATIONS
Lake water quality and eutrophication are important considerations in the regulation and planning of
recreation areas. Cold water trout Usheries are important for recreation and for the economics of western
mountainous areas of the US. and Canada A general warming of streams and lakes could eliminate or reduce
spawning success of resident and anadromous salmonids at lower elevations and allow for invasion of warm
water-tolerant species. Water quality degradation could necessitate policy changes regarding in-streamflows
and the extent and type of water use.
1-35
-------
Byron
CHAPTER?
FUTURE RESEARCH
Several areas of applied research are suggested by the results of these initial studies. We have touched
upon some of the potentially valuable resources of the long-term Tahoe and Castle Lake data bases, but much
more can still be learned concerning the effects of climate change on mountain lake and stream water quality.
In the case of Castle Lake, we are now in the process of inputting and validating extensive data sets of zoo-
and phytoplankton abundance estimates and the results of stratification on water chemistry in both lakes. Both
plankton and water chemistry are indicators of water quality that are expected to show important relationships
to climate. Investigation of these with predictive models should further elucidate the effects of climate change.
An interpretation of this data should allow us to better understand the mechanisms by which climate influences
production in lakes and streams.
In the case of Ward Creek, we have established in-stream models of water quality that only require better
climate model outputs to be useful in modeling potential effects of climate change. By working with hvdrologjsts
to refine the snow accumulation and melt predictions of the climate models we should be able to provide good
projections of nutrient and sediment loads under different climatic scenarios.
1-36
-------
Byron
REFERENCES
Axler, R.P., G.W. Redfield, and C.R. Goldman. The importance of regenerated nitrogen to phytoplankton
productivity in a subalpine lake." Ecology. 62:345- 354, 1981.
Brimblecombe, P. and D.H. Stedman. "Historical'evidence for a dramatic increase in the nitrate component of
acid rain." Nature. 298:460-462, 1982.
Byron, E.R. and C.R. Goldman. "A Technical Summary of Changing Water Quality in Lake Tahoe: The First
Five Years of the Lake Tahoe Interagency Monitoring Program." Tahoe Research Group, Inst. of Ecology,
Univ. Calif., Davis, 1986. 62 pp.
Cohen, S J. 1986. Impacts of CCL-induced climatic change on water resources in the Great Lakes basin.
Climatic Change. 8:135-153.
Cohen, S J. 1987. Projected increases in municipal water use in the Great Lakes due to COj-induced climatic
change. Water Resources Bulletin. 23(1):
91-101.
Edmondson, W.T. and J.T. Lehman. "The effect of changes in the nutrient income on the condition of Lake
Washington." Limnol. Oceanogr.. 27:272-293, 1981.
Gleick, P.H. 1987. Regional hydrologic consequences of increases in atmospheric CO- and other trace gases.
Climatic Change. 10:137-161.
Goldman, C.R. "Eutrophication of Lake Tahoe emphasizing water quality." EPA- 660/3-74-034. U.S. Gov.
Printing Office, 1974. 408 pp.
Goldman, C.R. "Lake Tahoe: two decades of change in a nitrogen deficient oligotrophic lake." Verb. Int.
Verein. Limnol.. 21:45-70, 1981.
Goldman, C.R. "Lake Tahoe: an oligotrophic lake's response to nutrient loading." Lake Pollution and Recovery.
Proceedings of an International Congress. European Water Poll. Control. Assoc., 1986. pp. 329-334.
Goldman, C.R. "Primary productivity and transparency: the early onset of eutrophication in ultra oligotrophic
Lake Tahoe, California-Nevada" Limnol. Oceanogr.. In Press.
Goldman, C.R. and de Amezaga, E. "Primary productivity and precipitation at Castle Lake and Lake Tahoe
during twenty-four years, 1959-1981" Verb. Int. Verein. Limnol.. 22:591-599, 1984.
Goldman, C.R., A. Jassby, and T. Powell. "Interannual fluctuations in primary production: impact of climate and
weather at two subalpine lakes."
Submitted.
Kuusisto, E. 1987. An analysis of the longest ice observation series made on Finnish lakes. Aqua Fennica.
17(2):123-132.
Leonard, R.L., LA. Kaplan, J.F. Elder, R.N. Coats, and C.R. Goldman. "Nutrient transport in surface runoff
from a subalpine watershed, Lake Tahoe basin, California." Ecol. Monogr.. 49(3):281-310,1979.
Meisner, J.D., J.L. Goodier, HA. Regier, BJ. Shuter and WJ. Christie. 1987. An assessment of the effects
of climate warming on Great Lakes basin fishes. J. Great Lakes Res.. 13(3):340-352.
1-37
-------
Byron
Ramanathan, V. 1988. The greenhouse theory of climate change: a test by an inadvertent global experiment.
Science. 240:293-299.
Rannie, W.F. 1983. Breakup and freezeup of the Red River at Winnipeg, Manitoba, Canada in the 19th
century and some climatic implications, riimatir ffiang»T 5:283-296.
Semtner, A J., Jr. 1984. On modeling the seasonal thermodynamic cycle of sea ice in studies of climatic
change. Climatic Change. 6:27-37.
Smith, RA., R.B. Alexander, and M.G. Wolman. "Water quality trends in the nation's rivers." Science.
235:1607-1615, 1987.
Soderlund, R. "NO pollutants and ammonia emissions - a mass balance for the atmosphere over Northwest
Europe." Ambio. 6(2-3):118-122,1977.
Strub, P.T., T. Powell, and C.R. Goldman. "Climatic forcing: effects of El Nino on a small, temperate lake."
Science. 227:55-57, 1985.
Vincent, W.F. and C.R. Goldman. "Evidence for algal heterotrophy in Lake Tahoe, California-Nevada."
Limnol. Oceanogr.. 25(l):89-99,1980.
Wetzel, R.G. Limnology. 2nd Edition. W.B. Saunders Co., Philadelphia, Pennsylvania, 1983. 767 pp.
Wetzel, R.G. "Long-term dissolved and participate alkaline phosphatase activity in a hardwater lake in relation
to lake stability and phosphorus enrichments." Verb. Int. Verein. Limnol.. 21:369-381,1981.
1-38
-------
POTENTIAL RESPONSES OF GREAT LAKES FISHES AND THEIR HABITAT TO
GLOBAL CLIMATE WARMING
by
John J. Magnuson
David K. Hill
Center for Limnology
University of Wisconsin Madison
Madison, WI 53706
and
Henry A. Regier
John A. Holmes
J. Donald Meisner
Department of Zoology
University of Toronto
Toronto, Ontario
Canada M5S 1A1
and
Brian J. Shuter
Ontario Ministry of Natural Resources
Maple, Ontario
Canada LOJ 1EO
Contract No. CR-814644-01-0
-------
CONTENTS
FINDINGS 2-1
CHAPTER 1: INTRODUCTION 2-2
DESCRIPTION OF ECOLOGICAL SYSTEM 2-2
ORGANIZATION OF THIS REPORT 2-4
CHAPTER 2: THERMAL NICHE 2-5
GREAT LAKES 2-5
Methods 2-5
Scenarios 2-6
Results 2-6
HEADWATER STREAMS 2-8
Methods 2-8
Scenarios 2-13
Results : 2-13
CHAPTER 3: GROWTH AND FOOD CONSUMPTION 2-15
METHODS 2-15
SCENARIOS 2-16
RESULTS 2-19
CHAPTER 4: POPULATION RESPONSE 2-24
METHODS 2-24
SCENARIOS 2-26
RESULTS 2-26
CHAPTER 5: ECOSYSTEM RESPONSE 2-28
METHODS 2-28
RESULTS 2-29
CHAPTER 6: INTERPRETATION OF RESULTS 2-32
GENERAL 2-32
SPECIFIC INTERPRETATIONS 2-33
Thermal Niche 2-33
Growth and Food Consumption 2-33
Population Response 2-34
Ecosystem Response 2-34
CHAPTER 7: IMPLICATION OF RESULTS 2-35
CHAPTER 8: POLICY IMPLICATIONS 2-36
TERMS, ABBREVIATIONS, AND SYMBOLS 2-37
REFERENCES 2-39
u
-------
Magnuson
FINDINGS1
Global climate warming will directly influence fishes in the Great Lakes by significantly wanning their
aquatic habitat. These effects act on those fishes which do best in coldwater (e.g., salmon, trout and whitefish),
in coolwater (e.g., yellow perch and walleye), and in warmwater (e.g., largemouth and smallmouth bass).
Three types of direct effects of altered thermal structure were examined for fishes: 1) changes in the size
of their optimum thermal habitat, 2) changes in growth and prey consumption, and 3) changes in reproductive
success and population size. Generally, all three thermal groups of fishes experience an increase in the size of
favorable thermal habitat integrated over the annual cycle and grow at faster rates and eat more, provided they
can seek out favorable temperatures and prey is available. These positive effects are apparent in all three
(OSU, GISS, and GFDL) global climate model scenarios and in cold regions (northern Lake Superior), cool
regions (southern Lake Michigan), and warm regions (Lake Erie) of the Great Lakes. Most of these positive
effects result from an increased length of growing season. However, in the warmest part of the summer the
size of favorable thermal habitat decreases for coldwater fishes like trout both in the lakes and in headwater
brook trout streams. During summer, species interactions like competition and predation should be intense and
may have negative effects on fishes. The increased demand for forage may also intensify species interactions
and alter food web structure and efficiency.
Smallmouth bass were used as an example of potential changes in reproductive success and population
size. This warmwater species, which lives in protected Great Lakes bays would experience marked increases
in reproductive success and population size in cold and cool regions and would not be harmed in warm regions.
Effects of climate change on the recruitment of other taxa were not evaluated.
The increase in thermal habitat for warmwater fishes may allow an increase in colonization and dispersal
of exotic fishes with either negative or beneficial effects.
To estimate whether there would be a general increase in primary and secondary production to meet the
increased forage requirements of the fishes, a simple ecosystem level model was applied to primary production
(algae), secondary production (zooplankton), and fishery production (fish yields). In all regions of the Great
Lakes primary production, zooplankton biomass, and fishery yields would be expected to increase significantly.
Thus, for most regions and most fishes, habitat and productivity would increase following climate wanning.
Even though productivity is expected to increase, surprises are expected involving (1) the loss of important
stocks and (2) intensified species interactions due to new colonizing species and habitat constriction among the
existing fish communities. The higher productivity and the potential for more permanent thermal stratification
increase the possibility of anoxia in some basins and bays, excluding such regions from fishes.
Policy issues include the continued management of cold, cool, and warmwater fisheries for what appears
to be higher recreational and commercial fishery opportunities, continuing to protect the water quality of the
lakes especially in regard to nutrients contributing to eutrophication and toxics that accumulate in fishes, the
continuing value of lake trout for rehabilitation of populations, intensified predator pressure on the forage base
and intensified species interactions, and the potential problem of negative species interactions between present
warm and coolwater fishes and exotic warmwater fishes that will find expanded habitats, especially in inshore
areas.
'Although the information in this report has been funded wholly or in part by the U.S. Environmental
Protection Agency under Contract No. CR 814644-01-0 under the Office of Policy, Planning and Evaluation, it
does not necessarily reflect the Agency's views, and no official endorsement should be inferred from it.
2-1
-------
Magnuson
CHAPTER 1
INTRODUCTION
DESCRIPTION OF ECOLOGICAL SYSTEM
The Great Lakes lie within a relatively small watershed basin (Figure 1) that has long been the industrial
heartland of North America and is now home to some 37 million people. In a world in which fresh water is
becoming a scarce resource, our wealth is reflected in the fact that these lakes contain about one-sixth of all the
surface fresh water of the globe, not including the glaciers and groundwaters. This chain of lakes drains through
the St. Lawrence River into the Gulf of St. Lawrence, where the freshwater outflow plays a role in the
reproduction and recruitment of the large herring fisheries of the gulf.
By 1950, the southern third of the basin had become a vast ecological slum because of careless agricultural,
forestry, industrial, and urban development in preceding decades. A series of studies then led to binational
conventions and agreements to rehabilitate the fisheries and remediate the bad practices that degraded water
quality. Many billions of dollars have been spent in these efforts by the various governments at different levels.
Some programs have been relatively successful: control of the predaceous sea lamprey, correction of overfishing,
reduction in loadings of old-fashioned industrial wastes and of plant nutrients from sewage, and discontinuation
of the use of persistent pesticides. Prospects are improving that loadings of hazardous contaminants and acidic
substances will progressively be contained and reduced.
By 1950 the ecological quality of the lakes, streams, and shores was generally offensive to humans, especially
in cities and towns. People had turned their backs to the lakes. Some four decades later, many people are
rediscovering these ecosystems as they are now recovering from past abuses. Shore properties are escalating in
value and are being redeveloped with sensitive uses in mind. Commercial fisheries have rebounded, but then-
value has been dwarfed by new recreational fisheries that are based, in part, on a new "manmade" association
of fish species thai is dominated by Pacific salmons. Parks, marinas, beaches, promenades, and condominiums
along the shorelines are increasing in number and quality as the demand for them grows.
Commitment to a healthy Great Lakes ecosystem as reflected in an abundance of valued fish is part of a
regional policy to redevelop, on a sustainable basis, the cities, industries, farmlands, and forests of the basin, as
key to the resurgence of prosperity in the basin. In the face of this optimistic prospect, concern is growing
about the likely effects of climate change.
Will the warming trigger relapses in the ecological recovery of degraded parts of the lakes? Will increased
temperatures foster undesirable ecological productivity and thus impede recovery from cultural eutrophication?
Will the higher temperatures create difficulties, in southern waters of the basin, for the coldwater salmons and
lake trout? Will higher temperatures facilitate invasion of new species with some of them interacting to the
detriment of some valued species of the present association? WU1 valued new species appear in the warmer
habitats? Will sea lamprey benefit in Lake Superior to the detriment of lake trout and other valued fish species?
Such questions come to mind as we consider the possible ecosystem impacts of climate change.
In its pristine state, the fish association of the Great Lakes could be separated into five major types:
i) The deeper, colder waters of the lakes and rivers were dominated by the salmon family (lake trout, lake
whitefish, lake herring, chubs), the cod family (burbot), and the cottid family (sculpins).
ii) Waters of intermediate depth that were cool even in midsummer were dominated by the perch family
(yellow perch, walleye, sauger, etc.) and the pike family (northern pike, muskellunge, etc.). The sturgeon family
(lake sturgeon) and the ictalurids (bullheads, channel catfish) frequented the bottoms of cool waters.
2-2
-------
Magnuson
Great Lakes Drainage Basin Boundary
Southern Basin Silt
Chicago o\X Eva5S'°n
Rgure 1. Great Lakes drainage basin and sites referred to in text Symbols: 0 « air temperature data site,
X = water temperature data site.
2-3
-------
Magnuson
iii) In a few of the wannest waters of moderate depth, the "percichthyid" family (white bass) fluctuated in
abundance over the decades.
iv) Shallow, warm inshore waters, especially of bays and streams, were the home of the sunfish family
(sunfishes, crappies, smallmouth bass, largemouth bass, etc.) and the cyprinid family (minnows).
v) In the forested coldwater tributary streams, again the salmon family (brook trout) and cottid family
(sculpins) predominated.
At the outset, it seems likely that each of these five types of associations will react somewhat differently to
climate wanning in a particular locale. Coldwater habitats may shrink toward the north in the lakes and
upstream into the headwaters; coolwater habitats may be shifted northward and upstream; wannwater habitats
may expand from the south northward and from the protected bays into the lake proper.
ORGANIZATION OF THIS REPORT
Ecological phenomena in our report are addressed in a hierarchic context across several levels of
organization, progressively from the organism level, to the population level, to the ecosystem level First, the
report considers the potential influence of altered thermal structure on the size of the habitat within the thermal
niche of fishes which prefer and do best in cold, cool, or warm waters. Second, our report considers how the
body growth and the prey consumption of coldwater, coolwater, and wannwater fishes may be changed by living
near-shore in Great Lakes neritic habitats altered by climate change scenarios. Third, for a wannwater fish
that lives in the shallowest inshore habitats and for which the best models were available, our report evaluates
the response of a population of fishes in respect to reproductive success and population size. Fourth, the report
steps back from the species level and estimates the response of phytoplankton production, zooplankton biomass,
and fishery yields as entire trophic levels to climate warming scenarios. To seek generality among global climate
models, three climate change scenarios were applied: Oregon State University's (OSU) general circulation model,
the general circulation model from the Geophysical Fluid Dynamics Laboratory (GFDL), and the general
circulation model from Goddard Institute of Space Studies (GISS). To seek generality across the Great Lakes,
simulations were made for a warm region in Lake Erie, a cool region in southern Lake Michigan or Lake Huron,
and a cold region in northern Lake Superior. Finally, the report discusses what we do not know or have not
addressed especially in regard to the possibility of surprises that could occur in the dynamic food webs perturbed
by altered fish species interactions. A recent review paper by Meisner et al. (1987) is appended for general
reference.
2-4
-------
Magnuson
CHAPTER 2
THERMAL NICHE
GREAT LAKES
Fishes of the Great Lakes have been grouped into three broad thermal guilds for which a central
temperature has been estimated by Hokanson (1977): coldwater ca 15.0°C; coolwater ca 24.0°C; and warmwater
ca 28.0°C. Body temperatures of fishes are determined by the temperatures of the water in which they live, and
changes in their environmental temperatures directly influence their survival, physiology, and behavior (Fry,
1971; Hokanson, 1977; Magnuson et al., 1979). Fish species within a thermal guild have preferred temperature
ranges in which growth is maximal. The median temperature of the preferred range of a species is where the
species will be found at greatest densities if food resources are not limiting and interspecific competition is not
high. The thermal niche of a species was defined as ± 2.0°C of the median temperature by Magnuson et al.
(1979).
The amount of thermal habitat within a fish's thermal niche has been shown to be a statistically significant
determinant of fishing yield (Christie and Regjer, 1988). Estimates of maximum sustained commercial fish yield
of lake trout, lake whitefish, and walleye were directly related to the amount of lake bottom area or pelagic
volume contained within the optimal temperature ranges of these species in a set of 21 large North American
lakes.
Altered thermal structure of the Great Lakes owing to climate warming could increase or decrease the
size of optimum thermal habitats available for important fishes. This would ultimately change commercial and
sport fishing yields of valued species. The above empirical model of thermal niche is used here to estimate the
amount of thermal habitat available for the thermal guilds of Great Lakes fishes before and after applying the
global climate warming scenarios. Expected changes in potential yield of several commercial fishes are estimated
from the Christie and Regier (1988) scaling of maximum sustained yields to summer thermal habitat.
Methods
Potential effects of climatic warming on the size of thermal habitats of fishes representing several of the
thermal guilds were simulated using thermal structure data from southern Lake Michigan (McCormick 1988)
and the central basin of Lake Erie (Blumberg, 1988). These results were subsequently used to estimate changes
in fishery yields.
Thermal habitat was estimated for individual species from three thermal guilds. The thermal niche was
defined in two alternative ways, one narrow and the other broad: (1) the optimum ±2°C, which includes the
temperatures within which a fish in a preference tank in the laboratory will spend 2/3 of its time; and (2) the
optimum ±5°C, which includes the temperature a fish in the laboratory chooses to spend all of its time, if such
temperatures are available (Magnuson et al., 1979; Crowder and Magnuson, 1983). Hokanson (personal
communication) has compiled data suggesting that stream fishes spend -90% of their time within the thermal
range of optimum ±5°C. In some cases we estimated the thermal habitat for the entire year, in others only
from June 5 to September 2 to be able to use the fishery yield prediction models of Christie and Regier (1988).
In some cases, the thermal niche was measured in terms of depth days or the number of meters of the water
column within the thermal niche summed across dates and expressed in terms of m-days of suitable space in the
water column. For southern Lake Michigan, the size of the thermal niche was estimated in terms of the volume
days of suitable thermal habitat and expressed as hm 3.7 days. For walleye and yellow perch, we calculated the
weekly ha of bottom area exposed to suitable temperatures based on the approach of Christie and Regier (1988).
The average and minimum weekly thermal niche for the summer period (weeks 23-35) was also calculated
for some species by computing the weekly mean and minimum weekly depth-days or volume-days of suitable
thermal habitat. This measure of habitat size provides insight to periods during which suitable thermal habitat
2-5
-------
Magnuson
may be most limited (Coutant, 1987a; Rudstam and Magnuson, 1985). In the central basin of Lake Erie, the
mass of cold water is small and usually anoxic in summer, so we did not estimate any thermal habitat in Lake
Erie for coldwater fishes.
Lake trout is an example of a coldwater fish with a lower optimum (10°C) (Stewart et al., 1983) than most
coldwater fishes. It is a species which, at first thought, should be most negatively influenced by warming of its
habitat. Coho salmon is perhaps a more representative example of a coldwater fish with its optimum of 15° C
(Hewett and Johnson, 1987). Yellow perch provide our example of coolwater species with an optimum of 23°C,
and largemouth bass serves as the wannwater fish with an optimum of 27.5°C (Hewett and Johnson, 1987).
Other individual species were used as well: lake whitefish at 12°C, and walleye at 20°C (Christie and Regier,
1988).
The statistical model equations relating summer thermal habitat to fishery yields from Christie and Regier
(1988) are:
Lake trout Iog10 SY = 0.807 loglO THV + 0.944, r2 - 0.86;
Lake whitefish 4og" SY = OJ76 loglO THV + 3.606, r2 = 0.69;
Walleye logjjj SY = 0.872 loglO THA + 0389, r2 - 0.72
where SY is an estimate of sustained yield in kg/year, THV is summer thermal habitat (±2°C) volume in hm3 • 10
days from June 3 to September 2, and THA is summer thermal habitat (±2°C) bottom area in ha. 10 days from
June 3 to September 2.
The summer period as defined by Christie and Regier (1988) omits habitat that might exist during the
spring and fall. Sufficient data for spring and fall, necessary for their statistical analyses, were not available;
hence, their possible contribution to production could not be assessed. During the fall and late spring, there are
periods during which water temperatures in their study lakes were very likely within the optimal ranges for lake
trout and lake whitefish. The thermal habitat for these species during these periods may have contributed to the
sustained yields they used in their analyses.
Scenarios
Thermal structure throughout the year was obtained from McCormick (1988) for Lake Michigan and from
Blumberg (1988) for Lake Erie. OSU-simulated thermal structure was not available for Lake Erie. The three
simulations we used are BASE, the present conditions (IxCO,); the OSU scenario from the OSU 2xCO2 general
circulation model; the GISS scenario derived from the GISS 2xCO, general circulation model; and the GFDL
scenario derived from the GFDL 2xCO, general circulation model. For Lake Erie we used the profiles
generated from 1975 (a cool year) and 19/0 (a warm year).
Results
A profile of the thermal habitat of lake trout throughout the year (Figure 2) clarifies the concept of thermal
niche as the m-days of water within ±2°C (black) and ±5°C (grey) of optimum. In these terms the thermal
habitat under base climate is 1*500 m-days for i2°C and 5322 m-days for ±5°C. Typically the temperature in
winter and spring mixis are too cold to be in the optimum range for lake trout as are the deeper waters in
summer. In summer the near-surface waters are too warm. During fall mixis, much of the upper 100 m is within
their thermal niche.
Under all three climate warming scenarios (OSU, GFDL, and GISS) the quantity of suitable annual
thermal habitat increases greatly for lake trout (Figure 2). The increases are most pronounced for the GFDL
and GISS scenarios where much of the upper 100 m of the water column reaches temperatures above 5°C in
winter. Temperatures between 5 and 15°C persist long into the fall after climate wanning. The size of
2-6
-------
50
100
Magnuson
BASE CLIMATE
D
E
P
T
H
(M) 50
100
92 183 274
DAY NUMBER
365
Figure 2.
Extent of favorable thermal habitat before and after climate wanning for a coldwater fish (lake
trout) in southern Lake Michigan based on WC*TC (black) and KrC±5°C (grey). Vertical lines
span the period (June 5 to September 2) used in the fishery yield prediction models of Christie and
Regier (1988). Thermal scenarios for Lake Michigan provided by McCormick (1988).
2-7
-------
Magnuson
unsuitably warm waters near the surface in summer increases with the climate warming scenarios. With the
GISS and GFDL scenarios, about 40% of the upper 100 m are within the thermal niche of lake trout compared
with 22% with the OSU scenario and only 14% under the base climate. Climate warming scenarios would
increase the amount of favorable thermal habitat over the year even for a coldwater fish like the lake trout.
The size of the annual thermal habitat (Figure 3) not only increases for lake trout (factor of 1.5 to 2.7 x
BASE) in Lake Michigan's southern basin under climate warming, but also for other coldwater fishes like coho
salmon (13 to 1.5 x BASE), and coolwater fishes like yellow perch and walleye (1.4 to 3.1 x BASE). A small
amount of thermal habitat also is opened up or expanded for warmwater fishes. In Lake Erie, climate warming
increases the thermal habitat both for coolwater (1.2 to 23 x BASE) and warmwater (23 or greater x BASE)
species (Figure 3). These results are robust and not sensitive to the particular method of estimation. The change
in the amount of thermal habitat for lake trout measured in various ways (Figure 4) indicates that when the
whole year is included, thermal habitat increases with climate warming regardless of whether ±2°C or ±5°C is
used as a criterion or whether depth days or volume days are used as a metric.
The average summer habitat within the thermal niche of various fishes in Lake Michigan remains about
the same or decreases (Figure 5). For the lake trout (coldwater) and the yellow perch (coolwater), the summer
niche space decreases slightly as delineated by ±2°C of optimum and the vertical lines in Figure 2. Decreases
(Figure 5) with climate warming for walleye (coolwater) and lake whitefish (coldwater) were large.
Sustained yields for Lake Michigan estimated by the Christie and Regjer (1988) model are predicted to
decrease only slightly for lake trout (2 to 6%) and lake whitefish (7 to 10%) with GISS and GFDL climate
scenarios but increase significantly for walleye (+29 to +33%) (Figure 6).
HEADWATER STREAMS
Waters of many headwater streams in the Great Lakes Basin, especially those which originate in sandy and
gravelly moraines with copious groundwater flows and springs, are inhabited by brook trout, a coldwater fish.
Groundwater temperatures track local average air temperatures and are expected to increase with climate
wanning (Meisner et al., 1988). The major impact of climate change on the streams maybe on the temperature
of the groundwater entering the streams and on the rate of warming of water flowing downstream from
headwater springs. A calibrated hydrometeorological model of stream temperature (Delay and Seaders, 1966)
can be used to estimate summer stream temperatures before and after applying global climate change scenarios.
Longitudinal distributions of brook trout in nondegraded streams normally are bounded between 19 and 24°C
(Hawkes, 1975; Ricker, 1934), but Barton et al. (1985) and Bowlby and Roff (1986) found that water
temperatures near 22°C marked the downstream habitat boundaries for Ontario. Combining the stream
temperature estimates along the course of a stream with the thermal requirements of brook trout can provide
estimates of changes in the size of thermal habitat for coldwater fishes in headwater streams.
Methods
A hydrometeorological model of summer thermal regime was implemented for the brook trout zone of
two first-order headwater streams: the Rouge and Humber Rivers, which ultimately empty into Lake Ontario
east of Toronto (Figure 1). The streams originate as groundwater discharge at about 10°C. The Humber River
is about 5 km long with a mean discharge of .06 m3/s and has forested banks, while the Rouge River is about
3 km long, has a mean discharge of .075 m3/s, and flows through open fields. The streams are similar to other
brook trout streams of southern Ontario.
The basic model (Delay and Seaders, 1966) simulates the change in temperature of the water as it flow
downstream:
2-8
-------
Magnuson
LAKE ERIE
(0
Z
I-
o
I
GC
LU
HI
500
400-
300
200
100
0-
•
COLDWATER
HABITAT
UNAVAILABLE
1
COLDWATER
HABITAT
UNAVAILABLE
• BASE
H QISS
m GFDL
__H.^B§1
COLD WARM COLD WARM
YEAR YEAR YEAR YEAR
DC
ULJ
z
h-
LAKE MICHIGAN
500
250-
COLDWATER COLDWATER COOLWATER
FISHES FISHES FISHES
(LAKE TROUT) (COHO SALMON)
WARMWATER
FISHES
Figure 3. Amount of thermal habitat (optimum ±5°C) in the water columns of the central basin of Lake Erie
and the southern basin of Lake Michigan for coldwater fishes (lake trout, coho salmon), coolwater
fishes (yellow perch) and warmwater fishes (largemouth bass) for the BASE climate and for the
OSU, GISS, and GFDL climate change scenarios using the thermal structure simulated by Blumberg
(1988) for Lake-Erie and by McCormick (1988) for Lake Michigan. For Lake Erie, the cold year
was 1975 and the warm year was 1970.
2-9
-------
Magnuson
DEPTH-DAYS DEPTH-DAYS VOLUME-DAYS
*c ,
THERMAL HABITAT
FOR ENTIRE YEAR
C
VOLUME-DAYS KG / YEAR
*~2 ° YIELD
THERMAL CHRISTIE
HABITAT & REGIER
SUMMER MODEL
Figure 4. Change in the size of southern Lake Michigan thermal habitat expressed as a ratio of the GISS or
GFDL estimates to the estimates from the BASE climate for a coldwater fish (lake trout) measured
by various criteria; optimum ±2°C for the entire year or ±2°C for the summer months without and
with the correction for lake hypsometry, and the estimated change in maximum sustained yields
using the summer months and ±2°C criteria with the Christie and Regier (1988) model The
temperature structure of Lake Michigan is from McCormick (1988).
2-10
-------
Magnuson
AVERAGE
THERMAL
HABITAT
(HM?7 DAYS OR
HA-7 DAYS)
400000
300000
200000-
100000-
• BASE
Bl GISS
m GFDL
# HM3-7DAYS
* HA -7 DAYS
LAKE LAKE WALLEYE YELLOW
TROUT WHITEFISH • PERCH
n #
Figure 5. Average size of thermal habitats averaged over the summer months (weeks 23-35) for four Lake
Michigan fishes based on a thermal niche of optimum ±2°C for the BASE climate and GISS and
GFDL climate warming scenarios. Temperature structure of southern Lake Michigan is from
McCormick (1988), hypsometry is for all of Lake Michigan.
2-11
-------
Magnuson
SUSTAINED
YIELD
1,000,000
KG/YR
LAKE TROUT
LAKE WHITEFISH
WALLEYE
Figure 6. Maximum sustained yields of lake trout, lake whitefish, and walleye in southern Lake Michigan
based on summer optimum ±2°C thermal habitat volumes for the BASE climate and the GISS and
GFDL climate change scenarios. Yields estimated from the Regier and Christie (1988) model,
thermal structure for southern Lake Michigan from McCormick (1988), hypsometry for the whole
lake.
2-12
-------
Magnuson
AT,.. = V where,
w
ATW = change in temperature of parcel of water;
TW = initial temperature of parcel of water;
A = surface area of parcel of water;
V = volume of parcel of water;
v = volume of inflow (e.g. groundwater, tributary);
ty = temperature of inflow;
and S = change in energy stored in parcel of water.
S quantifies the energy flux across the air/water interface in a series of equations not shown here. This
term treats the effects of air temperature, solar and longwave radiation, evaporation, conduction and wind on
stream temperature explicitly. The model adapts to varying hydrologjcal conditions along the stream and
accounts for the thermal effects of shade.
The model was used to estimate the farthest downstream extent of temperatures suitable for brook trout
in the wannest part of the diel cycle for a date in late July before and after applying the 2xCO2 climate wanning
scenarios. Observations indicate that the thermal barrier defining the downstream limit of the brook trout zone
is 23°C in the Humber River and 24°C in the Rouge River.
Scenarios
Simulations of maximum July water temperatures were made using July 24,1987, water temperatures from
the sites and 1987 meterological data from the sites for BASE conditions and applying the 2xCO2/lxCO2
adjustment statistics for July and August for the OSU, GISS, and GFDL models. Groundwater temperatures
in the 2xCO2 simulations were derived using the mean annual air temperature for each scenario.
Results
The length of stream suitable for brook trout declines markedly in the 2xCO, simulations of the Rouge
and Humber Rivers (Figure 7). The loss of thermal habitat is least pronounced in the OSU scenario and most
significant in the GFDL scenario, where suitable length in both streams is about half that found in BASE
conditions. Although changes in groundwater discharge were not included in these simulations, any drying of
the climate would likely exacerbate the habitat losses displayed here.
2-13
-------
Magnuson
O
LU
ROUGE RIVER
NUMBER RIVER
BASE
OSU
GISS
GFDL
Figure 7. Length of the brook trout zone in two river headwaters in southern Ontario during July for the
BASE climate and the OSU, GISS, and GFDL climate warming scenarios.
2-14
-------
Magnuson
CHAPTERS
GROWTH AND FOOD CONSUMPTION
The physiological processes of fishes depend strongly on water temperature (Fry, 1971), and coupled with
the rate of food consumption temperature determines the rate at which fish grow. The fish bioenergetics model
of Kitchell et al. (1977) as given in Hewett and Johnson (1987) allows the computation of growth given the
food consumption and the water temperature environment of the fish. Alternatively prey consumption can be
estimated from knowledge of water temperature and growth using the bioenergetic model. This methodology
has been applied to a variety of individual fish species (Kitchell et aL, 1977; Rice et aL, 1983; Stewart et al.,
1983; Rice and Cochran, 1984; Stewart and Binkowski, 1986) as well as of interspecific interactions (Kitchell and
Breck, 1980; Stewart et al., 1981, Lyons and Magnuson, 1987).
The impact of increased water temperatures on fish growth and food consumption in the Great Lakes will
be assessed using the bioenergetic models. The potential effects of a temperature change on growth and food
consumption are estimated by-applying the model to temperature scenarios representing conditions before and
after climate wanning. Simulations using different proportions of the maximum consumption rate were used in
an attempt to bracket a range of potential energetics responses to climate warming.
METHODS
The bioenergetics model computes energy budget statistics for individual fish using the equation G-C
(R + F+U), where G is the growth rate, C the rate of food consumption, R the sum of metabolic costs, and F
and U the waste products due to egestion and excretion (Kitchell et al. 1977). All of the terms in this equation
are strongly dependent on temperature. The physiological processes (R, F, and U) are known functions of
temperature and feeding rate. Therefore, fish growth may be computed if the temperature and feeding rate are
known.
To evaluate the range of potential energetics responses of fish to climate warming, two sets of energetics
simulations were run using different assumptions about prey consumption rates. The first set of simulations
were run under the assumption that as the climate changes, the fish will continue to feed at the same fraction
of their maximum consumption rate as in the base period. The assumption that the proportion of maximum
consumption (called the "P-value") is constant implies that prey availability will increase with climate warming
at a rate such that actual consumption remains the same fraction of maximum potential consumption. P-values
were chosen to produce growth in the range observed in the Great Lakes for baseline climatic conditions.
Baseline P-values were 0.7 for largemouth bass, 0.45 for yellow perch, and 0.75 for lake trout. A P-value of 0.7
means, for example, that a fish consumes prey at a rate of 0.7 times the maximum consumption rate, which is
a function of temperature and body weight. In the second set of simulations, annual prey consumption was
assumed to remain constant at baseline levels in the 2xCO« scenarios. This assumption means that the P-value
in the 2xCO, simulations is lower than in the base scenarios.
In both sets of simulations, growth was computed for representative "cold," "cool," and "warm" water fishes:
lake trout, yellow perch, and largemouth bass. Each individual simulation consisted of a yearling fish exposed
to one year of the water temperature scenarios, beginning in February, the coldest month in the annual cycle.
Starting weights were 8 g for yellow perch and 20 g for both lake trout and largemouth bass. For all species,
a generic prey was used in the models with an energy density corresponding to a mixed diet of fish and
invertebrates. The lake trout prey was given a higher energy density to reflect the importance of the energy-
rich alewife in their diet (Stewart et al., 1983).
An important assumption used in our bioenergetic simulations is that fish behaviorally thermoregulate; they
seek out a preferred temperature which is close to their optimal temperature for growth (Neill and Magnuson,
2-15
-------
Magnuson
1974; Magnuson et al., 1979). Thus, the maximum water temperature in our scenarios was limited to the
preferred temperature of the species being modeled. The assumption of behavioral thermoregulation includes
the implicit assumptions that the preferred temperature is available to the fish, that oxygen concentration is
adequate at that temperature, and that sufficient prey are available to support the assumed feeding rates. The
following preferred temperature values were used: 10°C for lake trout, 23°C for yellow perch, and 27.5°C for
largemouth bass (Hewett and Johnson, 1987).
SCENARIOS
The energetics simulations with prey consumption held constant at baseline levels in 2xCO2 climate
conditions were run using McCormick's (1988) water temperature scenarios for the southern basin of Lake
Michigan (see Figure 1). The simulations with prey consumption increasing with climate warming were applied
to three Great Lakes near-shore sites representative, respectively, of "cold," "cool," and "warm" thermal
environments in baseline climate conditions: Thunder Bay, Ontario (Lake Superior), Evanston, Illinois (Lake
Michigan), and Put-In Bay (Lake Erie) (see Figure 1). EPA's standard procedure was followed to create air
temperature scenarios for the three sites. Daily average air temperature observations from Cleveland, Ohio
(Put-In Bay site), Chicago, Hlinois'(Evanston site), and Thunder Bay, Ontario (Thunder Bay site) were used to
compute monthly average air temperatures. These BASE scenarios covered the period 1951-1980 for Chicago
and Cleveland, and 1953-1980 for Thunder Bay. The OSU, GISS, and GFDL air temperature scenarios were
created by multiplying temperatures in the base period by the appropriate 2xCO2/lxCO2 ratio from the climate
model output, using the closest model gridpoint to each site of interest. The air temperature scenarios were
then used to create water temperature scenarios representing inshore epilimnetic conditions at each site.
Multiple regression models were developed relating the observed monthly averaged water temperature to the
current month's air temperature and the previous month's air and water temperatures (see Table 1 for
equations). Three different seasonal models were developed for each site: a winter model and models for the
periods of rising and falling water temperatures. Water temperature data for Put-In Bay were obtained from
the Put-In Bay fish hatchery. The years 1954-1973 and 1976 were used to calibrate the model The city of
Evanston water supply intake was the source of Lake Michigan temperature data, spanning the period 1974-1987.
Lake Superior data between 1958 and 1977 were obtained from the city of Thunder Bay water intake. For each
site, daily water temperatures were averaged to monthly values.
The base period water temperature scenarios (BASE) were derived by applying the regression models to
the monthly average observed air temperatures for 1951-1980. The model was run in an iterative fashion, using
the predicted water temperature for the current month in the computation of the next month's water
temperature. This approach requires in addition to the air temperature data only an initial guess of the water
temperature for January 1951 to generate the entire 30-year scenario. Similarly, the OSU, GISS, and GFDL
water temperature scenarios were derived by applying the regression equations to the respective air temperature
scenarios. Water temperature scenarios were then modified for thermoregulation by cutting off the temperature
at the fishes' preferred value.
The water temperature scenarios are summarized in Figure 8. The left column displays the average and
extreme values for the water temperature scenarios over the entire 30-year period. For every site, the GFDL
climate change scenario is the warmest, the GISS scenario intermediate, and the OSU scenario the coolest In
each case, Lake Erie experiences the greatest increase in annual mean temperature, while Lake Superior changes
the least. The GFDL mean annual water temperature is 33°C warmer than the BASE scenario for the Lake
Superior site, 3.4°C warmer for Lake Michigan, and 5-8°C wanner for Lake Erie. The seasonal plot of average
monthly temperatures in the right column of Figure 8 shows that the GFDL scenario is uniformly warmer than
the OSU and GISS scenarios, with the largest difference occurring in midsummer. The horizontal lines bisecting
the annual temperature cycle are the optimal temperatures for growth used in the bioenergetics simulations.
Under the behavioral thermoregulation assumption, the fish follow these lines as soon as the near surface waters
warm above their preferred temperature. The increased width of the thermoregulation lines for the OSU, GISS,
and GFDL scenarios over those for the BASE scenario means that the fish have a longer growing season wherein
they may, with behavioral thermoregulation, operate in optimal temperature conditions for growth.
2-16
-------
Table 1. Regression Equations Relating Water Temperatures to Air Temperatures for Warm,
Cool, and Cold Great Lakes Sites3
Thunder Bay (Lake Superior)
Winter: WT=0.418+0312WTL+0.036AT-0.023ATL R2=038
Warming: WT=0.440+ 0.157WTL+0357AT+0.239ATL R2-0.90
Cooling: WT=2.534+0324WTL+0.164AT+0.174ATL R2=0.91
Evanston (Lake Michigan)
Winter: WT=1.475+0.363WTL+0.106AT-0.027ATL R2=0.57
Warming: WT=3347+0304WTL+0.217AT+0.208ATL R2=0.90
Cooling: WT=1.928+0377WTL+0382AT+0.067ATL R2=0.%
Put-In Bay (Lake Erie)
Winter: WT=0.959+0.154WTL+0.140AT-0.053ATL R2=0.49
Warming: WT=2.896+0.451WTL+0.524AT-0.009ATL R2=0.%
Cooling: WT=-7.642 + 1.022WTL+0386AT-0.173ATL R2-0.94
aSymbols and units: WT = current month's mean water temperature; WTL = last month's mean
water temperature; AT = current month's mean air temperature; ATL = last month's mean air
temperature (all temperatures °C); R2 = coefficient of determination.
2-17
-------
Magnuson
SUMMARY
MONTHLY MEANS
30-
20-
10-
0-
COLD AREA
COLD COLDWATER
FISH
GISS
UJ
GC
QC
UJ
Q.
UJ
H
QC
UJ
i
30-
20-
10-
COOL AREA
I T I
COLO COLDWATER
FISH
30-
COLD COLDWATER
FISH
JAN
Figure 8. Water temperatures in cold, cool, and warm neritic areas of the Great Lakes (Thunder Bay, Lake
Superior; southern Lake Michigan; Put-In Bay, Lake Erie) for the BASE climate and the OSU,
GISS, and GFDL scenarios based on multiple regression relations between water intake and air
temperatures. Optimum temperatures for three thermal guilds of fishes are shown.
2-18
-------
Magnuson
Water temperatures in the base scenario for Evanston, Lake Michigan, are several degrees cooler in
midsummer than those found for the central Lake Michigan site by McCormick (1988). This difference is likely
owing to the fact that the water intake site at Evanston is in a nearshore site subject to rapid temperature drops
caused by upweUing events, internal waves, etc. These factors may also cause our scenarios at Thunder Bay and
Put-In Bay to underestimate actual epilimnetic temperatures.
RESULTS
The consequences of longer growing seasons under the 2xCO2 climate scenarios when prey consumption
is allowed to increase with climate warming are readily apparent in Figure 9. Coldwater and coolwater fishes
Qake trout and yellow perch) exhibit clear increases in annual growth over that in the base period The ratio of
growth of cold coldwater and coolwater fishes in 2xCO2 conditions to that in the base scenario ranges from 1.2
to 1.7 (mean 13). Wannwater fishes (largemouth bass) cannot grow in any of the cold area temperature
scenarios or in base conditions in the coolwater area They experience minor growth in the 2xCO, scenarios
in the cool area, and large increases in growth in the warm area (2xCOu/lxCO2 ratios of 3.2, 43, and 5.7).
Although in all cases the fishes under the GFDL scenarios grow faster than those under the GISS scenarios,
differences are minor compared with differences between the 2xCO2 and baseline scenarios.
For all three thermal guilds of fishes, the site with the largest simulated growth is the warm area. This
may seem inconsistent with the intuitive notion that a coolwater fish should do best in a coolwater environment.
However, Figure 8 shows that the water temperature scenarios for the warm area possess longer periods when
preferred temperatures may be available for all three fish species. In the case of Lake Erie lake trout, suitable
temperature and oxygen conditions are not found in the vicinity of the Put-in Bay site in summer. Figure 8
shows that Put-In Bay is potentially habitable for lake trout for 3-4 months during winter after climate warming.
Historically, lake trout in Lake Erie have used the eastern basin as a summer refuge and migrate into the central
basin during the fall. Also, other coldwater fishes (cisco and lake whitefish) use areas near Put-In Bay in fall
for spawning, but are not present during summer.
Results for prey consumption (Figure 10) closely parallel those for growth, because growth is an integrated
form of consumption. Changes in consumption between the BASE and 2xCO2 scenarios ranged from factors
of 1.2 to 1.7 (mean 13) for cold coldwater and coolwater fish. Wannwater fish consumption rates ranged from
1.6 to 2.8 (mean 2.1) times higher than the base period Similar to the result seen for growth, Lake Erie fishes
consumed more than their counterparts in the other lakes.
The effect of the two assumptions about prey availability on fish growth following climate warming is
readily apparent in Figure 11. When prey consumption is allowed to increase with climate warming, growth
increases. The 2xCO2/lxCO2 growth ratios are about 1.6 for coldwater fish and 1.4 for coolwater fish.
Wannwater fish, which lost weight in the baseline scenario, are able to grow in the 2XCCL scenarios. When
consumption is held constant in the climate wanning scenarios, growth for all three fish guilds declines.
Wannwater fish lose even more weight than in the base scenario. Coolwater and coldwater fishes also suffer
in the equal consumption simulations (2xCO2/lxCO2 growth ratios of 0.7, 0.9). Differences in growth caused
by the two assumptions concerning prey consumption are much larger than the differences in growth between
climate model scenarios. This implies that fish growth is more sensitive to changes in food web structure (Figure
12) than to differences in the temperature scenarios.
2-19
-------
Magnuson
COLD
AREA
COOL
AREA
WARM
AREA
400
CC
<
LJJ
COLDWATER
FISH
CO
LL
o
I
I-
o
DC
COOLWATER
FISH
200
z
<
150
100
50
WARMWATER
FISH
B 0 G G
A S I F
S U S D
E S L
B O G G
A S I F
S U S D
E S L
B 0 G G
A S I F
S U S D
E S L
Figure 9.
Annual growth (tlSD) of cold, cool, and warmwater fishes (lake trout, yellow perch, and largemouth
bass) in cold, cool, and warm neritic areas of the Great Lakes (Thunder Bay, Lake Superior; southern
Lake Michigan; Put-In Bay, Lake Erie) for the BASE climate and the OSU, GISS, and GFDL
climate change scenarios. These simulations assume prey consumption rises with climate wanning.
2-20
-------
COLD
AREA
COOL
AREA
WARM
AREA
Magnuson
COLDWATER
FISH
WARMWATER
Figure 10. Annual prey consumption (±1SD) by cold, cool, and warmwater fishes (lake trout, yellow perch, and
largemouth bass).in cold, cool, and warm neritic areas of the Great Lakes (Thunder Bay, Lake
Superior; southern Lake Michigan; Put-In Bay, Lake Erie) for the BASE climate and the OSU,
GISS, and GFDL climate wanning scenarios.
2-21
-------
400
Magnuson
300
LJLJ 200
S 1°°
GO 0
50
40
30
20
10
£ °
O
CC
O
cc
<
LU
LL
O
COLDWATER FISH
COOLWATER FISH
WARMWATER FISH
GISS GFDL , BASE ,
GISS GFDL
CONSUMPTION
INCREASES
CONSUMPTION
CONSTANT
Figure 11. Annual growth of cold, cool, and warmwater fishes (lake trout, yellow perch, and largemouth bass)
in southern Lake Michigan with prey consumption allowed to increase or stay constant as climate
warms. Lake Michigan thermal scenario provided by McCormick.
2-22
-------
Magnuson
LAKE TROUT
PACIFIC SALMON
ISCULPINS
J.
[ALE WIFE]
BENTHOS
A
ZOOPLANKTON
/\
PRIMARY PRODUCERS AND
DECOMPOSERS
Figure 12. Idealized Great Lakes food web diagram. Question marks and varied widths of energy transfer
arrows highlight the uncertainties about whether production at lower trophic levels will be sufficient
to meet the predatory demand of higher levels.
2-23
-------
Magnuson
CHAPTER 4
POPULATION RESPONSE
Smallmouth bass, a wannwater fish, was used to demonstrate population responses to global climate
wanning. These animals live in the shallowest, most protected areas of the Great Lakes and are, therefore,
likely to show the effects of climate wanning earlier than many other species. However, since the Great Lakes
watershed is near the northern limit of the natural range of the species, these effects will almost certainly be
positive. This must be kept in mind while reviewing our results. Our choice of smallmouth bass was dictated
solely by the fact that reliable recruitment and growth models for this species were already in existence and
could be used immediately to produce credible forecasts of the effects of climate change on year class strength
and population abundance.
The reproductive rate of any fish population is strongly dependent on the average survival of the young and
the age at which adults reach sexual maturity. For smallmouth bass, both survival and age of maturity are
strongly dependent on temperature. Overwinter survival of young bass depends on their size at the end of their
first summer, and this in turn depends on the length of the growing season (Shuter et aL, 1980). Similarly, age
of first maturity varies directly with growth rate and, therefore, with length of the growing season (e.g., Wrenn,
1980,1984). Thus, population responses of these fishes should be affected by a change in climate.
METHODS
Three sites (Figure 1) were chosen for the estimated population response of smallmouth bass: warm - a
shallow bay near Long Point on Lake Erie; cool — a shallow bay, Baie du Dore, Lake Huron; and cold ~ a
shallow bay near Thunder Bay on Lake Superior. Smallmouth bass populations presently exist at both the Lake
Erie and Lake Huron sites. Data from these populations were used extensively in developing forecasts of the
potential effects of climate change. The Lake Superior site is an imaginary bay physically similar to Baie du
Dore.
The smallmouth bass recruitment model of Shuter et al. (1980) was used to estimate winter mortality and
the length of the growing season for young-of-the-year fish. Growth for all age groups was estimated using the
Kitchell bioenergetics model (Kitchell et al., 1977; Hewett and Johnson, 1987). Age of first maturity (Tin) was
defined as the number of years of growth required to exceed a total weight of 260 g. This weight is
representative of the size at first maturity among the males of both the Lake Erie and Lake Huron populations.
It is also similar to the values reported for other smallmouth bass populations (Carlander, 1977). The food
availability estimates used in the bioenergetics model were derived from the actual growth rates observed in the
Lake Erie and Lake Huron populations in recent years. These estimates indicate that fish in these populations
are currently feeding at levels well below satiation. We assumed that these food availability parameters were
not affected by climate change.
For each site, the combined growth recruitment model was run using current water temperatures and using
estimated water temperatures derived from GISS 2xCO, air temperatures. The population effects produced by
this change in climate were evaluated in terms of the following ratios: (average YCS GISS 2xCO,/ average
YCS current) and (average adult abundance GISS 2xCO2/ average adult abundance current). These ratios
were calculated using equations derived from the simple age-structured population model described below.
The model we use is constructed directly from the cohort and recruitment models described in Ricker
(1975). Assume all population members aged 1 and over experience a constant instantaneous natural mortality
rate equal to M and that all fish aged Tin and over experience a constant instantaneous fishing mortality rate
equal to F. Assume a relation between YCS at age 1 (Nj) and the number of breeding adults (S) which has
the following form:
2-24
-------
Magnuson
(1)
where S= £Nf and 6<1.
i=Tm
Preliminary analyses of the Lake Opebngo smalhnouth bass data base show that such an Nl-S relation does
hold for that population. At steady state, the following equations hold:
(2)
(3)
Now, by substituting equations 2 and 3 into 1, we obtain:
N1=a
r -M(im-iy
N-)G V '
\
1-e
-z
which simplifies to:
/ -M
N1=a
1-3
1e
-z
(4)
Now, assume that a in equation 4 is proportional to winter survival (w). Further, assume that Z, M and ft are
constant. Under these conditions, equation 4 requires that:
1
Therefore, any change in first year winter survival (w) or age of first maturity (Tm) can be evaluated in
terms of the resulting changes in both steady-state YCS (using equation 5) and S (by substituting 5 into 3 and
then substituting the resultant equation for N- into 2). Equation 5 shows that the only other parameters that
affect the results are M and ft. The level of fishing mortality is not important. All climate change scenarios
that we explore here are based on values for M and ft (03 and 0.65, respectively) derived from preliminary
analyses of the long term behavior of the Lake Opeongo smallmouth bass population.
2-25
-------
Magnuson
SCENARIOS
For both the Lake Erie (Ontario Ministry of Natural Resources, unpublished data) and Lake Huron (Shuter
et al., 1985) sites, littoral temperature data were available from several locations within the overall habitat of
resident populations of smallmouth bass. These data were used to construct annual temperature curves
representative of the average temperatures experienced by members of these populations over the last 10-20
years. A current temperature curve for the Lake Superior site was derived using the Baie du Dore water
temperature curve, Baie du Dore and Thunder Bay air temperatures, and the equations of Shuter et al. (1983)
to translate observed air temperature differences into water temperature differences. The estimated GISS
2xCO2 effect on each water temperature annual cycle was determined by first determining the GISS-predicted
change in annual average air temperature and then using that change with the Shuter et al. (1983) equations to
estimate the resultant change in the annual water temperature cycle.
RESULTS
The smallmouth bass populations in shallow protected bay habitats experience an increase in reproductive
success and fishable population size in cold, cool, and warm areas with a GISS 2xCO~ climate wanning (Figure
13). As expected, the relative advantage is greater in the cold areas. The imaginary Lake Superior population
probably is not viable prior to climate change but after climate change, it appears quite viable. The population
in the warm area near Long Point, Lake Erie is already doing well and would neither gain nor lose from climate
warming. The cool area near Baie du Dore, Lake Huron exhibits a positive response, with year class strength
increasing by a factor of 9 and the fishable population by a factor of 23, both of which are intermediate between
the response at the warm and cold sites.
The increases in reproductive success and population size occur in the model simulations because the
length of the growing season increases, overwinter survival of young-of-the-year smallmouth bass increases, and
the fish spawn at earlier ages (Figure 13). The growing season is simulated to increase with climate change by
a factor of approximately 13, for both young-of-the-year and adults. Overwinter survival of young of the year
fish increases from 10% to 88% in cold areas and nears 100% in cool and warm areas. Age of maturity
decreases by 1 to 3 years, so generation time decreases.
Our forecasts for smallmouth bass were uniformly positive. The reader must keep in mind that similar
detailed modeling for cool or coldwater species could well produce very negative forecasts. It was our intention
that the smallmouth bass work fulfill two primary functions:
(1) provide a credible forecast, given our present state of knowledge, of the likely effects of climate change
on a typical warmwater species currently resident in the Great Lakes drainage basin, and
(2) provide an example of the kinds of detailed models necessary to forecast the effects of climate change
at the population level. Our smallmouth bass forecasts are in no way representative of the response expected
from a "typical" species currently resident in the Great Lakes area.
2-26
-------
Magnuson
SMALLMOUTH BASS POPULATIONS
YEAR CLASS STRENGTH
FISHABLE POPULATION
COLO
AREA
COOL
AREA
WARM
AREA
COLD
AREA
COOL
AREA
WARM
AREA
300
YOUNG OF YEAR
GROWING SEASON
COLO
AREA
COOL
AREA
WARM
AREA
ADULT
GROWING SEASON
WARM
AREA
150
YOUNG OF YEAR
WINTER SURVIVAL
COLO COOL WARM
AREA AREA AREA
AGE AT MATURITY
COLD COOL WARM
AREA AREA AREA
Figure 13. Changes in smallmouth bass populations with GISS climate wanning in shallow protected littoral
areas of Lake Superior (cold), Lake Huron (cool), and Lake Erie (warm).
2-27
-------
Magnuson
CHAPTERS
ECOSYSTEM RESPONSE
Richard A. Vollenweider (personal communication) reminded us of works by F. Ruttner (1931), A.
Thienemann (1930), and other early German limnologists comparing the tropical inland waters of Java, Sumatra,
and Bali and the temperate lake systems of Northern Europe. Lakes which appeared similar in all aspects except
temperature differed with respect to relative trophic status, with the tropical lakes being more productive and
appearing to be more eutrophic than the temperate lakes.
The rate of photosynthetic production in well-lit surface waters is dependent on light intensity, temperature,
and abundance of nutrient substances and phytoplanktonic organisms. The rate of decomposition in poorly lit
deep waters depends on availability of oxygen, temperature, and abundance of decomposer organisms. All of
these processes involve chemical reactions.
According to van't HofFs Law, the rate of chemical reactions is doubled or tripled by a 10°C increase in
temperature. Physiologists have used this law, which they call the Q1Q rule, to compare the effects of
temperature on physiological rates (Hoar, 1975). Ruttner (1931) invoked an ecological Q10 rule to explain the
differences in limnological features between tropical and temperate lakes as noted above. Regier and Henderson
(1973) also referred to an ecological Q1Q concept with respect to fisheries production in waters of different
latitudes. Recent comparative fisheries-related studies of Pauly (1980) and Schlesinger and Regier (1982)
implicitly used the Q1Q rule. George Woodwell (1983) used an ecological Q1Q approach to derive some first
approximations as to the likely impacts of climate warming on terrestrial ecosystems.
The Arrhenius relationship has been used to describe the effect of temperature change on the "specific rate
constant," k, in chemical reactions. For many physiological and ecological applications a first-order chemical
reaction has often been taken as an analogue of some living process. Here
k
dt x " k-
where x is the mass or some surrogate measure of living substance. If a particular ecological process can be
approximated by a first-order expression, then the relevant Arrhenius expression is
k=Ae'E/RT,
where T is absolute temperature, and A^E, and R are constants. The Arrhenius expression is soundly based in
theory and in practice and, therefore, preferable to van't Hoffs empiric expression (Regier et aL, 1988). Regier
et al. (1988) provide a more detailed rationale for the use of the Arrhenius relationship.
The Arrhenius relationship provides a simple model from which first approximations of some potential
changes in lake productivity in response to climate warming can be derived.
METHODS
Data from the literature at the ecosystem level of organization were fined to the Arrhenius relationship by
regression methods, and community Q10 values were then estimated for phytoplankton production, zooplankton
biomass, and fishery yields. No data were located that treated an ecosystem as an integrated whole, but data
sets were found which related three major ecosystem components - algae, zooplankton, and fish - to
temperature. The Arrhenius expression summarized the information available in the form:
2-28
-------
Magnuson
=a(=?VbV+c
where: R is an ecosystem response variable, such as production;
T is absolute temperature,°K;
V is another environmental variaole, such as lake depth or fertility; and
a, b, and c are estimated regression coefficients.
Here the Q10 measures the proportional change in R with a temperature increase of 10°K.
The multiple regressions relating primary production, zooplankton biomass, and fish yields to temperature
(T) and Ryder's morphoedaphic index as V in the above equation are either taken from the literature or
computed by us from literature values (Table 2).
To estimate the relation for primary production and temperature in Table 2, primary production and MEI
data compiled by Oglesby (1977) and mean annual air temperature data from Wernstedt (1972) were used in
a stepwise multiple regression (SAS Version 6.02). Temperature explained 51% of the variation in primary
production and MEI explained only 7%, and the model appears to have a good enough fit to estimate Q1Q.
To estimate the relation for zooplankton biomass and temperature in Table 2, zooplankton biomass,
mean epilimnetic temperature, and MEI from Patalas (1975) were used in stepwise multiple regression.
Temperature explained 89% of the variation in zooplankton biomass while MEI only explained 3%. The overall
model accounted for 92% of the variation in zooplankton biomass among lakes.
The relation for maximum sustained fish yields in Table 2 is a reparameterization of the relationship
reported by Schlesinger and Regier (1982).
In addition to the three relationships in Table 2, we also estimated two for marine taxa, one on shrimp
with data from Turner (1977), and another for fish stocks with data from Pauly (1980). The temperature
coefficients, not shown here, were roughly comparable to those shown in Table 2 which gives us further
confidence in our estimates of community Q1Q values.
RESULTS
Community Q1Q values ranged from 2.6 for fishery yields to 3.7 for primary production, and 4.0 for
zooplankton biomass (Figure 14). When these different processes were fitted by the Arrhenius expression, they
provided similar estimates of the temperature coefficients. We do not understand enough about the relevant
theory to be able to explain this observation, but from a practical viewpoint we infer that the relevant community
rate processes respond, qualitatively and quantitatively, in ways similar to temperature. Temperature effects on
ecosystem processes or properties appear greatest on the primary producers and progressively less as one moves
up the food web.
2-29
-------
Table 2. Multiple Regressions Relating Community-Level Parameters to Temperature and R. A. Ryder's
Morphoedaptic Index, Using the Arrhenius Relationship, and the Resulting Community Q1Q
Values for Phytoplankton Production, Zooplankton Biomass and Freshwater Fish Maximum
Sustained Yields*
Phytoplankton Productionb
(1) logel°=44.40-113(V?(l/T)+ 0.2908 lo^MEI^
N=32; R2=0.575; r12=0.5lO; r2=0.065;
QlO=3'7
Zooplankton Biomassb
(2) logeZB=40.91-12074(l/EPTEMP)+0.4080
N=14;R2=0.918; r2=0.887; r2=0.031;
Qio=4.o
Freshwater Fish Maximum Sustained Yield0
(3) logeMSY=29.74-8115.4(l/T)+0.4895 tog^MEI^
N=43; R2=0.826; r2=0.737; r2=0.089
aSymbols and units: 1° * primary production (g C/mz.yr); ZB = Zooplankton biomass during
the summer in wet weight (mg/L); MSY = maximum aggregate sustainable yield (kg/ha.yr);
T = mean annual air temperature (°K); EPTEMP = mean epilimnion temperature (°K);
morphoedaphic index (Ryder, 1982) for maximum mean depth of 25m; Qin.3 proportional change
in ecosystem variable with a 10°K temperature change; N = sample size; K2 = coefficient
of determination; r1 , r2 = fractions of variability explained by the first and second
variables, respectively.
Derived in this study, see text.
°From Schlesinger and Regier (1982).
2-30
-------
Magnuson
Q
10
PHYTOPLANKTON ZOOPLANKTON
PRODUCTION BIOMASS
FISH
YIELD
Figure 14. Increases in primary production, zooplankton biomass, and maximum sustained fishery yields
expected from estimated community Q10 values.
2-31
-------
Magnuson
CHAPTER 6
INTERPRETATION OF RESULTS
GENERAL
The initial interpretation of our results suggests that fishes in the Laurentian Great Lakes will benefit from
global climate warming as determined from the OSU, GISS, and GFDL 2xCO2 climate scenarios. Based on
our model simulations, the size of thermal habitat within their optimum temperatures will generally increase in
basins with oxygenated hypolimnia (Figures 2, 3, and 4X their growth will increase provided prey consumption
increases with climate warming (Figures 9 and 10), population size of an example wannwater fish will increase
(Figure 13), and the biological productivity will increase at all levels from from primary production, to secondary
production, to fish yields (Figure 14). These results were obtained not only for wannwater fishes like smallmouth
and largemouth bass but also for coolwater fishes like yellow perch and walleye, and even for coldwater fishes
like trout and salmon. Nor were the increases restricted to the northern cold waters of Lake Superior; similar
increases appeared in cool and warm regions including neritic areas of Lakes Michigan, Huron, and Erie.
Our simulations also revealed some potential negative effects of climate warming. During summer, the
simulated thermal habitat size for a number of coldwater and coolwater fishes decreased (Figure 5) suggesting
that during summer, species interactions such as competition and predation could intensify in the optimum
thermal habitats. This restriction in summer habitat could be a limiting factor for some species as it is for
certain species in southern U.S. reservoirs (Coutant, 1987b). Simulations of the thermal habitat in brook trout
streams during summer showed that suitable temperatures decreased in a similar manner (Figure 7) and could
have similar consequences. Increases in fish growth (Figure 9) and its corollary, increased prey consumption
(Figure 10), would intensify species interactions and put an increased demand on prey production. In some of
the lakes, forage for the salmonid fishery is believed to be at a premium (Stewart et al., 1981), and the increase
in demand for forage could result in an overexploitation of prey resources by the salmonids. If additional prey
are not available, then the increased temperatures will result in a decline in growth (Figure 11). We have no
ecosystem level models for simulating the nature of such surprises.
There are other unknowns for which models or model simulations are unavailable at present. (1) The
increase in wannwater habitats may encourage colonization by wannwater exotics (Mandrak, 1988), which could
destabilize fish communities in warm areas (Magnuson, 1976). (2) The reduction of ice cover in bays could
reduce the reproductive success of lake whitefish (Taylor et al., in press; Freeberg, 1985). (3) If the global
climate warming reduced vertical mixing of Lakes Michigan, Huron, and Ontario, and the deeper basin of Lake
Erie for a long period of years, the oxygen in these deep lakes could become limiting to the detriment of the
coldwater fishes. Our simulations of increased growth assume that the fishes can behaviorally thennoregulate;
that is, when inshore or surface waters become too warm, the fish can find suitable thermal habitat offshore or
downslope or to the north. This means that the water in and below the thermocline and below must contain
sufficient oxygen to support fish life. Presently oxygen is depleted annually in the hypolimnia of the shallower
basins of Lake Erie, and these waters cannot support coldwater fishes. The increase in primary production
suggested in our simulations could exacerbate this process of developing anoxia in the deep coldwaters by
increasing the annual hypolimnetic oxygen deficit. (4) Our analyses have not included any effects on fishes
from loss of wetlands caused by water level changes. (5) Our analyses have not included any simulation of the
influence of climate wanning on the availability and accumulation of microcontaminants in Great Lakes fishes.
(6) Our analyses have not included effects of stream wanning on the reproductive success of sea lamprey. (7)
Our analyses have not included predictions about the smaller inland lakes, which are sure to be affected by
climate warming. Our general conclusion about the direct influence of temperature changes on fishes in the
waters of the Great Lakes, as induced by global climate change, is that it will be positive. A number of the
indirect effects deserve research level attention.
2-32
-------
Magnuson
SPECIFIC INTERPRETATIONS
Thermal Niche
Thermal habitats for all thermal groups of fishes (coldwater, coolwater, and wannwater) increase from the
BASE climate with the OSU, GISS, and GFDL 2xCO2 climate wanning scenarios when the entire year is
considered. In some cases, such as for coldwater fishes, the increase is dramatic, averaging a factor greater
than 2 or a doubling of thermal habitat. The increase results because warmer waters are found in spring, fall,
or winter and at deeper depths than for BASE climate. However, during summer, the isotherms are compressed
and the size of the thermal habitat (±2°C) in summer decreases under both scenarios by 2 to 47% depending
on the species. Inter- and intraspecific competition and predation may be intensified during these months.
Fishes may as a consequence occupy a broader range of temperatures than optimum (Rudstam and Magnuson,
1985) and grow less rapidly. The interaction of warmer epilimnetic temperatures coupled with increased hypoxia
in the hypolimnia of some Great Lakes basins and bays may lead to a "summer squeeze" on the available habitat
space for some species (Colby and Brooke, 1969; Coutant, 1985; Coutant, 1987a,b).
The increased surface water temperature in open water opens up or expands thermal habitat for wannwater
fishes in Lake Michigan and in Lake Erie. These conditions would provide more suitable habitat for white bass
and white perch, two species which are currently expanding their ranges in Lakes Erie and Ontario. Interest in
stocking Lake Michigan with striped bass, which has been introduced to various fresh waters in the southern
United States, may increase.
The expansion of optimal habitat of all species in the spring and fall in both climate scenarios suggests that
these "shoulder" seasons, if important to yield, may evolve as major contributors to fish yield and production if
climate warms. This appears to be the case with another coldwater fish, the burbot, in Lake Opeongo (Hackney,
1973). Of course, this depends on similar changes in other factors that control growth and yield, such as food
supply and interspecific competition.
Growth and Food Consumption
The bioenergetics simulations indicate that individual fish will grow faster under 2xCOU climate conditions,
if sufficient and appropriate prey are available. If prey consumption does not increase with climate warming,
fish growth will decline. If the forage base were to decrease or remain the same following climate warming,
predator growth rates would be expected to decrease because of an inability to compensate for the increased
metabolic costs of operating at higher temperatures. However, if prey availability increased so that fishes could
function at or near their optimal consumption rates, increases in growth rates would be even higher than those
reported above.
The Qj0 estimates of potential changes in zooplankton biomass and fish production indicate that at least
part of the forage base may increase with climate warming. This result provides evidence that the assumption
of a constant P-value may be reasonable. However, because of the uncertainties in estimating prey availability,
the effects of climate change on growth may be over- or underestimated. The dynamics of the Great Lakes
food webs (Figure 12) subjected to climate wanning must be considered in detail in order to answer the question
of whether primary and secondary production will increase enough to meet the increased predatory demands of
fishes.
Although the results for consumption are quite similar to those for growth, total consumption is an
important parameter in its own right because it is a measure of the trophic interaction among fishes and between
fishes and their prey. Increased consumption implies an amplification of competition for food if the forage base
does not grow as fast as the predatory demand on it. Consumption may also be thought of as a measure of
the potential predation pressure on the forage base. Predation pressure increases significantly for the 2xCO2
water temperature scenarios (Figure 10).
2-33
-------
Magnuson
Population Response
Smallmouth bass populations in shallow protected bays in warm areas of the Great Lakes would not be
harmed and would continue to provide recreational fishing. The fishing would improve in cool areas, and new
fisheries would develop in cold areas.
Other warmwater species and warmwater exotics might experience similar changes and opportunities from
range expansion.
Ecosystem Response
We infer from our statistical studies that the early observations by Ruttner (1931) and his colleagues were
generally reliable. The rates of ecosystemic processes respond to temperature changes in an exponential way
broadly consistent with the Q10 rule. Most expressions in Table 2 from which Q10 values were estimated also
include data from the Great Lakes; hence, these broad generalization do apply specifically to the likely effects
of climate warming on ecosystemic processes in the Great Lakes.
We expect that climate wanning would cause an overall increase in primary production (phytoplankton),
secondary production (zooplankton), and maximum sustained yields of fisheries across the Great Lakes (Figure
14). Lower trophic levels would experience a greater increase than higher trophic levels. The increased
ecosystemic productivity may be associated with shifts in community structure favoring organisms having higher
temperature preferences. The increased biological production will also be reflected in a long-term increase in
biogenic materials in the lake basins, which will contribute to increased oxygen demand in the hypolimnia. In
basins with small hypolimnia or with permanent thermal stratification, this increased oxygen demand would
contribute to hypoxia, with negative consequences for fishes.
2-34
-------
Magnuson
CHAPTER?
IMPLICATION OF RESULTS
The five Great Lakes are inhabited by cold, cool, and warm water fishes and provide a range of thermal
habitats characteristic of a large, deep, cold lake (Lake Superior) to a large, shallow, and warmer lake with
anoxic hypolimnia in some basins (Lake Erie). The lakes provide important recreational and food fisheries to
heavily populated regions of Canada and the United States. For most regions, the climate warming would
increase the amount of optimum thermal habitat not only for warm water fishes, but also for cool and cold
water fishes. Consequently, the productivity and potential yields from these thermal guilds is predicted to
increase in general if climate warming occurs. The productivity of zooplankton and phytoplankton is also
expected to increase. New warmwater fish species are expected to invade, especially into the littoral zones of
the more southern habitats. Even though productivity is expected to increase in general, surprises involving
water quality interactions and loss of important stocks are expected from invasion of new species as well as
from intensified species interactions in the fish communities.
2-35
-------
Magnuson
CHAPTERS
POLICY IMPLICATIONS
1. Increased fish production should provide greater recreational, commercial, and artisanal fishery opportunities.
2. Increased value of the Great Lakes in a warmer and dryer time should attract more people to the basin and
increase the consumptive use of the Great Lakes' water and living resources.
3. Increased predation pressure on the forage base should require management agencies to pay even closer
attention to the management of food webs and prey-predator interactions.
4. The potential for the further invasion of warmwater exotics and expansion of their thermal habitat should
require even greater attention to the effects of exotics on the stability and productivity of Great Lakes fisheries.
5. Increasing value of the fishery, should continue to make the reduction in toxins that bioaccumulate a high
priority for the Great Lakes.
6. Heightened eutrophication effects due to greater primary production and greater isolation of the hypolimnia
should require continued and increased attention to nutrient reductions in some lakes.
7. Water for the Great Lakes themselves should be more of an issue for fishery management agencies.
8. Coldwater fishes: thermal habitat for lake trout should increase more than for Pacific salmon; thus
rehabilitation of lake trout populations should remain an important policy issue.
9. Coolwater fishes: production of coolwater fisheries such as perch and walleye should increase and should be
expected to move northward and offshore.
10. Warmwater fisheries: production of warmwater fishes should increase in cool areas and develop in cold
areas. Managing the species structure of the inshore habitat should be an important policy issue.
2-36
-------
Magnuson
TERMS, ABBREVIATIONS, AND SYMBOLS
Climate Scenarios
BASE = Scenario using actual air temperatures from 1951-1980.
OSU = Scenario produced by the Oregon State University general circulation model.
GISS = Scenario produced by the Goddard Institute for Space Studies general circulation
model.
GFDL = Scenario produced by the Geophysical Fluid Dynamics Laboratory general
circulation model.
Areas
Cold = Northern Lake Superior
Cool - Southern Lake Michigan, Baie du Dore (Lake Huron)
Warm = Lake Erie
Fish Thermal Guilds and Temperature Preferences
Coldwater = Lake whitefish, coho salmon (-15 C), lake trout (-10 C)
Coolwater - Yellow perch, walleye (-24 C)
Warmwater = Largemouth bass, smalhnouth bass, bluegill (28 C)
Thermal Niche Model
Depth day = The depth of the water column within the thermal niche summed over days
Volume day * The volume of water within the thermal niche summed over days
THV = Summer thermal habitat volume in hm3.7 days or hm3.10 days (hm=hectometers)
THA = Summer thermal habitat lake bottom area in ha.7 days (ha=hectares)
SY = Sustained yield of fishery in kg/yr
Brook Trout Stream Model
FTw = Change in temperature of parcel of water
Tw a initial temperature of parcel of water
A = Surface area of parcel of water
V = Volume of parcel of water
v = Volume of inflow (e.g. groundwater, tributary)
tv = Temperature of inflow
S = Change in energy stored in parcel of water
Bioenergetics Model (all units g/dav. e.g.)
G = Fish growth rate
C = Rate of food consumption
F - Egestion
U = Excretion
P-value = Proportion of maximum consumption rate
Smallmputh Bass Model
M = Natural mortality rate
Tm = Age of first maturity
2-37
-------
Magnuson
TERMS, ABBREVIATIONS, AND SYMBOLS (continued)
F = Fishing mortality rate
YCS = Year class strength
Nl = Year class strength at year 1
NTm = Year class strength at age of first maturity
S = Number of breeding adults
a,B = Empirically determined constants in the equation relating N, and S
w = First year winter survival
O10 Model
Q10 = Proportional change in R with a 10°K temperature increase
R = Ecosystem response variable
V = Environmental variable
a,b,c = Estimated regression coefficients
MSY = Maximum Sustainable Yield
T = Mean annual air temperature
EPTEMP = Mean epilimnion temperature
MEI25 * Morphedaphic index for maximum mean depth of 25m
lo = Primary production in g C/m2/yr
COMMON AND SCIENTIFIC NAMES OF FISHES CITED IN TEXT
Common Name Scientific Name
Lake trout Salvelinus namaycush
Lake whitefish Coregonus clupeaformis
Coho salmon Oncorhyncus kisutch
Yellow perch Perca flavescens
Walleye Stizostedion vitreum vitreum
Northern pike Esox lucius
Largemouth bass Micropterus salmoides
Bluegill Lepomis macrocbirus
Smallmouth bass Micropterus dolomieui
Sea lamprey Petromyzon marinus
Cisco (lake herring) Coregonus artedii
Chubs various Coregonus sp.
Burbot Lota lota
Sculpins family Cottidae
Sauger Stizostedion canadense
Muskellunge Esox masquinongy
Lake sturgeon Acipenser fulvescens
Bullheads various Ictalurus sp.
Channel catfish Ictalurus punctatus
White bass Morone chrysops
Sunfish family Centrarchidae
Crappies family Centrarchidae
Minnows family Cyprinidae
Brook trout Salvelinus fontinalis
White perch Morone americana
Striped bass Morone saxatalis
2-38
-------
Magnuson
REFERENCES
Barton, D. R., W. D. Taylor, and R. M. Biette. 1985. Dimensions of riparian buffer strips required to maintain
trout habitat in southern Ontario streams. N. Am. J. Fish. Mgmt. 5: 364-378.
Blumberg, A.F. 1988. The effects of climate warming on Lake Erie water quality. Draft report to USEPA
Office of Policy Analysis.
Bowlby, J. N. and J. D. Roff. 1986. Trout biomass and habitat relationships in southern Ontario streams.
Trans. Amer. Fish. Soc. 115: 503-514.
Carlander, K. D. 1977. Handbook of Freshwater Biology, Volume 2. Iowa State University Press, Ames, IA.
Christie, G.C. and HA. Regier. 1988. Measures of optimal thermal habitat and their relationship to yields for
four commercial fish species. Can. J. Fish. Aquat. Sci. 45: 301-314.
Colby, P. J. and L. T. Brooke. 1969. Cisco (Coregonus artedii) mortalities in a southern Michigan lake, July
1968. Limnol. Oceanogr. 14:958-960.
Coutant, C. C. 1985. Striped bass, temperature, and dissolved oxygen: a speculative hypothesis for
environmental risk. Trans. Amer. Fish. Soc 114: 31-61.
Coutant, C. C. 1987a. Thermal preference: when does an asset become a liability? Environ. BioL fish.
18:161-171
Coutant, C. C. 1987b. Poor reproductive success of striped bass from a reservoir with reduced summer habitat.
Trans. Amer. Fish. Soc 116:154-160.
Crowder, L.B. and J. J. Magnuson. 1983. Cost benefit analysis of temperature and food resource use: a
synthesis with examples from fishes. Pp. 189-221 in W. P. Aspey and S. I. Lustick (eds.X Behavioral Energetics.
Ohio State University Press, Columbus.
Delay, W. H. and J. Seaders. 1966. Predicting temperatures in rivers and reservoirs. J. Sanit. Eng. Div., Proc
Am. Soc Civil Eng. 92: 115-134.
Freeberg, M. H. 1985. Early life history factors influencing lake whitefish (Coregonus clupeaformis) year class
strength in Grand Traverse Bay, Lake Michigan. MS Thesis, Michigan State University, East Tanking, MI. 89
pp.
Fry, F. E. J. 1971. The effect of environmental factors on the physiology of fish. Chapter 1 in W. S. Hoar and
D. J. Randall (eds). Fish Physiology Vol VI Environmental Relations and Behavior. Academic Press, NY. 559
pp.
Hackney, PA. 1973. Ecology of the burbot (Lota lota) with special reference to its role in the Lake Opeongo
fish community. Pn.D. Dissertation. Department of Zoology, University of Toronto. Toronto, Ontario. 209
pp.
Hawkes, HA. 1975. River zonation and classification. In River Ecology, ed. B. A. Whitton, pp. 312-374.
University of California Press, Berkeley and Los Angeles.
2-39
-------
Magnuson
Hewett, S. W. and B. L. Johnson. 1987. A generalized bioenergetics model of fish growth for microcomputers.
University of Wisconsin Sea Grant Institute Technical Report No. WIS-SG-87-245.
Hoar, W. S. 1975. General and Comparative Physiology. Prentice Hall, Inc. NJ. 848 pp.
Hokanson, K. E. F. 1977. Temperature requirements of some percids and adaptations to the seasonal
temperature cycle. J. Fish. Res. Board Can. 34:1524-1550.
Kitchell, J. F. and J. E. Breck. 1980. Bioenergetics model and foraging hypothesis for sea lamprey (Petromyzon
marinus). Can. J. Fish. Aquat. Sci. 37: 2159-2168.
Kitchell, J.F., D. J. Stewart, and D. Weininger. 1977. Applications of a bioenergetics model to yellow perch
(Perca flavescens) and walleye (Stizostedion vitreum vitreum). J. Fish Res. Board Can. 34: 1922-1935.
Lyons, J. and J. J. Magnuson. 1987. Effects of walleye predation on the population dynamics of small littoral
zone fishes in a northern Wisconsin lake. Trans. Amer. Fish Soc. 116: 29-39.
Magnuson, J. 1976. Managing with exotics a game of chance. Trans. Am. Fish. Soc. 105: 1-9.
Magnuson, J. J., L. B. Crowder, and PA. Medvick. 1979. Temperature as an ecological resource. Amer.
Zool. 19: 331-343.
Mandrak, N. E. 1988. Potential invasion of the Great Lakes by fish species associated with climate warming.
Env. Biol. Fish. In Press.
McCormick, M. J. 1988. Potential climatic changes to the Lake Michigan thermal structure. Draft report to
USEPA Office of Policy Analysis.
Meisner, J. D. J. S. Rosenfeld, and H. A. Regier. 1988. The role of groundwater on the impact of climate
warming on stream salmonines. Fisheries 13(3). In press.
Meisner, J.D., J. L. Goodier, H. A. Regier, B. J. Shuter, and W. J. Christie. 1987. An assessment of the effects
of climate warming on Great Lakes basin fishes. J. Great Lakes Res. 13: 340-352.
Neill, W. H. and J. J. Magnuson. 1974. Distributional ecology and behavioral thennoregulation of fishes in
relation to heated effluents from a power plant at Lake Monona, Wisconsin. Trans. Am. Fish. Soc 103:663-710.
Oglesby, R. T. 1977. Phytoplankton summer standing crop and annual productivity as functions of phosphorus
loading and various physical factors. J. Fish. Res. Bd. Can. 34: 2255-2270.
Patalas, K. 1975. The crustacean plankton communities of fourteen North American great lakes. Verb.
Internac. Verein. Limnol. 19: 504-511.
Pauly, D. 1980. On the interrelationships between natural mortality, growth parameters, and mean environmental
temperature in 175 fish stocks. J. Cons. Int. Explor. Mer. 39:175-192.
Regier, H. A. and H. F. Henderson. 1973. Towards a broad ecological model of fish communities and fisheries.
Trans. Am. Fish. Soc. 102: 56-71
Regier, H. A., J. A. Holmes and D. Pauly. 1988. The influence of temperature on ectothermic aquatic living
systems: some elementary and general comments. Submitted to Trans. Am. Fish. Soc.
Rice, J. A., J. E. Breck, S. M. Bartell and J. F. KitchelL 1983. Evaluating the constraints of temperature,
activity and consumption on growth of largemouth bass. Env. Biol. Fish. 9: 263-275.
2-40
-------
Magnuson
Rice, J. A., and P. A. Cochran. 1984. Independent evaluation of a bioenergetics model for largemouth bass.
Ecology. 65:732-739.
Ricker, W. 1934. An Ecological classification of certain Ontario streams. Univ. Toronto Studies Biol. Series,
No. 37, Publ. Ont. Fish. Res. Lab. 49: 7-114.
Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Fish. Res. Bd.
Can. 191: 382 pp.
Rudstam, L. E. and J. J. Magnuson. 1985. Predicting the vertical distribution of fish populations: analysis of
cisco, Coregonus artedii, and yellow perch, Perca flavescens. Can. J. Fish. Aquat. Sci. 42: 1178-1188.
Ruttner, F. 1931. Hydrographische und hydrochemische Beobachtungen auf Java, Sumatra und Vali. Arch.
Hydrobiol., Suppl. 8: 197-454.
Ryder, R. A. 1982. The morphoedaphic index use, abuse, and fundamental concepts. Trans. Am. Fish. Soc.
111:154-164.
Schlesinger, D. A. and H. A. Regier. 1982. Climatic and morphoedaphic indices of fish yields from natural
lakes. Trans. Am. Fish. Soc. Ill: 141-150.
Shuter, B. J., J. A. Maclean, F. E. J. Frey, and H. A. Regier. 1980. Stochastic simulation of temperature effects
on first year survival of smallmouth bass. Trans. Amer. Fish. Soc 109: 1-34.
Shuter, B. J., D. A. Schlesinger, and A. P. Zimmerman. 1983. Empirical predictors of annual surface water
temperature cycles in North American lakes. Can. J. Fish. Aquat. Sci. 40: 1838-1845.
Shuter, B. J., Wismer, D. A., Regier, H. A., and Maruszek, J. E. 1985. An application of ecological modelling:
impact of thermal effluent on a smallmouth bass population. Trans. Amer. Fish. Soc. 114: 631-635.
Stewart, D. J. and F. P. Binkowski. 1986. Dynamics of consumption and food conversion by Lake Michigan
alewives: an energetics modeling synthesis. Trans. Am. Fish. Soc. 115: 643-661.
Stewart, D. J., J. F. Kitchell and L. B. Crowder. 1981. Forage fishes and their salmonid predators in Lake
Michigan. Trans. Amer. Fish. Soc. 110:751-763.
Stewart, DJ., D. Weininger, D. V. Rottiers and T. A. Edsall. 1983. An energetics model for lake trout,
Salvelinus namaycush: application to the Lake Michigan population. Can. J. Fish. Aquat. Sci. 40(6): 681-698.
Taylor, W.N., M. A. Smale, and M. H. Freeberg, 1988. Biotic and abiotic determinants of lake whitefish
(Coregonus clupeaformis) recruitment in northeastern Lake Michigan. Can. J. Fish. Aquat. Sci. In Press.
Thienemann, A. 1930. Die deutsche limnologische Sunda-Expedition. "Deutsche Forschung". Ausder Arbeit
der Notgemeinschaft der Deutschen Wissenschaft.
Turner, R. E. 1977. Intertidal vegetation and commercial yields of penacid shrimp. Trans. Am. Fish. Soc. 106:
411-416.
Wernstedt, F. L. 1972. World climatic data. Climatic Data Press, Lemont, PA. 523 pp.
Woodwell, G. 1983. Paper presented at Climate Change Conference, Washington, D.C., Proceedings edited by
John Topping. Climate Research Inst., Washington, D.C. In Press.
2-41
-------
Magnuson
Wrenn, W.B. 1980. Effects of elevated temperature on growth and survival of smallmouth bass. Trans. Am.
Fish. Soc. 109: 617-625.
Wrenn, W. B. 1984. Smallmouth bass reproduction in elevated temperature regimes at the species native
southern limit. Trans. Am. Fish. Soc 113: 295-303.
2-42
-------
ECOLOGICAL EFFECTS OF GLOBAL CLIMATE CHANGE:
WETLAND RESOURCES OF SAN FRANCISCO BAY
by
Michael Josselyn
John Callaway
Romberg Tiburon Centers
Center For Environmental Studies
San Francisco State University
PO Box 855
Tiburon, CA 94920
Contract Number CR-814594-01-0
-------
CONTENTS
Page
FINDINGS 3-1
CHAPTER 1: INTRODUCTION 3-3
OBJECTIVE OF REPORT 3-3
GENERAL DESCRIPTION OF THE ECOLOGICAL SYSTEM 3-3
SENSITIVITY TO PREHISTORIC CLIMATE CHANGE 3-6
SENSITIVITY TO PRESENT-DAY VARIABILITY 3-8
STUDIES APPLICABLE TO THIS PROBLEM 3-8
CHAPTER 2: METHODOLOGY 3-10
MARSH AND MUDFLAT SEDIMENTATION 3-10
SALINITY INTRUSION 3-14
CHAPTER 3: RESULTS 3-16
SEA LEVEL RISE 3-16
FLOW MODIFICATION AND SALINITY INTRUSION 3-21
Marsh Plant Distribution 3-21
Rare and Endangered Species Associated With Wetlands 3-21
Phytoplankton Biomass 3-22
Benthic Invertebrates 3-22
Fish Abundance 3-22
SUMMARY OF RESULTS 3-23
CHAPTER 4: IMPLICATIONS OF RESULTS 3-27
ENVIRONMENTAL IMPLICATIONS 3-27
SOCIOECONOMIC IMPLICATIONS 3-27
CHAPTER 5: POLICY IMPLICATIONS 3-29
ADDITIONAL RESEARCH 3-29
REFERENCES 3-30
11
-------
Josselyn
FINDINGS1
This report examined the effects of global climate change on the deep-water and wetland habitats of the
San Francisco Bay estuary. Global wanning will result in increased sea level and modification of the freshwater
run-off into the estuary.
Three scenarios were used for the sea-level rise rate: a 100-, 200-, and 300-cm rise by the year 2100. The
relative rise at any one location in the estuary is a combination of the land surface movement and sea level
change. For the San Francisco Bay region, the lowest rate of sea level rise occurs around the Golden Gate,
with greater rates of relative sea level rise occurring in the northern and southern extremes of the estuary.
Thus, the effects of predicted sea level rise will be magnified by the regional subsidence.
Under the 100-cm rise scenario, sediment entering the bay via the rivers is sufficient to meet the
accretionary balance required by the tidal marshes and mudflats, at least until 2040. However, under the 200-
and 300-cm rise, the sediment requirement by tidal marshes and intertidal mudflats win exceed the 50% of the
annual estuarine sediment input early in the 21st century. If the area of tidal inundation increases owing to levee
failure, it is expected that most of the current 100,000 acres of intertidal wetlands (marshes and mudflats) will
be converted to subtidal habitats owing to a negative accretionary balance.
The diked and managed wetlands behind levees face a double jeopardy. If the levees are maintained, the
diked wetlands will continue to decline in elevation relative to sea level and will require substantial pumping
to remove standing water accumulated from land run-off and precipitation. Otherwise the standing water will
inundate the marsh vegetation. The cost of levee repair will also increase substantially owing to greater
hydrostatic pressure exerted by sea level rise. If the levees fail, over 130,000 acres of diked wetlands (seasonal
wetlands, fanned wetlands, and salt ponds) will be converted to subtidal habitats with resulting declines in
migratory waterbirds and extinction of many of the region's rare and endangered species. In some regions (such
as Suisun Marsh), additional wetlands may be created from higher sea level, but in the south and central bay,
intertidal wetland habitat will be completely eliminated owing to measures that will be necessary to protect
existing development in low-lying areas.
The hydrologic changes in the estuary owing to the combined effect of rising sea level and changing
climate will be significant. Most of the inflow will occur in the winter, and spring flows will decrease owing to
smaller snowpacks in the Sierra The greater volume of water in the estuary will cause salt intrusion into the
Delta To repel this salt from the intakes of the federal and state water projects, additional water (carriage
water) would have to be released from upstream reservoirs. Nevertheless, despite increased freshwater inflows,
the average salinity within the estuary wul increase.
All of the current regression models relating species abundance to flow are based on a positive correlation
between flow and abundance. High flows result in regions of high plankton biomass in the San Pablo and Suisun
Bays. This region is called the null zone as it is found where the net landward bottom current is zero. The
null zone also supports high numbers of benthic invertebrates and juvenile fish. However, under the scenarios
projected for sea-level rise, increasing flows will not result in the positioning of the null zone within Suisun Bay,
but rather in the western Delta During two recent drought periods, plankton and fish biomass reached historic
lows when the null zone occurred in the western Delta
The result of increasing salinization of the estuary will be a decline in brackish marsh plant productivity,
waterfowl habitat, and estuarine invertebrate populations. In some cases, marine species will invade areas
formerly occupied by estuarine species. However, it is anticipated that overall productivity of the estuary will
'Although the information in this report has been funded wholly or partly by the VS. Environmental
Protection Agency under Contract No. CR-814594-01-0, it does not necessarily reflect the Agency's views, and
no official endorsement should be inferred from it.
3-1
-------
Josselyn
decrease owing to decline in phytoplankton biomass. Perhaps more critical to many anadromous fish and
nesting waterfowl will be the increase in salinity in important spawning and breeding areas. These species
require nearly freshwater conditions, and the predicted salinity increase will significantly reduce their ability to
successfully reproduce.
Society currently places a high value on both tidal and wetland habitats, investing money to protect,
enhance, restore, and manage these areas. Their loss would have significant economic effects and would
also threaten the stability of developed areas behind levees. A number of recommendations are made to
protect tidal and diked wetland habitats including dredge spoil disposal in diked areas, high tide barriers, and
salt barriers. Each of these engineering solutions have significant environmental trade-offs and must be
evaluated now before their large-scale implementation.
3-2
-------
Josselyn
CHAPTER 1
INTRODUCTION
OBJECTIVE OF REPORT
The objective of this study is to investigate the response of wetland habitats to two attributes of global
climate change: general sea level rise and modification of freshwater inflow to the San Francisco Bay estuary.
Sea level change affects the distribution of various wetland types by conversion to deep water habitats
due to inundation. While some wetland types are tolerant of long periods of submergence, others such as tidal
marshes are severely impacted. The impacts of changing freshwater supply and sea level rise are related ~ a
reduction in sediment supply from the rivers will increase the rate of tidal wetlands loss.
Freshwater supply to the estuary affects the longitudinal distribution of various wetland types, from
freshwater wetlands in the Delta region to saltwater marshes in San Francisco Bay. In addition, freshwater
inflow drives the circulation pattern within the estuary and establishes the salinity gradient, which is important
in the distribution of the estuary's fish and wildlife populations.
We have used available topographic, photographic, and sedimentation data to assess the ability of Bay
marshes to sustain themselves in light of increasing sea level. Sea level rise scenarios were develop?
increases of 100, 200, and 300 cm by the year 2100. Sedimentation data and accumulation of sediment
marshes were calculated using a modification of the model given in Hatton et al. (1983). The quantity ot
sediment required as a proportion of the riverine borne material is then estimated for various water flow
scenarios.
To determine the effects of salinization, we compared the abundance or distribution of organisms as
related to various flow scenarios analyzed by Williams (1988).
Based on these analyses, we have addressed the potential long-term impacts of global climate change
on the wetlands of the San Francisco Bay estuary and the policy implications which must be considered to
protect these habitats.
GENERAL DESCRIPTION OF THE ECOLOGICAL SYSTEM
The general features of the San Francisco Bay estuary have been described by Conomos (1979). San
Francisco Bay is the largest estuary on the west coast of the United States. The Bay is actually a complex of
interconnected embayments, sloughs, marshes, channels, and rivers (Figure 1). It consists of two main
components, the northern reach containing San Pablo and Suisun Bays and the southern reach containing San
Francisco Bay proper. Ninety percent of the mean annual river inflow to the estuary enters the northern reach
through the Sacramento-San Joaquin Delta. Most of the flow occurs in the winter and spring months at which
time the estuary functions as a partially mixed system, whereas during summer and fall low-flow periods, the
estuary is well mixed.
The estuary has been highly modified by man's activities (Nichols et al., 1986). Water development, in
particular, has had tremendous impact on the distribution of plants and animals. Inflow to the estuary is highly
controlled by state and federal water projects. The Sacramento-San Joaquin Delta is a critical element in the
water management program: freshwater from the Sacramento River flowing through the Delta from the north
is diverted to federal and state pumping stations in the southern part of the Delta near Tracy. The integrity of
the Delta levees and channels is critical to the pumping efficiency and the maintenance of freshwater flows to
the estuary. Nevertheless, at times during low-flow periods, the net flow in the Delta is directed toward the
pumping stations instead of seaward through the estuary. Since the inception of the water projects in 1945,
3-3
-------
Josselyn
Suisun Marah
Figure 1. Generalized map of San Francisco Bay estuary with tide station locations (from San Francisco Bay
Conservation and Development Commission, 1987).
3-4
-------
Josselyn
natural freshwater inflow has been reduced by over 50%, which has resulted in greater salinization of the estuary
with concomitant reductions in habitat for anadromous fish and landward shifts in salt marsh vegetation and
benthic invertebrates.
In addition to changes in freshwater inflow, other man-induced changes have occurred in the estuarine
flora and fauna. Loss of wetlands owing to agriculture and urbanization has resulted in lowered populations of
many resident species and in reduced habitat for migratory species. Introduced species of invertebrates and fish
have displaced many of the native species. Non-native marsh plants have invaded the remaining wetlands.
Pollution has also affected organisms in the estuary. Domestic waste discharge and agricultural drainage
have increased the nitrogen and phosphorus content of the Bay's water. Petroleum by-products have affected
the reproductive capacity of striped bass. Heavy metals such as silver and mercury are found in high
concentrations in bay invertebrates. Despite all these impacts, however, the San Francisco Bay estuary is
relatively low compared to other U.S. estuaries in its susceptibility to pollution and, therefore, still supports
important fish and wildlife resources despite significant urbanization (Klein et ah, 1988).
The primary focus of this study is on the estuary's wetlands habitats. Atwater et al. (1979) reviewed the
history of the tidal marshes around the Bay and their current condition. Josselyn (1983) provided a substantial
review of the ecology of tidal wetlands and the San Francisco Bay and Development Commission et al. (1982)
of the diked wetlands.
The urbanization of the San Francisco Bay estuary has greatly modified the wetlands around its margins.
Recent mapping by the U.S. Fish and Wildlife Service indicates that almost 200,000 acres of various wetland
habitats (exclusive of open water) are present, compared to an estimated 540,000 (including the Delta) acres
prior to 1850. However, the mixture and type of wetlands present today are primarily the result of human
activities such as diking, agriculture, and land development. In most cases, tidal marshes have been replaced
by non-tidal wetlands such as salt ponds and seasonal freshwater and brackish marshes. In addition, declines
in freshwater outflow (estimated at between 30 and 70% of the natural annual inflow) have resulted in a shift
of wetland vegetation from fresh to saline tolerant species. Despite this large-scale conversion and the
substantial loss due to urbanization, wetlands continue to provide substantial beneficial uses for flood control,
shoreline anchoring, sediment trapping, nutrient retention, and fish and wildlife habitat.
The Bay's wetlands are vital to migratory birds, commercial and recreational fisheries, and a large
number of federal and state rare and endangered species. Located along the Pacific flyway, the estuary's
mudflats and deep water habitats provide important feeding and roosting areas for migratory waterfowl and
shorebirds. Suisun Marsh, the largest brackish water marsh in the western United States, is a major over-
wintering habitat for waterfowl and is specifically managed to produce food plants suitable for ducks and geese.
The estuary supports several anadromous fisheries including striped bass and salmon. Key spawning areas
include the Delta and central valley rivers and streams. Important to the survival of the juvenile fish is the high
planktonic biomass occurring in the north bay associated with the null zone or region where fresh and salt
water masses meet. As a result of urbanization, a number of plants and animals are now listed as rare and
endangered by the federal government. These species, such as the salt marsh harvest mouse, clapper rail, salt
marsh bird's beak, and least tern, are largely restricted to marsh habitats. Any further loss or modification of
these wetlands will surely reduce these species.
The primary types of wetlands present in the estuary are given in Table 1. The wetland acreages were
determined from digitized National Wetland Inventory Maps prepared by the U.S. Fish and Wildlife Service
from aerial photography taken in 1985 (Josselyn et al., in press). While over 200 different wetland types were
identified, they were grouped into eight major classifications for the purposes of this work: (1) open water
(deep and shallow water submerged habitat); (2) mudflat (unvegetated tidal areas); (3) tidal marsh (vegetated
tidal areas); (4) diked wetlands (non-tidal wetland areas [exclusive of fanned areas] that are generally within
the former boundary of historic tidal wetlands); (5) agricultural lands (former tidal wetlands now used for
fanning); (6) salt ponds (including crystallizers and bittern ponds); (7) lakes and ponds (shallow open water
areas); and (8) other wetland habitats (freshwater and riparian wetlands). Three geographic locations were
3-5
-------
Josselyn
considered: (1) Suisun Bay between the Delta (at Collinsville) and the Napa River; (2) North Bay between
Mare Island and the Oakland-San Francisco Bay Bridge; and (3) South Bay between Treasure Island and San
Jose. The area! extent was determined for those areas generally within the 10-ft. elevational contour around
the Bay (the Bay Zone).
For Suisun Bay, the largest component is the diked seasonal wetlands, which are managed for waterfowl
habitat. The vegetation within these wetlands is typical of brackish water wetlands: alkali bulrush, cattails,
brass-buttons, and pickleweed dominate. There is a substantial amount of tidal marsh as well, especially along
the southern shoreline of Suisun Bay.
Open water and mudflats are the most abundant wetland types in the north bay, especially along the
northern edge of San Pablo Bay. It is also the location of the highest acreage of tidal marsh, much of it
created by hydraulic mining debris in the 1860*5. Much of the diked historic marshlands are now farmed,
although some of the diked areas still function as wetlands. The tidal marshes in this sector are dominated
by saline tolerant vegetation: cordgrass and pickleweed.
Table 1. Areal Extent in Acres for Eight Generalized Wetland
Types Within the San Francisco Bay Estuary. Based
on U.S. Fish and Wildlife Service National Wetland
Inventory and Josselyn et al. (in press).
HABITATSUISUN BAYNORTH BAYSOUTH BAY
Open Water
Mudflats
Tidal Marsh
Seasonal Wetland
Farmed Wetlands
Salt Ponds
Lakes and Ponds
Other
Total
25,439
6,114
10,541
47,159
8,263
27
1,459
963
99,965
101,994
28,131
16,201
8,606
25,828
9,059
2,926
507
193,252
93,220
30,378
8,604
4,592
1,317
27.497
5,331
1,412
172,352
In the south bay, non-vegetated wetland habitats make up the majority of the acreage. Salt ponds in
which water depths are controlled provide habitat for a number of fish-eating birds, long-legged shorebirds such
as avocets and stilts, and migratory waterfowL Mudflats provide important habitat for short-legged shorebirds
such as sandpipers and willets. Vegetated wetlands and diked wetlands comprise less than 13,000 acres, the
smallest acreage of these important habitats for any of the regions in the bay. Pickleweed, a plant tolerant of
high soil salinities, is the most prevalent plant.
SENSITIVITY TO PREHISTORIC CLIMATE CHANGE
The evolution of tidal wetlands in San Francisco Bay has been the result of a gradual inundation of
lowland areas balanced by sedimentation. The melting of glaciers has been the major cause for a rise in sea
level in the past 10,000 years. There have been two phases in the rate of rise during this period (Figure 2).
From approximately 10,000 years before present (YBP) to 7,000 YBP, sea level rose at an estimated rate of 20
mm/yr (Arwater et aL, 1977). At this rate, dryland was rapidly inundated and converted to tidal flats and
3-6
-------
Josselyn
Mean
tea
level
-10
w
IL
IU
I-
W
o
UJ
-i
UJ
-20
-30
-40
-50
EXPLANATION
Probable range of error in age
and elevation of former sea level.
Arrow indicates former
sea level below some point
in rectangle
_L
_L
I
I
I
I
I
123456789
AGE. IN THOUSANDS OF YEARS BEFORE PRESENT
10
Figure 2. Changes in sea level relative to land during the past 10,000 years at the site of southern San
Francisco Bay (after Atwater et aL, 1977).
3-7
-------
Josselyn
subtidal areas. Dated cores from San Francisco Bay indicate little marsh development during this period. After
7,000 YBP, the rise in sea level slowed to the estimated current rate of 1-2 mm/yr. During this phase of sea
level rise, extensive marsh development occurred.
Marsh development along the margins of San Francisco Bay has largely been a function of inundation
of dryland by rising sea level. As wetland plants colonized the newly inundated areas, they also retreated from
the bay edge as it eroded. Sediments accreted on the marsh surface from this erosion as well as from the
suspended sediment derived from river drainage. In the late nineteenth century, increased sediment supply
derived from mining and agricultural practices in the watershed resulted in the migration of marshes over former
mudflats. With the reduction of these sources and declining freshwater inflow, retreating shorelines along the
bay margin are again more prevalent than accreting ones (Atwater et al., 1979).
Current scenarios for global climate change place the rate of sea level rise at between 10 to 40 mm/yr,
which is an order of magnitude greater than at present. The assumption is that tidal marshes would be rapidly
inundated. However, the acreage of tidal wetland present today is only 10% of that which existed just 150 years
ago and, therefore, one might anticipate that sufficient sediment supply is available to support these wetlands.
SENSITIVITY TO PRESENT-DAY VARIABILITY
Two analogues are available to assess the response of present-day wetlands to variability in water flow
and sea level rise. The first is the severe drought of 1976-77, which drastically reduced water supply to the
estuary, and the second is the relative change in sea level experienced by areas that have subsided due to ground
water withdrawal
During the 1976-77 drought period, water supply to the estuary was reduced by 75% over normal
regulated inflow. A number of changes were observed in the vegetative composition of both tidal and non-tidal
(managed) wetlands. A substantial decrease in the productivity of brackish water species such as alkali bulrush
and tules were observed with a concomitant increase in the occurrence of saline tolerant species such as
cordgrass and pickleweed (Atwater et al., 1979; California Department of Fish and Game, 1987). Other
estuarine organisms were also affected. Phytoplankton productivity decreased within Suisun Bay (Ball and
Arthur, 1979), striped bass juveniles drastically decreased, and more saline-tolerant benthic organisms migrated
upstream. These changes occurred within the 2-year period of the drought, and the marine species that invaded
the upper reaches of the estuary were eliminated once flows returned to previous levels. However, the recovery
of some estuarine species has been much slower. For example, the striped bass population remains critically low
to this date.
The second analogue, groundwater withdrawal, has contributed to local subsidence in south San Francisco
Bay of up to 1.5 m within the recent past (Atwater et al., 1977). Many diked wetlands formerly at or above
MHHW are now at MLLW or below. Continued repair is required to maintain the levees around these areas
and pumps are necessary to remove accumulated precipitation. On the other hand, tidal marshes do not appear
to have been lost owing to inundation. Patrick and DeLaune (unpublished manuscript, Louisiana State
University) used isotopic data to determine sedimentation rates in a tidal marsh near Alviso Sedimentation over
a 30-year period averaged 56 mm/yr. At a tidal marsh near Foster City where subsidence is considerably less,
sedimentation in a tidal marsh was less than 3 mm/yr during the same time period. Thus, it appears that in
some locales, sedimentation can be sufficient to maintain tidal marshes despite large rises in relative sea level.
STUDIES APPLICABLE TO THIS PROBLEM
Most studies provide only a glimpse of the impacts of long-term climatic change on estuarine wetlands
and their associated flora and fauna. In general, the focus has been placed on sedimentation rates in wetlands.
Stevenson et aL (1986) provide an excellent review of sedimentation studies to date and found a strong positive
correlation between tidal range and sedimentation in wetlands experiencing annual sea level increases between
3-8
-------
Josselyn
2 to 10 mm/yr. Regions that have a very small tidal range, such as the Mississippi Delta, experience the greatest
accretionary deficit as sea level has risen whereas regions with higher tidal ranges have been able to sustain
elevations suitable for wetland vegetation.
Regional studies on the impacts of sea-level rise on the wetlands of Delaware and Chesapeake Bays
and the Mississippi Delta have also been completed (Hull and Titus, 1986; Stevenson et al., 1985; Salinas,
1986). These studies emphasize the complexity of the factors which contribute to wetland loss in these systems.
Salinity intrusion and sediment accretion deficits, however, are the most significant impacts resulting in either
modification or loss of wetlands.
The literature for San Francisco Bay is much more limited than that in other areas. Approximations
for sediment input to the estuary have been derived (Krone, 1979). The amount of sediment entering the
estuary is directly related to total river discharge. In the 1960's, the estimated annual sediment supplied to the
Bay via the Delta was 3 million metric tons/yr. By 1990, with projected water development, sediment supply
is estimated at only 1.6 million metric tons/yr, a 45 percent decrease.
Less is known about the distribution of sediment once it enters the estuary. Only a few studies have
been completed on sedimentation on mudflats or within marshes (Nolan and Fuller, 1986; Pestrong, 1972;
Josselyn and Buchholz, 1984). Much of this data is site specific and difficult to generalize to other areas.
Mudflats generally exhibit the highest rates of sedimentation, especially in regions where the tidal prism (volume
of water between high and low tide) has been reduced owing to filling of wetlands. Accretion rates of 0.6 to
2.8 mm/yr were observed in a shallow embayment, whereas rates of 70 mm/yr occurred in subtidal waters
(Nolan and Fuller, 1986).
Within marshes, seasonal variation in elevation may exceed long-term accretion (Josselyn and Buchholz,
1984). Generally, lower elevations within tidal marshes exhibit the greatest rates of sedimentation, which then
decrease rapidly with height above sea level (Krone, unpublished). Josselyn and Buchholz (1984) measured
annual fluctuation in sedimentation within tidal marsh channels ranging between 100 and 300 mm/yr, whereas
Martindale (1987) determined accretion rates of 1-5 mm/yr on the marsh surface.
Recently, a number of studies have been completed by the San Francisco Bay Conservation and
Development Commission on the impact of sea level rise on the shoreline of the estuary and the reduction of
freshwater flow (San Francisco Bay Conservation and Development Commission, 1985,1987a,b,c). These reports
provide useful recommendations concerning the land-use policy implications of global climate change.
3-9
-------
Josselyn
CHAPTER 2
METHODOLOGY
To date, large-scale biological data bases for the estuary have not been integrated into a model suitable
to study biological responses to changing water flow or sea level Therefore, the models used for this study are
based largely on empirical data, simple sedimentation budgets, and regression models.
MARSH AND MUDFLAT SEDIMENTATION
Scenarios used for sea level rise were taken from Titus (personal communication). Three scenarios were
used: 100-, 200-, and 300-cm rise by the year 2100. The equations used for calculation were:
100 cm scenario: RSL(t) • L(t) + 0.0012t + 0.000066422
200 cm scenario: RSL(t) = L(t) + 0.0012t + 0.00014342
300 cm scenario: RSL(t) - L(t) + 0.0012t + 0.00022032
where: t » the number of years since 1986
L(t) * the local component of relative sea level change
relative sea level at time t.
L(t) is further refined as LYLE*t - 0.0012t where LYLE is the local relative sea level trend. The value for
LYLE for the San Francisco region was obtained from Lyle et al. (1987).
The changes in sea level as predicted under these three scenarios and corrected for the San Francisco
Bay region are shown in Figure 3. This model assumes an increasing rate of rise throughout the next century.
However, many other models may be equally valid such as a linear rise, a stepped rise, and a precipitous rise.
Within San Francisco Bay, there is considerable variation in local sea level trend owing to differences in
crustal subsidence. San Francisco Bay Conservation and Development Commission (1987b) provides local land
surface changes for 11 locations throughout the Bay (Figure 4). These data vary considerably, ranging from
0.06 mm/yr in Sausalito near the Golden Gate to 29.23 mm/yr in the south bay at Alviso. To illustrate the
range of sea level change that might be expected to occur throughout the estuary, the 200-cm scenario for San
Francisco, Sausalito, Pittsburg, and Alviso is plotted in Figure 5. Thus, the actual rise within the estuary for
any one scenario may vary significantly depending upon the location. Recently, actions taken to replenish the
groundwater aquifers in south San Francisco Bay have halted further subsidence (Santa Clara Water Agency,
pers. comm.) and, therefore, the range of potential sea level rise is more appropriately restricted to that shown
for Sausalito and Pittsburg on Figure 5.
Sediment budgets for marsh and mudflat surfaces were calculated for each of the three scenarios using
the model equation from Stevenson et al. (1986):
Sediment necessary to maintain marsh surface = BD * PI * AR * A
where BD = Bulk density of soil
PI = Percent of inorganic matter
AR * Accretion rate
A » Area of tidal marsh or mudflats
3-10
-------
Josselyn
300
u
I
o
V
in
0
y
o
0
PREDICTED SEA LEVEL RISE
at San Francisco for 3 scenarios
1990 2000 2010
100 cm case
210
Year
200 cm cave
Figure 3. Predicted sea level rise at San Francisco for three scenarios: 100-, 200-, and 300-cm rise by the
year 2100.
3-11
-------
Josselyn
10
CURRENT RELATIVE SEA LEVEL RISE RATES
Son Francisco Boy Eatuory
29.2
E
"5
o
A
_
tr.
9 -
6 -
5 -
4 -
3 -
2 -
1 -
Pitts Ben Son Pt O Sous Pres
Suisun Bay I San Pablo & Central Bay
T
SF
I
Ala
H Pt
S.F. Bay
I
SM
Dum B Alv
(South S.F.
Bay
Figure 4. Current relative sea level rise rates for twelve sites in San Francisco Bay estuary. Abbreviations
refer to tide station locations mapped on Figure 1 (after San Francisco Bay Conservation and
Development Commission, 1987).
3-12
-------
Josselyn
PREDICTED SEA LEVEL RISE
200 cm Scenario
u
i
o
W
y
Q
9
600
500 -
400 -1
300 -
1990 2000 2010 2020 2030 2040 2050 2060 2070 2080 2090 2100
200 -
100 -
Year
Alviso
Pitt
SF
Sauiolito
Figure 5. The range of predicted sea level rise in San Francisco Bay estuary by the year 2100, including Alviso
Slough, Pittsburg, San Francisco, andSausalito.
3-13
-------
Josselyn
The data for bulk density and percent inorganic material were taken from Pestrong (1972), Mahall and Park
(1976), Perez (1981), Nolan and Fuller (1986), and Josselyn (unpublished) for a variety of geographic and
wetland types in the estuary. High, median, and low estimates were determined (Table 2).
Table 2. Ranges of Bulk Density and Inorganic Content Used
in Determining Sediment Budgets for Tidal Marshes
and Mudflats.
WETLAND TYPE FACTOR HIGH MEDIAN LOW
Tidal marshes
Mudflats
Bulk Density
Percent inorganic
Bulk Density
Percent inorganic
1.0
92
1.0
98
0.8
85
0.8
96.5
0.65 g/cm3
78
0.65
95
The sediment quantity required to maintain marsh and mudflat surfaces was then calculated for various
accretion rate scenarios within each region of the estuary. Finally, a comparison of the sediment required
under each of the scenarios against the sediment available was determined. The annual sediment supplied to
the Bay was determined from the relationship of flow to sediment supply as given in Krone (1979). Sediment
supply was estimated under four inflow scenarios: BASE, GFDL, GISS, OSU outputs from Williams (1988). An
important assumption in Williams's estimate of inflow was that sufficient water would be released by state and
federal water projects to retard salt intrusion to water project intakes within the Delta. This additional water
is referred to as carriage water. In addition, we assumed that levees around diked wetlands would be
maintained. If not, these areas would also serve as sinks for sediment, increasing the percent necessary for
marsh accretion.
In addition to sediment budget calculations, marsh erosion was also examined in locations with known
historic changes in relative sea level. We initially examined 1:24000 U.S.GS topographic maps from 1945 to
present to determine if any areas showed accretion or erosion. The scale of these maps prohibited accurate
measurements of shoreline change. Therefore, four locations in the estuary were selected for more detailed
analysis of shoreline change: Palo Alto and Alameda Creek Flood Control Channel in the south Bay and
Corte Madera and Southampton Bay in the north Bay. Aerial photographs from the 1950*8, mid-1960's, and
present were enlarged to a scale of 1"=400' and the shorelines overlain. Levees and permanent structures were
used for horizontal control. Erosion rates were compared to relative sea level changes for nearby stations.
SALINITY INTRUSION
Changes in water flow and average salinity for the estuary owing to climate change were taken from the
studies completed by Williams (1988). He examined the effect of a 100-cm sea level change within several
climate model scenarios. With the projected sea level rise, salt water penetrated further landward towards the
Delta. In order to maintain water quality standards for water exported from the Delta, greater quantities of
freshwater discharge through the Delta would be needed to repel the saltwater intrusion. The estimated
"carriage water" required to protect existing water intakes was determined to be twice current levels of discharge.
Despite the increased inflow to the estuary, the larger volume of seawater in the Bay resulted in higher average
salinities in all of the basins.
3-14
-------
Josselyn
Williams also examined the impact of levee failure within the Delta and along the margin of San
Francisco Bay. This scenario resulted in a tripling of the existing area of the Bay and a doubling of its volume.
The greater volume resulted in stronger tidal currents at restrictions in the estuary (e.g., Golden Gate and
Carquinez Straits) and reduced tidal ranges further inland. However, salinity did not increase significantly
over the "intact levee" scenario.
A number of linear relationships have been observed between flow and the abundance of certain
organisms; however, these regressions are dependent upon the current geomorphology of the estuary. If the
Delta islands were flooded, the inland sea created in this region would so alter the geomorphology of the
estuary that current regressions would no longer be valid. For example, studies on juvenile salmon migration
have shown that higher flows are necessary to assure high survival of smolt by moving them quickly through
the Delta region to avoid entrainment losses (Stevens and Miller, 1983). Under the scenario where carriage
water is doubled, the regression model shows a higher survival by salmon smolt. However, the hydrology and
circulation in the Delta would be entirely different from the conditions on which the regression was based.
Second, most of the biological relationships observed in San Francisco Bay are based on the indirect
relationship between flow and salinity. However, under conditions where sea level is rising, increased freshwater
inflow does not reduce salt intrusion into the estuary. Thus, existing models of organism response to changing
inflow are not applicable and any analysis of the changes which might occur within the estuary under various
climate scenarios must necessarily be qualitative.
3-15
-------
Josselyn
CHAPTERS
RESULTS
SEA LEVEL RISE
The rates of sea level rise for 5-year periods under each of the three scenarios is shown in Figure 6.
For the purposes of this report, the modei ^as ...ejected for a 50-year period only. The model predicts sea level
rat" .creases throughout the next century. For the 100-cm rise, rates will range from less than 2 mm/yr in
1990 to over 8 mm/yr hi 2040. Given the range of accretion rates observed for San Francisco Bay and other
wetland systems with similar tidal ranges (Stevenson et aL, 1986), it is probable that sediment accretion on the
marsh surface could maintain the vegetated wetland for this scenario.
On the other hand, for scenarios of 200- and 300-cm rise by the year 2100, the accretion rates necessary
to maintain the marsh surface relative to sea level rise are considerably higher than the average rates presently
observed. For the 200-cm scenario, sea level increase would exceed average rates for marsh surface accretion
by the year 2005 and, for the 300-cm scenario, by the year 2000. Under the 300-cm scenario, the rate of rise
by 2030 will have exceeded that of recent prehistoric rates when it is presumed that tidal marshes were not very
prevalent owing to rapid inundation (Atwater, 1979).
The amount of sediment necessary to sustain tidal marshes under various accretion rates is shown in
Figure 7. The low, mid, and high estimates are based on the range of data for bulk density and inorganic
content given in Table 2. An 11 mm/yr rise would require approximately 1.1 million metric tons (range of 0.8
to 1.5 me : tons) of sediment to be deposited on all tidal marsh surfaces. These estimates are based on
present-day tidal marsh acreage. As sea level rises, additional acreage would be flooded. If the levees remained
intact, the increase would be negligible. On the other hand, if the levees failed around diked baylands, the
potential area over which sediments could be deposited increases by 3 (Williams, 1988). In addition, many of
the diked areas are below sea level (compared to tidal marsh surfaces which are generally near MHW) and
will be a substantial sink for suspended sediments. Thus, under the assumption of levee failure, the amount of
sediment deposited along the margins of the estuary could well exceed 5 million metric tons for an 11 mm/yr
rise in sea level
These data can also be expressed as a percentage of total sediment input from the Sacramento River
under various climate model scenarios (Figure 8). The amount of suspended sediments supplied to the estuary
were estimated from the regression of flow against suspended sediments given by Krone (1979), multiplied by
the flow rates projected by Williams for each of the three climate models. For example, for an 11 mm/yr sea
level rise, 20% of the total annual sediment input under the GFDL inflow scenario would be necessary to
maintain tidal marsh surfaces relative to sea level rise. However, this is less than the 31% required under base
flow conditions because of the greater carriage water flows necessary to repel salt water intrusion.
Under present-day conditions, only 5% of the total annual sediment input is necessary to maintain tidal
marshes at equilibrium with sea level rise. Under the lowest projected sea level rise scenario, only approximately
20% of the total sediment input would be needed to sustain marshes (i.e., for the 8 mm/yr rate of rise hi year
2040). However, for the highest projected sea level rise scenario, between 35 and 50% of the total sediment
input would be needed depending upon the climate scenario used. Again, the amount of additional wetland
acreage formed when levees fail will add to the sediment budget requirements.
The preceding discussion did not consider accretion over mudflats. When mudflats do not accrete as
rapidly as increases in sea level, erosion of the marsh edge is often the result. Shoreline erosion is a presently
a common phenomenon throughout San Francisco Bay (Josselyn, 1983), and we examined the relationship
between shoreline erosion and relative sea level change. In all sites examined using aerial photography,
significant shoreline erosion had occurred between the early 1950's and the present (Figure 9). Except at Palo
Alto, marsh edge erosion rates were generally between 0.6 to 0.9 m/yr since the mid-1950's. Palo Alto exhibited
3-16
-------
?
O
o
<8
RELATIVE SEA LEVEL RISE RATES
at San Francisco
2O.UU —
24.00 -
22.00 -
20.00 -
16.00 -
16.00 -
14.00 -
12.00 -
10.00 -
B.OO -
6.00 -i
4.00 -
2.00 -
r> r\n —
p;
'1 XXX4 L' X
A^A v\
0 r
'i \
0 F? N
i \^ rN
y, p\^ /\
v* r s
4 K^4 /^
I
/y
r;
X
s
X
X\
x\
> s.
^ \
x\
1
^
7
/
s
s
/
x
\
\
\
\
X
\_
\
\
I
-£
7-
/
/
/
/
/
/
T
'x
\
\
\
\
\
\
\
\
X
\
7-
./
y
t
y
y
y
y,
y
y
y,
y
^/
•^
7:
7
/
/
/
/
s
/
*\
\
\
\
\
\
\
\
\
\
\
>
\
7
y
y
y
/
i
V
^
7
X
X
/
/
/
f
/
— -n
\
\
\
\
\
\
\
\
\
N
\
x
\
\
7
y
y
y
y
y
f
y
y
y
y
/
y
^
y
y
y
y
^/
/;
o-
/
/
/
/
/
;
s
/
x;
\
\
\
\
\
\
\
\
N
\
\
\
\
\
\
??
-------
Josselyn
SEDIMENT REQUIRED BY TIDAL MARSHES
In metric tone
in
O
x
t-
•o
O
o
a
•o
0
VI
0.2
Accretion Rates (Sea Level Rise) CM/YR
Low PCSI Mid E5Z1 High Estimate
Figure 7. Sediment required by tidal marshes in the San Francisco Bay estuary in metric tons under various
hypothetical accretion rates as calculated from Table 2.
3-18
-------
Josselyn
100
SEDIMENT REQUIRED BY TIDAL MARSHES
Aa a p«rc«ntag« of total ••dlm«nt Input
a'
£
|
E
&
o
+t
c
\
o
a.
0.2
GFDL
0.5
0.8
AccPBtlon Rates (Sea Level Rise) CM/TR
' GISS C/%>3 OREGON
BASE
Figure 8. Sediment required by tidal marshes in the San Francisco Bay estuary as a percentage of total
sediment input under various rates of predicted sediment input. The percentage of sediment input
is indicated by the total bar height for each model scenario. For example, at the 1.1 cm/year rise,
the Oregon model predicts 31 percent (the sum of GFDL, GISS, and Oregon bars).
3-19
-------
Josselyn
Alomedo Creek
BAY
LAND
Palo Alto Baylonds
BAY
1958
1966
1985
South Hampton Bay
BAY
LAND
—1957
—1966
1987
Corte Modera Marsh
BAY
LAND
Figure 9. Changes in land margins at selected locations in San Francisco Bay.
3-20
-------
Josselyn
a rapid rate between 1958-66, but it slowed considerably after this period, partially owing to accretion along
the marsh channel. No relationship was observed between relative sea level change and rate of erosion. Thus,
given the scale of analysis, a general rate of erosion of 0.75 m is assumed under the most recent past when sea
level rise in the Bay area has generally ranged between 2 and 5 mm/yr.
Under the 100-cm scenario, shoreline erosion would probably continue at the present rate up to 2040.
However, both the 200- and 300-cm rise scenarios double the probable erosion rate by 2030 and 2010,
respectively. This would have serious .consequences for the narrow fringing tidal marshes that surround much
of San Francisco Bay, especially as many are outboard of levees that do not allow for marsh advance landward
FLOW MODIFICATION AND SALINITY INTRUSION
As explained above, it is much more difficult to predict the the changes in estuarine communities resulting
from modifications of flow or salinity. The detailed modeling does not exist and one must rely on regression
analyses. The results discussed below focus on marsh plant distribution, rare and endangered species associated
with marshes, phytoplankton biomass, invertebrate distribution, and fish abundance.
Marsh Plant Distribution
The longitudinal distribution of wetland plants within San Francisco Bay is directly related to the salinity
gradient. Suisun Bay is considered the transitional zone between the freshwater wetlands of the Delta and the
salt marshes of San Pablo and San Francisco Bays. Spring salinities are most crucial to the development of
wetland vegetation. Using the 1976-77 drought analogue model for wetland plant distribution, the spring
salinities predicted for Suisun Bay with the OSU hydrology and levee failure would reduce the productivity of
brackish and freshwater species such as cattails, tules, and alkali bulrush and favor more salt tolerant species
such as pickleweed and saltgrass. Salinities predicted under the GFDL and GISS models are within the
acceptable range for brackish water vegetation.
The submergence tolerance of many brackish water species decreases as salinity increases (Atwater et
al., 1979). Thus, they cannot extend as far down the intertidal. The net effect is a restriction in the intertidal
range of species such as hardstem tules, Scirpus spp., the major shoreline plant in Suisun Bay. This shift will
reduce available habitat, primary production, and shoreline stabilization.
The impact of the changes predicted under the OSU model would be a reduction in waterfowl use of
Suisun Marsh. Approximately 3 to 5% of the state's migratory waterfowl utilize Suisun Marsh during the winter
months, feeding primarily on fresh and brackish water vegetation. If the levees around the managed wetlands
of Suisun Marsh failed, the existing elevations would not support emergent vegetation and, thus, waterfowl would
have to seek wintering grounds in the Central Valley. In addition, waterfowl nesting in Suisun Marsh would
decline owing to higher summer salinities projected under all scenarios.
Rare and Endangered Species Associated With Wetlands
A number of federal and state rare and endangered species are associated exclusively with the salt and
brackish wetlands of San Francisco Bay. These species include the salt marsh harvest mouse, the California
clapper rail, the least tern, Suisun shrew, and the black rail. In addition, other species which may be considered
candidates for listing such as the salt marsh song sparrow and the yellowthroat are found within vegetated
wetlands.
These species are generally restricted to small home ranges within vegetated wetlands. Other migratory
species such as the least tern and snowy plover nest in low-lying areas adjacent to wetlands. Loss of vegetated
wetlands owing to either gradual inundation by sea level rise or to catastrophic flooding following levee failure
3-21
-------
Josselyn
will cause local extinction of these species. As distances between the remaining populations increase, the chances
for long-term survival of the species are greatly reduced.
Phvtoplankton Biomass
There are two peaks of phytoplankton biomass in the San Francisco Bay estuary, a early spring peak in
San Pablo Bay and a summer maximum in Suisun Bay. Generally, the peaks are associated with a circulation
phenomena called the null zone: the region where net bottom flow is zero. The null zone has longer residence
time than other areas of the estuary and, therefore, phytoplankton biomass accumulates. Flows necessary to
sustain these peaks are approximately 20,000 cfs during April and May for San Pablo Bay and at least 15,000
cfs during June and July for Suisun Bay (Williams and Hollibaugh, 1987a,b). These flows generally result in
average water column salinities between 2 and 6 ppt in these basins during the period of peak biomass.
The projected salinities under the OSU hydrology indicate that the null zone and, hence, the
phytoplankton biomass maxima would be reduced substantially in the San Pablo and Suisun Bays. Spring and
summer peaks would, instead, shift upstream into the Delta. However, under GFDL and GISS conditions, it
is possible that biomass maxima could still occur depending upon whether or not the levees are maintained.
Levee failure would introduce vast expanses of shallow water and mudflats over which primary production may
be enhanced.
During the 1976-77 drought and the 1987-88 dry period, phytoplankton biomass in Suisun Bay was near
zero during the summer months. Zooplankton, especially a mysid shrimp, Neomysis mercedis and a copepod,
Eurytemora sp., were also low. These two zooplankters are the major food source for juvenile striped bass,
Morone saxatilis. As a result, survival of juvenile striped bass during both dry periods was at historic lows.
Based on these two analogues, we conclude that positioning of the null zone landward of Suisun Bay will result
in reduced primary and secondary productivity with a direct effect on an economically important fishery.
Benthic Invertebrates
Many of the invertebrates now dominant in San Francisco Bay are introduced, non-native species. They
are considered estuarine, opportunists in that they are tolerant of relatively low salinities and do not occur at
all in the marine environment outside the Bay.
One explanation given for a decline in phytoplankton biomass during drought years is the invasion of
benthic marine filter feeders into San Pablo and Suisun Bays (Thompson and Nichols, 1987). These species
are very effective in removing phytoplankton, and it was estimated that they are capable, in the densities
observed, of filtering the entire volume of the estuary in a single day.
In addition to increasing abundance of marine infauna, epifauna such as dungeness crab also increased
during the periods of higher bay salinities.
Fish Abundance
Two sets of fish data exist: commercial catches of species such as salmon, striped bass, and American
shad and shorter-term data on fish populations taken under the State Board's Bay-Delta program. Most of
the commercial data is for the pre-water project development period. Only salmon have a long-term record
comparable to the base period selected for this study (1950-81). Rozengurt et al., (1988) examined a number
of flow/catch relationships and determined that the best correlations occurred with lag periods of 4 to 5 years.
That is, the flows occurring in the previous years had the best relationship to the catch in any given year. The
lag time is related to the age at which the fish reached spawning maturity. Two significant direct relationships
were observed: the preceding year(s) annual flow with total fish catch and preceding year(s) spring flow (Apr,
3-22
-------
Josselyn
May, June) with total fish catch. Though the specific regression equations differed among species, the
relationships were similar for salmon, striped bass, and shad
When examining the flow scenarios produced by the hydrologic models, all scenarios provided in Williams
(1988) require higher total annual flow to reduce salt intrusion into the Delta due to sea level rise (Figure 10).
These annual flows are significantly greater than occurring presently and the regression using annual flow
predicts greater fish abundance. However, the Williams report also predicts lower flows in the spring months
(April through June). High spring flows have been shown to be the most critical to young fish survival and to
greater abundance of spawning fish i-» subsequent years. Rozengurt et aL (1987) and Kjeldson et aL (1981)
recommend spring flows greater than 40,000 cfs to support anadromous fish. None of the climate scenarios
provide sufficient flows to meet these criteria, and these fish populations would likely decline.
A complicating factor to these scenarios is the probably of levee failure in the Delta creating a large
inland lake with fresh to brackish water quality. Despite the larger winter flows predicted by Williams, average
salinities in the western Delta are predicted to increase over base conditions. Striped bass and shad spawn in
essentially fresh water conditions, which under base conditions occur during winter and spring months in the
Delta. The predicted increase in salinity would eliminate this region as a spawning ground.
A short-term data base (1979-1987) for non-commercial fish species illustrates the difficulty of assessing
long-term changes in the estuarys fauna (California Department of Fish and Game, 1987). This report examined
the relationship between inflow and monthly fish abundances at over 70 deep and shallow-water sampling sites
in San Francisco Bay. High correlations between flows and abundance were observed for a number of
crustaceans and fish. For these organisms, winter and spring flows (February-May) were commonly correlated
with abundance. However, for a great many species, no correlation or negative correlations were observed.
In addition, distribution in relation to salinity varied between years for many of the species and, therefore, no
reliable regression with salinity could be developed.
Comparing the flow and salinity scenarios developed by Williams with the Fish and Game data base
indicates that despite increasing flows, the higher salinities predicted for the Suisun and San Pablo Basin will
increase the abundance of marine species occurring within these basins while decreasing freshwater and
anadromous species. Specifically, species such as California halibut, jacksmelt, and walleye surfperch will
increase their use of the estuary, while species such as starry flounder, threespine stickleback, and longfin smelt
will decrease. The data are insufficient to determine whether this species shift will result in lower or higher
fish biomass within the estuary.
SUMMARY OF RESULTS
Specific responses of estuarine organisms to global climate change are difficult to assess from current
models. Presently, increased flows result in decreased salinities within the estuary and the available regressions
between flow and plant or animal distribution are influenced by this correlation. Under the scenarios
investigated by Williams, the necessity to increase carriage water flow to retard salt intrusion to the state's water
supply does not result in decreased salinities in the estuary. Instead, average salinity in the basins also increases
owing to the larger volume of water present. The resulting conditions are not within the bounds of current
regression models. Therefore, our conclusions are primarily qualitative (Table 3).
3-23
-------
Josselyn
AVERAGE MONTHLY DELTA OUTFLOWS
Under various climate scenarios
120
38
U- 3
DEC JAN FEB MAR APR MAY JUN JUL AUG SEP
OCT
BASEAVG
GFDLAVG
MONTH
GISSAVG
OREGAVG
Figure 10. Projected monthly Delta outflows under various climate scenarios (from Williams).
...•*
3-24
-------
Josselyn
Table 3. Summary of conclusions on the impact of global climate
change on estuarine habitats and organisms
HABITAT OR
ORGANISM(S)
SCENARIO
LIKELY RESPONSE OR CHANGE
DURING PERIOD 1990-2040
Tidal marsh
100 cm rise
Diked baylands
200/300 cm rise
All sea level
rise scenarios
and with levee
maintenance
Rare and
endangered
species
Phytoplankton
biomass
With levee
failure
All models
GISS, GFDL
with levee fail-
ure
OSU with levee
failure
Benthic
invertebrates'
All models
Some decline in acreage, but
marshes able to maintain
accretionary balance; landward
shift in distribution of plant
species; decrease in production
of freshwater/brackish water
plants.
Likely inundation of many of
the fringing tidal marshes by
end of period.
Increasing degradation of
marsh vegetation and water
quality due to inability to
circulate water within diked
wetland channels, decline in
brackish water vegetation
used by wintering and nesting
migratory waterfowl.
Conversion to open water
habitat, loss of waterfowl
nesting habitat.
Decreased wetland habitat due
to tidal marsh loss from
increased shoreline erosion,
localized and regional
extinction.
Some reduction due to shift
upstream during some years
Shift of peak biomass
upstream in all years, decline
in zooplankton biomass and
lower survivorship of juvenile
striped bass.
Invasion of marine species
and displacement of existing
estuarine species.
3-25
-------
Josselyn
Table 3 (continued)
HABITAT OR
ORGANISM(S)
SCENARIO
LIKELY RESPONSE OR CHANGE
DURING PERIOD 1990-2040
Shift upstream in distri-
bution, greater abundance
of marine species.
Decline in spawning area
within delta due to higher
salinities; likely decline
in abundance due to lower
spring flows.
Fish distri-
bution
All models
Anadromous fish All models
3-26
-------
Josselyn
CHAPTER 4
IMPLICATIONS OF RESULTS
ENVIRONMENTAL IMPLICATIONS
Under the moderate and highest sea level rise scenarios, tidal marshes will be inundated as sediment
supply will be insufficient to maintain an accretionary balance. This will result in significant losses of animals
and plants associated with tidal marshes and adjacent mudflats. Migratory waterfowl and shorebird habitat will
be completely inundated. A number of rare and endangered species will become regionally extinct.
On the other hand, the implications to the estuary as a whole may not be significant. This is largely
due to the substantial historical losses of tidal marshes which have already occurred. At present, only about
10% of the historical tidal marsh acreage surrounding San Francisco Bay exists. In the past, the area of
marshland around the bay was twice as large as the open water area; ow it is less than 10%. Unlike the vast
tidal marshes of the southeastern United States that export primary production to the estuary, phytoplankton
are responsible for the majority of primary production within the San Francisco Bay estuary (Spiker and
Schemel, 1979).
Levee failure will allow transgression of the Bay over former tidal marshes. However, it is unlikely that
new emergent wetlands will be created for several reasons. First, most of the diked baylands are currently at
or below sea level at present. In the Delta, some lands are 20 feet or more below sea level. Given the sediment
deficit predicted, these lands may not accrete rapidly enough to create emergent wetlands. Second, many of the
low-lying areas at elevations which might support emergent wetlands upon inundation are currently urbanized.
If levees are maintained, the diked wetlands behind these levees will degrade owing to the inability to
flush water through the channels without the use of large pumping systems. These areas will pond water during
the winter and then gradually dry out during the summer with a concomitant build-up of salts in the soils.
Presently, salt build-up in low-lying diked wetlands of Suisun Marsh and in south San Francisco Bay is causing
a die-back in salt marsh vegetation and anoxic conditions within marsh channels.
Regardless of whether or not levees are maintained, the primary fauna! impact will be the loss of
vegetated tidal marsh habitat that supports rare and endangered species.
The changes in flow and salinity due to modification of freshwater inflow are more difficult to assess.
Marine and estuarine organisms will be distributed more landward due to the increases in average salinity
throughout the basin. Circulation of the estuary will change dramatically as sea level rises and if levees fail.
Changes in sedimentation, retention times, and gravitational circulation are unknown, yet are crucial to
understanding the biological responses. The predicted reductions in spring and early summer flow are the most
important to the functioning the estuary in its current configuration, especially to the transport of juvenile
anadromous fish such as salmon and striped bass.
SOCIOECONOMIC IMPLICATIONS
Extensive public and private resources are invested in tidal and managed wetlands. State and federal
agencies require creation or restoration of wetlands as mitigation for development hi wetlands. To date, over
SO projects involving wetland creation have occurred in the estuary, ranging from 1 to 500 acres. In addition,
the 44,000 acres of managed wetlands in Suisun Marsh represent a significant investment by public and private
interests. These wetlands are used primarily for recreational hunting and fishing, itself an industry valued at
over $150 million annually (Meyer, 1987). The state and federal governments are investing over $60 million
dollars to construct new water supply facilities to the managed wetlands to reduce current salinity problems.
3-27
-------
Josselyn
There are also indirect values attributable to public use of wetlands such as tourism, day recreation, and
non-consumptive uses. Finally, non-marketable values such as rare and endangered species and heritage values
must be considered.
The losses of wetlands, especially diked wetlands if levee failure occurs, would result in substantial
economic losses for the region. If levees are to be maintained, the costs would also be significant. The State
of California is currently embarking on a 10-year, $100 million levee stabilization project for the Delta. If the
higher ranges of sea level rise prove to be true, the cost of levee repair may be insurmountable.
3-28
-------
Josselyn
CHAPTERS
POLICY IMPLICATIONS
San Francisco Bay conservation and Development Commission (1987b) has reviewed some of the local
policy decisions which must be made to deal with projected impacts to the estuarine shoreline. Generally, these
deal with implementing designs that take into account projected increases in sea level. The report also suggests
the establishment of monitoring stations around the bay to determine long-term changes in sea level and
provide data to estimate future changes.
The most significant impact to the estuary will be the decision to maintain or abandon the current levee
system. There is no choice for areas which are currently developed such as Foster City and Corte Madera.
Levees will require replacement and pumping systems will have to be constructed The levees surrounding
less valuable land (agriculture and diked baylands) may not receive the same priority. Many of the levees are
poorly engineering (if engineered at all) and are subject to increasing rates of failure due to high tides and
floods. Placing more rip-rap or higher levels of fill on these levees will not be an acceptable solution. However,
maintenance of the levee system will be essential to reduce the amount of carriage water necessary to repel
salt water intrusion and to reduce flooding hazards to developed areas behind the diked baylands.
Several engineering alternatives are possible. One is the construction of high tide/flood barriers which
will reduce the impact of high tides. Similar barriers are under construction on the Thames River and the
Venice Lagoon. The cost would be substantial for San Francisco Bay and would have significant economic,
environmental, and aesthetic impacts. However, they would provide protection to both developed and
undeveloped areas.
A second alternative involves the construction of a salinity barrier to retard landward movement of salt
as sea level increases. The barrier could be a subsurface sill, a perforated wall, or a constricted channel located
along the Carquinez Straits. Whatever design is used, its function would be to regulate the location of the null
zone. Although there are a number of environmental problems associated with such a barrier, its primary
benefit would be a reduction in the amount of carriage water necessary to repel salt intrusion into the Delta.
A third engineering alternative is to build up the elevations of diked baylands using dredge spoil or
other fill material. This alternative would require making decisions which have severe short-term impacts such
as the temporary loss of seasonal wetlands (some of which harbor endangered species) in order to manage for
long-term impacts of inundation if the levees fail Given the critical shortage of both tidal and seasonal wetlands
in the bay, few agencies will have the ability to suggest any alteration. Nevertheless, it is important that some
initial projects be started to demonstrate the feasibility of altering elevations in diked wetlands with subsequent
restoration to functioning wetlands. Otherwise, the ultimate fate of these areas will be complete loss.
ADDITIONAL RESEARCH
All phases of estuarine research need to be continued, especially in the area of circulation modeling
and factors affecting the distribution of organisms in the estuary. However, one area which is most lacking is
an understanding of sediment dynamics in the San Francisco Bay estuary. Sedimentation and sediment
movement will play a crucial role in whether marshes, mudflats, and diked baylands will be sustained under
projected climate scenarios. Most of the models are based on relatively low rates of marsh accretion
determined under current sea level rise. However, several studies reviewed in this report showed that in areas
of subsidence, 'sedimentation can be extremely rapid. Very few measurements of marsh accretion have been
conducted hi such conditions. In addition, the rate of input of sediment relative to freshwater inflow needs to
be determined.
3-29
-------
Josselyn
REFERENCES
Atwater, B.F., C.W. Hedel, and EJ. Helley. "Late Quaternary Depositional History, Holocene Sea Level
Changes, and Vertical Cmstal Movement, Southern San Francisco Bay." U.S. Geological Survey Professional
Paper 1014. U. S. Geological Survey, Menlo Park, California, 1977. 15 pp.
Atwater, B.F., S.G. Conard, J.N. Dowden, C.W. Hedel, R.L. MacDonald, and W. Savage. "History, Landfonns,
and Vegetation of the Estuary's Tidal Marshes." In: San Francisco Bay: the Urbanized Estuary, T J. Conomos,
ed Pacific Div. Amer. Assoc. Advance. Sci., San Francisco, California, 1979. pp. 347-386.
Ball, M.D., and Arthur, J.F. Tlanktonic Chlorophyll Dynamics in the Northern San Francisco Bay and Delta"
In: San Francisco Bay: the Urbanized Estuary, TJ. Conomos, ed Pacific Div. Amer. Assoc. Advance. Sci.,
San Francisco, California, 1979. pp. 265-285.
California Department of Fish and Game. "Delta Outflow Effects on the Abundance and Distribution of San
Francisco Bay Fish and Invertebrates, 1980-1985." Exhibit 60, entered by the California Department of Fish
and Game for the State Water Resources Control Board 1987 Water Quality/Water Rights Proceeding on the
San Francisco Bay/Sacramento-San Joaquin Delta. California Department of Fish and Game, Sacramento,
California, 1987. 345pp.
California Department of Fish and Game. Testimony on the Suisun Marsh." Exhibit 5, entered by the California
Department of Fish and Game for the State Water Resources Control Board 1987 Water Quality/Water Rights
Proceeding on the San Francisco Bay/Sacramento-San Joaquin Delta. California Department of Fish and
GAme, Sacramento, California, 1987b.
Conomos, T J. "Properties and Circulation of San Francisco Bay Waters." IirSan Francisco Bay: the Urbanized
Estuary, TJ. Conomos, ed Pacific Div. Amer. Assoc. Advance. Sci., San Francisco, California, 1979. pp47-84.
Davis, G.H. "Land Subsidence and Sea Level Rise on the Atlantic Coastal Plain of the United States." Environ.
Geol. Water ScL, 10(2):67-80,1987.
DeLaune, R.D., R.H. Baumann, and J.G. Gosselink. "Relationships Among Vertical Accretion, Coastal
Submergence, and Erosion in a Louisiana Gulf Coast Marsh." J. of Sediment. Petrol., 53(1): 147-157,1983.
DeLaune, R.D., S.R. Pezeshki, and W.H. Patrick, Jr. "Response of Coastal Plants to Increase in Submergence
and Salinity." J. of Coastal Res., 3(4):535-546,1987.
Hackney, C.T., and W J. Cleary. "Saltmarsh Loss in Southeastern North Carolina Lagoons: Importance of Sea
Level Rise and Inlet Dredging." J. of Coastal Res., 3(l):93-97,1987.
Harrison, E.Z., and AX. Bloom. "Sedimentation Rates on Tidal Salt Marshes in Connecticut" J. of Sediment.
Petrol., 47(4):1484-1490,1977.
Hatton, RS~, R.D. De Laune, and W.H. Patrick, Jr. "Sedimentation, Accretion, and Subsidence in Marshes
of Barataria Basin, Louisiana." Limnol. Oceanogn, 28(3):494-502,1983.
Hoese, H.D. "Effect of Higher than Normal Salinities on Salt Marshes." Contrib. Mar. Set, 12:249-261,1967.
Hull, C.HJ., and* J.G. Titus, eds. "Greenhouse Effect, Sea Level Rise, and Salinity in the Deleware Estuary."
EPA-230-05-86-010. U.S. Environmental Protection Agency, Washington, D.C., 1986. 88 pp.
Josselyn, M. "The Ecology of San Francisco Bay Tidal Marshes: A Community Profile." FWS/OBS-83/23.
U.S. Fish and Wildlife Service, Division of Biological Services, Washington, D.G, 1983.102 pp.
3-30
-------
Josselyn
Josselyn, M., and J.W. Buchholz. "Marsh Restoration in San Francisco Bay: A Guide to Design and Planning."
Technical Report No. 3. Tiburon Center for Environmental Studies, San Francisco State University, Tiburon,
California, 1984. 104 pp.
Josselyn, M., L. Handley, M. Quammen, and D. Peters. The Distribution of Wetland Habitats in the San
Francisco Bay Region." US. Fish and Wildlife Service, Office of Biological Services. Washington, DC. In press.
Kjelson, MA., P.F. Raquel, and F.W. Fisher. "Influences of Freshwater Flow on Chinook Salmon in the
Sacramento-San Joaquin Estuary. In: Proceedings of the National Symposium on Freshwater Inflow to
Estuaries, R.D. Cross and D.L. Williams, eds. United States Department of the Interior, Washington, DC, 1981.
Vol. 2. pp. 88-108.
Klein ffl, CJ., S.P. Orlando, Jr., C. Alexander, J.P. Tolson, F. Shirzad, R.B. Biggs, and E. Zolper. "How
Representative Are the Estuaries Nominated for EPA's National Estuary Program? Strategic Assessment
Branch, National Oceanic and Atmospheric Administration. Washington, DC. 1988. 9pp.
Krone, R.B. "Sedimentation in the San Francisco Bay System." In: San Francisco Bay: the Urbanized Estuary,
TJ. Conomos, ed. Pacific Div. Amer. Assoc. Advance. Sci., San Francisco, California, 1979. pp. 85-96.
Krone, R.B. "A Method for Simulating Historic Marsh Elevations." Unpublished 8 pp.
Letzsch, W.S., and R.W. Frey. "Deposition and Erosion in a Holocene Salt Marsh, Sapelo Island, Georgia." J.
of Sediment. Petrol., 50(2):529-542,1980.
Lyle, SJX, L.E. Hickman, Jr., and HA. Debaugh, Jr. Sea Level Variations for the United States, 1855-1986.
National Oceanic and Atmospheric Administration, Rockville, Maryland, 1987.
Mahall, B.E., and Park R.B. "The Ecotone between Spartina foliosa Trin. and Salicornia virginica L. in Salt
Marshes of Northern San Francisco Bay: H. Soil Water and Salinity." J. EcoL, 64:793-809, 1976.
Mall, R.E. "Soil-Water-Salt Relationships of Waterfowl Food Plants in the Suisun Marsh of California."
California Department of Fish and Game, Wildlife Bulletin No. 1. California Department of Fish and
Game, Sacramento, California, 1969. 59 pp.
Martindale, M. "Salicornia Europea L. and Salicornia virginica L. on a San Francisco Bay Salt Marsh: A
Study of Factors Contributing to their Zonation Pattern." MA. Thesis, San Francisco State University, San
Francisco, California, 1987. 52 pp.
Meyer, PA. The value of wildlife in San Francisco Bay." Exhibit 38, entered by the Bay Institute of San
Francisco for the State Water Resources Control Board 1987 Water Quality/Water Rights Proceeding on the
San Francisco Bay/Sacramento-San Joaquin Delta Bay Institute of San Francisco, Sausalito, California, 1987.
22pp.
Nichols, F.H., J.E. Cloern, S.N. Luoma, and D.H. Peterson. The modification of an Estuary." Science
231:525-648. 1986.
Nolan, K.M., and Fuller, C.C. "Sediment accumulation in San Leandro Bay, Alameda County, California,
During the 20th Century: A Preliminary Report." US. Geological Survey Water-Resources Investigations
Report 86-4057. U.S. Geological Survey, Sacramento, California, 1986. 25 pp.
Orson, R., W. Panageotou, and S.P. Leatherman. "Response of Tidal Salt Marshes of the U.S. Atlantic and
Gulf Coasts to Rising Sea Levels." J. of Coastal Res., l(l):29-37, 1985.
3-31
-------
Josselyn
Patrick, W.H., Jr. and R.D. DeLaune. "Capacity of south San Francisco Bay Salt Marshes to Accrete Rapidly
Enough to Compensate for Subsidence and Sea Level Rise." Manuscript. Center for Wetland Soils, Louisiana
State University. Unpublished.
Pearcy, R.W., and S.L. Ustin. "Effects of Salinity on Growth and Photosynthesis of Three California Tidal
Marsh Species." Oecologia, 62:68-73, 1984.
Perez, RJ. "Salt Marsh Restoration from Former Salt Evaporators: Changes in Sediment Properties." MA.
Thesis, San Francisco State University, San Francisco, California, 1981. 41 pp.
Pestrong, R. "Tidal-Flat Sedimentation at Cooley Landing, Southwest San Francisco Bay." Sediment. GeoL,
8:251-288,1972.
Pethick, J.S. "Long-Term Accretion Rates on Tidal Salt Marshes." J. of Sediment. PetroL, 51(2):571-577,1981.
Phillips, J.D. "Coastal Submergence and Marsh Fringe Erosion." J. of Coastal Res., 2(4):427-436,1986.
Richard, GA. "Seasonal and' Environmental Variations in Sediment Accretion in a Long Island Salt Marsh."
Estuaries, l(l):29-35,1978.
Rosen, PJS. "A Regional Test of the Bruun Rule on Shoreline Erosion." Mar. GeoL, 26:M7-M16,1978.
Rozengurt, MA., MJ. Herz, and S. Feld. "The Role of Water Diversions in the Decline of Fisheries of the
Delta-San Francisco Bay and Other Estuaries." Technical Report No. 87-8. Romberg Tiburon Center for
Environmental Studies, San Francisco State University, Tiburon, California, 1987.158 pp.
Salinas, L.M., R.D. DeLaune, and W.H. Patrick, Jr. "Changes Occurring along a Rapidly Submerging Coastal
Area: Louisiana, USA." J. of Coastal Res., 2(3):269-284, 1986.
San Francisco Bay Conservation and Development Commission. "Diked Historic Bayiands of San Francisco
Bay." Staf Report. San Francisco, California, 1981 38 pp.
San Francisco Bay Conservation and Development Commission. "An Overview of the Impact of Accelerated
Sea Level Rise on San Francisco Bay." Prepared by P.B. Williams. San Francisco, California, 1985. 25 pp.
San Francisco Bay Conservation and Development Commission. "An Overview of Flow and Salinity Standards
Required to Protect the Ecosystem of San Francisco Bay." Prepared by P.B. Williams and M. Josselyn. San
Francisco, California, 1987a. 45 pp.
San Francisco Bay Conservation and Development Commission. "Future Sea Level Rise: Predictions and
Implications for San Francisco Bay." Prepared by Moffatt and Nichol, Engineers, and Wetlands Research
Associates, Inc. San Francisco, California, 1987b. 98 pp.
San Francisco Bay Conservation and Development Commission. "Recommendations for Salinity Standards to
Maintain the Wetlands of Suisun Marsh." Prepared by P.B.Williams and M. Josselyn. San Francisco, California,
1987c. 35 pp.
Spiker, E.C. and L.E. SchemeL "Distribution and Stable-Isotope Compositon of Carbon in San Francisco Bay."
In: San Francisco Bay: the Urbanized Estuary, T J. Conomos, ed. Pacific Div. Amer. Assoc. Advance. Sci., San
Francisco, California, 1979. pp. 195-212.
Stevens, D.E. and L.W. Miller. "Effects of river flow on abundance of young chinook salmon, American shad,
longfin smelt, and delta smelt in the Sacramento-San Joaquin River system. N. Amer. J. Fish. Manage., 3:425-
437,1983.
3-32
-------
Josselyn
Stevenson, J.C., L.G. Ward, and M-S. Kearney. "Vertical Accretion in Marshes with Varying Rates of Sea Level
Rise." In: Estuarine Variability, D. Wolf, ed. Academic Press, New York, 1986. pp. 241-260.
Stevenson, J.C., L.G. Ward, MS. Kearney, and T.E. Jordan. "Sedimentary Processes and Sea Level Rise in
Tidal Marsh Systems of Chesapeake Bay." In: Wetlands of the Chesapeake, HA. Groman, EJ. Meyers, D.M.
Burke, and JA. Kusler, eds. The Environmental Law Institute, Washington, D.C., 1985. pp. 37-62.
Stumpf, R.P. "The Process of Sedimentation on the Surface of a Salt Marsh." Estuarine Coastal Shelf Sci.,
17:495-508, 1983.
Thompson, J JC and F.H. Nichols. "Food Availability Controls Seasonal Cycle of Growth in Macoma balthica
(L.) in San Francisco Bay, California." J. Exp. Mar. Biol. Ecol., 116:43-61, 1988.
Titus, J.G. "Greenhouse Effect, Sea Level Rise, and Coastal Zone Management." Coastal Zone Mngmnt. J.,
14(3):147-171, 1986.
Titus, J.G., C.Y. Kuo, MJ. Gibbs, T.B. LaRoche, M.K. Webb, and J.O. Waddell. "Greenhouse Effect, Sea
Level Rise, and Coastal Drainage Systems." J. Water Res. Planning Mngmnt., 113(2):216-227, 1987.
United States Fish and Wildlife Service. "The Needs of Chinook Salmon, Oncorhvnchus tshawvtscha. in the
Sacramento-San Joaquin Estuary." Exhibit 31, entered by the US. Fish and Wildlife Service for the State Water
Resources Control Board 1987 Water Quality/Water Rights Proceeding on the San Francisco Bay/
Sacramento-San Joaquin Delta. U.S. Fish and Wildlife Service, Sacramento, California, 1987. 179 pp.
Williams, P.B. "The Impact of Climate Change on the Salinity of San Francisco Bay." Philip Williams and
Associates, San Francisco, California. Environmental Protection Agency, Washington, DC. 1988. 38pp.
Williams, P.B. and J.T. Hollibaugh. "A Salinity Standard to Maximize Phytoplankton Abundance by Positioning
the Entrapment Zone in Suisun Bay." Exhibit 1, entered by the Contra Costa County Water Agency for the
State Water Resources Control Board 1987 Water Quality/Water Rights Proceeding on the San Francisco Bay/
Sacramento-San Joaquin Delta. Philip Williams and Associates, San Francisco, California, 1987a. 40 pp.
Williams, P.B. and J.T. Hollibaugh. "A Flow Standard to Maximize Phytoplankton Abundance by Positioning
an Entrapment Zone San Pablo Bay." Exhibit 3, entered by the Contra Costa County Water Agency for the
State Water Resources Control Board 1987 Water Quality/Water Rights Proceeding on the San Francisco Bay/
Sacramento-San Joaquin Delta. Philip Williams and Associates, San Francisco, California, 1987b. 31 pp.
3-33
-------
PROJECTED CHANGES IN ESTUARINE CONDITIONS BASED ON MODELS OF
LONG-TERM ATMOSPHERIC ALTERATION
by
Robert J. Livingston
Center for Aquatic Research and Resource Management
Florida State University
Tallahassee, FL 32306
Contract No. CR-814608-01-0
-------
CONTENTS
Page
ACKNOWLEDGMENTS iii
FINDINGS 4-1
CHAPTER 1: INTRODUCTION 4-4
EXISTING DATA BASES: PROJECT SCOPE 4-4
THE APALACfflCOLA RIVER-BAY SYSTEM: BACKGROUND 4-4
CHAPTER 2: METHODS 4-13
GENERAL APPROACHES 4-13
SCENARIOS FOR EFFECTS MODELING 4-13
CHAPTER 3: RESULTS 4-17
RIVER FLOW SCENARIOS 4-17
Water Quality Changes 4-22
Long-Term Biological Changes 4-23
Projections 4-28
TEMPERATURE 4-30
Scenarios 4-30
Biological Response 4-30
SEA LEVEL CHANGES 4-37
Projected Effects 4-37
CHAPTER 4: DISCUSSION 4-44
RIVER FLOW PROJECTIONS AND EFFECTS 4-44
TEMPERATURE CHANGES AND EFFECTS 4-45
PROJECTED CHANGES IN SEA LEVEL 4-46
REFERENCES 4-48
11
-------
ACKNOWLEDGMENTS
This report is based on the work of literally hundreds of people over an 18-year period of study. The
funding of this work has been diverse, with various agencies (local, state, and federal) contributing over the years
to the final product. The Apalachicola data base is the product of the cumulative efforts of students
(undergraduate and graduate), faculty, independent researchers, staff personnel, and a long line of volunteers
who have given their time in a most generous fashion. A series of project managers and patrons have allowed
the continuous accumulation of field and experimental data during this period with the result that we are now
in the final stages of analysis, modeling, and publication of the results. Appendices cited in the text of this paper
were not included in this volume because of space considerations; however, they may be obtained from Jim Titus,
OPPE, EPA. The appendices are:
I. Analysis of existing (long-term, interdisciplinary) field data concerning a series of river-estuaries in the
southeastern United States (South Carolina, Georgia, Florida, Mississippi, Alabama).
II. Curriculum vitae: Robert J. Livingston.
HI. Statistical analyses of epibenthic macroinvertebrates of the Apalachicola estuary.
IV. Statistical analyses of fishes of the Apalachicola estuary.
V. Data concerning sea level changes in the Apalachicola region. (Dr. Richard Park, Butler University)
VI. Use of existing field data to evaluate the effects of projected long-term climatological changes in a south
temperate gulf estuary. (R J. Livingston, Florida State University)
ui
-------
Livingston
FINDINGS1
This paper addresses the problem of the potential effects of projected climatological changes on a
river-estuary in the northern Gulf of Mexico. A series of scenarios was developed by other researchers
concerning long-term climatological changes due to atmospheric contaminants. Such scenarios involved changes
in precipitation/river flow, temperature, and sea level in the region encompassing the Apalachicola-
Chattahoochee-Flint drainage system in Alabama, Georgia, and Florida Using long-term data sets (1972-1987)
from the Apalachicola estuary, it was possible to make estimates of potential changes in estuarine productivity
and dominant (commercially valuable) populations based on the long-term scenario predictions.
RIVER FLOW
According to models provided by Dr. C. F. Hains (Northport, AL) concerning altered Apalachicola
River flow, the following projections (percent losses or gains) could be estimated concerning the response of
estuarine functions in the Apalachicola Bay system:
SCENARIO
FACTOR GFDLx2 GISSx2 OSUx2
River flow -24.2 -6.2 -13.9
Mean salinity +113 +3.1 +6.5
Detritus/nutrients -10-25 -3-7 -5-15
Detrital food web -10-25 -3-7 -5-15
White shrimp -15-35 -5-10 -10-30
Pink shrimp +10-20 +3-5 +8-20
Blue crabs -15-35 -5-10 -10-30
Finfishes -15-35 -5-10 -10-30
Oysters -20-35 -7-10 -15-30
The above estimates are based on the combined losses of organic carbon and nutrients to the system and on
models of increased predation rate on nurserying juveniles of the top dominant estuarine populations due to
increased salinity. Oysters would be particularly vulnerable as the salinity increased owing in part to increased
predation and disease. These changes are predicated on the assumption that existing populations would not be
replaced either by adaptation of existing species or invasion by species new to the region. Such an assumption
is somewhat unlikely but remains difficult to estimate in a quantifiable fashion. The increase in pink shrimp
would be due to the increased salinity and probable increases in the submerged aquatic vegetation that would
accompany the projected changes in river flow and estuarine habitat.
TEMPERATURE
According to the GFDLx2 projections of changes in ambient temperature in the tri-river region, there
would be a mean increase of 3-5" C with even greater changes in the mean maximum temperatures of the estuary
'Although the information in this report lias been funded wholly or partly by the US. Environmental
Protection Agency under Contract No. CR-814608-01-0, it does not necessarily reflect the Agency's views, and
no official endorsement should be inferred from it.
4-1
-------
Livingston
with time. Such projections are based on the assumption that the estuarine water temperatures would closely
follow the air temperatures. Currently, the summer high temperatures of the Apalachicola estuary are close to
the upper thermal tolerance limits of most of the dominant (commercially important) species in the system.
With such assumptions in mind, the following projections were made based on the established thermal tolerance
limits and projected changes of water temperature in the Apalachicola estuary:
SCENARIO
FACTOR GFDLx2 GKSx2 CURRENT (AMBIENT)*
Blue crab larvae -100 -100 -77
Blue crab juveniles -100 -100 -27
Spotted seatrout -91 -59 -5
Oyster larvae -91 -59 -5
Adult oysters ILC. n.c. ILC.
Pinfish -91 -59 -5
Flounder -91 -59 -5
Redfish -55 -9 ILC.
Mullet -5 ILC. ILC,
White shrimp -5 n.c. n.c.
•current loss of habitat use due to summer high temperatures
The above estimates are indications of a loss of estuarine habitat to various commercially important
species due to increased water temperature. Currently, juveniles of such species use the estuarine habitat as a
highly productive nursery and as a haven from offshore (stenohaline) predators. Such a loss of habitat could not
be replaced and the movement of the nurserying populations to offshore regions in the gulf would probably lead
to major losses of such populations owing to offshore predation.
The possibility of adaptation and/or replacement by tropical species is quite real but is not easily
estimated in the above models. There are currently some very productive tropical estuaries which could serve
as models for projected changes in the Apalachicola region. Species such as mullet and white shrimp would not
necessarily be lost if temperature is the only variable under consideration. The above projections should also
be qualified by the various assumptions implicit in the analysis. However, if the projected temperature changes
do occur, there will be some major changes in the structure and productivity of the Apalachicola Bay system with
considerable shifts in population dominance and fisheries potential. By analogy, such changes could be projected
to other important river-estuaries along the northern gulf coast.
SEA LEVEL RISE
Paniculate organic carbon (POC) can be used as an indicator of allochthonous and autochthonous sources
of production in the Apalachicola estuary. There are considerable data concerning the production, distribution,
and utilization of POC in the Apalachicola drainage system. According to a series of models developed by Dr.
Richard Park (Indianapolis, Indiana), various scenarios based on different levels of sea level rise were produced
concerning the changes in wetlands currently within the Apalachicola drainage system. Such changes were then
translated into overall gains or losses of POC over the period 1987-2100. The following is an estimate of the net
input of organic carbon (metric tons/year) in the Apalachicola system based on current conditions and projected
sea level rises in the region:
4-2
-------
Livingston
SCENARIO (by 2100)
FACTOR Current 1.2mm/yr OJm/yr l.Om/yr 2.0m/yr
Freshwater wetlands 30,000 26,100 24,000 21,300 4,980
Seagrasses 27,200 28,700 28,800 30,100 31,035
Saltmarsh 46,905 23,500 4,690 940 780
Phytoplankton 233,280 144,640 71,450 58,790 15,160
Total 337,385 222,940 128,940 111,130 51,955
Since there are various other limiting factors that contribute to the recreational and commercial fisheries
of the region, it is difficult to estimate the potential losses of such populations. However, the various wetlands
serve both as habitats and sources of productivity to the bay system. Any loss of such wetlands would certainly
lead to reductions of these estuarine populations. One of the key issues in this estuary concerns the fate of
phytoplankton production which is the single most important factor in the high level of productivity in the bay.
Such production is largely a measure of freshwater input (e.g., a major nutrient source) to the system and the
shallowness of the estuary. The relationship of the phytoplankton and the surrounding wetlands is poorly
understood The above results are predicated in part on the correlation of phytoplankton productivity and the
extent of wetland areas; this assumption is further complicated by the lack of an adequate forecast of microbial
activity in the estuary (based on increasing losses of wetlands with time). Such twin uncertainties would thus
serve to qualify the above projections and would limit any meaningful extrapolation to detailed projections of
population response to such changes. Nevertheless, the progressive loss of wetlands to the region would lead
to substantive changes in the habitat availability and nutrient/POC budgets in the Apalachicola drainage system.
It is quite likely that a substantial, perhaps even proportional, loss would occur of species that depend on
the detrital food webs of the estuary. Such species would include white shrimp, blue crabs, spot, Atlantic
croakers, and white seatrout. If, as postulated, the phytoplankton productivity were adversely affected by such
losses, species such as anchovies and oysters would also be adversely affected. There would be a replacement
of some of the primary productivity of the estuary through increased sea grass bed distribution. This would
improve the production of species such as pink shrimp and spotted seatrout.
There are certain shortcomings to the application of correlations, regressions, and multivariate statistical
techniques to a complex data set that includes many (diverse) biological variables. Accordingly, this analysis has
included various assumptions that must be taken into account when evaluating projected changes in the system
according to anthropogenous alterations in the various climatic features of the system. In addition, changes in
both water quality and freshwater input to the system due to future human activities associated with municipal
developments, industrial expansion, and agricultural activities could significantly alter these projections. An effort
has been made to preclude such impacts, although various forms of deterioration due to human activities around
the bay have already taken a serious toll on the Apalachicola fisheries. Such effects can be addressed. By
contrast, climatological changes would eventually be beyond any efforts of amelioration. It is almost a certainty
that the Apalachicola estuary, as it currently exists, would be severely altered by a global warming trend.
Although it could be transformed into a productive tropical estuary dominated by seagrass productivity, it is
highly likely that there would be an overall decrease in fisheries production and that qualitative changes in the
commercially valuable populations would lead to far-reaching changes in fisheries equipment and practices and
the availability of seafood to people in the region.
4-3
-------
Livingston
CHAPTER 1
INTRODUCTION
EXISTING DATA BASES: PROJECT SCOPE
This project was designed to evaluate the possible effects of projected long-term changes in the climate
(e.g., temperature, precipitation) and associated variables (river flow, sea level changes) related to increased
atmospheric CO2 and other greenhouse gases. The analysis was based on a 15-year data base taken in the
Apalachicola Bay system (East Bay, Apalachicola Bay: Figure 1). Data were taken continuously on a weekly
to monthly basis from March 1972 through October 1987. Station groups were summed in East Bay
(oligohaline) and Apalachicola Bay (mesohaline to polyhaline). The Apalachicola River system (regional) and
Tale's Hell Swamp (local) constitute the two major sources of fresh water to the system.
THE APALACHICOLA RIVER-BAY SYSTEM: BACKGROUND
The Apalachicola River is the terminus of a major drainage system that is located mainly in Georgia and
Alabama Seasonal river flow patterns reflect Georgia precipitation and seasonal changes in evapotranspiration
rates of the tri-river (Apalachicola-Chattahoochee-Flint) system (Meeter et aL, 1987). The winter-spring river
peak flows strongly influence various ecologically important estuarine habitat variables (Livingston, 1984a).
Emergent and submergent vegetation (freshwater to marine), along with phytoplankton production, provide a
year-round source of energy for the system (Livingston, 1983, 1984a) as mediated by river flow and microbial
activities (Figure 2). The relationship of river flow and microdetritus is well established as a function of the
winter-spring flood events (Table 1). Long-term river flow changes follow interannual periods that approximate
5-7 year cycles (Meeter et al., 1979). Such trends determine salinity changes in the bay, which, in turn, have an
important impact on the biological organization of the estuary (Livingston, 1984a).
Organisms at different levels of biological organization follow different seasonal patterns of numerical
abundance and species richness (Figure 3). Peak numbers of fishes, dominated by the sciaenids, occur in the
winter-early spring months whereas epibenthic invertebrates peak in the late summer and fall months and the
infaunal macroinvertebrates peak in the winter. Specific dominant fish and invertebrate populations, some with
great commercial value, follow highly individualistic patterns that vary seasonally and yearly (Figures 3, 4, and
5). Most such populations spawn offshore (Figure 6). Others, such as oysters, spend their entire lives in the
estuary. As a consequence of the river input and basic climatological conditions, the Apalachicola system
produces a major fishery with a dock value of oysters, penaeid shrimp, and finfishes of over $12 million in 1980.
According to Prochaska and Mulkey (1983), application of multipliers to such output would increase the regional
value of such fisheries to between $17 and $23 million for that year. Projected values of the estuary by the year
2000 have varied from $34 million (Colberg et al., 1968) to $67 million (Mulkey and Maturo, 1983). The
Apalachicola Bay system also serves as a habitat for various rare, endangered or threatened species as well as
a major flyway for diverse birds (Livingston and Joyce, 1977).
The basic model (Figure 7) for the ongoing analysis of the Apalachicola system indicates the extreme
complexity involved in any prediction of future changes based on climatological alterations. Because of this
complexity, there are various assumptions that must be made when applying statistical models to the field data.
Such assumptions severely restrict a truly quantitative approach to the problem so that the following analysis
is basicly a first-cut attempt to project relatively broad responses to the projected changes in the system based
on the climatological scenarios provided by the Environmental Protection Agency.
4-4
-------
Livingston
jpalachicola
t'nv
lull ol
Figure 1. The Apalachicola estuary showing loeal drainage areas and the river model showing relationships
of detrital flow through the Apalachicola estuary (according to Livingston, 1984a). Also shown are
sampling stations used in the long-term monitoring program.
4-5
-------
Livingston
RUVMOI-FLOW
PRIMARY
PRODUCTION
SALINITY
DISSOLVED NUTRIENTS
DETRITUS ^
substrate
MICROBIAL
?' SUCCESSION
.*
GRAZING
ORGANISMS
TIDAL SUBSIDY
WIND
DISTURBANCE]
INCREASED MICROBIAL
BIQMASS
PHB/LIPIDS
14CO2 RELEASE
O2UTILIZATION
BASE,DETRITAL FOOD WEBS
Figure 2. Model showing relationships of detrital flow through the Apalachicola estuary (according to
Livingston, 1984a).
4-6
-------
Livingston
Table 1. Linear Regression (log/log) of Total Microdetritus (Ash-Free Dry Weight) and Riverflow (m3/sec)
by Month/Year by Season (8/75-4/80)
Station 7 (Surface.)
December-February 0.70 0.49
March-May 0.77 0.60
June-August 0.08 0.01
September-November 0.48 0.23
a (significance)
0.00188
0.00057
0.39863
0.03469
Station 7 (Bottom)
December-February 0.77 0.60
March-May 0.55 0.30
June-August 0.08 0.01
September-November 0.21 0.04
0.00037
0.02253
0.40243
0.22867
Station 8 (Mid-depth)
December-February 0.64 0.40
March-May 0.68 0.46
June-August 0.35 0.12
September-November 0.19 0.04
0.00570
0.00397
0.11809
0.25542
4-7
-------
Livingston
i
$
i
JFMAMJJASONO
JFMAMJJASONO
** .200-1
8 "S
w v
| | «x>-
ii
£ ^ «oo-
1 1
£ | 300-
2 9
t z
1,1
UJ ^^
0J
<
,
I
,1
1
1
1
1
!<
.
^
^
i
CNJ )••
2 •
A "G 12 -
5 OT »-
- 0 •'
.a * ,.
| 1 4.
I a 5 2-
n 0
1 !
I'lliit'r1 i
i • • i
«*
I
&
JFMAMJJASONO
JFMAMJ JASONO
- l600'
"J I4OO-
g 1200-
1 IOOO-
§ < aoo-
1 1 6°°"
« 400-
3 20O-
" 0-
(
1
1
1
(I
1 IT*
1 I i I
1 1
i
20-
•< 19-
.3 *"
c W 16-
•2 *
•^
Z '»
~* 12-
1 I*'
<
i
1
1
'
<
(
i
JFMAMJJASONO
Month
JFMAMJJASONO
Month
K
I
Figure 3. Mean monthly values of numerical abundance and species richness of infaunal macroinvertebrates,
epifaunal macroinvertebrates, and fishes taken in the Apalachicola estuary from 1972-1982.
4-8
-------
3000
2000
1000
Livingston
216
Micropogon undulatus
Anchoa mitchilli
I
Cynoscion
arenarius
Leiostomus xanthurus
hloroscombrus
chrysurus
Bairdiella cnrysura
240
120
MJSDMJSDMJSOM MJSDMJSDMJSOM
TIME-MONTHS (3/72-3/75)
Figure 4. Monthly numerical abundance and mean lengths of dominant fishes in the Apalachicola estuary from
3/72-3/75. Populations include anchovies, croaker, spotted seatrout, spot, Atlantic bumper, and silver
perch.
4-9
-------
Livingston
D
O
2000-
1600-
1200-
800-
400-
cc
UJ
CQ
480-
360-
240-
120-
480-
360-
240
120
0
Penaeus setiferus
"if
Penaeus duorarum
42-
„ s s s
Penaeus aztecus
Palaemonetes
pugio
s s s
Cailinectes sapidus
s s s
Lolliguncula brevis
210
"168
126
84
42
-0
234
•156
-78
•84
• 42
MJ S OMJ S DMJ S DM
MJ S DMJ SDMJ SDM
E
E
X
H
CD
uu
TIME-MONTHS
Figure 5. Monthly numerical abundance and mean lengths of dominant invertebrates in the Apalachicola
estuary from 3/72-3/75. Populations include white shrimp, paleomonetid shrimp, pink shrimp, blue
crab, brown shrimp, and least squid.
4-10
-------
Livingston
OVIGEROUS
FEMALES
DEVELOPING
STAGES
OFFSHORE
CURRENTS
megalops
V k.
'first crab'
Figure 6. Schematic diagram of the life history movements of blue crabs along the Florida gulf coast. Ovigerous
females move north to spawn in the waters offshore from the Apalachicola-Ochlockonee River
systems. After spawning, the young move south, possibly using existing currents in offshore areas.
The young crabs nursery in inshore waters such as the Apalachicola estuary.
4-11
-------
REGIONAL
RAINFALL
Soil Compoiillon
Floodpioin Inundation
EVAPOTRANSPIRATION
VEGETATION DISTRIBUTION
MICROBIAL
SUCCESSION
t
Leaf Production
Decomposition
Fragmentation V
RIVER PEAKS
/ it\
NulrUnt/Ddrltut Flow
Tidal
Wind
Biological
Oitiufbonc*
Subwdy
I LOCAL RAINFALL
Eimro«it V»g«laHon ^^V*
Nulriwili
SUBMERGED VEGETATION PRODUCTIVITY
Salinity Gradltnli
PHYTOPLANKTON PRODUCTIVITY
Conditioning: Porllculot* Organic
DETRITUS- FEEDING
INVERTEBRATES
ASSOCIATED
INVERTEBRATES
OLIGOCHAETE
kMPHIPODS and POL
NKTON—t
\. INVERT!
EEDING
BRATES
RS (Inclu ling man )
DEC.
APR. MAY JUN.
TIME (months)
NOV.
Figure 7. Model showing possible relationships of different variables in the Apalachicola estuary. The model incorporates the various
sources of energy to the system in addition to the various forms of climatological factors that control the biological processes in
a the estuary. The dominant populations are shown during months when organisms reach peak numbers in the estuary.
-------
Livingston
CHAPTER 2
METHODS
GENERAL APPROACHES
Ideally, to develop the appropriate models for an assessment of the potential impacts of projected
climatological changes on the Apalacnicola estuary, we would need to build a comprehensive model of the
entire ecosystem and food web. Such a model would have to involve both time series analysis and simulations
that would incorporate observed changes, temporal dynamics, and species interactions. Such an effort is outside
the scope of this effort. Instead, we have used simple regression and multivariate analyses to evaluate the
responses of various portions of the estuary over 15 years of continuous observation. Thus, this analysis can be
considered preliminary to a more comprehensive modeling effort using more sophisticated modeling techniques.
The Apalacnicola river-estuary has been studied for the past 18 years (15 years of field work, 3 years of
analysis) by scientists associated with the Center for Aquatic Research and Resource Management (Florida
State University). A complete list of the data available for this analysis along with a summary of quality
assurance and analytical protocols is given in Appendix I (appendices are available from the U.S. Environmental
Protection Agency). Background publications are presented in Appendix n. The field data base consists of
monthly collections of important variables taken over a 13-year period (Table 2). The oyster populations were
studied for an additional 2 years. The exact methods of data collection have been given in Livingston (1983,
1984a) and associated papers (Appendix II). Apalacnicola River flow was a primary variable in the evaluation.
During the study, there was a major river flooding episode (1973-1975) and a prolonged drought (1980-1981).
The response of the estuary to these extreme conditions was evaluated relative to background levels of river flow.
The Apalacnicola estuary was the subject of the present analysis which included station-specific data
(Figure 1), grouped stations (East Bay, Apalacnicola Bay), and estuary-wide averages. A series of scattergrams
was developed using the monthly field data. Various transformations were made to determine the best fit for
a normalized distribution when running the different statistical analyses. Such transformed data were used for
all calculations. Moving averages were constructed to eliminate specific aspects of the short-term variability.
Multiple correlation matrices were constructed using all variables. Such correlations were made with and without
lags of key factors. Regressions were used to determine specific relationships among the climatological, physical,
chemical, and biological variables. In this way, certain relationships between and among the different dependent
and independent variables were ascertained.
SCENARIOS FOR EFFECTS MODELING
This analysis required scenarios of future temperature, river flow, and wetland loss. Estimates of wetland
loss due to sea level rise were provided by Park (Volume B). Mains (Volume A) provided scenarios of future
river flow based on climate model projections of temperature, rainfall, and evapotranspiration. In addition, we
developed scenarios of river flow based solely on precipitation scenarios from the models. The climate change
scenarios provided by the Environmental Protection Agency (EPA) were used as inputs to the effects models.
We used the GISS 2CO2 and the GFDL 2xCO2 models for our calculations. Data provided by the EPA were
used (e.g., means by month) for surface air temperature (°C) and precipitation (mm day*1) levels in our region
(GFDL, 82^°W, 3333°N; GISS, 80°W, 35.22°N) (see Table 3). Rainfall data (ambient) were used from
Columbus, Georgia (85°W, 32J°N). The NOAA data from Columbus, Georgia included minimum temperature
(°F), maximum temperature (°F), and precipitation (0.01 inch) on a daily basis from 1/1/48 to 12/31/83. Data
bases were constructed for the entire data set and for data from 1951 to 1980. The period 1/84 to 7/84 was
added from our own data base so that the final numbers could then be applied Table 2.
4-13
-------
Livingston
Table 2. Data Base Used in the EPA Climate Project
I. Climatological features
a. temperature
b. precipitation
c. river flow
II. Physical/chemical factors
a. temperature
b. salinity
c. dissolved oxygen
d.pH
e. color
f. turbidity
g. oxygen anomaly
III. Productivity features
a. microparticulates
b. macroparticulates
1. leaf matter-detritus
2. sea grasses
IV. Biological variables
a. infaunal macroinvertebrates
1. numbers
2. biomass
b. epibenthic macroinvertebrates
1. numbers
2. biomass
c. fishes
1. numbers
2. biomass
d. fisheries
4-14
-------
Table 3. Decadal Monthly Values for GISS and GFDL Models at Georgia Grid Points*
HONTH-
TYPE
GISST
GISSP
GFDLT
GFDLP
1
1.0128
1.0873
1.0122
.7689
2
1.0122
1.2439
1.0127
1.0537
3
1.0218
1.2369
1.0151
1.1465
4
1.0114
1.1783
1.0184
1.1493
5
1.0149
1.1392
1.0127
.8283
6
1.0128
1.3120
1.0209
.4408
7
1.0102
1.4133
1.0245
.6514
8
1.0073
1.2309
1.0148
.4257
9
1.0173
.7970
1.0206
.8082
. 10
1.0183
.9699
1.0161
.6737
11
1.0128
.7464
1.0190
1.0781
12
1.0097
.9169
1.0190
1.0716
T
P
temperature
precipitation
* Values for Columbus, GA (85°N, 32.5'W)
-------
Livingston
to our long-term (Apalachicola) data base, which ran from March 1972 through July 1984. The calculated
adjustment factors for the GISS and the GFDL models were then applied to the data for the predicted effects
of the doubled CO2 scenarios. There was a different adjustment factor for each month in the respective models
for the surface air and precipitation values:
Temperature: [(average TX 273.15)Xadj] - 273.15
Precipitation: total precipitationXadj.
The ambient data (precipitation, temperature maxima and minima) were then compared with the projected
(GISS.GFDL) results for the different time periods (1/1/48-12/31/83, 1/1/51-12/31/80, 3/1/72-7/31/84).
The GISS precipitation model predicted generally higher levels of precipitation, whereas the GFDL
precipitation model predicted somewhat less. The study period (3/72-7/84) appears as part of an overall
(long-term) decline in precipitation with the drought of 80-81 shown clearly. When viewed as differences in the
respective moving averages, the GISS model predicts average increases of a little less than 2 cm/month whereas
the GFDL model predicts average decreases of just under 2 cm/month.
Using the previously developed regressions Unking precipitation to Apalachicola River flow at Blountstown,
Florida, we used the various data projections to calculate new river flows (based on the GISS and GFDL
scenarios). The regressions were also used to calculate the salinity regimes in the Apalachicola estuary over the
period of observation (3/72-7/84). The various relationships (e.g., correlation, regression) were then calculated
for the different physical, chemical, and biological variables (Table 2). Regressions were applied to the model
river data to get projected river flows according to the EPA models. Such models were then applied to the
various biological variables taken hi the Apalachicola estuary (3/72-7/84).
The various biological indices (e.g., population, community) were then analyzed in the existing data base
with particular attention to responses to changes in river flow, salinity, and temperature. Such data were
supplemented with reviews of the literature concerning temperature ranges and optima for estuarine dominants
and commercially valuable species. Using the above-outlined scenarios of climate change and the combination
of quantitative and qualitative analyses of biological response to such variables, a series of tentative conclusions
was outlined relative to the ranges of predicted events.
4-16
-------
Livingston
CHAPTERS
RESULTS
RIVER FLOW SCENARIOS
An application of the models to the Columbus, Georgia precipitation data over the 13-year study period
indicated differential seasonal differences. The GFDL model predicted considerable reductions in rainfall that
are most evident during January and summer-fall periods. The GISS model predicted general increases in
precipitation that approach 8 cm in extreme cases. The overall trends showed differences approaching or
± 2 cm in the two models. There is a pronounced reduction in the differences in the GFDL model during the
latter half of the record, which reflects general reductions of Georgia precipitation.
The Apalachicola River flow data (Figure 8) show the extreme high and low level years quite clearly.
Generalized peak flows were most consistent during 1973; the overall drought months (20) occurred during the
period 1980-1981. These overall trends are evident in the 12-month summed data. Peak flows occurred during
1973; this was followed by an overall decline after a secondary peak in 1975. This decline ended in the drought
year 1981, which was followed by a moderate increase in 1983-1984. During the 24-month period of 1980-1981,
15 months had monthly river flow averages that were below 500 m3; this was the most prolonged drought during
the period of observations (3/72-7/84). The seasonal distribution of such flow conditions showed that, with the
exception of the spring of 1979 and 1980, the overall decline of Apalachicola River flow from 1973 to 1981 was
real. Summer-fall periods varied the least with the general trend following the winter-spring flows.
A statistical examination of the relationship between Apalachicola River flow and rainfall (Columbus,
Georgia; Apalachicola. Florida; East Bay, Florida) indicates that only Georgia rainfall is significantly associated
with such river flow (r= 0.207, p < 0.0001). There is a close relationship of Georgia rainfall and river flow with
some evidence of an averaged lag period of 3 months between the Columbus, Georgia, precipitation and the
Blountstown, Florida, Apalachicola River flow. Using the established regression between these two variables,
the GISS and GFDL models were run to determine the new river flow conditions under the projected
precipitation change. The GISS model predicted an average increase in river flow conditions from the observed
(from 722.277 to 827307 m per second). The minimal levels were higher, whereas the maximum levels were
lower. The GFDL model, however, predicted an average decrease (from 777369 to 722.277 m3 per second).
Minimal values increased, while maximal values were less. In both cases, the extremes were moderated with less
high flow in the winter-spring months and higher flow rates during the summer-fall months.
The mean salinities for the Apalachicola estuary (Figure 10) show trends that are related to the above
described long-term changes in meteorological conditions. During the 1980-1981 period, 10 mean salinity points
were above 15 ppt, the highest such number in the entire survey. The spring flooding in 1979 and 1980 was
particularly effective in driving down the bay salinity levels. A 12-month moving average of the data (Figure 11)
shows the general decline in salinity prior to the major increases during the drought of 1980-1981. The spatial
distribution of such salinity during the drought showed considerable vertical stratification and distinct differences
between the oligohaline East Bay and the meso- to polyhaline Apalachicola Bay (Livingston, 1984a). Such
differences are crucial to an understanding of the biological organization of this estuary.
The GISS and GFDL models were run with the Apalachicola River flow regressions (Table 4). According
to the GISS model, mean salinity (overall) is decreased from 8.9 ppt to 8.4. Model projections of minimum and
maximum salinities are considerably higher and lower, respectively, when compared to the actual data. According
to the GFDL model, mean salinity increases from 8.9 ppt to 92 ppt. The relationship of the minimum and
4-17
-------
Livingston
5000
1972 1973 1974 1975 1976 1977 1978 1979 1980 1981 1982 1983 1984
YEAR
WINTER
SPRING
SUMMER
FALL
Figure 8. Annual quarterly (seasonal) means of Apalachicola River flow from 1972-1984. Data show peak levels
during 1973-1975 and the drought during 1981-1981
4-18
-------
Livingston
y « -2.913X * 10.996, R-squared: .461
f
T3
M
•3
f
2.6 2.8 3
tog(1+x) of river flow
3.2
3.4
Figure 9. Regression of the square root of salinity on the log (x+1) Apalachicola River flow (monthly data;
1972-1984). Salinity represents monthly means of all stations in the estuary, surface, and bottom.
River'flow data include monthly means taken at Blountstown, Florida. There is a strong inverse
relationship between these two variables.
4-19
-------
APALACHICOLA ESTUARY: TOTAL
Livingston
• J« <• «• •• !•• 12* 14*
MONTH
Figure 10. Average salinity data taken monthly in the Apalachicola Estuary from 3/72-7/84. Monthly means
of all stations (surface and bottom) were used for this analysis.
APALACHICOLA SYSTEM: MA-12
I
e
z
14
12
10
1200
1000
400
O
HI
9.
DC
UJ
E
— SALJIA-12
-*• R1VERI1A-12
200
ooooo*oooooooo«
TIME IN MONTHS
Figure 11. Twelve month moving average of river flow and average salinity data taken monthly in the
Apalachicola estuary from 3/72-7/84. River slow data are monthly mean values taken at
Blountstown, Florida. Salinity data include monthly means of data taken from all estuarine stations
(surface and bottom).
4-20
-------
Livingston
Table 4. A: Regression Statistics Concerning the Relationship of Apalachicola River Flow and Mean Salinity
Values in the Apalachicola Estuary
Mean Std. Dev. Min. Max. Ranoe # cases
River flow 8.9 5.7 0 25.0 25.0 149
GISS 8.4 2.1 2.2 12.0 9.8 149
GFDL 9.2 1.9 2.7 12.0 9.2 149
B: Regression Statistics Concerning the Relationship of Apalachicola River Flow and Salinity
According to Hains' GISS, GFDL, and OSU Projections
OSU
Data projections are made monthly from January, 1948
through July.1984.
Nat, flow 9.9
GISS
Mean Std. Dev Min.
9.9
10.
11.
10.
2
0
6
3.
3.
3.
2.
1
0
1
5
0
0
0
2.1
Max.
1
1
3.
3.
14.
1
3.
7
9
3
6
Ranae
1
1
3.
3.
14.
1
1.
7
9
3
5
# cases
372
372
372
372
4-21
-------
Livingston
maximum salinities parallels that described for the GISS model. In both models, the peaks and valleys of the
salinity levels are largely eliminated. Such results are obvious artifacts of the methods (e.g., regression) used to
arrive at the final projections.
River flow projections according to the Hains models indicated a 24.2% decrease in river flow projected
by the GFDLx2 model; lessor decreases were noted with the GISSx2 and OSUx2 models. Such changes in river
flow would result in a maximum (GFDLx2) increase of 113% of the mean salinity in the Apalachicola estuary
(Table 4). Such a change would result in generally higher salinities over tune with such increases particularly
evident in the summer-fall highs. These differences approach a range of + /-10 ppt in a given month. The
differences can go either way over the tune of projection although most such differences are positive.
Water Quality Changes
Various water quality factors in the Apalachicola estuary are strongly affected by river flow conditions
(Livingston, 1984a). Mean turbidity values for the estuary reflect the major river flooding during the
winter-spring of 1973. The decreased turbidity in the following years indicates reduced flows; the same is true
for the latter portions of the 1980-1981 drought. Color levels follow a different pattern with the lowest levels
noted during the 1980-1981 drought. The long-term pattern of Secchi data reflects the above-noted episodes of
river flow; the highest Secchi readings (e.g., the clearest water) were noted during the drought period. The
oxygen anomaly (AO2), denoted as excesses of dissolved oxygen above or below solubility, has been associated
by some authors with productivity.
AO2 = O2 - O2 (calculated from the equation of Weiss 1970 as the difference between the measured
oxygen and the oxygen solubility at ambient temperature and salinity)
Analyses of the oxygen anomaly in the Apalachicola estuary (Justic and Livingston, in review) indicate a
weak positive association with temperature and a weak negative association with river flow in East Bay and
Apalachicola Bay. Relatively high correlations were noted in surface waters hi both systems between the oxygen
anomaly and salinity; lesser such correlations were noted at the bottom. This would indicate possible increases
in phytoplankton activity during periods of high temperature, low river flow, and high salinity. Such trends
would be consistent with known patterns of phytoplankton productivity in the estuary (Livingston, 1983,1984a).
This would indicate that reduced river flow could be associated, at least during the short term, with increases in
phytoplankton productivity.
River-driven estuarine systems such as the Apalachicola are often dependent, to a considerable degree, on
organic detritus that comes from upland wetlands and submerged aquatic vegetation. From January 1976 through
March 1982, a survey was conducted to determine the relative amounts of POM that were occurring in the bay.
Water from the lower end of the river was pumped through a series of sieves and analyzed for organic carbon
(total particulate organic carbon; sized fractions of POM from 2 mm to 0.1 /im). Such matter forms the substrate
for microbial activity that is directly linked to the detritus-based food webs of the bay (Livingston, 1984a,
unpublished data with D.C. White). The regional rainfall (Georgia and Florida) along with the Apalachicola
River flow was low from June 1980 through December 1981. Although concentrations of TPOC were high during
certain months of the drought, the overall mass transport of TPOC was relatively low throughout this period
Such mass flows are directly related to river flow levels. Most of the POC coming down the river is composed
of very fine particulate organic carbon (VFPOC). The TPOC (concentration and mass flow) was negatively
associated with temperature and positively associated with river flow. In East Bay during the drought, salinity
was uniformly high. Color was low and Secchi readings were high, and the peak oxygen anomaly readings
occurred at this time. The bay salinity was very high during the drought; in most months, mean levels exceeded
20 ppt. Once again, Secchi readings were high and color was low. The POM was particularly low in all its forms
during the drought. A summary of these data (Appendix VI) indicates that all forms of POM except the SAV
were lower during the drought than at other times when rainfall/river flow rates were higher.
4-22
-------
Livingston
The data indicate that, during periods of low rainfall and river flow, the bay becomes clearer, more saline,
and has less POM derived from the winter/spring winter flooding. Such conditions could lead to short-term
increases in phytoplankton productivity and the proliferation of submerged aquatic vegetation. When viewed
on the basis of overall productivity in the bay, there would be a net increase in the spring to fall productivity at
first. However, as nutrients became limiting, there would probably be a trend toward lower phytoplankton
production and higher SAV production with conditions resembling the seagrass systems of Apalachee Bay (see
Livingston, 1986, for a review of the long-term conditions in this system). Those forms that utilize the estuary
in the winter-early spring months (sciaenid fishes) or are dependent on phytoplankton productivity (oysters)
would be adversely affected by such conditions. Other species, such as the spotted seatrout (Cvnoscion
nebulosus) and other forms that depend on the seagrass beds would be favored. The data indicate that if river
flows are reduced, the major commercial fisheries of the Apalachicola Bay system would be adversely affected
whereas the sports fisheries associated with increased seagrass beds would be enhanced. Prolonged increases
in regional rainfall and river flow would have the opposite effect.
Long-Term Biological Changes
Salinity is one of the most important ecological variables in estuarine systems. Low and variable salinities
tend to reduce the number of offshore species that can live in the estuary; this makes this kind of system a haven
or sanctuary for those species that are euryhaline. The complex community structure and food web patterns that
are associated with the high productivity and salinity relationships tend to explain why estuaries are important
centers of sports and commercial species. However, such complexity precludes the direct use of
correlation/regression analyses to predict specific relationships with estuarine habitat features and important
forcing functions. Thus, the following is a preliminary approximation of such relationships since it is based
primarily on such statistics.
The overall trends of the infaunal macroinvertebrates are shown in Figure 12. Such data indicate that both
species richness and numerical abundance vary together (r2* 0.603; p < 0.0001). When viewed as a function of
river flow (Figure 12), the increases in both indices are associated with the drought period even though the
general trends are not significantly correlated. Salinity does not appear to be a major factor in these trends, at
least directly; increased salinity could have an indirect effect by allowing predators into the area. In a series of
experiments (Mahoney and Livingston, 1982; Livingston, in press; Livingston, unpublished data), it was shown
that direct predation has little impact on the infaunal associations in East Bay. There is some indication,
according to these analyses, that the East Bay infaunal assemblages are physically controlled by temperature and
river /rainfall conditions. The considerable drop in numbers after the drought could be an indication of a delayed
response to the reduction of detrital matter associated with the drought.
The epibenthic invertebrates represent some of the most important species in the estuary in terms of their
commercial value. The two top dominants are the white shrimp (Penaeus setiferus) and the blue crab
(Callinectes sapidus) which together with the oyster (Crassostrea virginica). provide the basis for a multimillion
dollar industry in the region. Invertebrate numbers were highest during the years of highest rainfall, whereas
invertebrate biomass was highest during the 1980-1981 drought (Figure 13). The penaeid shrimp tended to have
peak numbers during periods of high river flow (Figure 14), whereas the blue crabs were in the highest numbers
during the drought (Figure 14). The regressions of invertebrate indices with physical/chemical factors (Appendix
VI) indicate reductions of numbers and biomass with increasing river flow, POC influx, and decreasing salinity.
The fish associations of the Apalachicola estuary are well developed with numerical dominants including
the bay anchovy (Anchoa mitchilli). spot (Leiostomus xanthurus). croaker (Micropogonias undulatus). and white
seatrout (Cvnoscion arenarius). The numbers of such species usually peak during winter-early spring periods.
The 13-year trend (Figure 15) indicated the highest numbers of fishes during the 1980-1981 drought, whereas
fish biomass was highest during and immediately following the period of high river flooding. Fish biomass
tended to follow the long-term trends of river flooding with particularly low biomass before,during, and after the
drought of 1980-1981. Fish species richness followed a cyclical pattern with peaks in 1975 and
4-23
-------
30
Livingston
Cd
O
U
M
ta
O
&
Cd
20-
INFAUNAL NUMBER OF SPECIES
INFAUNAL NUMBERS
n i
HA"^
200
100
C/5
«
U
a
z
20 30 40 50 60 70 80 90 100 110 120 130 140 ISO 160
MONTHS
100
Cfl
C£
W
ea
2
z
fa.
Z
80-
60
40-
20-
INFNMAI2
RIVERKLOWMA12
1200
1000
•800
600
400
U
Cd
I
fig
U
200
40 SO 60 70 80 90 100
MONTHS
110 120 130 140 150
Figure 12. (a) Species richness and numerical abundance of infaunal macroinvertebrates in East Bay (all
stations) taken monthly from 3/75-7/84. Data represent summed data. Peak levels occurred during
the drought months and low levels were reached during the heat following the drought, (b) Moving
averages of infaunal macroinvertebrate numerical abundance and species richness from 3/75-7/84.
4-24
-------
APALACHICOLA ESTUARY: TOTAL
Livingston
20 40 60 10 100 120 14»
MONTH
100
g
oc
CD
QC
UJ
z
o
o
y = 10.0263 + 0.513x - 0.0096x*2 + 4.369e-5x*3 fl = 0.27
10 10* 120 140
MONTH
Figure 13. (a)Trawl-susceptible invertebrates (numbers, dry-weight biomass) taken monthly in the Apalachicola
estuary (3/72-7/84). Monthly samples were summed Log data were also analyzed using third order
polynpial distribution showing a trend of abundance during the early years with low levels reached
during the drought of the early 1980*8. (b) Trawl-susceptible invertebrates Gog-transformed) taken
monthly in the Apalachicola estuary 3/72-7/84.
4-25
-------
APALACHICOLA ESTUARY: TOTAL
Livingston
a.
E
a
UJ
ui
a.
I
y . 1.5504 » 0 4759x • 0.008«*2 + 3.51 5o-5x*3 R • 0.21
15
•10
a
i
5
a
ui
w
a
~ 4
CO
E
UI
ffi
3 3
u
UI
II* 111 141
— TOTPEN-UA13
• II 4« M M IM 111 141
•*• CALSAPSH-9
MONTHS
Figure 14 (a) Numbers of total penaeid shrimp taken monthly in the Apalachicola estuary (3/72-7/84). A third
order polynomial distribution was used to show the trend of high numbers during the early years of
the program. Also shown are moving averages of shrimp numbers (13-month) and moving averages
of blue crab numbers (5-month) over the period of study, (b) Moving average of numbers of total
penaeid shrimp taken in the Apalachicola estuary, 3/72-7/84. (c) Numbers of total blue crabs taken
monthly in the Apalachicola estuary, 3/72-7/84.
4-26
-------
APALACHICOLA ESTUARY: TOTAL
APALACHICOLA ESTUARY: TOTAL
1500
1000
in
E
500
I* II* II* 14*
300
<-. 200
O
100
FN/TWL-SM13
MONTH
O
ffi
(/)
u.
at
<
O
m
y = 41.9058 + 1.6468x - 0.0334x"2 + 1.556e-4x*3 R . 0.26
-*• MEAN B TH
20
MONTH
Figure IS. (a) Fish numbers taken monthly in the Apalachicola estuary from 3/72-7/84. Mean numbers summed over all stations were <•
used, (b) Moving average of fish numbers taken monthly in the Apalachicola estuary, 3/72-7/84. (c) Total fish dry weight 5'
biomass taken monthly in the Apalachicola estuary, 3/72-7/84. (d) Moving average of total fish dry weight biomass taken *S
monthly in the Apalachicola estuary, 3/72-7/84. §
-------
Livingston
1984. These cycles tended to follow a very complex series of climatologjcal changes and are currently under
study. Low levels of species richness occurred during the drought. The statistical basis for the above
relationships is given in Appendix IV. These results, while extremely complex, are consistent with known patterns
of fish distribution in the Apalachicola system. Fish numbers are high during winter-early spring periods of high
river flow, high POC, and organic carbon input from the river. By the summer-early fall months of high
temperature and low dissolved oxygen, fish numbers decrease but biomass per trawl tow increases with seasonal
growth and the arrival in the estuary of the larger predators. Fish species richness increases at this time. A
multiple correlation matrix (with and without lags, Appendix IV) tends to corroborate this interpretation of the
data. The various relationships shown with the lagged data indicate that a time series analysis may be more
effective in delineating the cyclic associations of these variables. This is currently under study.
A review of the long-term climatologjcal changes in the tri-river system and changes in the important
fisheries of the Apalachicola estuary is given by Meeter et al. (1979). A spectral analysis of long-term trends
of Apalachicola river flow and Georgia rainfall indicated cycles on the order of 6-7 years with possible longer
term cycles that may reflect variations such as the low river flow during the mid-1950's. Cross-spectral analyses
of these two variables showed that river flow and precipitation in Georgia are in phase. When commercial
landings taken in the Apalachicola estuary are compared with the long-term trends of annual river flow, there
were diverse results. Penaeid shrimp catches were not correlated (p <0.05) with any of the other variables. Blue
crab catches were positively correlated with river flow and negatively correlated with oyster catches. The oyster
landings showed negative correlations with river flow and Apalachicola rainfall.
Such data reflect complex biological processes and socio-economic events which cannot be accounted for
by use of the relatively simple statistics of correlation and regression. The considerable seasonal variations of
the various commercial populations may mask the possible associations of such catches with longer term
climatological changes and cycles. Optimal oyster growth occurs at salinities ranging from 15.0 to 225 ppt;
during periods of high salinity, oysters are preyed upon by various stenohaline organisms that are usually excluded
from the estuary by the low and varying salinities. However, extended periods of low salinity can have an adverse
impact on oyster growth. Young blue crabs have a preference for low salinity, whereas the euryhaline white
shrimp have no overt relationship to salinity. All such populations respond to complex combinations of variables
such as river flow, temperature, salinity, productivity, predation, and other factors. There is little doubt that
without the input of the Apalachicola River, these fisheries would largely disappear. However, the direct and
indirect effects of changes in regional precipitation and river flow on the local fisheries are still not entirely
clear owing in large part to the socio-economic variables that influence the commercial catch statistics.
Projections
A decrease in river flow of 24.2% (Hains' estimated GFDLx2 scenario) would cause direct losses of
nutrients and POC that eventually would lead to proportionate losses in the detrital- and phytoplankton-based
food webs of the estuary. A possible impact scenario for fishes is given in Figure 16, where the regressions
determined during the 13-year field studies are applied to the Hains GFDLx2 scenario. Fish numbers were
reduced in excess of 50% due to the loss of river flow according to this scenario. Once again, such projections
are mere approximations owing to the problems associated with the statistical methods. If invertebrates are
considered, after a brief increase of invertebrate numbers due to increased salinity and reduced predation, there
would be an eventual reduction of infaunal and epibenthic macroinvertebrates with shifts in species representation
as the estuary became less turbid and more saline. The reduction of food and increase of predators would lead
to reductions of the oyster productivity; more oysters would be eaten by offshore stenohaline predators and there
would be an increased toll due to oyster diseases that often accompany higher salinities. Such changes would
probably be proportional to the change in river flow over the prolonged period of adjustment. Thus, the
applications of the various scenarios of river flow changes could lead to proportional changes in the productivity
of the estuary.
4-28
-------
300
Livingston
FN/TWL-NAT RIVER
FWTWL-GfDLX2
100
200
400
100
SO
UJ
w
<
UJ
cc
o
ui
o
o-
-50-
-too
•
if. *•*•/*.*,.*
*•• V • • • • »
v-.j /•.{.•/. v/'
100
200
MONTHS
—i—
300
400
Figure 16. (a) Comparison of numbers of fishes taken per trawl under natural river flow conditions and
numbers calculated from regressions of the GFDL river flow projections. Data are based on monthly
averages of fishes taken from all stations in the Apalachicola estuary from 3/72-7/84. Also shown
are the percent decrease of ratios of fish numbers taken under actual conditions and numbers
produced from regression of GFDL projections of river flow, (b) Comparison of numbers of fishes
taken per trawl under natural river flow conditions and numbers calculated from regressions on the
GFDL river flow projections. Figure is presented as the percent decrease of fish numbers.
4-29
-------
Livingston
TEMPERATURE
Scenarios
The EPA scenarios (GISS, GFDL) were run with ambient data taken within the tri-river drainage system.
Comparisons were made during the 13-year period of study (3/72-7/84). Previous studies (Livingston, 1984b)
indicated that monthly average air temperatures in Apalachicola, Florida (NOAA station) are correlated (r =«
0.90, p < 0.00001) with water temperatures in the shallow estuaries along the north Florida gulf coast. The use
of air temperature as a predictor of estuarine water temperature in the Apalachicola Bay system is thus valid.
However, unlike river flow and salinity, no regressions were run with the biological data. The temperature
changes would affect the bay thoughout any given year; however, the summer monthly temperatures would reach
lethal levels which differ from species to species. This is a threshhold effect that is difficult to quantify by the
use of correlations and regressions. Instead, the projected changes in temperature were calculated according to
the transformation of existing data by use of the various models. A literature search gave the various
temperature tolerances of the important estuarine species in the Apalachicola estuary. From this analysis, a
frequency diagram was constructed giving the projected changes in the estuary based on the temperature
tolerances of indigenous species.
The ambient (bay-wide mean) water temperature trends over the period of study (Figure 17) indicated
that winter low temperatures showed progressive declines from 1975 to 1977 at which time the lowest
temperature levels of the period were observed. From 1977 to 1980, there was a progressive increase in low
winter temperature levels. A time-series analysis of air temperatures in Apalachicola from 1930 to 1979
(Livingston, 1984b) showed that the lowest monthly minimum air temperatures occurred during winter months
in 1940, 1948, 1958, 1968, and 1977. The approximate period of such low winter temperatures appears to be
within a range of 8-10 years. Summer high mean temperatures varied less widely with a range approximating
28-31°C. The highest summer mean temperatures in the Apalachicola system occurred in 1972,1976,1977,1980,
and 1983. No long-term patterns of recurring high summer temperatures were evident.
A comparison of the ambient temperature vs the GISS transformation showed relatively little difference
in the summer high temperatures (Figure 18); winter lows would increase somewhat more in terms of absolute
increases. The GFDL model, however, showed substantial increases in the summer highs. Such differences
range between 5 and 9°C (Figure 18). According to the GFDL model, summer high average temperatures in
the Apalachicola estuary would range between 15 and 16°C.
Biological Response
Thermal stress operates at both ends of the temperature spectrum. However, because the high
temperatures (summer) along the upper gulf coast approach the upper lethal levels of various estuarine species,
and because the GISS and GFDL projections indicate general increases in such temperature levels, the upper
lethal levels will be emphasized in this review.
There is a considerable literature concerning the effects of temperature on poikilothermk (aquatic)
organisms (Collins, 1985; Darcy, 1985; Lassuy, 1983; Muncy, 1984; Reagan ,1985; Reagan and Wingo, 1985;
Stanley and Sellers, 1986; Sutler and McHwain, 1987;Van Den Avyie, 1984). Brett (1970) and Warren (1971)
have reviewed various aspects of the effects of temperature on fishes and invertebrates. Temperature is a major
environmental determinant of the distribution of aquatic organisms in space and time. Factors such as
acclimatization and adaptation complicate the interpretation of direct effects of temperature on the distribution
of aquatic organisms. Temperature tolerance is genetically controlled with modifying features including salinity,
photoperiod, season, size, diet, reproductive state, and the acclimatization history of the population in question.
The rate of temperature change along with the exact nature of the state variable in question (e.g., respiration,
behavior, reproduction, mortality) and the thermal history of the subject all interact to determine the mechanism
4-30
-------
Livingston
APALACHICOLA ESTUARY: TOTAL
o
,0,
UJ
DC
UJ
Q.
UJ
20
60
80
100
120
140
MONTH
Rgure 17. Mopthly temperature changes in the Apalachicola estuary from 3/72-7/84. Numbers are based on
mean values taken from all stations (surface and bottom) in the Apalachicola estuary.
4-31
-------
Livingston
U
o
Cd
«
3
<
at
cd
cd
H
U
o
Cd
d
I
cs
Cd
eu
Cd
40
30
10-
10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
0 10 20 30 40 50 60 70 80 90 100 110 120 130 140 150
MONTHS
Figure 18. Comparisons of ambient water temperature and (a) GISS transformations and (b) GFDL
transformations (projections) of water temperature in the Apalachicola estuary (3/72-7/84). Data
are based on monthly means of all stations.
4-32
-------
Livingston
of thermal tolerance in a given population. Indirect effects of associated variables such as dissolved oxygen (e.g.,
temperature is a primary determinant of the level of dissolved oxygen in water) complicate the direct and indirect
effects of temperature on aquatic organisms.
A review was made concerning the upper thermal collection limits (e.g., the highest temperature at which
an aquatic species was taken in the field) and the upper critical or incipient lethal temperature (e.g., that
temperature level which will eventually cause the death of a stated fraction of the test organisms, usually 50%)
(Table 5). There are species-specific trends with organisms such as the lutjanids having relatively low upper
thermal collection limits. Important species in terms of abundance and biomass in the Apalachicola estuary
include oysters (upper thermal collection limit or U.T.C. of 36° C), blue crabs (upper thermal tolerance or TJ.T.T.
= 30-33°C), penaeid shrimp (U.T.C. = 35-35.5°C), stone crabs (U.T.C. = 32°C), seatrouts (U.T.C. = 34.9-35°C),
catfishes (U.T.C. = 34.9-37°C), mullet (U.T.C. = 345*0, and southern flounder (U.T.C. = 35°C). Between
the temperatures of 30-37S"C (usually not exceeding 35°C), there could be a progressive reduction of the
dominant estuarine species found during summer periods in the Apalachicola Bay system.
An analysis was made of the major dominant species in the Apalachicola estuary with particular emphasis
on commercially important species (Tables 6 and 7). According to the GFDL projections of summer high
temperatures, blue crab larvae and juveniles, spotted seatrout, oyster larvae, pinfish, flounder, and croaker would
be virtually eliminated. Severe impacts would occur with the redfish population. In the GISS model projections,
only blue crab larvae and juveniles would be eliminated, although spotted seatrout, oyster larvae, pinfish, and
flounder would be severely affected. Under current (ambient) conditions, only blue crab larvae have been
adversely affected by high summer temperatures; since larvae move into the bay mostly during winter periods,
such an impact would be hardly felt. In any case, the GFDL and GISS projections of temperature increases
would have a major adverse impact on the estuary.
The general increase of winter low temperatures would also have a profound influence on the species
composition of aquatic organisms in the Apalachicola estuary. The biota would probably become more tropical
in terms of temperature-tolerant species with general increases in species richness and diversity and general
reductions of overall fish and invertebrate production in terms of dry-weight biomass. Such projections would
be strongly influenced by potential (co-existing) changes in other variables such as salinity and dissolved oxygen.
According to Yanez-Arancibia et al. (1985), tropical estuaries in the Mexican portion of the Gulf of Mexico have
the highest phytoplankton productivity in areas receiving high levels of river flow. Thus, the overall production
of the Apalachicola estuary would depend to a considerable degree on changes in the mass flow of freshwater
into the estuary regardless of the ambient temperature of the system. Commercial fish catches in Mexican
estuaries are directly dependent on the river discharge. Coastal wetlands are also an important component in
such productivity (Yanez-Arancibia et al., 1985).
Overall, an average increase of 4-5°C in the Apalachicola estuary would probably cause a major disruption
of the aquatic fauna at both ends of the temperature spectrum. It is doubtful that there would be enough time
for species tolerant to the high temperature and accompanying low dissolved oxygen to adapt to such conditions,
leaving only the organisms already tolerant to the projected temperature increases. When other more restrictive
functions such as growth and reproduction are taken into consideration, the impact of the projected temperature
increases in the Apalachicola region become even more meaningful since few organisms are adapted to maintain
their population numbers at temperatures higher than 35-36°C. The major break point in the Apalachicola
estuary would be sustained summer maximum temperatures of 35°C and higher.
The question of acclimatization is difficult to address. Temperature increases are projected to occur over
a relatively short time period. There could be some adaptive response of existing species, although this is
unlikely since there is some evidence (Livingston, unpublished data) that many of these species are already
temperature limited. This is not a matter of succumbing to a single factor, since the effects of such temperature
changes would result in habitat alterations that would simply render the estuary unfit for certain species.
Another complicating factor would be the synergjstic effects of temperature and other state variables such as
4-33
-------
Livingston
Table 5. Thermal Tolerance Levels of Selected Estuarine Species*
Reference)
Species
Upper Thermal
Collection Limit
Upper Thermal Tolerance
(Upper Critical or Incipient Lethal
Temperature)
INVERTEBRATES
1) Caltinectes sapidus
2) Menippe mercenaria 32°C
3) Palaemonetes pugio 38° C
4) Penaeus douranun 35S"C
5) Penaeus setiferus
6) Penaeus axtecus 35° C
7) Crassostrea virginica 36°C
FISH
8) Anchoa milchelU 39.8°C
8) Anchoa hepsetus 34.9° C
9) Archosargus probatocephalus 3S.1°C
VS)Ariusfetis 37DC
10) Bagre marinus >34.9°C
11) Brevoortia patronus 34.9°C
12) Cynoscion arenarius 35°C
13) Cynoscion nebulosus 34.9°C
12) Cynoscion nothus 30°C
14) Fundulus heteroclUus 40-42°C
14) Fundulus majalis 40-42°C
15) Lagodon rhomboides 33-34°C
16) Lutjanus analis 27.8°C
16) Lutjanus griseus 32J°C
16) Lutjanus synagris 28.9°C
17) Micropogonias undulatus
18) A/ugi/ cepAoftu 34.5°C
16) Ocyurus chrysums
19) Orthopristes chrysoptera 36°C
20) Paralichtys lethostigmata 35°C
21) Pogonias cromis 33°C
22) Sciaenops oceUatus 375PC
30°COarvae)
33°C (juveniles)
38°C
3S°COarvae)
42°C
48°C
>3S°C inhibits larval growth
near39°C
35-36°C avoided
36°C
38°C
413°C (juveniles)
* Data are given as the Upper Thermal Collection Limits and as the Upper Thermal Tolerances.
Refemces and significance of the numbers are given in the text.
4-34
-------
Livingston
Table 6. Distribution of Upper Thermal Tolerances of Dominant Species in the Apalachicola Estuary from
1972-1977. Temperatures that Eliminate Species from the Bay are Listed Along With Projections
that Would be Eliminated According to Different Scenarios During the Summer Months from 1972-
1977 (Months Start in June of Each Year)
AVERAGE MAXIMUM MONTHLY TEMPERATURE
°C °C °C
DATE
summer
1972
1973.
GFDL
37.
39.
37.
38.
37.
40.
3
6
6
1
7
6
36.2
1974
1975
1976.
37.
36.
40.
36.
38.
37.
36.
6
5
1
4
0
9
7
36.6
40.0
. ELIMINATE GISS
bcl.bcj.
St.p.C.f
bcl.bcj,
st.p.c.r
bcl.bcj,
st.p.c.r
bcl.bcj,
st, p.c,
bcl.bcj.
st.p.c.r
bcl.bcj,
st.p.c.r
bcl.bcj,
st.p.c.f
bcl.bcj.
st.p.c.r
bcl.bcj.
st.p.c.f
bcl.bcj.
st.p.c.r
bcl.bcj.
st.p.c.f
bcl.bcj.
st.p.c.r
bcl.bcj.
st.p.c.r
bcl.bcj,
st.p.c.f
bcl.bcj,
st.p.c.f
bcl.bcj.
ol
ol
,f
ol
,f
ol
r.f
ol
,f
ol
,f
ol
ol
,f
ol
ol
,f
ol
ol
,f
01
,f
ol
ol
ol
34
35
35
37
35
36
33
36
.8
.2
.3
.1
.2
.2
.9
.6
33.4
34
35
34
.1
.7
.1
33.2
35
34
35
.5
.1
.6
st.p,c,r,f
1977
36.
35.
4
1
33.0
34.6
40.9
42.5
bcl.bcj.
st.p.c.f
bcl.bcj.
st.p.f
bcl.bcj
bcl.bcj
bcl.bcj.
st.p.c.r
bcl.bcj.
ws.st.p
m.r.f
ol
ol
ol
,f
ol
,c
34
34
35
38
38
.1
.1
.2
.4
.1
35.2
ELIMINATE ACTUAL
bcl.bcj
bcl.bcj,
st.p.f
bcl.bcj.
st.p.f
bcl.bc).
st.p.c.f
bcl.bcj,
st.p.f
bcl.bcj.
st.p.c.f
bcl.bcj,
bcl.bcj.
st.p.c.f
bcl.bcj
bcl.bcj
bcl.bcj.
st.p.f
bcl.bcj
bcl.bcj
bcl.bcj,
st.p.f
bcl.bcj
bcl.bcj,
st.p.f
bcl.bcj
bcl.bcj
bcl.bcj.
st.p.f
bcl.bcj,
st. p. r.f
bcl.bcj.
st.p.c.r
bcl.bcj.
st.p.f
ol
ol
ol
ol
ol
ol
ol
ol
ol
ol
ol
ol
.f
ol
30
32.
33
31
.9
.1
.1
.9
31.3
33.1
31.
31.
28.
33
32
30
32
28
30
32
31.
28
30
34
35
33
.7
3
9
.1
.6
.3
.5
.9
.3
.5
9
.9
.7
.5
.0
.0
ELIMINATE
bcl
bcl
bcl.bcj
bcl
bcl
bcl.bcj
bcl
bcl
bcl.bcj
bcl
bcl
bcl
bcl
bcl
bcl
bcl
bcl.bcj
bcl.bcj.ol
P.f
bcl.bcj
SPECIES UPPER THERMAL TOLERANCE (QC)
blue crab larvae (bcl) 30 (TEMPERATURES THAT ELIMINATE
blue crab juveniles (bcj) 33
white shrimp (ws) 42
oyster larvae (ol) 35
spotted seatrout(st) 34.9
pinfish(p) 35
croaker (c) 36
mullet (m) 41.3
redfish(r) 37.5
flounder (f) 35
SPECIES FROM BAY)
4-35
-------
Livingston
Table 7. Elimination of Species (%) Based on Projected Increases in Temperature According to a Comparison
of the GFDL and GISS Scenarios With Ambient Conditions in the Apalachicola Estuary (Summer
Months: 1972-1977)
afdl olss ambient
blue crab larvae 22 22 17
blue crab juveniles... 22 22 6
spotted seatrout 20 13 1
oyster larvae 20 13 1
pinfish 20 13 1
flounder 20 13 1
croaker ! 19 4 0
redfish 12 2 0
mullet 100
white shrimp 100
4-36
-------
Livingston
salinity. Since we know little about how the system would respond to each of the projected changes as sole
entities, projections of interacting effects are even more problematical. The interactions of such changes on
single populations should also be related to altered levels and timing of nutrient input and primary productivity
in the Apalachicola estuary.
According to Yanez-Arancibia et al. (1985), the temperature regime of the Terminos Lagoon, a tropical
estuary on the Mexican Gulf of Mexico, approximates that of the Apalachicola estuary with high temperatures
around 30-31° C. The projected estuarine temperatures of the future Apalachicola system were predicated on
an equivalence with increased air temperature. It is possible that conditions in the Gulf of Mexico could alter
such a situation whereby the water temperature would be cooler than the air temperature. Such equivalence
should thus be acknowledged as an untested assumption subject to further review. In any case, such studies of
tropical estuaries indicate that these systems are highly productive in terms of fisheries output and are dependent
on factors (e.g., freshwater input, wetlands vegetation) similar to the temperate estuaries to the north.
SEA LEVEL CHANGES
Projected Effects
One complicating feature of the projected impact of climatic changes on the Apalachicola system includes
sea level alterations. A series of projected changes have been developed by Dr. Richard Park (Table 8). Such
changes should be placed in the perspective given by the various sources of POC for the Apalachicola estuary
as summarized in Table 9 (Livingston, 1984a). The Apalachicola Bay system receives input in terms of organic
carbon and nutrients from various portions of the drainage system. The river flooding during winter-early spring
periods has been associated with input of POC and nutrients. Phytoplankton blooms during the spring and fall
months combine with input from the seagrass beds and fringing salt water marshes during the summer and fall
months to provide the bay with POC and dissolved and participate forms of nutrients. Such inputs are thought
to be mediated by the activities of microorganisms, which transform the dissolved components into the enriched
(e.g., N, P) participate organic matter that forms the basis of the detrital food webs of the estuary. There is a
certain equivalence of input from the freshwater wetlands, the coastal marshes, and the seagrass beds.
Phytoplankton productivity accounts for over 200% of the total of the above sources. The river has been
associated with such production along with the movement of nutrients and POC within the river-bay system.
The models provided by Dr. Richard Park (Butler University, Indianapolis, Indiana) will serve as the basis
of this section of the report. No quantitative (statistical) projections can be made with such data, since there is
little precedence for the major loss of wetlands in this region. However, there are enough data concerning the
function of such areas that impacts can be evaluated concerning losses of major portions of the Apalachicola
wetlands.
Four scenarios of sea level change are presented (Tables 8 and 9): baseline projected changes (no change
in current rate of sea level increase); sea level increase of 0.5 m by 2100; sea level increase of 1.0 m by 2100;
and sea level increase of 2.0 m by 2100). The projected increases in sea level by the year 2100 (File FLAAPLAC;
Scenario 1) would be associated with some decrease in swamp habitat; high marshes would be hard hit and low
marshes would also suffer some losses. No significant changes would occur in the freshwater marshes or the
overall water surface.
The second scenario (0.5-meter increase by 2100) would be associated with more significant changes.
About 1/3 of the swamps and nearly 1/5 of the freshwater marshes would be gone along with nearly all the
low and high marshes around the bay. The increase of mangrove swamps is difficult to factor into these
evaluations; if mangroves would eventually grow in the Apalachicola region, there would be some amelioration
of the effects owing to the projected losses of the existing wetlands. However, due to the uncertainty involved
in this projection, the mangroves will not be taken into account in this analysis.
4-37
-------
Livingston
Table 8. Projected Changes in the Apalachicola Region Due to Sea Level Rises (Based on Data Provided by
Dr. Richard Park: Butler University, Indianapolis, Indiana)
I. File FLAAPLAC
A. Scenario 1:
Time period: 1987-2100
Sea level rise: 1.2 mm/year (baseline).
Projected changes:
1. Swamos: 9.4616 to 6.7053: decrease of 2.7563= 29.1% decrease.
2. Freshwater marshes: 1.4595 to 1.2708; decrease of 0.1887=12.9%
decrease.
3. High marsh: 1.1858 to 0.3656: decrease of 0.8202 = 69.1%
decrease.
4. Low marsh: 3.4213 to 2.3268: decrease of 1.0945 - 32.0%
decrease.
5. Mangrove: no change.
6. Water: 77.4019 to 81.5257: increase of 4.1238 - 5.3% increase.
B. Scenario 2:
Time period: 1987-2100.
Sea level rise: 0.5 m increase.
Projected changes:
1. Swamps: 9.456 to 6.2579; decrease of 3.1981; = 33.8% decrease.
2. Freshwater marshes: 1.4595 to 1.1715 - decrease of 0.2880=
19.7% decrease.
3. High marsh: 1.1858 too.0353; decrease of 1.1505 = 97.0%
decrease.
4. Low marsh: 3.4213 to 0.3879 - decrease of 3.0334 - 88.9%
decrease.
5. Mangrove: 0 to 3.0575 - increase of 3.0575.
6. Water: 77.4019 to 81.857 - increase of 4.4551 - 5.8% increase.
4-38
-------
Livingston
Table 8. (continued)
C. Scenario 3:
Time period: 1987-2100.
Sea level rise: 1 met-ar increase.
Projected changes:
1. Swamps: 9.456 to 5.4695 - decrease of 3.9865 - 42.2% decrease.
2. Freshwater marshes:1.4595 to 0.999 = decrease of .4215 » 28.9%
decrease.
3. High marsh: 1.1858 to 0.0350 = decrease of 1.1508 = 97.0%
decrease.
4. Low marsh: 3.4213 to 0.0616= decrease of 3.397 » 98.4% decrease.
5. Mangrove: 0 to 2.1305= increase of 2.1305.
6. Salt marsh: 4.6071 to 0.0966- decrease of 4.5105= 97.9%
decrease.
7. Water: 77.4019 to 85.2671; increase of 7.8652 = 10.2 % increase.
D. Scenario 4:
Time period: 1987-2100.
Sea level rise: 0.0756 to 2.004 = 2 meter increase.
Projected changes:
1. Swamps: 9.456 to 4.1158: decrease of 5.3402 = 56.4% decrease.
2. Freshwater marshes: 1.4595 to 0.2461: decrease of 1.2134 =
83.4% decrease.
3. Hioh marsh: 1.1858 to 0.0199: decrease of 1.1659 = 98.3%
decrease.
4. Low marsh: 3.4213 to 0.0322: decrease of3.3891 = 99.0% decrease.
5. Mangrove » 0 tp 1.8006 « increase pf 1.8006.
6. Salt marsh: 4.6071 to .0521: decrease of 4.555 - 98.9% decrease.
7. Water: 77.4019 to 88.3808: increase of 10.9789 - 14.1% increase.
4-39
-------
Livingston
Table 9. Compendium of Data Concerning Different Terrestrial and Aquatic Habitat/Productivity V,
in the Apalachicola Drainage Basin
. Distribution and area of Mjor bodies of water along the coast of Franklin
County (north Florida) with areas of oysters, grassbeds, and contiguous Marshes.
Area
Hater, body tha)
St. Tincent Sound 5,539.
lay 20,959.
East Bay 3,980.
St. George Sound (West) 14,746.
St. George Sound (East) 16.015.
Alligator Harbor 1.637.0
Total 62.879.3
Percent of total water area 100
Terrestrial habitats and land-use
Apalachicola Bay system (Florida Bureau of
Category
Residential
Commercial, services
'Transportation, utilities
Mixed urban or built-up areas
Other urban or built-up areas
All urban or built-up areas
Cropland and pasture
Other agriculture
All agricultural land
Herbaceous range land
Range land
Evergreen forest land
Nixed forest land
All forest land
Streams and canals
Lakes
Reservoirs
Bays and estuaries
All water
Forested wetland
Hoof ores ted wetland
All wetlands
Beaches
Quarries and pits
Transitional areas
All barren land
Oysters Grassbeds Marshes
(ha) (ha) (ha)
1,096.5
1.658.5 1,124.7
66.6 1.433.5
1.488.8 624.3
2.6 2.767.3
36.7 261.3
4,349.7 6.211.1
7 10
patterns In the limed I ate
Land and Water Management
Total area (ha)
2.461
178
218
27
39
2.922
78
4
82
13
H.
68.598
36.396
104,994
1.469
452
10
62.879
64,810
25.562
8.465
34,027
1.441
25
110
1,575
i.noe
703
4.606
751
810
144
8.850
14
watershed of
1977).
X of total
113
0.1
O.I
0.0
0.0
1.5
0.0
0.0
0.0
0.0
0.0
35.7*
19.9
54.6
0.8
0.2
0.0
24.3
25.4
13.3
4.4
17.7
0.7
0.0
0.1
0.8
.9
.4
.1
.9
.8
.3
.4
the
Total area of Franklin County: 198.398
4-40
-------
Livingston
Table 9. (continued)
Apalachlcola estuary
Net In situ
Vegetation productivity
mt C yr-1
Freshwater
wetlands
Coastal
marshes
Phyto-
p lank ton
Seagrass
beds
Mapping
category
Pine
Sweet gum-
Sugar berry-
Water oak-
Loblolly Pine
Water hlckory-
Sweetgum-
Overcup oak-
Green ash-
Sugarberry
Tupelo-cypress
with mixed
hardwoods
Tupelo-cypress
Pioneer
Marsh
Open water
Unidentified
Total
360,000
37.714
Apalachlcola Bav system
Net Input Net In situ Net Input Season of
mt C yr-1 productivity mt C yr'1 maximum input
mt C yr-1
30,000 360.000
37,714(7) 46,905
30,000 winter/spring
46.905(7) late summer.
fall(7)
103.080
8.953
Upper
river
136
642
12.500
1.170
2.420
0
0
2.730
1,020
20,600
103,080 233,284(7) 233,284(7) spring
fall
8,953
Middle
river
672
1.440
32.200
1,860
2,270
150
0
3,110
748
42.500
27,213
Acres
Lower
river from
Wewahltchka .
to Sumatra
0
154
15.800
8,310
6.240
19.2
0
1.540
81.3
32,100
27,213
Lower
river from
Sumatra
to mile 10
204
474
1,770
15,800
10,300
0
0
2.010
76.8
30.600
and
sunmer-fall
Lower
river from
mile 10
to mouth
0
0
48.0
6.920
456
0
9,030
1.260
19.2
17,700
Total
1,010
2,710
62.300
34,100
21.700
169
9,030
10,700
1,950-
144,000
Annual output minus Input
(metric tons)
•
Drainage basin
Apalachlcola-
Chattahoochee-
Fllnt
Chattahoochee-
Fllnt
Apalachlcola-
Chlpola
t Apalachlcola
Chlpola
Area
(km?)
50.800
44,600
6.200
3,100
3.100
• 9
Carbon Nitrogen
213
142
71
41
29
,800 21.480
.700 17,860
.100 3.620
.500 1.060
.600 2.560
Phos-
Areal yield
(q m-2 yr-M
Phos-
phorus Carbon Nitrogen chorus
1.652
1.340
312
237
75
4 0.4
3 0.4
12 0.6
13 0.3
10 0.8
0.03
0.03
0.05
0.08
0.02
4-41
-------
Livingston
A one meter increase in the sea level would be accompanied by further reduction of the freshwater marshes
(28.9% decrease). The rest of the marsh systems would be virtually gone with the exception of the mangrove
areas. The 2-meter increase by the year 2100 would be associated with the loss of virtually all the wetlands
systems (freshwater, brackish, and marine) according to the model projections.
An estimate of how such progressive losses of the wetlands in the Apalachicola region would translate
into a loss of POC to the system is given in Table 10. It should be acknowledged that these calculations are
based on major assumptions that cannot be effectively evaluated in terms of their impact on the final results.
With the exception of the mangroves, the model assumes no replacement of marshes as the sea level rises.
The exact relationship of the marshes to the seagrass beds remains undetermined; it is possible that an adverse
impact on the marshes and wetland systems associated with the river-estuary could adversely affect the SAV in
the estuary, since such plants may be dependent on nutrients from such areas. The SAV, on the other hand,
would have more area to grow as the water surface increased; hence the direct increase of SAV as the water
areas expanded with time (Table 10). The direct loss of wetlands may have an equivalent effect on phytoplankton
productivity which may, in turn, be inversely associated with the SAV production (inversely due to competition
for nutrients and light). The projected changes in the carbon regime of the bay are also complicated by unknown
effects on the microbiota: microbial activity is critical to any evaluation of POC input to the estuary. The
extremely complex relationships of such elements have not been determined. However, for the sake of simplicity,
I related the direct loss of phytoplankton productivity with an average of the loss of the freshwater marshes and
the high/low marshes of the system. There is little doubt that the nutrients from such systems, as mediated by
freshwater flow into the estuary, are an important component in the high levels of phytoplankton productivity
in the bay system.
Based on the above assumptions, the calculations (Table 10) would indicate a considerable loss of POC
to the estuary by the year 2100. The salt water marshes would disappear quickly. The phytoplankton losses
would be considerable as the surrounding wetlands disappeared. Even the baseline decreases would be
associated with about one-third of the POC of the system. The projected losses from the OS- to 2,0-m increases
in sea level would be considerable. The increase in SAV over this period would not appear to make up for such
losses. Overall, the input of POC due to the primary allochthonous and autochthonous sources would go from
337,385 metric tons annually in 1987 to 51,955 metric tons in 2100. This loss could then be projected to those
elements of the estuary that are directly involved with the detrital foodwebs. Since most of the major fisheries
are tied in some way to the detrital food webs of the estuary, the projected loss could be applied to similar losses
of the shrimp, oyster, and finfish production of the estuary.
The projected changes in the upper regions of the tri-river basin are equivalent to those treated above
and would not appreciably change the estimates of carbon loss by the year 2100. The speculation concerning
the appearance of mangroves in the Apalachicola region was not addressed here and can be considered to be
a neutral issue by the year 2100 owing to the eventual reduction and loss of such vegetation.
4-42
-------
Livingston
Table 10. Projected Changes of the Net Input of Organic Carbon (mtons/year) to the Apalachicola Bay system
(Livingston, 1984a) According to Various Scenarios of Sea Level Rise (Table 8). (Seagrass Changes
are Related to Increases in Water Areas; Phytoplankton Changes are Related to Changes in the
Average Decreases of Freshwater Marshes, High Marsh Areas, and Low Marsh Areas)
FACTOR FRESH SEA- SALT PHYTO- TOTAL
WETLANDS GRASS MARSHES PLANKTON
Scenario
(1987-2100)
CURRENT 30,000 27,200 46,905 233,280 337,385
BASELINE 26,100 28,700 23,500 144,640 222,940
INCREASES
(1.2MM/yr)
0.5 M 24,000 28,800 4,690 71,450 128,940
INCREASE
1.0 M 21,300 30,100 940 58,790 111,130
INCREASE
2.0 M 4,980 31,035 780 15,160 51,955
INCREASE
4-43
-------
Livingston
CHAPTER 4
DISCUSSION
RIVER FLOW PROJECTIONS AND EFFECTS
The GISS model of precipitation changes predicted an average increase of 2 cm of rainfall per month,
whereas the GFDL model predicted an average decrease of 2 cm per month when applied to the Columbus,
Georgia, precipitation records over the period of study (3/72-7/84). However, the Hains model (GFDLx2)
projected a reduction of over 25% of the river flow. Such an alteration could be associated with proportional
losses of nutrients and POC to the estuary in addition to changes in the salinity regime that could cause shifts
in the biotic relationships and losses of some part of the fisheries potential of the system.
The Apalachicola River delivers considerable quantities of particulate organic matter and nutrients to the
estuary. Such substances are associated with the various food webs in the estuary, especially during the
winter-early spring peak flows. Phytoplankton production during the warmer periods is thought to be related
to nutrients deposited in the bay during various times of the year. The methods used to evaluate changes in
flow preclude an exact measure of the seasonal distribution of possible changes in the delivery of organic carbon
and nutrients to the system. In addition, the bay system, in terms of actual utilization of such substances, is
characterized by various complicating features such as species-specific lags in consumption, changes in microbial
utilization and availability of enriched detrital material to the estuarine food webs, and differences in
phytoplankton production with time. Consequently, the overall changes in the amounts of such products
delivered .10 the bay can only be calculated hi a gross fashion. In general, the loss of the estuarine food webs
would be roughly proportional to the reduction of river flow delivered to the bay.
Specific variables such as turbidity and color usually reflect the level and time of Apalachicola River flow
so that projected increases or decreases would result in similar changes in such variables in the estuary.
Short-term reductions in the river flow could be generally associated with increases in the oxygen anomaly,
although dissolved oxygen per se probably would not be affected.
Reductions in river flow would result in proportional decreases in nutrient delivery and phytoplankton
productivity. At the same time, organisms that depend on phytoplankton, such as oysters, would be adversely
affected by such reductions. The associated decrease of color and turbidity would, all other factors remaining
equal, probably allow an expansion of the submerged aquatic vegetation (SAY) which, in turn, would provide
habitat and food for those organisms inhabiting the estuary during the warmer months of the year. With a shift
to higher salinity, white shrimp, oysters, and various sciaenid fishes such as spot and croaker would be at a
disadvantage relative to the pink shrimp, spotted seatrout, and other organisms that are associated with the
submerged aquatic vegetation. Because of the complexity of the interrelationships of habitat, physical constraints,
productivity, predator-prey relationships, and competition, no exact sequence of response can be predicted
Overall, with reductions of the river flow, the major fisheries that currently are important in the Apalachicola
estuary would give way to fisheries associated with the seagrass beds; the Apalachicola system would begin to
assume the characteristics of Apalachee Bay, all other factors remaining the same. Increases in river flow would,
within limits, expand the current estuarine habitat and increase the level of fisheries production in the
Apalachicola Bay system.
A review of the long-term biological responses of the Apalachicola estuary to periods of drought and flood
indicate that the response to prolonged increases or decreases in precipitation and river flow would be extremely
complex and would, in large part, be based on the the responses of individual populations to such changes.
There would also be considerable differences between the short- and long-term responses of such organisms to
prolonged changes in state variables. The complex relationship of productivity and salinity in the estuary
complicates this situation. Reductions in river flow would result in increases in the numerical abundance and
species richness of infaunal macroinvertebrates. However, the relative dominance and overall production of this
4-44
-------
Livingston
group of organisms would be reduced by such changes. Increases in river flow would expand the area of the
highly opportunistic species that dominate the oligohaline and mesohaline infaunal associations. As river flow
decreased over time, there would be a reduction in the physically controlled infaunal groups with a corresponding
expansion of the biologically controlled associations. In the long run, such changes would probably lead to an
overall reduction in the standing crop biomass and numbers of such organisms.
Increased salinity due to a reduction in river flow would allow oyster predators into the estuary; enhanced
disease vectors and parasites could also take a toll of the oyster productivity. Coupled with a reduction in the
phytoplankton productivity, the oyster1 industry would be adversely affected by the projected changes in river flow
(hains GFDLx2). The same could be said for those organisms, such as white shrimp, blue crabs, and sciaenid
fishes, which currently use the estuary as a nursery area and a haven from the stenohaline gulf predators that
would be able to enter the bay as salinities increased. Although the habitat alterations would differ in various
parts of the bay, the overall effects of reduced river flow would probably be consistent with general reductions
in the various invertebrate and fish species which currently form the basis of the commercial fisheries of the
region. The immediate estuarine response to changes in the river flow would probably be somewhat different
than the long-term responses to such changes, and such differences would be species-specific in terms of the
exact timing of the response. There is some evidence that after an initial increase of certain populations relative
to the increased salinity of the system, there would be overall decreases in the numbers and biomass of the
dominant epibenthic macroinvertebrates and fishes. Trends of numbers and biomass in the individual populations
would not necessarily be parallel in such long-term responses.
TEMPERATURE CHANGES AND EFFECTS
Since the early 1940's, winter low temperatures in the estuarine systems along the north Florida gulf coast
have varied considerably from year to year with long-term cycles (8-10 years) of the recurrence of very cold
winters. Summer high mean temperatures are less variable from year to year with a range approximating
28-31°C. In terms of changes in the average water temperature of the Apalachicola estuary, the GISS 2xCO2
scenario predicted minimal increases of summer high temperatures and increases of 2-3° C of the winter lows.
The GFDL 2xCO2 model predicted increases of 4-5° C in the average summer high temperatures with projected
differences in the summer maximum temperatures of 5-9°C. Average winter low temperatures would increase
4-5°C according to the GFDL projections. This would mean that there will be future summer temperatures
(mean) along a range of 31-36°C with summer maximum temperatures occasionally exceeding 40°C.
Temperature is an important factor in the seasonal and interannual variation of organisms in the
Apalachicola estuary. Infaunal macroinvertebrates of East Bay have peak numbers/biomass during winter-early
spring months of high river flow and low winter temperatures. Species richness/diversity tends to follow similar
trends as the numerical abundance. Such indices are associated with winter low temperatures and drought
periods. Epifaunal macroinvertebrates, dominated by the commercially important penaeid shrimp and blue
crabs, are most abundant during summer-fall periods of high temperature hi the estuary. Temperature-related
trends of such species appear to be less important than precipitation/river flow variation over long (interannual)
periods, although overall invertebrate biomass was associated with periods of very low winter temperatures. Fish
numbers tend to peak during winter-early spring periods due to the migrations of juvenile spot and croakers into
the estuary at that time. Fish species richness peaks during the fall when a number of species such as the
spotted seatrout are abundant: such species are important to the sports fisheries of the region. Absolute fish
numbers peaked during the drought period, while fish dry weight biomass was highest during the period of
abundant river flow. Thus, the stereotypic seasonal temperature response of the dominant fishes and
macroinvertebrates (infaunal and epifaunal) is complicated by long-term trends of regional precipitation,- there
are distinct interactions of low winter temperatures and drought (e.g., rainfall/river flow) conditions that tend
to drive the biological response. There is considerable variation in such response from one level of biological
organization to the next.
Based on.reviews of upper thermal collection limits (e.g., the highest temperature at which different
estuarine species have been taken in the field) and the incipient lethal temperatures (e.g., that temperature
4-45
-------
Livingston
causing the death of a specific fraction of test organismsX there is good evidence that as the upper summer
mean and maximum temperatures increase, there will be a reduction and/or elimination of some of the most
important populations in the Apalachicola estuary. Blue crabs and stone crabs will no longer be in the estuary
at mean temperatures of 30-33°C. By the time the mean temperatures reach 35-36°C, oysters, penaeid shrimp,
catfishes, mullet, the seatrouts, and southern flounder will no longer be able to use the estuary as a habitat if
such populations follow the experimentally determined sensitivities to high temperatures.
Should the GFDL scenario of temperature increase be correct, there is a high probability or likelihood
of a major change in the biotic composition and fisheries potential of estuaries across the northern Gulf of
Mexico. The main uncertainty in this generalization rests on the possibility that various species will adapt to
the temperature changes and/or be replaced by other species that may or may not take their place. Such
alternatives are possible and cannot be determined here with any degree of certainty. The relatively short
period over which such temperature increases would occur would tend to preclude any significant adaptation
of these species to the projected summer high temperatures. Within the context of these uncertainties, it is
highly likely that temperature increases in the Apalachicola estuary of 4-5°C (mean high summer values) would
have a significant effect on the biota of this system. Since many of the organisms that inhabit the estuary during
summer periods are already close to limiting high temperatures, any sustained increase in such temperature levels
could have a negative impact on the fisheries potential of the estuary. The projected increases of the winter low
temperatures with time could have a profound influence on the species composition of the invertebrate and fish
faunas in the estuary with an overall trend toward an increase in tropical species, general increases in species
richness and diversity, and associated reductions of fish and invertebrate production. Such projected changes
would, of course, be dependent on associated changes in variables such as dissolved oxygen (which would be
reduced as the temperatures increased) and salinity, which is dependent on other state variables such as regional
precipitation and Apalachicola River flow.
PROJECTED CHANGES IN SEA LEVEL
According to projected changes of sea level from current (baseline) changes to a 2-m increase by the year
2100, there would be progressive and almost total losses of freshwater marshes, the high and low marshes, and
the salt marshes of the Apalachicola system. Over half the swamps would be lost with a 2-m increase hi sea level
by the year 2100. Based on a series of basic assumptions, such losses could be associated with considerable
reductions of autochthonous and autochthonous sources of POC to the estuary. Seagrass beds would undergo
moderate increases, whereas phytoplankton productivity would be severely reduced The organic carbon input
from the freshwater and saltwater marshes would be virtually wiped out. Overall POC levels, in metric tons/year,
would go from 337,385 hi 1987 to 51,955 by 2100. Such losses could be directly related to losses of specific
populations within the important detrital food webs throughout the Apalachicola Bay system. Overall, the system
would be transformed into a seagrass-dominated estuary with relatively minimal input of POC from the
surrounding wetlands. The phytoplankton productivity, presently a primary source of organic carbon, would be
severely reduced to a relatively insignificant level The bay system would be much less turbid with most of the
nutrients held hi the seagrass vegetation.
Ultimately, the various projected changes in the Apalachicola estuary over the next century would include
a simultaneous yet gradual alteration of the system, whereby river flow would decrease as the wetlands were
inundated by sea level rise. During this time, the water temperature would approach and possible exceed that
currently observed in the tropics. Based on the combined effects, it would be virtually impossible to predict the
exact sequence of the biological response of the estuary to such changes. However, the estuary would change
hi terms of species composition, the source and level of productivity, and the fisheries potential Relative to the
current conditions, it would be a warm water seagrass system with less overall fisheries production.
There are certain shortcomings to the analysis presented in this report. A complete review of this problem
is given in Appendix VI. The use of correlations, regressions, and other multivariate analyses with long-term,
multidisciplinary data is extremely limited in terms of scope, comprehensiveness, and the development of
predictive models. In essence, such analysis should be used as a preliminary review of the data so that the state
4-46
-------
Livingston
variables can be identified for more comprehensive modeling efforts. In this case, the weaknesses of the rather
simple statistical tests were obvious in that specific lags in the biological response to important climatological
features could not be included in the final interpretation of results. It should be emphasized that, from the
beginning, this analysis was considered preliminary to a more rigorous application of more sophisticated statistical
models to the data base.
Accordingly, it is appropriate that a more comprehensive analysis be made of the data with inclusion of
similar (currently available) information from other river-estuaries along the gulf coast. The data are sufficient
for the application of time-series analysis (ARIMA models, spectral analysis, intervention analysis); such models
do not have the rather restrictive assumptions that are associated with most multivariate statistical techniques.
In addition, there should be an application of simulation models. Again, the data base is sufficient for the
analysis of results with different portions of the available information, both from the standpoint of within-system
and between-system comparisons. A comparison of the various estuaries in which long-term field data have been
taken would thus be an important part of such an analysis. The use of the above techniques in the development
of predictive models could be considered the end result of a 20-year effort to understand river-estuaries along
the Florida gulf coast.
4-47
-------
Livingston
REFERENCES
Brett, J. R. Temperature. la- Marine Ecology. Ed O.Kinne. Wiley-Interstience, New York, 1970. pp. 513-616.
Colberg, M.R., TJS.Dietrich, and D.M.Woodham. 1968. The social and economic values of Apalachicola Bay,
Florida. Final report. Contract NO. 14-12-117. U.S. Environmental Protection Agency, Washington, D.C.
Collins, M.R.1985. Species profiles: life histories and environmental requirements of coastal fishes and
invertebrates (South Florida), striped mullet. U&Fish and Wildlife Service. FWS/OBS-82/11.21. U.S. Army
Corps of Engineers, TR EL-82-4.
Dubravko, J. and R. J. Livingston. Long-term oxygen cycling as an indicator of productivity in a river-dominated
estuary, (in manuscript).
Lassuy, D.R.1983. Species profiles: life histories and environmental requirements of coastal fishes and
invertebrates (Gulf of Mexico), spotted seatrout. U.S.Fish and Wildlife Service. FWS/OBS-82/11.21. U.S. Army
Corps of Engineers, TR EL-82-4.
Lassuy, D.R.1983. Species profiles: life histories and environmental requirements of coastal fishes and
invertebrates (Gulf of Mexico). Atlantic croaker. U^.Fish and Wildlife Service. FWS/OBS-82/11.21. U.S. Army
Corps of Engineers, TR EL-82-4.
Livingston, R. J. Resource atlas of the Apalachicola estuary. Florida Sea Grant College Rep. No. 55. 1983. 64
pp.
Livingston, R. J. The ecology of the Apalachicola Bay system: an estuarine profile. U. S. Fish Wildl. Serv.
FWS/OBS 82/05.1984a. 148pp.
Livingston, R. J. Trophic response of fishes to habitat variability in coastal seagrass systems. Ecology 65(4),
1984b. pp. 1258-1275.
Livingston, R J. and EA Joyce, Jr. Proceedings of the conference on the Apalachicola drainage system. Florida
DNR Marine Research Publication No. 26. St. Petersberg, Florida.
Meeter, D. A., R. J. Livingston, and G. C. Woodsum. Long-term climatological cycles and population changes
in a river-dominated estuarine system. la* Ecological Processes in coastal and marine systems. Ed R. J.
Livingston. Plenum Press, New York. 1979. pp. 315-338.
Mulkey, D. and F. Mature. 1983. A socioeconomic profile of the Apalachicola area of northwest Florida.
Apalachicola oyster industry: Proceedings of a conference. Ed Scott Andree. Sea Grant ReportNo. SGR-57.
Gainesville, Florida.
Muncy, RJ. 1984. Species profiles: life histories and environmental requirements of coastal fishes and
invertebrates (Gulf of Mexjco)--white shrimp. U^.Fish and Wildlife Service. FWS/OBS-82/11.21. US. Army
Corps of Engineers, TR EL-82-4.
Prochaska, FJ. and D. Mulkey. 1983. The Apalachicola oyster industry: some economic considerations.
Apalachicola oyster industry: Proceedings of a conference. Ed Scott Andree. Sea Grant Report No. SGR-57.
Gainesville, Florida.
4-48
-------
Livingston
Reagan, R.E.1985. Species profiles: life histories and environmental requirements of coastal fishes and
invertebrates (Gulf ofMexico). red drum. U.S.Fish and Wildlife Service. FWS/OBS-82/11.21. VS. Army Corps
of Engineers, TR EL-82-4.
Reagan, R.E. and W. M. Wingo.1985. Species profiles: life histories and environmental requirements of coastal
fishes and invertebrates (Gulf of Mexico), southern flounder.U^.Fish and Wildlife Service. FWS/OBS-82/11.21.
US. Army Corps of Engineers, TR EL-82-4.
Stanley, J.G. and MA. Sellers. 1984. Species profiles: life histories and environmental requirements of coastal
fishes and invertebrates (Gulf of Mexico). American oyster. U.S.Fish and Wildlife Service. FWS/OBS-82/11.21.
US. Army Corps of Engineers, TR EL-82-4.
Sutler, F.C. and T.D.Mdlwain. 1984. Species profiles: life histories and environmental requirements of coastal
fishes and invertebrates (Gulf of Mexico), sand seatrout and silver seatrout. U.S.Fish and Wildlife Service.
FWS/OBS-82/11.21. US. Army Corps of Engineers, TR EL-82-4.
Van Den Avyle, M J. 1984. Species profiles: life histories and environmental requirements of coastal fishes and
invertebrates (South Atlantic)~blue crab. U-S.Fish and Wildlife Service. FWS/OBS-82/11.21. US. Army Corps
of Engineers, TR EL-82-4.
Warren, C. E. Biology and water pollution control W. B. Saunders Company, Philadelphia 1971. 434 pp.
Yanez-Arancibia, A., A.L.Laral-Dominguez, and H. Alvarez-Guillen. 198S. Fish community ecology and
dynamics in estuarine inlets. In; Fish commumity ecology in estuaries and coastal lagoons: toward an ecosystem
integration. Ed.: A. Yanez-Arancibia. UNAM Press, Mexico.
Yanez-Arancibia, A., G. Soberon Chavez and P. Sanchez-Gil. 198S. Ecology of fish production in the coastal
zone. In; Fish commumity ecology in estuaries and coastal lagoons: toward an ecosystem integration. Ed: A.
Yanez-Arancibia, UNAM Press, Mexico.
4-49
------- |