United States
Environmental Protection
Agency
r Office of Policy
& Resource Management
Washington, DC 20460
JtiJ JT//V. ,
October 1983
EPA 230-09-007
Revised
ENVIRONMENT,
PROTECTION
Projecting
Future Sea Level Rise
Methodology,
Estimates to the Year 2100,
& Research Needs
DALLAS, TBM<
-------
C_
PROJECTING FUTURE SEA LEVEL RISE
Methodology, Estimates to the Year 2100,
and Research Needs
By
John S. Hoffman, EPA
Dale Keyes, Consultant
James G. Titus, EPA
A Report of:
The Strategic Studies Staff
Office of Policy Analysis
Office of Policy and Resource Management
U.S. Environmental Protection Agency
401 M Street, S.to. (PM-221)
Washington, D.C. 20460
(202) 382-5484
2nd Edition, Revised
October 24, 1983
-------
TABLE OF CONTENTS
Page
PREFACE iii
ACKNOWLEDGMENTS v
REPORT SUMMARY AND FINDINGS vi
CHAPTER 1
INTRODUCTION 1
CHAPTER 2
THE SCIENTIFIC BASIS FOR PROJECTING SEA LEVEL RISE 4
CHAPTER 3
THE APPROACH, ASSUMPTIONS, AND MODELS USED TO
ESTIMATE SEA LEVEL RISE 13
CHAPTER 4
SEA LEVEL SCENARIOS TO THE YEAR 2100 38
CHAPTER 5
IMPACTS OF SEA LEVEL RISE 41
CHAPTER 6
RESEARCH NEEDED TO IMPROVE
ESTIMATES OF SEA LEVEL RISE 51
APPENDIX A — SUMMARY OF SEA LEVEL RISE SCENARIOS,
INCLUDING SPECIAL CASES OF INCREASED
VOLCANIC ACTVITY AND CHANGES IN
SOLAR IRRADIATION 59
APPENDIX B — MODELS, ANALYTICAL METHODS, AND
ASSUMPTIONS USED FOR ESTIMATING
THE SCENARIOS OF SEA LEVEL RISE 63
APPENDIX C — METHODS OF ESTIMATING
SNOW AND ICE CONTRIBUTION 101
BIBLIOGRAPHY 115
-------
- Ill -
PREFACE
The primary objective of this report is to estimate the
range of future sea level rise. This information should help
coastal engineers, planners, coastal zone managers, water supply
and quality planners, and site planners to make better decisions
in coastal areas. Scientists and federal research policy makers
can use this report in choosing research to improve sea level
rise estimates.
This report has undergone an extensive peer review to ensure
its accuracy and completeness. Nevertheless better estimates
of sea level rise will be forthcoming as scientific knowledge
improves.
No EPA policy is implied in this document. Comments are
welcome and should be sent to John Hoffman, PM-221, Strategic
Studies Staff, EPA, Washington, D.C. 20460.
-------
- V -
ACKNOWLEDGMENTS
Several people made valuable contributions to this study.
Dr. Alan Truelove of Pechan and Associates did much of the early
computer programming. Dr. Sergej Lebedeff, Dr. Gary Russell,
Dr. Andrew Lacis, and Dr. James Hansen of the Goddard Institute
for Space Studies (GISS) provided computer programs and technical
advice on the temperature-flux equations, thermal-expansion
model, and ice-sheet melting. Dr. Robert Thomas and Dr. David
Thompson of the Jet Propulsion Laboratory provided information
on glacial processes. Dr. William Emanuel of Oak Ridge National
Laboratories provided the computer program and technical advice
on the carbon-cycle model. C02 projections were derived from
runs of a modified version of the World Energy Model provided by
the Institute of Energy Analysis, Oak Ridge Associated Universities,
Loren Dunn of EPA made valuable contributions to earlier versions
of this effort. Wynne Cougill and Steven Seidel made editorial
suggestions. Tom Glover provided graphics and Carolyn Acklin typed
numerous drafts.
We would like to thank more than one hundred peer reviewers,
too numerous to name, who took time to comment on this report.
All errors, omissions and inaccuracies, however, are the
responsibilities of the authors.
-------
- VI -
REPORT SUMMARY AND FINDINGS
Concentrations of atmospheric CC>2 and other greenhouse gases
will continue to increase in coming decades. Two National Academy
of Sciences panels have concluded that higher levels of these
gases will almost certainly produce a large global warming.
That warming, by thermally expanding the oceans and by causing
the transfer of ice and snow resting on land to the oceans,
should raise sea level substantially faster than the rise that
has taken place during the past century.
Although current knowledge is inadequate to make a precise
prediction of future sea level rise, it is sufficient to predict
the likely range. Many factors were considered in generating
the estimates of sea level rise contained in this report:
population and productivity growth, atmospheric and climatic
change, and oceanic and glacial response. High and low assump-
tions for these principal determinants of sea level rise were
derived from the literature. When linked together the various
assumptions allowed the estimation of high and low paths of future
sea level rise. Based on this work, the following findings can
be stated:
(1) GLOBAL SEA LEVEL WILL ALMOST CERTAINLY RISE IN COMING DECADES.
A global rise of between 144 cm (4.8 feet) and 217 cm
(7 feet) by 2100 is most likely.
A global rise as low as 56 cm (1.9 feet) or as high as
345 cm (11 feet) by 2100 cannot be ruled out.
Along most of the Atlantic and Gulf Coasts of the United
States, the rise will be 18 to 24 cm (0.6 to 0.8 feet)
more than the global average.
(2) ESTIMATES OF FUTURE SEA LEVEL RISE CAN BE USED TO REDUCE
ITS ADVERSE IMPACTS.
Sea level rise will increase shoreline retreat, erosion,
flooding, and saltwater intrusion in coastal areas.
Important economic impacts could result.
Many, if not most of the adverse economic impacts of sea
level rise can be avoided if timely actions are taken in
anticipation of these effects.
Professionals and policymakers need to assess the
vulnerability of forthcoming decisions and the existing
infrastructure to sea level rise.
-------
- VI 1 -
3) THE RANGE OF FUTURE SEA LEVEL RISE ESTIMATES COULD BE NARROWED
BY ACCELERATING RESEARCH.
- Differences in estimates of sea level rise are due to
deficiencies in scientific knowledge and in the methods
used for constructing estimates. These deficiencies
can be corrected.
The most important research opportunities for improving
sea level rise estimates are in estimating the transfer
of ice and snow from land to sea, understanding the
growth of the minor trace gases, and narrowing estimates
of the sensitivity of the climate system to increased
greenhouse gases. Many opportunities exist for improving
knowledge in these areas.
Under current funding, estimates of sea level rise will
improve very slowly.
- Greater and more secure funding of interdisciplinary
teams could produce a more rapid improvement in
estimates.
-------
CHAPTER 1
INTRODUCTION
Sea level can be expected to rise substantially throughout
the coming decades as a result of a global warming caused by
rising concentrations of atmospheric carbon dioxide,£/ methane,
nitrous oxide, and chlorofluorocarbons. The anticipated sea
level rise would increase coastal erosion and flooding;
groundwater tables would rise; and saltwater would intrude
into rivers, bays, and aquifers.^/
Most coastal planners and decision makers are not considering
the effects of sea level rise. Consequently, facilities are
being located without adequate consideration of the costs of
protecting them from erosion, flooding, or storm waves; important
environmental protection decisions are being based on inadequate
assessments of their future effectiveness; and entire communities
are being planned without adequate consideration of the
coastal works that will be necessary to protect their physical
infrastructures and capital investments.^/ Decision makers
and technical personnel need estimates of sea level rise to
evaluate the vulnerability of their decisions to changing risks.
A tremendous reserve of knowledge has been accumulated by
natural and social scientists about the factors that will
determine sea level rise - population growth, climatic change,
oceanic heat absorption, glacial discharge, and atmospheric
composition. This study uses that knowledge to project low and
-------
- 2 -
high ranges for future sea level rise. Because the range of
assumptions used to project sea level rise varied from the very
conservative to the less restrictive, the estimates produced
here are likely to encompass the true rate of future sea level
rise.
This report is organized as follows:
o Chapter 2 reviews the scientific basis for the
belief that global warming will occur and cause sea
level to rise.
o Chapter 3 presents the approach, the assumptions, and
the models used to estimate sea level rise.
o Chapter 4 presents the results of our analysis: the
low, high, and mid-range sea level rise scenarios.
o Chapter 5 reviews the research conducted on the impacts
of sea level rise and discusses the need for further
research on this subject.
o Chapter 6 presents a review of the research under way
to improve estimates of sea level rise, along with
options for accelerating the efforts to improve these
estimates.
o Appendix A provides a summary of a larger set of sea level
rise scenarios, including "special case" scenarios.
o Appendix 8 provides technical details of the models and
assumptions used to generate the scenarios on thermal
expansion.
o Appendix C provides details on the methods used for
estimating snow and ice contributions, including a
detailed analysis of the melting estimated under doubled
CC>2 in a three-dimensional global climate model.
-------
- 3 -
Each chapter is followed by end notes that provide references
and explanations of the text. It should be noted that the
estimates of sea level rise presented in this report differ
from those of the peer review and first edition (July, 1983)
because of improvements made in the treatment of trace gases.
END NOTES TO CHAPTER 1
1. Charney, Jules, et al., 1979. Carbon Dioxide and Climate:
A Scientific Assessment. Washington, D.C.: National
Academy of Sciences Press. The warming predicted
in this report is 1.5°C to 4.5°C for a doubling of
C02.
2. For an overview of the impacts of sea level rise and
decisions that may be influenced, see:
Titus, James G., et al., 1983. Sea Level Rise Conference
Document (draft). Washington, D.C.: EPA., or
Earth, Michael C. and James G. Titus, (eds). Sea Level
Rise to the Year 2100. Stroudsburg, PA: Hutchinson Ross
(in press).
For example, the 100-year floodplain defined in Federal
Emergency Management Agency (FEMA) maps is being used at
EPA in making many decisions for coastal areas. The maps
have not, however, included any sea level rise in setting
the flood boundaries. Consequently, facilities may face
greater hazards or have shorter useful lives than planned
for.
-------
- 4 -
CHAPTER 2
THE SCIENTIFIC BASIS FOR PROJECTING SEA LEVEL RISE
Global sea level depends primarily on three factors:
(1) the total amount of water that rests in ocean basins rather
than on land; (2) ocean temperatures at various depths, which
determine ocean density and volume; and (3) the shape (bathymetry)
of the ocean floors.^/ Global warming could increase the
water resting in the oceans and the volume of that water, and
thereby raise sea level. Because changes in bathymetry are
slow and unlikely to accelerate, this report considers only
the first two factors.
This chapter discusses the scientific basis for expecting
a global warming to occur, and the linkages between that warming
and sea level.
GREENHOUSE GASES HELP DETERMINE THE PLANET'S TEMPERATURE
The earth's temperature is determined by three factors:
the sunlight it receives, the sunlight it reflects, and the
infrared radiation absorbed by the atmosphere ._2/ Without the
influence of the atmosphere, incoming visible radiation (in
the form of sunlight) and outgoing radiation (in the form of
invisible infrared radiation) would balance to yield a certain
surface temperature. However, the atmosphere contains gases
such as CO2 and water vapor that absorb some of the infrared
radiation. These gases are warmed by the radiation and radiate
energy back to the earth's surface, raising its temperature.
-------
- 5 -
The larger the percentage of infrared radiation blocked by
the atmosphere, the warmer the earth's surface temperature.
This feature of CC>2 and certain other gases is known as the
"greenhouse effect."
The greenhouse effect is an important factor determining
a planet's temperature, as observations of the temperatures of
other planets confirm (see Table 2-1).£/
TABLE 2 - 1
THE GREENHOUSE EFFECT ON THE INNER PLANETS: PREDICTION AND OBSERVATIONS
VENUS
EARTH
MARS
Sunlight
Received
(watts per
square meter)
2613
1367
589
Sunlight
Reflected
(percentage)
75
30
15
Temperature
Without An
Atmosphere
-40°C
-18°C
-56°C
"Opacity"
Of Atmosphere
To Infrared*
~ 100
~ .1
Predicted
Temperature
With Simple
Models**
429°C
17°C
-52°C
Actual
Temperature
427°C
15°C
-53°C
Higher number indicates greater ability to trap infrared radiation.
*Average of two simple models: radiative equilibrium and corrective
equilibrium.
Source: Modified from Hansen, J. Lacis, A. and Rind, D., "Climate Trends Due to
Increasing Greenhouse Gases", Coastal Zone 83. New York: ASCE, 1983.
Contrary to popular belief, Venus is not hotter than Earth
because it is nearer to the sun, but because its atmosphere is
97 percent CC>2. Although Venus receives more sunlight
(2613 watts per square meter) than Earth (1367 watts per square
meter), it reflects 75 percent of this radiation (which explains
its brightness), compared with 30 percent for Earth. If the
atmospheres of the planets did not differ in their ability to
absorb infrared, Earth, with its lower reflectivity, would be
23°C degrees warmer than Venus. Similarly, with its very low
-------
- 6 -
reflectivity, Mars would be only 15°C cooler than Earth.
However, these planets have atmospheres with very different
capabilites of absorbing infrared. The atmosphere of Mars has
almost no water vapor and a low level of CO2. According to
simple models of the greenhouse effect, its temperature should
rise only slightly. In contrast, according to the same models,
Venus's atmosphere of 97 percent CC>2 should raise its temperature
to a scorching 432°C (775°F), and Earth's atmosphere of 0.03
percent CC>2 and considerable water vapor should warm it by
30°C to an average of 17°C (63°F). Space probes to Mars and
Venus and measurements on Earth confirm that their actual
temperatures rise as predicted.
This and other evidence has led scientists to believe
that as atmospheric levels of CC>2 and other greenhouse gases
increase, Earth's temperature will rise. Using more comprehensive
models of the greenhouse effect, it has been relatively simple
to estimate that if no other changes take place in atmospheric
composition (i.e., concentration of other greenhouse gases) or
the albedo (reflectivity) of the planet, a doubling of atmospheric
CC>2 concentrations will warm the earth's surface temperature
1.2°C.f/
Projecting the extent of the future warming is complicated
by the fact that the initial warming will change reflectivity
and atmospheric composition in a way that will almost certainly
amplify the direct warming. For example, the 1.2°C temperature
-------
- 7 -
rise will increase water vapor levels in the atmosphere,
trapping more infrared radiation, and will also melt snow and
ice, decreasing the reflectivity of the earth. Other climatic
effects, such as changes in cloud cover and cloud heights,
could increase or diminish these amplifications by amounts
that cannot be predicted accurately.
The 1979 report of the National Academy of Sciences (NAS),
prepared by a review panel of leading climatologists, reflected
the agreement of the scientific community on the effects of
these feedbacks. The panel members summarized their view by
stating: "We have tried but have been unable to find any
overlooked physical effect that could reduce the currently
estimated global warming due to a doubling of C02 to negligible
proportions. . . . "jV Unfortunately, the panel's agreement
that a large warming will occur did not permit it to develop
a narrow range for the estimated global temperature increase.
However, it was able to conclude that the earth's equilibrium
temperature increase for doubled CC>2 would be at least 1.5°C
(2.7°F) and not more than of 4.5° (8.1°F).
Since 1979, scientific consensus has continued to grow
that the warming will be significant. In 1982, a second NAS
panel reviewed new evidence and confirmed the conclusions of
the first panel._£/ More recently, the World Meteorological
Organization and other researchers have concluded that other
gases whose atmospheric concentrations are increasing could
double the warming from CO9 alone. / Thus, by the time CO.-,
^ — ^
-------
doubles, the earth's total equilibrium temperature increase
is likely to be between 3°C to 9°C (5.4°F to 16.2°F).
CONCENTRATIONS OF GREENHOUSE GASES WILL ALMOST CERTAINLY DOUBLE
IN THE NEXT CENTURY
In the past 180 years, atmospheric CO2 appears to have
risen approximately 20 percent, from between 260 and 290 ppm
to 340 ppm. Very accurate monitoring began in 1958; since then,
atmospheric C02 has increased 8 percent, from 315 ppm to
340 ppm._^/
Atmospheric levels of other gases have also risen. Methane
increased annually by about 1 to 2 percent from 1970 to 1980,
chlorofluorocarbons by about 6 percent over the same period.
and nitrous oxide by about 0.2 percent per year from 1975
to 1980.V
Economic activities have caused most of these increases.
For CO2, the most important cause of emissions has been the
combustion of oil, gas, and coal, with deforestation probably
contributing a small percentage. Burning fossil fuels (hydro-
carbons) oxidizes carbon (C + 2O -> CO2), inevitably
releasing CO2-
Future energy use and fuel selection will thus be the
primary determinants of the rate of C02 emissions. Because
fossil fuels have important competitive advantages and play a
critical role in existing energy systems, their use is expected
to grow even if radical policies are undertaken to curtail their
use. For example, Seidel and Keyes found that even a 300 percent
worldwide tax on fossil fuel use would delay a 2°C temperature
-------
- 9 -
rise only five years.10/
Because only a fraction of the emissions will remain in the
atmosphere, concentrations of atmospheric CC>2 cannot be predicted
by emissions alone. Nevertheless, although disagreement exists
about the exact fraction, an almost universal consensus exists
that it will be great enough to ensure that atmospheric CC>2
levels will at least double in the next century.H/
The future rates of increase for the other greenhouse gases
are less certain. Much less is known about the sources
of methane, nitrous oxide, and chlorofluorocarbonsj^/ than the
sources of CC>2. Furthermore, scientific understanding about
the fate of these gases once they enter the atmosphere is also
insufficient. Nevertheless, atmospheric concentrations of
these gases are likely to increase substantially, and in the
case of chlorofluorocarbons, probably at a rate faster than CC>2.
THE OCEANS WILL ALSO INFLUENCE THE RATE OF FUTURE WARMING
The rate of future warming will depend on more than the
earth's equilibrium temperature sensitivity and the rate of
increase in atmospheric C02 and other greenhouse gases. It
will also depend on the time it takes for the climatic system
to reach the equilibrium temperature.
The most important factor delaying the warming will be
the oceans' capacity to absorb heat that would otherwise warm
the atmosphere.^/ As surface air temperatures increase, a
temperature difference will develop between the surfaces of
the atmosphere and the oceans, causing heat to be moved to
-------
- 10 -
the oceans. This heat will slowly pass downward to the lower
and cooler layers. In this way, the oceans will act as heat
"sinks", somewhat delaying the full atmospheric warming.
EXPECTED TEMPERATURE INCREASES WILL BE SIGNIFICANT
The expected global warming will be large compared to
historical temperature changes, and even more significant
compared to historical rates of temperature change. In the last
two million years, the earth nas never been more than 2 to 3°C
warmer than it is today.^/ In the last hundred thousand years,
it has been at most 1°C warmer, and in the last thousand years,
at most 0.5°C warmer. Since the Wisconsin Ice Age (18,000 years
ago), the earth has warmed about 4°C,££/ and in the last century,
about 0.4°C._J_^/ The projected warming for the next century would
be ten times as rapid as the Historical warming trend.
The expected warming will also be large compared to geo-
graphical temperature differences. A 3°C warming would leave
San Francisco as warm as San Diego is today. A 9°C warming
would raise New York City's average temperature to the current
temperature of Daytona Beach, Florida.
THERMAL EXPANSION AND GLACIAL DISCHARGES COULD RAISE SEA LEVEL
As global temperature rises, the sea level can be raised
in two ways:
o Thermal expansion
Warming will decrease the density of ocean water, and
thus increase its volume. Because the same water will
take up more space, the levels of the oceans will
rise. The rate of this rise directly depends on the
amount of heat the oceans absorb.
-------
- 11 -
o Transfer of snow and ice from land to sea
Warming can transfer snow and ice from land to sea by
melting (if the meltwater runs off into the ocean) or
deglaciation (glaciers breaking up or moving more
rapidly into the ocean). Sea level will rise as a result
if these effects are not offset by the additional
accumulation of snowfall on land. Both of these effects
are likely to occur with the warming predicted for the
next century. The rate at which they would occur,
however, is less certain than that of thermal expansion
and may not be proportional to the rate of warming.
Deglaciation, in particular, may be a phenomenon in
which a threshold exists. Once this threshold is
passed, the deglaciation may become a self-reinforcing
process, whose timing depends on many things other
than the magnitude of the warming.
Geologic history indicates that these physical mechanisms
describe the behavior of the earth. During warmer periods in
earth's history, sea level has been higher than in colder
periods, varying by over one hundred meters.
END NOTES TO CHAPTER 2
_!/ Winds, currents, and land subsidence and emergence may
cause local sea level to change at rates different from
global sea level. For specific applications, the global
estimates in this document should be adjusted by comparing
historical local and global sea level trends.
_2/ Hansen, James E., A. Lacis, and D. Rind, 1983. "Climatic
trends due to increasing greenhouse gases," in Coastal
Zone '83, Orville T. Magoon, ed. New York: American
Society of Civil Engineers. 3:2796-810.
V Ibid.
V Ibid.
_5/ Charney, Jules, et al., 1979. Carbon Dioxide and Climate:
A Scientific Assessment. Washington, D.C.: National
Academy of Sciences Press.
6/ Smagorinsky, J. , 1982. Carbon Dioxide and Climate; A
Second Assessment. Washington, D.C.: National Academy of
Sciences Press.
7/ World Meteorological Organization, Sept. 1982. WHO Global
Ozone Research and Monitoring Project, Report No. 14 and
Lacis, A., et al., 1982. "Greenhouse effect of trace gases,
1970-1980." Geophysical Research Letters. 81:10:1035-8.
-------
- 12 -
Q_/ Keeling, C.D., et al., 1976. "Atmospheric carbon dioxide
variations at Mauna Loa Observatory, Hawaii." Tellus, 28,
and Rotty, Ralph, 1979. "Energy demand and global climate,"
in Man's Impact on Climate, Wilfred Bach, et al., eds. New
York: Elsevier Scientific Publishing.
9_/ World Meteorological Organization, op. cit.
10/ Seidel, Stephen and Dale Keyes, 1983. Can We Delay A
Greenhouse Warming? Washington, D.C: U.S. Environmental
Protection Agency. These authors demonstrate that strong
efforts to curtail fossil fuel growth, such as a tax that
quadruples the costs of these fuels, will delay the increase
in emission levels and temperature rise by only a few years.
An unconventional counterview that argues that it is theo-
retically possible to reverse the growth is provided in
Lovins, A., et al., 1981. Least Cost Energy. Andover,
Mass: Brick House Publishing Company.
ll/ Keeling, C.D., op. cit., and Rotty, R., op. cit.
12_/ Chlorofluorocarbons will have two effects on climate.
Their radiative effect will be warming. Although ozone
absorbs infrared radiation, the depletion of ozone in the
upper stratosphere by CFCs will have a slight warming
effect; the increased incoming ultraviolet radiation will
be greater than the increased outgoing infrared radiation.
Other perturbants, such as nitrogen oxides from airplanes,
may increase ozone in the lower stratosphere, which because
of pressure broadening, would make ozone a better infrared
obsorber, thus making increases at this lower altitude pro-
duce a net warming. Although the total temperature effect
of the changes in stratospheric ozone are still uncertain,
the warming contributed by this factor may be significant.
See Wuebbles, D., 1983. "Effect of coupled anthropogenic
perturbations on stratospheric ozone," Journal of Geophysical
Research. 88:C2:1444-1456.
13/ Charney, J., op. cit., and Smagornisky, J., op. cit. It
is estimated that the oceans take several decades to feel
the full effect of atmospheric warming. Thus, rnucn of
the equilbirium warming from greenhouse gases added to
the atmosphere in the 1970s has not yet occurred.
j.4/ Flohn, H., 1981. Life on a Warmer Earth. Laxenburg,
Austria: International Institute for Applied Systems
Analysis.
15/ Hansen, James E., A. Lacis, and D. Rind, 1983. op. cit.
16/ Hansen, James E., A. Lacis, and D. Rind, 1981. "Climate
impact of increasing atmospheric carbon dioxide," Science.
213:4511:957-66.
-------
- 13 -
CHAPTER 3
THE APPROACH, ASSUMPTIONS, AND MODELS
USED TO ESTIMATE SEA LEVEL RISE
Estimating global sea level rise requires considering all
of the possible changes in the social and natural systems that
could influence future climate and sea level.£/ Existing
scientific knowledge is inadequate to make a precise prediction
of the extent of these changes. Consequently, we generated a
range of estimates, called scenarios.
Each scenario used a different set of assumptions (see
Figure 3-1). For each factor that could affect sea level, a
range of assumptions was developed by consulting the literature
and the appropriate scientists. Thus, the full set of
assumptions covered the likely ranges for each factor, although
not necessarily the most extreme possibilities.
Models were selected that allowed various assumptions
to be combined to generate yearly estimates of sea level
rise. The models were chosen for their ability to accomplish
this task in a reliable, flexible, and cost-effective manner.
Many scenarios were generated, but four critical ones
were named. The "low scenario" consists of the most conservative
assumptions for each factor, while the "high scenario" consists
of the least conservative. Although the determinants of sea
level rise are not likely to be all at the high ends or low
ends of their ranges, these possibilities cannot be ruled out.
-------
- 14 -
FIGURE 3-1
SUMMARY OF HIGH ALS1D LOW ASSUMPTIONS
USED TO ESTIMATE SEA LEVEL RISE
Assumption
Low
High
Population Growth All scenarios assumed the world
will reach zero population growth
by 2075
Productivity Growth 2.2% per year; 3.5% per year;
decreases to decreases to
1.7% by 2100 2.2 % by 2100
Energy Technologies Best estimate, Best estimates
Nuclear costs halved
arbitrarily.
Unexpected Additions None None
To Fossil Fuel Base
Energy Conservation All Countries Move Toward High Efficiency
(60% improvement in energy efficiency)-
Fraction Airborne (CO2) — 53% ORNL Model; 60%
increases to 80 %
Nitrous Oxide 0.2% per year 0.7% per year
growth growth
Chlorfluorocarbons Emissions increase Emissions increase
0.7% of 1980 level 3.8% of 1980 level
per year per year
60-year half-life for CFC13 and 120-year
half-life for CF2Cl2
Methane 1% per year growth 2% per year growth
Temperature Sensitivity 1.5°C for CO2 4.5°C for CO2
doubling doubling
Heat Diffusion of Ocean 1.18 cm2/sec 1.9cm2/sec
Glacial Discharge Equal to Thermal Twice Thermal
Expansion Expansion
-------
- 15 -
Two mid-range scenarios were developed. They differ only
in their estimates of snow and ice contributions to sea level
rise; the estimates of the mass transfer of ice and snow from
land to sea are much less reliable than estimates of thermal
expansion. The mid-range scenarios used assumptions that fell
between those used in the low and high scenarios (except in
the case of fraction airborne, where we used the ORNL model).
Although it is currently impossible to estimate probabilities,
we believe that the actual sea level rise will probably fall
between the two mid-range scenarios.
The remainder of this chapter presents the assumptions
and models used to develop the scenarios. Technical details
about the assumptions and models are provided in Appendix B.
YEARLY RISES IN SEA LEVEL WERE SIMULATED
The yearly rise in sea level can be estimated by project-
ing yearly changes in the natural systems that determine it.
Four systems must be considered: atmospheric composition,
the climate, the oceans, and the cryosphere (the part of the
world that is ice or snow). Changes in the atmosphere will
produce changes in temperature, which in turn produce thermal
expansion and the transfer of snow and ice from land to sea
(Figure 3-2). Below, we also discuss possible changes in
other factors that determine global temperature, such as changes
in solar irradiation or volcanic aerosols. (A permament
increase in the average level of aerosols would block sunlight
and cool the earth.)
-------
- 16 -
Atmospheric
Composition
Other
Climate
Forcings
Figure 3-2
Yearly Modeling Sequence
Temperature
Change
Thermal
Expansion
Snow/Ice
Transfer
To Sea
Global Sea
Level Rise
Atmospheric Composition
Projecting changes in the atmospheric concentrations of
CC>2 and other greenhouse gases required projecting emissions of
these gases and their fate as a result of atmospheric chemistry
and other biogeochemical cycles that can create and destroy
them or remove them from the air. In general, only a percentage
of the emissions of each gas is retained in the atmosphere in
a chemical form that absorbs infrared radiation.
Estimating future C02 emissions required estimating future
economic growth and fuel use. Because economic growth is
determined by the rate of population increase, productivity
growth, and technological change, it was necessary to make
assumptions about each of these factors and to integrate them
in a consistent manner.
-------
- 17 -
The following assumptions were used:
0 Worldwide Population Growth was assumed to level
off by 2075 for all scenarios, based on the work
of Keyfitz et al .£/
0 Productivity Growth for labor was assumed to decline
for both the high and low scenarios.
- In the high scenario, the growth rate
began at 3.5 percent (per year) in 1980
and linearly decreased to 2.2 percent by 2100.
In the low scenario it linearly declined from
2.2 percent in 1980 to 1.7 percent by 2100.
The past 30 years have experienced a higher rate of
productivity growth than either of these assumptions.
0 Energy Production Technologies for various
fuels were assumed to improve:
for the low scenario, a best guess of future
technologies and costs was used.
for the high scenario the same estimates
were used except the cost of nuclear
energy was assumed to be cut in half in 1983.
0 Fossil Energy Resources were assumed to stay at
current levels with no new major discoveries.
° Energy Use per unit of output (measured in GNP)
was assumed to decline by over 60 percent between
1975 and 2100, reflecting two assumptions:
(1) as economies develop, they will become more
service oriented, and (2) energy technologies
will become more efficient, partly in response to
higher prices.
A world energy model developed by the Institute for Energy
Analysis at Oak Ridge Associated Universities was used to inte-
grate the assumptions.^/ The model divides the world into nine
regional sectors. Market mechanisms are used to select energy
sources and trade supplies in and between each region. C02
emissions are generated by using well-established coefficients
for each fuel.
-------
As mentioned, not all the CC>2 emissions remain in the
atmosphere. Some are removed by green plants, and some are
absorbed by the oceans. Unfortunately, knowledge of these
percentages is uncertain because the movement of carbon between
various "storage compartments" in the earth (the oceans,
biosphere, and the atmosphere) is only partially understood.
Two assumptions were made for predicting the fraction airborne:
o For the high scenario, a model of the carbon cycle
developed at Oak Ridge National Laboratories (ORNL)
was used ,_4/ The model represents, at a very high
level of aggregation, many of the physical and
chemical mechanisms that regulate the exchanges of
carbon between the various natural compartments that
absorb and release it. The fraction airborne increased
over time from 60 to 80 percent.
o For the low scenario, the percentage of carbon dioxide
remaining in the atmosphere was assumed to be equal to
its historical average of 53 percent.^/ This assump-
tion is conservative for several reasons. Like a
sponge, the upper layers of the oceans have a limited
absorption capacity.^/ Thus, as the emissions of
CO2 increase, the percentage taken up by the oceans
should decrease. Also, warming of the oceans and the
resulting changes in circulation could lower the capacity
of the oceans to absorb C02._
o For the mid-range scenario, the ORNL model was used.
High and low assumptions about CO2 emissions and the
fraction airborne were used to generate high and low estimates
of the yearly increase in atmospheric CO2» Under the low scen-
ario, atmospheric concentrations of CO2 will double by 2085,
for the high scenario, by 2055.
-------
- 19 -
Projections of the concentrations of chlorofluorocarbons
and other trace gases involved much less sophistication than
our projections of CO2 concentrations.^'
o The low scenario assumed that emissions of chloro-
fluorocarbons grew at 0.7 percent of the 1980 level
(1.8 million kg) annually until being capped in 2020
and that CFC13 (CFC-11) and CF2C12 (CFC-12) have
half-lives of 60 and 120 years, respectively.
o The mid-range scenario assumed that emmissions of
CFC13 and CF2C12 grew at 2.5 percent of the 1980 level
(6.4 million kg) per year until 2020 and then are
capped.
o The high scenario assumed that emissions of CFC13
and CF2C12 grew at 3.8 percent of the 1980 level
(9.8 million kg) per year until being capped in 2020.
The low assumption implies that atmospheric concentrations
will gradually approach an equilibrium concentration by 2100.
Although the high assumption implies that concentrations will
grow faster, it is not really an upper bound for likely future
concentrations. It is possible that chlorofluorocarbon
emissions will never be capped, or that their use in developing
countries will parallel that of the developed countries
(a growth not anticipated in the high scenario).
Nitrous oxide concentrations are also assumed to increase:^/
o The low scenario assumed 0.2 percent per year growth.
o The mid-range scenario assumed 0.45 percent per year
growth.
o The high scenario assumed 0.7 percent per year growth.
-------
- 20 -
Finally, methane is also assumed to grow:10/
o The low scenario assumed 1.0 percent growth per year.
o The mid-range scenario assumed 1.5 percent growth per
year.
o The high scenario assumed 2.0 percent growth per year.
Our analysis did not consider changes in ozone concentra-
tions at different altitudes, which may prove to be important.
Ozone depletion in the upper stratosphere caused by chlorofluro-
carbon emissions will tend to have a warming effect because
the loss of ozones' infrared absorbing capibility at this
altitude will be more than offset by the additional U-V energy
allowed to penetrate to the earth's surface. Higher ozone
levels in the lower stratosphere and upper troposphere, caused
by NOX emissions from airplanes, will cause additional warming
because at those altitudes the additional infrared that is
absorbed outweighs the reduction in ultraviolet that penetrates
to the surface._/ other minor greenhouse gases, such as CCl/,
CF4, NH3, and C2C13F3 (CFC-113) could also contribute to the
warming, but were not considered in this analysis.
Climatic Responses
As the atmospheric abundance of gases that absorb infrared
radiation increases, the earth's temperature will rise. The
extent of the ultimate warming will depend on how much the
initial warming alters the levels of other infrared-absorbing
gases such as water vapor, or its reflectivity, by changing
ice and cloud cover.l^/ These "feedbacks" may amplify the
-------
- 21 -
initial warming considerably. Because the magnitudes of these
factors are not precisely known, a large range of temperature
changes was used to represent all of these uncertainties:
o The high scenario used the National Academy of Sciences'
(NAS) estimate of a 4.5°C rise for a CC>2 doubling.
o The mid-range scenario used the NAS middle estimate
of 3.0°C.
o The low scenario used the NAS low estimate of 1.5°C.
To integrate these assumptions about thermal sensitivity
and the changes in atmospheric concentrations, we used an equa-
tion obtained from the Goddard Institute for Space Studies
based on a one-dimensional radiative convective model. (See
Appendix B). Coupled with a box-diffusion model that
calculates the heat absorbed by oceans, the equation allowed a
consistent integration of changes in greenhouse gases, thermal
sensitivity, and the oceans' absorption of heat.13/
Thermal Expansion of Ocean Waters
Because ocean waters circulate slowly, the deeper layers
of the ocean will warm very slowly, and for our purposes, can
be ignored. As the top layers warm, however, the ocean's
volume will expand. Although the percentage expansion is
small, the resulting sea level rise could be significant from
the human perspective.
-------
- 22 -
As discussed, the downward movement of heat was projected
with a simple box diffusion model. While this model does not
represent the ocean circulation processes that actually
transport heat, it is probably a good surrogate for the likely
effects of such processes for the time span of 100 years.
Nevertheless, to ensure the validity of. the sea level rise
estimates, we used different assumptions:
o For the high assumption, a rate of heat diffusion
compatible with higher estimates of observed movement of
chemical and radioactive tracers was used (1.9 cm^/sec).
o E'or the low assumption, a rate of 1.18 cm^/sec was used,
which is compatible with a more conservative
interpretation of tracer studies.14/
o For the mid-range assumption, the average of the high
and low assumptions (1.54 cm^/sec) was used.
The expansion of the seas was then computed using known
coefficients of expansion for the temperature, salinity, and
pressure of each ocean layer. We did not disaggregate thermal
expansion to reflect the geographical variation of ocean
temperatures. Because water expands by different amounts at
different temperatures, disaggregation would improve sea level
rise estimates. However, the one-dimensional description of
temperature and salinity provides a good first approximation.
The Cryosphere
Most discussions concerning sea level rise have focused
solely on the possible contributions of ice and snow resting
below sea level in West Antarctica moving into the sea,
-------
- 23 -
while only a few articles have discussed thermal expansion.
Yet no glaciologist has estimated the potential contributions of
ice and snow to sea level in the next century. In this section,
we discuss the several methods developed to perform this task.
Because the contributions of ice and snow to sea level rise could
be more important than thermal expansion, these methods should
be improved as soon as possible.
Size and Potential Source of the Land-Based Contributions
The amount of water contained in the ice and snow of
Antarctica, Greenland, Northern Europe, Asia, and North America
or in alpine glaciers is equivalent to about 70 meters of sea
level rise (see Table 3-1). The Arctic Ocean consists of sea
ice; therefore, its melting or breakup would not directly
raise sea level.
Global warming can influence the percentage of the earth's
water resting on land by several means:
o It can melt ice and snow. Melt water that runs off
into the sea will raise sea level.
o It can increase the rate of flow of land-based ice
sheets toward the sea. This deglaciation would add
ice to the oceans, thereby raising sea level.
o It can cause the atmosphere to carry more moisture to
cold areas in the form of snow, thereby increasing
snowfall accumulation and decreasing sea level.
Estimates of ice and snow contributions to sea level rise
can be made for each of these sources separately, or through
an aggregate estimating method. Both approaches are used
here.
-------
- 24 -
TABLE 3-1
DISTRIBUTION OF ICE AND SNOW
Area
Volume
Sea-Level
Equivalent
Snow and Ice on Land
Antarctica 1,2
Greenland
Small Ice Caps and Mountain Glaciers
Ground Ice (Excluding Antarctica)
Continuous
Discontinuous
Sea Ice
Arctic:
Antarctic:
Max.
Min.
Max.
Min.
106km2
13.9
1.8
0.5
7.6
17.3
15.5
8.4
20.0
2.5
106km3
water
28.0
2.7
0.24
5.0x104
2.0x104
3.0x104
5.0x103
Total Land Ice, Sea Ice and Snow
m
70
7
0.3
1.0
Jan.
July:
Global Mean
N.
S.
N.
S.
Annual
Hemisphere
Hemisphere
Hemisphere
Hemisphere
58
18
14
25
59
1 Excludes peripheral, floating ice shelves (which do not affect
sea level).
2 Roughly 10 percent of the Antarctic ice is in West Antarctica,
and 90 percent is in East Antarctica.
Source: N. Untersteiner, "Sea ice and ice sheets role in climatic
variations, Appendix 7," of GARP Pub. series 16: Physical
Basis of Climate and Climate Modeling. World Meteorological
Council of Scientific Unions, 206-24. April 1975.
-------
- 25 -
Predicting Snow and Ice Contribution Using Climate Models
The most comprehensive tool for understanding climatic
change is the three-dimensional general circulation model
(GCM). By solving a series of equations that represent the
fundamental laws of atmospheric structure and motion, such
models simulate the earth's weather on an hour-by-hour basis,
moving moisture and heat upwards and downwards, north and south,
east and west. Although the models represent the actual atmos-
phere, solar irradiation, and the reflectivity of the earth with
a fair degree of accuracy, the topography used is coarse;
variations in large geographic regions are averaged together,
and each grid is treated as a plateau.
By accumulating hourly and then daily weather statistics,
these models can generate years of data that represent the
climate of different regions. Since the models do not yet
represent all of the processes that determine climate, one
finds some discrepencies. Nevertheless, in most aspects, a
GCM simulates weather fronts realistically: rain, snow, heat
and cold are generated by the same mechanisms active in the
actual climate system. By comparing projections of these
models (assuming today's CC>2 levels) against observed weather
and climate, GCMs ahve been shown to accurately reproduce
the world's weather patterns.
-------
- 26 -
Several research groups have developed GCMs to explore
the effects of changes in the atmosphere's composition:
NOAA Geophysical Fluid Dynamics Laboratory (GFDL), the National
Center for Atmospheric Research (NCAR), and the Goddard Insti-
tute for Space Studies (GISS) have undertaken three well-known
efforts in the United States.^/ Each group has run its GCM
assuming the current atmosphere and doubled CC>2 concentrations.
While some of the model predictions differ about climatic
change, they are qualitatively similar.
The GISS group has used its model to estimate the size
and geographical distribution of snow and ice in a warmer
world. In fact, the group has done two doubled CC>2 experiments,
which differed only in their techniques for computing sea ice,
a key factor in determining the effects of atmospheric change
on Antarctica. One model run started with somewhat more sea
ice, and one with somewhat less sea ice than currently exists in
Antarctica. This provided a method to check the sensitivity
of the models to this initial condition.
Both doubled C0'2 runs showed a substantial net melting
of land-based snow and ice. Both "control" runs (i.e., using
current atmosphere and climate) showed nearly stable amounts
of snow and ice. Thus, the results of the four runs increase
our confidence that global warming will significantly decrease
snow and ice resting on land. The run deemed most accurate
showed a melting equivalent to a sea level rise of about one
-------
- 27 -
centimeter per year.^2/ (See Appendix C for details and
possible sources of error.) Snowfall accumulation increased
in the run, but was overwhelmed by additional melting. This
contrasted sharply with the results of the model runs using
today's temperature and CC>2, which showed almost stable
levels of ice and snow on land.
These results need to be analyzed in greater depth than
was possible for this study, and model experiments should be
made with a more realistic treatment of ice sheet decay.
Studies should also be made using other GCMs, and collateral
evidence should be reviewed and integrated into the analysis.
Nevertheless, these two GCM experiments provide evidence that
a global warming would decrease the mass of the ice sheets
by the melting and ensuing runoff.
One aspect of the GISS results that particularly needs
more analysis is the runoff of meltwater, not all of which
will escape to the sea. In many fringe areas of the ice sheets,
where much of the melting will take place, most of the water
probably will run off into the ocean; however, the percentage of
runoff will be low in the antarctic interior. Water that does
not run off will percolate into the ice and refreeze, which may
cause ice softening and crevasing.
Glaciologists should use GCM outputs and information about
glaciers and the topography of the land underlying ice to
estimate the percentage of meltwater that will run off. They
should also consider the degree to which changes in the physical
-------
- 28 -
characteristics of the ice sheet induced by refreezing will
lead to faster deglaciation. (A brief discussion appears in
the next section and in Appendix C.)
Another important feature of the GISS modeling is that the
estimated melting of one centimeter per year was computed on
the basis of a world whose atmospheric levels of C02 have
already doubled. Thus, the GCM results do not directly provide
a realistic time trend of sea level rise as CC>2 and other
greenhouse gases approach and then exceed this level of warming.
Computation of a time trend would have required estimating
(1) the percentage of the meltwater that runs off into the
oceans; (2) the degree to which the meltwater that does not
run off accelerates glacial movement; and (3) the scale of
melting that occurs as the earth warms. These steps could not
be undertaken in this study. Thus, instead of directly using
the one centimeter per year of melting as a scenario, the GISS
output was taken as a qualitative confirmation that ice and
snow can be expected to contribute significantly to sea level
rise.
Deglaciation as a Source of Land-to-Sea Transfer of
Snow and Ice
The possibility of a major deglaciation (the removal of
ice by a process other than melting in place) cannot be ignored.
Geological evidence suggests that changes in global temperatures
can cause the large ice sheets that rest on Antarctica to break
up and "fall" into the oceans. For example, during the last
-------
- 29 -
interglacial (120,000 years ago), which was about 1°C warmer
than today, the West Antarctic ice sheet may have completely
disappeared.^/
Several ice sheets that contain significant quantities of
water are thought to be vulnerable to warming. Thousands of
feet thick and resting mainly on land, these sheets are held
in place and prevented from entering the ocean by floating
"ice shelves" and pinnings below sea level. As global warming
occurs, the waters around these shelves and pinnings may become
warmer and melt them, allowing the ice sheets to begin to slide
into the oceans. Furthermore, meltwater that does not run off,
but percolates into the sheets and then refreezes, will tend to
soften the ice, making rapid movement of ice more likely.
Polar warming, which may be several times the magnitude of the
average global increase, thus makes it possible for "the bottle
to be uncorked."
Because it is marine-based (resting on the ocean floor),
the ice sheet most vulnerable to such a deglaciation is the
West Antarctic ice sheet. If all of that ice enters the ocean,
sea level will rise five to six meters. Bentley has estimated
500 years, and Hughes 200 years, as the earliest time this
could possibly happen.^/ Although both estimates were
made in the absence of detailed information about melting
rates, sea ice retreat, and ocean and air temperatures, the
possibility of a complete disintegration in 200 to 500 years
cannot be ruled out.
-------
- 30 -
Accurate estimation of the partial deglaciation of the
West Antarctic, East Antarctic, and Greenland ice sheets for
the next 120 years will require detailed studies of the specific
ice sheets. Such studies should consider such factors as the
predicted temperature of the upper surface of the ice, surface
precipitation rates, ocean water temperatures, melting of ice
shelves from the bottom, speeds of ocean currents and their
ability to remove ice, the specific topography of the "gates"
(narrow areas that constrict the flow of ice), and the specific
location of grounding lines (land on which marine ice sheets
rest)._^0/ These factors will determine the speed of discharges.
Unfortunately, such studies have not yet been made for deglaciation
in the next century.
Method Used to Estimate Snow and Ice Transfers from Land
to Sea
In the absence of appropriate studies of the various ice
fields, another approach had to be used that was much less
direct:
o A range of historic sea level rise estimates was
gathered from the literature.
o An estimate of the historical sea level rise
attributable to thermal expansion of the oceans was
obtained from the literature.
o High and low ratios of ice and snow contributions to
thermal expansion were computed using different estimates
of the historical sea level rise with the single estimate
of past thermal expansion.
o These ratios were assumed to remain constant in the future
(implying that snow and ice contributions will proceed
in marginal steps and that deglaciation, when averaged
across all ice sheets, will also be a proportional
phenomenon).
-------
- 31 -
Estimates of the Last Century's Sea Level Rise
In the last century, measurements of tidal heights have
been collected in harbors around the world. On the basis of these
measurements, several researchers have concluded that worldwide
sea level has been rising. (See Table 3-2 on the next page.)
Differences between sea level rise estimates can be attri-
buted to several factors that make precise interpretation of the
data difficult. Tidal gauges are influenced by local conditions
that do not influence worldwide sea level, such as river flow,
weather, and emergence or submergence of land. Furthermore,
they are not distributed uniformly, and large parts of the
oceans are unmeasured.
To overcome these problems, the various researchers employed
different approaches. Some researchers averaged all available
stations. Others chose a few stations in each typical geographic
zone and assumed them to be "representative." At least one
group of researchers also attempted to factor out local tectonic
influences.^/ Despite the differences, however, the estimates
of sea level rise have been remarkably similar. Most of the
researchers have concluded that worldwide sea level has risen
between 10 and 15 centimeters in the last century.22/
-------
-32-
TABLE 3-2
ESTIMATES OF GLOBAL SEA LEVEL RISE
(cm per century)
Author
Thorarinsson (1940)
Gutenberg (1941)
Kuenen (1950)
Lisitzin (1958)
Wexler (1961)
Fairbridge and Krebs (1962)
Hicks (1978)
Emery (1980)
Gornitz et al (1982)
Barnett (1983)
Estimate
> 5 cm
11 + 8 cm
12 to 14 cm
11.2 +_ 3.6 cm
11.8 cm
12 cm
15 cm (U.S. only)
30 cm
12 cm (10 cm excluding
long-term trend)
15 cm
Sources: Adopted from Barnett (1983) and Hicks (1978)
-------
- 33 -
Determining the Ratios
Using available estimates on past temperatures and the
rates of heat absorption in the oceans, Gornitz et al. estimated
past sea level rise due to thermal expansion to be 4 to 5 cm
the last century.23/
o For the low scenario, a 10-centimeter estimate of
historicalsea level rise was used, thus implying a
one-to-one ratio of the contribution of snow and ice
to thermal expansion.
o For the high scenario, the higher sea level rise
estimate of 15 cm was used, thus implying a ratio
of two-to-one.
Deficiencies in Our Estimates of Snow and Ice Contribution
Although the GISS 3-D experiments yield a similar magnitude
for the rate of sea level rise as the ratio method, we have no
illusions about the adequacy of these projections. If estimates
based on process models of deglaciation had been available, we
would have used them. The ratios depend on estimates of change
which are themselves subject to mis-estimation. Futhermore,
the physical basis for extrapolating the historical ratio
into the future is weak at best. Past "associations" may
not continue in the face of much larger temperature increases.
To remedy this situation, EPA and NASA have assembled
a team of glaciologists, oceanographers, and climatologists,
who, with limited funding, will use "process models" and
judgment to estimate a range for possible meltwater runoff
-------
- 34 -
and deglaciation. A comprehensive well-funded effort, however,
is still not on the research agenda of any federal agency.
CHANGES IN VOLCANIC AEROSOLS AND SOLAR IRRADIATION COULD
INFLUENCE CLIMATE
From decade to decade solar irradiation and volcanic
aerosols in the atmosphere may vary. An increase in aerosols
can have a cooling effect, while an increase in solar irradiation
will have a warming effect. In coming decades it is likely
that there will be fluctuations for both factors, but there is
no reason to expect a consistent long-range trend. Nevertheless,
to test tne possible influence of a strong shift in these
factors, these changes were examined:
o For the minimum case, we assumed high volcanic
activity and a reduction in solar radiation. A level
of optical thickness was chosen that was five times
the historical average over the last 80 years.24/
A linear trend equal to a 0.5 percent reduction in
solar radiation by 2100 was assumed.25/
o For the maximum case, we assumed solar irradiation
increased by a linear trend to a 0.5 percent increase
by 2100, with no change in volcanic activity.
o For the high volcanic case, we assumed high volcanic
activity, but no change in the solar constant.
These and other special-case scenarios are reported in
Appendix A. They suggest that the greenhouse effect is not
likely to be overwhelmed by volcanoes or changes in solar
irradiation.
-------
- 35 -
END NOTES CHAPTER 3
The figure below shows a schematic relationship of the
factors considered in estimating the sea level rise:
BASIS FOR SCENARIOS
POPULATION
GROWTH
PRODUCTIVITY
GROWTH
COST OF
ENERGY
TECHNOLOGIES
ENERGY
USE PER
UNIT OUTPUT
ATMOSPHERIC
GROWTH TRACE
GASES
1
ATMOSPHERIC
co2
CLI
SOLAR IRRADIATION
MATE
RESPONSE
VOLCANIC
AEROSOLS
I
THERMAL SENSITIVITY OF CLIMATE
SYSTEM TO CHANGES IN GREENHOUSE
GASES AND OTHER FORCIf
IGS
COEFFICIENTS
OF HEAT
DIFFUSION
THERMAL
EXPANSION
OF OCEAN'S
VOLUME
1
"ICE/SNOW-
WATER
MASS ADDED
TO OCEANS
LEVEL RISE
«
ESTIMATED
RATIO OF
"ICE/SNOW^
TO THERMAL
EXPANSION
SEA LEVEL RISE
For wch factor or r.lrtor«hlp high ml tow
wn itovrtopM imMg th. publW.«l Hwratur.
2.
3.
4.
5.
Keyfitz, Nathan, et al., 1983. Global Population (1975-2075)
and Labor Force (1975-2050). Oak Ridge, TN: Oak Ridge
Associated Universities, Institute for Energy Analysis.
Edmonds, Jae, and John Reilly, 1982b. "Global energy and
C02 to the year 2050." Washington, D.C.: Oak Ridge Asso-
ciated Universities.
Emanuel, W. R., and Killough, G. G., computer tape provided
with 17-layer ocean model. Oak Ridge, TN: Oak Ridge
National Laboratories.
Siegenthaler, Ulrich, 1983. "Uptake of excess C02 by an
outcrop diffusion model of the ocean." Journal of
Geophysical Research. 88:06:3599-608.
-------
- 36 -
6. Broecker, Wallace Smith, and Tsung-Hung Peng, 1982.
Tracers in the Sea. Palisades, N.Y.: Lament - Doherty
Geological Observatory, Columbia University.
7. Ibid.
8. Gibbs, Michael, Current and Future CFG Production and Use,
Washington, U.C.: ICF, 1983.
9. Weiss, R.F., Keeling, C.D., and Craig, H., "The Determi-
nation of Tropospheric Nitrous Oxide," Journal of
Geophysical Research, Vol. 86, No. 68, August 20, 1981,
pp. 7197-7202.
10. Rasmussen, R.A., Khalil, M.A.K., "Increase in the
Concentration of Atmospheric Methane" Atmospheric
Environment, Vol. 15, No. 5, 1981, pp. 883-886.
11. Lacis, A., personal communication, October, 1983.
12. Cloud responses would affect the total amplification
because cloud cover, heights, and their optical properties
(the amount of sunlight they reflect) could influence
and change the radiation balance and the earth's
temperature positively or negatively.
13. Hansen, James E., et al. 1981., "Climate impact of increasing
atmospheric carbon dioxide." Science. 213:4511:957-66.
The equation was modified by A. Lacis to allow the trace
gases to be considered. A box diffusion model portrays
the ocean as a column of water and treats the movement of
heat as a passive tracer. The GISS group adpoted the
model from Oeschger, H., et al., 1975. Tellus. 27:168.
14. Broecker, Wallace Smith, Tsung-Hung Peng, and R. Engh, 1980.
"Modeling the carbon system," Radiocarbon. 22:3:565:-98.
15. For example, an article in Newsweek, ("Is Antarctica
Shrinking", October 5, 1981) focused solely on the rise
in sea level that might be caused by ice transfers,
completely ignoring thermal expansion.
16. Manabe, Sykuro, and Richard T. Wetherald, 1975. "The
effects of doubling the CO2 concentration on the climate
of a general circulation model." Journal of the Atmospheric
Sciences. 32:1:3.
Hansen, James, E., et. al., 1983. "Efficient three-dimensional
global models for climate studies: models I and II."
Monthly Weather Review. 111:4:609:62.
-------
- 37 -
Washington, W.M. and D.L. Williamson, 1977. "A description
of the NCAR global circulation models in methods of
computational physics," in General Circulation Models
of the Atmosphere, J. Chang, ed. New York: Academic
Press, 17:111-172
17. Russell, G.; see Appendix C of this report.
18. Mercer, J.H., 1978. "West Antarctica ice sheet and CC>2
greenhouse effect: a threat of disaster." Nature.
271:5643:321.
19. Bentley, Charles R., 1983. "West Antarctic ice sheet:
diagnosis and prognosis," in Proceedings; Carbon Dioxide
Research Conference; Carbon Dioxide, Science and Consensus
1983, Washington, DC: Department of Energy. U.S. CONF-820970
Hughes, T., 1983. "The stability of the west Antarctic ice
sheet: what will happen," in Proceedings; Carbon Dioxide
Research Conference: Carbon Dioxide, Science and Consensus
1983. Washington, DC: U.S. Department of Energy. CONF-820970,
20. Ibid.
21. Gornitz, Vivian, S. Lebedeff, and J. Hansen, 1982. "Global
sea level trend in the past century." Science. 215:26:1611-4
22. Barnett, T.P., 1983. "Global sea level: estimating and
explaining apparent changes," in Coastal Zone '83, Orville
T. Magoon, editor. New York: American Society of Civil
Engineers. 3:2777-84.
23. Gornitz, V., S. Lebedeff, and J. Hansen, op. cit.
24. Lamb, H.H., 1970. "Volcanic dust in the atmosphere."
Philosophical Transactions of the Royal Society. London
Series A, 255:425.
25. Sofia, S., and A.S. Endal, 1982. "Solar variability."
Reviews of Space Physics.
-------
- 38 -
CHAPTER 4
SEA LEVEL SCENARIOS TO THE YEAR 2100
This chapter presents our estimates of sea level rise, using
the methods described in the previous chapter. It also discusses
alternative approaches for pro3ecting sea level rise that yield
similar estimates, increasing our confidence in these results.
Table 4-1 presents the four major scenarios. The low and
high scenarios use low and high assumptions, respectively, for
each of the factors influencing sea level rise. The two mid-range
scenarios use the medium estimate of sea level rise attributable
to thermal expansion. However, given the greater uncertainty
about snow and ice contributions to sea level rise, the mid-range
high and mid-range low use the two-to-one and one-to-one
assumptions for glacial contributions to sea level rise,
respectively.
TABLE 4.1
SCENARIOS OF FUTURE SEA LEVEL RISE
(centimeters)
Year
2000 2025
Scenario
High
Mid-range high
Mid-range low
Low
Current Trends 2.0-3.0 4.5-6.8 7.0-10.5 9.5-14.3 12.0-18.0
2000
17.1
13.2
8.8
4.8
2025
54.9
39.3
26.2
13.0
2050
116.7
78.9
52.6
23.8
2075
211.5
136.8
91.2
38.0
2100
345.0
216.6
144.4
56.2
-------
- 39 -
We believe that the actual rate of sea level rise is more
likely to fall between the two mid-range estimates than outside
of them. Because the high and low scenarios employ more extreme
assumptions, neither is likely to occur. However, the possibility
can not be ruled out.
SEA LEVEL RISE WILL ACCELERATE UNDER ALL ASSUMPTIONS
Even in the low scenario, sea level will rise twice as
fast as its historical rate in the next 20 years, and from
2000 to 2025, at three times the historical rate. In the high
scenario, the sea will rise about ten times its historical
rate from 1980 to 2025. Over the last quarter of the 21st
century, the high scenario predicts the sea to rise at over 40
times the historical rate.
The mid-range scenarios predict that from 1980 to 2000,
sea level will rise almost as much as it has in the last century.
Given local trends on much of the East and Gulf Coasts of the
United States, shoreline retreat in the next 20 years could
be one-half that of the past century. In the following 25 years,
worldwide sea level would rise at about eight times the historical
rate, creating much greater erosion and flooding problems.
Thus, in the next four decades (within the lifetimes of many
projects now under design) the sea is likely to rise twice as
much as it has in the last century.
AN ALTERNATIVE METHOD PRODUCES SIMILAR ESTIMATES OF SEA LEVEL RISE
To cross check our projections, we estimated sea level
changes by another method: extrapolating past associations
-------
- 40 -
between temperature and sea level. Sea level rise in the last
century has been estimated at 10 to 15 cm (4 to 6 inches).
The surface temperature rise for the same period has been
estimated at 0.4°C. Thus, the ratio of sea level rise to
temperature is somewhere between 25 and 37 cm for each degree.
Including the effects of trace gases, global warming should be
equivalent to at least a quadrupling of C02 by 2100, which
would raise surface air temperatures by 3.0°C to 9.0° (based
on the National Academy of Sciences' range for climate
sensitivity, ignoring delays caused by the heat absorbing-
capacities of the oceans). Using the 25 cm to 37 cm ratio for
the 3°C to 9°C range yields sea level rises of 75 cm to 333 cm,
These estimates are in line with those produced by our more
elaborate approach.
One other result is worth noting. Using a time series
regression, Chylek and Kellogg estimated a 10 cm to 25 cm sea
level rise per 1°C temperature rise. For a temperature rise
of the magnitude projected, here, their result is consistent
with a sea level rise of 30 cm to 225 cm.^/
END NOTES TO CHAPTER 4
1. Chylek, Peter, and William Kellogg, 1982. "The Sea Level
Climate Connection," unpublished paper, National Center
for Atmospheric Research, Boulder, Colorado.
-------
- 41 -
CHAPTER 5
IMPACTS OF SEA LEVEL RISE
The objective of this report is to project sea level
rise, not to evaluate its impacts. Nevertheless/ because our
reason for estimating sea level rise is to help decision makers
and professionals anticipate its effects and evaluate its
importance, we provide a brief discussion of the impacts in
this section.
THE MOST SIGNIFICANT DIRECT IMPACTS WILL BE SHORELINE RETREAT,
INCREASED FLOODING, AND SALT INTRUSION
Sea level rise will have three major types of physical
effects: shoreline retreat, increased flooding, and landward
movement of saltwater. Shorelines will retreat because very
low land will be innundated and other land along the shore
will erode. For example, a thirty centimeter (one foot) rise
in sea level would erode most sandy beaches along the U.S.
Atlantic and Gulf Coasts at least thirty meters (one hundred
feet).
Even the historical sea level rise trend may be causing
significant erosion._£/ Most shorelines are maintained by the
balance between sediment eroded by storms and sediment deposited
back on the beach by waves during calm periods. However, a
rise in sea level allows storm waves to strike and erode the
beach farther inland, and makes calm waves less effective at
"dredging" sand from the ocean floor and redepositing it onto
the beach.
The Bruun Rule (illustrated by Figure 5-l)^/, a method
-------
- 42 -
for quantifying the erosion impact, has been validated by
field tests on the Great Lakes.jV The shore maintains the
equilibrium profile shown by the solid line. When sea level
rises, this profile must also rise, which requires additional
sediment equal to b'. Unless sediment is transported into the
area by currents or by mechanical dredges, ocean waves will
provide the sediment by eroding enough sand from the upper
portion of the beach to equal _b, producing the new profile
shown by the dotted line. Applications of this theory have
concluded that a rise in sea level of one centimeter can cause
the shore to erode between one hundred and one thousand
centimeters on the Atlantic and Gulf coasts of the United States,
FIGURE 5-1
BRUUN RULE
BEACH
, x SEA LEVEL INITIAL SEA LEVEL
\b\ XAFTER RISE / \
BOTTOM PROFILE/
AFTER SEA
LEVEL RISE
INITIAL BOTTOM
PROFILE
LIMITING DEPTH —
BETWEEN PREDOMINANT
NEARSHORE AND
OFFSHORE MATERIAL
SHORE EROSION FOLLOWING A RISE IN SEA LEVEL ACCORDING TO
THE BRUNN RULE (AFTER SCHWARTZ 1967) "THE BRUNN THEORY OF
SEA LEVEL RISE AS A CAUSE OF SHORE EROSION, " ADOPTED FROM
THE JOURNAL OF GEOLOGY, 75, 76-92.
-------
- 43 -
Low-lying areas not lost to a rising sea will experience
increased flooding, for a number of reasons. The higher sea
level will provide a higher base on which storm surges can
build. Beach erosion and deeper water may allow large waves
to strike further inland. Finally, higher water tables will
decrease the land's drainage capacity, increasing runoff
during storms.
Sea level rise will also cause salt water to move landward,
intruding into groundwaters, rivers, and estuaries. In some
rivers, salt may move upstream tens of kilometers. This effect
may alter local availability of fresh water and alter ecosystems
in some areas.^/
IMPACTS WILL DEPEND ON HOW PEOPLE ANTICIPATE SEA LEVEL RISE
Economic and environmental impacts will depend on how well
people anticipate and plan for the physical changes associated
with sea level rise. As busineses, governments, and individuals
make decisions in coastal and low-lying areas, they have the
opportunity to adjust to sea level rise before it occurs, which
will decrease the eventual impacts of erosion, flooding, and
saltwater intrusion.
Many decisions have outcomes that last long enough to be
affected by sea level rise, e.g., where to locate roads,
wastewater treatment plants, and chemical and nuclear waste
storage facilities, (See Table 5-1.) A coastal highway may
determine development patterns long after the pavement and the
structures along the road have been replaced. Similarly, although
a nuclear power plant built today may only be designed to operate
-------
-44-
TABLE 5-1
CATEGORIES OF DECISIONS SEA LEVEL RISE WILL INFLUENCE
LOCATIONAL DECISIONS
• WHERE TO PUT PRIVATE DEVELOPMENT AND REDEVELOPMENT
-HOUSING
- FACTORIES
- RESORTS
- ENERGY FACILITIES
- HAZARDOUS WASTE SITES
• PUBLIC DEVELOPMENT DECISIONS-ROADS
- UTILITIES
- PORT INFRASTRUCTURES
- PARKS
- BRIDGES
• PURCHASE OF LANDS FOR CONSERVATION
STRUCTURAL AND SITE DESIGN DECISIONS
• HOW TO BUILD FACILITIES
- THEIR MOVABLILITY
- SITE CONTOURING
- CONSTRUCTION TYPE AND QUALITY
- PLANNED LIFETIME OF STRUCTURE
• R&D ON HOW TO IMPROVE OPTIONS
- SUCH AS MAKING STRUCTURES MORE "SEA LEVEL RESISTANT"
- MAKING STRUCTURES MOVE MOVABLE
• HOW TO MAKE LOW-COST DESIGN CHANGES TO REDUCE ADVERSE EFFECTS
PROTECTIVE MEASURES AGAINST FLOODS AND EROSION
• PROTECTIVE FACILITIES SUCH AS SEAWALLS
- HEIGHT
- TYPE
- FOUNDATION SIZE (SO THEY CAN BE EXPANDED LATER)
• BEACH NOURISHMENT DECISIONS
• VEGETATION PLANTING AND MAINTENANCE DECISIONS
• RIVER CHANNELING AND RECHANNELING DECISIONS
• LAND ACQUISITION AND SET ASIDE FOR PUBLIC AND PRIVATE WORKS
FOR FUTURE PROTECTION
• LOCAL ZONING AND OTHER LAND-USE CONTROLS TO REDUCE DEVELOPMENT
IN WRONG AREAS
• FLOOD PROTECTION REQUIREMENTS FOR HAZARDOUS FACILITIES
DECISIONS ABOUT FLOOD MITIGATION PLANNING
• EVACUATION PLANS
• POST-DISASTER PLANS
• INSURANCE POLICIES, SUBSIDIES AND COSTS
-------
- 45 -
for 30 years, its location may become the only available site
for power generation in a given region for hundreds of years.
The wide range of sea level rise scenarios makes planning
more difficult than it would be if we were certain about a
particular forecast. Nevertheless, prudence demands that
decision makers plan for at least the low sea level rise scenario,
which would change many decisions. For other decisions, the
low scenario presents few risks, but the high scenario would
pose major problems. In such cases risk analysis can be used to
balance the benefits and costs of various planning options for
for each sea level rise scenario.
QUANTITATIVE ESTIMATES SHOW LARGE PHYSICAL EFFECTS
AND ECOLMOMIC IMPACTS FROM ANTICIPATED SEA LEVEL RISE
Few studies of the physical consequences of sea level
rise have been undertaken. Schneider and Chen concluded that
a 4.7-meter (15-foot) rise in sea level would innundate over
one-fourth of Louisiana and Florida, as well as one-eighth of
Delaware, Maryland, and the District of Columbia.5/
EPA has funded case studies on the impacts of sea level
rise for the areas around Charleston, South Carolina and
Galveston, Texas. Detailed results of these studies are
available in Sea Level Rise Conference Document.6/ A forth-
coming book entitled Sea Level Rise to the Year 2100, will
contain papers by the same researchers, as well as the reactions
of coastal decision makers who attended that conference.^/
-------
- 46 -
Both case studies found that sea level rise will have
significant impacts. Research Planning Institute concluded
that a 1.5 meter (five foot) rise in sea level would innundate
one-quarter of Charleston if no additional bulkheads or seawalls
were constructed.^/ They also calculated that areas in
Charleston that are now flooded once every 100 years would be
flooded once every 10 years. Leatherman concluded that a
1.5 meter rise would claim much less land around Galveston,
provided that the existing network of levees and seawalls were
maintained.^/ However, he estimated that such a rise would
double the area flooded every 15 years. Because Galveston
is already vulnerable to storms, Gibbs concluded that the
resulting annual storm damage in this area would increase to
$105 million from $23 million._W
Quantitative projections of the effects of sea level rise
on salinity in surfacewater and groundwater have only been made
in limited cases. The Delaware River Basin Commission has
estimated the effects of historical sea level trends on salt
concentrations in the Delaware River.££/ it plans to examine
the effects of the scenarios reported here, which may threaten
aquifers recharged by the river and possibly Philadelphia's
current water intakes during droughts However, no comprehensive
analyses have been undertaken of salinity intrusions from sea
level rise.
-------
- 47 -
CURRENT AND FUTURE EFFORTS TO STUDY SEA LEVEL RISE IMPACTS
The National Academy of Engineering's Marine Board plans
to assess the engineering implications of sea level rise. -*-2/
Port and coastal structures are designed for lifetimes of 100
years and possibly longer. The Marine Board is interested
in determining whether the evidence of possible sea level rise
justifies designing structures to withstand such a rise. It
also intends to determine whether the existing network of tidal
gauge stations is sufficient to detect a CO2~induced rise
in global sea level. EPA is working with the Delaware River
Basin Commission to determine the impact of sea level rise on
drinking water. Projects in Maryland, New Jersey, and New
York are being considered, and others are being sought, j-3/
The prospect of sea level rise of the magnitude estimated
in this study is so new, however, that much more basic research
and study needs to be done on its effects, the methods for
estimating them, and on the economic importance of anticipating
the rise.
A variety of academic and professional disciplines need to
define research and action agendas for understanding and
preparing for sea level rise. B'or example, planners should
review zoning standards and architects should review building
codes. Coastal geologists should develop better means of
projecting erosion. Paleo-sea-level researchers should consider
how their substantial body of knowledge can be used for
-------
- 48 -
understanding the response of the land to sea level change,
for validating climatic models, and for adjusting global sea
level estimates by regional influences.^/
The need for additional research is further underscored
by EPA's pilot studies, which demonstrate that half the damages
can be prevented with adequate planning. Ignoring this
opportunity could be very costly to society. In the next
chapter, we discuss the research necessary to improve sea
level rise projections. However, preparing for sea level rise
will also require a much better understanding of its effects.
END NOTES TO CHAPTER 5
1. Pilkey, Orrin, et al., 1981. Saving the American Beach;
A Position Paper by Concerned Coastal Geologists. Results
of the Skidaway Institute of Oceanography Conference on
America's Eroding Shoreline. Savanah: Skidaway Institute
of Oceanography.
2. Bruun, Per, 1962. "Sea level rise as a cause of shore
erosion." Journal of Waterways and Harbors Division.
New York: American Society of Civil Engineers. 1:116-30.
Schwartz, M.L. 1967. "The Bruun theory of sea level rise
as a cause of shore erosion." Journal of Geology. 75:76-92,
3. Hands, Edward B., 1980. Prediction of Shore Retreat and
Nearshore Profile Adjustments to Rising Water Levels on
the Great Lakes, Technical^ Paper Number 80-7. Fort
Belvoir, VA: US Army Corps of Engineers, Coastal
Engineering Researcn Center.
Weggel, J.R., 1979. AMethod for Estimating Long-Term
Erosion Rates from a Long-Term Rise in Water Level,
Coastal Engineering Aid Number 79-2. Fort. Belvoir, VA:
U.S. Army Corps of Engineers, Coastal Engineering Research
Center.
-------
- 49 -
4. Haskin, H.H., and Tweed, S.M., 1976. Oyster Setting and
Early Spat Survival at Critical Salinity Levels on Natural
Seed Oyster Beds of Delaware Bay. New Brunswick, NJ:
Water Resources Research Institute, Rutgers University.
5. Schneider, Stephen H. and Robert S. Chen, 1980. "Carbon
dioxide flooding: physical factors and climatic impact,"
Annual Review of Energy. 5:107-40.
6. This Document is Available from EPA.
7. Barth, Michael and James Titus, eds. Sea Level Rise to The
Year 2100. Stroudsburg, PA: Hutchinson Ross, (in press).
8. Kana, Timothy w., et al., 1983. "Shoreline changes due
to various sea level rise scenarios," in Coastal Zone '83
Orville T. Magoon, ed. New York: American Society of
Civil Engineers. 3:2768-76.
Kana, Timothy W., et al., "The Physical Impact of Sea Level
in the Area of Charleston, South Carolina," in Barth, Michael
and James Titus, eds., op. cit.
9. Leatherman, Stephen, 1983. "Geomorphic effects of projected
sea level rise: a case study of Galveston Bay, Texas,"
in Coastal Zone '83, Orville T. Magoon, ed. New York:
American Society of Civil Engineers. 3:2890-901.
Leatherman, Stephen, "Coastal Geomorphic Responses to
Sea Level Rise: Galveston Bay, Texas," in Barth, Michael
and James Titus, eds., op. cit.
10. Gibbs, Michael, 1983. "Economic effects, and value of
information on sea level rise," in Coastal Zone '83,
Orville T. Magoon, ed. New York: American Society of
Civil Engineers. 3:2754-67.
Gibbs, Michael, "Economic Analysis of Sea Level Rise:
Methods and Results," in Barth, Michael and James Titus,
op. cit.
11. Hull, C.H.J., and R.C. Tortoriello, 1979. "Sea level
trend and salinity in the Delaware estuary." Staff paper.
West Trenton: Delaware River Basin Commission.
12. Gallagher, Aurora, personal communication to James Titus,
Marine Board, National Academy of Sciences, 2101
Constitution Avenue, N.W., Washington, D.C. 20418.
-------
- 50 -
13. Interested parties should contact: James Titus.
U.S. EPA, (PM-220) Washington, D.C. 20460.
14. Lisle, L.D., 1982. "Annotated bibliography of sea level
changes along the Atlantic and Gulf coasts of North America."
Shore and Beach. 50:3.
Shennan, Ian, University of Durham, personal communication
with John S. Hoffman
-------
- 51 -
CHAPTER 6
RESEARCH NEEDED TO IMPROVE
ESTIMATES OF SEA LEVEL RISE
Little research has been conducted to estimate future sea
level rise. Improving the scenarios developed in this report
will require appropriate research and accurate monitoring of
the underlying physical systems. Although some of the scientific
problems may resist solution, a larger and better-focused
research effort could solve most of them in time for the resulting
information to be useful to coastal decision makers. A key
determinant of progress will be the priority society attaches
to this research.
This chapter examines the sources of variation in the
estimates of sea level rise, to help identify the most promising
opportunities for additional research. It also discusses
principles for managing this research.
MONITORING SEA LEVEL WILL NOT BE SUFFICIENT
The low and high scenarios differ by a factor of six.
This range makes it difficult for managers to decide how to
reduce the adverse impacts of sea level rise. A "wait and see"
approach would eventually reveal which scenario is most accurate.
But because most forthcoming decisions cannot be postponed
the several decades that this might require, these managers
need a smaller uncertainty range. This can be achieved by
reducing the uncertainties surrounding the individual assumptions
that must be made to project sea level rise.
-------
- 52 -
Interpreting the observed sea level rise will be difficult.
Many factors that cause short-term variations in global
temperature and sea level have not been modeled, including the
internal dynamics of the climate system, changes in ocean
circulation, and year-to-year fluctuations in volcanic activity.
Thus, if the sea rises 9 to 13 centimeters by 2000 (the most
probable range), then it will be difficult to determine the
amount of this rise caused by the greenhouse warming, as
opposed to temporary fluctuations caused by other factors,
unless better monitoring and research is undertaken. It will
be even more difficult to determine the percentage of the rise
that should be attributed to thermal expansion versus glacial
contribution. Better research and monitioring could help
ensure that our ability to forecast sea level rise improves as
additional data becomes available.
UNDERSTANDING THE SOURCES OF DIFFERENCES BETWEEN SCENARIOS CAN
HELP FOCUS RESEARCH PRIORITIES
The estimates of thermal expansion vary by a factor of four.
However, the greatest source of uncertainty is the rate of
transfer of snow and ice from land to sea, which varies by a
factor of eight between the low and high scenarios. Unfortunately,
there has been insufficient funding of coordinated work between
glaciologists, climate modelers, and Southern Hemisphere
oceanographers to estimate this important source of sea level
rise.
Examining the sources of uncertainty is a first step
toward developing research priorities. Figure 6-1 shows the
-------
- 53 -
contribution of each major factor to the current uncertainty
of sea level rise from thermal expansion, using the low scenario
as a baseline. Assumptions were changed one-at-a-time from
low to high for each factor. The first four factors (CC>2
emissions, chlorofluorocarbons, nitrous oxide, and methane)
were approximately additive in their effects. Because the
other factors were not, changing them in a different order
would change these estimates somewhat. Nevertheless, this
figure conveys the relative importance of the uncertainties
for each of the various factors in 25-year time spans.
This figure reveals some important insights for setting
research priorities. The best way to improve estimates of sea
level rise to the year 2025 would be to improve estimates of
temperature (climate) sensitivity, and next, the concentrations
of the trace gases. In the longer term, the level of CC>2 becomes
quite important.
The atmospheric fraction airborne for CC>2 and diffusivity
do not appear to be the major sources of uncertainty. It is
possible that increasing temperatures may stratify the ocean,
and increase fraction airborne and decrease diffusivity beyond
the ranges of our assumptions. However, although these effects
would both raise atmospheric temperatures, their effects on sea
level rise would tend to offset one another.
Appendix A shows that in the long run, the impact of
extreme and unexpected forms of natural variation (volcanoes
and solar radiation) are unlikely to be important, except in
the early stages when they could mask the overall trend.
-------
- 54 -
Unless these factors are carefully monitored, their variability
could reduce the ability of scientists to understand the
other factors.
Figure 6-1
RELATIVE CONTRIBUTION OF FACTORS TO THE DIFFERENCE
IN SEA LEVEL RISE BETWEEN THE HIGH AND LOW SCENARIOS
C02
Fraction
Airborne
QC
_C
05
O
£
I
Q
o>
Q_
30 -
10-
Thermal
Sensitivity
Heat
Diffusivity
CFCs
C0?
Emissions
2000
2025
2050
2075
2100
NOTE: Factors were evaluated in the order that they appear starting at the top.
-------
- 55 -
OPPORTUNITIES EXIST FOR MORE RESEARCH
The greatest needs are:
o projecting the transfer of ice and snow from land
to sea;
o estimating the atmospheric changes in the trace gases
methane, nitrous oxide, and the chlorofluorocarbons;
and
o estimating the sensitivity of the climatic system to
atmospheric and other changes.
As discussed earlier, in response to these findings, EPA,
in cooperation with NASA, has already initiated work to improve
our understanding of snow and ice melting, oceanic warming
around Antarctica, and glacial ice discharge. An inter-
disciplinary team of scientists from the Goddard Institute
for Space Studies (GISS), NASA, and Lamont Doherty Geological
Observatory has formed to improve the scenarios. They will
use output from the GISS general circulation models; ice sheet/ice
shelf process models; estimates of oceanic changes provided by
a southern hemisphere oceanographer; and a thorough literature
review of ice fields.
In early 1984, another scenario report will be issued
that incorporates the results of this project and refines the
thermal expansion model so that it reflects a geographically
disaggregated absorbtion of heat by the world's oceans. That
effort should increase the reliability of the scenarios somewhat.
In the longer term, however, more precise and reliable estimates
of sea level rise will become available only if efforts are
made to improve our underlying knowledge of the relevant
natural systems.
-------
- 56 -
Several projects need to be undertaken to improve the
basic scientific understanding of glacial processes:
o Better Monitoring of the Ice Packs
The amount of ice and snow in all ice fields must be
carefully tracked, probably with a combination of
observations from space and land.
o Experiments with Ice Sheets
Basic aspects of ice processes must be studied,
especially those that relate friction and ice
movement and the influence of warming and refreezing
on the forces that determine ice movement.
o BetterTopographical Surveys
Land under ice sheets that could affect the rate
of discharge must be mapped better.
o Parametric Studies with ice Models
Studies that consider different interpretations
of ice processes and environmental conditions need
to be conducted with ice sheet and ice -shelf models.
o Improvements in Polar Climatology, Sea Ice, and
Oceanography in Climatic Models
Representations of the climatic processes, melting
of sea ice, and the oceanographic processes need to
be improved, with interdisciplinary teams contributing
to ensure that general circulation models treat these
areas realistically.
A major effort should be undertaken to narrow the uncertain-
ties regarding trace gases.
o Develop Better Estimates of Sources and Sinks
Research has been insufficient to even determine
the origins and fates of these gases
o Develop a Better Understanding of Atmospheric Chemistry
The interactions of these gases in the atmosphere
greatly influence their abundance, and should be better
assessed.
-------
- 57 -
o Integrate Atmospheric Chemistry and Climatic
Modeling
Because atmospheric chemistry depends on temperatures
and atmospheric mixing, atmospheric chemists and climate
modelers must work in closer contact with each other.
Knowledge of the speed of climatic change and the sensi-
tivity of the climatic system to changes in atmospheric composi-
tion can be greatly enhanced in several ways.
o Develop a Better Representation of the Oceans
Dynamic ocean models can be integrated into general
circulation models in realistic ways in the next 10
years only if a major, well-funded effort is undertaken.
This is critical to producing a more valid geographical
distribution of precipitation and temperature changes
needed to estimate ice and snow changes.
o Improve Data on Clouds
Better observational data, a better theoretical under-
standing, and a better computational representation
of cloud processes can reduce tne uncertainty of the
effect that clouds have on thermal sensitivity.
o Expand Computing Capability
The acquisition and use of appropriately sized computers
would allow more experiments, with different
representations of various processes, to be run in
general circulation models and also provide greater
geographical resolution. These efforts would allow
quantification of the uncertainties in regional
precipitation, which would considerably improve estimates
of ice melting and snowfall.
Sufficient personnel and resources are available to sustain
present efforts. Over the longer term, however, progress will
require training new people. While not particularly costly,
this process takes time. Delaying its start would diminish
society's ability to accelerate research later.
-------
- 58 -
RESEARCH CAN BE MANAGED MORE EFFECTIVELY
Research to improve estimates of sea level rise should
have three primary objectives:
o Establish interdisciplinary Teams of Sufficient Size
Progress will be impeded if interdisciplinary teams
are not developed and brought to work together on a
long-term basis. They must be large enough to address
difficult long-term problems.
o Provide Long-Term and Secure Financial Support
Interaisciplinary scientific teams will not evolve
unless financial support is secure. The development
and maturation of research efforts depends on a steady
source of funding. These teams should be directed by
eminent scientists.
o Develop Long-Term and Geographically Extensive Data
Sets
Efforts to collect data must ensure that geographically
extensive data sets are collected over long periods of
time without interruptions. The greater the extent
and detail of the data, the more scientists can test
their models for realism.
Although success in scientific research cannot be guaranteed,
it can easily be thwarted by failing to develop the conditions
that make it possible. The interdisciplinary nature of projecting
sea level rise makes management of research a particularly
important part of ensuring progress. The opportunity exists;
the decision facing society is whether or not to employ the
resources needed to meet this challenge.
-------
-59-
APPENDIX A
SUMMARY OF SEA LEVEL RISE SCENARIOS INCLUDING SPECIAL
CASES OF INCREASED VOLCANIC ACTIVITY AND
CHANGES IN SOLAR RADIATION
The abbreviations used in the following tables are explained below:
CO2 Scenario
no growth -- constant (1975) emissions
low growth — low (1.674 %) annual growth in emissions
medium growth -- medium (2.074 %) annual growth in emissions
high growth — high (2.345 %) annual growth in emissions
Thermal Sensitivity (Te)
low — 1.5°C per doubling of CO2
medium — 3.0°C per doubling of CO2
high — 4.5°C per doubling of CO2
CH4 Scenarios
no growth -- constant 1980 concentrations
vl (very low growth) — 0.5 % annual compound growth
Ig (low growth) -- 1.0 % annual compound growth
med (medium growth) — 1.5 % annual compound growth
hg (high growth) -- 2.0 % annual compound growth
vh (very high growth) — 2.5 % annual compound growth
N?0 Scenarios
no growth -- constant 1980 atmospheric concentrations
vl (very low growth) — 0.1 % annual compound growth
Ig (low growth) — 0.2 % annual compound growth
med (medium growth) -- 0.45 % annaul compound growth
hg (high growth) — 0.70 % annual compound growth
vh (very high growth) — 0.90 % annual compound growth
CFC Scenarios
no growth -- constant 1980 emissions
Ig (low growth) — 1.8 million kg annual growth to 2020
med (medium growth) -- 6.4 million kg annual growth to 2020
hg (high growth) — 9.8 million kg annual growth to 2020
vg (very high growth) — 11.6 million kg annual growth to 2100
All Scenarios assume that emissions are constant after 2020 and that
the half-lives of CF2Cl2 (CFC-12) and CFC13 (CFC-11) are 120
and 60 years (i.e., "lifetimes" are 150 and 75 years) respectively.
-------
-60-
Volcanic Activity
average -- Historical average optical depth value of .007
high -- optical depth of .037 (average of highest two decades
in last century. Both values are dimensionless; see
Appendix B)
Solar Luminosity
average -- no net change in historical luminosity
low -- Linear decrease from historical average to 0.5 % less
in 2100
high -- Linear increase from historical average to 0.5% more
in 2100
Diffusivity Coefficient (k)
very low — 0.20 cm2/sec
low — 1.18 cm2/sec
medium — 1.54 cm^/sec
high — 1.9 cm2/sec
very high — 4.0 cm^/sec
CO? Retention (Fraction Airborne)
ORNL — estimates from the ORNL carbon cycle model (.6 to .8)
low -- a constant traction of 0.53
-------
-61-
o
o
CM
in
r-
o
CM
O
in
o
CM
CM
O
CM
O
o
o
CM
CM 0) C
8
as
>1 4-1
4-> C
•H 0
> -H
•H U
in -H
D 'HH
LM 4-1
'HH 0
•H O
Q U
Trace Gases
P tn -H
0 C >
£ 0 -H
EH W 4-1
o
•H
VH
CM 03
8 0
u
_C£
=#:
0
in
03
u
Cd
U
CO
o
CN
Cd
O
K
i
M
s
CM
CM
^
i-D
m
^
CO
CM
rH
on
rH
^
^
>
rH
C
o
c
o
• H
in
c
03
a
X
Cd
rH
0
,£
EH
rJ
O
•H
T3
0
g
•
rH
0
>
03
rH
0
>
03
•H
•H
•8
g
rH
^r
^r
rH
CM
rH
(Ti
CM
m
CM
CM
00
00
0
S1
03
-C
O
in
H
Q
0
U
M
rH
• •
—I
5
H
3
^O
^o
i~H
CM
00
IJO
00
rH
oo
^
oo
(Ti
OO
CM
CO
rH
0
r?
03
G
O
in
•H
Q
0
U
M
rH
CM
4-J
•H
£
o
m
rH
m
0
Q*\
9
oo
oo
oo
oo
rH
r-~
m
rH
O
O
03
a
X
Cd
rH
0
in
0
cn
G
M
0
03
M
0
03
cn
G
en
•rH
cn
-C
cn
,C
•rH
x:
CM
0
m
OO
m
rH
CM
•
rH
rH
(Ti
m
rH
r-
rH
0
2
03
U
in
Q
0
U
M
rH
CM
5
rH
00
CM
0
•
•
rH
M
m
•
i-D
^r
•
CM
Thermal Expansion Only
rH
rH
aver aver
rH
rH
I-H
rH
rH
OO
CM
•
VD
m
0
•
oo
00
00
•
oo
CM
0
•
co
rH
oo
•
^r
with 1:1 Ice Discharge
MAXIMUM (Major solar increase acompanies greenhouse warming)
o
•
CO
^r
rH
•31
•
>£>
oo
CTi
•
^
^r
^D
•
rH
CM
in
•
^D
Thermal Expansion Only
j
O
.c
Oi
•H
.C
-C
aver
^
>
-C
j^
>
r;
01
•rH
-C
-C
CTi
• H
.c
•^
-H
in
c
03
volcanic activity simuJ
c
•H
m
0
10
0
rH
B
•rH
03
O
•H
4->
03
Hi
rH
•rH
rH
03
o
UJ
C
• H
in
0
in
iU
0
u
0
rH
O
03
S
i
2
M
S
O
•
r-
rH
(Ti
O
rH
m
00
oo
rH
rH
rH
O
C
o
• H
03
-H
03
0
EH
rH
O
rH
0
rH
-C
en
• H
G
^
rH
rH
1
1
m
1
o
Cd
1
g
CN
8
en
o
CO
rH
rH
i— 1
•
O
rH
m
CO
•
CM
Thermal Expansion Only
j
O
•H
aver
aver
?
H1
S1
•H
-C
rH
vo
M
i-^
DIFFUSE
%
Cd
U
2
CM
2
M
rH
vD
ro
CM
CM
rH
OO
CM
•
OO
Thermal Expansion Only
D
rH
0
aver
aver
1
1
1
• H
medium
-
00
vo
CM
oo
0
•
m
CM
m
CM
rH
CM
•
Thermal Expansion Only
O
-H
aver
rH
0
03
I
-o
"O
medium
medium
oo
CM
m
in
:M
CO
rH
m
•
O
C
o
H
03
Cd
rH
0
H
0
.C
cn
•H
si
aver
aver
1
1
1
• H
medium
<*
'N
00
OO
OO
m
oo
•
0
00
m
m
rH
•
Thermal Expansion Only
D
en
H
-C
0
0
03
rH
0
03
1
i
1
medium
medium
o
H
M
M
EH
M
rn
z
1
Cd
CM
O
Cd
U
2
Cd
H
2
M
O
0
cn
rH
•
^O
in
o
•
CM
oo
r-
•
m
i — i
o
•
m
Thermal Expansion Only
j
O
cn
x:
aver
0
1
1
1
•H
-C
medium
rH
rH
m
m
OO
•
CM
CM
rH
00
•
oo
rH
•
CO
Thermal Expansion Only
O
rH
aver
0
TD
1
1
rH
• H
CM
rH
-------
-62-
§
Cd
y;
i
Cu
tt
tt
2
W
ro
CM
m
m
J1
OO
>
Expansio
e
(0
00
CTl
y
l
o, a
x
e
e
30
y
a
e
in
00 O
00 00
m
m
in
"0
CTi
oo
in
oo
oo
oo
oo
>i
rH
>1
in
>i
a
a
a
a
x
e
e
e
ave
e
0)
.HI
CH
ro
r
^
1
oo
CN
t
CN
CN
m
CN
to
I
rH
£
ro
0)
ro
Oi
ro
C
C
H
x;
c
2
fa
"8
fa
CN
CN
in
m
O
m
&!
' X
c
-C
g
(0
med
W
y
W
fa
2
CN
CN
-------
- 63 -
APPENDIX B
MODELS, ANALYTICAL METHODS, AND ASSUMPTIONS
USED FOR ESTIMATING THE SCENARIOS OF SEA LEVEL RISE
Estimating the sea level rise produced by the temperature
increases that will be associated with rises in atmospheric
concentrations of C02 and other greenhouse gases required
integrating a variety of data, assumptions, and physical and
behavioral relationships (see Figure B-l). For this integra-
tion, three models were used in this study. The first is a world
energy and C02 emission model obtained from the Institute for
Energy Analysis (IEA) of the Oak Ridge Associated Universities.
The second is a regression equation that simplifies a one-dimen-
sional radiative/convective atmospheric temperature model obtained
from Dr. James Hansen of the Goddard Institute for Space Studies
(GISS). This model also represents heat flux into the ocean and
computes sea level rise from thermal expansion of the ocean
layers. The third, used in only some scenarios, is a global
carbon cycle model obtained from Dr. William Emanuel of the
Environmental Sciences Division, of the Oak Ridge National
Laboratory (ORNL).
All three models were received from their developers in the
form of FORTRAN source programs and were installed on the EPA
computer located in North Carolina. Throughout this appendix,
the models will be referenced as the IEA, GISS, and ORNL models.
-------
FIGURE B-1
KEY COMPONENTS OF THE ANALYTICAL METHODOLOGY
POPULATION
GROWTH
ASSUMPTIONS
ENERGY COST
ASSUMPTIONS
IEA ENERGY/CO2 MODEL
• ENERGY - C02 EMISSION RELATIONSHIPS
• ENERGY SUPPLY FUNCTIONS
• ENERGY DEMAND FUNCTIONS
PRODUCTIVITY
CHANGES
ASSUMPTIONS
FOSSIL FUEL CARBON
EMISSIONS OVER TIME
NATURALC02SOURCES
C02 SINKS
FLOWS BETWEEN SOURCES
AND SINKS
ASSUME CONSTANT
ATMOSPHERIC
C02RETENTION
ATMOSPHERIC CO
/ATMOSPHERIC (SEA SURFACE)
\TEMPERATURE OVER TIME
HEAT
DIFFUSION
ASSUMPTIONS
SEA LEVEL THERMAL
EXPANSION
C ICE/SNOW
ONTRIBUTION
TOTAL SEA LEVEL
RISE
TRACE GAS
GROWTH
ASSUMPTIONS
ASSUMPTIONS
ABOUT OTHER
FORCINGS
GISS MODEL
ATMOSPHERIC C02 - TEMPERATURE
RELATIONSHIPS
OCEAN RESPONSES TO TEMPERATURE
THERMAL
BOX DIFFUSION OCEAN MODEL - EXPANSION
EQUATION
-------
- 65 -
In addition to changing some of the parameters and operational
procedures of both the GISS and ORNL models, additional programs
were developed to provide simple interfaces between these two
models and to permit more rapid and extensive sensitivity analyses.
Separate sections of this appendix describe the individual models
and procedures for operating and combining them in the context
of this study, as well as the specific assumptions used to run
the scenarios.
THE ORGANIZING FRAMEWORK USED
The assumptions, relationships, and models used to estimate
sea level rise are shown in Figure B-l. The key variables are
described as primary inputs or outputs of the three models or
are subsumed in the internal structures of the models.
As shown in Figure B-l, the three models are operated in
sequence. The IEA model is first run to generate future energy
use and carbon emission scenarios. The ORNL model (or an assumed
constant fraction of CO2 airborne) is next used to translate
fossil-fuel carbon emissions into atmospheric CO2 levels. The
GISS model is finally employed to translate increases in C02
concentrations into atmospheric temperature increases and sea
level rise. (The temperature of the atmosphere at the earth's
surface is assumed equal to the temperature of the ocean's top
layer.) The GISS model uses a variety of key assumptions as
inputs: trace gas growth, thermal sensitivity of the atmosphere,
and changes in other forcings.
-------
- 66 -
The process of running the models is not perfectly sequen-
tial. The ORNL and GISS models are coupled in an iterative
fashion for those analyses where the ORNL model is run. (This
is discussed below under the heading "Coupling the ORNL and GISS
Models.") Endnote 1 presents the rationale for coupling ORNL
and GISS models and compares results using the two basic analytical
approaches.
MODEL DESCRIPTIONS
Three models will be described: the IEA world energy model,
the ORNL carbon cycle model, and the GISS climate equation ocean
heat diffusion model.
The IEA Long-Term Global Model
The IEA/ORAU model was developed by Jae Edmonds and John
Reilly at the Institute for Energy Analysis as an assessment
tool for policy analysis._2/ It provides a consistent represen-
tation of economic, demographic, technical, and policy factors
as they affect energy use and production and CO2 emissions.
The global model is dissaggregated into nine regions (see
Figure B-2). Its time horizon is long-term — from 1975 to
2100 — with benchmark years 2000, 2025, 2050, and 2075.
Projections beyond 2050 have not been examined in detail for
reasonableness of parameters and model structure. Eleven
types of primary energy in six major categories are currently
considered (see Table B-l). In addition to the supply and demand
for energy by region and forecast period, the model also estimates
-------
-67-
FIGUKE B-2
GEOPOLITICAL DIVISIONS IN THE IEA MODEL
K.y:
1. USA
2. OECO WEST (Carudl Hid W«Um Europ*)
3 JANZ (Japan. Auilr«ln >nd Ntw Zwlandl
4 EUSSR (EnternEurop* wid USSR)
6. ACENP (AiiHt Camrilly PUnned)
6. MlDEST (M«ldli E«l)
7 AFR (Alricj)
6 LA Ilitin Anwricl)
9. SEASIA (South md Eul Ami
world and regional energy prices consistent with overall global
energy balance. The structure, data base, output, and usage of
the model are extensively documented.^/
Energy Demand
Energy demand for each of the six major fuel categories is
developed for each of the nine regions separately. Five major
exogenous inputs determine energy demand: population, economic
activity, technological change, energy prices, and energy taxes
and tariffs.
-------
-68-
TABLE B-1
PRIMARY FUEL TYPES IN THE IEA MODEL
OIL
• CONVENTIONAL
• UNCONVENTIONAL (ENHANCED RECOVERY, SHALE OIL,
AND TAR SANDS)
• SYNTHETIC (FROM COAL AND BIOMASS)
GAS
• CONVENTIONAL
• UNCONVENTIONAL (DEEP WELLS, TIGHT FORMATIONS)
• SYNTHETIC (FROM COAL AND BIOMASS)
SOLIDS
• COAL
• BIOMASS
RESOURCE-CONSTRAINED RENEWABLES
• HYDROELECTRIC
NUCLEAR
SOLAR
SOLAR ELECTRIC (OTHER SOLAR IS ASSOCIATED WITH
CONSERVATION)
-------
- 69 -
An estimate of GNP for each region is used as a proxy for
the overall level of economic activity and as an index of income.
While the level of GNP is an input to the system, it is derived
from assumptions on demographic growth and levels of labor
productivity.
World population was assumed to stabilize by 2075, based on
work done by Keyfitz et al.^/ Labor productivity growth for
the high scenario was assumed to be 3.2 percent per year in 1980,
declining linearly to 2.2 percent per year by 2100.
In the low scenario, labor productivity growth was assumed to
start at 2.2 percent per year in 1980 and to decline linearly to
1.7 percent per year by 2100. Both the high and low estimates of
productivity growth utilize rates below the average rate of increase
around the world in the last thirty years. The mid-range estimate
was an interpolation of the two curves.
The energy technology parameter is a time-dependent index of
energy productivity, given constant energy prices and real incomes.
That is to say, it reflects improvements in energy efficiency
beyond those stimulated by increases in real prices or decreases
in real income. In the past, technological progress has had an
important influence on energy use in the manufacturing sector of
advanced economies. The inclusion of an energy technology para-
meter allows scenarios to be developed which incorporate either
continued improvements or technological stagnation as an integral
part of energy use scenarios. A constant improvement of 25
-------
- 70 -
percent for each 25-year period was assumed for the industrial
sector in all OECD countries (no change was assumed for the
other sectors for these countries), and a 10 percent improvement
was assumed for the single aggregate sector in other countries.
The final energy factor which influences demand is energy
price. Each region has a unique set of energy prices which are
derived from world prices (determined in the energy balance
component of the model), transportation costs, and region-specific
taxes and tariffs. The model can be modified to accommodate
non-trading regions for any fuel or set of fuels. It is assumed
that no trade is carried on between regions in solar, nuclear,
or hydroelectric power, but that all regions trade other types of
fuels.
The four secondary fuels (refined oil, refined gas, refined
coal, and electricity) are consumed to produce energy services.
In the three OECD regions (Regions 1, 2, and 3 in Figure B-2),
energy is consumed by three end-use sectors: residential/commer-
cial, industrial, and transportation. In the remaining regions,
final energy is consumed by a single aggregate sector.
The demand for energy services in each region's end-use
sector(s) is determined by the cost of providing these services,
and the levels of income and population. The mix of secondary
fuels used to provide these services is determined by the relative
costs of providing the services using each alternative fuel.
-------
- 71 -
The price of secondary fossil fuels is a function of the
regional price of primary fuels and the cost of refining:
Pjr = Pir(9ij) + hj
Pir = (Pi + TRir) TXir
Where: Pjr is the price of secondary fuel j in region r;
Pir is the price of primary fuel i in region r;
gij is the efficiency of refining i into j ; hj is
the nonenergy cost of refining; P^ is the world
market price of fuel i; TRir is the cost of
transporting fuel i to region r, and TXir are
taxes on fuel i in region r.
The demand for fuels to provide electric power is then
determined by the relative prices of the alternative electricity-
generating fuels. Likewise, the demand for synthetic oil and
gas from coal and biomass are functions of oil and gas prices.
Finally, production levels of primary energy sources are derived
from the demand for secondary fuels.
Energy Supply
Three generic energy supply categories are distinguished:
resource-constrained conventional energy, resource-constrained
renewable energy, and unconstrained energy resources. There are
eight different supply modes across these categories, as shown
in Table B-2. Production of conventional gas and oil are repre-
sented by a logistics curve which reflects historical supply
levels and estimates of remaining deposits:
F(t) = exp(a + bt)
l-F(t)
Where: F(t) is the cumulative fraction of the total resource
exploited by time t, and a and b are empirical parameters.
-------
- 72 -
Production rates of these fuels are thus insensitive to price
levels.
TABLE B-2
DISTRIBUTION OF IEA FUEL TYPES
ACROSS SUPPLY CATEGORIES
Supply Categories
Resource-
Constrained
Conventional
Energy
Conventional Oil
Conventional Gas
Resource-
Constrained
Renewable
Energy
Hydro, Bioniass
Unconstrained
Energy
Resources
Unconventional Oil
(primarily oil shale)
Unconventional Gas
(primarily gas)
Synthetic Gas
(from solids)
Synthetic Oil
(from solids)
Coal
Solar
Nuclear
Production levels of unconventional oil and gas, nuclear,
solar, and solids (coal and biomass) are modeled as "backstop
technologies." That is, a base level of production is assumed to
be generated over time if real prices remain constant. Shorter-
term supply schedules are then superimposed on these long-term
trends to reflect the increase or decrease in production due to
-------
- 73 -
price rises or declines. If the price falls below a breakthrough
price level, then production ceases. These relationships are
encompassed in the following general expression:
P = a [exp(g/b)C]
Where: P is the price of the backstop fuel; a is the break-
through price; b is a parameter which determines the
"normal" backstop price; c is a price-elasticity con-
trol parameter; and g is the ratio of output in year t
to the base level associated with the backstop price.
Production levels of synthetic oil and gas are driven by the
cost of solids, the cost of producing the synthetic fuel, and
associated nonenergy costs. The share of coal or biomass
allocated to the production of synfuels is specified by a logit
share equation, with the relative cost of synfuels versus other
sources of gas and oil and the price elasticity of production as
key terms.
Resource-constrained renewable fuels are considered constant-
flow sources. That is, the rates of energy production are limited
by the availability of the resource.
For the high scenario the price of supplying each energy
source was the "best guess" of the experts familiar with the
respective sources. For the low scenario the cost of nuclear
energy was halved starting in 1980.
Energy Balance
The supply and demand modules each generate energy supply
and demand estimates based on exogenous input assumptions and
-------
- 74 -
energy prices. If energy supply and demand match when summed
across all trading regions in each group for each fuel, then the
global energy system balances. Such a result is unlikely at any
arbitrary set of initial energy prices.
The energy balance component of the model is a set of
rules for choosing energy prices which, on successive attempts,
brings supply and demand nearer a systemwide balance. Successive
energy price vectors are chosen until energy markets balance
within a prespecified bound. Figure B-3 displays the interactions
necessary to achieve a global energy balance.
CO2 Release
Given the solution from the energy balance component of the
model, the calculation of CC>2 emissions rates is conceptually
straightforward. The problem merely requires the application of
appropriate carbon coefficients (carbon release per unit of energy)
at the points in the energy flow where carbon is released. Carbon
release is associated with the consumption of oil, gas, and coal.
Large amounts of CO2 are released from the production of shale
oil from carbonate rock and from the production of synthetic
fuels from coal. A zero-carbon-release coefficient is assigned
to biomass, nuclear, hydro, solar, and conservation. Table B-3
shows the CC>2 emission coefficients used in the IEA model.
-------
- 75 -
FIGURE B-3
SUMMARY OF INFORMATION FLOWS AND PROCEDURES IN THE IEA
MODEL TO BALANCE ENERGY SUPPLIES AND DEMANDS
r
REGIONAL
POPULATIONS
REGIONAL
LABOR
FORCES
REGIONAL
LABOR
REGIONAL
GNP
TECHNO-
PRODUC-
TIVITIES
LOGICAL
CHANGE
REGIONAL
ENERGY
DEMANDS
GLOBAL
SUPPLIES
AND
DEMANDS
REGIONAL
ENERGY
SUPPLIES
Model Results
Three separate scenarios were examined in conjunction with
this study:
1. High economic growth with the more expensive nuclear
power production costs produced a 2.345% annual growth
in carbon emissions.
2. Slow economic growth with the more expensive nuclear
power production costs produced 2.074% annual growth
in carbon emissions.
-------
- 76 -
TABLE B-3
CO 2 COEFFICIENTS FOR CARBON-PRODUCING
FUELS IN THE CO? MODEL
FUEL CARBON RELEASED
(TERAGRAMS OF CARBON PER EXAJOULE)
LIQUIDS 19.7
GASES 13.8
COAL 23.9
CARBONATE ROCK (SHALE) MINING AND PROCESSING 27.9
COAL LIQUIFACTION 18.9
COAL GASIFICATION 26.9
3. Slow economic growth with halved nuclear power production
costs produced a 1.674% annual growth in carbon emissions.
The CC>2 emissions used for the three scenarios are depicted
in Figure B-4.
Reasonableness of the Assumptions
The forces that determine emission trends -- long-term
population, labor force, and productivity growth, changes in energy
supply and demand, additions to the resource base, and the inter-
action of these factors in a supply-demand framework -- are
difficult to project. For example, energy technologies could
change much more radically than currently foreseen. Large-scale
biomass production might become much more feasible in the next
century as an energy source if recombinant DNA and other biotech-
nologies produce supercrops; large fossil-fuel reserves might be
discovered, which would radically decrease fossil-fuels costs;
-------
• 17-
FIGURE B-4
GLOBAL CO2 EMISSIONS, 1975-2100
(Billions of Tons of Carbon)
m
cc
3
CO
O
LL
O
CO
O
en
90
80
70
60
50
40
30
20
10
1975
2000
2025 2050
YEAR
2075
2100
-------
- 78 -
new technologies might be developed that vastly reduce the costs
of extracting hydrocarbons from shale or tar sand; or nuclear
fusion might become a commercial reality sooner than expected.
Any of these "surprises" would alter fuel use and emissions in
ways unanticipated by our analysis. Although counting on such
surprises would be foolish for a study like this, our conclusions
must be tempered by the realization that the future may be very
different from the extensive range of possibilities currently
envisioned as possible.
With respect to the scenarios generated, the assumption that
nuclear energy costs could be instantaneously halved in 1980 is
overly optimistic. It was used simply to test the implications
of a radical shift in the cost of a major energy source that does
not emit CO2- Other experiments with the model, such as doubling
or tripling the cost of fossil fuels, had negligible effects on
atmospheric temperatures.^/
Although the IEA energy model was designed for long-range policy
analysis and satisfies this overriding objective, it does not con-
sider the capital constraints that may limit the rates of substi-
tution among competing fuels. By ignoring potential bottlenecks
that could result, the model tends to overestimate the degree
to which rapid change in energy supplies is possible. However,
since the results of this study (projections of sea level rise)
were relatively insensitive to energy projections, this limitation
of the IEA model is not critical.
-------
- 79 -
THE ORNL CARBON CYCLE MODEL
The Oak Ridge National Laboratory carbon cycle model repre-
sents flows and stocks of carbon on a global scale._5/ Terrestrial
carbon is modeled in considerable detail, while ocean carbon is
represented by a simple box-diffusion model. A general overview
of the model is presented here.
Figure B-5 depicts the overall structure of the model.
Carbon in living material is divided between ground vegetation,
the wood parts of trees, and the non-wood parts of trees. Dead
organic matter is divided between detritus/decomposer and active
soil carbon. Finally, carbon in the oceans is subdivided into
surface (less than 260 meters below sea level) and deep layers.
For each of the carbon reservoirs, equations specify maximum
stocks of carbon and rates of flow into and out of the reservoirs.
In general, terrestrial flows are modeled as linearly dependent
on carbon content of the i donor reservoir (F^- = a— C^) or as
a more complicated logistics function. (Details of the structural
equations can be found in Emanuel, et al., 1981.)
The logistic functions for trees include terms for carbon
release to the atmosphere from forest clearing. Permanent forest
clearing is reflected by reductions in the parameters governing
storage capacity. Where clearing is temporary, reestablishment
of forests occurs in a delayed exponential fashion. (No future
clearing option was assumed in this study.)
-------
- 80 -
FIGURE B-5
STRUCTURE OF THE ORNL CARBON CYCLE MODEL
ORNL - DWG B1 6057 ESD
ATMOSPHERE
GROUND
VEGETATION
NONWOODY PARTS
TREES
WOODY PARTS
TREES
DETRITUS/
DECOMPOSERS
ACTIVE SOIL
CARBON
SURFACE OCEAN
DEEP OCEAN
MNM,cM FROM AN UpDATED MANUSCRIPT BY W.R.
EMANUEL AND OTHER AT OAK RIDGE NATIONAL LABORATORY.
Net uptake of carbon by the oceans is limited by the avail-
ability of carbonate ions. The rate at which gaseous CC>2 dissolves
in seawater depends on the rate of reaction with CO3, which is
the first step of a process by which CC>2 is incorporated into
various seawater carbonate compounds.
The rate at which carbon flows among the various reservoirs
is generally dependent on atmospheric temperatures. Sensitivity
to temperature is significant, because a rise in temperature is
likely to accompany an increase in atmospheric CO2. The original
-------
- 81 -
version of the ORNL represented this temperature-CC>2 relationship
by a simple exponential equation, rather than through a series of
heat-flux equations. This portion of the model was replaced by
the relatively sophisticated heat-flux relationships in the
GISS model, by coupling the two models (see the discussion under
"Coupling the GISS and ORNL Models").
The model inputs are estimates of fossil-fuel carbon emis-
sions on a yearly or longer time period basis from 1980 until
some future year. (The IEA model estimates are on a 25-year
basis. Estimates were provided on a 5-year basis by interpola-
tion of the 25-year estimate.) Outputs are atmospheric CO2 for
5-year intervals.
The model simulates historical patterns of carbon cycling
from 1740 to 1980 on an annual basis. The resulting time trend
in CO2 was modified slightly so that the concentration in 1980
was 339 ppm. This corresponds to the level estimated by the GISS
model in 1980, given estimates of fossil fuel carbon emissions in
1975 (the starting point of the IEA model) used in this study.
The model was then run to produce estimates of atmospheric C02
from 1980 to 2100.
Reasonableness of the Assumptions
The fraction of carbon emissions retained in the atmosphere
is the product of a complex biogeochemical cycle that evolves
over time. Evidence indicates that as the atmospheric concentration
-------
of CC>2 increases, the fraction retained in the atmosphere
will also increase. Consequently, using a constant fraction
airborne is a very conservative assumption.
The number chosen for the low estimate was 53%, the best
estimate of the historical fraction airborne.^/ A lower number
could have been chosen, based on the work of Woodwell et al.,
which states that past estimates of the biosphere's contribution
were too low arid thus caused fraction airborne estimates to be
too high.V
Lugo and other biologists, along with Broecker and other
oceanographers, however,- have challenged the Woodwell estimates.£/
Broecker has shown that the oceans do not have the physical
capacity to absorb CC>2 contributions from the biosphere as quickly
as Woodwell had originally estimated, and Lugo has shown that
Woodwell underestimated tropical forest regrowth, thereby over-
estimating net forest contributions.
Although it is a strong challenge to the Woodwell estimate,
the rebuttal is not absolutely conclusive. Nevertheless, because
all carbon cycle models show fraction airborne increasing over
time, use of 53 percent probably constitutes a very conservative
estimate of retained CC>2 emissions. The ORNL model modifies
the fraction airborne over time by taking into account water
heating and changes in ocean chemistry as well as temperature-
sensitive changes in carbon flows among all compartments.
-------
- 83 -
Even the URNL carbon cycle model, however, is oversimpli-
fied, representing the oceans as a layered vertical column or box,
and the biosphere by only a few types of vegetation. Ocean dynamics
are far more complex, containing complicated and interdependent
vertical and horizontal transports. Ocean chemistry is not well
represented in models of the carbon cycle, and sources and sinks
may be over- or underestimated.
The impacts of CO2 and temperature change on photosynthesis
and changes in ecosystems are not considered by these models.
The inaccuracies produced by such emissions have not been estimated.
INCREASES IN CONCENTRATIONS OF TRACE GASES
Increases in the concentrations of trace gases will add
significantly to the warming from CO2- This report considers
four of the most important gases: methane, nitrous oxide, CF2C12
(CFC-12), and CFC13 (CFC-11).
Scientific knowledge of these gases is sufficient to separately
project emissions and atmospheric residence time only for chloro-
fluorocarbons. We assume that the half-lives of CFC13 and CF2Cl2
will be 60 and 120 years, respectively. The low, mid-range, and
high scenarios assume that emissions will increase annually by
0.7, 2.5, and 3.8 percent of the 1980 level until 2020. Because
these emissions will also cause depletion of stratospheric ozone,
we assume that additional restrictions will be placed on their use
and that emissions will remain constant after 2020. Because of
the long residence times for these gases, their concentrations
-------
- 84 -
will continue to increase throughout the twentyfirst century in
all scenarios. Table B-5 shows the resulting concentrations
of these gases in 25-year intervals.
TABLE B-5
CONCENTRATIONS OF CHLOROFLUOROCARBONS
(parts per billion)
Year
1980
2000
2020
2040
2060
2080
2100
Low Scenario
CFC13 (CFC-11)
CF2Cl2 (CFC-12)
Mid-Range Scenario
CFC13
CF2C12
High Scenario
CFC13
CF2C12
.18
.31
.18
.31
.18
.31
.37
.69
.41
.76
.44
.84
.54
1.06
.69
1.34
.79
1.64
.68
1.44
.95
1.94
1.14
2.51
.79
1.73
1.15
2.46
1.40
3.27
.88
2.00
1.30
2.91
1.60
3.93
.94
2.25
1.42
3.32
1.76
4.51
Existing knowledge of the environmental fates of methane
and nitrous oxide did not permit us to model the biogeochemical
cycles for these gases. Instead, we projected their concentrations
directly. For the low scenario, we assumed that the historical
trends in the concentrations of methane and nitrous oxide would
continue; for the mid-range and high scenarios, we assumed that
they would accelerate. The concentration of nitrous oxide was
assumed to increase from the 1980 level (0.3 ppm) geometrically
by 0.2, 0.45, and 0.7 percent per year for the three scenarios
-------
-85-
(Weiss, 1981). The concentration of methane was assumed to
increase annually by 1.0, 1.5, and 2.0 percent for the three
scenarios from the 1980 level of 1.6 ppm (Rasmussen, 1981).
SEA LEVEL RISE MODEL AND THE GISS ATMOSPHERIC TEMPERATURE
The GISS model used in this study was based on a one-dimen-
sional radiative-convective (RC) model for estimating temperature
increases associated with atmospheric CC^ rises.^/ A routine for
estimating sea level rises in response to temperature fluxes
generated by atmospheric forcings was added for this study.
The RC model computes vertical temperature profiles over
time from net radiative and convective energy fluxes. Radiative
fluxes, in turn, depend on changes in atmospheric gases, especially
CC>2, and on the associated feedback effects. The GISS equation
used here is based on an empirical fit to the RC model.
The fitted equation, described in Hanser. et al., was modified
by Lacis for this study to incorporate trace greenhouse gases.
Heat Flux Computations
The heat flux into the earth's surface is estimated by an
equation that contains all key temperature-related terms:
F(t) = 2.6xlO~5( A C02) - 5.88xlQ-3( A T) + 3.685xlO"4 ( A T)2
[1 +.0022( n C02)]°'6 T0 "TT2
t^ ^
- 4.172xlO~7 ( A C02) ( A T) + 1.197x10-3 ( A CH4)°-5
Te
+ 5.88x10-3 ( A N20)°-6 + S.lSxlO"4 ( A CCS) + 3.78xlO~4 ( A CC2)
- 1.197xlO-4 ( A CH4)( N20) - 2.40x10-2 ( A V) - 2.10xlO~3 ( A V)2
- 1.17x10-3 ( A T)( A V) + 3.184X10-1 ( A S)
-------
- 86 -
Where: F(t) is the heat flux as a function of time in cal min 1 cm 2
A C02 is the change in atmospheric CC>2 from the 1880
value (293 ppm) in ppm.
A T is the change in atmospheric temperature (surface
level) from the 1880 value in °C
A CH4 is the change in atmospheric CH4 from the 1880
value (1.6 ppm) in ppm.
A N20 is the change in atmospheric N20 from the 1880
value ( .300 ppm) in ppm.
A CC3 is the change in CC^F from the 1880 value
(0 ppb) in ppb.
A CC2 is the change in CC12F2 from the 1880 value
(0 ppb) in ppb.
^ V is the change in atmospheric optical depth from
a baseline level due to volcanic activity in dimen-
sionless units.
A S is the change in solar luminosity from a baseline
level in fractional units.
A Te is the temperature equilibrium sensitivity -- the
assumed temperature rise when CC>2 doubles from the
1880 level (from 293 to 586 ppm).
The heat flux is estimated for time periods ranging from
each month to each year (a semimonthly time step was used in this
study). The appropriate A T value for calculating F(t) in each
time period (t=n) is the value estimated for the previous period
(t=n-l). For a simple one-layer ocean model, A T is obtained by
solving the following differential equation:
d A T = F(t)
dt C0
Where: Co is the heat capacity of the mixed layer of the
ocean per unit area (cal cm~2).
Estimates of heat flux from the empirical equation described
above were compared by Lacis with the RC model calculations. The
two estimates agreed to within one percent for CC>2 values of 0 to
1220 ppm, and to within 5 percent for CC>2 values of 1220 to
-------
- 87 -
1700 ppm.^£y Values for the parameters in the empirical
heat-flux equation were obtained as follows:
A CC>2: value obtained from the IEA and ORNL models;
A CH4-. 1.6 ppm from 1880 to 1980; concentrations after 1980
increase as described in section on trace gases.
A N20: 0.300 ppm in from 1880 to 1980; concentrations after
1980 increase as described in section on trace gases,
A CC2 and A CCS: See Table B-5.
A V: constant 0.007 or 0.037 each year for 1980 to 2100
A S: 0 for 1880 to 2100; or 0 for 1880 to 1980 and increased
or decreased by 0.005% per year for 1980 to 2100.
Te: 1.5°C, 3.0*C, and 4.5*C.
Changes in Volcano Aerosols and Solar Irradiation
For special cases changes in aerosols and solar luminosity
were considered. Aerosols and dust that are produced by volcanoes
can lower global temperatures by increasing reflection of incoming
energy. A higher estimate of optical depth than the baseline was
used as a special case.
Baseline optical depth = .007 each year (dimensionless units)
High optical depth = .037 each year (dimensionless units)
An optical depth of 0.037 is the highest twenty-year
average in the last century, while 0.007 is the average over
the period 1900-1980. No basis exists for assigning a different
value across a whole century. To sustain the high value over
the next 120 years would require a continuation of levels of
volcanic activity that have been sustained for decades, at most,
in recorded history.
Recent evidence indicates some possible shifts in solar
luminosity making the term "solar constant" a bit of a misnomer.
-------
-88-
Two special case scenarios were considered:
Special Cases: 0.5% increase over 100 years and
0.5% decrease over 100 years
A one-half percent decrease in solar luminosity is much
greater than would be likely in the time period under considera-
tion. However, even such an improbably large variation would not
overwhelm the greenhouse effect.
Diffusion of Heat in the Ocean
The ocean model consists of a mixed layer of depth Hm = 100m
and a thermocline with 62 layers and depth H = 900m. The mixed
layer temperature is assumed to be independent of depth, while
the thermocline temperatures are defined by a diffusion equa-
tion with constant thermal diffusivity. The layering is different
from that used in the ORNL model, but the difference is not of
importance in generating the expansion.
The temperature change in the mixed layer •( A Tm) is a solu-
tion of the equation:
cHm d A Tm = F(t) - FD(t)
dt
where c is the heat capacity of water, Hm is the depth of the
mixed layer, F(t) is the heat flux from the atmosphere into the
ocean and
FD(t) = - x aA T
3 z z=Hm
is the heat flux from the thermocline into the mixed layer.
-------
- 89 -
Note that our z-axis is directed toward the bottom of the ocean.
Also, we use g, cm, sec and cal, so the heat conductivity lamda
is n.umercially equal to the heat diffusivity K. Values for
diffusivity were set at either 1.18 or 1.9 cm2/sec. (Special
scenarios investigated values of 0.2 and 4.0 cm^/sec.)
The temperature change in the thermoclir.e ( ^ T) is determined
by the diffusion equation:
c SA T(z,t) = k a2A T(z,t)
3 t 6 z2
The boundary conditions for A T are:
A T = A Tm at z = Hm
and zero heat flux at the bottom of the thermocline:
ISA T = 0 at z = H + Hm.
3 z
Thus it is assumed that no energy escapes through the lower
boundary of the thermocline. Note that A Tm and A T are
temperature changes of the mixed layer and the thermocline
between the initial time (1880) and time t. It is assumed that
in the year 1880 A Tm = A T = ° and thus that the ocean
temperature was in a state of equilibrium with the atmosphere at
that time.
Required input data are atmospheric CC>2 levels on a yearly
or longer time period. Five-year estimates were obtained from
the ORNL, or alternatively, by multiplying a fixed-retention
factor times the interpolated five-year estimates of carbon
-------
- 90 -
emissions from the IEA model. GISS outputs are estimates of
ocean temperature increases on a five year basis.
The use of diffusivity coefficients as a surrogate for all
circulation processes that transport heat may fail to describe
the time paths well, with the downward heat transport in the
latter period probably being overestimated because of increasing
oceanic stability. Nevertheless this method simplified the
problem substantially and a range of different possible coeffi-
cients was used to investigate the sensitivity of the overall
estimate to the rate of downward heat transport.
These coefficients are:
High: = 1.9 cm2/sec
Low: = 1.18 cm2/sec
Special Scenarios: = 0.2 cm2/sec (minimum)
= 4.0 cm^/sec (maximum)
The range of coefficients used, 1.18 to 1.9 cm^/sec
is representative of mean oceanwide mixing rates as determined
by the NSF-sponsored transient tracer experiment and others.
Variation in estimates of the exact value of the data depend
on the tracer used and the statistical method of computing
the global mean. We tested 0.2 and 4.0 cm2/sec to account
for the possibility of dramatic changes in ocean transports due
to deglaciation and climate change.
-------
- 91 -
Sea Level Rise
Given a time trend in ocean temperature rise, the sea
level change is estimated as:
Hm+H
A H = r dz A a
o a
where a is the specific volume and ^ a is its change due to
thermal expansion. For this computation an ocean temperature of
10°C and salinity of 35 per mille are used. In other words, the
mixed layer and each of the 62 layers in the thermocline expand to
a degree that reflects the average pressure, and the rise in
temperature of the layer (as reflected by the change in density).
The average global rise in sea level is the output from this
component of the model.
Crustal Movements
Systematic changes in the topography of earth's land or
oceans can alter apparent sea level. No changes were assumed
however, because global variations occur over a vastly longer
time scale than the 120 year period considered here. Clark et al.
estimated that long-term isostatic adjustment of the earth's
surface has caused a worldwide sea level change of 1 meter per
5000 years (2 centimeters per century).^/ In certain regions,
coastal movements can substantially increase or decrease local
sea level rise. However, such trends are likely to balance out
worldwide.
-------
- 92 -
COUPLING THE QRNL AND GISS MODELS
The general .methodology for estimating thermal expansion of
the oceans was outlined in Figure B-l. Of the two pathways for
translating carbon emissions from fossil fuel combustion into
atmospheric CO2 concentrations (assuming a retention factor and
running the ORNL model), this section focuses on the use of the
ORNL model. As discussed previously, using the ORNL model to
estimate atmospheric concentrations of C02 requires the coupling
of the ORNL and GISS models and an iterative procedure using the
GISS model alone. The discussion in this section emphasizes the
sequence of actions employed to couple the two models, and the
procedures for iterating within the GISS model.
Calculating Initial Temperature vs. Time Curves
The starting point for the analyses that use the ORNL
model is the estimation of a family of temperature-time curves
using the GISS model. These curves are employed in the ORNL
model to describe a range of future increases in sea surface
temperature (in response to rising atmospheric CO2) which, in
turn, will affect the rate of CO2 exchange among sources and
sinks. Essentially, the time trends of temperature, generated
by a fairly sophisticated treatment of heat flux in the GISS
model, replace the very simple CO2~temperature relationships in
the ORNL model. The treatment of heat flux in the GISS model is
still a highly simplified representation of real ocean transports,
-------
- 93 -
as discussed above.
To generate the GISS temperature-time curves, it is first
necessary to estimate an atmospheric CC^-time curve, since this is
the main driving function for the GISS model. The most straight-
forward approach is to apply a constant retention factor to the
time trend of CC>2 emissions generated by the IEA model. In other
words, the first step in this approach appears identical to the
other methodological approach. in fact, a modification was
employed for these analyses: changing retention factors (from
about 0.6 in 1980 to about 0.8 in 2100) were used to generate
the preliminary atmospheric C02~time curves.
The GISS model was then run with high and low temperature/
C02 sensitivity (Te) values, with and without trace gases, to
produce a family of curves of sea-surface temperature versus time.
(Values for solar luminosity, volcanic activity, and heat diffu-
sivity were specified in each scenario.) These curves are repre-
sented by a quadratic equation of temperature versus time with
parameters a and b. The parameter a is the temperature difference
between 2100 and 1980; b is the temperature difference between
2040 and 1980 divided by a. These parameters are then used to
calculate coefficients for the quadratic equation T = ct + dt2 + 293,
where T is the temperature rise after 1980 (°K), and t is the number
of years after 1980.
-------
- 94 -
Four new curves reflecting the extreme values observed for a
and b were then specified:
a
Highest
Lowest
Highest
Case 1
Case 3
Lowest
Case 2
Case 4
These curves bound all possible changes in temperature, given the
time trend in atmospheric CC>2 specified previously. The four
combinations of values for a and b are then transmitted to the
ORNL model.
Calculating CO2 v. Time Curves with the QRNL Model
The ORNL model is next employed to estimate future increases
in atmospheric CC-2, using the fossil-fuel CC>2 emission scenarios
from the IEA model and the four sea-surface-temperature time
trends from the GISS model. Since four separate temperature/
time curves are employed, four separate CC^-time curves are
generated as output. Each CC>2 curve is represented by estimates
of atmospheric CC>2 concentrations in 10-year intervals from 1980
to 2100. Thus, a 4 by 13 matrix of values (corresponding to the
four time v. temperature curves and the thirteen inclusive
decades between 1980 and 2100) is generated as output.
Estimating Final Temperature v. Time Curves
The four "refined" time projections of CC>2 are returned to
the GISS model to obtain a consensus temperature versus time
-------
- 95 -
curve. This is accomplished by selecting one of the CC>2 curves
as a starting point. (Tests have demonstrated that starting from
any one of the four curves will yield tne same result.) The GISS
model is then run to obtain a corresponding temperature curve
from 1980 to 2100. This temperature curve is compared with the
four temperature curves originally generated by the GISS model
and transmsitted to the ORNL model. The new temperature curve is
composed by interpolating among the four previous curves, and
a new C02 curve is estimated corresponding to the interpolated
temperature curve. (In essence, this is a two-way interpolation
for each 10-year interval.) The whole process is repeated until
two successive temperature curves closely approximate each other.
Usually, this takes no more than two iterations.
To check the "accuracy" of such an iterative approach, the
final temperature-time curve from GISS was used as input to the
ORNL model for a small sample of runs. The resulting CO2~time
curve from ORNL was compared with the final C02~time curve in GISS.
In all cases, the two agreed within 2 ppm for each 10-year interval
Estimating the Contribution of Thermal Expansion to Sea Level Rise
Once the final projections of expansion to sea level rise are
obtained, the equations in GISS relating ocean volume, temperature,
and pressure described previously are employed to estimate thermal
expansion. The increase in ocean volume is then translated into
global mean sea level rise.
-------
- 96 -
END NOTES TO APPENDIX B
The rationale for coupling the GISS and ORNL models in this
fashion is based on (a) the sensitivity of many C02 uptake
and release equations to temperature in the ORNL model, and
(b) the more sophisticated treatment of heat flows and
temperature changes in response to C02 increases in the GISS
model. An alternative approach would be to replace the simple
temperature equations in the ORNL model with the GISS equa-
tions, in essence combining the two models. However, the
ORNL model is much more expensive to run than the GISS model.
Thus, the coupled approach employed here is a cost-effective
way to upgrade the treatment of temperature in the carbon
cycle model.
To illustrate the difference in results between the two basic
analytical approaches (fixed-retention factor with the GISS
model v. coupled ORNL-GISS modeling), we have listed
estimated atmospheric CO2 levels and sea level rise from
1980 to 2100 in the following table for each approach.
Two sets of results are listed for the coupled ORNL-GISS
modeling results — the initial outputs from the ORNL and
the final outputs from the GISS model. As shown, the
coupled ORNL-GISS approach produces higher estimates than
the fixed-retention-factor approach. This is reflected
in the retention factor inferred from the ORNL-GISS modeling,
which varies from 0.53 in 1980 to 0.72 in 2100.
A second observation concerns differences between the two
sets of ORNL-GISS outputs. The initial ORNL CO2 (or sea
level rise) time curve is higher than the final iteration
from GISS, reflecting the fact that the ORNL results repre-
sent one of the four extreme temperature-versus-time curves.
The final iteration from GISS is a C02 time curve interpolated
among the extremes, as discussed in the section on coupling
the ORNL and GISS models.
-------
- 97 -
ESTIMATED ATMOSPHERIC C02 LEVELS AND SEA LEVEL RISE
USING THE TWO BASIC ANALYTICAL APPROACHES
(High-Growth Scenario, High-Temperature Sensitivity,
No Trace Gases)
Fixed-Retention Factor
Coupled ORNL-GISS Modeling
(0.
53A)*
Ini
tial
ORNL Results
1980
1990
2000
2010
2020
2030
2040
2050
2060
2070
2080
2090
2100
CO2 Level
(ppm)
339
354
370
392
420
457
511
584
676
787
919
1083
1281
Sea Level
Rise (cm)
0
0.99
2.22
3.72
5.58
7.88
10.80
14.52
19.13
24.68
31.19
38.71
47.31
CO 2
Level
(ppm)
339
359
382
412
451
502
578
681
814
979
1181
1439
1756
Sea
Level
Rise
(cm)
0
1.06
2.47
4.27
6.57
9.45
13.15
17.86
23.71
30.77
39.04
48.63
59.59
Final
Iteration
CO2
Level
(ppm)
339
356
375
401
436
482
555
654
784
947
1147
1403
1720
GISS
Results
Sea
Level
Rise
(cm)
0
1.02
2.32
3.95
6.03
8.66
12.09
16.56
22.19
29.05
37.16
46.62
57.48
* A is a conversion factor to convert tons of carbon retained in
the atmosphere into ppm of CO2. A = 0.472 Gigatons of Carbon
per ppm of CO2.
-------
- 98 -
2/ Edmonds, Jae and John Reilly, 1982b. Global Energy and CQ2
to the Year 2Q50. Oak Ridye, TN: Institute for Energy
Analysis, Oak Ridge Associated Universities.
Edmonds, Jae, John Reilly, and Rayola Dougher, 1981.
Determinants of Global Energy Demand to the Year 2050.
Oak Ridge, TN: Institute for Energy Analysis, Oak Ridge
Associated Universities.
3/ Keyfitz, Nathan, et al. 1983. Global Population (1975-2075)
and Labor Force (1975-2050). Oak Ridge, TN : Oak Ridge
institute for Energy Analysis, Oak Ridge Associated
Universities.
4/ Seidel, Stephen and Dale Keyes, 1983. Can We Delay A
Greenhouse Warming?. Washington, DC: U.S. Environmental
Protection Agency.
jj/ Emanuel, William R. , G.G. Killough, and J.S. Olson, 1981.
"Modeling the circulation of carbon in the world's
terrestrial ecosystems," in Modeling the Global Carbon
Cycle—Scope 16, Bert Bolin, et al., eds., New York: John
Wiley.
6/ Rust, Bert W., Ralph Rotty, and Gregg Marland, 1979.
"Inferences drawn from atmospheric CO2 data." Journal
of Geophysical Research. 84:C6:3115-22.
!_/ Woe-dwell, G.M. 1978. "The carbon dioxide question."
Scientific American. 238:34-43.
_§_/ Lugo, Ariel, 1980. "Are tropical forest ecosystems a new
sink?" in The Role of Tropical Forests in the World
Carbon Cycle, Sandra Brown, Ariel Lugo, and Beth Liegel,
editors. Washington, D.C.: U.S. Department of Energy
#CONF 800350. 1-18.
Broecker, W.S., Tsung-Hung Peng, and R. Engh, 1980. "Modeling
the carbon system." Radiocarbon. 22:3:565-98.
9_/ Hansen, James E., et al. 1981. "Climate impact of increasing
atmospheric carbon dioxide." Science. 213:4511:957-66.
10/ Lacis, A., personal communication. Lacis, A., et al. 1981.
"Greenhouse effect of trace gases, 1970-1980." Geophyscial
Research Letters. 81:10:1035-8.
-------
ll/ Sofia, S., and A.S. Endal, 1982. "Solar variability."
Reviews of Space Physics.
12/ Clark, J.A., W.E. Farrell, and W.R. Peltier, 1978. "Global
changes in postglacial sea level: a numerical calculation."
Quarternary Research. 9:265-87.
-------
- 101 -
APPENDIX C
METHODS OF ESTIMATING SNOW AND ICE CONTRIBUTION
This appendix reviews the results and methods used to gene-
rate the estimates of ice melting in the doubled C02 experiments
with the GISS general circulation model (GCM) and presents the
methods used to estimate snow and ice contributions in the
scenarios.
GLACIAL MASS BALANCE IN GISS GCM
by Gary Russell
Goddard Institute for Space Studies
The GCM doubled C02 experiments imply a sea level rise of
13.5 mm/yr due to melting of land ice. The melting is caused
by a mean addition of heat of 3.27 Wm~2 to the surface of the
land ice. The increment of sea level caused by reduced snow on
the continents is insignificant compared to the glacial metling.
The following is a short description of algorithms applied
to land ice in the GCM, tables of results from the control run
and the doubled C02 experiment, and calculations based on those
tables.
Precipitation that falls on the ground is calculated from
the atmospheric conditions. Precipitation is snow if the first
atmospheric layer temperature is below 0°C; otherwise it is rain,
EPRCP (W/m2) = 0 if T1 > Q.
[LHM = 334000 (J/kg)]
-PRCP*LHM if T1 < 0
PRCP is the precipitation in kg/m^S
-------
- 102 -
The energy heat flux falling on the ground depends only on the
phase of the precipitation, not additionally on the temperature
of the rain or snow; this is an inaccuracy, but the error is a
relatively small portion of the total flux. Rain over ice
immediately runs into the ocean. SNOW (kg/m2) accumulates on the
surface. Over land ice, when the snow exceeds 100 (kg/m2), some
of it is compacted into ice so that the snow depth is reset to 90
(kg/m2). The compacted ice is accumulated (with a minus sign) in
the diagnostic MANDC.
Evaporation is calculated in the surface interaction routine
and depends on the surface drag, surface wind magnitude, surface
specific humidity, and the ground temperature.
EEVAP (W/m2) = EVAP * LH
LH = 2500000 (J/kg) over water
LH = 2834000 (J/kg) over ice
EVAP is the evaporation in kg/m2s
The temperature at which evaporation occurs is ignored.
Runoff over ice is the summation of rain plus melting of
snow or ice at the surface, represented by RUNOFF 9kg/m2s).
Runoff carries no energy with it. Over earth, runoff depends on
whether the ground is saturated and the energy of runoff is
ERUN (W/m2) = RUNOFF * T (°C) * SHW [SHW = 4185 (J/kg°c)]
-------
- 103 -
Other Energy Fluxes depend on the ground temperature, the
atmospheric situation, the solar zenith angle, and surface albedo.
ESENS (W/m2) sensible heating
ESOLR (W/m2) net solar heating into the ice
ETHRM (W/m2) net thermal heat out of the ice
Net Energy Fluxes into the ground (or ice) are calculated as
ENET (W/m2) = EPRCP + ESOLR - ETHRM - ESENS - EVAP - ERUN
If ENET is negative, the temperature of the ground snow or ice is
reduced. If ENET is positive, snow and ice are warmed until they
reach 0°C. At that point, snow is melted; after all the snow is
melted, the remaining energy melts ice. The melted ice is
accumulated in the diagnostic MANDC.
Melting and Compacting of glacial ice is a diagnostic of the
GCM called MANDC (kg/m2s), which is positive for melting. The
following equality holds:
d (SNOW + WATER + ICE) = PRCP - EVAP - RUNOFF + MANDC.
dt
SNOW, WATER, and ICE pertain to the first 4 meters in the ground.
MANDC comes from the ice below the first 4 meters. Averaged over
an annual cycle in a run which has achieved equilibrium, c3 (SNOW
dt
+ WATER + ICE) should be nearly zero. There is energy associated
with MANDC.
EMANDC (W/m2) = MANDC * [T(°C) * SHI - LHM], [SHI = 2060 (J/kg°C)]
and an equation (for ice).
-------
- 104 -
d
dt [(SNOW + ICE) * T * SHI] = ENET + EMANDC.
The time derivative was not calculated but is several orders
of magnitude smaller than ENET or EMANDC.
Sources of Error
The results of the model should be interpreted in light of
the following sources of error:
1. The major source of error is that the climate model uses
fractional grid, i.e., each grid box can can contain ocean,
sea ice, land ice, and earth. The composite surface
albedo is an area-weighted average of the albedos of the
separate surface types, but once the composite surface
solar absorption is calculated, it is distributed into
the surface types uniformly. When ocean ice is replaced
by open ocean, the composite albedo will decrease and the
increased solar absorption should primarily go into the ocean.
However, the model erroneously sends part of the absorbed solar
energy into the land ice and ocean ice. Correction of this
feature of the model could significantly decrease the estimated
melting rate of glaciers.
2. Another aspect of the fractional grid is that sensible heating
is proportional to the ground temperature minus the surface
air temperature, but the air temperature is a composite of the
entire grid box. Because land ice is at a higher altitude than
sea ice and the ocean, the sensible heating over land ice is
underestimated in both runs. Because the net melting is the
-------
-105-
difference between these two runs, the sign of the resulting
error is not known.
3. Areas of land ice in the model include Antarctica, Greenland,
islands along the Arctic Ocean, and glaciers in southern
Alaska. Had alpine glaciers been included as well, the
sea level rise from melting would have been somewhat greater.
General Comments on the Tables
1. The hemispheric and global numbers for water and ice were
produced with a resolution of hundredths of a mm/day or
3.65 mm/yr. The latitudinal numbers used a resolution
of tenths of mm/day or 36.5 mm/yr.
2. The hemispheric and global numbers for energy used a
resolution of tenths of W/.m2 except for EMANDC which is
know to a hundredth of W.m2. The latitudinal numbers
used W/m2.
3. ENET is calculated by subtracting large numbers and is
therefore less accurate. EMANDC is a diagnostic accumulated
directly in the GCM and should be much more accurate.
(The latter is also true for MANDC).
4. The numbers are based on a 10 year average (years 26-35)
of the control run 882 and doubled CC>2 experiment 886.
5. All numbers are per net area [1 (mm/yr) = 1 (kg/m2yr)].
The associated area is all land area.
6. .0106 (W) added to a square meter of ice will melt 1 (mm)
of ice during a year.
-------
- 106 -
7. The 33 (mm/yr) of melting in the difference 886-882
applies to all land. The average melting over land ice is
300 (mm/yr) = 36 .291 (land area)
.032 (land ice area).
Similarly, the energy of melting in the difference 886-
882 which applies to all land is .36 (W/m2). Applying
this to the land ice only it is 3.27 (W/m2) = .36 .291
.032.
8. The 3 (mm/yr) of melting of land ice in the control run
882 would cause a sea level rise of
1.23 (mm/yr) = 3 .291 (land area)
.709 (water area)
The 33 (mm/yr) of melting in the difference 886-882 would
cause a sea level rise of
13.54 (mm/yr) = 33 .291
.709
9. The sea level rise implied by the reduction of snow on
the continents in 886, compared to 882, is
1.59 (mm) = 3.75 .291
.709
This contribution is negligible, because it is a one-time
change rather than a rate as in the above cases.
-------
Glob
NH
- 107 -
TABLE C-l
Water and Ice Units: (mm/yr)
Control run: 882 Quantities refer to Land
SH 82°N 74°N 67°N 59°N -67°S -74°S -82°S -90°S
PRCP
EVAP
RUNOFF
d SNOW
dt
d( WATER + ICE)
dt
RUN+EV-PR
MANDC
Land
Percentage
Land Ice
Percentage
SNOW (mm)
960 931 1022
719 686 788
245 270 190
.05 .06 .02
-.4 -.6 0
3 25 -44
3 26 -44
29.1 39.4 18.8
3.2 4.3 2.1
17.6 15.7 21.6
329
73
292
1.1
0
36
37
15.0
13.2
95
438
110
402
.2
0
74
73
30.6
10.7
114
621
256
475
-.9
-.6
110
146
71.5
4.6
61
694
402
402
.2
-1.8
110
110
57.9
1.0
30
730
219
840
.2
0
329
329
15.4
15.4
55
475
73
73
-.2
0
-329
-329
66.0
66.0
82
183
37
0
.3
0
-146
-146
100.0
100.0
85
110
37
0
-.4
0
-73
-73
100.0
100.0
86
-------
- 108 -
TABLE C-2
Water and Ice
Difference: 886 - 882
Units: (imi/yr)
Land
Glob NH
SH
82°N 74°N 67°N 59°N -67°S -74°S -82°S -90°S
PRCP
EVAP
RUNOFF
d SNOW
dt
d( WATER + ICE)
dt
RUN+EV-PR
MANDC
SNOW (mm)
157 131
110 95
80 55
212 73 110
146 0 0
128 475 255
182
110
182
Numbers are
33 19
33 21
-3.75 -4.8
62 402 145
62 365 146
-1.6 -21 2
110
73
-15
146
110
110
nearly
74
36
-15
146
-37
1314
0.
1131
1095
-11
146 110 73
0 37 0
219 73 0
73 0 -73
73 0 -37
-5 -4 1
-------
- 109 -
TABLE C-3
Energy
Control Run: 882
Units: (W/m2)
Quanitities Refer to Land
Glob
NH
SH 82°N 74°N 67°N 59°N -67°S -74°S -82°S -90°S
ESOLR
ETHRM
ESENS
EEVAP
EPRCP
ERLJN
ENET
EMANDC
154.5
+48.0
47.5
57.3
-1.3
.4
-.07
-.01
150.2
+48.0
45.6
54.7
-1.4
.4
.17
-.29
163.3
+48.1
51.6
62.6
-1.2
.4
-.56
+.57
32
+23
1
5
-5
0
-2
0
42
23
6
8
-5
0
-1
-1
70
30
14
21
-3
0.1
1
— 1
96
37
23
32
-3
0.1
1
-1
57
27
3
16
-7
0
3
-3
38
30
2
5
-5
0
-4
+4
38
33
1
4
-2
0
-2
+2
42
39
-1
4
-1
0
-1
+1
-------
- 110 -
TABLE C-4
Energy
Difference: 886 - 882
Units: (W/m2;
Land
Glob NH
SH 82°N 74°N 67°N 59°N -67°S -74°S -82°S -90°S
ESOLR
ETHRM
ESENS
EEVAP
ERPCP
ERLJN
ENET
EMANDC
0.7
-4.0
-4.2
8.6
0.1
0.1
.37
-.36
.6
-3.6
-3.3
7.4
.3
.1
.24
-.22
1.2
-5.0
-6.3
11.4
-.2
.2
.63
-.64
3
-2
0
0
-1
0
4
-4
2
-2
0
2
-1
0
2
-1
_1
-4
-5
7
0
0.1
1
-1
-1
-4
-5
8
1
0.1
0
-1
3
-4
-1
-3
1
0
12
-12
-1
-3
0
0
-1
0
1
-1
-2
-2
0
0
-1
0
0
0
__"J
-3
0
0
-1
0
-1
+1
-------
- Ill -
INTERPRETATION OF GISS RESULTS
by John S. Hoffman
The problems with interpreting the GISS model results
are twofold:
1. The model looks at the effects of doubled CO2 / while
a projection would require looking at the system in
transition from current CO2 levels, to the doubled
level and beyond. For the mid-range scenario,
temperatures reach 4.1°C in the GISS doubled equilibrium
run by 2050. A method needs to be developed to
determine melting before this doubling and after it.
2. Melting in situ may not all run off. Some percentage
will refreeze, raising the temperature of the ice and
changing its physical characteristics. It is undetermined
at this time whether those changes cause ice discharges
greater than or less than if the melting had resulted
in runoff.
These problems are now being analyzed by a team from GISS,
NASA, and Lamont-Doherty Geological Observatory, along with
several other issues. In the preceeding analysis the results
were used only to check the general validity of the numbers
developed through our procedure of extrapolating historical
ratios, rather than to generate an ice and snow land-to-sea
transfer scenario.
-------
- 112 -
THE USE OF THE HISTORICAL METHOD FOR GENERATING SNOW-ICE
CONTRIBUTIONS
by John S. Hoffman
A variety of authors have estimated sea level rise in the
last century on the basis of tidal gauge stations. The esti-
mates vary between 10 cm and 15 cm. Using historical changes in
temperature, Hansen et al. and Gornitz et al. have varied CO2
increases and the heat absorption capacity of the ocean as
represented by the diffusitivity coefficent "K"._ For a Te
of 3°C, and a K of 1.2 cm2sec~l, almost half of the total sea
level rise could be accounted for. Sea level rose approximately
5 cm due to thermal expansion; thus, a 10 cm total rise implies
5 cm due to ice and snow transfer from land to sea, assuming
changes in other water reservoirs to be negligible. However,
if the historical sea level rise was actually 15 cm, then
10 cm was due to ice and snow contributions, giving the higher
two-to-one historical ratio. Barnett, the source of the 15 cm
estimate, attributes two-thirds of the rise to this source.^/
The scenarios of future sea level are simply extrapolated past
ratios, one to one as the low, two-to-one as the high. This
ratio generates estimates within the contribution capacity of
the ice sheets, alpine glaciers, and other water sources on
land based on a review of several strands of evidence, including
the GISS melting estimates.
-------
- 113 -
END NOTES APPENDIX C
_!/ Gornitz, V., S. Lebedeff, and James Hansen, 1982. "Global
sea level trend in the past century." Science. 215:26 March;
1611-4.
_2/ Barnett, Timothy P., 1983. "Global sea level; estimating
and explaining apparent changes," in Coastal Zone '83,
Orville T. Magoon, editor. New York: American Society of
Civil Enginners. 3:2777-84.
-------
- 115-
BIBLIOGRAPHY
Barnett, Timothy P. [1982]. "Recent changes in sea level and their
possible causes." Climate Research Group, Scripps Institution
of Oceanography, LaJolla, CA.
,1983. "Global sea level: estimating and explaining
apparent changes," in Coastal Zone '83, Orville T. Magoon, editor
New York: American Society of Civil Engineers. 3:2777-84.
Barnola, J.M., D. Raynaud, and H. Oeschger, 1983. "Comparison
of CC>2 measurements by two laboratories on air from bubbles
in polar ice." Nature 303:2 June:410-13.
Barth, Michael and James Titus, eds. Sea Level Rise to the Year 2100.
Stroudsburg, PA: Hutchinson Ross, (in press).
Bell, P.R. 1982. "Methane hydrate and the carbon dioxide question,"
in Carbon Dioxide Review 1982, William C. Clark, ed. New York:
Oxford University Press. 401-5.
Bently, Charles R. 1980. "Response of the west Antarctic ice sheet
to C02 induced climate warming." Alenarch Plan, Washington, DC:
American Association for the Advancement of Science.
1983. "West Antarctic ice sheet: diagnosis and prognosis,"
in Proceedings; Carbon Dioxide Research Conference; Carbon Dioxide,
Science, and Consensus 1983, Washington DC: Department of Energy.
U.S. CONF-820970.
Bloom, Arthur L. 1967. "Pleistocene shorelines: a new test of
isostasy." Geological Society of America Bulletin. 78:1477-1494.
Broecker, Wallace Smith, Tsung-Hung Peng, and R. Engh, 1980.
"Modeling the carbon system." Radiocarbon. 22:3:565-98.
Broecker, Wallace Smith, and Tsung-Hung Peng, 1982. Tracers in
the Sea. Palisades, N.Y.: Lamont-Doherty Geophysical Observatory,
Columbia University.
Brunn, Per. 1962. "Sea level rise as a cause of shore erosion."
Journal of Waterways and Harbors Division. New York: American
Society of Civil Engineers. 1:116-30.
Charney, Jule, et al., 1979. Carbon Dioxide and Climate;
A Scientific Assessment. Washington, DC: National Academy of
Sciences Press.
-------
- 116-
Chylek, Peter and Vvilliam Kellogg, 1982 "The sea level climate
connection." Unpublished paper. Boulder, CO: National Center
for Atmospheric Researcn.
Clark, J.A., W.E. F'arrell, and W.R. Peltier, 1978. "Global changes in
postglacial sea level: a numerical calculation." Quarternary
Research. 9:265-87.
Council on Environmental Quality, 1981. Global Energy Futures and
the Carbon Dioxide Problem. Washington, DC: Superintendent of
Documents.
Edmonds, Jae and John Reilly, 1982a. A Long-Term, Global, Energy,
Economic Model of Carbon Dioxide Release From Fossil Fuel Use.
Oak Ridge, TN: Institute for Energy Analysis, Oak Ridge
Universities.
, 1982b. Global Energy and CO? to the Year 2050.
Oak Ridge, TN: Institute for Energy Analysis, Oak Ridge
Universities.
Edmonds, Jae, John Reilly, and Rayola Dougher, 1981. Determinants
of Global Energy Demand to the Year 2050. Oak Ridge, TN:
Institute for Energy Analysis, Oak Ridge Associated Universities,
Emanuel, William R., G.G. Killough, and J.S. Olson, 1981. "Modeling
the circulation of carbon in the world's terrestrial ecosystems,'
in Modeling the Global Carbon Cycle — SCOPE 16, Bert Bolin,
et al. eds. New York: John Wiley.
Emery, K.O. 1980. "Relative sea levels from tide-guage records."
Proceedings of the National Academy of Sciences USA.
Washington, DC. 77:12:6968-72.
Emery K.O. and Robert H. Mead, 1971. "Sea level as affected by
river run-off, eastern United States." Science. 173:3995:425.
Etkins, 8., and E. Epstein, 1982. "The rise of global mean sea
level as an indication of climate changes." Science. 215:287-9.
Fairbanks, Richard G., and R.K. Matthews, 1978. "The Maine oxygen
isotope record in pleistocene coral, Barbados, West Indies."
Quaternary Research. 10:181-96.
Flohn, H., 1981. Life on a Warmer Earth. Laxenburg, Austria:
International Institute for Applied Systems Analysis
-------
- 117 -
Gallagher, Aurora, personal communication to James Titus.
Marine Board, National Academy of Sciences, 2101 Constitution
Avenue, NW, Washington, DC, 20418.
Gibbs, Michael, 1983. "Economic effects, and the value of informa-
tion on sea level rise," in Coastal Zone '83, Orville T. Magoon,
ed. New York: American Society of Civil Engineers. 3:2754-67.
Gornitz, V., S. Lebedeff, and J. Hansen, 1982. "Global sea
level trend in the past century." Science. 215:26 March:1611-4.
Gutenberg, Beno, 1941. "Changes in sea level, postglacial uplift,
and mobility of the earth's interior." Bulletin of the
Geological Society of America. 52:721-72.
Hands, Edward B., 1980. Prediction of Shore Retreat and Nearshore
Profile Adjustments to Rising Water Levels on the Great Lakes,
Technical Paper Number 80-7. Fort Belvoir, VA: US Army Corps
of Engineers, Coastal Engineering Research Center.
Hansen, James E. [1982]. "Global climate change due to increasing
C02 and trace gases." Statement presented at the joint hearing
on carbon dioxide research and the greenhouse effect to the
U.S. House of Representatives Subcommitte on Natural Resources,
Agriculture Research and Environment and the Subcommittee on
Investigations and Oversight of the Science and Technology
Committee.
Hansen, James E., A. Lacis, and D. Rind, 1983. "Climate trends due
to increasing greenhouse gases," in Coastal Zone '83, Orville
T. Magoon, ed. New York: American Society of Civil Engineers.
3:2796-810.
Hansen, James E., et al. 1981. "Climate impact of increasing atmos-
pheric carbon dioxide." Science. 213:4511:957-66.
, 1983. "Efficient three-dimensional global models for
climate studies: models I and II." Monthly Weather Review.
111:4:609-62.
Hey, R.W., 1978. "Horizontal quaternary shorelines of the
Mediterranean." Quaternary Research. 10:197-203.
Hicks, Steacy D., 1978. "An average geopotential sea level series
for the United States," Journal of Geophysical Research.
83:C3:1377-9
Hicks, Steacy D. 1981. "Long-period sea level variations for
the United States Through 1978." Shore and Beach. 26:9.
Hollin, John T. 1983. "Climate and sea level in isotope stage 5:
an east Antarctic ice surge at 95,000 BP?" Nature.
283:5748:629-33.
-------
- 118 -
Hollin, John T. and R. Barry, 1979. "Empirical and theoretical
evidence concerning the response of the earth's ice and
snow cover to a global temperature increase." Environmental
International. 2.
Hughes, T., 1983. "The stability of the West Antarctic ice sheet:
what has happened, and what will happen," in Proceedings:
Carbon Dioxide Research Conference: Carbon Dioxide, Science
and Consensus 1983. Washington, DC: U.S. Department of
Energy. CONF-820970.
Hughes, T., J.L. Fastook, and G.H. Denton, 1979. Climatic Warming
and the Collapse of the West Antarctic Ice Sheet. Maine:
University of Maine Press.
Hull, C.H.J., and R.C. Tortoriello, 1979. "Sea level trend and
salinity in the Delaware estuary." Staff paper. West Trenton:
Delaware River Basin Commission.
Idso, S., 1980. "The climatological significance of a doubling
of the earth's atmospheric carbon dioxide concentration."
Science. 207:4438:1462-63.
Kana, Timothy W., et al., 1983. " Shoreline changes due to
various sea-level rise scenarios," in Coastal Zone '83,
Orville T. Magoon, editor. New York: American Society of
Civil Engineers. 3:2768-76.
Keeling, C.D., et al., 1976. "Atmospheric carbon dioxide vari-
ations at Mauna Loa observatory, Hawaii." Tellus, 28.
Keu, Richard A., 1981. "Whither the shoreline." Science.
214:4519:428.
Keyfitz, Nathan, et al., 1983. Global Population (1975-2075)
and Labor Force (1975-2050). Oak Ridge, TN: Oak Ridge
Associated Universities, Institute for Energy Analysis.
Kopec, Richard J.f 1971. "Global climate change and the impact
of a maximum sea level on coastal settlement." The Journal
of Geography. Dec:541-50.
Ku, Teh-Lung, et al., 1974. "Eustatic sea level 120,000 years ago
on Oahu, Hawaii," Science. 183:4128:959-62.
Kukla, G., and J. Gavin, 1981. "Summer ice and carbon dioxide.
Science. 214:4520:497-503.
-------
- 119 -
LaBrecque, John L. , and Peter Barker, 1981, "The age of the
Weddell basin." Nature. 290:5806:489-92.
Lacis, A., et al., 1981. "Greenhouse effect of trace gases,
1970-1980." Geophysical Research Letters. 81:10:1035-8.
Lamb, H. H., 1970. "Volcanic dust in the atmosphere." Philosophical
Transactions of the Royal Society. London Series A, 255:425.
Leatherman, Stephen, 1983. "Geomorphic effects of projected
sea level rise: a case study of Galveston Bay, Texas," in
Coastal Zone '83, Orville T. Magoon, editor. New York: American
Society of Civil Engineers. 3:2890-901.
Lisle, L.D., 1982. "Annotated bibliography of sea level changes
along the Atlantic and Gulf coasts of North America."
Shore and Beach. 50:3.
Lovins, A., et al., 1981. Least Cost Energy. Andover, MA:
Brick House Publishing Co.
Lugo, Ariel, 1980. "Are tropical forest ecosystems a new sink?"
in The Role of Tropical Forests in the World Carbon Cycle,
Sandra Brown, Ariel Lugo, and Beth Liegel, eds. Washington,
D.C.: U.S. Department of Energy. #CONF 800350. 1-18.
Magness, Lt. Col. Thomas H., Ill, 1983. "Sea level rise: impli-
cations for the Corps of Engineers." Statement presented at the
EPA conference on the effects of sea level rise on the
coastal economy and environment. Washington, DC: March 30.
Available from EPA.
Manabe, Syukuro, and Richard T. Wetherald, 1967. "Thermal
equilibrium of the atmosphere with a given distribution of
relative humidity." Journal of Atmospheric Sciences. 24:3:241-59
, 1975. "The effects of doubling the CC>2 concentration
on the climate of a general circulation model." Journal of
Atmospheric Sciences. 32:1:3.
Marshall, J.F., and B.C. Thorn, 1976. "The sea level in the last
interglacial." Nature. 263:5573:5573:120-21.
Mercer, J.H. 1970. "Antarctic ice and interglacial high sea
levels." Science. 168:3937:1605-6.
, 1976. "Glacial history of southernmost South America."
Quaternary Research. 6:125-66.
-------
- 120 -
, 1978. "West Antarctic ice sheet and C02 greenhouse effect
a threat of disaster." Nature. 271:5643:321.
Nardin, Thomas R., et al., 1981. "Holocene sea level curves for
Santa Monica Shelf, California continintal borderland."
Science. 213:4505:331-3.
Newsweek, October 5, 1981. "Is Antarctica Shrinking." 72-4.
Oerlemans, J., 1981. "Effect of irregular fluctuations in Antarctic
precipitation on global sea level." Nature. 290:5809:770-2.
Oeschger, H., et al., 1975. Tellus. 27:168
Reilly, John, Rayola Dougher, and Jae Edmonds, 1981. "Determinants
of global energy supply to the year 2050." Draft. Oak Ridge,
TN: Institute for Energy Analysis, Oak Ridge Universities.
Rasmussen, R.A., Khalil, M.A.K., "Increases in the Concentration
of Atmospheric Methane" Atmospheric Environment, Vol. 15,
No. 5, 1981, pp. 883-886
Rotty, Ralph, 1979. "Energy demand and global climate," in Man's
Impact on Climate, Wilfred Bach, et al., eds. New York:
Elsevier Scientific Publishing.
Rotty, Ralph, and Gregg Marland, 1980. "Constraints on carbon
dioxide production from fossil fuel use." Oak Ridge, TN:
Institute for Energy Analysis, Oak Ridge Associated
Universities. ORAM-IEA-80-9(m).
Rust, Bert W., Ralph Rotty, and Gregg Marland, 1979. "Inferences
drawn from atmospheric C02 data." Journal of Geophysical
Research. 84:C6:3115-22.
Schneider, Stephen H. and Robert S. Chen, 1980. "Carbon dioxide
flooding: physical factors and climatic impact." Annual
Review of Energy. 5:107-40.
Schwartz, M.L. 1967. "The Brunn theory of sea level rise as a
cause of shore erosion." Journal of Geology. 75:76-92.
Seidel, Stephen and Dale Keyes, 1983. Can We Delay A Greenhouse
Warming?. Washington, DC: U.S. Environmental Protection Agency,
Siegenthaler, Ulrich, 1983. "Uptake of excess C02 by an outcrop
diffusion model of the ocean." Journal of Geophysical
Research. 88:C6:3599-608.
-------
- 121 -
Smagorinsky, J., 1982. Carbon Dioxide and Climate: A Second
Assessment. Washington, DC: National Academy of Sciences
Press.
Sofia, S., and A.S. Endal, 1983. "Solar variability." Reviews
of Space Physics, in press.
Sullivan, Walter, 1981. "Study finds warming trend that could raise
sea levels." New York Times. August 22.
Thomas, Robert H., T.J.O. Sanderson, and Keith E. Rose, 1979.
"Effect of climatic warming on the west Antarctic ice sheet."
Nature. 277:5695:355-8.
U.S. Department of Energy, 1982. 1982 DOE Research Conference.
Coolfont, WV: Office of CC>2 Research.
Wang, W.C. et al., 1976. "Greenhouse effects due to man-made
pertubations of trace gases." Science. 194:4266:685.
Weiss, R.F., Keeling, C.D., and Craig, H., "The Determination of
Tropospheric Nitrous Oxide," Journal of Geophysical Research,
Vol. 86, no. 68, August 20, 1981, pp. 7197-7202.
Washington, W.M. and D.L. Williamson, 1977. "A description
of the NCAR global circulation models in methods of
computational physics," in General Circulation Models
of the Atmosphere, J. Chang, ed. New York: - Academic
Press, 17:111-172
Weggel, J.R., 1979. A Method for Estimating Long-Term Erosion
Rates from a Long-Term Rise in Water Level, Coastal Engineering
Aid No. 79-2. Fort Belvoir, VA: U.S. Army Corps of Engineers,
Coastal Engineering Research Center.
Woodwell, G.M., 1978. "The carbon dioxide question." Scientific
American. 238:34-43.
World Meteorological Organization, 1982. WMO Global Ozone Research
and Monitoring Project, Report No. 14.
Worsley, Thomas R., and Thomas A. Davies, 1979. "Sea level
fluctuations and deep-sea sedimentation rates." Science.
203:455-6.
-------
PROTE.CTV.W
AGiNtY
DALLAS, f IX&*
-------
United States Postage and Fees Paid
Environmental Protection Environmental Protection Agency
Agency EPA-335
Washington DC 20460
Official Business SP«cial
Penalty for Private Use $300 Fourth-Class
Rate
Book
------- |