United States
             Environmental Protection
             Policy, Planning,
             And Evaluation
December 1989
The Potential Effects
Of Global Climate Change
On The United States


  Editors:  Joel B. Smith and Dennis Tirpak

United States Environmental Protection Agency
  Office of Policy, Planning and Evaluation
    Office  of Research and Development

            December 1989
                              Printed on Recycled Paper


This document has been reviewed  in accordance with the U.S.
Environmental Protection Agency's and the Office of Management
and Budget's peer and administrative review policies and approved
for publication. Mention of trade names or commercial products
does not constitute endorsement or recommendation for use.


                              TABLE OF CONTENTS
Foreword	  ™
Acknowledgments	  XX111

             Sensitivities	  2
             Direction and Magnitude	  2
             Linkages	  2
             National Impacts	  2
             Regional Impacts	•	  3
             Uncertainties	  3
             Policy Implications	  3
             Research Needs	  3
             Important Systems	  3
             Regional Case Studies	  3
             National Studies 	  4
              Step 1:  Initial Scoping of the Report	  4
              Step 2:  Preparatory Workshops	  5
              Step 3:  Identification of Potential Projects	  5
              Step 4:  Reviews of Proposals 	  5
              Step 5:  Planning and Integration	  5
              Step 6:  Analysis  	  5
              Step 7:  Preliminary Project Review	  5
              Step 8:  Project and Report Peer Review	  5
              National Research and Policy Activities	  6
              International Activities	  6
       REFERENCES .  ,	  7

              Greenhouse Gases	•	   12
                     Carbon Dioxide (CO2)	   12
                     Methane (CH4)  	   13
                     Chlorofluorocarbons (CFCs)	   13
                     Nitrous Oxide (N2O) 	   15
                     Ozone (O3)	   15
              Solar Variations  	   15
              Volcanoes	   15
              Tropospheric Aerosols	   16
              Surface Properties	•	   16

              Internal Variations  	          16
              Water Vapor - Greenhouse Effect		  16
              Snow and Ice  	  17
              Clouds	  17
              Biogeochemical Feedbacks	  18
              Oceans	,	  18
              Observational Evidence of Climate Change	         19
        CLIMATE MODELS  	  21
              Model Projections of a Doubled-CO,, World		  24
        REFERENCES	  25

        FINDINGS	  29
              Temperature	..-..	  32
                     Maximum Temperatures	  32
                     Minimum Temperatures	  33
              Precipitation	 .  . .	  34
                     Droughts	  . .	  34
                     Floods	  35
              Severe Storms - Hurricanes	  35
              Empirical Studies  	  36
              Modeling Studies	     37
              The GISS Study	  38
                     Interannual Variability	  39
                     Daily Variability	  40
                     Variability of the Diurnal Cycle	  40
              The NCAR Study	  40
                     Methods	  40
                     Comparison of Observed Versus Chervin Control Runs  	  43
                     Variability Comparisons of Three CCM Versions and Observed Data	  45
                     Intercomparisons of the Three CCM Versions and Observed Data	  45
              Control Versus CO2-Perturbed Runs	  45
              Interannual Variability	  47
              Daily Variability	  47
                    Comparison of Climate Change	  49
              Limitations of the Two Studies	  49
       RESEARCH NEEDS	  51
              Further Investigation of Variability GCMs	 .	 .	..-...•	  51
              Improvements in GCMs	  52
              Sensitivity Analyses  of Reports	           52
       REFERENCES	  52

       SCENARIO  COMPONENTS	        57
       TYPES  OF SCENARIOS	   58
             Arbitrary Changes	   58

             Analog Warming	• •	  <-q
             General Circulation Models (GCMs)	  ^
             GCMs Used	  ฃJ
                    Limitations	  *ฃ.
             Arbitrary Changes	  ,
             Analog Warming	•	  ,.
             GCM Transient Runs 	  ™
             Limitations 	  ,,
             Sea Level Rise Scenarios	  ฐฐ
             GCMs	  *'
             Scenarios  	•	•	  ,-„

      FINDINGS 	  71
            ' Range Shifts	,	   ' j"
             Productivity Changes	•	  71
             Combined  Impacts With Other Stresses . .	  if
             , Policy Implications	  „
             Distribution and Ownership	   'J1
             Value of U.S. Forests	   L\
             Magnitude	   7(-
             Rates  	•	   yg
             Mechanisms	   _,.
                    Temperature	   „
                    Precipitation	•	   _„
                    CO2 Concentration	   ''
                    Ligfit  	   "
                    Nutrient Status	   ''
                    Atmospheric Chemistry   		   "
                    Disturbances	   ''
                    Landscape Processes	   ^
                    Multiple Stresses	• • • • • ••  • •   /s
              FORESTS	   ^
              Design of the Studies	   ,1ฐ
              Limitations	•	   S1
              Results	  "{
                     Magnitude  	  ฐ^
                     Rates of Decline and Migration	  ฃj
                     Mechanisms of Migration 	• • •  <ฃ
              Ecological Implications 	
                     Tree Distribution and Biomass Productivity	  o^
              Socioeconomic Implications	

                     Quality of the Human Environment	  87
                     Recreation	  37
                     Wood Products	                         37
       FOREST POLICY AND CLIMATE CHANGE	 .  . .........   88
              How Much Land Should Be Forested?	I'.'.'.'.'.'.'.'.  88
              How Much Should Be Withdrawn From Timber Production?  	:.'.'.'.  88
              How Should We Manage Federal Forests?	  89
              How Can We Ensure National Goals?	  89
                     Reforestation		                       89
              Who Should Pay? 	           	  90
       RESEARCH NEEDS 	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.  90
              Effects of Climate Change	  90
              Methods	               	  90
              Forest Management	  9j
              Timing of Research	                                   91
       REFERENCES	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.  91

CHAPTER 6: AGRICULTURE	                                                93
       FINDINGS  	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.  93
              Crop Yields	  93
              Economic Impacts	  93
              Irrigation Demand	I'.'.'.'.'.';'.'.'.'.'.'.'.'.'.  94
              Agricultural Pests  	I'.'.'.'.'.'.'.  94
              Farm-Level Adjustments 	'.'.'.'.'.'.'.'.'.'.'  94
              Livestock Effects	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.  94
              Policy Implications	                 94
             Structure of and Rationale for the Studies	   97
                    Variability	•...'.'.   99
                    Timing of Effects 		   99
       RESULTS OF AGRICULTURAL STUDIES  		'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.' 100
             Regional Crop Modeling Studies  	[\\  100
                    Design of the Studies	  100
                    Limitations	'  IQQ
                    Results	  100
                    Implications  	  102
             Regional and National Economics Study	  102
                    Study Design  	'.'.'.'.'.  103
                    Limitations	           103
                    Results	   104
                    Implications	'.'.'.'.'.  106
             Demand for Water for Irrigation	  107
                    Irrigation Requirements in the Great Plains	  107
                    Water Resources for Agriculture in California .	  108
                    Implications for Demand for Irrigation Water 	  109
             Direct Effects of CO, on Crops	'.'.'.'.'.  109
             Climate Impacts on Pest-Plant Interactions	'...'.'.'.'.'.'.'.'.'.  110
                   Study Design and Results 	  HO
                   Limitations	  HI
                   Implications  	''   m
            Effects of Climate Change on Water Quality	'..'.-'.'.'.'.'.'.'.'.'.'.  Ill
                   Study Design   	  HI

                    Results	  .~
                    Implications 	   ....ป
             Climate Variability 	   -.,,
             Farm-Level Management Adjustments to Climate Change	   ^
                    Study Design	   -^
                    Results	   ^~
                    Implications 	   ....
             Livestock	   ^4
                    Design of Studies  	   j-rj
                    Limitations	•	   ^.
                    Results .	   115

             Costs and Timing of Adjustment	   ||
             Effects of CO? 	   ^6
             Environmental Quality	 . . .	
             Global Agriculture	•	   117
             Commodity Policies	   ^-,
             Land-Use Programs	   118
             Water-Resource Management Programs	   f^ฐ
             Water Quality Policy	   tL
             Risk Management and Drought Policy	
             International Trade Agreements	
             Agricultural Contributions to the Greenhouse Effect	  j-Jฐ
             Agricultural Research	•	  1 ig
       RESEARCH NEEDS  	  |~0

       FINDINGS 	   123
              Policy Implications	   104
       CAUSES, EFFECTS, AND RESPONSES	 - - -	   gj
              Causes  	   ^25
              Effects	   17r
                     Destruction of Coastal Wetlands	<	   ^
                     Inundation and Erosion of Beaches and Barrier Islands	   u&
                     Flooding	   127
                     Saltwater Intrusion	•	    „
              Responses	    10Q
              Loss of Coastal Wetlands and Dryland	    ^ j-
                      Study Design	
                      Lunitations	• • • •	    ^2
                      Results	    1 ,,,
              Costs of Defending Sheltered Shorelines	•	    f;"
                      Study Design	    J34
                      Limitations	•	   ~~.
                      Results	   ..-,,-
              Case Study of the Value of Threatened Coastal Property . .	   f^o
                      Study Design	   ^

                      Results	   137
               Nationwide Cost of Pumping Sand Onto Recreational Beaches	   137
                      Study Design	   137
                      Limitations	   137
                      Results	   137
               Synthesis of the Three National Studies 	   138
                      Study Design 	'. .  .   138
                      Limitations	   139
                      Results	               140
       POLICY IMPLICATIONS	  .   141
               Wetland Protection	  ....   141
                      Options for State and Local Governments	   141
                      The Federal Role	'	   142
               Coastal Protection	   142
                      State and Local Efforts	   142
                      Federal Efforts	   143
                      Sewers and Drains  	        144
       RESEARCH NEEDS	   145
       REFERENCES	   145

       FINDINGS	   149
              Species Diversity	   149
              Marine Ecosystems	   149
              Freshwater Ecosystems 	   149
              Migratory Birds	   149
              Policy Implications	      149
              The National Resource 	   150
              Species' Diversity	   151
              Stressed Biological Diversity	   152
              Genetic Diversity	   152
              Community and Ecosystemic Diversity	   152
              CHANGE	  152
              Rate of Climate Change	  153
                     Effect on Genetic Diversity	  154
              Barriers to Response	  154
                     Reserve and Island Species	  154
                     Mountain Species	     155
              Forest Ecosystems	   155
              Tropical Forest Ecosystems	   156
              Freshwater Ecosystems	   156
              Saltwater Ecosystems	   156
              Coral Reef Ecosystems	   157
              Arctic Ecosystems	   157
              Migratory Birds	   157
              Endangered Species	   158
              Other Direct and Indirect Stresses	      158
              Management Options to Maintain Biological Diversity	   158
                     Maintenance of Native Habitats	   158
                     Maintenance of Species in Artificial Conditions  .	   159

                     Restoration of Habitat 	   16Q
              Planning Options	-	   160
       RESEARCH NEEDS	   160
              Identification of Biological Diversity	
              Species Interactions and Biological Diversity	   ^

       FINDINGS 	•	   165
              Impacts on Water Uses	   165

                              CHANGE ON TOE WATER RESOURCES IN THE UNITED*

              STATES	'	   166
              Current Status of Water Resources 	
              Climate Change, Hydrologic Conditions, and Water Resources	   |_ป
              Impacts of Climate Change on Water Uses	   J^j-
                      Irrigation  	'   -. 72
                      Thermal Power Generation	• •
                      Industrial Uses	
                      Domestic Water Uses	• •
                      Navigation 	' ' •	
                      Hydropower	   ^2
                      Recreation	•	   YJT.
               Climate Change and Water Quality	    „.
               Climate Change and Flood Hazards	     .
               Climate Change and Conflicts Among Water Uses	   |'*
               The West  	   176
               Pacific Northwest  	•	•  • • '	   176
               California	•••••••	
               Colorado, Rio Grande, and Great Basins	   177
               Great Plains	   -^77
               Great Lakes	'	
               Mississippi River	•	'
               Northeast  	•	
               Southeast  	'	   170
               Supply and Structural Policy Approaches 	
                      Design for Uncertainty  . . .	   ._„
                      Surface Water Development 	•	     _
                      Optimization of Water Resource Systems	• • •   *'*
               Demand Management and Nonstructural Policy Approaches		   |o"
                      Water Pricing, Water Markets, and Water Conservation . .	   fปu
                      Drought Management Policies	•	•	
                      Water  Quality	 ^	
                      Policies for Floodplains	• •	   ^
        RESEARCH NEEDS  	' '	   184
        REFERENCES	'	

        FINDINGS	'	" '   187
                Policy Implications	   1Ro

              Study Design	  189
              Limitations 	  190
              Results	  	  191
       POLICY IMPLICATIONS	              	  196
       RESEARCH NEEDS  		  196
       REFERENCES	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.  197

 CHAPTER 11:  AIR QUALITY  ...	              199
       FINDINGS	'.'.'.'.'.'.'.'.	  199
       RELATIONSHIP BETWEEN CLIMATE AND AIR QUALITY ...... .. ...... '. '. '. '. '. '. '.  200
              Ventilation	  200
              Circulation 	.	!!!!!!  200
              Precipitation	                        200
       PATTERNS AND TRENDS IN AIR QUALITY .	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.  201
              Total Suspended Particulates	. . .	; . . . .	'.'.'.  201
              Sulfur Dioxide	  201
              Ozone	  202
              Acid Deposition  		                  202
       STUDIES OF CLIMATE CHANGE AND AIR QUALITY	'.'.'.'.'.'.'.'.'.'.'.'.'.'.   202
              Climate Change and Its Interactions with Air Chemistry .. ..	   205
                    Effect of Climate Change on Ozone Formation	         205
              Study Design	]   210
              Limitations	   210
              Results	   211
                    Central California Study		   211
                    Midwest and Southeast Study	   211
              Population Exposure	            213
              Ozone	 .	   213
             Acid Rain	                                      215
       POLICY IMPLICATIONS	!!.'.'	'"	   216
       RESEARCH NEEDS  	         	   216
       REFERENCES	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.   217

CHAPTER 12: HUMAN HEALTH 	                    219
       FINDINGS	 . . . .   	   219
       CLIMATE-SENSITIVE ASPECTS OF HUMAN HEALTH . . . '. '. .. ...... .... ........   219
             General Mortality and Illness	' . '   220
             Cardiovascular, Cerebrovascular, and Respiratory Diseases	   221
             Vector-Borne Diseases	  222
             Human Reproduction 	                        222
       POTENTIAL HUMAN HEALTH EFFECTS OF CLIMATE CHANGE' ...'.'.'.'.'.'.'.'.'.'.'.'.  222
             General Mortality	..	  223
             Cardiovascular, Cerebrovascular, and Respiratory Diseases	  225
             Vector-Borne Diseases  	,	  .......  226
                    Tick-Borne Diseases	  227
                    Mosquito-Borne Diseases 	  227
             Other Diseases	                   231
      SOCIAL AND ECONOMIC IMPLICATIONS	 ... .  .   	   232
      POLICY IMPLICATIONS	           .               	   233

      RESEARCH NEEDS  	• • •	            235
      FINDINGS	   237
             Northern and Southern Cities	   237
             Coastal Cities  	   237
             Water Supply and Demand  	   237

                                                                          ' ' ' ' ' ' ' '
             Impacts on Miami, Cleveland, and New York City  	
                     Study Design 	'   240
                     Limitations	   24Q
                     Results and Implications	•• •	   . ,.
              Implications Arising From Other EPA Studies in This Report	  **>
              Metropolitan Water Supply	  245
                     Washington, DC	  246
                     New Orleans	•	  246
                     New York City		•	  246
                     Tucson	• • •	•	
              Water	•	   246
              Drainage and Wastewater Systems	   247
              Coastal Defenses	]   247
              Roads	• •:	   247
              Bridges	   247
              Mass Transit	   247
              Electricity and Air-Conditioning		• •	•	   „,-
        POLICY IMPLICATIONS			   247
              Investment Analysis Methods	   248
              Water Supply	   248
              Infrastructure Standards	   24g
        RESEARCH NEEDS 	   248
        REFERENCES	'	

        FINDINGS			••••-	'	  251
               Water Resources	•	'	  251
               Wetlands and Fisheries	 ..	•	  ^52
               Agriculture	 • • •	  252
               Natural Vegetation  	'  252
               Air Quality	•	•	' '	   252
               Electricity Demand . .		   252
               Policy Implications	   7r^
               Current Climate	•	   253
               Water Resources	   253
                      Water Distribution
                      Flood Control and Hydroelectric Power	•
                      Sacramento-San Joaquin River Delta ...		

       Forestry	'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.	  256
       Natural Vegetation 	                             	  ฃ7
       Wetlands	          '	  f'
       Wildlife and Fisheries	'.'.'.'.'.'.'.'.'. '.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.	  257
       Recreation and Nature Preservation	               	  9<7
       Forests	  257
       Water Resources		  7ro
CALIFORNIA STUDIES IN THIS REPORT 	'.'.'.'.' '.'.'.'.'.'.'.'. '. '. '. '. '..'.".' " ' .'	'  258
       Analyses Performed for This Study	             	'	  258
       Hydrology of Catchments in the Central Valley Basin	" '  262
             Study Design	      262
             Limitations	.[........	  262
             Results	!!!!!.'!.'.'!!	   262
             Implications	   263
       Water Resources in the Central Valley Basin	       264
             Study Design  	.	   264
             Limitations	• • • •   ^
             Results	...................'.'.'.'.'.'.'.'.'.   264
             Implications  	_             ' ' ' '  255
       Salinity in San Francisco Bay	  266
             Study Design	  266
             Limitations	!.!!!!....!.	  266
             _  ,.   .               •	•	  26/
             Implications 	                  267
      Wetlands in the San Francisco Bay Estuary	 .  . .	  268
             Study Design	  268
             Limitations	          	  2fiซ
             Results	                 '	  0--o
             T   ,.  ..       	•	  2oo
             Implications 	           269
      California Agriculture	'.'.'.'.'.'.'.	  269
             Study Design	'•'-'•'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'."' 270
             Limitations	          '	 270
             Results		.................               270
             Implications	' '	 271
      Regional Implications of National Agriculture Changes . .	...... . .	 273
             Results	 273
      Water Quality of Subalpine Lakes	 273
             Study Design	    273
             Limitations	 >          	 273
             Implications  	;	 273
      Summary of Effects on Water  Resources	     ••••••••..,.......
      Vegetation of the Sierra Nevada		'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'.'"•""  275
             Study  Design	  275
             Limitations	       ] •	  275
             Results	]   	  276
             Implications  		  j7f-
      Electricity Demand	  277
             Results	.		  277
            Implications	 .                      	  27o
     Air Pollution			'.'.'.'.'.'.'.'.'.'.'.'.'.'.'. '.'. '.'.'.'.'.'.	  278
            Results	......'!	  278

                    T   ,.  t.                                               	  278
                    Impucations 	               27ป
      POLICY IMPLICATIONS	• •  • • •	  ,78
             Water Supply and Flood Control	•	
                    Approaches for Modifying the Water Resource System	
                    Options for Allocating Water Shortages	
             Sacramento-San Joaquin River Delta	•	
                    Delta Island Land Use
                    Water Quality of the San Francisco Bay Estuary
             Water Quality of Freshwater Systems
             Terrestrial Vegetation and Wildlife  ...		•
             Agriculture	•283
             Wetland Vegetation and Fisheries	
              Shoreline  Impacts of Sea Level Rise	 ^
              Energy Demand	 283
              Air Quality	 284

CHAPTER 15: GREAT LAKES	• • • •	 287
       FINDINGS 	•	'...... 287
              Lakes	•	:	•	'	 287
              Water Quality and Fisheries	•	 288
              Forests	  288
              Agriculture	•	  288
              Electricity Demand	  288

                                       RKOURCEs'lN 'THE GREAT'LAKES REGION' '. '. '. '.  289
              Current Climate  	•••'•;	''	  290
              The Lakes	• •	  290
                     Lake Regulation
                     Climate-Sensitive Uses of the Lakes 	
                     Climate and Water Quality	
                     Fluctuating Lake Levels	•	
              Land Around the Lakes	•	•290
                     Land Uses 	•	;	•	
              Direct Effects on Lakes	
                      Impacts of Lake Changes on Infrastructure  	
              Water Quality	•	 294
              Forests	•	 295
              Agriculture  . .	 295
               Energy	" ' 295
               Policy	•	•	
               Lakes		, •	• • • •	*296
                       Lake Levels	
                       Effects of Lower Lake Levels
                       Ice Cover	302
                       Shipping	  3Q4
               Water Quality .	;	' ' ' '
                       Thermal Structure of Southern Lake Michigan
                       Eutrophication of the Lake Erie Central Basin

              Forests	t	 309
                     Potential Range Shifts	 309
                     Transitional Effects	 . ;	 399
                     Forest Migration	 311
                     Implications of Forest Studies	 311
              Agriculture	[[[ 312
                     Crop Yields	     313
                     Regional Shifts	,, ...	 314
                     Adjustments by Illinois Corn Producers	 315
              Electricity Demand	-.....,	 . . . 316
                     Study Design	 316
                     Results	'..'..'.'.'.'.'.'. 316
                     Implications 	                                316
       POLICY IMPLICATIONS	'.'.'.'.'.'.','.'.'.'.'.'.'.'.'.'.'.'. 316
              Water Supply Issues		           317
                     Lake Regulation	r	 317
                     Withdrawals	 .   317
                     Shipping		 317
                     Pollution Control	 317
                     Fisheries	      317
              Land Use	 318
                    Shorelines	,	  318
                    Forestry	 .	  318
                    Agriculture	'.....-•		  318
              Demographic Shifts	                       	  318
       REFERENCES	'.'.'.'.'.'.'.-'.'.'.'.I'.'.'.'.'.'.'.'.'.'.  319

CHAPTER 16:  SOUTHEAST	                              323
       FINDINGS 	.'	'.'.'.'.'.'.'.'.'.'.'.'.['.'.  323
              Agriculture	t	[	'.'.'.'.'.'.'.'.'.  323
              Forests	             323
              Water Supplies	               323
              Sea Level Rise  	' . . ]	'.','.'.'.'.'.'.'.'.'.'.'.'.  324
              Marine Fisheries	  324
              Electricity Demand	 ', '.  324
              Policy Implications	                                      324
       CLIMATE AND THE SOUTHEAST 	'.'.'.'.'.'.'	  325
       CLIMATE-SENSITIVE RESOURCES OF THE SOUTHEAST .....  ^ .................  325
             Water Resources	              325
             Estuaries	;	               326
             Beach Erosion and Coastal Flooding		' '  327
             Agriculture	              327
             Forests	            327
             Indoor and Outdoor Comfort	     	     328
             SOUTHEAST		.	  328
             Flooding	          32g
             Wetlands	  329
             Infrastructure  	          	329
      RESULTS OF SOUTHEASTERN STUDIES  	'.'.'.'.'.'.'.'.  333
             Coastal Impacts	t	_         333
                   Coastal Wetlands  	..'.....':'...'.'.'.'.'.'.'.'.'.'.'.'.  333
                   Total Coastal Land Loss	  334

                    Cost of Protecting Recreational Beaches  	  334
                    Cost of Protecting Calm-Water Shorelines	
             Tennessee Valley Authority Studies 	
                    TVA Modeling Study  	
                    Results  . 	
                    Tennessee Valley Policy Study
             Studies of the Impacts on Lake Lanier and Apalachicola Bay
                    Lake Lanier 	•
                    Apalachicola Bay 	•	
                  '  Crop Modeling Study
                    Shifts in Production
                    Agricultural Pests
                    Implications of Agriculture Studies
                    Potential Range Shifts	•
                    Transitional Effects	
             Electric Utilities
             Agriculture and Forests	
             Water Resources
                     Impacts of Wetter Climate
                     Impacts of Drier Climate  	•
                     Is Current Legislation Adequate?	
              Beach Erosion	• • •

       FINDINGS 	• • •359
              Agriculture	      359
              Ogallala Aquifer	•	
              Water Quality
              Electricity Demand	
              Policy Implications
              Dryland Agriculture
              Irrigated Agriculture	
              Water Quality  .,	-
              Electricity Demand, . 	
              Crop Production	  ^,,
                      Study Design  	  *™
                      Limitations	•	• •    _,
                      Results		• •	' '
                      Implications ...........		•	•	•
              Agricultural Economics	
                    •  Implications  	

              Irrigation	 370
                     Study Design 	 370
                     Limitations	 370
                     Results	 370
                     Implications 	 371
              Water Quality	 371
                     Study Design	 372
                     Results	 372
                     Implications 	 372
              Livestock	 373
              Electricity Demand	 373
                     Results	 373
                     Implications 	         374
              Land-Use Management	 375
              Water Resource Management  	 375
              Risk Management	               375
       REFERENCES		 376

CHAPTER 18:  RESEARCH NEEDS	         379
              Institutional Response to Climate Variability and Climate Change	 381
              Climate System	 383
              Research Scales	 383
              Socioeconomic Impacts  	 383
              Data	 383
              Objectives of Federal Global Change Program	 383
              Three Major Scientific Objectives	  385
       IMPACT ASSESSMENT METHODOLOGY	                       386
       REFERENCES	  388

              Strategic Assessments 	  389
                    Decision-Oriented Assessments	  390
                    Program-Oriented Assessments	  390
                    Problem-Oriented Assessments	  390
              Criteria for Choosing a Strategy	            390
              No Immediate Action 	  394
                    Reservoir Operating Rules  	  394
                    Choice of Crops	  394
              Anticipatory Action	  394
                    Modifying Ongoing Projects to  Consider Climate Change	  394
              Undertaking New Projects Primarily Because of Future Climate Change	  395
              Planning: Changing the Rules of the Game	  396
                    Land Use	  395
                    Water Allocation	  398

            Research and Education: Increasing Our Understanding
                   Research and Development . .	
                   Education	•	







      D: FORESTS



      G: HEALTH



      J: POLICY




        I am pleased to transmit the attached Report to Congress:  The Potential Effects of Global Climate
Change on the United States.  This report, written in response to a congressional request in the Fiscal Year 1987
Continuing Resolution Authority to prepare two reports on  climate  change,  focuses on the health and
environmental effects of climate change.  A second draft report, Policy Options for Stabilizing Global Climate,
is being revised in preparation for delivery to Congress.

        This report is one  of the  most comprehensive published studies of the potential impacts  of  the
greenhouse effect.  It examines national effects and, more specifically, impacts on four regions of the United
Statees:  California, the Great Lakes, the Southeast, and the Great  Plains.  Fifty studies  conducted by
government, academic, and  consulting scientists to examine impacts are included.  EPA provided common
scenarios of climate change to the scientists for use in their analyses.  This report is an overview of the results
of those studies.

        I invite you to carefully read the Executive Summary and  the chapters that follow. Although it is difficult
to summarize such a large and comprehensive project in a few words, it is fair to say that climate change could
lead to  significant changes in many ecological and socioeconomic  systems. The environmental impacts of a
relatively rapid climate change may be particularly acute.  Sea level rise could lead to the loss of many coastal
wetlands, while a rapid wanning could reduce the populations of many plants and animals and,  in some cases,
lead to  extinction of species.

        The socioeconomic effects, especially on a regional scale, also  may be  quite important. Significant
expenditures may be needed for such measures as protecting  areas from  sea level rise, building dams  and
reservoirs for flood and drought protection, modifying infrastructure, and adding electricity capacity.

        I urge caution in interpreting  the results of these studies.  Since we cannot predict regional  climate
change  or extreme events such as hurricanes or droughts, we cannot predict impacts. The work done for this
study was based on scenarios of climate change and is indicative  of what could occur in the future. So, too, this
work does not identify all of the impacts of climate change, the interactions, or the economic damages that could

        In examining a study such as this, there is often a temptation to identify "winners" and "losers."  One
must be careful in drawing such conclusions.  The scenarios are  based on a certain point in time (when carbon
dioxide levels have doubled); and they assume that climate stops  changing.  If emissions are not stabilized,
climate change will not stop at this carbon dioxide doubling, but will continue to warm. With continued wanning,
what was a positive effect could become negative. Responding to climate change would be a matter of keeping
up with increasing rates of change.

        I feel this report is a significant contribution to our understanding of climate change impacts. More
work needs to be done on understanding impacts on other systems and  regions.  Yet, this information will be
helpful as we address the difficult problems associated with climate change.
                                                Terry Davies
                                                Assistant Administrator
                                                Office of Policy, Planning and Evaluation


    This report was made possible because of the hundreds of people who participated in workshops, conducted
research projects, reviewed draft manuscripts, and contributed ideas that shaped the final product. They shared
a common belief that an objective analysis of global climate change could be undertaken despite the uncertainties
in scientific information.  We are grateful for their support and encouragement.  It was the difference that
sustained us through this effort. In particular, we wish to acknowledge the authors who organized and integrated
the following chapters.

    The following people  contributed to this report:
          Chapter 1: Introduction

          Chapter 2: Global Climate Change

          Chapter 3: Climate Variability

          Chapter 4: Methodology

          Chapter 5: Forests

          Chapter 6: Agriculture

          Chapter 7: Sea Level Rise

          Chapter 8: Biological Diversity

          Chapter 9: Water Resources

          Chapter 10:  Electricity Demand

          Chapter 11:  Air Quality

          Chapter 12:  Human Health

          Chapter 13:  Urban Infrastructure

          Chapter 14:  California
          Chapter 15: Great Lakes

          Chapter 16: Southeast
Joel B. Smith

Alan Robock

Linda O. Mearns

Joel B. Smith

Jack K. Winjum
Ronald P. Neilson

Cynthia Rosenzweig
Margaret M. Daniel

James G. Titus

Lauretta M. Burke
Ross A. Kiester

Mark W. Mugler
Michael  C. Rubino

Kenneth P. Linder

Joseph J. Bufalini
Peter L. Finkelstein
Eugene C. Durman

Janice A. Longstreth

Ted R. Miller

George A. King
Robert L. DeVelice
Ronald P. Neilson
Robert C. Worrest

Joel B. Smith

James G. Titus

          Chapter 17:  Great Plains

          Chapter 18:  Research Needs

          Chapter 19:  Preparing for Climate Change
Cynthia Rosenzweig
William E. Riebsame

Anthony Janetos

James G. Titus
     Special thanks are extended to Roy Jenne of the National Center for Atmospheric Research.  Mr. Jenne
and Dennis Joseph of his staff collected data from the GCMs and assembled them in a way that could be easily
used by the effects researchers.  In addition, he collected historic weather data for 1951-80 and distributed them
to the researchers  as needed.

     We wish to thank James Hansen, Syukuro Manabe, Richard Wetherald, and Michael Schlesinger for providing
us with the results from their GCM runs.  Special thanks are also necessary to Joan O'Callaghan and Karen
Swetlow for editing; Roberta Wedge for assistance on production of the report; and Margaret Daniel, Michael
Greene, and Chris Parker for research and administrative assistance.

     This work was conducted within EPA's Office of Policy Analysis, directed by Richard Morgenstern, within
the Office of Policy, Planning and Evaluation, administered by Linda Fisher, and most recently by Terry Davies.
Support was provided by EPA's Office of Environmental Processes and Effects Research, directed by Courtney
Riordan, within the Office of Research and Development, administered by Eric Bretthauer.

                          EXECUTIVE  SUMMARY
     Scientific theory suggests that the addition of
greenhouse gases to the atmosphere will alter global
climate,  increasing  temperatures  and  changing
rainfall and other weather patterns.  In 1979, the
National Academy of Sciences estimated the most
probable global warming from a doubling of carbon
dioxide concentrations over preindustrial levels to be
between  1.5 and  4.5ฐC.   In 1985,  the  World
Meteorological Organization (WMO), the  United
Nations Environment Programme (UNEP), and the
International Council of Scientific Unions (ICSU)
reaffirmed these estimates. Such a climate  change
could have significant implications for mankind and
the  environment: it could  raise sea level,  alter
patterns of water availability, and affect agriculture
and global ecosystems.

     Although  there is consensus  that increased
greenhouse  gas concentrations will change global
climate, the rate and magnitude of change are not
certain   (see  box   entitled  "Climate  Change").
Uncertainties about climate feedbacks from clouds,
vegetation, and other factors make it difficult to
predict the exact amount of warming that a given
level of greenhouse gases, such as doubled carbon
dioxide (CO2) concentrations, would  cause. How
quickly  climate may change also  is not known,
because  scientists are  uncertain both about how
rapidly heat will be taken up by the oceans and
about some climate feedback processes. Generally,
scientists assume that current trends  in emissions
will continue and that climate will change gradually
over the next century, although at  a much faster
pace than historically. At this rate, the full effect of
the  equivalent  doubling  of CO,  concentrations
probably would not be experienced until after 2050.
It is possible,  however, that sudden changes in
ocean circulation could cause abrupt changes in
global climate.  Indeed, if climate  changed more
rapidly than estimated, adapting to the  effects would
be more difficult and more  costly.  Furthermore,
continued emissions of greenhouse gases could raise
atmospheric concentrations beyond doubled  CO2
causing greater and more rapid  climate changes,
and  larger effects.

     To explore the  implications of  climate change
and  ways to control it,  Congress asked the U.S.
Environmental  Protection  Agency  (EPA)  to
undertake two studies on the greenhouse effect: the
first study was to address "The potential health and
environmental effects of climate change including,
but not be limited  to, the potential impacts on
agriculture, forests, wetlands, human health, rivers,
lakes, estuaries as well as societal impacts;" and the
second study was to examine "policy options that if
implemented would  stabilize  current levels  of
greenhouse gas concentrations."  The second study,
"Policy Options for Stabilizing Global Climate," is a
companion report to this document.

     EPA responded to this request by first holding
workshops with atmospheric scientists to discuss the
use of global climate change models  for impact
analyses   and  then  meeting  with  ecologists,
hydrologists,   geographers,   and   forestry  and
agricultural specialists to identify topics for this
study.  A major purpose was to bridge the gap in
our ability to relate a rise in average annual surface
temperatures to regional climate changes. Based on
these and  other discussions, EPA decided to use
common scenarios of climate change to analyze the
sensitivities of coastal resources, water resources,
agriculture, forests, biodiversity, health, air pollution,
and electricity demand  to  climate   change on
regional and national scales (see Figure 1).  These
systems were chosen for analysis because they are
sensitive to climate  and significantly affect  our
quality of life.  EPA decided  to conduct regional
analyses  for the  Southeast,  the  Great  Plains,
California, and the Great Lakes, because of  their
climatological,   ecological,   hydrological,   and
economic  diversity.     Leading  academic  and
government scientists  in the relevant  fields  used
published models to estimate the impacts on both
the regional and national scales. As a common base
for  conducting  these  analyses,  they  used  the
scenarios specified by EPA.

     After consulting with scientific experts, EPA
developed  scenarios  for  use  in effects analysis.
Regional data from atmospheric models known as
General Circulation Models (GCMs) were used as
a basis for climate change scenarios (see box on
"Scenarios and Methodology").  The  GCMs are
large models of the ocean-atmosphere  system that
simulate the fundamental physical relationships in
the  system.  GCMs  provide  the best scientific

Executive Summary
                                      CLIMATE CHANGE

        A panel of experts convened by the National Academy of Sciences (National Research Council,
   1987) recently gave the following estimates of scientific confidence in predictions Of the climate response
   to increased greenhouse gas concentrations. This table summarizes only their conclusions concerning "the
   possible climate responses to increased greenhouse gases," The full report should be consulted for the

        Large Stratospheric Cooling (virtually certain).  The combination of increased cooling by
        additional  CO2 and other trace gases, and reduced heating by reduced ozone, "will lead to  a
        major lowering of temperatures fa the upper stratosphere,"

        Global-Mean Surface Warming (very probable).  For an equivalent doubling of CO^, "the
        long-term global-mean surface warming is expected to be in the range 1,5 to 45ฐC."

        Global-Mean Precipitation Increase (very probable)* ^Increased heating of the (Earth's] surface
        will lead to increased evaporation and, therefore, to greater global mean precipitation. Despite
        this increase in global average precipitation, some individual regions might welt experience
        decreases in rainfall,"

        Reduction of Sea Ice (very probable). This will be due to melting as the climate warms,

        Polar Winter Surface Warming (very probable). Due to the sea ice reduction, polar surface
        air may warm by as much as 3 times the global average.

        Summer Continental Drvness/Warming (likely in the long term). Found in several^ but not all,
        studies, it  is mainly caused by earlier termination  of winter storms.  "Of course,  these
        simulations of long-term equilibrium conditions may not offer a reliable guide to trends over
        the next few decades of changing atmospheric composition and changing climate.*

        Rise in Global Mean Sea Level (probable). This will be due to thermal expansion of seawater
        and melting or calving of land ice.


Sea Level Rise
Water Resources
Electricity Demand
Air Quality
Human Health
Urban Infrastructure

— *•


Great Lakes
Great Plains


Sea Level Rise
Electricity Demand



Data Bases

                               Figure 1. Elements of the effects report.

                                                                           Effects of Climate Change
                            SCENARIOS AND METHODOLOGY

       A number of scenarios were specified by EPA to help identify the sensitivities of natural and
   manmade systems to climate change. Scenarios were used as inputs with models of natural resources,
   Most researchers used GCM*based scenarios*  Some used  analog scenarios or expert judgment*

      - Regional outputs from three General Circulation Models: {GCMs) were used: the Goddard Institute
   for Space Studies (GISS); the Geophysical Fluid Dynamics  Laboratory (GFPI/); and Oregon State
   University (OSU).   All  of  these models estimate climate  change caused  by a doubling of CCL
   concentrations in the atmosphere. The regional estimates of doubled CO~changes were combined with
   1951*80 climate  observations to create doubled  CO, scenarios*  This GISS model has been used to
   estimate how climate may change between BOW anrithe middle of the next century.  This is called a
   transient run* the outputs of which were used to  create a transient scenario.

        Other approaches were used to supplementthe GCMs* Weather observations from the 1930s were
  -used as an analog for global warming, although greenhouse warming may raise temperatures much, higher
   than they were in that decade* In some cases, pateoclimatic warmings were studied to provide evidence
   of how species respond to climate change.  In addition, the  use of scenarios was supplemented by expert
   judgment (gathered thoughJiterature reviews and workshops with scientific experts) to provide the best
   opinions on.potential effects,     ,_                                                       —
                                           '  -                               -'.
        Since we cannot predict the exact nature of climate change, we cannot predict its impacts. All these
   analytic approaches help us-determine the potential sensitivities and vulnerabilities of systems to climate
   change, ,                      ~ '                   "  ,        -  ,
estimates of the impacts of increased greenhouse
gas concentrations  on  climate.   Yet,  they use
relatively simple models of oceans and clouds, both
of which will be very critical in influencing climate
change.  The  GCMs generally agree concerning
global and latitudinal increases in temperature, but
they disagree and are less reliable concerning other
areas, such as  regional changes in rainfall and soil
moisture.   The  GCM data were compared with
historic meteorologic data.  In addition, the decade
of the 1930s  was  used as an  analog for global

    In Figure 2, the temperature changes from the
three GCMs used to create;scenarios are shown for
both the United States and  four  regions of the
United States for  a  doubling of carbon  dioxide
levels.   The  GCMs  agree on  the  direction of
temperature changes,  but differ hi the magnitude.
Estimates of precipitation  changes are shown in
Figure 3.   The GCMs agree that annual rainfall
would  increase across the country,  but disagree
about  the  direction  of  regional and seasonal
changes.  All models show increased evaporation.
     The GCM results should not be considered as
predictions,  but as plausible scenarios  of future
climate change. Ideally, one would like to use many
regional  climate change  scenarios to reflect the
potential range of climate change.    Resource
constraints allowed us to use only a limited number
of regional  climate scenarios.   It would also be
useful to estimate the probabilities of occurrence for
each scenario.  Given the state of knowledge, it is
difficult to assign probabilities  to  regional climate
change. Because the regional estimates of climate
change by GCMs vary considerably, the scenarios
provide a range of possible changes  in climate for
use in identifying the relative sensitivities of systems
to higher temperatures and sea level rise. Hence,
the results of the studies should not be considered
as predictions, but as indications of the impacts that
could occur  as a result of global warming.

     There are two other major limitations in the
GCM scenarios.  First, the scenarios assume that
climate  variability does not change from recent
decades.  Second, the scenarios did not change the
frequency of events, such as heat waves, storms,
hurricanes, and droughts in various regions, which

Executive Summary
                                                              2xCO2 less 1xCO2
                SoutlMt! GraHPttis Gallon*  UntedSMes-
                      * Lower 48 States
Great Lakes Southeast Great Plains California  United Stales*

             Goddard Institute for Space Studies

     jg$jfl%;g] Geophysical Fluid Dynamics Laboratory

     I	| Oregon State University
                                                                                                               Southeast Great Plains California  United Stales'
                                                  Figure 2.   Temperature scenarios.
                                                              2xCO2 less 1xCO2
         GiMILlliM  Southwl  CntlPWns  California l>iปdSaas'
                       •Lower 49 States
Great Lakes Southeast Great Plains California  United Slates'

              Goddard Institute for Space Studies

      [%$jjjjjjjjj|  Geophysical Fluid Dynamics Laboratory

              Oregon State University
                                                                                                       Great Lakes Southeast Great Plains California  United Stales*
                                                                                                            NC = No Change
                                                  Figure 3.   Precipitation scenarios.

                                                                            Effects of Climate Change
           Climate Scenarios
           **   Differences Between Scenarios. The <3CM and other scenarios do not provide consistent
               estimates of climate change,

           -   Variability, The scenarios assume ao change in variability,

           ~   Ma|or Climate Events,  Tjbe scenarios assume no changes It* hurricanes, droughts, etc,

           Societal Changes. Most studies did not consider changes in population technology, and other
           areas. There was only limited consideration of responses and adaptation measures, which could
           mitigate some of the results presented here,

           Linkages, Many indirect effects (e.g., effect of increased irrigation demand on water resources)
           were not quantitatively analyzed*

           Limited Effects Analyses.  Many effects and regions in the United States were not analyzed.
           In addition, this report did not analyze the impacts of climate change on other countries,
           Compared to the United States, it may be much more difficult for poorer and Jess mobile
           societies to respond to climate change.  It is not unreasonable to assume that climate change
           could have important: geopolitical consequences,, which could have subsequent impacts on the
           United States.

           Effects Models. These models were calibrated for historic  climate conditions and may not
           accurately estimate future response to climate change.
would have affected the results presented in this
report  (see  "Limitations"  box).    Changes  hi
variability as estimated by GCMs were examined for
this report. We found that no firm conclusions can
be drawn about how global warming could affect

     The methods used  to estimate impacts  (for
example, how  forests  might  change) also have
limitations because our scientific understanding of
physiological processes is limited and subject  to
uncertainties. We have no experience with the rapid
wanning of 1.5 to 4.5ฐC projected to occur during
the next century. Many of the effects are estimated
based on knowledge  of the response of systems to
known climate conditions.  We cannot be certain
that  a forest would be able to migrate, how higher
atmospheric concentrations of CO2 would affect
vegetation, whether fish would fina new habitats,
how agricultural pests  would  proliferate, or  how
impacts would combine to create or  reduce stress.
     With some exceptions, we did not generally
examine human responses and adaptations to effects
of climate  change.   The report was  intended  to
examine sensitivities and potential vulnerabilities of
current systems to  climate change.  Many other
changes will  also take place in  the world at the
same time that global climate is changing.  We
cannot anticipate how changing technology, scientific
advances, urban growth, and changing demographics
will affect the world of the next century.  These
changes and  many others  may singularly,  or  hi
combination,  exacerbate or  ameliorate the  impacts
of global climate change on society.

     The results are also inherently limited by our
imaginations. Until a severe event occurs, such  as
the drought of 1988, we fail to recognize the close
links between our society,  the environment, and
climate.  For example, in this report we  did not
analyze the reductions in barge shipments on the
Mississippi River due to  lower river  levels, the

 Executive Summary
 increases in forest fires due to dry conditions, or the
 impacts of disappearing prairie potholes on ducks;
 all these impacts were made vivid during 1988.  The
 drought reminded us  of our vulnerability  as  a
 nation, but it cannot be viewed as a prediction of
 things to come.

     The  findings collectively  suggest  a  world
 different from the world that exists today, although
 there are many uncertainties about specific effects.
 Global  climate  change  could  have  significant
 implications for natural ecosystems; for where and
 how we farm; for the availability of water to irrigate
 crops, produce power, and support shipping; for
 how we live in our cities;  for the wetlands that
 spawn  our fish;  for  the  beaches we  use for
 recreation; and for all levels of government and

     The rate of global warming may be the most
 important factor affecting both natural and managed
 systems. The faster the warming, the harder it will
 be to  adapt.   The ability of natural  ecosystems
 (forests, wetlands, barrier islands, national parks) to
 adapt to a rapidly warming climate is limited.  Rates
 of natural migration and adaptation could be much
 slower than the rate of climate change. Populations
 of many  species and  inhabited  ranges  could
 decrease, and  many  may face  extinction.   The
 ultimate effects could last for centuries and would
 be  virtually   irreversible.     Whether  human
 intervention could mitigate  these effects was not

     Managed systems may show more resilience.
 For example,  although  sea level  rise may  put
 additional stresses on  coastal cities and  although
 changes in temperature and rainfall patterns may
 require new strategies for managing water resources
 and agriculture, we could adapt to changing climate
 relatively  quickly, if  we  have  enough  financial
 resources.     We would   expect   that  basic
 requirements for food  and water  could be met in
 the United States  (as crops  are shifted and water
 management  systems  are  modified),  and  that
 developed areas with high economic value could be
protected against  sea level rise (as bulkheads and
levees  are built).   The total cost of adapting to
global  climate change is beyond  the scope of this
report.   It appears  it could be expensive, but
affordable, fora highly industrialized  country like
 the United States to  adapt managed systems in
 response to gradual global warming.   If  change
 comes more quickly, adaptation by managed systems
 will be more difficult  and expensive.  If it comes
 more slowly, the cost  and difficulty of adaptation
 will be less.

     In  many  cases, the results of our  analysis
 appear to be consistent  across scenarios, because
 either increasing temperatures or higher sea levels
 dominate the  systems that  were  studied.   For
 example, higher temperatures would cause earlier
 snowmelt, a northward migration of forests, and a
 northward shift in crops, and higher sea levels could
 inundate wetlands and low-lying areas.  In other
 cases,  however, only  a  range of values  can  be
 presented because uncertainties in  an important
 variable, such as precipitation, make the direction of
 change highly uncertain.

     The main findings and policy implications of
 this report are presented in national and regional
 chapters. They are summarized in the following
 pages, but the reader is  urged to explore the full
 report to understand the  complete context of these

Natural Systems

     The location and composition of various plants
and animals in the natural environment depend, to
a great extent, on climate.  Trees grow in certain
areas and fish exist in streams and lakes because the
local climate and other conditions are conducive to
reproduction and growth. A  major focus of this
report was to identify what may happen to plants
and animals,  as a result  of climate change  —
whether they would survive in their current locations
or be able to migrate to new habitats, and how soon
these ecosystems could be affected.  The following
descriptions of  impacts  on natural systems are
subject to uncertainties about climate change and
the responses of natural systems to such change.

Natural Systems  May Be Unable to Adapt Quickly
to a Rapid Warming

    If current trends continue, climate may change
too quickly for many natural systems to adapt.  In
the past, plants  and animals adapted to historic

                                                                              Effects of Climate Change
climate changes over many centuries.  For example,
since the last ice age 18,000 years ago, oak trees
migrated northward from the southeastern United
States as the ice sheet receded.   Temperatures
warmed  about 5ฐC (9ฐF) over thousands of years,
but they rose slowly enough for forests to migrate at
the same rate as climate change.  In the future, the
greenhouse effect may lead to similar changes in the
magnitude of warming, but the changes may take
place within a century.  Climate zones may shift
hundreds of  miles  northward, and animals and
especially  plants  may  have  difficulty  migrating
northward that quickly.


     Forests occupy one-third of the land area of
the United States.  Temperature and precipitation
ranges  are  among  the determinants  of forest
distributions.  Forests are also sensitive to soils, light
intensity,  air  pollution,  pests   and  pathogens,
disturbances  such   as  fires   and  wind,  and
management practices.

     Several  approaches were used to  examine
geographic shifts in  forests.   Potential  ranges of
                  forests were estimated for eastern North America
                  using temperature and  precipitation correlations
                  from pollen data.   Changes in composition  and
                  abundance of particular forests were estimated for
                  particular sites in the Great Lakes and Southeast
                  using site-specific  models.  These  regions were
                  chosen to represent a diversity of forest types  and
                  uses. Finally, the ability of trees to migrate to  new
                  habitats was analyzed using shifts in climate zones
                  from GCMs and historic  rates of tree migration.
                  This study focused on several species that are widely
                  dispersed across the northeastern  United States.
                  The direct  effects of CO2, which  could  change
                  water-use  efficiency,  pest interactions,  and  the
                  competitive  balance  among   plants,  were  not
                  modeled, nor were reforestation or the suitability of
                  soils and sunlight considered.  It is not clear  how
                  these results would have been  affected if such
                  factors had been included.

                  The Range of Trees May Be Reduced

                       Figure 4 shows the potential  shifts in forest
                  ranges in response to climate change. The scenarios
                  assume  that climate  change  could  move  the
                  southern boundary  northward  by 600-700  km
                                                                                  Potential Range

                                                                                  Inhabited Range
                  Present Range

             Sugar Maple
                                     Range After 2050: GISS
                                                          Range After 2050: GFDL
                  Present Range
Range After 2050: GISS       Range After 2050: GFDL

                            Scale 0 400Km
          Figure 4.  Shifts in range of hemlock and sugar maple under alternative climate scenarios.

 Executive Summary
 (approximately  400  miles),  while  the  northern
 boundary would move  only  as fast  as the  rate
 ofmigration of forests.  Assuming a migration rate
 of 100  km (60  miles) per century, or  double the
 known historic rate, the inhabited ranges of forests
 could be significantly reduced because the southern
 boundary may  advance  more quickly than  the
 northern  boundary. Even if climate stabilizes, it
 could take centuries for migration to reverse this
 effect.  If climate continues to warm, migration
 would continue to lag behind  shifts in climate zones.
 If elevated CO2 concentrations increase the water-
 use efficiency of tree species and pest infestations
 do not worsen, the declines of the southern ranges
 could be partly alleviated.  Reforestation could help
 speed the migration of forests into new areas.

 Changes in Forest Composition Are Likely

     Climate change may significantly alter forest
 composition and reduce the land area of healthy
 forests.  Higher temperatures may lead to drier soils
 in many parts of the country.  Trees that  need
 wetter soils may die, and their seedlings could have
 difficulty surviving these  conditions.  A  study of
 forests in northern Mississippi and northern Georgia
 indicated  that seedlings  currently in  such areas
would not grow  because of high temperatures and
dry soil conditions. In central Michigan, forests now
dominated by sugar maple and oak may be replaced
by grasslands, with some sparse oak trees surviving.
These analyses did not consider the introduction of
species  from areas south of these  regions.   In
northern Minnesota, the mixed boreal and northern
hardwood  forests could become  entirely  northern
hardwoods. Some areas might experience  a decline
in productivity, while  others (currently saturated
soils) might have an  increase.   The process  of
changes in species  composition would most likely
continue for centuries.   Other studies of the
potential effects of climate change in forests imply
northward  shifts in ranges and significant changes in
composition,   although  specific   results   vary
depending  on sites and scenarios  used.

Changes May Begin in 30 to 80 Years

     Forest change may be visible in a few decades
from now.   This would involve  a faster rate  of
mortality among mature  trees and  a decline  in
seedlings and growth of new species. The studies  of
forests in the Southeast and  Great Lakes indicate
that these forests could begin to die back in 30 to 80
years.   Figure, 5 displays possible reductions  in
balsam fir trees in northern Minnesota and forests
in Mississippi in response to two different scenarios


f 140
w 120

i 10ฐ
S 80
8 80

— — — GtSSA
— "- '"s. — "1^.
i I I I I \ 	 !-*• 	
a_ 4000
5 2000



_^ — i.^
""" *^^^***^^
VN^====^- 	 — 	 "
x "x.
1 > " 	 -..o^-t. 1

1980 2000 2020 2040 2060 1980 2000 2020 2040 2060 2080
* Assumes constant exponential
growth in emissions
" Assumes constant arithmetic
growth In emissions
                         Figure 5.  Forest declines due to temperature increases.

                                                                             Effects of Climate Change
of wanning.  At the same time in Minnesota, for
example, sugar maple could become more abundant.
These forests appear to be very sensitive to small
changes  in climate,  because  dieback starts  to
become noticeable after an approximate 1 to 1.5ฐC
warming. Once this process starts, major dieback
may occur rapidly.  The timing of a decline is
sensitive to the rate of climate  change; a warming
slower than that assumed in the scenarios would
delay the dieback.

Other Factors Will Influence Forest Health

     The health of forests will not be determined by
climate  change alone.  The drier soils expected to
accompany climate change  could  lead  to  more
frequent fires, warmer climates  may cause changes
hi forest pests  and pathogens,  and  changes  in air
pollution levels could  reduce  the  resilience  of
forests.  Continued depletion of stratospheric ozone
would also further stress  forests.  None of these
outcomes was considered by the forest studies in
this  report,  although  they could  speed  forest


     Biological diversity  can be defined as  the
variety of species in ecosystems, and  the genetic
variability within each species  and the variety of
ecosystems around the world. Over 400 species of
mammals,  460  species of reptiles, 660 species of
freshwater fishes, and  tens  of  thousands  of
invertebrate species can be found hi this country, in
addition to some 22,000 plant species.  About 650
species of birds reside in or pass through the United
States annually.  Biological diversity is needed to
provide food, medicine, shelter, and other important

     This report examined the  impacts of climate
change  on specific plants and animals by using
climate  change scenarios and models of particular
species  or systems within  a region.  Analyses have
been performed for impacts on  finfish and shellfish
in the Apalachicola Bay in the  Florida Panhandle,
fish hi the Great Lakes, and marine species in San
Francisco Bay.  Additional information on potential
impacts  on biodiversity was gathered from  the
published literature.

Extinction of Species Could Increase

     Historic climate changes, such as the ice ages,
have  led to  extinction of many species.  More
recently, human activities, such as deforestation,
have  greatly  accelerated the  rate  of species
extinction.  The faster rate of climate warming due
to the greenhouse effect, absent an active program
to preserve species, would most likely lead to  an
even greater loss  of species.  The uncertainties
surrounding the rate of warming, the response of
individual species, and interspecies dynamics make
it difficult to assess the probable impacts, although
natural  ecosystems are likely to be  destabilized in
unpredictable ways.

    As with trees, other plants and animals may
have difficulty  migrating at  the  same  rate  as a
rapidly  changing climate, and many species may
become extinct or their populations maybe reduced.
The presence of urban areas, agricultural lands, and
roads would restrict  habitats and block  many
migratory pathways.  These obstacles may make it
harder  for plants  and  wildlife  to  survive future
climate  changes. On the other hand, some species
may benefit from  climate change  as a result  of
increases in habitat size or reduction hi population
of competitors.  The extent  to which society can
mitigate negative impacts through such  efforts as
habitat restoration is not clear.

Impacts on Fisheries Would Vary

    Freshwater fish populations may grow in some
areas and  decline  in others.  Fish hi such large
water bodies as the Great Lakes may grow faster
and may be  able  to migrate  to  new habitats.
Increased amounts of plankton could provide more
forage for fish.  However, higher temperatures may
lead to more aquatic growth, such as algal blooms,
and decreased mixing of lakes Gonger stratification),
which would deplete oxygen levels hi shallow areas
of the Great Lakes, for example Lake Erie, and
make them less habitable for fish.   Fish hi small
lakes  and  streams  may  be unable  to  escape
temperatures beyond  their  tolerances,  or  then-
habitats may simply disappear.

    Warmer temperatures could also exceed the
thermal  tolerance  of many marine finfish and
shellfish hi some southern locations, although some
marine  species  could benefit.  The full impacts  on
marine species are not known at this tune. The loss
of  coastal wetlands  could  further  reduce fish
populations,  especially  shellfish.    And  while
increased salinity  hi  estuaries could reduce the
abundance of freshwater species, it could increase
the presence of marine species. Whether finfish and

 Executive Summary
 shellfish  could migrate  to new  areas  and  the
 effectiveness of restocking were not studied.

 Effects on Migratory Birds Would  Depend on
 Impacts on Habitats

     Migratory birds are likely to experience mixed
 effects from  climate change, with some  arctic-
 nesting  herbivores  benefiting,  and  continental
 nesters and shorebirds suffering.  Some  winter
 habitats could  experience  increased  productivity.
 On the other hand, the loss of wintering grounds,
 which may result from sea level rise and changing
 climate, could harm many species, as would the loss
 of inland prairie potholes resulting from potentially
 increased midcontinental dryness.

 Sea Level Rise

     A rise in sea level is one of the more probable
 impacts   of  climate   change.     Higher  global
 temperatures  will expand ocean  water and melt
 some mountain glaciers, and may eventually cause
 polar ice sheets to discharge ice.   Over the last
 century, global sea level has risen 10 to 15 cm (4 to
 6 inches), and along the U.S. coastline, relative sea
 level  rise  (which includes land subsidence) has
 averaged about 30 cm (1 foot). Published estimates
 of sea level rise due to global warming generally
 range from 0.5 to 2.0 meters (1.5 to 7 feet) by 2100.
 Sea level rise could be greater than or less than this
 range because uncertainties exist regarding the rate
 of atmospheric wanning, glacial processes, oceanic
 uptake of heat, precipitation in  polar areas,  and
 other variables.

     The studies estimate the potential nationwide
 loss of wetlands, and the cost of defending currently
 developed areas from  a  rising  sea, for three
 scenarios  (50,  100, and 200 cm) of sea level rise by
 the year  2100.   The scenarios  are based  on
 quantitative estimates  of  sea  level rise, but no
 probabilities have been attributed to them.  Wetland
 loss estimates were based  on remote-sensing data
 and topographic maps for  a sample of sites along
the U.S. coast.  The cost of holding back the sea
was based on (1) the quantity of sand necessary to
 elevate beaches and coastal barrier islands  as sea
level  rises; (2)  rebuilding  roads  and elevating
structures;  and (3)  constructing  levees  and
bulkheads  to  protect developed  lowlands  along
sheltered waters.
 Protectins Developed Areas May Be Expensive

     Given the high property values of developed
 coastlines  in the United  States, it is likely that
 measures would be taken to hold back the sea along
 most  developed shores.   Preliminary estimates
 suggest that the cumulative capital cost (including
 response to current sea level rise) of protecting
 currently developed  areas would be $73 to $111
 billion (in 1988 dollars) through 2100 for a 1-meter
 global rise (compared with  $4  to $6 billion to
 protect developed areas from current trends in sea
 level rise).  A 1-meter sea level rise would lead to
 a cumulative inundation of 7,000 square  miles of
 dryland —  an area the  size of Massachusetts (see
 Table  1). If the oceans continue to rise at current
 rates,  approximately 3,000 square miles of dryland
 would be lost.

 Most Coastal Wetlands Would Be Lost

     Historically, wetlands have  kept pace with a
 slow rate of sea level rise.  However, in the future,
 sea  level will  probably  rise too  fast for  some
 marshes and swamps to keep pace.  Although some
 wetlands can survive by migrating  inland, a study on
 coastal wetlands  estimated that for a 1-meter rise,
 26 to  66%  of wetlands would  be lost, even if
 wetland migration were not blocked. A majority of
 these losses would be in the South (see Table 2).
 Efforts to  protect  coastal development  would
 increase wetland losses,  because  bulkheads and
 levees  would prevent new wetlands from forming
 inland. If all shorelines are protected, 50 to 82% of
 wetlands would be  lost.  The different amounts of
 dryland lost for different regions and scenarios  are
 shown in Figure 6.

     The loss of wetland area would have adverse
 ecological impacts, with the ability of ecosystems to
 survive a rising sea level depending greatly on how
 shorelines are managed. For many fish and shellfish
 species,  the fraction of  shorelines along  which
wetlands can be found is more important than  the
total area of wetlands. This fraction could remain
at approximately present levels if  people do not
erect additional bulkheads and levees. In Louisiana,
with 40% of U.S. coastal wetlands, large areas of
wetlands are already being converted to open water
as a result of natural subsidence and the effects of
human activities,  and most could be lost by 2030 if
current trends continue.

                                                                            Effects of Climate Change
                            Table 1. Nationwide Impacts of Sea Level Rise
                                                         Sea Level Rise by 2100
If Densely Developed Areas
 Are Protected

Shore protection costs
 (billions of 1986 dollars)
 Dryland lost (mi2)
 Wetlands lost (%)

If No Shores Are Protected

 Dryland lost (mi2)
 Wetlands lost (%)

If All Shores Are Protected

 Wetlands lost (%)
    N.C.      3,300-7,300    5,100-10,300  8,200-15,400
    N.C.         17-43         26-66        29-76
N.C. = Not calculated.
*Baseline assumes current global sea level rise trend of 12 cm per century. Given coastal subsidence trends, this
 implies about a 1-foot rise in relative sea level along most of the U.S. coast.
Source: Assembled by Titus and Greene.
                 Table 2. Loss of Coastal Wetlands from a One-Meter Rise in Sea Level

South Atlantic
South and West
Other Gulf
United States

area (mi2)

(% loss)

(% loss)


(% loss)

77 :
75 :
• b
 aLouisiana projections do not consider potential benefits of restoring flow of sediment and freshwater.
 bPotential gain in wetland acreage not shown because principal author suggested that no confidence could be
 attributed to those estimates. West Coast sites constituted less than 0.5% of wetlands in study sample.
 Source:  Adapted from Park et al.

  Executive Summary
                                                                             SEA LEVEL RISE
                                              Louisiana  other Gull














                                     Figure 6. Dryland loss by 2100.
 Estuaries May Enlarge and Become More Saline

     Although future riverflows into estuaries are
 uncertain, a rise in sea level would increase the size
 and salinity  of estuaries and would increase the
 salinity of coastal aquifers.  For example, sea level
 rise may result  in  a more saline  and  enlarged
 Sacramento-San  Joaquin Delta, and Miami,  New
 York, and other coastal communities would have to
 step up current efforts to combat salinity increases
 in surface water and groundwater supplies.


     The  temperate climate and rich soils in the
United  States, especially in the  Midwest,  have
helped  make  this country the world's  leading
agricultural   producer.    Agriculture,  a  critical
component of the U.S. economy, contributed 17.5%
                         of the gross national product in 1985, with farm
                         assets  totaling $771 billion.   Crop production  is
                         sensitive to climate, soils, management methods,
                         and many other factors.  During the Dust Bowl
                         years of the 1930s, wheat and corn yields dropped
                         by up to 50%, and during the drought of 1988, corn
                         yields declined about 40%.

                             The  agricultural  analyses  in  this  report
                         examined potential impacts  on  crop yields and
                         productivity from changes hi climate and  direct
                         effects  of CO2.  (Higher CO2 concentrations may
                         increase  plant growth and  water-use efficiency.)
                         The studies used high estimates of the beneficial
                         effects  of CO2 on crops.  Changes in dryland and
                         irrigated corn, wheat, and soybean yields and  hi
                         irrigation demand were estimated for the Southeast,
                         Great Plains, and Great Lakes regions using widely
                         validated crop growth models.  Crop yield changes

                                                                             Effects of Climate Change
were  estimated  for  California  using a  simple
agroclimatic index. The studies  did not examine
effects on yields of introduction of crops, such as
citrus,  into  new   areas;  changes   in  weed
growthcaused by higher CO, concentrations; or new
technologies, such as biotechnology. Some of these
changes could enhance the ability of agriculture to
adapt to global warming.

     The  estimated yield changes  from the four
regional crop modeling studies and runoff changes
from the  GCMs were  used  in a  nationwide
agricultural economic model to estimate regional
and national changes  in crop production, land use,
and demand for  irrigation.  The economic model
did not consider  the introduction of new crops,
changes in government  policies on  agriculture,
change in demand for water  for  nonagricultural
uses, and  global  agricultural changes.    Both a
modeling study and a literature review were used to
estimate changes in  plant-pest interactions.  An
agricultural runoff and leaching model was used to
estimate potential changes in water quality in the
Great  Plains.    Some  farm-level  adjustments,
including the effects of changed planting dates and
use of different varieties, were  investigated  in
various  studies,   and  the   potential  national
implications on  livestock  were analyzed  using
modeling studies and a literature review.

Yields Could Be Reduced. Although the Combined
Effects of Climate and CO2 Would Depend on the
Severity of Climate Change

      In most regions  of the country, climate change
 alone could reduce dryland yields of  corn, wheat,
 and soybeans, with site-to-site losses ranging from
 negligible  amounts to 80%.  These decreases would
 be primarily the result  of higher temperatures,
 which would  shorten a  crop's life cycle.  In very
 northern areas, such as Minnesota, dryland yields of
 corn  and  soybeans could increase as  warmer
 temperatures extend the frost-free growing season.
 The combined   effects  of climate  change and
 increased  CO2 may result in net increases in yields
 in some cases, especially in northern areas or in
 areas where rainfall is abundant.  In southern areas,
 however, where heat stress is already a problem,
 and in areas where rainfall is reduced, crop yields
 could decline.

 Productivity Mav Shift Northward

      Under all of the scenarios (with and  without
 the  direct effects of increased CO2), the  relative
productivity of northern areas for the crops studied
was estimated to rise in comparison with that of
southern areas.  In response to the shift in relative
yields, grain  crop acreage  in Appalachia,  the
Southeast,  and the southern  Great Plains  could
decrease, and acreage in the northern Great Lakes
States, the northern Great Plains, and the Pacific
Northwest could increase (see Figure 7). A change
in agriculture would affect not only the livelihood of
farmers  but  also agricultural infrastructure  and
other support services.  The sustainability of crop
production in  northern areas  was not  studied.
Changes in foreign demand for U.S. crops, which
would likely be altered as a result of global warming
and could  significantly alter the magnitude  of the
results, were not considered in this analysis.

The National Supply  of Agricultural Commodities
Mav  Be Sufficient to Meet Domestic  Needs. But
Exports Mav Be Reduced

     Even under the more extreme climate change
scenarios,   the  production  capacity  of   U.S.
agriculture was estimated to be adequate to meet
domestic needs. Only small to moderate economic
losses  were  estimated  when climate  change
scenarios  were modeled  without the beneficial
effects of CO2 on crop yields.  When the combined
effects of climate and CO2 were considered, results
were positive with a relatively wetter climate change
scenario and negative with the hotter, drier climate
change scenario. Thus, the severity of the economic
consequences could depend on the type of climate
change that occurs and  the  ability of the direct
effects of CO2 to enhance yields. A decline in crop
production would reduce exports, which could have
serious implications for food-importing nations. If
climate change is severe, continued and substantial
improvements in crop yields  would be needed to
 fully offset  the negative  effects.   Technological
 improvements, such as improved crop varieties from
 bioengineering, could be helpful in keeping up with
 climate change. These results could be affected by
 global changes  in  agriculture, which  were not
 considered in the analysis.

 Farmers Would  Likely  Change  Many  of Their

      Farm practices would likely change in response
 to different climate conditions. Most significantly,
 in many regions, the demand for irrigation is  likely
 to increase as a result of higher temperatures.  If
 national productivity declines, crop prices may rise,
 making irrigation more economical and increasing

Executive Summary
                  r   i GISS
                  II GFDL
                      GISS +Dlrect Effects of CO2
                      GFDL+Dlrect Effects of CO2
                          Figure 7.  Percent change in regional agricultural acreage.
                   I   I GISS
                   II GFDL
                       GISS +Dlrect Effects of COg
                       GFDL+Dlrect Effects of CO2
Southern Plains
                     Figure 8.  Change in regional irrigation acreage (100,000 of acres).

                                                                            Effects of Climate Change
the use of it (see Figure 8). Irrigation equipment
may be installed in many areas that are currently
dryland farms, and farmers already irrigating may
extract more water from surface and groundwater
sources.  Changes in competing demands for water
by municipal and industrial users, which could raise
the cost of irrigation, were not considered. Farmers
may also switch to more heat- and drought-resistant
crop varieties, plant two  crops during a growing
season, and plant and harvest earlier.  Whether
these adjustments would  compensate for climate
change depends on a number  of factors, including
the severity of the climate change.  Under extreme
climate change conditions, some farms could be

Ranees  of   Agricultural  Pests   May  Extend

     Warmer  temperatures  may  result  in  the
northward extension of the range  of diseases and
pests that  now afflict livestock in the South, and
could  make conditions  more favorable  for the
introduction of new livestock diseases  into the
southern United  States.   This  extension  could
reduce crop yields and affect livestock.

Shifts in Agriculture Mav Harm the Environment in
Some Areas

     Expansion of irrigation and shifts in regional
production patterns imply more competition for
water resources, greater potential for surface water
and groundwater pollution, loss of some wildlife
habitats, and increased soil erosion. A northward
migration of agriculture would increase the use of
irrigation  and  fertilizers on sandy  soils,  thus
endangering the quality of underlying groundwater.
Chemical pesticide usage may change to control
different crop and livestock pests.   Thus, climate
change could exacerbate environmental pollution
and increase resource use from agriculture in some

Water Resources

     The United States is endowed with a bountiful
 supply of water, but the water is not always in the
 right place at the right time or of the right quality.
 In some regions,  such as the  Great Basin and the
 Colorado River Basin, the gap between demand and
 supply of water  is narrow.  In these basins, such
 offstream  uses  as   irrigation   and  domestic
 consumption often conflict with each other and with
other needs, such as maintaining flow to preserve
environmental quality.

    Although global  precipitation  is  likely  to
increase, it  is not known how regional rainfall
patterns will be affected. Some regions may have
more  rainfall,  while  others  may  have  less.
Furthermore, higher temperatures would most likely
increase evaporation. These  changes would likely
create new stresses  for many water management

     To discuss  the  potential impacts of climate
change on water resources, this report studied water
resources in California, the Great Lakes, and the
Southeast, estimated the demand for  irrigation in
the Great Plains, and drew on information from the
literature.  These studies  focused on changes in
runoff and,  for California  and  the Southeast,
considered management responses.   The  studies
examined the water management systems as they
are currently configured and  did not examine new
construction.  Among other factors not considered
were changes in demand for water resources (which
would most likely lead to greater  changes in water
management systems) and changes in vegetation due
to climate change and increased CO^, which could
affect runoff.  The studies did not estimate impacts
on groundwater.

The Direction of Change  in Some Water Bodies
Can Be Estimated, but Total Impacts in the United
States Cannot Be Determined

     Results of hydrology studies  indicate that it is
possible in some regions to identify the direction of
change in water supplies and quaiity due to global
wanning.   For example, in  California, higher
temperatures would  reduce the snowpack and cause
earlier melting.   Earlier  runoff from  mountains
could increase winter flooding and reduce deliveries
to users. In the Great Lakes, reduced  snowpack
combined with potentially higher evaporation could
lower lake levels (although certain combinations of
conditions could lead to higher levels).   In other
areas, such as the South, little snowcover currently
exists, so riverflow and lake levels depend more on
rainfall patterns. Without  better rainfall estimates,
we  cannot determine whether riverflow and  lake
levels in the South would rise or fall.

Water Quality in Many Basins  Could Change

      Changes in water supply could  significantly
 affect water  quality.  Where riverflow  and lake

  Executive Summary
  levels decline, such as in the Great Lakes, there
  would be less water to dilute pollutants.  On the
  other hand,  where there  is  more  water, water
  quality may improve.  Higher temperatures  may
  enhance  thermal stratification in some  lakes and
  increase algal production, degrading water quality.
  Changes  in runoff and leaching from farms and
  potential increases  in the use  of irrigation  for
  agriculture could affect surface and groundwater
  quality in many areas.

  Water Use Conflicts May Increase

      In some regions, decreased water availability
  and increased  demand for  water,  such as for
  irrigation and powerplant cooling, may intensify
  conflicts among offstream uses.  Conflicts between
  these  offstream uses and  instream uses such as
  flood  control and wildlife  habitat  also  may  be

  Electricity Demand

      The  demand for  electricity is influenced  by
  economic growth, by  changes in industrial and
  residential/commercial technologies, and by climate.
  The principal  climate-sensitive electricity end uses
  are space heating and cooling and,  to a lesser
 degree, water heating and refrigeration. These uses
 of electricity may account for up to a third of total
 sales for some utilities and may contribute an even
 larger portion of seasonal and daily peak demands.

      This report analyzed potential changes hi the
 national demand for electricity in 2010 and 2055,
 using the relationship between demand and climate
 for several  major  utility systems.   The study
 estimated  changes in demand due to nonclimate
 factors, such as increases in population and GNP.
 The impacts of climate change are expressed as an
 increase over non-climate growth, and results are
 given on nationwide and regional bases. The study
 did not  consider changes  in  technology and
 improvements in energy efficiency; the impacts of
 higher temperatures on the demand for natural gas
 and oil for home heating, which will  most likely
 decrease;  changes in electricity supplies,  such  as
 hydropower;  or changes  in demand for electricity
 for such uses as irrigation.

National Electricity Demand Would Rise

     Global warming would increase annual demand
for  electricity  and  total  generating  capacity
  requirements in the United States. The demand for
  electricity for summer cooling would increase, and
  the demand for electricity for winter heating would
  decrease.  Annual electricity generation in 2055 was
  estimated under the transient scenarios to be 4 to
  6% greater than without climate change.  The
  annual costs of meeting the increase due to global
  warming,  assuming no change in technology or
  efficiency, was estimated to be $33-$73 billion  (in
  1986 dollars).  These results differ on  a regional
  basis and are shown in Figure 9. States along the
  northern tier  of the United States could have net
  reductions in annual demand of up to 5%, because
  decreased heating demand would exceed increased
  demand for air-conditioning. In the South, where
  heating needs are already low, net demand was
  estimated to rise by 7 to 11% by 2055.

      Generating   capacity  requirements  are
  determined largely by peak demand, which occurs
  in the summer in all but the  far northern areas of
  the country.  By 2010, generating requirements to
  meet increased demand could rise  by 25 to  55
  gigawatts (GW), or by 9 to 19% above new capacity
  requirements,  assuming no  climate  change.   By
  2055, generating requirements could be up by 200 to
  400 GW, or 14 to 23% above non-climate-related
 growth.  The cumulative cost of such an increase in
 capacity, assuming no  change  in technology  or
 improvements hi energy efficiency, was estimated to
 be between $175 and $325 billion (hi 1988 dollars).
 The  South would  have a greater need than the
 North for additional capacity, as shown hi Figure 10.
 Increases in capacity requirements could range from
 0 to 10% in the North, to 20  to 30% in the South
 and Southwest.  U.S. emissions of such greenhouse
 gases  as  CO2  could  increase  substantially  if
 additional powerplants are built to meet  these
 capacity requirements, especially if they burn coal.
 Improvement in the efficiency of energy production
 and use would reduce these emissions.

 Air Quality

    Air   pollution  caused  by  emissions  from
 industrial and transportation sources is a subject of
 concern hi the  United States.  Over the last two
 decades,  considerable progress has been made hi
 improving air quality by reducing emissions.  Yet
 high temperatures hi the summer of 1988 helped
raise tropospheric ozone levels to all-tune highs hi
many U.S. cities.  But air quality is also directly
affected  by other weather  variables, such   as

                                                                           Effects of Climate Change
   Figure 9.  Changes in electricity generation by state, induced by climate change scenarios in 2055.
Figure 10.  Changes in electricity capacity additions by state, induced by climate change scenarios in 2055.

 Executive Summary
 windspeed  and  direction,  precipitation  patterns,
 cloud cover, atmospheric water vapor, and global
 circulation patterns.

     A literature review of the relationship between
 climate and air  pollution was conducted for this
 report. In  addition, air quality models were used
 for a preliminary analysis of the changes  in ozone
 levels in several regions. The latter analysis did not
 consider reduction in emissions of air pollutants due
 to enforcement of the Clean Air Act.
Climate Changes Could Increase Air  Pollution.
Especially Smog
     A rise in global temperatures would increase
manmade and  natural emissions of hydrocarbons
and manmade  emissions of sulfur  and  nitrogen
oxides over what  they would be  without climate
change.  Natural  emissions of sulfur would also
change, but the  direction is uncertain. Although the
potential magnitude of the impacts of the increased
emissions on  air   quality is uncertain, higher
 temperatures would speed the reaction rates among
 chemicals in the atmosphere, causing higher ozone
 pollution in many urban areas than would occur
 otherwise.  They would also increase the length of
 the summer season, usually a time  of  high air
 pollution levels. As shown in Figure 11, preliminary
 analyses of a 4ฐC temperature increase in the San
 Francisco  Bay  area  (with no  changes in other
 meteorologic variables, such as mixing  heights),
 assuming no  change in  emissions  from  current
 levels, suggest that maximum ozone concentrations
 would increase by 20%, and that the area exceeding
 the National Ambient Air Quality Standards would
 almost double.  Studies of the Southeast also show
 expansion of the areas violating the standards, but
 they show smaller changes in levels.  Although the
 impacts of higher temperatures on acid rain were
 not analyzed, it is likely that sulfur and nitrogen
would  oxidize   more    rapidly  under   higher
temperatures.    The  ultimate  effect  on  acid
deposition is difficult to assess because changes in
clouds, winds,  and  precipitation  patterns   are
                                                       Exceeds Standard
                               <6   >6    >8
                    Base Case
                  August 6, 1981
      Climate Sensitivity Scenario No. 1
          4ฐC Temperature Increase
                      Figure 11. Changes in maximum daily ozone concentrations.

                                                                             Effects of Climate Change
Health Effects

     Human illness  and  mortality are linked in
many ways to weather patterns.  Weather affects
contagious  diseases  such   as   influenza   and
pneumonia, and allergic diseases such as asthma.
Mortality rates, particularly for the elderly and the
very ill, are influenced by the frequency and severity
of extreme temperatures. The life cycles of disease-
carrying insects, such as mosquitoes and ticks, are
affected by changes in temperature and rainfall, as
well as by habitat, which is itself sensitive to climate.
Finally, increased air pollution, which is related to
weather patterns,  can heighten the  incidence and
severity of such respiratory diseases as emphysema
and asthma.

     Both expert judgment and modeling were used
to study the potential impacts of climate change on
human health.  A literature review and  workshop
were  conducted to identify potential changes  in
vector-borne  diseases caused by ticks, fleas, and
mosquitoes (such as dengue and malaria).  Models
were used to estimate potential geographic shifts in
the prevalence of Rocky Mountain spotted fever and
 malaria. Potential changes in mortality from heat
 and  cold  stress were  quantitatively estimated,
 although such estimates did not consider changes in
 air  pollution  levels.  The total impacts of climate
 change on human health are difficult to assess; these
 analyses looked at  a limited number of potential
 effects and are only indicative of possible changes in
 mortality and morbidity.

 Summer Mortality  Could Increase. While Winter
 Mortality Could Decrease

      Global  warming  may  lead to changes  in
 morbidity and increases in mortality, particularly for
 the  elderly during the  summer.   Morbidity and
 mortality may decrease because of milder winters,
 although net  mortality  may increase.    If the
 frequency or intensity of climate extremes increases,
 mortality is likely to rise. If people acclimatize by
 using air-conditioning, changing  their  workplace
 habits, and altering the construction of their homes
 and cities, the impact on summer mortality rates
 may be substantially reduced.

 Regional Morbidity Patterns Could Change

       Changes in climate as well as in habitat may
  alter  the regional  prevalence  of  vector-borne
  diseases.  For example,  some forests may become
grasslands, thereby  modifying the  incidence  of
vector-borne diseases. Changes in summer rainfall
could alter the amount of ragweed growing  on
cultivated land, and changes in humidity may affect
the incidence  and severity of skin infections and
infestations  such  as  ringworm,  candidiasis, and
scabies. Increases in the persistence and level of air
pollution  episodes associated with climate change
would have other harmful health effects.

Urban Infrastructure

     The  value of municipal infrastructure  in  the
United States,  excluding buildings  and  electric
power production, probably approaches one trillion
dollars.  The  majority of the nation's investments
are in water  supply,  wastewater transport  and
treatment facilities,  drainage,  roadways,  airports,
and mass transit facilities.  Like the regions studied
for this report, urban areas would feel a variety of
impacts from climate change. This report examined
the  potential impacts  of  climate  change  on
Cleveland, New York City, and Miami. These areas
encompass  a  diversity  of climates  and uses of
natural resources.

     Much  of  the  current  inventory  in urban
 infrastructure will most likely turn over in the next
 35 to 50 years.  A warmer global climate would
 require changes in the capital investment patterns of
 cities for water  supplies,  peak electric generating
 capacity, and storm sewer  capacity.  Urbanized
 coastal areas might have to invest additional billions
 of  dollars  into coastal protection  to  defend
 developed areas from a rising sea.  In Miami, for
 example, this could imply an increase of 1 to 2% in
 the city's capital spending over the next 100 years.
 Generally,  northern cities such  as Cleveland  may
 fare better, since reductions in  the operating and
 maintenance costs  associated with heating public
 buildings, snow removal, and road  maintenance
 should offset increasing costs for air-conditioning
 and port dredging (see Table 3).

      Studying the national impacts of climate change
  may disguise  important differences  in regional
  effects across  the country.   Shifting demands for
  economic and natural resources may cause stresses
  that  cannot  be   seen  at  a  national  level.
  Furthermore, changes in one system, such as water
  supply, may affect other systems such as irrigation

  Executive Summary
  Table 3. Estimated Impacts of Doubled CO2 Scenarios on Cleveland's Annual Infrastructure Costs (millions of
          1987 dollars)
                     Cost category
             operating costs


                 Snow and ice control

                 Frost damage to roads

                 Road maintenance

                 Road reconstruction

                 Mass transit

                 River dredging

                 Water supply

                 Stormwater system







         summer increase offsets
         winter savings

         less than $0.5



         -1.6 to +1.1
 Source: Walker et al.
 for agriculture.  These combined effects  may be
 most evident on a regional scale.  The designs of
 the regional studies on agriculture, forests, and
 electricity were described above.

     The studies discussed below considered only
 some  of the potential  regional impacts.   Many
 potential impacts were not analyzed — for example,
 demographic shifts into or out of the Southeast,
 recreational  impacts in the Great Lakes, direct
 effects on such aquifers as the Ogallala in the Great
 Plains,  and impacts  on many specialty crops in
 California   In  addition, current  GCMs  often
 disagree  significantly about  simulated regional
 changes, particularly about such key variables as
precipitation. Their spatial resolution is roughly of
the same  size  as the  regions  of concern; for
example,  there  are  two  simulation points in
 California.  The discussion that follows should not
 be viewed as comprehensive, but rather as providing
 examples of important issues for each region.


     California contains a highly managed water
 resource  system and one of the most productive
 agricultural regions  in the world.    The  state
 produces  14% of the  nation's cash receipts  for
 agriculture. California's water resources are poorly
 distributed in relation to its needs. Precipitation is
 abundant  in the north, with the highest levels in the
winter,  while  water is  needed in the south for
agriculture and domestic consumption. The Central
Valley Project (CVP)  and State Water Project
(SWP) were built basically to capture runoff from
the north  and deliver it to uses in the south. These

                                                                                     Effects of Climate Change
projects also provide flood protection, hydroelectric
power, and freshwater flows to repel salinity (known
as carriage water) in the Sacramento-San Joaquin
River Delta.    Islands  in the  delta  are  highly
productive farmlands and are protected by levees.

      The  California  case study focused on  the
Central Valley. First, changes in runoff in the valley
were estimated.   These  results were then used to
estimate changes in deliveries from the CVP and
SWP and in agricultural water use.  These  results
were combined  with  sea level rise estimates and
were used to  model how the salinity and shape of
the San Francisco Bay estuary may change and how
the demand  for carriage water may be affected.
The estimated changes in salinity and sea level rise
were used to  examine impacts on the ecology of the
bay. Yield changes for a number of crops grown in
the state were estimated, as were changes in ozone
levels in central California and changes in electricity
demand (see Figure 12).

California's  Water  Management  System  Would
Have to Be Modified

     Warmer  temperatures  would change  the
seasonalhy  of   runoff   from   the   mountains
surrounding  the Central Valley. Runoff would be
                      TEMPERATURE SCENARIOS

                         WINTER SPRING  SUMMER  FALL
                      PRECIPITATION SCENARIOS
                          2xCO, lesslxCO,
                                                                      Water Resources
                                                                      Regional warming could cause:
                                                                      • higher winter, lower summer runoff
                                                                      • decreased deliveries from Central
                                                                       Valley Project and State Water Project
                                                                      • decreased water quality in subalpine
         Wetlands and Fisheries
         Sea level rise could cause:
         • gradual inundation of wetlands
         • increased salinity in and size of
          San Francisco Bay
         • shift from brackish and freshwater
          species to marine species.
         Increases in temperature and CO2
         concentrations could cause:
         • variable crop responses
         • a northward shift in agricultural
         • increased irrigation demand
           resulting in groundwater extraction
           and decreased water quality
                                                                      Air Quality
                                                                      Higher temperatures would increase
                                                                      ambient ozone levels in central California
                                                                      Higher temperatures could increase
                                                                      electricity demand
                         WINTER SPRING SUMMER  FALL
                                              Figure 12.  California.

  Executive Summary
  higher  in  the  winter months as a  result of less
  snowpack  and more precipitation in the form of
  rain. Consequently, runoff would be lower in the
  late spring and summer.  Under these conditions,
  the current reservoir system in the Central Valley
  would not have the capacity to provide adequate
  flood protection  in  the winter and  store enough
  water to meet deliveries in  the  summer.  Thus,
  much of the earlier winter runoff would have to be
  released. This would leave less water in the system
  for late spring and summer deliveries, when runoff
  would be lower.  Under the three GCM scenarios,
  annual  water  deliveries   from  the SWP  were
  estimated to decrease by 200,000 to  400,000 acre-
  feet (7 to 16% of supply).  In contrast, the increase
  in statewide demand for water from the SWP due to
  non-climate factors, such as population growth, may
  total  1.4  million acre-feet by  2010.   Reduced
  snowpack and earlier runoff could occur throughout
  the West, exacerbating water management problems
 in a region that is currently short of water.

 Climate  Change  Is Likely  to  Increase  Water

      On the whole, California's water demand could
 increase with a warmer climate.  Twice  as much
 carriage water  may  be needed to  repel higher
 salinity levels resulting from a 1-meter  sea level rise.
 In addition, consumptive uses may also increase.
 Irrigation, which may come from groundwater, may
 increase in some parts  of the  state.    If new
 powerplants are built,  they will need water for
 cooling,  which  could come from surface water
 supplies, depending on the location.   Although it
 was not  studied, municipal demand for water may
 also rise.

 Sea  Level  Rise  Would  Affect  the  Size and
 Environment of San Francisco Bay

     A  sea  level rise  would increase  the  salt
 concentrations of San Francisco Bay. It is estimated
 that a 1-meter rise could cause the salt front in the
 Sacramento-San Joaquin River Delta to  migrate
 upstream 4 to 10 km (2.5 to 6 miles). Sea level rise
 would also increase the difficulty of maintaining the
 Sacramento-San Joaquin Delta islands.  If the levees
 around  the  delta  islands  were strengthened and
 raised, a 1-meter rise could increase the volume of
 the San Francisco Bay estuary by 15%  and the area
by 30%.  If the levees were not maintained and the
islands were flooded, there would be a doubling and
tripling, respectively, of the volume and area of the
  bay.  As a result of these changes, some wetlands
  would be lost, marine aquatic species would become
  relatively more abundant,  and freshwater species
  would decline.

  Climate Change Could  Degrade  Air Quality in

      Air quality is currently a major concern in
  California.   The  area  of  central California in
  violation of ozone  quality standards could increase
  as a  result  of  higher  temperatures.  Under  one
  climate scenario,  with a  4ฐC rise and  current
  emission levels, the maximum size of the area with
  ozone levels in  excess of the EPA standard of 0.12
  ppm could double.  This scenario assumed that such
  climate variables as  windspeed and mixing height
  (the volume of  air  in which pollutants are diluted)
  would not change.

  Great Lakes

     The Great Lakes contain 18% of the world's
 supply and 95% of the U.S.  supply  of  surface
 freshwater, and  they are an important source of
 commerce and recreation for the region. In recent
 years,  reductions  in  pollutant   loadings  have
 significantly unproved the  quality  of  such water
 bodies as Lake Erie.  The  Great Lakes States
 produce 59% of the country's corn  and 40% of its
 soybeans,  and   their  forests   have   important
 commercial, recreational,  and conservation uses.

     Models were  used to  estimate the potential
 impacts of climate  change on lake levels and  ice
 cover.  Results  from these studies  were used to
 analyze  impacts on navigation and  shorelines.
 Changes in the  thermal structure of the Central
 Basin  of Lake Erie and southern Lake Michigan
 were estimated.  Output  from these studies was
 used along with scenario  temperatures to analyze
 potential impacts on fishes in the lakes.  Changes in
 crop yields were estimated for corn and soybean,
 and changes hi forest composition were analyzed for
 Michigan and  Minnesota (see Figure 13).

 Lake Levels Could  Drop and Ice Cover Duration
 Could  Decrease

     Higher  temperatures  would  likely reduce
 snowpack and could  increase evaporation,  which
would lower lake levels. The level of Lake Superior
was estimated to be  reduced under the climate
scenarios by 0.4 to 0.5 meters (1.2 to 1.5 feet), and

                                                                                          Effects of Climate Change
                      TEMPERATURE SCENARIOS
                          2xCO. lesslxCO,


                        WINTER  SPRING  SUMMER  FALL
                      PRECIPITATION SCENARIOS
                         WINTER  SPRING SUMMER   FALL
       Climate change could:
       . cause average lake levels to fall by
         0.5 to 2.5 meters
       . reduce ice cover duration by 1-3 months
       Adjustments may be required, including:
       . increased dredging of harbors and
         channels, or
       • lower cargo capacities on ships

       Water Quality
       Changes in temperature and precipitation
       could cause:
       • greater stratification in lakes and
         increased growth of algae, which in turn
         could cause lower dissolved oxygen
         levels in shallow areas
       • an increase in pollutants resulting from
         more dredging
        Wetlands and Fisheries
        Higher temperatures could cause:
        •  an increase in fish habitats in fall, winter,
          and spring, and a decrease in summer
        •  accelerated growth for some fish species
        •  potential invasion by new species
        Higher temperatures could result in:
        • loss of mixed northern hardwood and
          oak in southern areas
        • shifts of mixed northern hardwood and
          boreal forests in northern areas to all
          northern hardwood
        • forest declines evident in 30 to 60 years
                                                                         Higher temperatures could cause:
                                                                          • corn and soybean yields to increase in
                                                                           North, decline in Cornbelt; mixed results
                                                                           under climate change and COg
                                                                          • acreage could expand in the North,
                                                                           leading to increased erosion and runoff
                                               Figure 13.  Great Lakes.
that of Lake Michigan by 0.9 to 2.5 meters  (3 to 8
feet). Diversions out of the lakes for irrigation or
to  supply other basins would further  lower  lake
levels, although these  impacts were not analyzed.
These results  are  very sensitive  to assumptions
made   about    evaporation   and   under    some
circumstances, lake levels could rise.

      Higher temperatures  would  also reduce ice
cover on the lakes.  Specifically, they could cut ice
duration by 1 to 3 months on Lake Superior and by
2 to 3 months on Lake Erie, although ice still would
form on both lakes.  Changes in windspeed would
affect the reduction in duration of ice cover.  In
response  to lower  lake levels, either  ships would
have  to sail with  reduced  cargoes  or ports  and
channels would have to be dredged.  On the other
hand, a shorter ice season  would allow  a longer
shipping season.

 Executive Summary
 Water Quality May Be Degraded in Some Areas

     Higher   temperatures  could  lengthen
 stratification   of  the   lakes   (where  summer
 temperatures warm  the  upper part  of lakes  and
 isolate the cooler lower layers of lakes). Analysis of
 the Central Basin of Lake Erie showed that longer
 stratification,   combined  with   increased  algal
 productivity, would most likely reduce dissolved
 oxygen levels in the lower layers of the lake (see
 Figure 14). Reducing pollutant loadings in the lake
 would likely result in less severe impacts. One study
 raised the possibility that the annual mixing of a
 lake such as Lake Michigan may be disrupted.  If
 winds  and storms increased, such outcomes would
 be less likely.  Disposal  of contaminated dredge
 soils could increase water pollution.

 Fish Productivity in Open Areas Mav Increase

     The  average annual thermal habitat  would
increase with a warmer climate (see Figure 15).  If
 sufficient  oxygen  is  present,  growth  rates  and
 productivity for such fish as bass and lake trout in
 open areas of large  lakes may increase, provided
 that the  forage base also increases.   However,
 reduced ice cover and decreased water quality could
 harm some species in shallow basins of the Great
 Lakes.  The effects of increased  species interaction,
 changes in spawning areas, and possible invasion of
 exotic species were not analyzed.

 Northern Agriculture Mav Benefit

     As a result of the relative increase in northern
 agricultural  productivity,  agriculture   could  be
 enhanced in Minnesota, Wisconsin, and northern
 Michigan with  additional  opportunities for  the
 agriculture  support  sector.    The presence of
 relatively   poor   soils,   however,  could   limit
 agricultural expansion.  Increased  cultivation in
northern areas could increase erosion and runoff,
with negative impacts on surface and groundwater
                                 AUGUST 1970*

                            BASE CASE
   AUGUST 1975*
                               * Base Case Years     \///X  Area That Is Anoxic (Has No Oxygen)
    Figure 14. Area of central basin of Lake Erie that becomes anoxic under doubled CO2 scenarios.

                                                                           Effects of Climate Change
                                                                         ฑ 2ฐC OF OPTIMUM TEMPERATURE

                                                                         ฑ 5ฐG OF OPTIMUM TEMPERATURE
Figure 15.
Increases in thermal habitat for lake trout in southern Lake Michigan under alternative climate
Abundance  and  Composition  of Forests  Could
     Northern  hardwood forests in dry  sites  in
Michigan may die back and could become oak
savannas or grasslands.  In  northern Minnesota,
mixed boreal and  northern hardwood forests may
become   completely   northern  hardwoods.
Productivity in some wet sites in Michigan could
improve. Commercially important softwood species
could be replaced  by hardwoods used for different
purposes. Changes in forests could be evident in 30
to 60 years.  Whether reforestation with southern
species  not  currently in  the  region  and  CO2
fertilization would mitigate these impacts was not


     The Southeast is distinguished from the other
regions in this study by its warm  temperatures,
abundant rainfall, large coastal plain, and productive
marine fisheries. The region supplies about half of
                                           the nation's softwood and hardwood  timber, and
                                           tobacco,  corn, and soybeans are among its major
                                           crops.  Over 85% of the nation's coastal wetlands
                                           are in  the Southeast, and over 43% of the finfish
                                           and 70% of the shellfish harvested in the United
                                           States are caught hi the region.

                                                This report focused on two regions within the
                                           Southeast:   the  Tennessee  Valley  and   the
                                           Chattahoochee  and Apalachicola Rivers.   The
                                           Tennessee Valley Authority examined the potential
                                           vulnerability  of  its  water management  system to
                                           high and low riverflow scenarios (based on runoff
                                           estimates from GCMs).  Flow in the Chattahoochee
                                           River Basin was  estimated using hydrologic analysis
                                           to study impacts  on  the  management  of Lake
                                           Larder, which supplies water to Atlanta.   The
                                           estimates of outflow from  the lake, along with
                                           estimates of the flow  in the  Apalachicola River,
                                           were  combined with  potential  wetland  losses
                                           attributable to sea level rise to identify impacts on
                                           finfish and shellfish in Apalachicola Bay.  Sea level
                                           rise  impacts for the entire  Southeast  were derived

                                        CHAPTER 1
    Since the beginning of the Industrial Revolution,
 human   activities   have   led  to   increased
 concentrations  of  greenhouse  gases   in   the
 atmosphere.  Fossil fuel burning, which  releases
 CO2> CO, N2O, and other pollutants, has expanded
 many times over.  Changes in agriculture have led
 to  increased  emissions  of  CH4  and
 Population growth has contributed to deforestation
 in many areas  of  the globe, which  in turn has
 affected the global carbon  cycle.   Atmospheric
 concentrations   of  tropospheric   ozone   and
 chlorofluorocarbons have also increased, primarily
 because of industrial activity.

    Scientists have concluded that the increase in
 greenhouse  gas concentrations  will  eventually
 change  global climate.   In  1979, the National
 Academy of Sciences stated  that a doubling of
 carbon dioxide levels would lead to an increase of
 1.5 to 4.5ฐC (2 to 8ฐF) in global air temperatures.
 Since  then, other researchers have examined the
 increase in all greenhouse gases and have concluded
 that a greenhouse gas increase equivalent to CO2
 doubling could occur  as early as the 2030s, with a
 hypothesized commensurate global warming lagging
 by several decades.

   The Earth's atmosphere has  undergone many
 cycles   of  warming  and  cooling  in  the  past.
 Paleoclimatologists have estimated that at the glacial
 maximum  of the last ice  age, which was  about
 18,000 years ago, the Earth was approximately 5ฐC
 (9ฐF)  cooler than at  present.  This is  generally
 attributed  to  changes  in  orbital characteristics
 combined with lower trace gas concentrations and
 different climate feedbacks.

   Two aspects may make the current greenhouse
warming different from past climate changes.  First,
it will raise temperatures higher than the planet has
experienced in the last 125,000 years.  (During the
Pliocene Epoch (2 to  5 million years ago), global
temperatures were several degrees higher than they
are  now.)   Second,  past  climate  changes of
comparable magnitude have generally occurred over
tens of thousands of years. Estimates are that the
            greenhouse   effect   may   raise   atmospheric
            temperatures several degrees in less than a century.

    The significant implications of the greenhouse
 effect have been the subject of discussion within the
 scientific community for the past three decades. In
 recent years, Members of Congress  have held
 hearings and have begun to explain the implications
 for public policy.  This interest was accentuated
 during a series of hearings held hi June 1986 by the
 Senate  Subcommittee   on   Pollution  of  the
 Environment  and   Public Works  Committee.
 Following  the hearings, members  of the Senate
 Environment and Public Works Committee sent a
 formal request to the EPA Administrator, which
 asked the Agency to undertake two studies on
 climate change due to the greenhouse effect. (The
 letter is reprinted in Appendix C of this report.)

    One of the studies we are requesting should
    examine   the   potential  health  and
    environmental effects  of  climate  change.
    This study should include, but not be limited
    to, the potential impacts  on agriculture,
    forests, wetlands, human health, rivers, lakes,
    and estuaries, as  well  as other  ecosystems
    and societal impacts. This study should be
    designed  to include original analyses,  to
    identify and fill in where important research
    gaps exist, and to solicit  the opinions  of
    knowledgeable   people  throughout   the
    country through a process of public hearings
    and meetings.

   Congress  also requested that EPA prepare a
study on policy options to stabilize current levels of
atmospheric greenhouse  gas concentrations.  That
study analyzes  policy  options  for  limiting gas
concentrations   including  energy   efficiency,
alternative  technologies,  reforestation  options,
chlorofluorocarbon (CFC) reductions, and  other
options for limiting CH4 and N2O.  It is entitled

                                                                            Effects of Climate Change
These changes could be evident in 30 to 80 years.
The forest studies did not consider whether more
southern species could be transplanted and survive
in the region, nor did they account for higher CO2
concentrations,  which could mitigate some losses.
The combined  effects of reduced  agriculture and
forestry could lead to significant economic losses in
the Southeast.

Some Coastal Fish Species Would Be Harmed

    Sea  level  rise could  inundate  most of the
coastal  wetlands and raise salinity levels,  which
could reduce the populations of gulf coast fisheries.
In addition, higher temperatures may exceed the
thermal tolerances of many species of shellfish in
gulf  coast   estuaries,   further  reducing   fish
populations. Whether these species would be able
to migrate to  cooler water was not considered.
Some   species,  however,  could  increase  in
abundance,  while others may  migrate  into the

The Studies Were Unable to Determine Regionwide
Impacts on Water Resources

     The Southeast  currently  has  little  winter
snowcover.   Therefore,  seasonal  runoff depends
much more  on changes in rainfall than on changes
in temperature that affect the size  of snowpack.
Analysis of  the rivers managed  by the Tennessee
Valley Authority showed that increased runoff could
lead  to  higher  riverflow   and  higher   flood
probabilities, while less runoff could reduce flood
probabilities, but could lead to lower riverflow and
problems  maintaining  adequate  supplies  for
industrial use, powerplants, and dilution of effluent.
Use  of  climate  change  scenarios  produced
inconclusive results concerning the potential change
in flow in the Chattahoochee River. A study of the
management of Lake Lanier concluded that changes
in  operating rules  would be  sufficient  to handle
higher or lower flows estimated in the scenarios,
although some uses would be restricted.

The  Great Plains

     Agriculture is  one of the main  sources  of
income in the Great Plains. The States of Kansas,
Nebraska, Oklahoma, and Texas produced 80% of
the nation's sorghum and 30% of the wheat crop in
 1982.  In recent years, increased use of water from
the Ogallala Aquifer has reduced groundwater levels
in the region, with potential long-term consequences
for agriculture and the economy.

     The studies in this report focused on Nebraska,
Kansas, Oklahoma, and Texas, and concentrated
mainly  on  agriculture-related impacts.    They
estimated changes in corn, wheat, and soybean
yields and in the demand for irrigation. Changes in
runoff and leaching of chemicals from farms were
also  examined (see Figure 17).

Crop Acreage Could Decline

     The  crop yield and  economic adjustment
studies  indicate that grain crop  acreage  could
diminish in the region. The direction of changes in
wheat and corn yields depends  on the direct effects
of CO2 on crop growth and the severity of climate
change.  If climate becomes hotter and relatively
drier, yields could decrease.  Whatever the climate
change, relative productivity may decline compared
with northern areas.  As a result, crop acreage was
estimated to drop by 4 to 22%. Such a reduction in
agriculture could adversely affect  the economy of
the region.  These studies did not consider use of
new technologies or introduction of new crops.

Demand for Irrigated Acreage Would Increase

     The demand  for irrigation on the farms that
continue  to  grow grain  crops  could  increase.
Irrigated acreage, which currently makes up about
10% of the  total  acreage  and is growing, could
increase by 5 to 30%. This report did not examine
how this demand would be satisfied, although the
Ogallala Aquifer  could  be a candidate.    Other
impacts of global warming could change ground and
surface water supplies and, possibly, surface water
quality.  Changes  in precipitation could affect the
leaching of pesticides into groundwater and runoff
to  surface waters in  some  cases, although  the
direction of change cannot  be  determined because
runoff and leaching of pesticides and soils are very
sensitive to rainfall variability.

     Because this is the most comprehensive study
 to address the issue of the environmental effects of
 climate change in the United States, we expect that
 a sizable  debate  will  follow  its  publication.

 Executive Summary
                   TEMPERATURE SCENARIOS
                      ZxCO, less 1xCO.

                     WINTER SPKHQ SUMMER FALL
                                                                    Higher temperatures could:
                                                                    • reduce corn and wheat yields, and
                                                                      could have mixed effects on yields
                                                                      when considering both climate change
                                                                      and increased CO2
                                                                    • reduce crop acreage
                                                                    Irrigation Demand
                                                                    Changes in agriculture are likely to
                                                                    result in increased irrigation demand
                                                                    and acreage
          Water Quality
          Changes in rainfall, runoff, pesticide
          loadings, erosion, and irrigation
          could affect water quality
                                                                    Higher temperatures could increase
                                                                    electricity demand
                     WINTER SPttHO SUMMER FALL
                                        Figure 17. The Great Plains.
Considerable additional research and analyses are
likely to amplify,  improve,  and  challenge these
findings.  We expect further research  to develop
new insights into the role of climate, but precise
forecasts must await more advanced climate models,
which may require many years to develop.  For
some time to come, our ability to provide national
and local officials with guidance may be limited to
effects driven primarily by temperature and sea level

     Apart  from strategies to  limit emissions  of
greenhouse gases  (discussed  in  the  companion
report), policymakers should consider policy options
for adapting to global warming.  Consideration of
these options  is complicated by the uncertainties
identified in this report by delays in the onset of
climate change, and by the pressure to solve today's
problems.  Many adaptations would undoubtedly
occur as climate changes, but some decisions being
made today  have a  long  enough lifetime and
sufficient risk  to support  consideration of  the
potential range of impacts of the greenhouse effect.
These decisions  should  be  made if they  make
economic  and environmental  sense for today's
conditions and are sufficiently flexible to handle
changing climate.  Given the uncertainty about the
timing,  magnitude, and regional scope  of climate
change,  we  cannot  plan  for specific   climate
conditions in the future,  but we can strive to be
ready to respond to significantly changed climate
conditions in the future.

                                                                            Effects of Climate Change
     Conversely,  natural  resource  management
should not assume that climate will not change. All
managers of natural resources that are sensitive to
climate should consider the vulnerabilities of their
systems to climate change and whether anticipatory
steps are prudent. In some cases, no anticipatory
action  would be needed  —  the  systems can be
adjusted and adapted  as climate changes. In other
areas,  where  long-term  decisions  on  sensitive
systems  may  result  in   irreversible  impacts,
anticipatory actions   to mitigate these  potential
effects may be required.  It may make sense in
some instances to change the rules under which
long-term planning is  done, such as zoning laws, to
allow for consideration of climate change in private-
sector decisions.  Finally, research and education
are needed in many areas to improve our ability to
respond to these changes.  In any case, managers
should reexamine their systems to consider ways to
improve the flexibility and resiliency of the systems
to handle these and other changes.  The criteria to
guide decisions should include consideration of the
following factors:

     •    the uncertainties in the magnitude and
         timing of effects;

     •    whether the lifetime of the plan, project,
         or policy is long enough to be affected by
         climate change;

     •    whether effects of climate change  are

     •   whether the policy or project will increase
         flexibility and resilience or restrict future

     •   whether  a  policy or   action  makes
         economic or  environmental sense,  even
         without climate change;

     •   the  uniqueness of the  ecosystems  or
         manmade   structures  that  may need
         protection; and

     •   whether the impacts would be greater if
         no anticipatory action were taken.
     The U.S. government is strongly supporting the
Intergovernmental  Panel  on  Climate   Change
(IPCC) under the auspices of the United Nations
Environment   Programme   and   the    World
Meteorological Organization.   The  IPCC has
established a  process for governments to follow
when reviewing scientific information and policy
options.   The  federal government is conducting
other activities on  global  climate change.   The
Global Climate Protection Act of 1987 calls for a
scientific assessment of climate change, which is to
be completed by 1989. This work will be sponsored
by EPA and  other federal agencies such as the
National Aeronautics and Space Administration, the
National Oceanic  and Atmospheric Administration,
and   the  National  Science  Foundation,  and
coordinated  through  the  IPCC.    Also, the
Department of Energy and EPA have been asked to
report to Congress  on policy options for reducing
CO, emissions in the United States.  In addition,
various federal agencies conduct significant research
programs on climate.  These research efforts on
climate change are coordinated by the  National
Climate  Program Office and the Committee on
Earth  Sciences.  The latter has produced a plan
called  Our Changing Planet:   A  United States
Strategy  for  Global  Change  Research,  which
outlines federal research activities.

     The federal government can also take the lead
in pursuing  prudent policies in  anticipation  of
climate change, and many agencies can play a role
in preparing the  country for the impacts.  These
include the Departments of the Interior, Energy,
Health and Human Services, and Agriculture; the
U.S. Environmental Protection  Agency;  and the
U.S. Army Corps of Engineers (see box on "Federal
Activities"). However, adaptation should not occur
just  at the federal level, for there will likely be a
need to  involve  other  nations,  state  and  local
governments,  industry, and  even individuals. The
regional studies  in this report  demonstrate that
climate change cuts across  manmade and natural
systems, geographic boundaries,  and government
agencies.  Research, technical guidance,  planning,
and creative approaches to resource management
will  be needed in  the  future to prepare for the
impacts of climate change on the United States.

Executive Summary

        Sample questions relating to climate change impacts that federal agencies should consider;
   XJ.S. Environmental
    Protection Agency
      . department of
    the Interior
   U.S. Department of
  tf-S. Army Cdrps
    of Engineers
  Federal Emergency
    Management Agency

  U.S. Department of
    Health and Human
                         Policy Question

 How should current wetlands protection programs be modified to
 accommodate future sea level rise and precipitation, changes?

 Should regulatory approaches to air pollution 1>e supplemented with incentive
 systems, new chemicals, or relocation policies?

 Should national parks and wildlife refuges purchase land to accommodate the
 migration necessitated by climate change? Should additional parks and refuges
 be created?

 Are current activities increasing the vulnerability of species that might be
 threatened by climate change?

 Should the U.S.  Geological Survey produce  coastal area maps with finer
 contour intervals? How will climate change alter projected groundwater levels?

 Will current water policies in the West prove to have been ill-advised if the
" climate changes?

 Do price support programs help or hinder the adjustments that climate change
 may necessitate?                         % ,

 To what extent could irrigation be increased on a sustainable basis if climate
 became drier?

 What actions would be necessary to maintain  national forests as the climate
 changes?                           '    '"

 How does a consideration of future climate change alter the relative merits
 of alternative approaches to coastal protection, flood control, and navigation?

 Will climate change  affect the success: of wetlands protection efforts in
 Louisiana as administered under Section 404 of the Clean Water Act?

 Will current rate caps on premiums enable the National Flood Insurance
 Program to remain solvent if climate changes?

 Are current programs adequate to address potential changes in. mortality and
 shifts in diseases resulting from climate change?

Executive Summary
from  the  national  studies.    Crop yields were
estimated for corn and soybeans,  and changes  in
forest composition were analyzed at several sites
across the region (see Figure 16).

Adverse Impacts on Agriculture and Forests Could
Hurt the Region

      Decreases  in  the  relative  productivity  of
southeastern agriculture were estimated under the
scenarios to lead to the abandonment of 10 to 50%
of  the  agricultural  acreage in  the  region.   The
studies did not consider introduction of new crops,
such as citrus, or the use of new technologies, such
as biotechnology.

      Most forests in the Southeast were estimated
to  have difficulty surviving the  assumed  climate
change.  Dieback of existing forests in such areas as
Georgia and Mississippi could be particularly large.
                       TEMPERATURE SCENARIOS
                           2xCO Jess 1xCO,
                         WINTER  SPRING  SUMMER  FALL
                       PRECIPITATION SCENARIOS
                         WINTER  SPRING  SUMMER  FALL
                                                                            Climate change could:
                                                                            • decrease corn and soybean yields in
                                                                             hotter areas and could have mixed
                                                                             results elsewhere
                                                                            • decrease cultivated acreage
                                                                            • increase need for irrigation
                                                                            • increase pest infestations
           Higher temperatures could result in:
           • significant dieback of southern forests
            with declines evident in 30 to 80 years
           • regeneratbn of species becoming
                                                                            Water Resources
                                                                            Increased temperature and changes
                                                                            in precipitation could:
                                                                            • produce uncertain effects for water
                                                                             resource availability
                                                                            • affect water quality and flood risks
                                                                            • lower levels in some recreational lakes
           Sea Level Rise
           Rising sea level could:
           • inundate a significant proportion of the
             region's coastal wetlands
           • flood some dry land areas
           • create significant costs for protecting
             coastal resources
                                                                            Higher water temperatures and rising
                                                                            sea level could reduce fish and
                                                                            shellfish populations
                                                                            Higher temperatures could increase
                                                                            electricity demand
                                               Figure 16.  The Southeast.


 Chapter 1
 Policy Options for Stabilizing Global Climate and is
 a companion to this report. Congress requested the
 studies  in  the   Fiscal  Year  1987  Continuing

    This report builds on the past contributions of
 many scientists throughout the world, most notably
 the reports by the National Academy of Sciences
 (1979,  1983,  1987), the  World Meteorological
 Organization  and  the  International  Council of
 Scientific  Unions  (1986),  the  United  Nations
 Environment Programme (1986), Scope 29 (1986),
 and the U.S. Department of Energy (1985a,b).  It is
 an attempt to identify some of the sensitivities,
 direction  and   magnitude,  linkages,   regional
 differences, national impacts, policy implications,
 and uncertainties associated with the effects of
 global climate warming.

    We hope it will provide useful information to
 climate modelers and effects researchers.  We  also
 hope that officials, at all levels of government, will
 be encouraged  to examine the  implications of
 climate change for long-term policies. Since this is
 the first study of this type, we expect that a great
 deal more research, analysis, and planning will be
 needed in the future. We do not pretend to have all
 the answers.

   This  report has been designed to identify the


   Since the rate and extent of climate change on
 a regional level  are uncertaia  we cannot predict
 effects. However, we can identify the sensitivities of
 systems to climate change.  Our goal was to use a
 variety  of scenarios to  determine  what climate
 variables are important in causing impacts and the
 degree to which systems are sensitive to changes in
 these variables. Specifically, we were interested in
identifying the sensitivity  of systems to  higher
temperatures and sea level, which are  among  the
 changes  most likely to  occur following increased
greenhouse gas  concentrations.    (For  further
discussion, see Chapter 2: Climate Change.)
 Direction and Magnitude

    Since the  scenarios  do  not encompass  all
 possible combinations of climate change  due to
 increased greenhouse gases,  the results  do not
 represent the entire range of possible effects.  For
 example, there could be more or less rainfall, or
 higher  or lower  temperatures than estimated by
 climate  models.   Yet, the results  from  various
 scenarios help define the direction and magnitude of
 effects.  First, we  examined  them  to  see if a
 direction of change  (e.g., more water, lower crop
 yields)  is  evident.   Second,  we attempted  to
 determine if the magnitude of change is significant.
 Third, we asked whether the results are consistent
 with scientific theory. Outcomes outside the bounds
 of our results cannot be ruled out at this  time.


    Individual environmental systems will  not  be
 affected by climate change in isolation.  Water
 resources, for example, maybe affected not only by
 changes in water  supply but  also by changes in
 demand for water for such purposes as irrigation.
 Wildlife may be  directly affected by changes in
 climate and indirectly affected by changes in habitat
 due to  climate change.  This report attempts to
 identify linkages among effects, quantitatively where
 possible and qualitatively elsewhere.  Linkages are
 identified mainly in regions.  Quantitative analysis
 of all linkages would change the numerical results of
 this report, in many cases exacerbating impacts.
National Impacts

   Impacts were analyzed on a national scale to see
how the country as  a whole may be affected by
climate change and  to see if latitudinal patterns
(such as northward shifts in species) are detectable.
Some analyses, such as coastal wetland impacts and
changes in electricity demand, were conducted on a
national basis.  Other national analyses, such as
forests, were based on results from regional studies.
In some cases, national analyses  estimated total
costs over the next century. No attempt was made
to assess  the  total national impact from climate
change, and conclusions about the total costs and
benefits of climate change should not be made.

Regional Impacts

   Effects were examined in several regions of the
United States for a number of reasons. As pointed
out above, linkages exist among many of the effects,
and these are likely to be seen on a regional scale.
For example, the supply of water in a river basin
may change as a result of climate change.  The
water resource in that basin may also be affected by
changes in the demand  for water for irrigation,
powerplant cooling, and  other uses.  Analysis of
similar  systems  in different regions allows for
comparison of impacts among regions. This report,
however, does not attempt to identify "winners and


   Many uncertainties are related to our knowledge
about the  rate and  magnitude  of warming and
changes in regional weather patterns. As discussed
in Chapter 2: Climate Change, we do not know how
much and how quickly climate may change and how
regional climates may change. Uncertainties also
exist about how ecological and other systems will be
affected by climate  change.  We  do not  have
empirical  evidence on how these systems will
respond to  higher temperatures and CO2 levels, as
well as  to different rainfall amounts.   These
uncertainties are reflected in the models  used to
estimate climate change and impacts. This report
attempts to clearly state these limitations.

Policy Implications

   The management of most natural resources has
generally been undertaken assuming that climates
will not  change.  A change in climate could affect
many of these resources and raise implications for
resource management.  This report discusses some
policy implications of climate change, but it does not
lay out a prescriptive  policy agenda.

Research Needs

   The analysis in this report should provide climate
modelers with information concerning how general
circulation  models could be  improved.  It should
also help define research needs for future analysis
of the potential impacts of climate change.
   Fundamentally,  these goals  center  on  the
identification of important issues and state-of-the-
science investigations in each environmental system.
Because  each component of  science and  policy
development is at an early stage,  the goals  of the
report are to develop insights and estimates of the
ranges of possible future effects  and to use that
information for identifying where  the policies and
research programs  of EPA and other agencies
should be reexamined.

Important Systems

   This report focuses on the following systems,
which are important, are sensitive to climate, and
may be particularly affected by climate change:

   Sea Level Rise
   Water Resources
   Electricity Demand
   Air Quality
   Human Health
   Urban Infrastructure

Regional Case Studies

   Four regional case studies were  selected:  the
Southeast,  the  Great Lakes, California,  and the
southern Great Plains.  These regions were picked
because each is important for economic, social, and
environmental reasons, and each offers some unique
current characteristics that make it  an interesting
example of the  range  of possible environmental
issues  that may need to  be considered.   The
Southeast   depends  heavily  on   forestry   and
agriculture, and has extensive and fragile wetlands
and coastal ecosystems.  The Great Lakes are the
dominant natural resource in their region, supplying
freshwater, fishery resources, and a pathway for
shipping and transportation, and providing a natural
laboratory for environmental issues that affect both
the United States and Canada. California already
must carefully manage its water  supplies, and its
agricultural industry provides many  crops for the
United States and a large share of the international

Chapter 1
market; it is among the most productive agricultural
regions in the world. The Great Plains is one of the
largest  producers of  grain  crops in  the  world.
Although these regions are diverse, they do not
encompass the entire range of regional  differences
in the United  States.   The  analysis of effects  in
these regions does not cover all potential impacts hi
the United States.

National Studies

   The  effects on  a  number of  systems were
quantitatively analyzed on a national scale. National
agricultural markets were analyzed with respect  to
their sensitivities to changes in yield derived from
our agricultural models. Options for adapting to a
sea level rise were examined on a national scale,  as
were possible  health  impacts.   Forestry,  water
management, air  quality,  and biodiversity  issues
were explored by analyzing the results of several  of
the regional case studies with a broader perspective.
In each case, the national-level analyses  provide an
additional  level of qualitative integration  that  a
purely regional analysis  could not.  The structure  of
the regional and national studies is displayed  hi
Figure 1-1.
                            climate scenarios we used were based on outputs
                            from  general  circulation models  (GCMs)  (see
                            Chapter 4: Methodology). Where possible, we tried
                            to  obtain  quantitative  estimates  of  effects.
                            However, the development of quantitative estimates
                            was  constrained  by  the  availability  of  well-
                            documented models that included some interaction
                            of the particular effect hi question and climatic
                            variability.  We obtained additional information on
                            sensitivities  by reviewing the  literature and by
                            gathering expert judgment.  The approach of using
                            existing  models,  all  of  which  were  originally
                            constructed  for   other  purposes,   makes   the
                            interpretation of results instructive but  somewhat
                            limited with respect to the full range of climatically
                            relevant questions that could be asked.
                            PROCESS   FOR   CONDUCTING
                            THIS REPORT

                               We used an eight-stage process to define  the
                            scope of this report, select the projects, write  the
                            chapters, and review the results.

                            Step 1: Initial Scoping of the Report

   Since we do not know how climate will change,
this report used scenarios of possible climate change
to identify sensitivities of systems to climate. The
                               This stage immediately followed the  request
                           from the Senate Environment and  Public Works
                           Committee.  We agreed on using the regional case
                           study  approach,  on  the  four  regions  to  be
                           investigated, and on using climatic scenarios. We
Sea Level Rise
Water Resources
Electricity Demand
Air Quality
Human Health
Urban Infrastructure
                           Great Lakes
                           Great Plains
                                                      Sea Level Rise
                                                      Electricity  Demand
to Congress


Data Bases
                               Figure 1-1. Elements of the effects report.

also  decided   not  to   attempt  to   analyze
environmental effects outside the United States in
this report.   Our rationale; for this  decision was
based on available time and funds, and on the lack
of suitable  models  that  would be immediately
accessible to us.

Step 2: Preparatory Workshops

   We held two workshops in February-and April
1987 in Boulder, Colorado, to prepare the report.
In the February workshop, sponsored and organized
by the National Center for  Atmospheric Research,
general circulation modelers convened to discuss
some of the problems inherent in attempting to
understand the regional results from global models.
Several major  topics  were  discussed  from  the
standpoint of how the results from GCMs should be
used in impact studies.  A list of. variables that
would be available for use by effects researchers
was  produced at the end  of the  workshop.   In
addition, several potential studies on aspects of the
frequency  of  extreme   weather  events  were

   The April workshop  was  organized with the
assistance   of   the  University   of   Colorado.
Approximately  100  scientists explored  the major
climate change-related issues in agriculture, forest
effects, water resources, and sea level rise. Working
groups  in each discipline  discussed  the potential
impacts that climate change might have and the
most important uncertainties to explore to arrive at
better predictions. The working groups  were then
rearranged into regionally  oriented groups.  They
identified a series of studies that would address the
major scientific issues in each region.

Step 3: Identification of Potential Projects

   From  the lists identified in the  two Boulder
workshops, and from additional studies on urban
and regional  air quality   subsequently  identified
internally by EPA, we arrived at list of investigators
from  whom we  would  solicit. proposals.   The
decision to solicit proposals was based primarily on
the potential coverage of environmental issues in
each region.

Step 4: Reviews  of Proposals

   At least one  intramural  and two  extramural
reviewers examined each proposal. We responded
 to  all  comments  and  modified  proposals  as
 appropriate.     EPA  used   a  combination  of
 cooperative  agreements,  existing contracts,  and
 interagency agreements to fund  projects for  this

 Step 5: Planning and Integration

    All  the researchers  met  with EPA staff in
 October 1987 to  discuss scenarios,  goals,  and
 approaches for the  studies. Researchers discussed
 integration of projects within regions as well as the
 commonality of approaches within disciplines.

 Step 6: Analysis

    The National Center for Atmospheric Research
 assembled the scenarios  and  distributed them to
 researchers  in  the fall  of  1987.   Researchers
 conducted their analysis over  the winter  and
 prepared draft reports in March and April 1988.

 Step 7: Preliminary Project Review

    In April  1988, EPA assembled  panels  of
 scientists to provide a preliminary review of most of
 the agriculture, forestry,  and hydrology projects.
 The  principal investigators  of  the  appropriate
 projects were asked to present their work orally and
 in written  drafts. EPA project managers used the
-comments from   the  review panels   to  make
 corrections in the conduct of a few projects, and as
 a guide to interpreting the  results of  individual
 projects and  to writing this report.

 Step 8: Project  and Report Peer Review

    At least two to  three peer reviewers examined
 the final reports from all principal investigators
 before the EPA project managers accepted them.
 During this time, EPA staff on the report project
 team wrote the overviews that are reflected in this
 final report.   In November  1988,  a  special
 subcommittee of EPA's  Science  Advisory Board
 (SAB) was convened and asked to review the entire
 report.   Following  the SAB's written review, the
 EPA project team responded to comments  and
 produced  the final  version of the Effects Report.
 The draft  of the report was sent to other federal
 agencies and the Office of Management and Budget
 for review and comment, and these comments were
 also taken into account in the final version.

 Chapter 1

   This  report  is divided into  several sections.
 Section  I  consists  of Chapter  2 on trends  in
 emissions of greenhouse gases and potential impacts
 on climate; Chapter 3 on changes in variability; and
 Chapter  4  on the choice of scenarios and effects
 modeling.  In Section II, the results of national
 analyses  are  presented.   Each chapter covers  a
 different system. The chapters include an overview
 of relevant regional studies, and they present results
 from national analyses. Each chapter discusses the
 current state of resource, reviews previous literature
 on climate  change  and  the resource,  discusses
 studies used for this report, presents national results
 from regional and national studies, and discusses
 broader socioeconomic and policy implications. The
 design and limitations for each study are presented
 only once — in a regional chapter if it is a regional
 study or in a national chapter if it is a national
 study.  Section m contains results from the regional
 case studies, with each chapter devoted to different
 regions.   Each  regional chapter describes  the
 climate-sensitive systems in  the  region; reviews
 previous  studies  on impacts  of climate change on
 the  region; describes  the structure of regional
 studies for the report; discusses regional climate
 change scenarios; reviews the design, results, and
 limitations of the studies; and discusses the broader
 socioeconomic and policy implications of climate
 change for the  region.   The regional chapters
 include relevant  regional results  from national
 studies.  Not all regionally relevant results  are
 presented in the appropriate regional  chapters.
 Results for health are presented only in the health
 chapter  in  Section  n.    Section IV  includes
 conclusion chapters. Chapter 18 discusses directions
 for future research on climate change effects, and
 Chapter  19  discusses policy implications  and

   This report is designed to be an overview of the
 individual studies. Those studies are printed  in
 appendix volumes. In this report, the studies are
 referenced  by the author's name or  names  in
parentheses and volume letter. Previously published
work is referenced by the author's name and the
year of publication.

National Research and Policy Activities

   The Global  Climate Protection Act of  1987
requested EPA  to develop a national policy on
global climate change and to prepare an assessment
of scientific information.  The very scope  of this
issue suggests that this request can be fulfilled only
in cooperation with other federal agencies; hence,
EPA is working with these agencies to formulate a
process to  achieve  this  goal.    The  scientific
assessment will be  conducted in coordination with
the National Aeronautics and Space Administration,
the   National   Oceanic   and    Atmospheric
Administration, the National Science Foundation,
and other agencies.  To the extent possible, this
scientific assessment will also be developed on an
international basis and should be available in 1990.

   The development of a national policy will be
coordinated with the Department of Energy and
other natural resource departments.  The goal will
be  to build  on this report  and others  under
development by federal agencies  to identify the
adoptive policies and other measures that may be
appropriate to deal with this issue.  The nature  of
this issue suggests that a continuous review  of
domestic policy will be required for many years.

International  Activities

   In  1987,  the  United  Nations  Environment
Programme (UNEP) and the World Meteorological
Organization (WMO)  were  asked by  member
governments  to  establish an  Intergovernmental
Panel on Climate Change (IPCC)  for the specific
purpose of reviewing the scientific information and
potential response strategies.   The WMO has
primary responsibility  for  the  World  Climate
Research Programme, and UNEP has responsibility
for the World Climate Impacts Programme.  The
UNEP  was  the  primary  international   agency
responsible for negotiations leading to the Montreal

Protocol To Protect the Ozone Layer.  The first
meeting  was  held  in  November  1988,  and
subsequent meetings have been held in 1989 to
organize activities. It is expected that the IPCC will
be the primary forum for multilateral discussions
between governments on this issue.

   Other governments and  international  agencies
are also examining this issue. Italy, Japan, and the
Netherlands held conferences in 1989.  The United
States has bilateral activities with the Soviet Union
and  China   The  Organization  for  Economic
Cooperation and Development and the International
Energy Agency   are  examining their   potential

National Academy of Sciences.   1979.   Carbon
Dioxide  and Climate:  A Scientific Assessment.
Washington, DC: National Academy Press.

National Academy of Sciences.  1983.  Changing
Climate.   Washington,  DC: National Academy

National Academy of Sciences.   1987.  Current
Issues in Atmospheric Change.  Washington, DC:
National Academy Press.
Scope 29.  1986.  The Greenhouse Effect, Climatic
Change, and Ecosystems (SCOPE 29). Bolin, B., B.
Doos, J. Hager, and R. Warrick (eds). Chichester,
England: John Wiley and Sons.

United Nations Environment Programme.  1986.
Effects of Changes in Stratospheric Ozone  and
Global Climate. Titus, J.G. (ed). Washington, DC:
U.S.  EPA  and  United  Nations  Environment

U.S. Department of Energy.  1985a.  Atmospheric
Carbon  Dioxide and the  Global  Carbon  Cycle.
Trabalka, J.R. (ed). Washington, DC: Government
Printing Office (DOE/ER-0239).

U.S. Department of Energy. 1985b. Projecting the
Climatic Effects of Increasing Carbon Dioxide.
MacCracken,  M.C.,   and  F.M.  Luther  (eds).
Washington,  DC:  Government  Printing  Office

World Meteorological Organization.  1986. Report
of the International Conference on the Assessment
of the  Role of Carbon Dioxide and  of  Other
Greenhouse  Gases in  Climate  Variations  and
Associated Impacts, Villach, Austria, 9-15 October
1985. World Climate Programme Report No. 661.
Geneve,   Switzerland:  World   Meteorological
Organization, The International Council of Scientific
Unions,  and the  United Nations  Environment


                                      CHAPTER 2
                      GLOBAL CLIMATE CHANGE
    The Earth's climate has changed continuously
over the entire lifetime of our planet as a result of
various natural causes.  Recently, we have come to
the realize that human activities may, in the near
future,  produce  effects  powerful  enough   to
overwhelm these natural mechanisms and dominate
the changes of climate. By early in the next century,
the planet's temperature may rise to a range never
before experienced by our  species, at a rate faster
and to temperatures warmer  than the Earth  has
experienced in the past  million years.   This
anticipated temperature increase would be  caused
by an enhancement of the greenhouse effect.

    Although  the  overall effect  of  increased
greenhouse gases  is understood, many details  are
less clear, including both the timing of the predicted
warming and its spatial distribution. This is because
the response of the climate  system to the additional
greenhouse gases, including all the feedbacks and
interactions that   would  take  place,  is very
complicated and not completely understood.   In
addition, while  the human-induced component of
the greenhouse effect increases in magnitude, other
causes of climate change remain important, such as
changes in the amount of energy emitted by the sun,
changes in the atmospheric composition due  to
volcanic eruptions and human input of aerosols,
internal redistributions of energy by El Ninos, and
random, unpredictable variations. Thus, the task of
predicting the future evolution of climate involves
not only understanding the  response of the climate
system to  increased concentrations of greenhouse
gases but also predicting the concentrations of these
gases and the effects of other  causes of climate

    Several detailed assessments of the current state
of our knowledge of these projected climate changes
have  been conducted  recently.   These  include
studies by the National Research Council  (NRC,
1979, 1983,  1987),  the  World  Meteorological
Organization  (1986a,b), and the "state-of-the-art"
reports of the Department of Energy (MacCracken
and Luther, 1985a,b; NRC, 1985; Trabalka, 1985;
Strain and Cure,  1985; White,  1985).  Excellent
shorter summaries include Ramanathan (1988) and
Chapters 2 and 3  ofLashof  andTirpak (1989).
These studies should be consulted for more detailed

    This chapter describes the climate system, the
important causes  of climate change for the  next
century, and the so-called climate forcings, and it
summarizes  the various trace gases that human
           The Greenhouse Effect

          Gases in the atmosphere are virtually
      transparent   to   sunlight  {shortwave
      radiation), allowing it to pass through the
      air and to heat the Earth's surface,  The
      surface absorbs the sunlight and emits
      thermal radiation (longwave radiation)
      back to the atmosphere. Because several
      gases  in the  atmosphere,  particularly
      water vapor (ELO)  and  carbon dioxide
      (CO2),   are  not transparent  to  the
      outgoing thermal radiation, they absorb
      some of it and heat the atmosphere,  The
      atmosphere emits thermal radiation, both
      upward to outer space and downward to
      the surface, further warming the surface.

          This phenomenon  is  called  the
      greenhouse   effect  because  in  some
      respects  it  describes how  an  actual
      greenhouse  works.  Even without  any
      human  impacts, this natural greenhouse
      makes  the Earth's surface about 33ฐC
      (596F) warmer than it would be without
      the  atmosphere.     Gases  that  are
      transparent to sunlight, but not to thermal
      radiation, are called greenhouse gases.

          If either the concentration Of existing
      greenhouse gases increases or greenhouse
      gases that were not there  before  are
      added to the atmosphere, more thermal
      radiation will be absorbed and re-emitted
      downward, making the surface warmer
      than before.

Chapter 2
activities put into the atmosphere.  It then describes
important feedbacks in the climate system that act
to amplify or dampen the climate change induced by
the forcings. Uncertainties in our understanding of
these feedbacks are an important component of our
current uncertainty of the timing  and amount of
future climate change. Next, it discusses the recent
history  of  climate  change,   compares   these
observations with theory, and presents theoretical
models of the climate and their projections of future
climate change.  Finally, the concluding section
summarizes the extent of our knowledge about the
future climate and discusses future research needs.

    The climate system includes all the interactive
components  of our  planet  that  determine the
climate. This includes the atmosphere, oceans, land
surface, sea  ice, snow,  glaciers,  and  biosphere.
Climate change can be measured hi terms of any
part of the system, but it is most convenient to use
surface air temperature as a measure of climate,
since it is the parameter for which we have the best
record,  and  it is measured  where  the  most
important component of the biosphere — humans —
lives. Other components of the climate system, such
as precipitation, cloudiness, evaporation, windspeed
and direction, and sea level,  also have important
impacts on human activities.

    Figure 2-1 shows a schematic representation of
the climate system.   Changes in the amount of
energy  emitted  by  the  sun,  changes  hi  the
atmospheric composition (such as from volcanic
eruptions   and human  input  of aerosols  and
greenhouse gases),  and  changes in the  Earth's
surface  (such  as  deforestation)  can affect  the
Earth's  energy balance.  Atmospheric and oceanic
circulation can redistribute the energy.

    The radiative balance of the planet, as shown in
Figure 2-2, determines the  global average vertical
distribution of temperature.  If the concentration of
certain trace gases (carbon dioxide (CCO, water
vapor (H2O), methane (CH4), nitrous oxide (NjO),
tropospheric ozone (O3), and chlorofluorocarbons
(CFCs)) increases, the atmosphere's absorption of
longwave  radiation  (thermal  radiation  from  the
Earth's surface) will increase.  Some of this energy
will be radiated downward, heating the surface and
increasing the surface temperature.  Because the
concentrations of all these gases are projected to
increase in the future, this effect and its timing must
be compared  to  the other projected causes of
climate change  (forcings), and the response of the
climate  system, to  project  the  future  climate.
Uncertainties are associated with all these factors.

Solar Radia
SPACE ft |


('Clouds Kj
H20, N2, 02, C02, 03, etc. Air-Biomass
Aerosols Precipitation, -Land
Air-Ice Coupling Evaporation Coupling
4 ft Heat Exchange ft Wind Stress BIOMASS
T/^-JL™ ft L j :_ ****

jy 7 mKv^utft
Changes of \ Coupling ft Atmosphere-Ocean Coupling / LAND
Atmospheric Composition \ OCEAN ^ __/

Changes of Land Features,
Orography, Vegetation,
Albedo, etc.

Changes of Ocean Basin,
Shape, Salinity, etc.

Figure 2-1. The climate system. The principal interactions among components of the atmosphere, ocean, ice,
and land surface, and some examples of external forcings are indicated (Gates, 1979).

                                                    Global Climate Change

Chapter 2

RADIATION Shortwave Longwave
100 8 17 6 9 40
t f t \


Absorbed by ^/
Water Vapor. 19 -*^
Dust, O3
X^5^—. ,/
( 4 ^~)
Absorbed by
^v / / /
\V\ Backscattered / /
\\\bV/Air / /
\\ //
\ Reflected /
\ by Clouds /
\ ^^^\ /
\(ฃ* \/
\ Reflected
\ by Surface
>rbed \ / LONGWAVE
\ V

' t
Net Emission _r* ^TN
f^ '"5

Water Vapor. Emission
CO2,O3 bY Clouds
by Clouds 1
Water Vapor, |


* t Latent
Heat Flux

46 115 100 7 24

Figure 2-2. The Earth's energy balance.  If the average amount of solar radiation received by the Earth (342
watts per meter2) is represented as 100 units, then the amplitudes of the various components of the energy flux
are shown proportionately (MacCracken, 1985).

    Both  the past and future courses of climate
change are determined by a combination of external
forcings, unforced internal fluctuations, and  the
response of the climate system. This section briefly
discusses the forcings that will be important in the
next century.

Greenhouse Gases

    If the Earth  had no atmosphere, its average
surface  temperature, determined  by the balance
between incoming  solar  radiation and emitted
longwave radiation at the surface, would be about
0ฐF (-18ฐC), the same as the current temperature of
the moon. The average temperature is actually a
hospitable 59ฐF  (15ฐ C) because  of the natural
greenhouse effect of H2O, CO2, and O3. Because
a large  amount of the radiation in the wavelength
band 7 to  13 micrometers is not absorbed by these
gases, it is referred to as the "atmospheric window,"
and is a region where longwave radiation can escape
relatively unimpeded to space.

    The concentration of a number of trace gases in
the atmosphere is increasing as a result of human
activities. Because the trace gases are very effective
absorbers of longwave radiation in the atmospheric
window region, small (trace) amounts can have
large  effects  on the radiation balance, in effect
"dirtying" the atmospheric window.  Trends and
concentrations of some of these gases are shown in
Table 2-1 and Figure 2-3.  The projected relative
effects of these gases are shown in Figure 2-4.  Each
of the gases is discussed in more detail below.

Carbon Dioxide (CO2)

    Combustion of fossil fuels and deforestation are
increasing the concentration of CO2- Since Keeling
began   detailed   measurements   during   the
International Geophysical Year in 1958 at Mauna
Loa, Hawaii, the atmospheric concentration of CO2
has  risen   from  315  ppmv (0.0315%)  to  a

                          Table 2-1.  Trace Gas Concentrations and Trends
                                                                              Global Climate Change
 Current annual        Mid-21st
observed trends (%)     century

275.00 ppmv3 348.00 ppmv
  0.70 ppmv    1.70 ppmv
               0.34 ppmv
               0.22 ppbv6
               0.39 ppbv6
               0.13 ppbv6
0.29 ppmv
               0.08-0.10 ppbv6
              10.00-100.00 ppbv^
400.00-550.00 ppmv
  1.80-3.20 ppmv
  0.35-0.40 ppmv
  0.20-0.60 ppbv
  0.50-1.10 ppbv
aUnits of ppmv are parts per million by volume; 1 ppmv = 0.0001% of the atmosphere. Units of ppbv are
 parts per billion by volume; 1 ppbv = 0.001 ppmv.
bValue given is for 1986.
ฐTropospheric ozone only (below 12 kilometers).  Values (below 9 km) for before 1850 are 0 to 25% less than
 present-day; values (12 kilometers) for mid-21st century are 15 to 50% higher.
 Value given is for 1985.
Source: Ramanathan (1988), Lashof and Tirpak (draft 1989).
current level of 350 ppmv. About half of the CO2
put into the atmosphere each year remains in the
atmosphere, with the rest absorbed  in the ocean.
Because society's basic energy sources (combustion
of coal, oil, and natural gas) produce CO2, unless
strong energy conservation measures and shifts to
other energy sources take place, it is  projected that
the atmospheric concentration of COj will continue
to increase. As climate changes, the effectiveness of
the  oceanic  sink for CO2  may  also  change,
increasing or decreasing the fraction of CO2 that
remains in the atmosphere. CO2 contributes about
half of the total  anthropogenic greenhouse forcing.

Methane (CH4)

    Although the  methane concentration is now
increasing at a rate of about 1% per year and was
much lower during the ice ages, the basic cycle is
not completely  understood.  Sources include rice
paddies,  cows,  termites,  natural  gas  leakage,
biomass burning, landfills, and wetlands. Although
methane   has   a  much  lower   atmospheric
concentration than CO2 (currently 1.7 ppmv), it is
more effective at dirtying the atmospheric window
and  accounts   for   about   18%   of  current
anthropogenic greenhouse forcing.
                               Chlorofluorocarbons (CFCs)

                                   These completely anthropogenic gases, the most
                               important of which are known by the trade name
                               Freon, have been implicated not only in greenhouse
                               warming but  also in  chemical destruction  of
                               stratospheric ozone (O3). Because of this, nations
                               agreed to limit production of these gases  in  an
                               international agreement signed in Montreal in 1987.
                               The most important of these  gases  are CFC-11
                               (CFC13) and CFC-12 (CF^L). CFCs are used in
                               refrigerants,  aerosol  propeEants, foam-blowing
                               agents, and solvents. Substitutes for CFCs are being
                               developed that  are not  as stable chemically and,
                               therefore, would  not  accumulate as fast  in  the
                               atmosphere.  The resulting lower concentration
                               would produce  a smaller greenhouse effect and
                               would be less  effective at destroying O3.   The
                               current   fractional  greenhouse  contribution  of
                               CFC-11  and CFC-12  of 14%  would  probably
                               decrease  in the  future,  but  the  total  CFC
                               greenhouse  effect would most likely increase for
                               some time because of the long  lifetime of these

Chapter 2
                              AND ATMOSPHERIC SOURCES
                   ICE CORE DATA
              ATMOSPHERIC DATA


    I 1.2-

    g 1.0-

       1750            1850

    Source: Stauffer el al., 1985
             1978         1983

     Source: Blake & Rowland, 1988
             Mauna Loa and Ice Core Data
   — 330-
                                    x Mauna Loa
                                    v Ice Core

                                                    z | 340
                                                    O —
                                                           Monthly Concentrations of Carbon Dioxide
                                                           at Mauna Loa, Hawaii
1 290-
8 270-
250 •

40 1860 1980
Source: Neftel et al., 1985; Keeling et al., 1982

I 350


8 S 325
ฐ| 320
rt A/l/li
- AA/iM/l/F
i i I* i r
>5 1970 1985
Source: Keeling. 1984; Keeling, unpublished, 1988
1. 308
. ง 306

N2O ^
                  t '

    Source: Poarman et al..

     Source: Khahll, 1987
           Figure 2-3. Greenhouse gas trends in ice cores and atmospheric instrument data.

                                                                               Global Climate Change
Figure 2-4. Greenhouse gas contributions to global
warming; estimated values based on concentration
changes (1880-1980: Ramanathan et al., 1985; 1985,
1980s:  Hansen et al., 1988).
Nitrous Oxide (N2O)

    This gas, with both natural and anthropogenic
sources,  contributes about 6% to the enhanced
greenhouse effect, although its concentration is only
'about 0.31 ppmv. Its concentration is increasing at
a rate of about 1 ppbv per year, and sources include
oceans,  fossil   fuel  and  biomass  combustion,
agricultural fertilizers, and land disturbances.

Ozone (CO

    In addition to its role in the stratosphere as an
absorber of ultraviolet shortwave radiation, O, has
an  important impact  on climate.  This role is
complicated by its dependence on the altitude where
O3 occurs. Both ozone increases in the troposphere
and lower stratosphere and ozone decreases in the
upper stratosphere would tend to warm the surface.
Although the ozone concentration is believed to be
increasing in the troposphere, it is active chemically
and has highly variable concentrations in time and
space.  Responding to local air pollutants, such as
nitrogen oxides  (NO )  and hydrocarbons, ozone
provides a complex link between local air pollution
and global climate change.  Other  gases,  such as
carbon   monoxide  (CO)  and  volatile  organic
compounds,  also   play  important   roles   in
atmospheric  chemistry  and  hence affect  the
greenhouse problem.

Solar Variations

    The sun provides the energy  source for all
weather  on the  Earth,  and the balance between
incoming sunlight and outgoing longwave radiation
determines the  climate.  Small variations  in solar
radiation have  the potential for causing climate
changes  as  large  as those caused by  projected
increases of greenhouse gases. Precise observations
of the sun have been taken only for the past decade
(Willson and Hudson, 1988). They show, however,
that solar variations during this period have been so
small that they would not be important  compared
with the other  forcings discussed in this  section.
Since these  high-quality observations have been
taken only for a short period, they do not rule out
past  or  future variations of the sun that would be
larger.  But  on the time scale of centuries, solar
variations do not  now seem to be  an  important


    Large volcanoes can significantly increase the
concentration of stratospheric aerosols, decreasing
the amount of sunlight reaching the surface and
reducing surface temperatures  by several tenths of
degrees for several years (Hansen et al., 1978,1988;
Robock, 1978, 1979, 1981, 1984). Because of the
thermal  inertia of the climate system  (discussed
below),   volcanoes can  even  be responsible  for
climate   changes   over  decades.    It   has  been
suggested that a significant part of the observed
global climate change of the past 100 years can be
attributed  to the effects  of  volcanic  eruptions
(Robock, 1979).  Since large eruptions occur fairly
frequently, this component of  climate change will
have to  be considered when searching past climate
for a greenhouse signal and when projecting future
climate  change.

 Chapter 2
 Tropospheric Aerosols

     Natural sources, such as forest fires and sea
 spray, and human activities generate atmospheric
 aerosols in the troposphere.  The concentrations
 vary greatly in space and time, and local sources are
 important.   Furthermore,  these  aerosols  can
 produce either warming or cooling, depending on
 their  concentration,  color,  size,  and  vertical
 distribution.  It is not now possible to definitively
 determine their role in global climate.

 Surface Properties

     The Earth's  radiative  balance can  also  be
 changed by variations of surface properties. While
 interactions with the oceans which cover 70% of the
 Earth's  surface, are  considered internal to  the
 climate system, land surfaces can  exert a strong
 influence on the climate. Human activities, such as
 deforestation, not only provide a source of CO2 and
 CH4 to the atmosphere but also change the surface
 albedo and rate of evaporation of moisture into the
 atmosphere.    Detailed  land  surface  models,
 incorporating the effects of plants,  are now being
 developed  and  incorporated into climate model
 studies (Dickinson, 1984; Sellers et al., 1986).

 Internal  Variations

    Even with no changes in the external forcings
 discussed above, climate exhibits variations due to
 internal rearrangements of energy both within the
 atmosphere and between the atmosphere and the
 ocean. The total amplitude and tune scales of these
 variations are not well understood; this contributes
 to the difficulty of interpreting the past record and
 projecting the level of future climate change.

    Some  studies  suggest  that  these  random
 variations  can have amplitudes  and time scales
 comparable to climate changes  expected to  be
 caused  by greenhouse warming in the  coming
 decades (Lorenz, 1968; Hasselmann, 1976; Robock,
 1978; Hansen et al., 1988). A large El Nino, such as
 that observed in 1982-83, can take large amounts of
 energy out of the oceans and warm the surface
 climate  for a few years;  this warming is  then
 superimposed  on  any  warming  due  to  the
greenhouse effect. Our understanding of these El
Nino/Southern Oscillation variations is improving,
allowing us to account for this factor in interpreting
past global  climate change (Angell, 1988).

     Any imposed imbalance in the Earth's radiative
 balance, such as discussed above, will be translated
 into  a  changed  climate  through   feedback
 mechanisms that can amplify or decrease the initial
 imposed forcing.  A feedback in  which the final
 temperature is higher than what it would have been
 without  the   feedback  is  termed  a  "positive
 feedback."  If the effect of the initially imposed
 forcing  is  reduced,  it  is  termed  a "negative
 feedback."  This section describes several of these
 mechanisms that are internal to the physical climate
 system and that involve  the planet's  biology and

     Although   important   climate   feedback
 mechanisms have been  identified, we may not
 understand or even know about all the mechanisms
 involved in  climate feedbacks.  Figure 2-5 shows
 that even with the known physical climate feedbacks
 involved in  changing surface temperature, the
 potential  interactions are   complex.    Current
 state-of-the-art  climate   models   attempt  to
 incorporate most of the physical feedbacks that have
 been identified but are  forced, for example, to
 provide a very crude treatment for one of the most
 important ~ ocean circulation — because of large
 computer demands and inadequate  ocean climate
 models.   Another important  and  inadequately
 understood  feedback  — clouds —  has been the
 subject of  recent  climate  calculations  but,  as
 described below, is also treated crudely owing to
 inadequate understanding of cloud physics and the
 small  spatial  scale  on  which clouds  form  as
 compared with the resolution of the climate models.

Water Vapor -- Greenhouse Effect

    When the climate warms, more water (H2O)
 evaporates into the atmosphere from the warmed
 surface.  This  enhances the  warming because it
increases the greenhouse effect of the water vapor,
producing still  more  evaporation.   This positive
feedback acts  to approximately double imposed
forcings.  Thus, an important greenhouse gas, H2O
vapor,  is controlled by the climate system  itself.
Transformations of HLO between vapor and other
phases, liquid  and  soUd, provide other important
climate feedbacks discussed below.

                                                                              Global Climate Change
                                          NET ENERGY BALANCE
           Planetary albedo
                                             mospheric moisture
                                                capacity    t
                 V-*g=^//  \
                  \^m/./   \
                                      Atmospheric moistur
Figure 2-5. Physical climate feedback relationships. External forcings are indicated in underlined italics (Robock,
Snow and Ice

    When climate warms, snow and ice cover are
reduced, exposing land or ocean with a lower albedo
than the snow or ice. In addition, the albedo of the
remaining  snow  and  ice  is  reduced owing to
meltwater puddles and debris on the surface. This
acts to absorb more energy at the surface, further
enhancing the warming. This albedo feedback was
originally  thought  to  be the dominant  positive
feedback  effect of snow and  ice, but  we now
understand that the thermal inertia feedback of sea
ice plays a much more important role (Manabe and
Stouffer, 1980; Robock, 1983).

    The thermal inertia feedback acts  to increase
the  thermal  inertia of the  oceans when climate
warms by melting sea ice and exposing ocean waters
to the atmosphere.  Since imposed climate change
must then affect the ocean and atmosphere together
rather than  the  atmosphere alone, this acts to
reduce the seasonal cycle of surface temperature
and is  the prime reason for the enhancement of
imposed climate change in the polar regions in the
winter (Robock, 1983).

    Clouds respond  directly  and immediately to
changes in climate  and may represent  the  most
important uncertainty in determining the  sensitivity
of the climate system to the buildup of greenhouse
gases.  Fractional cover, altitude, and optical depth
of  clouds  can  change  when  climate  changes
(Schlesinger, 1985).  At the present  time, clouds
have a large greenhouse effect, but this is offset
(averaged over the globe) by their even stronger
cooling effect,  because  clouds reflect sunlight back
to space (Ramanathan et  al., 1989).   Since the
current greenhouse effect of clouds is larger than
the effect of an increase of CO2 by a factor of 100,
small changes  in clouds as climate changes can be
very important  in  affecting  the overall  climate
response to increases in trace gases.

    If climate becomes warmer,  more water will
evaporate into the  atmosphere.   Coupled  with
warmer  surface temperatures, this may produce
more upward motion of air, which would produce
more clouds.   One way clouds could increase is to

 Chapter 2
 increase in area. This would raise the albedo of the
 planet (except over polar snow and ice fields, which
 have an albedo larger than clouds), reflecting more
 sunlight back to space and having a cooling effect.
 Thus,  the  initially imposed warming is reduced,
 producing  a negative feedback.  Clouds already
 increase the planetary albedo from  about 17% (if
 there were no clouds) to 30% (Ramanathan et al.,
 1989). An increase of planetary albedo of only 0.5%
 would cut in half the warming imposed by doubled
 CO2 (Ramanathan, 1988).

     Other  studies suggest that, especially in the
 tropical  regions,   convection  could   increase,
 producing taller but narrower clouds.  This would
 produce additional warming in two ways: (1) by
 reducing the  cloud  area,  thus  decreasing the
 planetary albedo; and (2)  by decreasing the cloud
 top temperature and reducing longwave radiation to
 space.   This  mechanism would  be  a positive
 feedback.   In addition, convective  clouds in the
 tropical regions  (thunderstorms) tend to produce
 large shields of high cirrus  clouds,  which have a
 large  greenhouse  effect  further enhancing  the
 warming.   Cirrus  clouds allow much sunlight to
 penetrate because  they are so  thin,  but  the cloud
 particles absorb the  outgoing longwave radiation
 from the surface,  efficiently trapping much of it
 (Ramanathan, 1988).

    In the latest climate model simulations, it was
 found that clouds have a net positive feedback  on
 global climate (Schlesinger,  1988),  but  the final
 answer will  be known only after more research. It
 is not possible to be certain of the net effect of
 cloud feedbacks because of the complexity of clouds
 and their   response  to  climate  change.   The
 complexity is because all the above properties of
 clouds can change  simultaneously, because clouds
 affect  both  longwave and  shortwave radiation,
 because clouds affect precipitation (which affects
 land temperatures), and because the net  effect
 depends on the location  of the cloud, surface
 albedo, tune of day, and time of year.

 Biogeochemical  Feedbacks,

    In addition to  the physical climate feedbacks
 discussed   above,   a  number   of  positive
biogeochemical   feedbacks  may  be  important
 (Lashof, 1989).   These  feedbacks  can influence
future concentrations of greenhouse gases, especially
CO2 and CH4, through changes in sources and sinks
  of these gases induced by climate change, and they
 can  influence the climate change  itself through
 changes in vegetation,  and hence the surface heat
 and  moisture balance.  Such processes  include
 changes in releases of methane hydrates from ocean
 sediments, changes of land albedo due to shifting
 ecosystems, and changes in the ability of the oceans
 to absorb CO2 (this process is discussed in the next

     Methane  hydrates  are  combinations  of  a
 methane molecule trapped in a lattice of water
 molecules. They are found in ocean sediments and
 are stable under current pressure and temperature
 conditions in many ocean shelf  regions.   As the
 climate  warms,  these  conditions  may  change,
 releasing more methane into  the atmosphere and
 enhancing the greenhouse effect.

    As the climate warms, forests may shift closer
 to the pole, producing a region with a lower albedo.
 The  surface will thus  absorb  a larger fraction of
 sunlight,  warming  the  Earth and producing  a
 positive feedback, further enhancing the warming.


    Oceans play an important role  in the climatic
 response to changed forcings because they absorb
 and emit both heat and CO2, and because changing
 ocean circulation can change the redistribution of
 energy internal to the climate  system, as discussed
 above. When any of the above climate forcings are
 applied to the climate system, the climate will start
 to change. Since both the climate forcings and the
 climatic response are time-dependent, and since the
 climate system has a certain amount of inertia built
 in owing to the response times of the ocean, the
 exact relationship between the timing of the forcings
 and the timing of the response  is complex. Much of
 the lag  between the  imposed forcing  and  the
 climatic  response depends on the  oceans.   The
 upper 50 to 100 meters (164  to  328 feet) of the
 ocean, called the mixed layer, responds relatively
 rapidly to imposed forcings. The deep ocean is also
 important because its interactions could impose lags
 of as  much as 100 years.

    The relative depth and role of the mixed layer,
 as well as the circulation of the ocean, will change
in a complex way in response  to changed climate.
Broecker (1987) has suggested  that a rapid shift in
ocean currents, such as the Gulf Stream, may occur

                                                                               Global Climate Change
as the climate warms, producing large regional and
relatively  rapid  global  climate  changes.    In
preliminary  tests  with  the  Geophysical  Fluid
Dynamics Laboratory models, when CO2 is doubled,
the oceanic circulation around Antarctica changes so
as to increase the upwelling of cold bottom water.
As  a  result,   cooling  occurs in  the  Southern
Hemisphere high latitudes for a period  of several
decades as the rest of the globe warms! These two
examples  suggest that unforeseen climate events
may be possible in the future and that until  the
ocean response is well understood, the timing and
amplitude of the climatic  response to increased
greenhouse gases and the other forcings will need to
remain the subject of additional research.

    Oceans  are  also the  dominant  sink  of
atmospheric CO2, absorbing about half of all CO2
that is put into the atmosphere each year by the
combustion of  fossil fuels and deforestation.  The
amount of absorption is a strong function of oceanic
temperature, and shifts in oceanic circulation  and
temperature may shift the fraction of CO2 absorbed
in the future and, hence, change the  rate of CO2
accumulation in the atmosphere.  As the oceans
warm, they may absorb a  smaller fraction of the
excess CO2 in the atmosphere, thereby enhancing
the warming (Lashof,  1989).  In addition, oceanic
chemical  reactions  change  as  climate changes.
Oceanic production of dimethyl sulfide particles
could also change as climate changes  (Charlson et
al.,  1987).    These  particles  serve  as  cloud
condensation nuclei and may change the reflectivity
of marine  clouds  by changing the  number of
droplets in the clouds.

Observational Evidence  of Climate

     Thermometers have been  used to actually
 measure global climate change for more than 100
 years in enough locations to provide an estimate of
 how the planet's climate has changed during this
 period.  The most complete and up-to-date global
 surface air temperature record available is shown in
 Figure 2-6 (Wigley et al.,  1989).  Other analyses,
 including Hansen and Lebedeff (1988) and Vinnikov
 et al.  (1987), give   similar  results.   Problems
 common to all  these data sets include possible
 contamination from urban heat islands, inadequate
 spatial coverage  of   the Earth,  and  corrections
 necessary to counteract the effects of changing the
 methods used to take observations from ships.
    While the gradual warming seen in Figure 2-6
during the  past  century  is  consistent with the
increasing greenhouse gases during this period, most
scientists suggest that a clear link has not yet been
established between observed temperatures and the
greenhouse effect.  The large interannual variations
and the relatively flat curve from 1940 to 1975 show
that there are also other important causes of climate
change. For example, large volcanic eruptions, such
as Hekla in  1947 and Agung in 1963, and El Ninos
certainly have  produced  some  of  the variations
shown in this record.   Because of the projected
future emissions  of  greenhouse  gases,  global
warming is likely to dominate these factors during
the next century.

    The global temperature record shown hi Figure
2-6 can also be compared with the  record for the
United States for the same period shown in Figure
2-7 (Hanson et al.,  1989). While the globe as a
whole has been generally warming, the lower  48
states of the United  States  have  actually  been
cooling for  the past 40 to 50 years, although the
high  temperatures in  the 1980s are  among the
warmest on record. Since the lower 48 states of the
United States cover only 1.5% of the planet, this
indicates that regional climatic variations, which may
be caused by changes in  sea surface temperature
and wind circulation patterns, can be an important
factor in the climate of small regions of the Earth.
These factors  will continue   to be important  as
global climate warms.  For example, such regional
events as the midwestern drought of 1988 may be
related to changes hi ocean temperature (Trenberth
et al., 1988) and can be greater than the  effect of
greenhouse gases on a national or larger scale.

    On a longer time scale, proxy climate variables
 can  indicate how  climate   has  changed.   An
intriguing record comes from a core drilled in the
 antarctic icecap at Vostok and is shown in Figure
 2-8 (Barnola et al., 1987). The temperature record
 is deduced  from the deuterium isotope ratio. The
 past  CO2 concentration is actually measured from
 bubbles of ancient air trapped in the ice. The warm
 period of  the past 10,000  years  is  called the
 Interglacial and represents an anomalously warm
 period compared  with the   climate of  the past
 100,000  years. It is projected that because of the
 greenhouse effect, our climate will warm to a level
 much above even  the level  of the  Interglacial,
 warmer in  fact than the Earth has experienced for
 the  past million years.  The rate of warming will

 Chapter 2
                                                  1920     1940
Figure 2-6.  Hemispheric and global surface air temperatures, 1861-1988.  The 1988 value is preliminary and
includes data only through November. This record incorporates measurements made both over land and from
ships. The smooth curve shows 10-year Gaussian filtered values. The gradual warming during this period is not
inconsistent with the increasing greenhouse gases during this period, but the large interannual variations and the
relatively flat curve from 1940 to 1975 show that there are also other important causes of climate change (Wieley
etaL, 1989).
f 12-
S. 11-
I •

n nL
h A /\A A m / !/ W VW\A / W\ A n J\ n 1
i V
J .ป
T • 1 * '•
1 | * /V J J ^ f * i
95 1905 1915 1925 1935 1945 1955 1965 1975 1985
-800 3
-700 i
-650 ง
Figure 2-7. Annual average surface air temperature (solid) and precipitation for the contiguous United States,
1895-1987. Note that the United States has been cooling for the past 50 years (Hansen et al., 1988).

                                                                               Global Climate Change
also be unprecedented. From Figure 2-8, it appears
that the warming from the chill of the ice age 18,000
years ago to the Interglacial was very rapid, but in
fact a warming of even 2ฐC in one century would
be much faster than this warming.
               (Thousands of years before present)
 Figure 2-8.   Temperatures and  carbon dioxide
 concentrations for the past 160,000 years at Vostok,
 Antarctica.  Since these  observations were taken
 near the South Pole, they show larger temperature
 variations  (by a factor of 2 or 3) than took place
 averaged over the whole globe (Barnola et al.,
     Figure  2-8  shows  that  during the  entire
 160,000-year   period,   the   atmospheric   CO2
 concentration varied along with the  temperature.
 When it was warmer, the CO2 concentration was
 higher, although it never approached the current
 level of 350 ppmv.   It is not  known whether the
 climate  change  preceded  the increase  in  CO2,
 whether the increase in CO2 preceded the warming,
 or whether they both happened simultaneously.  It
 is well accepted that the changing orbit of the Earth
 produced   the   ice   ages   (the   Milankovitch
 Hypothesis), and this recently  discovered variation
 of CO2 certainly worked to enhance the climate
 changes  caused by  the changing orbit.   These
 natural processes are now being overwhelmed by
 the  human impact  of fossil fuel  burning and
    Two recent studies of CH4 concentration in
ancient  air found in Greenland and Antarctic ice
cores also have  shown  that CH4 concentration
varied with climate in prehistoric times (Stauffer et
al., 1988; Raynaud et al., 1988).  Although the CH4
concentrations were not  large enough to have an
appreciable impact  on the greenhouse effect,  the
CH4 did vary in the same sense as CO^ and climate
(see Figure 2-8).  The CH, variations indicate that
sources of CH4  increased in a warmer climate,
which suggests that natural sources of CH4 may also
increase in the future as  global  climate warms,
further amplifying the greenhouse effect.

    In many sciences, such as biology, chemistry, or
physics, it is possible to investigate new phenomena
by doing research in a laboratory.  In the field of
climate, this is not possible. One cannot bring the
Earth's climate  system into a room and perform
experiments   on  it,  changing  the  trace  gas
concentration or increasing the amount  of sea ice.
It is not possible to have two identical systems, one
a control and one that is changed to  compare the
outcomes. There is only one climate system, and
humans   are  now  performing an  uncontrolled
experiment on it by polluting it with CO2, CH4,
CFCs, and other trace gases.

     To try to understand  how the global climate
will  change  in response  to  human  activities,
researchers have applied various approaches.  The
climates of other planets,  particularly Venus and
Mars which are the most  Earth-like, can give us
some ideas about  climate under very different
conditions.  However, their atmospheres are not
similar enough to  Earth's to  give  us definitive
 answers about the next 100 years here. The history
 of the Earth's  climate is  another area we could
 study, but since many different forcings of similar
 strengths  have  been acting,  and  since the  data
 coverage is imperfect, it has not been  possible to
 definitively isolate the roles of the different forcings.
 Attempts have been made to use rotating tanks of
 water or other fluids (called dishpan experiments)
 as  models  for  the atmosphere,  but  these  are
 imperfect as  they cannot simulate realistic heating
 profiles or the detail of the real climate system.

  Chapter 2
      The  most  useful  tool  to investigate future
  climate  is  the computer model  of  the climate
  system.  In a climate model, the various physical
  laws  that   determine   the  climate,   such   as
  conservation of energy, conservation of mass, and
  the gas  law,  are  expressed as mathematical
  equations  that  specify the  relationship  between
  different variables, such as temperature,  pressure,
  wind, and precipitation.  By  specifying the various
  climate forcings,  it is possible to calculate the
  climate.   An  experiment can be performed by
  doubling CO2,  for instance, and  comparing the
  resulting climate to the current CO, concentration.
  Many theoretical calculations can be made to test
  the importance of various assumptions and various
  proposed feedback mechanisms.

     The   simplest  climate  model   is   the
  zero-dimensional global average model, which can
  be used to give a global-average measure of climate
  but cannot consider many important processes and
  cannot give regional distribution of climate  changes.
 Models that are one-dimensional  in  the  vertical,
  called  radiative-convective   models,   or   in  the
 horizontal, called energy-balance models,  are very
 useful for quickly and inexpensively testing various
 components of the climate system.  However, to
 calculate the location of future climate change, and
 to  incorporate   all  the   important   physical
 interactions, especially with atmospheric circulation,
 fully three-dimensional general circulation models
 (GCMs) are necessary.  These sophisticated models
 solve simultaneous equations  for all the important
 climate variables in three dimensions. The  world is
 broken up into a discrete grid of boxes  placed side
 by side and stacked to cover the globe.  The biggest
 and fastest supercomputers available are used, but
 computer speed and size constraints limit the size of
 these grid boxes to 250 to 1,000 kilometers (150 to
 600  miles) on  a side and to  a height of 1 to 5
 kilometers (0.6 to 3 miles). Thus, in one of these
 grid boxes,  all  the  complexity of weather  and
 horizontal variation is reduced to one number for
 temperature, one for cloudiness, and so  forth.  The
 equations used  to represent  the  physical  and
 chemical processes involved are also simplifications
 of real-world processes.

    Different climate modelers represent physical
processes hi different ways.  In all the models, the
radiation  schemes  attempt to account for  the
radiatively significant gases, aerosols, and clouds.
  They generally use different schemes for computing
  cloud height, cloud cover, and optical properties.
  The models also differ in their treatment of ground
  hydrology, sea ice, surface albedo, and diurnal and
  seasonal  cycles (Schlesinger and Mitchell, 1985).
  Perhaps the most important differences lie in the
  treatment of oceans, ranging from prescribed sea
  surface temperatures to "swamp" oceans with mixed-
  layer thermal capacity but  no heat transport, to
  mixed layers with specified heat transport, to full
  oceanic GCMs.  Models are constantly  becoming
  more   complex  and   sophisticated   as  new
  understanding  of  the physics  evolves and faster
  computers become available.

     One of the first experiments used to test any
  climate model is its  ability to simulate the current
  climate. In these tests, the various state-of-the-art
  climate models have differences.  Crotch (1988) has
  recently compared the simulations of surface air
  temperature and precipitation of four recent GCM
  simulations  and found that  although they do  a
 reasonable job  of simulating  global values,  the
 simulations at the regional scale are  poor.  He
 compared model simulations and observations on
 gridpoints, where  each  gridpoint  "represents a
 region of about 400 kilometers  (250 miles) by 400
 kilometers or  larger, or roughly  the  size  of
 Colorado, even though regions of this size may have
 very diverse local  climates"  (Crotch,  1988). He
 found differences between models and observations
 (see Table 2-2),  and between models, particularly
 for smaller regions. Crotch concluded that GCMs
 cannot  currently  project  regional  changes  of
 precipitation or temperature.

     Given the current state of the art, how can
 these models be used?  As discussed in Chapter 4,
 model simulations can be of use even in their crude
 state.  In the first place, even if the models do not
 exactly reproduce the current climate, perhaps the
 differences between their simulations of current and
 future climates  provide an estimate  of potential
 future changes.  In  addition, the models produce a
 data set of all  the variables needed  for impact
 assessment that are physically consistent within the
 physics of  the model.  Thus, although the actual
 model projections can not be taken as predictions of
 the future, they are useful in providing scenarios for
 impact assessment.  As model projections become
more accurate in the future, the scenarios they
generate will become  more accurate.

                                                                                Global Climate Change
Table 2-2.   Differences Between Winter and Summer Temperature Estimates for Four GCMs and Observed
Variable and model
                                                             Domain of  comparison
                                           North America
              Contiguous U.S.
                                                                                 Midwestern U.S.
Observed median
temperature (ฐC)
Difference in median
 Chapter 2

 Model Projections of a Doubled-CCX, World

     Several  climate  modeling  groups   have
 conducted  GCM experiments  to calculate  the
 equilibrium  climate  response to  doubled CO2.
 These include researchers at  the National Center
 for Atmospheric Research (NCAR), Oregon State
 University  (OSU),  NOAA's Geophysical  Fluid
 Dynamics Laboratory (GFDL),  NASA's Goddard
 Institute for Space Studies (GISS), and the United
 Kingdom Meteorological Office (UKMO).  The
 results from the different  experiments depend  on
 the assumptions made, especially on the treatment
 of clouds and of oceans.  The  models predicted
 global temperature increases  of 2.8 to  5.2ฐC and
 global precipitation  increases of 7 to  16%  (see
 Table 2-3).

     Attempts have also been made to  determine
 climate sensitivity from  past  data.   If we  could
 accurately determine the strength and timing of all
 the climate forcings  that have competed with the
 greenhouse effect in the past, we could account for
 them, and the residual warming would be a measure
 of the greenhouse effect to date. Unfortunately, our
 knowledge of both past climate change and  the
 responsible   forcings   is  too  poor  to  reliably
 determine the sensitivity of climate to greenhouse
 warming. Wigley and Raper (1987)  estimate that if
       all of the warming of the past 100 years were due to
       greenhouse gases, a doubling of CO2 would warm
       climate by about 2ฐC.   If, however, we allow for
       other   possible   forcings   (including   natural
       variability), for uncertainties in ocean  heat uptake
       and  the  timing of the climate response, and for
       uncertainties   in  preindustrial  greenhouse   gas
       concentrations  (Hansen et al., 1985;  Wigley and
       Schlesinger, 1985; Wigley et al., 1986), then from
       past  data we  can only  say that a CO2 doubling
       might produce a global climate change  anywhere in
       the  range  of  0  to  6ฐC  (Wigley,  personal
       communication). Wigley et al. (1989) point out that
       while the global warming of the past  137 years is
       highly significant statistically, it is not possible to
       definitively attribute this  warming  to a specific

           The actual path that the climate system would
       take  to approach the equilibrium climate would be
       determined by the tune  scales of the forcings and
       the various elements of  the climate system  and is
       referred to as the transient response.  Because the
       climate system response  lags behind the forcing, a
       built-in unrealized warming will always  occur in the
       future, even if no more greenhouse gases are added.
       Thus,  some   future  climate  response  to   the
       greenhouse gases that were put into the  atmosphere
       in the past will certainly occur, even if emissions
       were stopped  today.
Table 2-3. General Circulation Model Predictions of Globally Averaged Climate Change Due to Doubled CO2
 Surface air
increase (ฐC)
increase (%)












Source: Karl et al. (1989).

                                                                           Global Climate Change

Angell, J.K.  1988.   Impact of El Nino on  the
delineation of tropospheric cooling due to volcanic
eruptions.   Journal  of  Geophysical  Research

Barnett, T.P.  1985.  On long-term climate change
in observed physical properties of the ocean.  In:
MacCracken, M.C., and F.M. Luther, eds.
Detecting the Climatic Effects of Increasing Carbon
Dioxide.  Washington, DC: U.S. Department  of
Energy, (DOE/ER-0235), pp. 91-107.

Barnola, J.M., D. Raynaud, Y.S. Korotkevich, and
C.  Lorius.    1987. Vostok ice  core  provides
160,000-year record of atmospheric CO2. Nature

Chapter 2
Blake, D.R., and F.S. Rowland. 1988. Continuing
worldwide increase in tropospheric methane, 1978-
1987.  Science 239:1129-1131.

Broecker, W.   1987.  Unpleasant surprises in the
greenhouse? Nature 328:123-126.

Charlson, R.,  J. Lovelock,  M. Andreae, and S.
Warren. 1987. Oceanic phytoplankton, atmospheric
sulphur,  cloud  albedo  and  climate.    Nature

Dickinson, R.E. 1984. Modeling evapotranspiration
for three-dimensional global climate models.  In:
Hansen, J.E.,  and T.  Takahashi, eds.   Climate
Processes and Climate Sensitivity. Washington, DC:
American Geophysical Union, pp. 58-72.

Dickinson, R.E.  1986. The climate system and the
modelling of future climate. In: Bolin, B., B. Doos,
J. Hager, and R.  Warrick,  eds.  The Greenhouse
Effect, Climatic Change, and Ecosystems (SCOPE
29).  Chichester, England:  John Wiley  and Sons,
pp. 207-270.

Gates, W.L.  1979. The physical basis of climate.
In: Proceedings of the World Climate Conference.
WMO  No. 421.   Geneva,  Switzerland:  World
Meteorological Organization, pp. 112-131.

Grotch, S.L.  1988. Regional Intercomparisons of
General  Circulation  Model  Predictions  and
Historical Climate Data.  Washington, DC:  U.S.
Department of Energy. (DOE/NBB-0084).

Hansen, J., I. Fung, A. Lacis, S. Lebedeff, D. Rind,
R. Ruedy, and G.  Russell.   1988.  Global climate
changes as  forecast by the  GISS  3-D  model.
Journal of Geophysical Research 93:9341-9364.

Hansen, J., and S. Lebedeff.  1988. Global surface
air temperatures: update through 1987. Geophysical
Research Letters 15:323-326.

Hansen, J., G. Russell, A. Lacis, I. Fung, and D.
Rind.  1985. Climate response times: Dependence
on climate sensitivity and  ocean mixing. Science

Hansen, J.E., W.C. Wang, and AA. Lacis.  1978.
Mount Agung eruption provides test of a global
climatic perturbation. Science 199:1065-1068.
Hansen, K., GA. Maul, and T.R. Karl.  1989. Are
atmospheric "greenhouse" effects  apparent in the
climatic record of the contiguous U.S. (1895-1987)?
Geophysical Research Letters 16:49-52.

Hasselmann, K.  1976. Stochastic climate models.
I. Theory. TeUus 28:473-485.

Karl, T.R., H. Diaz, and T. Barnett. 1989. Climate
variations of the past century and the greenhouse
effect (A report based on the First Climate Trends
Workshop).  Rockville, MD:   National Climate
Program Office/NOAA.

Keeling, C.D. 1984. The global carbon cycle: What
we  know  and  could know from  atmospheric,
biospheric, and oceanic observations.  In: Institute
for Energy Analysis.  Proceedings: Carbon Dioxide
Research Conference: Carbon Dioxide, Science and
Consensus. Washington, DC:  U.S. Department of
Energy, (DOE CONF-820970).

Khalil,  MA.K.,  and  RA. Rasmussen.   1987.
Nitrous oxide: Trends and global mass balance over
the last 3,000 years. Annals of Glaciology 10:73-79.

Lashof, D.   1989.   The dynamic  greenhouse:
Feedback  processes   that  may  influence  future
concentrations  of  greenhouse  gases.  Climatic

Lashof, D., and D.  Tirpak, eds.   1989.   U.S.
Environmental Protection Agency.  Policy Options
for Stabilizing Global Climate.   Draft  Report to
Congress.  Washington, DC:  U.S. Environmental
Protection Agency, Office of Policy, Planning and
Evaluation.  February.

Lorenz,  E.N.    1968.    Climatic  determinism.
Meteorological Monographs 30:1-30.

MacCracken, M.C.   1985.  Carbon  dioxide and
climate change:  Background  and  overview.   In:
MacCracken,  M.C.,   and  F.M.  Luther,  eds.
Projecting the Climatic Effects of Increasing Carbon
Dioxide.  Washington, DC:  U.S. Department of
Energy, (DOE/ER-0237), pp. 1-23.

MacCracken, M.C., and F.M. Luther, eds. 1985a.
Projecting the Climatic Effects of Increasing Carbon
Dioxide.  Washington, DC:  U.S. Department of
Energy. (DOE/ER-0237).

                                                                              Global Climate Change
MacCracken, M.C., and F.M. Luther, eds. 1985b.
Detecting the Climatic Effects of Increasing Carbon
Dioxide. Washington, DC:  U.S. Department of
Energy. (DOE/ER-0235).

Manabe, S., and RJ. Stouffer.  1980.  Sensitivity of
a global climate  model to  an increase of CO,
concentration in  the  atmosphere.   Journal of
Geophysical Research 85:5529-5554.

National Research Council. 1979.  Carbon Dioxide
and Climate: A Scientific Assessment. Washington,
DC: National Academy Press.

National Research  Council.   1983.   Changing
Climate.  Washington, DC:   National Academy

National Research  Council.  1985.  Glaciers, Ice
Sheets, and Sea Level: Effect of a  CCylnduced
Climatic  Change.     Washington,   DC:     U.S.
Department of Energy. (DOE/EV/60235-1).

National Research Council.  1987.  Current Issues
in  Atmospheric  Change.    Washington,   DC:
National Academy Press.

Neftel, A., E. Moor, H. Oeschger, and B. Stauffer.
1985.   Evidence from  polar ice cores for  the
increase in  atmospheric CO2 in  the past two
centuries.  Nature 315:45-47.

Pearman, G.I., D. Etheridge,  F. deSilva, and PJ.
Eraser. 1986. Evidence of changing concentrations
of  atmospheric CO2, N2O, and  CH,  from air
bubbles in Antarctic ice. Nature 320:248-250.

Ramanathan, V.  1988. The greenhouse theory of
climate  change: A test  by  an inadvertent global
experiment.  Science 240:293-299.

Ramahathan, V., R.D.  Cess, E.F.  Harrison, P.
Minnis, B.R.  Barkstrom,  E. Ahmad, and D.
Hartmann.  1989.   Cloud-radiative forcing  and
climate: Results from the Earth Radiation Budget
Experiment. Science 243:57-63.

Ramanathan, V., RJ. Cicerone,  H.B. Singh, and
J.T. Kiehl.  1985,   Trace gas trends and  their
 potential role in  climate  change.    Journal of
 Geophysical Research 90:5557-5566.
Raynaud,  D.,  J. Chappellaz, J.M. Barnola, Y.S.
Korotkevich, and C. Lorious. 1988. Climatic and
CH. cycle implications of glacial-interglacial CH4
change in the Vostok ice core. Nature 333:655-657.

Robock, A.  1978.  Internally and externally caused
climate change.   Journal  of  the Atmospheric
Sciences 35:1111-1122.

Robock, A.  1979.  The "Little Ice Age": Northern
Hemisphere  average  observations  and   model
calculations. Science 206:1402-1404.

Robock,  A.   1981.   A  latitudinally  dependent
volcanic dust veil index, and its effect  on climate
simulations.     Journal   of   Volcanology  and
Geothermal Research 11:67-80.

Robock, A.  1983.  Ice and snow feedbacks and the
latitudinal and  seasonal  distribution  of climate
sensitivity.  Journal of the Atmospheric Sciences

Robock, A. 1984.  Climate model simulations  of
the effects of the El ChichSn eruption.  Geoffsica
International 23:403-414.

Robock, A.  1985.  An updated climate feedback
diagram.  Bulletin of the American Meteorological
Society 66:786-787.

Schlesinger, M.E.  1985.  Analysis of results from
energy balance and radiative-convective models. In:
MacCracken,   M.C.,  and  F.M.  Luther,  eds.
Projecting the Climatic Effects of Increasing Carbon
Dioxide.  Washington, DC:  U.S. Department  of
Energy, (DOE/ER-0237), pp. 280-319.

Schlesinger, M.E.  1988.  How to make models for
behaviour of clouds. Nature 336:315-316.

Schlesinger, M.E., and J.F.B. Mitchell. 1985.  Model
projections of the equilibrium climatic response to
increased carbon dioxide. In: MacCracken, M.C.,
and F.M. Luther,  eds.   Projecting the Climatic
Effects of Increasing Carbon Dioxide. Washington,
DC: U.S. Department of Energy, (DOE/ER-0237),
pp. 81-147.

Sellers, PJ., Y. Mintz, Y.C. Sud, and A. Dalcher.
 1986.  A simple biosphere model (SiB) for use
within general circulation models. Journal of the
Atmospheric Sciences 43:505-531.

 Chapter 2
 Somerville, R.CJ., and LA. Remer.  1984.  Cloud
 optical thickness feedbacks in the  CO2 climate
 problem.  Journal   of   Geophysical  Research

 Stauffer, B., G. Fischer, A. Neftel, and H. Oeschger.
 1985. Increase of atmospheric methane recorded in
 Antarctic ice cores. Science 229:1386-1389.

 Stauffer, B., E. Lochbronner, H. Oeschger, and J.
 Schwander.  1988. Methane concentration  in the
 glacial  atmosphere  was  only half  that of the
 preindustrial Holocene.  Nature 332:812-814.

 Strain, B.R., and J.D. Cure, eds.  1985.   Direct
 Effects of Increasing Carbon Dioxide on Vegetation.
 Washington,  DC:  U.S.  Department  of Energy.

 Trabalka, J.R., ed.  1985.  Atmospheric Carbon
 Dioxide and the Global Carbon Cycle.  Washington,
 DC:      U.S.   Department   of   Energy.

 Trenberth, K.E., G.W. Branstator, and PA. Arkin.
 1988. Origins of the 1988 North American drought.
 Science 242:1640-1645.

 Vinnikov, KYa., P.Ya. Groisman, K.M. Lugina, and
AA.  Golubev.  1987.   Variations in Northern
 Hemisphere mean surface air temperature over
 1841-1985. Meteorology and  Hydrology  No. 1:
45-55. In Russian.

White, M.R., ed.    1985.  Characterization  of
Information Requirements for Studies of  CO2
Effects:  Water Resources, Agriculture, Fisheries,
Forests and Human Health. Washington, DC: U.S.
Department of Energy. (DOE/ER-0239).
 Wigley, T.M.L., P.D. Jones, and P.M. Kelly.  1986.
 Empirical climate studies.  In:  Bolin, B., B. Doos,
 J. Hager, and R. Warrick, eds. The Greenhouse
 Effect, Climatic Change, and Ecosystems (SCOPE
 29).  Chichester, England: John Wiley and  Sons,
 pp. 271-322.

 Wigley, T.M.L., P.O. Jones, P.M. Kelly, and S.C.B.
 Raper.   1989.  Statistical significance  of global
 warming.  Proceedings of the  Thirteenth Annual
 Climate Diagnostics Workshop. Washington, DC:
 National   Oceanographic  and   Atmospheric
 Administration,  pp. A-l -A-8.

 Wigley, T.M.L.,  and S.C.B. Raper.  1987. Thermal
 expansion  of sea  water  associated with global
 warming.  Nature 330:127-131.

 Wigley, T.M.L., and  M.E.  Schlesinger.    1985.
 Analytical solution for the effect of increasing CO2
 on global mean temperature.  Nature 315:649-652.

 Willson,  R.C.,  and H.S.  Hudson.    1988.   Solar
 luminosity variations  in solar cycle  21.  Nature

 World  Meteorological  Organization.     1986a.
 Atmospheric  Ozone  1985. Assessment of  Our
 Understanding  of the Processes  Controlling Its
 Present Distribution and Change.  Global Ozone
 Research and Monitoring Project - Report No. 16.
 Geneva,   Switzerland:     World  Meteorological

World Meteorological   Organization.  1986b.
Report of the  International  Conference on the
Assessment of the Role of Carbon Dioxide and of
Other Greenhouse Gases in Climate Variations and
Associated Impacts,  Villach, Austria, 9-15 October
1985.  World Climate Programme Report No. 661.
Geneva,  Switzerland:     World  Meteorological

                                     CHAPTER 3
     A changecT-climate variability (defined in the
following section of this chapter) associated with
climate  change could significantly affect natural
resources.   However, lack  of information on
potential changes in climate variability has limited
the completeness of climate change impact studies
presented in this report.   It  is not possible to
definitively state how climate variability will change
with a changed climate because model results are
mixed. At this time, there is not a strong case for
altering the assumption of no change in variability
used in the scenarios for this report.

     Analyses of changes in climate variability for a
CO2 doubling  estimated by two general circulation
models  (GCMs) — Goddard Institute for Space
Studies   (GISS)   and   National  Center   for
Atmospheric  Research  (NCAR)  —   are   not
conclusive.  Some overall trends, but also  some
inconsistencies, are obtained when comparing the
changes in  climate variability associated with a
changing climate calculated by the two GCMs for
four U.S. regions.

•    The model results suggest that daily and year-
     to-year temperature variability could decrease
     and precipitation variability could  increase.
     However, the results for temperature are not
     statistically significant. Furthermore, the two
     models produce some inconsistent results.

•    Results indicate  that, the diurnal (day and
     night) cycle may be reduced in the summer,
     although  results  for  the  other seasons are

     To determine  the validity of the  variability
statistics of greenhouse gas-perturbed experiments,
investigators  examined  how  well  the GCMs
reproduce  present-day  climate  variability.    A
comparison of observed and model results for the
current climate for the two GCMs for selected U.S.
regions reveals interesting contrasts and similarities
regarding  the reproduction of climate variability.
Simulation of variability is reasonably  good  in
several cases.

•   Although some  discrepancies  exist  between
    actual   and   estimated   temperature  and
    precipitation values, the models simulate the
    seasonal cycles of temperature and precipitation
    reasonably well in the four regions investigated.

•   The  models   make   errors   (generally
    overpredictions) in predicting daily and year-to-
    year temperature and precipitation variability.

    Explanations  for some discrepancies, such as
why the daily temperature variances are  too high,
relate to how the  surface hydrology is modeled in
both  GCMs   (NCAR  and  GISS).    More
investigations of model results are necessary  to
improve understanding of future climate variability

    Global warming can change the variability of
climate.  Although less is known about variability
than about most other aspects of climate change, it
may have greater impacts on some systems  than
changes in average climate conditions.

    Variability  is  an inherent  characteristic  of
climate (Gibbs et al., 1975) and is closely related to
the concept of climate change. However, no clear
universally accepted distinction is made between the
terms  "climate  variability1 and  "climate change."
Both terms refer to fluctuations in climate from
some expected or previously defined mean climate
state.  Berger  (1980) makes the distinction that
climate  change refers  to a secular trend  that

 Chapter 3
 produces  a  change  in  the  average,  whereas
 variability refers to the oscillations about that mean.
 Distinctions can only be made relative to the time
 scales of concern. The climate change discussed in
 this report refers to a change from the mean global
 climate conditions we have experienced in roughly
 the past few centuries. On a longer time scale (i.e.,
 thousands of years), however, this climate "change"
 would be viewed as an instance of climate variability
 (i.e.,  as one  of  many fluctuations  around mean
 conditions prevailing over several thousand years).

     For the  purpose  of  this  report,  climate
 variability is defined as the pattern of fluctuations
 about some specified  mean value  (i.e.,  a  time
 average) of a climate element.  Hence, in regard to
 the  climate  change  considered  here,   climate
 variability refers to fluctuations of climate around
 the new mean condition that constitutes the climate
 change, and is expressed on time scales shorter than
 the time scale of the climate change.  For example,
 if it is assumed that the average annual global
 surface temperature will be 3ฐC warmer than it is
 currently, then the climate variability on a year-to-
 pattern of departures from this mean increase.

     One of the  main concerns regarding climate
 change is whether and how climate variability will
 change (i.e., will the pattern of fluctuations around
 the new mean at any given location be the same as
 that around  the  "old mean").   This concept of
 changing climate variabilities is illustrated in Figure
 3-1, which displays three simulated time series of
 daily maximum July temperature for  Des Moines,
 Iowa.   In all  three cases, the mean maximum
 monthly temperature is the same (i.e., 86.2ฐF), but
 the patterns of daily fluctuations about this mean
 differ significantly.  Changes in climate variability
 refer to the differences in these patterns.

     The causes of climate variability depend largely
 on time scales  and may be divided into two major
 categories: (1) those arising from internal dynamics
 that produce stochastic (random) fluctuations (and
possibly chaotic behavior) within the climate system,
 and (2) those arising through external forcing of the
system.  Table  3-1 summarizes different causes of
 climate variability on different time scales.  On very
long time scales (e.g., 100,000  years),  astronomical
factors   account  for  much  variability  (orbital
parameters in Table 3-1).

















	I    |
             1   5  9  13  17  21  25  29 33
Figure  3-1.    Simulated  July  daily  maximum
temperature time series at  Des Moines, Iowa. All
assume the same  average temperature  but use
different  statistical   estimates   (first-order
autocorrelation coefficient ) of variability (Mearns
et al., 1984).
    Variations of climate on a year-to-year basis
(interannual  variability)  can  arise from  external
forcings, such as volcanic eruptions, or from slowly
varying internal processes including, as part of the
internal   system,   interactions   between   the
atmosphere and oceans, soils, and sea ice fields.
These interactions can result in shifts in locations of
major circulation features  or  changes  in  their
intensity  (Pittock,  1980).    The  largest  effect,
presumably,  is due  to variations in sea  surface
temperatures, such as  those occurring in  El Nino
Southern Oscillation (ENSO) events.

                                                                                  Climate Variability

          Table 3-1.  Major Processes Involved in Climate Fluctuations for Different Time Scales
TIME SCALES related to Earth's history
Years 1010 109 108 1
INTERNAL (related to)
ci g.o galactic dust
'ฃ jo -g +j Sun's evolution
ซ^|^ solar variability
'S S .5 ฃ orbital parameters


          Source: Berger (1980).
     Daily variability of a nonperiodic nature largely
results from variations in synoptic scale weather
processes, such as high- and low-pressure cells and
upper-atmosphere wind  streams, which direct the
movement  of   such   features   (atmosphere
autovariation in Table 3-1) (Mitchell, 1976). These
features interact  with local topography to provide
location-specific variability.  (Variations caused by
these weather  processes  are largely stochastic and
internal to the climate system.)

     This report mainly discusses variations on time
scales of several years  or less  ~ that is, from
interannual to  daily variability. Climate variability
does not have  a  specific operational statistical
definition, but can be described by a constellation of
statistical properties other than the mean. The most
commonly used measure is the variance (which is
the mean of the sum of squared deviations from the
mean of  a time series) or its positive square root,
the standard deviation.

    Climate variability is experienced on an impact
level mainly through  the  occurrence  of  extreme
climate events.  The impact of extreme variability
may be the first indication of climate change. It is
important to note,  however,  that  change in the
frequencies of extreme  events (e.g., heat waves,
drought) is not synonymous with change in climate

    To illustrate this point, an example is presented
of a change in the frequency of heat waves in Des
Moines hi July, defined as 5 consecutive days in the
month with maximum temperatures exceeding 95ฐF.
Just changing the monthly mean of the series by
3ฐF, without changing variability  (as measured by
the standard deviation  and/or  autocorrelation),
increases the probability of experiencing a heat wave

 Chapter 3
in July from the  current level of 6%  to  21%.
However, the increase can be even more  dramatic
if the variability is altered as well as the mean.  By
increasing the persistence in the time series (i.e., the
day to day dependence of the daily temperatures) as
well as the mean,  the probability of a heat wave
increases from 6% to 37% (see Mearns et al., 1984,
for  further  details).    Hence,  changes  in  the
frequencies  of  extreme events will  occur with
changes in the  mean climate conditions, but this
change can be reduced or rendered more extreme
by changes in variability.*

     The  impacts  of  climate  change  on society
accrue  not necessarily  from the  relatively slow
trends in the mean of a climate variable, but rather
from the  attending  shifts in  the frequency of
extreme events.  This issue  has already received
some attention in the literature where the nonlinear
relationship  between changes  in the mean and
extreme events has been examined (e.g., Schwarz,
1977; Parry, 1978; Mearns et al., 1984). However,
less  is known about this factor  than about most
other aspects of climate change.

     For the purposes of climate impact analysis,
extreme  climate  events  may  be  considered
perturbations of climate that result in conditions
outside normal  ranges that  exceed some critical
threshold.   What  constitutes  "normal"  (i.e., the
averaging period) is, of course, a central issue in
defining extremes.

     Extreme events relevant  to  climate impacts
function on  different time scales, depending upon
the climate variable involved and the impact area of
interest. Thus, events can range from the length of
time  (in   minutes  and  hours)  that  minimum
temperatures  in Florida remain below  a critical
value, resulting  in damage to citrus crops, to the
length  of time  (in  months  and  years) that
precipitation  is  particularly  low  in California,
resulting in serious water shortages for industry and
agriculture. The probability of extreme events can
also vary considerably -- for example, from that of
extreme snowfall in the Buffalo, New York area
such as that  of the 1976-77 winter (P = 0.0002)
* Although the scenarios created for this study assume no
change in variability (see Chapter 4: Methodology) they do
assume, for example, increases in heat waves and decreases in
cold waves that result from changes in mean climate conditions.
 (Policansky,   1977),  to   that   of   heat  waves
 (temperatures above 100ฐF for 5  consecutive days)
 in Dallas, Texas (P = 0.38).

    What defines an event as extreme is not only
 a certain statistical property (for example, likelihood
 of occurring less than 5% of the time), but also how
 prepared a particular system is  to  cope with an
 event of such magnitude. Hence, very few extreme
 events have a fixed absolute  value independent of
 particular response systems at a particular location.
 This implies that what constitutes an  extreme event
 can also change over time because of  changes in the
 relevant response system (Heathcote, 1985).

    It  is  thus very difficult  to  comprehensively
 review all climate extremes of importance to society,
 and what is presented here is far from an exhaustive
 catalog. Because one of the purposes of this review
 is to highlight the extreme events of importance that
 can serve  as  guides for  choosing what  extreme
 events should be quantitatively analyzed in GCM
 experiments, priority is given to  events related to
 variables that can be relatively easily analyzed.

    This review  considers the two most important
 climate variables — temperature and  precipitation -
 - and their extremes (maxima and minima), and one
 type of major meteorological disturbance  —  severe
 storm effects. Extremes in these variables affect the
 areas  of  energy use  and  production,  human
 mortality  and   morbidity,   agriculture,  water
 resources, and unmanaged ecosystems (although not
 all areas are discussed under each  climate extreme).


    Given  the  scientific  consensus that  higher
 atmospheric concentration of greenhouse gases will
 raise   average  global  temperatures,   extreme
 temperature  effects  are   given  priority  in this

 Maximum Temperatures

    Extreme  temperature  effects  on   human
 mortality and morbidity have received the most
 attention in the scientific literature (e.g., Kalkstein,
Volume G; Becker and Wood, 1986; Jones et al.,
 1982; Bridger et al., 1976; Ellis, 1972). This is partly
because the relevant climate factors (i.e., maximum

                                                                                   Climate Variability
daily temperatures and relative humidity) are readily
available for analysis.

    A heat wave is defined as a series of days with
abnormally high  temperatures  (i.e.,  temperatures
exceeding  some  critical  threshold).  Examples
include  the 1980 heat  wave in the United States
when Kansas City had 17 consecutive days above
39ฐC (102ฐF) (Jones et al., 1982),  and Dallas had
42 consecutive days with temperatures above 38ฐ C
(100ฐF) (Becker and Wood,  1986). The death toll
that year was several times above normal (1,265

    Studies have  specifically tried  to  pinpoint the
most significant meteorological factors associated
with heat-related death and illness.   Jones  et  al.
(1982)   determined   that   high   maximum
temperatures,   the  number  of   days  that the
temperature is elevated, high humidity,  and low
wind velocity  contributed to excess  mortality  in
Kansas  City and St. Louis in the 1980 heat  wave.
Kalkstein et al. (1987) established that runs of days
with high minimum  temperatures,  low  relative
humidities, and maximum temperatures above 33ฐ C
(92ฐF) contributed  to heat-related deaths in New
York City.

    Increases in heat waves  are virtually certain,
assuming global warming. But how they increase
(longer  or greater departure from  the mean) very
much depends on changes in variability that would
affect the persistence of high temperatures.

    Such  crops as corn, soybeans,  wheat, and
sorghum are sensitive to high temperatures during
their bloom phases.  For example,  Shaw (1983)
reported that  severe temperature  stress  during a
10-day  period  around  silking  (a  critical period
during which the number  of kernels  on the ear is
determined) will result in crop failure.  McQuigg
(1981) reported that the  corn  crop was severely
damaged in July 1980 as a result of temperatures
exceeding 38ฐC (100ฐF). The destructive effects of
runs of hot days on corn yields were particularly
apparent  during 1983  in the  U.S. Corn  Belt.
Although the  damage  from  high temperatures is
best documented for corn, it has also been noted in
wheat   and soybean  yields  (e.g.,  Neild,  1982;
Mederski, 1983).

    Although  not  as  much research has  been
performed on the effects of temperature extremes
on  natural ecosystems, some research has been
done on forest responses to temperature extremes.
Solomon and West (1985) indicate in their summary
of climate effects on forests that the frequency,
intensity, and lengths of heat waves under climate
change conditions are important factors influencing
seedling survival and can contribute to the loss of a
species from an ecosystem.  A run of warm years
can affect the location of tree lines.  Shugart et al.
(1986) established that a period of warm summers
at high altitudes during the 1930s, when the mean
annual temperature was no more than 1ฐC higher
than average, resulted in a burst of regeneration in
boreal forest trees near polar and altitudinal limits
in North America.

    High temperatures have their most immediate
impact on energy by causing increased electricity
demand  for  air-conditioning.    Using  climate
scenarios  similar to  those  in  this  report  (see
Chapter 4: Methodology), Linder et al. (1987) found
that energy demand in New York would significantly
increase  in summer (on the  order of 3% for an
average August day in 2015 for the downstate area).

Minimum Temperatures

    Extreme  minimum  temperatures will not
necessarily be less of a problem with CO2-induced
climate warming.  For  example, changes will most
likely occur in the growing areas of certain crops,
where risks of frost damage may not be clearly

    The best example of frost damage to  crops is
the effect of low minimum temperatures on citrus
trees. This problem has been studied in depth for
the citrus crop in  Florida.  (See Glantz, Volume J,
for  a discussion  of the  Florida  citrus industry's
responses to freezes in the early 1980s.)  The most
striking  aspect  of these freezes is the very short
freezing time necessary for damage to occur. New
citrus growth (i.e., bloom buds) can be completely
killed during a 30-minute exposure to -3.3ฐC (26ฐF)
or a 3-hour exposure to -2.2ฐC (28ฐF).  The effect
of freezes  is exacerbated if the  crops have not
hardened with the cold. Thus, if a freeze follows a
warm period (i.e.,  indicating high daily temperature
variability) when dormancy has been broken, more
damage will occur  at less  extreme temperatures.
For example, the December 24-26,  1983, freeze
caused the Florida citrus yield to be 30% lower than
it had been the previous year (Mogil et al., 1984).

Chapter 3
    Extreme lows on a seasonal basis tend to most
directly affect winter energy use for heating.  In the
United States,  the  difference in heating fuel use
for a warm as compared with a cold winter can vary
by as much as 400 million gallons of oil. During the
extremely cold winter of 1976-77, heating degree
days (calculated on a base of 18ฐC (65ฐF)) were
10% greater than normal for the nation as a whole
(Dare, 1981).


    Anticipated changes in precipitation resulting
from climate change  are not well  known at this
point.    However,  geographic  shifts  in rainfall
patterns will   likely  occur.    Changes  in  the
frequencies of extremes of both droughts and floods
must be considered.

    Drought is  of particular interest at the time of
this  writing because  of the  1988  drought in the
United States and the energetic speculations being
made  concerning  its possible  connection with
CO2-induced climate  change (Wilford, 1988).  It
cannot be said  that the summer 1988 drought was
caused by CO2-induced climate warming, but rather
that such droughts would be possible and perhaps
more frequent with such, a warming. (In fact, most
recent evidence presented by Trenberth et al. (1988)
indicates that the cause of the drought was primarily
temperature anomalies in the Pacific (i.e., cool
temperatures  along  the  Equator  and  warmer
temperatures to the North), which led eventually to
the  anomalous displacement of the  jet  stream
northward.  These causes are  considered  to be
natural variations in the coupled atmosphere-ocean


    The most basic, general definition of drought
may be lack of sufficient water  to  meet essential
needs  (Gibbs,  1984).    From  a  more  strictly
climatological point of view, it may be considered a
condition determined relative to some long-term
average condition of balance between rainfall and
evapotranspiration in  a particular region (Wilhite
and Glantz, 1987).  Different types of drought are
recognized,  such  as  meteorological drought  (a
departure of precipitation from normal), agricultural
drought (insufficient soil  moisture based on crop
growth needs),  or hydrological drought (based on
departures from  normal  or  relevant hydrologic
parameters, such as streamfiow).  These "types" of
drought are not completely independent, but  can
show up at different time lags one from the other.

     Drought  of any  kind is  anomalous  as  an
extreme climatological event in that it is a "creeping"
phenomenon; neither its onset nor its end is clearly
punctuated  in  time.   It  is  difficult  to measure
drought severity, since drought is a combination of
factors:   duration,  intensity,  and areal  extent.
Drought also can be one of the longer-lived extreme
events  in that  it can  be measured in terms of
seasons or, more frequently, years.

     In the United States,  major droughts have
usually been defined in terms of several years,  and
the rate of occurrence is most strongly influenced by
interannual variability of precipitation.

     The effect of drought on  crop production is
perhaps the impact of drought that has received the
most research  attention.    The  occurrence  of
droughts  has  been a major  cause  for  yearly
variability in crop production in the United States
(Newman, 1978). During the 1930s, drought yields
of wheat and corn in the Great Plains dropped to as
much as 50% below normal, whereas the drought in
the 1950s brought less dramatic declines in yields
(Warrick et al.,  1975). In 1988, national corn yields
were  40%  below normal   (see  Chapter   6:

     Soil moisture deficits affect natural vegetation
as well as crops.  Much of the research in natural
ecosystems has been on forests.  Soloman and West
(1985)  identify  drought as the cause for death of
seedlings and  for slowed or stopped growth of
mature trees.

     Aside from the direct  effects  of  insufficient
moisture on unmanaged ecosystems, indirect effects
also  result  from  increased incidence of  fires.
During the drought of  1988, forest fires broke  out
across  the  country; the  most notable was  the
devastating August  fire in Yellowstone National
Park, which blackened 60% of its land area.

     The  effects  of  drought  on U.S.  energy
resources  are  most  apparent with  regard   to
hydroelectric power generation.  Linder et al. (1987)
discussed the effect of decreased streamflow due to
drought on the production of hydroelectric power in
New York (see Chapter 10: Electricity Demand).

                                                                                     Climate Variability
    The possibility of combined effects of higher
maximum temperatures and drought on electricity
demand and supply should  be noted.   Increased
demand (due primarily to increased temperature)
would very likely occur when drought would limit
generating capacity in regions such as  New York
and the Pacific Northwest.


    On average,  200 people  die each  year from
flooding; flash floods  account for  most of these
deaths (AMS, 1985). Floods also destroy property,
crops, and natural vegetation, and disrupt organized
social systems.

    Floods   result  from   a  combination   of
meteorological extremes (heavy precipitation from
severe   storms,   such   as   hurricanes  and
thunderstorms), the  physical characteristics  of
particular drainage basins, and  modifications  in
drainage basin characteristics  made  by  urban
development. Loss of life and property is increasing
as use of vulnerable floodplains increases.
   Major recent floods include the following:
        Rapid City, South Dakota (June 1972),
        231 deaths and more than $100 million
        in property damage;
   2.    Northeastern UnitedStates (June 1972),
        120 deaths  and about  $4 billion  in
        property  damage —  inundation from
        Hurricane Agnes;

   3.    Big Thompson Canyon, Colorado (July
        1976), 139 deaths .and $50' million  in
        property damage — a result of a stalled
        thunderstorm system that delivered 12
        inches (305 millimeters) of rain in less
        than 6 hours (Henz and Sheetz, 1976);
   4.    Johnstown, Pennsylvania (July 1977), 76
        deaths and $200 million in property
        damage -- a  result of slowly  moving
        thunderstorms that deposited 11 inches
        (279 millimeters) of rain in 9 hours.
      The recurrence  interval of flooding is most
 important  in  applying  effective  control  and
 protection mechanisms.   These include  building
 dams,   reservoirs,  and  levees,  and  improving
 channels and floodways (White  et al.,  1975).  For
 example, flood control reservoirs are designed to
 operate at a certain level of reliability,  and the
 reliability  is  determined  by  a  certain  flood
 magnitude that the reservoir  can handle, such as a
 100-year flood.  The statistics of flooding are vital
 for designing for protection  and are based on a
 certain  climate  variability determined from the
 historical record.   As that variability changes, the
 reliability of the protection system will change.

      Floods in the 1980s have been less serious in
 terms of loss of life,  but changing frequencies of
 severe   storms,  such   as   thunderstorms  and
 hurricanes,  as  well  as  general  shifting  of
 precipitation patterns could result in unprecedented
 losses from floods in a climate-changed world.

 Severe Storms - Hurricanes

      Three important kinds of weather extremes
 are present in hurricanes:  strong winds, intense and
 high precipitation  amounts,  and extreme  storm
 surges.   A hurricane is an  extreme  form of a
 tropical  cyclone, characterized by torrential  rains,
 typically as much as 127 to 254 millimeters  (5 to 10
 inches) in one storm;  high windspeeds, which can
 exceed  160 kilometers per hour (100 miles per
 hour); very steep pressure gradients, with pressure
 at the center as low as 915 millibars (27 inches);and
 diameters  of 160  to 640 kilometers (100 to 400

     Hurricanes are classified according to their
 severity on the Saffir/Simpson Scale  (categories 1
 through 5), taking into account the central pressure,
 windspeed,  and  surge.   Major hurricanes are
 considered to be all those  of categories 3 through 5
 wherein central pressure is less than 945 millibars
 (27.9 inches), windspeeds exceed 176 kilometers per
 hour (110 miles per hour), and the surge is greater
 than 2.4 meters (8 feet) (Herbert and Taylor, 1979).

     From 1900 through 1978,  53  major  hurricanes
 (averaging two  major hurricanes every 3 years)
directly   hit  the  United States.  Overall,  129
hurricanes  of any strength hit the United States
(averaging approximately two each year).  In recent

Chapter 3
decades, the  number of  major hurricanes has
declined. From 1970 to 1978, only three hurricanes
occurred,  compared with six  or more  in  earlier
decades. The last hurricane of category 4 or 5 to
strike the United States was Hurricane Camille in
1969. In 1980, Hurricane Allen, which at one time
reached  force  5, weakened before it struck a
relatively unpopulated segment of the Texas coast
(Oliver, 1981).  Since then, the population of the
south coastal regions of the United States has grown
tremendously, and most  inhabitants  have  never
experienced a major-force  hurricane.  Building in
coastal areas has also increased with population,
which raises the potential for high property damage.
Thus, the population may be more vulnerable and
less prepared to handle this particularly devastating
extreme event (Sanders, 1982).

   Any increase  in the frequency and/or intensity
of these storms,  which could result from climate
change, would  be of great concern to southern
coastal regions of the United States.  Hurricane
Gilbert,  which   occurred  in  September  1988,
reinforced this concern, even though it did not cause
major damage   to  the  coastal United  States.
Hurricane Gilbert may well prove to be the most
powerful hurricane of the 20th century;  its lowest
central pressure (883 millibars or 26.13 inches) was
the lowest ever measured in the Atlantic Gulf and
Caribbean   regions  of tropical storm  activity.
Serious damage did occur primarily in Jamaica, the
Cayman  Islands,  and the  northern  tip  of the
Yucatan Peninsula (Ludlum, 1988).

   Coleman (1988) has found in the historical
record  some  limited  evidence  for   increased
frequency for the number of storms formed in the
North Atlantic during years of warmer-than-average
sea surface  temperatures.   Emmanuel  (1987) has
found through a hurricane modeling experiment that
the intensity of hurricanes increases under warmer
conditions.   The extreme  intensity of Hurricane
Gilbert in September 1988 is  consistent with the
findings.     Emmanuel  (1988)  also  asserts the
importance  of  establishing a general   theory of
hurricane  development independent  of  current
atmospheric conditions, so that scientists can predict
changes in frequency and intensity of storms with
climate change.

Empirical Studies

     One of the methods available for gaining some
insight into how climate variability may change in a
generally warmer climate  is  to investigate  the
climate record for past relationships between mean
climate change and changes in variability. However,
past research  efforts  to  determine  changes in
climate variability and relationships with changes in
mean climate conditions have not resulted in a clear

     Van  Loon   and   Williams  (1978)  found
significant differences in interannual temperature
variability in North America during two different 51-
year periods.   However, they  found no single
connection between trend hi temperature and trend
in its interannual variability.  Specifically, they assert
that their results do not support  the  postulated
association between cold periods and high variability
of temperature.  Diaz  and Quayle (1980), in a
thorough analysis  of the U.S. climate (temperature
and precipitation), found no systematic relationship
between   changes  in   mean   temperature  and
precipitation and their corresponding variances.

     Brinkmann (1983)  analyzed the relationship
between  mean  temperature  and  variability in
Wisconsin using climate data for three stations.  She
found no relationship between mean temperature
and interannual variability, but did find a negative
correlation between winter mean temperatures and
the day-to-day variability,  and  a  corresponding
positive relationship for summer conditions. What
this means is that cold winters are more variable
than warm winters, but that cool summers are less
variable than warm ones. Brinkmann explains these
relationships on the basis of Wisconsin's location
with respect to general circulation patterns.

     Lough et al. (1983) analyzed the association
between  mean temperature and precipitation  and
variability in Europe by using the analog approach
to create climate change scenarios (the analog

 approach  is  further  discussed  in  Chapter  4:
 Methodology).  They  selected two periods when
 arctic temperatures were particularly warm and cold
 (1934-53 and  1901-20).  Results indicate that the
 regions  of lower  winter  temperatures roughly
 coincide with the region of increased variability, but
 the coincidence is far from perfect.

     These studies indicate that significant changes
 have occurred in both interannual and  day-to-day
 climate variability in historical times, but that simple
 or distinct relationships between changes in mean
 climate conditions  and changes in variability have
 not been established.    Moreover, the value of
 seeking such relationships in the past as a key to the
 future is potentially limited, since the causes of very
 short-term warming or cooling in the  past are not
 known, but in any event, are not caused by increases
 in greenhouse  gases.

     The failure of the analog approach to provide
 an empirically  consistent and  causally coherent
 scenario of possible changes in climate variability
 contributes to  the  necessity of examining climate
 variability in climate modeling experiments.  As
 discussed  in   Chapters  2  and  4, GCMs  have
 limitations,  but they have one clear strength over
 empirical attempts to analyze future climate change:
 the modeling experiments are constructed such that
 the response of the climate system to the true cause
 of the change  (increased greenhouse  gases in the
 atmosphere) is simulated.

 Modeling Studies

    Studies   comparing  variability statistics   of
 observed time  series with variability  statistics of
 GCM-generated  time series of  climate  variables
 relevant to climate impacts are not numerous in the
 atmospheric sciences literature,  although studies
 first appeared in the early 1980s (e.g., Manabe and
 Hahn,  1981; Chervin,  1981).   Such  studies  are
 critical  if climate change research is to determine
 whether the variability statistics  of doubled CO2
 experiments with GCMs are valid. To accomplish
 this, the ability of GCMs to reproduce  present-day
 climate variability statistics must be examined, and
 a thorough understanding of discrepancies must be

    Chervin (1986) used the National Center  for
Atmospheric Research Community Climate Model
 (NCAR CCM) to investigate interannual climate
                               Climate Variability

 variability and climate prediction.  He focused on
 the  additional variability  attributed to  external
 boundary conditions (i.e., in this modeling context,
 external  boundary  conditions refer  to important
 conditions  outside  the  atmosphere  that  cause
 changes  to the  atmosphere  but are not in turn
 affected by it, such as sea surface temperatures).
 He eliminated sources of external variability in the
 model, such  that discrepancies  between modeled
 and observed variability would reflect this external
 component.   The  variability of mean sea level
 pressure and 700-millibar geopotential height (which
 roughly corresponds to the height above the surface
 where  the   atmospheric  pressure   equals  700
 millibars, and is related to large-scale wind patterns)
 were analyzed for the Northern Hemisphere, with
 particular focus  on the United States.   Results,
 however,  indicated  no  significant  differences
 between modeled and observed variabilities of mean
 sea level  pressure over the United States and only
 limited areas of differences  in the  variability of
 700-millibar geopotential height.

     Bates and Meehl (1986) also used the CCM to
 investigate changes in the  frequency of blocking
 events (stationary pressure systems that block the
 flow of upper air currents hi the atmosphere) on a
 global  scale   under  doubled   CO,  conditions.
 Blocking  events are strongly  relatecf to persistent
 surface temperature anomalies, such as heat waves
 in  the  summer.    They found that the  model
 generally produces too few extreme blocking events.
 Under doubled CO2 conditions, standard deviations
 of blocking activity were found to mainly decrease
 in all seasons (i.e., the variability of blocking events

    Two studies were recently conducted on local or
 regional scales using the U.K. Meteorological Office
 five-layer  GCM.  Reed (1986) analyzed observed
 versus model control run results for one gridpoint in
 eastern England.  Compared with observations, the
 model tended to produce temperatures  that were
 too  cool  and variability that  was  too high as
 measured  by the   standard  deviation.     For
 precipitation, the model produced  too many rain
 days but did not successfully simulate  extreme rain
 events of greater than 20 millimeters per day.

    More recently,  Wilson and Mitchell  (1987)
 examined  the modeled distribution of extreme daily
 climate events  over Western Europe, using the same
model. Again, the model produced temperatures

Chapter 3
that were too cold, and hence, extreme minimum
temperatures were  overestimated.  This problem
was most pronounced in grid boxes away from the
coasts.   The model  also produced too  much
precipitation  in  general,  did  not   successfully
reproduce  observed  highest  daily  totals, and
overestimated the number of rain days. Wilson and
Mitchell examined  changes under quadrupled CO2
conditions and found that variability of temperature
generally decreased.

    Hansen et al. (1988) used the Goddard Institute
for Space Studies (GISS) general circulation model
to  simulate  the  global  climate   effects  of
time-dependent variations  of atmospheric  trace
gases and aerosols.   It was determined that the
model only slightly underestimates the  observed
interannual variability across the  globe. However,
the model's variability tends to be larger than that
observed over land (i.e., only considering land areas,
not ocean areas).

    Among the calculations made with output from
the transient run were changes in the frequencies of
extreme temperature events. This was accomplished
by adding the model-induced  temperature  change
with  climate warming to observed  local  daily
temperatures, assuming no change in variability.
Results indicate that  predicted changes  in the
frequency of extremes beyond the 1900s at locations
such  as  New York,  Washington,  and Memphis
become quite large and would have serious impacts.

    The  studies reviewed  above  indicate some
important shortcomings of GCMs  with regard to
their  ability to  faithfully  reproduce  observed
variability statistics.   More  research is  clearly
needed to further determine the sensitivity of the
models to  changes in physics, resolution,  and so
forth, with regard to the determination of variability.
Moreover,  only one  of these  studies  explicitly
concerns variables of importance  to climate impact
analysis.    Studying the higher  moments  (e.g.,
variance) of climate variable statistics,  and carefully
verifying the models' ability to reproduce observed
variability on regional scales, are the necessary
prerequisites  to   rigorously  analyzing  possible
changes  in  these  statistics under doubled  CO2

    Two research efforts were undertaken for this
report to attempt to increase knowledge concerning
how climate variability may change.  The climate
change scenarios used in the climate change impact
studies   reviewed   in   this   report   excluded
consideration of changes in variability (see Chapter
4:  Methodology).  The following two studies on
GCM estimates  of current and future variability
were performed for this report:

•   Variability  and  the  GISS  Model  -  Rind,
    Goldberg, and Ruedy, Goddard Institute for
    Space Studies (Volume I); and

•   Variability and the NCAR  Model - Mearns,
    Schneider, Thompson, and McDaniel, National
    Center for Atmospheric Research (Volume I).

    It should be recalled that scenarios of climate
change generated by the GISS GCM are used in
most  of the impact  studies  summarized  in this
report. The  results of these two studies are directly
compared in a later section.

The GISS  Study

    Rind et al. (1989) examined how well the GISS
GCM simulates the observed variability of climate
by  comparing the  model  and  the  observed
interannual and daily  variations of temperature and
precipitation. They described the model assessment
of changes in variability for these two major climate
variables, under climate  change using the GISS
doubled CO2 run (8ฐ  x  10ฐ  resolution) and the
transient climate change experiment in which trace
gases were increased gradually.   The analysis was
conducted for the Great Plains,  the Southeast, the
Great Lakes region, and California (see Figure 3-
2).    Observed data  consist  of the  average  of
observations at nine different stations per grid box.

    First, mean conditions were compared for actual
weather observations with the GCM control run (or
single CO2),  the doubled CO2 run, and the transient
run. The model values for mean temperatures for
four months  in the four regions are generally cooler
than observations (particularly in summer and fall),

                                                                                     Climate Variability
                         Figure 3-2. The locations of the four GISS model grids.
but  only by a  few  degrees  Celsius.   Model
precipitation values are fairly close to observed
values in the Great Lakes and Southeast grid boxes,
but model values are higher than observed for the
other two regions  (e.g., January in the southern
Great Plains: model =  2.1  millimeters per day,
observed  =  0.46 millimeters per day). Under the
doubled CO2 scenarios, temperatures increase over
the control run by 4 to 6ฐC (7 to 11ฐF) in the winter
and 3 to 4ฐC (5 to 7ฐF) in the summer.  Warming
in  the  transient  scenarios  is progressive,  but
temperature changes  more  gradually  than with
simply doubling the CO, amount.  Winter warms
more than  summer, and so the annual seasonal
cycle is reduced under climate change. Precipitation
changes are not statistically significant at individual
grids, but there is an overall tendency for increased

Interannual  Variability

    Standard  deviations   of  temperature   and
precipitation of observed and  modeled data were
compared for all  months.  In most months, the
model year-to-year temperature variability is similar
to the observed variability in the four regions, but in
summer the variability was overestimated by 0.3 to
0.6ฐC (0.5 to 1.1ฐF). Precipitation variability is
overestimated in half the cases where precipitation
amount is also overestimated.  The relative annual
variability of  precipitation (that  is, the standard
deviation relative  to  the  mean)  of the model is
generally in agreement with observations.

    Under  conditions of climate  change (doubled
CO2),  comparing  control  versus climate change,
there is generally reduced variability of temperature
from January through April.   Results  for other
seasons of the year  are  more  ambiguous.   For
precipitation, the doubled CO2 climate resulted in
increased variability hi most  months at the four
grids (hi 31  of 48 cases), but  was  particularly
striking at  the Southeast  grid.  These changes,
however, were often of the  same order  as the
model's natural variability (from examination of the
100-year  control run).  The  sign of the change in

 Chapter 3
 mean value and the sign of change in interannual
 variability are highly correlated.

 Daily Variability

    Daily variability of temperature was analyzed by
 taking the daily departures from monthly means and
 comparing the resulting model distribution with the
 distribution formed in the same manner  from the
 observational data

    Ten  years  of control  run  for the  transient
 experiment for four months (January, April, July,
 and October)  were compared with 30  years of
 observations.    Distributions of observed  versus
 modeled  daily temperature data were, in general,
 not significantly different.  Comparisons were also
 made  by  calculating the standard deviations of the
 departures  from  the mean for  the four months
 (Table 3-2). These results indicate that the model's
 values are significantly greater than the  observed
 values, which   demonstrates that  the model  is
 producing too many extremes.

    Results in Table  3-2, comparing  standard
 deviations, indicate that although changes with time
 are not strictly progressive, most cases by the end of
 the climate change experiment show reductions in
 the standard deviation although these reductions are
 not statistically significant.  (Note in Table 3-2 that
 standard  deviations for the future decades are
 changes in standard deviation (SD): model current
 SD minus future decade SD)  Since the results are
 not statistically significant,  a decrease  of daily
 temperature variability is not demonstrated.

    For   precipitation,  comparisons  are  more
 complex.  For example, the  number of observation
 stations used to represent a grid box does affect the
 results.    Model  rainfall   distributions  differ
 significantly from observed distributions in half the
 cases  (in three  seasons  for California  and the
 southern Great Flams). The model also produces
 fewer  days of  light rain in  general and  more
 extreme values  hi the winter  in all four regions
 (Table 3-3).

    In the transient experiment,  the precipitation
 distributions differ from the control  climate  about
one-fourth of the tune with no general progression
over the decades.  Figure 3-3 presents a sample set
of  distributions for precipitation during several
decades of warming for the West Coast in April. In
comparing standard deviations (Table 3-3),  the
 warmest tune period exhibits increases in standard
 deviations in half of the cases.  These results are
 again  consistent  with  those  for  interannual

 Variability of the Diurnal Cycle

    It would be expected that the diurnal cycle
 would  decrease under  changed  climate as  the
 additional greenhouse gases could limit nighttime
 cooling. Comparisons of control model results with
 observations are reasonable hi the four regions.
 Under  doubled CO2 conditions, it was found that
 the amplitude  of the diurnal cycle very definitely
 decreases in summer but changes inconsistently in
 the  other seasons.   The  reason for this  is  the
 dominance of radiative heating in the summer and
 of other forms of heating and cloud cover change
 in other seasons.

 The  NCAR Study

    In this  study, Mearns et al. (1989)  analyzed
 mean and variance of climate variable time series
 from selected empirical stations and those produced
 by general circulation model control and doubled
 COp  runs. They attempted first to determine how
 faithfully the GCMs reproduce these measures of
 the present variability and then to examine how the
 variability is estimated to change in CO2-perturbed
 cases. By comparing the relative performance (i.e.,
 model versus observations) of various versions  of
 the  NCAR  CCM (i.e.,  versions  with different
 physical parameterizations or formulations), Mearns
 et al. helped to determine what formulations maybe
 needed  for  forecasting  certain  measures   of
 variability and  how much  credibility  to  assign  to
 those forecasts.

 Methods                  .

    This study used the output from control runs of
 three different versions of the NCAR Community
 Climate  .Model  (CCM).   These versions  use
 different parameterizations of important physical
 processes in the model, such as surface hydrology.
 The Chervin version (Chervin, 1986) is the primary
 one used for comparison of observed and model
 control  output  (i.e.,  model runs to simulate the
 actual present-day climate), since it has the longest
 tune integration (20 years).

   The  CCM is  a  spectral  general circulation
model  originally  developed  by  Bourke   and

                                                                                   Climate Variability
                     Table 3-2. Daily Temperature Standard Deviations (SD) (ฐC)




location SD
Great Plains
West Coast
Great Lakes
Great Plains
West Coast
Great Lakes
Great Plains
West Coast
Great Lakes
Great Plains
West Coast
Great Lakes
* ASD = Change in standard deviation (model current - future decade).
Source:  Rind et al. (Volume I).
collaborators (Bourke, 1974; Bourke et al., 1977),
which has been modified by the incorporation of
radiation and cloud parameterization schemes. The
model has a resolution for physical processes (i.e.,
grid box size) of  approximately 4.5 degrees in
latitude and 7.5 degrees in longitude, and has nine
levels in the vertical.

    The other two versions of the CCM used are
the Washington version  (Washington and Meehl,

 Chapter 3
                     Table 3-3. Daily Precipitation Standard Deviations (SD) (mm/day)




Location SD
Great Plains
West Coast
Great Lakes
Great Plains
West Coast
Great Lakes
Great Plains
West Coast
Great Lakes
Great Plains
West Coast
Great Lakes
*ASD = Change in standard deviation (model current - future decade).
Source: Rind et al. (Volume I).
1984), which includes an interactive thennodynamic
ocean and surface hydrology; and the Dickinson
version  (Dickinson et al., 1986), a version of the
more sophisticated CCM1 containing a diurnal cycle
and a very sophisticated land surface package, the
Biosphere-Atmosphere Transfer Scheme (BATS).
This model calculates the transfer of momentum,
heat, and moisture between the Earth's surface and
atmospheric layers, and  includes  a very detailed
surface  hydrology  scheme  that   accounts  for
vegetation type and amount, and water use by the

                                                                                      Climate Variability
                       ,6.0   9.0   12.0   15.0
                       PRECIPITATION (mm/Day)
                       6.0    9.0   12.0    15.0
                       PRECIPITATION (mm/Day)
                                            18.0   21.0
^ 50
ฃ 40
| 30
"• 20


                8.0   12.0   16.0   20.0   24.0   28.0
                PRECIPITATION (mm/Day)
                 6.0    9.0    12.0   15.0
                 PRECIPITATION (mm/Day)
Figure 3-3.  Sample set of precipitation distributions for the West Coast in April for specified years of the
transient run (Rind et al., Volume I).
    The four regions of the United States chosen
for investigation were roughly the same as those
chosen for the GISS  study:  the Great  Plains
(GP;  represented  by  three  grid   boxes),  the
Southeast  (SE), the Great Lakes (GL),  and the
West Coast (WC). The locations of the grid boxes
and observation stations are indicated on Figure 3-4.

Comparison of Observed versus Chervin Control

    Four variables deemed particularly relevant  to
climate  impact  analysis  were   chosen for  this
analysis:   daily  mean temperature,   daily  total
precipitation,  mean  daily relative humidity,  and
mean daily absorbed solar radiation.

       Figure 3-5 displays the time  series  of daily
   average temperature for modeled and observed data
   for the  four regions  investigated.   The  model
   successfully simulates the annual cycle for the four
   regions, which represents the seasonal variability.

   Solar Radiation and Relative Humidity

       Simulation of solar radiation ranges from very
   good (the Great  Plains region) to  only fair at the
   Southeast,   where   the   model   consistently
   overestimated absorbed solar radiation during all
   months.  The Chervin CCM is poor at simulating
   the annual cycle of relative humidity at all four

 Chapter 3
                     • Temperature and
                       Precipitation Stations
                     o Relative Humidity and
                       Radiation Stations
                             Figure 3-4.  NCAR mbdel grid cells and station locations.
            g  15

            3  5

            1  ซ


            I  "
            i  10
            &  5
                             GREAT PLAINS I, II, III
                             120     180     240     300
                       60     120     180    240     300

g  15
 .  15
 — 10
                                                                                   GREAT LAKES
                                                                          60     120
                                                               180     240
                                                            WEST COAST
                                                                          60     120
                                                                             300    360
                                                                         	1	I	1	11
                                                                                             240     300     360
 rigure 3-5. Average temperature for a 20-year average year (NCAR model and observations) (Mearns et al.,
Volume I).

                                                                                     Climate Variability

    The  Chervin CCM consistently overestimates
precipitation, although  the  seasonal  cycle is  well
simulated in the Great Plains region and the West
Coast  grid.  The authors do not  know  why the
model overestimates precipitation, but  speculate
that it may partly be a result  of  a  precipitation
parameterization criterion of 80% relative humidity^
Variability Comparisons of the Chervin CCM

    Interannual  variability  of  temperature  is
generally underestimated  by the  Chervin CCM in
all   four   regions.    Interannual  variability  of
precipitation (i.e., relative variability, the standard
deviation relative to  the mean) is  generally  in
reasonable agreement with observed data, although
it  is  occasionally  overestimated.    This  is  a
particularly encouraging result for the credibility of
predicting climate changes, given how inaccurate the
control precipitation results are in terms of absolute

    In terms of daily variance, the model's relative
humidity  tends to  be much  less  variable  than
observed values at all locations and in most months.
Results  for  temperature for  January  and  July
indicate   that  the   Chervin   model   generally
overestimates daily temperature variance.

Intercomparisons  of Three CCM  Versions  and
Observed Data

    Comparing different model versions' simulations
of present-day climate facilitates understanding of
the  possible ranges of errors and  the effect  of a
model's  structural differences.   The present-day
climate  runs  of  models  incorporating physics
different from those of the CCM version of Chervin
(1986) are compared.  Both the Washington and
Dickinson runs consist of 3-year integrations.

    There is considerable  variability in how well the
models reproduce mean total precipitation for the
four grids, ranging from the relatively good results
of  Dickinson's  model,  to the  fair  results  of
Washington's  model,   to the overestimation  of
Chervin's model. On the  basis  of mean annual and
seasonal  comparisons, no  one  model  is  clearly
superior  to   the   other  two   in   accurately
reproducing  mean   climate  (temperature   and
precipitation) at the four locations.

    The   Dickinson   model   most   accurately
reproduces daily variability of temperature, while
the other two models overestimate it. This result is
graphically illustrated in the temperature histograms
(three models and observed) for two key months for
the Southeast grid (Figure 3-6).

    The reasons for these discrepancies have yet to
be  explored  in depth, but are  likely related to
different land surface packages in the models. A
possible   explanation  for  the  lowered   daily
temperature  variability of  the  Dickinson model
concerns  the  more  sophisticated surface energy
balance used, which includes consideration of soil
heat capacity.

Control Versus CO2-Perturbed Runs

    The authors included a preliminary analysis of
changes in precipitation and temperature, under a
scenario of doubled CO2, using the output from
Washington's control and doubled CO2 runs for the
four regions.  Interannual  variability could not be
analyzed because  the time series  are too short.
However, they  examined  the daily variability of
temperature and precipitation.

    An annual  temperature increase of about 2 or
3ฐC (4 to 5ฐF) occurs at  all locations.  Annual
total  precipitation increases between 22 and 26%
at three locations but decreases slightly (2%) in the
Southeast.  There are also potentially important
changes in the seasonal distribution of precipitation.
For example,  at  the Southeast grid a smaller.
percentage of the annual  total  occurs during the
summer in the CO2-perturbed  case (from 13 to/
6%).        :                                 ;

    Statistics  comparing  the  daily temperature
variance of the control and  perturbed runs for
January, April, July, and October indicate that the
temperature   variance  in  general   does  not
significantly change (at the 0.05 level of significance)
at  these  four  grids.   Without  consideration  of
statistical significance levels, results are mixed with
both increases and decreases.

    The percentage  of rain days decreases in the
summer under climate change in three of the four

Chapter 3

2 60
M 40






•30 -20 -10 0


" X
 i • i -i — i — i — i — i — i — i — i — r~
10 20 30 40 -30 -20 -10 0 10


^ X

i ^
s /




30 40

^ 60
| 40





-30 -20 -10 0




80 -

2 60 -
85 40 •
20 -


i i i -i 	 1 	 1 	 1 	 1 	 1 	 1 	 1 	 1 	 1
10 20 30 40 -30 -20 -10 0 10







30 40
Figure 3-6. Histograms of daily temperature, observations and three model versions, for two key months of the
Southeast grid (Mearns et al., Volume I).
grids.  Overall, there is a tendency for increased
daily precipitation variability at the four locations,
based  on  analysis  of  precipitation distribution

    It is difficult to compare the two studies.  The
modeling experiments were conducted partly with
different purposes in mind using two  different
models  (which differ not only in how physical
processes  are modeled but  also in their  spatial
resolutions). They also use different qualitative and
statistical methods for making comparisons.  The
GISS experiment was aimed primarily at examining
the changes  in variability with  climate change,
whereas the  immediate purpose  of the NCAR
experiment   was   primarily   to  examine  and
explain discrepancies in variability between model
control runs and observations.  Since the  spatial
resolutions of the models differ, the grid boxes of
the models do not coincide,  and so the regions
analyzed differ.   These  are  only some of the
problems  that  would affect  these  comparisons.
Nevertheless, an attempt is made here to compare
some of the results that roughly coincide.  Some
regions, such as the Great Lakes grids, coincide
fairly well  (see Figures  3-2 and 3-4), and some
similar analyses were conducted.

    A brief comparison is made of how the models
reproduce the observed mean climate. In general,
the GISS model is too cool and the NCAR model(s)
too  warm.    The  GISS  model  overestimates
precipitation at two grids, and  the Chervin version
of the NCAR model overestimates precipitation at
all grid boxes (although this is not true of two other
versions of the  NCAR CCM).

                                                                                   Climate Variability
    The following sections compare the observed,
control, and perturbed runs of interannual and daily
variability of temperature and precipitation. Table
3-4  summarizes  the comparisons between  the
modeled  control   runs  and  observations  for

Interannual Variability

    Rind et al.  used a 100-year control  run for
interannual   variability   calculations.      Their
observational data set consists of 30 years (1951-80).
The NCAR study uses a 20-year control run of
Chervin (1986) and a 20-year observational data set
(1949-68). The differences in sample size should be

    Table 3-5 presents the relevant results, winter
and summer  standard deviations for temperature,
and annual coefficients of variation (i.e., a measure
of relative variability) for precipitation for  the four
regions for both studies.  Relative variability values
(standard deviation relative to the mean) for the
GISS study were provided by  its authors (Rind,
personal   communication).      Both   models
overestimate  the temperature  variability of  the
Great Plains  region in winter.   (However,  the
difference in the NCAR study was deemed to be
statistically   insignificant.)      Both   models
underestimate the temperature variability  (but the
NCAR model much more so than the GISS) for the
                              West Coast winter.  In summer, the GISS model
                              overestimates, and the NCAR model underestimates
                              temperature variability at all locations.

                                  Regarding the relative variability of precipitation
                              (measured by the coefficient of variation),  the
                              results for the two models are rather similar. The
                              differences between observed and model values are
                              very close (from 1 to 6 percentage points) in each
                              study.  The NCAR model slightly underestimates
                              the variability at  each  location, whereas the slight
                              errors in the GISS results are mixed.

                                  The reasons for the lack of agreement  in the
                              two studies are far from obvious, and speculation
                              can only be rough. Certainly the difference in how
                              the atmosphere-ocean interaction is modeled may
                              play a role (i.e.,  the NCAR model uses fixed sea
                              surface temperatures,  whereas  the GISS  model
                              computes sea surface temperatures from a simple
                              ocean mixed-layer model).

                              Daily Variability

                                  Daily  variability  of  temperature   can  be
                              compared for two season months (January and July)
                              at the four locations using the standard deviations
                              (Table 3-6).    Because   of   certain problems
                              concerning  necessary  statistical assumptions for
                              quantitative  testing, these comparisons must be
                    Table 3-4.  Variability Results for Control Runs vs. Observations3
Temperature    Precipitation            Temperature     Precipitation
                (Relative/Absolute)13                      (Relative/Absolute)
aValues in chart refer to how the model estimates compare to the observations.
bRelative/absolute refers to comparison of coefficients of variation (relative) and standard deviation (absolute).
ฐChervin version of the NCAR model.
dValues are good or slightly low for the Dickinson version of the NCAR model.

 Chapter 3

 Table 3-5.   Interannual Standard Deviations, Temperature and Coefficient of Variation, Precipitation, GISS,
              and NCAR Control Runs
      Model and region
                                           Temperature (ฐC)
                                           standard deviation
   coefficient of
   variation (%)
(standard deviation/
    GISS fn =  1001
NCAR rn -
SGP = Southern Great Plains; SE = Southeast; WC = West Coast; GL = Great Lakes; GP = Great Plains.
Source: Rind, personal communication; Mearns et al. (Volume I).
viewed strictly qualitatively.  In seven of the eight
cases,   the   studies   agree   that   the   models
overestimate daily temperature variability.

    In _ both   studies,  explanations  for  the
overestimations are  related to  the  modeling of
surface  hydrology  (i.e.,  both  models   fail  to
completely   account   for    important
surface-atmosphere interactions that would tend to
                reduce daily temperature variability). (The relative
                success of the Dickinson version of the CCM in
                reproducing daily temperature variability partially
                supports such an explanation, since it has a more
                sophisticated surface hydrology scheme  compared
                with the Chervin version.)

                   The models produce, in the majority of cases,
                too few light rain days. The GISS model produces

                                                                                    Climate Variability
                       Table 3-6. Daily Temperature Standard Deviations (ฐC)
Obs.  Model
                                                                    Obs.  Model
     Great Plains
     Great Lakes
     West Coast
     Great Plains
     Great Lakes
     West Coast
Source: Rind et al. (Volume I); Mearns et al. (Volume I).
too many  extreme rain  events  in  winter  at  all
locations. The NCAR model tends to produce too
many high extremes in all four seasons.  Neither
study accounts for these discrepancies.

Comparison of Climate Change

    Comparison of climate change results of the
two  models  is  restricted  to  changes in  daily
temperature   variability  and   daily  precipitation
variability for four months for the four locations,
since  the  NCAR  study  includes  a quantitative
analysis of only daily variability change.

    The two studies do not agree on the direction
of change  of daily temperature variability.  The
NCAR results are mixed, showing both increases
and decreases, although most  of these changes are
statistically insignificant.  Rind et al. conclude that
in general, there is a decrease in daily temperature
variability  on the  basis  of changes in standard
deviations  (but  the changes  are not  statistically
significant).   On the basis of the  two research
reports,  no clear statement may be made about
changes in daily temperature variability under CO2
warming conditions.
                           A  slightly  clearer  picture  is  gained from
                        comparison of results for daily precipitation.  The
                        results of both  models point to  increased daily
                        precipitation  (although  not  from  analysis of  the
                        same  statistic).   This is not true for all locations
                        during all seasons, however.

                           Table 3-7   summarizes  the  very  tentative
                        conclusions that can be drawn  given all climate
                        change  results  regarding  changes  in  climate
                        variability from the GISS and NCAR studies.  The
                        degree of uncertainty in these conclusions should be
                        noted, as should the observation that many of the
                        results are from only one model (GISS).

                        Limitations of the Two Studies

                           Both  studies  underline  the  importance  of
                        viewing the climate change results of the models in
                        the context of how well they reproduce the present
                        climate. Model deficiencies can be expected to limit
                        the reliability of climate  change results, and faith in
                        quantitative results is probably misplaced.

                           A major model deficiency is inability to resolve
                        subgrid-scale   atmospheric   phenomena   that

 Chapter 3

     Table 3-7. Summary of GISS and NCAR Model "Scenarios" for Direction of Variability
               Changes from Present Climate to Doubled CO2 Climate for Four U.S. Regions3
                                          Variability Results
                                         CO2-Perturbed Runs
  Question marks indicate degree of uncertainty:
     ? -  results of only one model;
     ?? =  results of two models, but some conflicting results.
 contribute to climate variability, such as fronts and
 intense  cyclones   (hurricanes),  and  important
 variations in atmosphere-ocean coupling, such as El
 Nino  Southern  Oscillation  (ENSO)   events.
 (However, it appears that more sophisticated GCMs
 incorporating complete  ocean models do produce
 ENSO-type events (Meehl, 1989).) However, model
 results do give crude estimates as to the importance
 of some physical processes responsible for variability
 and what must be done  to improve them.  Further
 testing is needed to determine how the models'
 deficiencies  in  reproducing  present-day  climate
 affects "predictions" for  a  CO2-warmed future

    As  indicated in the  second section of  this
chapter, virtually all systems affected by climate are
affected by climate variability, although some are
more affected than others.  The relative importance
of climate variability and changes in variability, as a
result of climate change, to particular impact areas
is reflected in the results and limitations of some of
the studies summarized in  this report.

    Of greatest concern is  the lack of information
regarding changes in the variability of temperature
and precipitation that would attend climate change.
 The  lack  of this  information  resulted  in the
 formation of climate scenarios wherein the temporal
 variability of both precipitation  and temperature
 were not changed (see Chapter 4: Methodology).
 This was considered a limitation or concern in many
 studies, some of which are discussed in this section.

    In the Johnson et al. study on agricultural runoff
 and leaching (reviewed hi Chapter 6: Agriculture),
 the results were considered to be limited by the
 failure to consider changes in storm frequency and
 duration that would result from  climate change.
 The results  of this study could be vastly different
 from those presented, depending upon assumptions
 concerning precipitation duration, frequency,  and
 intensity, all of which would change if a changed
 daily variability were assumed.

    Several studies on hydrology summarized in this
 report also are highly dependent upon assumptions
 about precipitation variability.  These  include the
 Lettenmaier et  al.  study  on the  hydrology  of
 catchments hi the Central Valley and the Sheer and
 Randall study on  the impact of climate scenarios on
water deliveries,  both reviewed  in Chapter  14:
 California. The scenarios assumed that  the number
 of days of rainfall remains the  same under  the
 climate  change.    Model  results in  terms  of
predicting runoff amounts would be quite different
if  more  rainfall  events  of lower intensity were

                                                                                   Climate Variability
assumed compared with the same number of rainfall
events of (generally) higher intensity.

    The studies for the Southeast (Chapter 16) did
not consider changes in the frequency of droughts
or severe  storms such as hurricanes, which could
certainly affect the likelihood of flooding for some
coastal communities. However, these concerns are
considered to be secondary to changes in sea level
that would dominate  in terms of changing  the
likelihood of floods.

    Crop  yields  are  very  dependent   on daily
variability.  For example, heat waves  occurring
during the grain filling process lower wheat yields.
Whether  a drought occurs early or late  in  the
growing season  has differential effects  on yields.
Changes in variability were not considered in the
Rosenzweig, Peart et al., Ritchie, and Dudek studies
(see Chapter 6: Agriculture).

    Changes in the frequencies of extreme events
are  considered  to be  of  great  importance  to
potential forest disturbance, as discussed in Chapter
5: Forests.   The possibility  of increases  in  the
frequencies of events such as droughts, flooding,
wind,  ice, or snowstorms  may  be of greater
significance to forest survival than the gradual mean
change in climate that has been studied so far.

    The  Kalkstein study,  which  is reviewed in
Chapter 12: Human Health, is strongly dependent
upon  the determination  of  certain  maximum
temperature threshold values beyond which human
mortality increases. In applying the death/weather
effects statistical models to  scenarios of climate
change,  Kalkstein held temperature  variability
constant,  so that temperatures  that exceed  the
threshold  values are determined unrealistically.

    Changes in the variability of temperature both
seasonally and  daily  are important  to  studies
concerned with the effect of temperature change on
electricity demand  (discussed  in Chapter  10).
Although new generating capacity requirements for
the  nation for  2010 and beyond  are  calculated
assuming  climate  change, the numbers generated
could be  considerably different for any particular
year, depending mainly on air-conditioning needs,
which would be  the  major use  increase  for
electricity. Such needs are sensitive to extremes in
daily maximum temperatures and the persistence of
such temperatures (i.e., heat waves).

    It would be impossible to quantitatively or even
qualitatively estimate how different the results of
these studies would  be  if changes  in  climate
variability had formed part of the climate scenarios
made available as input for the various climate
impact models used.   Primarily, it is impossible
because  the  variability changes  are  not known;
second, it is impossible because most of the studies
are so complex that the effect of a change in one
variable (a complex change at that) is not intuitively
obvious in most cases. Analyses of the sensitivity of
the impact models involved to changes in variability
would be required to  provide  specific answers.
What can be said at this point is that the lack of
information on climate variability  has limited  a
number of studies in this report and has limited the
completeness of the answers they could provide.

    The research reported above clearly indicates
that research  of  changes  in  climate  variability
associated with climate change is truly hi its infancy.
Much needs to be  done.  Future research needs
may be broken into  three  categories:  further
analysis of GCMs;  improvements hi  GCMs; and
sensitivity analysis of impacts.

Further  Investigation  of  Variability  in

    Results summarized here  represent  only  an
initial effort at looking at variability in GCMs. We
need to examine in more models and at many more
grid boxes the daily and interannual variability of
many climate variables (such as relative  humidity,
solar radiation, and storm frequency) in addition to
temperature and precipitation.  Other time scales
of variability also should be examined,  such as 7- to
10-day scales, which correspond to the lifetime of
many  frontal storms.    Moreover, the  most
sophisticated statistical techniques must be used or,
where  needed, developed, such that  uniform
quantitative indicators are available to evaluate both
how well the  current models  reproduce  present
variability  and  how  they  forecast the  change

 Chapter 3
 in variability under climate change conditions. The
 causes  for  discrepancies  in present-day climate
 variability and control run variability must be better
 understood  to attain a  clearer understanding of
 future climate changes.

 Improvements  in GCMs

    The results of Rind et al. and Mearns et al. give
 some indications that oversimplifications in the land
 surface   packages  of  GCMs   contribute   to
 overpredictions  of daily temperature variability.
 This possibility is further underlined by the better
 results  obtained  with Dickinson's  model, which
 includes a more sophisticated land surface package.
 More  detailed  analyses of  current GCMs  are
 necessary to confirm this speculation, as well as to
 determine the causes of other errors in variability,
 such as for  precipitation.  Other known causes of
 error,  such as the  models'  relative inability  to
 simulate  subgrid-scale  phenomena,   must  be
 investigated further. The next step involves altering
 the GCMs so that  variability is properly simulated.
 Only then can much faith be put in GCM forecasts
 of variability changes with a perturbed climate.

 Sensitivity Analyses of Impacts

    It  also  must be determined how  important
 changes in variability will be  to different areas of
 impact.  Since the variability of climate variables
 produced from GCMs cannot be "trusted" or even
 easily  analyzed at  this point, these  sensitivity
 analyses of impact models should be performed with
 statistically  simulated  time   series  of  climate
variables, as has been performed by Schwarz (1976)
 and Mearns  et al. (1984).  By simulating time series,
 different levels of  autocorrelation and variance  in
the  time series  may   be  controlled for  and
systematically  varied.  By  this  means, important
thresholds  of  variability  change  for  different
variables as they affect the output of impact models
can be determined. Moreover,  ranges of possible
impacts of variability change can be determined and
can serve as  guides  until better information  is
available  on how variability will  change  in  a
COg-warmed world.
American  Meteorological Society.   1985.   Flash
floods:   A statement of concern  by the AMS.
Bulletin of the American Meteorological Society

Barbecel, O., and M. Eftimescu.  1973.  Effects of
Agrometeorological  Conditions on Maize Growth
and Development. Bucharest, Romania:  Institute
of Meteorology and Hydrology, pp. 10-31.

Bates, G.T., and GA. Meehl.  1986. The Effect of
CO2 concentration on the frequency of blocking in
a general  circulation model  coupled  to  a simple
mixed layer ocean model. Monthly Weather Review
Becker, R.J., and RA. Wood.
Weatherwise 39(4):195-6.
1986.  Heatwave.
Berger, A. 1980. Spectrum of climate variations and
possible   causes.  In:  Berger,  A.,  ed.  Climatic
Variations and Variability:  Facts and Theories.
Dordrecht, Holland: D. Reidel. pp. 411-432.

Bourke, W. 1974.  A multi-level spectral model.  I.
Formulation and hemispheric integrations. Monthly
Weather Review 102:687-701.

Bourke, W., B. McAvaney, K. Puri, and R. Thurling.
1977.  Global  modeling of atmospheric flow by
spectral  methods.   Methods  in  Computational
Physics 17:267-324.

Bridger,  CA., P.P.  Ellis, and H.L. Taylor.   1976.
Mortality in St. Louis, Missouri, during heat waves
in 1936,1953,1954,1955, and 1966. Environmental
Research 12:38-48.

Brinkmann,   WA.R.    1983.     Variability  of
temperature  in Wisconsin.    Monthly Weather
Review 111:172-180.

Chervin,  R.M.  1986.  Interannual variability and
seasonal predictability.  Journal of the Atmospheric
Sciences 43:233-251.

                                                                                  Climate Variability
Chervin,  R.M.   1981.   On the comparison  of
observed  GCM  simulated  climate  ensembles.
Journal of the Atmospheric Sciences 38:885-901.

Clark, R A.  1987. Hydrological design criteria and
climate variability. In: Soloman, S.I., M. Beran, and
W. Hogg, eds.  The  Influence of Climate Change
and Climatic Variability on the Hydrologic Regime
and Water Resources. International Association of
Hydrological Sciences Publ. No. 168.  Oxfordshire:
IAHS Press.

Coleman, J.E.    1988.   Climatic warming and
increased summer aridity in Florida, U.S A. Climatic
Change 12:164-178.

Dare, P.M.  1981. A study of the severity of the
midwestern winters of 1977 and 1978 using heating
degree days determined from measured and wind
chill temperatures.    Bulletin of  the American
Meteorological Society 62(7):674-682.

Diaz, H.F. and R.G. Quayle 1980. The climate of
the United States since 1895:  Spatial and temporal
changes.  Monthly Weather Review 108:249-266.

Dickinson,  R.E.,  A.   Henderson-Sellers,  PJ.
Kennedy,   and   M.F.   Wilson.      1986.
Biosphere-Atmosphere Transfer Scheme (BATS)
for the NCAR Community Climate Model. NCAR
Technical Note 275. Boulder, CO: National Center
for Atmospheric Research.

Ellis, F.P.  1972. Mortality from heat illness and
heat-aggravated illness in the U.S.  Environmental
Research 5(l):l-58.

Emmanuel,  KA.    1987.   The dependence  of
hurricane intensity on climate. Nature 326:483-485.

Emmanuel, KA.  1988. Toward a general theory of
hurricanes.  American Scientist 76(4):370-379.

Gibbs, WJ.  1984.  The  great Australian drought:
1982-1983. Disasters 8/2/84:89-104.

Gibbs, WJ., J.V. Maher,  and MJ. Coughlan.  1975.
Climatic variability and extremes. In: Pittock, A.B.
LA. Frakes, D. Jenssen, JA. Peterson, and J.W.
Zillman, eds.  Climatic Change and Variability.
Cambridge, United Kingdom: Cambridge University
Press, pp. 135-150.
Hansen, J., I. Fung, A. Lacis, S. Lebedeff, D. Rind,
R. Ruedy, G. Russell, P. Stone.   1988.   Global
climate changes as forecast by the Goddard Institute
for Space Studies three-dimensional model. Journal
of Geophysical Research 93(D8):9341-9364.

Heathcote, R.L. 1985. Extreme event analysis. In:
Kates, R.W., J.H. Ausubel, and M. Berberian, eds.
Climate   Impact   Assessment,   SCOPE   27.
Chichester, United Kingdom: John Wiley and Sons.

Henz, J.F.,  and V.R. Scheetz.   1976.  The big
Thompson flood of 1976 in Colorado. Weatherwise

Herbert, PJ., and G. Taylor. 1979. Everything you
always  wanted  to  know   about   hurricanes.
Weatherwise 32(2):61-67.

Jones, T.S.,  et al.  1982.  Morbidity and mortality
associated with the July 1980 heat wave in St. Louis
and Kansas  City,  MO.  Journal of the American
Medical Association 247:3327-3331.

Kalkstein, L.S., R.E. Davis, J A. Skindlov, and K.M.
Valimont.  1987.   The impact of human-induced
climate  warming upon human mortality:  A New
York  case   study.   Washington,  DC:    U.S.
Environmental Protection Agency.

Katz, R.W.  1984.  Procedures for  Determining the
Statistical Significance of Changes in Variability
Simulated by an Atmospheric General Circulation
Model.  Climate Research Institute Report No. 48.
Corvallis, OR: Oregon State University.

Linder, K.P., MJ. Gibbs, and M.R. Inglis.  1987.
Potential Impacts  of Climate Change on Electric
Utilities.   Albany, NY:  New York State Energy
Research and Development Authority.  Report 88.2.

Lough,  J.M.,  T.M.L.  Wigley,  and J.P. Palutikof.
1983.  Climate and climate impact scenarios for
Europe in a warmer world.  Journal of Climate and
Applied Meteorology 22:1673-1684.
 Ludlum,   D.M.   1988.     Weatherwatch
 September).  Weatherwise 41(6):354-355.
 Manabe, S., and D.G. Hahn. 1981. Simulation of
 atmospheric variability.  Monthly Weather Review

 Chapter 3
 McQuigg, J.D.  1981.  Climate variability and crop
 yield in high and low temperature regions. In: Back
 W.,  J.  Pankrath,  and  S.H.   Schneider,  eds.
 Food-Climate Interactions. Dordrecht, Holland: D.
 Reidel Publishing Company,  pp. 121-138.

 Mearns, L.O., S.H. Schneider, S.L. Thompson, and
 L.R.  McDaniel.    1989.    Analysis  of climate
 variability in general circulation models: comparison
 with observations  and  changes in  variability  in
 2xCO2  experiments.    Journal of  Geophysical
 Research (accepted).

 Mearns, L.O.,  R.W. Katz,  and S.H. Schneider.
 1984.  Extreme high-temperature events:  Changes
 in   their probabilities   with changes  in  mean
 temperature. Journal of Climate and Applied
 Meteorology 23:1601-1613.

 Mederski,  HJ.    1983.   Effects of   water and
 temperature stress on soybean  plant growth and
 yield in humid temperature climates.  In: Raper,
 C.D.,  and PJ. Kramer,  eds.  Crop  Reactions  to
 Water  and Temperature  Stresses  in  Humid,
 Temperate  Climates.  Boulder, CO:   Westview
 Press, pp. 35-48.

 Meehl, G. 1989. Seasonal cycle forcing of El Nino
 phenomena in a coupled  ocean-atmosphere model.
 Submitted to Science.

 Mitchell,  J.M.  1976,  An overview of climatic
 variability and its causal mechanisms.  Quaternary
 Research 6:481-493.

 Mogil, H.M., A. Stern, and R. Hagen.  1984.  The
 great freeze of '83: Analyzing the causes and the
 effects. Weatherwise 37(6):304-308.

 Neild, R.E.  1982.   Temperature  and rainfall
 influences on the phenology and yield  of  grain
 soybean and maize.   Agricultural  Meteorology

 Newman, J.E. 1978. Drought impacts on American
 agricultural  productivity.  In:  Rosenberg, N.J., ed.
 North American  Droughts.     Boulder, CO:
 Westview Press,  pp. 43-63.

 Oeschsli, F.W., and R.W. Buechley.  1970. Excess
 mortality  associated  with  three  Los  Angeles
September hot spells.   Environmental  Research
 Oliver, J.   1981.   The nature  and impact  of
 Hurricane  Allen - August  1980.   Journal  of
 Climatology 1:221-235.

 Parry, M.L. 1978.  Climatic Change, Agriculture
 and  Settlement.   Folkstone,  United  Kingdom:

 Pittock,  A.B.   1975.   Patterns  of  variability  in
 relation  to  the  general circulation.  In:   Pittock,
 A.B., LA. Frakes, D. Jenssen, JA. Peterson, J.W.
 Zillman,  eds.   Climatic Change and Variability.
 Cambridge,   United   Kingdom:     Cambridge
 University Press, pp. 167-178.

 Policansky, D. 1977. The winter of 1976-77 and the
 prediction of unlikely weather.  Bulletin  of the

 Quirk,  WJ.     1981.     Climate  and   energy
 emergencies.      Bulletin  of  the  American
 Meteorological Society 62(5):623-631.

 Reed, D.N.  1986.  Simulation of time series  of
 temperature and precipitation over eastern England
 by an  atmospheric  general  circulation  model.
 Journal of Climatology 6:233-257.

 Reibsame, W.E. 1988.   Adjusting water  resources
 management to  climate change. Climatic Change

 Rind, D.,  R. Goldberg,  and  R.  Ruedy.   1989.
 Change in climate variability in the  21st century.
 Climatic Change 14:5-37.

 Rosenberg, N.J., ed. 1979. Drought in the Great
 Plains:    Research on Impacts  and Strategies.
 Littleton, CO: Water Resources Publications.

 Rosenberg,  N.J., ed.   1978.   North American
 Droughts. Boulder, CO: Westview Press.
Sanders, J.F.   1982.   The
Weatherwise 35(4):174-178.
hurricane  dilemma.
Shugart, H.H., M.Y. Antonovsky, P.G. Jarvis, and
A.P. Sandford.  1986.   CO2> climatic change, and
forest ecosystems.   In:  Bolin, B., B.R. Doos, J.
Jager, and RA. Warrick,  eds.  The Greenhouse
Effect, Climatic Change, and Ecosystems.  SCOPE
29.  Chichester, United Kingdom: John Wiley and
Sons. pp. 475-521.

                                                                                  Climate Variability
Schwarz, H.E.  1977.  Climatic change and water
supply:   how sensitive is the Northeast?   In:
Climate, Climate  Change  and  Water  Supply.
Washington, DC: National Academy of Sciences.
pp. 111-120.

Shaw, R.H.  1983. Estimates of yield reductions in
corn caused by water and temperature stress.  In:
Raper, C.D., and P J. Kramer, eds.  Crop Reactions
to Water and Temperature Stresses  in  Humid,
Temperate Climates.   Boulder, CO:   Westview
Press,  pp. 49-66.

Soloman, A.M.,  and D.C. West.  1985. Potential
responses of forests to CO2-induced climate change.
In:    White,  M.R.,  ed.    Characterization  of
Information Requirements for Studies of CO2
Effects:  Water Resources, Agriculture, Fisheries,
Forests, and Human Health. UJS. Department of
Energy. DOE/ER-0236.  pp. 145-170.

Thompson, L.M. 1968.  Weather and technology in
the production of corn.  In: Purdue Top Farmer
Workshop, Corn Production Proceedings.   West
Lafayette, IN:  Purdue University,  pp. 3-19.

Treriberth, K.E.,  G.W. Branstator,  and PA. Arkin.
1988. Origins of the 1988 North American drought.
Science 242:1640-1645.

Van Loon,  H.  and J. Williams.  1978.   The
association  between   mean  temperature  and
interannual variability.  Monthly Weather Review
Warrick, RA. et al. 1975. Drought Hazard in the
United States: A Research Assessment.  Boulder,
CO: University of Colorado. IBS Monograph No.

Washington,  W.M.,  and GA.  Meehl.    1984.
Seasonal cycle experiment on the climate sensitivity
due  to  a doubling of CO2  with an atmospheric
general  circulation  model  coupled to a  simple
mixed-layer ocean model.  Journal  of Geophysical
Research 89(D6):9475-9503.

White, G.F. et al.  1975.  Flood  Hazard in the
U.SA.  Research  Assessment.    Boulder,  CO:
University of Colorado.   IBS Monograph No.

Wilford, J.N.    1988.   Vast  Persistent  Pattern
Spreading Heat Wave. The New York Times, p. 1;
July 19.

Wilhite,  DA.,   and   M.H.  Glantz.     1987.
Understanding the drought phenomenon:  The role
of definitions. In: Wilhite, DA., and W. Easterling,
eds.    Planning for  Drought.    Boulder,  CO:
Westview Press,  pp. 11-27.

Wilson,  CA., and J.F.B. Mitchell.  1987. Simulated
climate  and  CO2-induced  climate change  over
Western Europe.  Climatic Change  10:11-42.


                                      CHAPTER 4

     As discussed in Chapter 2: Climate Change,
there is  a scientific  consensus  that  increased
atmospheric concentrations of greenhouse gases will
likely increase global temperatures, and that such a
global  temperature increase will likely increase
global precipitation and sea levels.  There is no
consensus on how regional climates may change.
We do not  know whether temperatures will rise in
all regions;  we do not know whether precipitation in
any particular region will rise or fall or whether we
will have seasonal changes, and we are uncertain
about the  rate and magnitude of  change.   As
discussed   in  Chapter  3:  Climate   Variability,
scientists do not know how variability — that is, the
frequency of droughts,  storms, heat waves, and
similar phenomena  —  may change.   Without
knowing  how  regional climate may change, we
cannot predict impacts.

     Despite these uncertainties, we can get a sense
of what the future may look like through the use of
scenarios.  Scenarios are plausible combinations of
conditions  that may be used to illustrate future
events.  They may be used to identify possible
effects of climate change and to evaluate responses
to those effects.   To  incorporate uncertainties
surrounding regional  climate  change,  regional
scenarios should  include  a variety of  potential
climate  changes  consistent  with  the  state  of
knowledge regarding global warming. By analyzing
many scenarios,  we may be able to identify the
direction and relative magnitude of impacts.  Yet,
unless scenarios have probabilities assigned to them,
predictions of future impacts cannot be made.  In
this report,  probabilities are not assigned and results
do not represent predictions.  Only the direction of
change and relative magnitude are identified.  The
scenarios used in this report do not represent the
entire range of possible climate change. Thus, the
range of effects identified does not represent the
entire range of potential effects.

     To assess the potential effects of global climate
change, regional scenarios of such change should
have the following characteristics:

     1.  The  scenarios  should   be  internally
        consistent with global warming caused by
        increases in greenhouse gas emissions.  A
        doubling of the CO2 concentration in the
        atmosphere is  thought to increase global
        temperatures by approximately 1.5 to 4.5ฐ C
        (3  to 8ฐF).   The  regional  temperature
        changes and seasonal distributions may be
        higher  or  lower, as  long as they  are
        internally consistent with the global range.

     2.  The  scenarios  must include a sufficient
        number of meteorological variables to meet
        the requirements for using effects models.
        These effect models include models of crop
        growth, forest succession, runoff, and other
        systems. Some models of the relationship
        between climate  and a system use only
        temperature  and precipitation as climate
        variables, while  others also  need solar
        radiation,  humidity,  winds,   and  other

     3.  The  meteorological  variables should be
        internally consistent.  While a scenario is
        not  a prediction, it  should  at  least be
        plausible.  The laws of physics limit how
        meteorological variables may change  in
        relationship to each other.  For example, if
        global   temperatures   increase,   global
        precipitation must  also  rise.  Regional
        changes should be internally consistent with
        these large-scale changes.

     4.  The   scenarios   should  provide
        meteorological variables on a daily basis.
        Many of the effects models  used in this
        study, such as crop yield  and hydrology
        models, need daily meteorological inputs.

 Chapter 4
     5.  Finally, the scenarios should illustrate what
         climate would look like on a spatial scale
         fine enough for effects analysis.   Many
         effects   models   consider   changes  in
         individual stands of trees or farm fields. To
         run them, scenarios  must illustrate how
         climate may change locally.

     Two questions should be answered in analyzing
 the  potential impacts of the  greenhouse effect:
 What would be the effects of a large climate change
 in the future? How quickly will the effects become
 apparent over time?  The first  question asks what
 the world will be  like in the future;  the second is
 about the speed of change and the sensitivity of the

      One way of examining the first question is to
 use  scenarios of  an equilibrium future  climate.
 Climate equilibrium is defined as climate in which
 average conditions are not changing (although year-
 to-year variations could still occur).

     A drawback of an equilibrium scenario is that
 it occurs at an  arbitrary  point in the future and
 assumes that the climate has reached a stable level
 corresponding with the higher concentrations  of
 greenhouse gases.  It does not indicate how climate
 may change between now  and  the equilibrium
 condition  or  how soon  effects  may be seen.
 Furthermore, a "stable" climate has never happened,
 nor is it likely to occur.

     To help identify sensitivities and give a sense
 of when effects may occur, this study uses transient
 scenarios of climate change. A transient scenario is
 a scenario of how climate  may change over time.

     The options for creating regional scenarios of
 global warming include the following:

     1.  arbitrary changes in climate;

     2.  analog warming; and

     3. use of general circulation models.

Arbitrary Changes

     A simple way of constructing a scenario is to
assume  that  climate variables change by some
 arbitrary amount.  For example, one could assume
 that temperatures increase by 2 or 4ฐC, or that
 rainfall rises or falls by 10% and all other variables
 are held constant. Such scenarios are relatively easy
 to  use and  can help to  identify the sensitivities of
 systems  to  changes  in  different  variables.   To
 determine how sensitive a system is to temperature
 alone, one  could hold other variables at current
 climate levels and change temperature by arbitrary

     A major  drawback  to using scenarios  with
 arbitrary changes is that  they may not be realistic,
 since evaporation, precipitation, wind, and  other
 variables will   most  likely   change   if  global
 temperatures change. A combination of unrealistic
 meteorological  changes  may yield  an unrealistic
 effect. We  are not sure how other meteorological
 variables  would change  on a regional scale if
 temperature rose a certain amount. Thus, scenarios
 with  arbitrary  changes  may  be  useful   for
 determining sensitivities to particular variables but
 not for  determining  the  possible magnitudes  of

 Analog Warming

     Many climatologists have advocated the use of
 historic warming periods as an analog of how a
 future  warming  may  affect  regional  climates
 (Vinnikov and Lemeshko, 1987). The instrumental
 weather record can be used by comparing a cool
 decade on record, such as the 1880s, with a warm
 decade, such as the 1930s (Wigley, 1987), or by
 comparing a decade such as the 1930s  with  the

     Paleoclimatic data may also be incorporated
 into an analog  warming scenario.  For example,
 6,000 years  ago the temperatures were about 1ฐC
 warmer.  Paleoclimatologists have determined how
 rainfall  and temperature  patterns on a  broad
 regional  scale differed in the past.  The changes
 associated with past climates that were warmer than
 now may be used as an analog warming scenario.

     The advantage of using an analog is that it
gives a  realistic sense of how  regional  and  local
weather patterns change as global climate warms.
For example, climate data from  1880 to 1930 show
how daily and  local  weather  changed  during a
warming period.

      However,  analogs have  several  drawbacks.
First, they  are  not consistent with the range  of
global  warming now  thought likely  under the
greenhouse effect:  1.5 to 4.5ฐ C.  The warmest
period  of the last  125,000 years  was 1ฐC  warmer
than  the present  temperature.    (Although the
Pliocene Epoch (2 to 5 million years ago) had
global temperatures several degrees higher than
now,  there  is  virtually no information  on the
regional distribution of temperature and rainfall
during that period.) In addition, the past warmings
were  not necessarily  caused  by changes  in the
concentration of greenhouse gases, but may have
been  due to such factors as shifts in the inclination
of the Earth's axis. These factors caused different
regional climate changes than would be associated
with  increases  in  radiative  forcing.    Second,
paleoclimatic and historic  records dp not  provide
enough detail to conduct comprehensive analysis of
the 1ฐC  warming.    Paleoclimatic records only
indicate broad regional patterns of change for a few
variables, such as temperature, rainfall, and solar
radiation.   We  cannot discern  local, daily,   or
interannual climate from these records.  Even using
the 1930s  data  presents some problems.  Daily
records  are  available  only for temperature and
rainfall.  Some effects models need more variables,
such  as wind  or  radiation.   Furthermore, the
number of  weather stations with 1930s  data  is
limited, which could present problems for creating
comprehensive regional scenarios.

General Circulation Models (GCMs)

      GCMs are dynamic models  that simulate the
physical process of the atmosphere and oceans  to
estimate global climate.  These models have been
developed over two decades and  require extensive
computations to run.  They can be run to estimate
current climates and the sensitivity of  climate  to
different conditions such as different compositions
of greenhouse gases. The GCMs  are often used  to
simulate climate caused by a doubling of carbon
dioxide levels, also referred to as doubled CO-.
Estimates of climate change caused by this effective
doubling of CO21 are referred to  as "doubled CO2
scenarios." Output is given in regional grid boxes.
 The "effective doubling of CO2" means that the total radiative
forcing of all greenhouse gases (CO2, CH4, N2O, CFCs, etc.) is
the same as the radiative forcing caused by doubling carbon
dioxide concentrations, over midcentury levels, alone. In other
words, the combination of all greenhouse gases has the same
radiative forcing as simply doubling CO2.
     GCMs have several advantages over the other
approaches for creating scenarios. First, the models
are used to estimate how global climate may change
in  response   to  increased  concentrations   of
greenhouse gases.   Thus, regional outputs  are
internally  consistent  with   a  global warming
associated with doubled COu. Second, the estimates
of  climate   variables  (for  example,  rainfall,
temperature,  and humidity levels)  are physically
consistent within the bounds of the model physics.
Third,   GCMs   estimate   outputs  for    many
meteorological variables (including wind, radiation,
cloud cover,  and soil moisture) providing enough
input for effects models.  Fourth, GCMs simulate
climate variability on at least a daily basis.

     Among the most important limitations are the
GCMs' simulations of the oceans. The oceans play
a critical role in determining the rate of climate
change, regional climate  differences, and climate
variability. The  GCMs, however, are coupled to
relatively simple models of ocean circulation, which
either treat the oceans as a "swamp"  or. only model
the  upper layers  of  oceans.     The  models'
assumptions oversimplify the transfer of heat to and
from the oceans.  In  addition, the GCMs simplify
other important factors that affect climate, including
cloud cover and convection, sea ice, surface albedo
(the amount of light reflected, rather than absorbed,
from the surface) and land surface hydrology (i.e.,
soil  moisture), which may  also   contribute  to
uncertainty about the estimates of climate change
(Dickinson, 1986; Schlesinger and Mitchell,  1985;
Gates, 1985).   For example,  some  of the GCMs
model soil moisture storage in a simple manner,
assuming the soils act like a "bucket.'1 (There have
been recent improvements on this method.)  This
method of modeling raises uncertainties concerning
estimates of  runoff from the  models.   The -way
GCMs simulate such important climate factors as
oceans, clouds, and other features casts some doubt
on the validity of the magnitude of global warming
estimated by the models. (For a further discussion
of the role of oceans in climate change, see Chapter
2: Climate Change. For a discussion of the GCMs'
ability to estimate climate variability, see Chapter 3:
Climate  Variability.)

     One  of  the major  disadvantages of using
GCMs  for  effects analysis  is  their low  spatial
resolution. GCMs give outputs in grid boxes that
vary in size from 4 by 5 degrees latitude to as much
as 8 by 10 degrees longitude.  Figure 4-1 shows the
grid boxes from  the Goddard Institute for Space

 Chapter 4
                              Figure 4-1.  GISS model of the United States.
Studies (GISS) model  overlaid on a map  of the
United States.  Each grid box is 8 by 10 degrees and
is an  area larger  than France  (Mitchell,  1988).
Within each grid box,  the actual climate may be
quite variable.  For example, although both are in
the  same grid box,   the  weather  in southern
Washington State may be quite different from the
weather  in  northern   California   The models,
however, do not account for variations within each
grid box.   For  any simulated time, they provide a
single value for temperature, for  rainfall, and for
other variables for the entire area of the box.

     A second disadvantage for  effects  analysis,
which may be  more critical  than  the first,  is that
GCMs generally do not accurately simulate current
regional   climate  conditions.   In   general,  the
accuracy of GCM climate estimates decreases with
increasing resolution. The GCMs do a reasonable
job  of estimating  observed  global  and   zonal
climates, but the estimates of regional climate are,
in many cases,  far from observed conditions. This
is shown  in  Table 2-2  (see  Chapter 2:  Climate
Change),   adapted  from  Grotch  (1988),  which
displays GCM temperature  estimates  and actual
observations on different scales. GCM estimates of
rainfall are less reliable on a regional  scale.   As
Grotch points out, the disparities between GCM
estimates of  current  regional climate and actual
conditions calls into question the ability of GCMs to
predict climate change on a regional scale.

     The disparities among GCM estimates on a
regional scale are due to a number of factors. One
of the most important is the simplified assumptions
concerning the oceans.  The assumptions on other
factors such as cloud cover, albedo, and land surface
hydrology also affect regional estimates. The GCMs
also simplify topographic features within grid boxes,
such as the distribution of mountains or lakes. The
large size  of the  grid boxes  means that these
features are oversimplified on a geographic scale.
This contributes to uncertainty regarding estimates
of regional climate change.  In  sum,  as  Grotch
concluded, GCM estimates  of  regional  climate
change should  not be  taken  as predictions  of
regional  climate   change.    They should   be
interpreted as no more than illustrations of possible
future regional climate conditions.


      GCM outputs were employed as a basis for
constructing the scenarios to be used in our report
because they produce the best estimate of climate
change  due   to   increased   greenhouse   gas
concentrations  and  they produce regional climate
estimates internally consistent with doubled CO2
concentrations. Yet, GCMs are relatively new tools
that need  a great  degree  of refinement.  Their
results must be applied with caution.  The regional
GCM estimates of climate change are considered
to  be  scenarios,  not  predictions.    Given  the
uncertainties about GCM estimates  of daily and
interannual variability (see Chapter 3: Variability),
a  conservative approach  involves using average
monthly changes for each grid box.

      The scenarios described in this chapter are a
hybrid between GCM outputs and historic weather
data. The estimates of average monthly change in
temperature,  precipitation,   and  other  weather
variables are used from GCM grid boxes. Model
simulations of monthly doubled CO2 conditions are
divided by model simulations of average monthly
current conditions in each  grid.  The ratios of
(2xCC>2):(lxCC)2) are multiplied by historic weather
conditions at weather stations in the respective grid
boxes. Parry et al. (1987) used this approach in an
analysis of impacts of climate change on agriculture.
Thus, if a grid box is estimated to  be 2ฐC warmer
under the GCM doubled CO2 run, all stations in
that grid are assumed to be 2ฐC warmer in  the
doubled CO2 scenario.   The effect of this is to
keep geographic variation from  station to station
within a grid the same as in the historic base period.
Furthermore, interannual (year to  year) and daily
variability remain the same.  If rainfall occurs 10
days in a month, hi the scenario it also occurs 10
days in the month,  and the amount  of rainfall is
adjusted by the GCM output. Since these scenarios
are  hybrids  between  GCM  average  monthly
estimates and daily historic weather records, these
scenarios are not strictly GCM  scenarios.  Each
scenario is referred to by the GCM, whose monthly
output serves as its base (e.g., the "GISS scenario").

     The years 1951-80 were chosen as the base
period to which average doubled CO2 changes were
applied. Several decades of data give  a wide range
 of warm, cold, wet, and dry years. Since the data
 are from the most recent decades, they are the most
 complete historic data available. A complete daily
 record  for a number of weather variables only
 began in 1948.

 GCMs Used

     To obtain a range of scenarios, output from
 three GCMs  was used:

     •  Goddard Institute for Space Studies (GISS)
        (Hansen et al., 1988);

     •  Geophysical Fluid Dynamics Laboratory
        (GFDL) (Manabe and Wetherald, 1987);

     •  Oregon  State   University   (OSU)
        (Schlesinger and Zhao, 1988).

     The  average  seasonal   temperature  and
 precipitation for the U.S. gridpoints for each model
 are displayed in Figure 4-2.   All three models
 estimate that  average temperatures over the United
 States   would rise,  but  they disagree  on the
 magnitude. OSU gives 3ฐC, GISS 4.3ฐC, and GFDL
 5.1ฐC.   The  seasonal patterns  are different, with
 GISS having  a larger warming in winter and fall,
 GFDL having the highest temperature change in the
 spring, and OSU having little  seasonal variability.
All three models estimate that annual precipitation
 over the United States would increase.  GISS and
 OSU estimate that annual precipitation would rise,
respectively, by 73 millimeters (2.92 inches) and 62
millimeters (2.48 inches), while GFDL estimates a
rainfall increase of only 33 millimeters (1.31 inches).
The first two models have precipitation increases in
 all  four  seasons, while GFDL has a decline in
summer rainfall.  As can be seen in the regional
chapters, the models show greater disagreement on
the  direction and  pattern of regional  rainfall
changes than  on regional temperature.  Overall,
OSU appears to be the "mildest" scenario, with the
lowest  temperature  rise  and  largest  increase in
precipitation.    GFDL appears to  be  the  most
"extreme," with the highest temperature rise, the
smallest increase hi precipitation, and a decrease in
summer rainfall.  Some of the important parameters
in the three GCMs are displayed in Table 4-1.

 Chapter 4
          Wtntor   Spring
                                                                Spring   Summer
Figure 4-2. Average changes in temperature (ฐC) and precipitation (mm/day) over the grid boxes of the lower
48 states (2xCO2 less lxCO2).
     The "extreme" values  in the GFDL doubled
CO, scenario are due, in part, to assumptions made
in the model run used in this report.  That run did
not constrain sea surface temperature and sea ice,
which yielded seasonal extremes  in  the northern
hemisphere.  A later run, produced too late for use
in this study, constrained sea surface temperature
and sea ice to observed values.  Both runs yield the
same average global warming of 4.0ฐC, while the
later run has  greater seasonal extremes in  the
southern hemisphere.  Both runs show a  large
decrease in  summer soil  moisture (Wetherald,
personal communication, 1988).


     A  major  limitation of the doubled  CO2
scenarios used for this study is the lack of temporal
and spatial variability. By applying average monthly
changes to the historic data set, it is assumed that
the daily and interannual patterns of climate remain
the same. This assumption is probably unrealistic,
since a change in average conditions will probably
lead to  a change in variability.   Furthermore,
holding variability constant can have an impact on
effects analysis.

     Most climate-sensitive systems are sensitive to
climate  variability.  For example, riverflow is  very
sensitive to the amount and intensity of rainstorms.
Certain crops are sensitive to consecutive days with
    temperatures above a certain level.  The studies do
    not identify how these and other systems could be
    affected by changes in temporal climate variability.
    Holding spatial variability within a grid box constant
    also affects the results of the analyses performed for
    this  report.  Climate change may also lead to
    changes in wind patterns, which could change storm
    patterns,   cloud  distribution,  deposition  of  air
    pollutants, and other systems. In addition, the years
    1951 to 1980 were a period of relatively low weather
    variability in the United  States.  Only adjusting
    average conditions  from  the base period in the
    scenarios  may  underestimate  potential  increases
    invariability.  (For further discussion, see Chapter
    3: Variability.)

         The choice of the three doubled CO2 scenarios
    does not necessarily bracket the range of possible
    climate change in the latter half of the next century.
    Due  to  the uncertainties  about  the rate  and
    magnitude of global  warming, it is possible  that
    average global  temperatures  could be  lower or
    higher than indicated by the models. Other climate
    variables   could  be  different too.    Thus,  these
    scenarios  should be interpreted as  illustrations of
    possible  future  conditions,  not  as predictions.
    Furthermore, we did not assign probability to these
    scenarios.    Currently,   there is  not  enough
    information or  a methodology for making such a

                           Table 4-1. Major Features for the Three GCMsa



(lat. x long.)




Temp for Increase
in global
7.83 x 10ฐ
4.44 x 7.5d

GISS Transient

4.00 x 5.0d

7.83 x 10d



(in 1958)
a All models are global in extent and have an annual cycle.  All models have a smoothed topography that varies
  between models. The later GFDL run has been added for information. All models (except the transient) give
  data for the present climate (lxCO2) and double CO2 climate (IxCOJ.
bAU models make calculations for surface conditions as well as for the fisted upper-air levels.
ฐA gridpoint model with stated resolution.
dThis is a spectral model that has 15 waves.

Note: Oceans in Models:
  GISS: This model has a slab ocean not over 65 meters deep; it has some variation of mixed depth over the
        seasonal cycle (for example, the depth is shallower in summer than winter in mid-latitudes).  It has a
        specified pseudo ocean heat transport designed to reproduce the present day sea surface temperature
        (SST) in the simulation of the present climate. Ice thickness is predicted. For the GISS transient runs,
        the ocean depth was not limited in this way. In it, the average annual maximum mixed-layer depth
        was 127 inches.
 GFDL: The slab ocean is 68 meters deep.  There is no horizontal heat transport that would make the present
        day SST come out exactly right.  Ice thickness is predicted.
  OSU: This model has a slab ocean that is 60 meters deep (only 5 meters deep during spin-up period).  It does
        not have heat transport that would force the model to reproduce the present day SST (this  is being
        added in 1989).
     If current  emission  trends continue,  the
effective doubling of CCu concentrations will occur
around the year 2030. However, that estimate does
not account for some recent developments that may
slow the increase in greenhouse gas concentrations.
If implemented, the Montreal Protocol would cut
emissions  of chlorofluorocarbons (CFCs) by 50%.
If  an  international  agreement  is  reached  on
reduction   of  nitrogen   oxides  (NOX),   the
concentration of nitrogen dioxide (N2O) may be
slightly reduced.   Pollution control  measures in
countries such as the United States may also reduce
concentrations  of  low-level   ozone,  another
greenhouse gas.  Thus, the effective doubling of
CO2 may happen  after 2030.

     As discussed in Chapter 2: Climate Change,
the change in  climate potentially caused by CO2
                                doubling would not occur at the same time as the
                                increase hi greenhouse  gas  concentrations.  The
                                oceans absorb greenhouse gases and heat from the
                                atmosphere and serve to delay the warming. The
                                full extent of climate change associated with CO2
                                doubling could  take several decades or more and
                                may not occur until the latter half of the next
                                     In this  report, results from doubled  CO.
                                scenarios  are generally not  associated with  a
                                particular year. When analysis is necessary, we have
                                generally assumed that the CO2 warming will occur
                                in 2060.   In some  cases, researchers assumed a
                                different time period for CO2 warming, and those
                                exceptions are noted as appropriate in the text.

Chapter 4
     The  doubled  CO2  scenarios   are  often
interpreted   as   estimates   of  future   static
(equilibrium) conditions.  The assumption that the
concentration of greenhouse gases becomes constant
at doubled CO2 levels is an arbitrary one. In fact,
if emissions are not limited, concentrations could
become  greater  and the  global  climate would
continue to change.  In many places in this report,
responses are presented as if the climate stabilizes
at doubled CO2 conditions.  Natural systems  and
society, however, may be responding and adapting
to continuing and perhaps, accelerating changes in

     The options for developing transient scenarios
are similar to  the options for the  doubled CO2

     1. arbitrary changes;

     2. analog warming; and

     3. GCM  transient runs.

Arbitrary Changes

     One could examine the manner in which a
system  responds  to   an  arbitrary  1  or 2ฐC
temperature warming and to small arbitrary changes
in other variables.   The  problems of physically
inconsistent  assumptions  about  changes  among
variables and regions pertain here also.  In addition,
the arbitrary warming scenario gives no indication
of when the warming may occur.

Analog Warming

     Wigley (1987) has suggested using analogs as
scenarios for climates  that may  occur within the
next several decades.  He noted that the warming
from the late 19th century to 1940 was about 0.4ฐC,
which may approximate the transient warming over
the next two decades.  The problem is that climate
may change faster in the future than in the early
20th century.  (The average decadal warming may
be  as  much as 0.5ฐC,  rather  than  the  0.1ฐC
identified for earlier  years.)  Furthermore, the
analog takes one only as far as a 0.5ฐ C warming or,
in  the case  of paleoclimatic  records,  a  1ฐC
warmkig.lt does not indicate what happens in the
decades after the 0.5 to 1.0ฐ C level is reached.  In
addition, the analog may not represent the regional
distribution of climate associated with greenhouse

GCM Transient Runs

     The Goddard Institute for Space Studies has
modeled  how  global  climate may  change  as
concentrations of greenhouse gases gradually rise
over the next century.  This is  called the transient
run.  GISS has modeled  climate change under
several  assumptions of trace  gas  growth.  The
transient runs start in 1958 with the atmospheric
concentrations  of greenhouse  gases  that  existed
then.   The  concentrations  of the  gases   and
equivalent  radiative  forcing  were  estimated  to
increase from 1958 until an arbitrary point in the
future  according to several different assumptions
regarding trace gas growth. The GISS transient run
yields daily climate estimates from 1958 until that
arbitrary point.

     For example,  one of the  transient scenarios,
which is known as GISS A, assumes that trace gas
concentrations continue to increase at historic rates
and net greenhouse forcing increases exponentially.
The scenario is run from 1958 to 2062.  The end of
the transient  corresponds  with a global warming
equivalent to  that of the equilibrium climate from
the  doubled  CO2  run.   This  scenario does  not
account for  the  potential reduction  in CFC
emissions due to the Montreal Protocol or for other
activities that may reduce the growth in emissions.
GISS  B   assumes   a   decreasing   trace   gas
concentration growth rate such that climate  forcing
increases linearly (Hansen et al., 1988).  It stops in
2029.  GISS B  includes volcanoes,  while GISS A
does not.

     Since  the  GCMs  are used to produce  this
transient run, the advantages and disadvantages of
using this approach are the same as those described
in the  discussion of doubled  CO2 scenarios.   In
addition, the timing of the  changes estimated by the
GCMs is complicated by the uncertainties regarding
the growth of greenhouse gas emissions and the
roles of the oceans and clouds  in delaying climate
changes (Dickinson, 1986).


      This study used transient scenarios based on
 the GISS transient run because, of all the different
 approaches,  only  this  one  provides  internally
 consistent estimates of climate change and allows
 examination of the entire range of climate change
 between  current  conditions  and doubled CO2

      In creating the transient scenario, an approach
 similar to that  used for the doubled CO, scenario
 was employed. Since  relatively little  confidence
 exists  in  the  GCM's  estimates of  changes  in
 interannual and daily variability, the monthly means
 were calculated for each decade of the transient.
 This  process gives average decadal temperature,
 precipitation, and  other  changes.  The average
 decadal temperature changes in GISS A and B for
 the United States are shown hi Figure 4-3.

      As in the doubled CO2 scenario, the average
 meteorological changes from the  transient are
 combined  with a  historic time series.   What  is
 different from  the doubled CO2 scenario  is that a
 gradual change in  temperature and other variables
 is mixed with  a historic time series with its own
 variability. This can produce a regular oscillation.

                                            In this study, the historic time series 1951-80 is
                                       used, and the transient monthly statistics are applied
                                      'to the time series.  The procedure for creating the
                                       transient scenario was to first linearly interpolate
                                       between decadal means.  This smoothes out the
                                       sharp decadal changes from the actual transient
                                       GISS results and is shown in Figure 4-4(a).  The
                                       baseline 1951-80 weather data were repeated for 80
                                       years,  with  the last  20  years consisting of a
                                       repetition of the 1951-70 data. Figure 4-4(b) shows
                                       the average U.S. temperatures for 1951-80 repeated
                                       for 80 years.  The data transformations displayed in
                                       Figures 4-4(a) and (b) were done for data for each
                                       month for each grid box, site, and climate variable.
                                       The smoothed month-by-month transient data were
                                       added to the repeated 1951-80  data for  each site
                                       and variable. Figure 4-4(c) displays the addition of
                                       the smoothed average U.S.  transient temperatures
                                       with actual U.S. 1951-80 temperatures, repeated.
                                       Although there is a cooling from the 1950s to the
                                       1960s, followed by  a warming  in  the 1970s, the
                                       underlying warming of the transient, which is 3.7ฐ C
                                       by the  middle of the 2050s hi  GISS A, is much
                                       greater than the variability in the base period.


                                            Since the transient scenarios were also derived
                                       from  GCMs,  the  same limitations  concerning








           1980s 1990s 2000s 2010s 2020s 2030s 2040s 2050s

                      TRANSIENT SCENARIO A


•-   1
                                                     1990s    2000s   2010s

                                                       TRANSIENT SCENARIO B
Figure 4-3.  GISS transients "A" and "B" average decadal temperature change for lower 48 states gridpoints.

 Chapter 4
            1990  2000  2010  2020  2030   2040  2050  2060
                1951 AND 1980 REPEATED
Figure  4-4.
Transient scenarios (temperature
temporal  and spatial variability pertain as in the
doubled CO2 scenario. An additional limitation in
the transient scenario is  the rate of change.  The
GISS transient  runs assume a gradual  rate  of
change in temperature. The simplistic treatment of
ocean circulation in the  GCM affects the rate of
warming estimated by the model. Broecker (1987)
has shown that past climate changes may have been
abrupt.    Broecker, however,  analyzed  a global
cooling, and the changes occurred  over a  much
longer period than greenhouse warming. A sudden
warming could mean that significant effects happen
sooner and more suddenly than the  results of the
transient analysis used in this  study indicate.  The
inclusion of the 1951-80 base period in the scenario
yields short-term oscillations.

     In a few cases, researchers used meteorologic
 data from the 1930s as an analog scenario.  This
 scenario was used to provide additional information
 on the sensitivity of systems to climate change.  In
 a few  other  cases,  researchers  only examined
 paleoclimatic records.  In these cases, the goal was
 to  determine how a  system responded  to past
 climate change.

     EPA specified that researchers  were to use
 three  doubled  CO2   scenarios,   two  transient
 scenarios,  and an  analog scenario in this  study.
 Many researchers, however,  did not have sufficient
 tune  or resources to  allow for  the use  of  all
 scenarios.  EPA  asked the researchers to run the
 scenarios in  the  following  order, going  as far
 through the list as time and resources allowed:

     1.  GISS doubled CO2;

     2.  GFDL doubled CO2;

     3.  GISS transient A;

     4.  OSU  doubled  CO2;

     5.  Analog (1930 to 1939); and

     6.  GISS  transient B;

 Most researchers  were  able to use at least the GISS
 and GFDL doubled CO2 scenarios. Comparison of
 results across  studies may be limited because  of
 inconsistent use of scenarios.

 Sea Level Rise Scenarios

     Unlike the climate scenarios, the alternative
 sea level rise scenarios were not based solely on the
 differences  between various general circulation
 models.  Instead,  they were based on the range of
 estimates that previous studies have projected for
 the year 2100 (Hoffman et al., 1983,1986; Meier et
 al., 1985; ReveUe,  1983; Thomas, 1986), which have
generally considered alternative rates of greenhouse
gas emissions, climate sensitivity ranging from 1.5 to
4.5ฐC for  a  CO2 doubling, and  uncertainties
regarding ocean  expansion  and  glacial melting.
Estimates for the year 2100 generally range from 50
to 200 centimeters.

     This report uses three scenarios for the year
2100  -  50,  100,  and 200 centimeters  -- and
compares  them  to  the  current  trend  of  12
centimeters per century. Because most studies have
not reported estimates for  the intermediate years,
we followed  the convention of a 1987  National
Research Council report (Dean et al., 1987) and
interpolated sea level rise using a parabola. The
rates of sea level rise  assumed in this report are
displayed in Figure 7-8 in Chapter  7: Sea  Level
Rise. Because various coastal areas are also sinking
(and in a few cases rising), relative sea level rise at
specific locations was estimated by adding current
local subsidence trends.  Note  that sea level rise
scenarios are presented for the year 2100,  while
doubled CO, scenarios are  presented for the latter
half of the  21st century.

     In  this study,  the preferred approach  for
analyzing potential impacts of climate change was to
develop quantitative  estimates.  Most researchers
estimated impacts by running models that simulate
the relationship between weather and the relevant
system. The climate scenarios were used as inputs
into the  models.  Since the researchers had only
several months to do the analysis, they used either
"off-the-shelf"  models or analytic techniques.  In
many cases, existing models were calibrated to new
sites. This lack of time also limited the gathering of
new data to a few studies.

     A  drawback of using empirical  models of
systems to  estimate sensitivities is that the models
are applied to climates for which they were not
developed.  The models estimate relationships with
observed  climate.    This  relationship  is  then
extrapolated to an unprecedented climate.  It is
possible  that  in  the new climate  situation, the
statistical relationship maybe different owing to the
crossing  of a threshold or for some other reason.
With the  drawbacks of  empirical  models, the
current statistical relationships are the best basis for
quantitatively estimating sensitivities.

     For the most part, researchers analyzed the
potential effects  of climate change on  systems as
they currently exist. Although these changes may be
quite substantial, potential changes in populations,
the economy, technology, and other factors were not
considered.    In  some  cases,  researchers  ran
additional  scenarios  with  assumptions  about
technological and other changes.   In  addition,
potential  responses  to  climate   change  were
considered in some,  but not all,  cases. For these
and  many other reasons,  the results should be
interpreted only as an indication of the sensitivity of
current systems  to  global warming,  not as  a
prediction of what the effects will be.

     In some situations, quantitative models of the
relationship between climate and a particular system
did not exist. In those cases, other approaches were
used  to  try  to  identify  sensitivities.    Some
researchers examined how systems responded to
analog warmings.  In other cases, expert judgment
was used.  This consisted of literature reviews to
assemble information on sensitivities as they appear
in the literature, and workshops  and interviews to
poll  experts on how they thought systems would
respond to global warming.

     The scenarios used in this report help identify
the  sensitivities of  systems  to climate  change.
Because  of the  lack  of confidence concerning
regional estimates of climate change from GCMs,
we cannot predict impacts.  In order to predict the
effects of climate change, major improvements need
to be made in GCMs.  These  could take many
years.  In the meantime, we will continue to use
scenarios to identify sensitivities.  As with GCMs,
scenarios can also be improved.


     To  produce  better estimates  of  regional
climate change, both the resolution of GCMs and
the modeling  of physical processes  need to be
improved.  The GCMs used for this report had
large grid boxes,  in  which  major  geographic
features,  such as the Great Lakes or the Sierra
Nevada Mountains, which have  large impacts on
local .climate, were not well represented.  Ideally,
the  higher  the   resolution,   the   better  the
representation of geographic  features.  But each
increase in resolution means a  large increase in
computations and computing power needed to run
the model. Furthermore, at high resolutions, the
GCMs may require  new parameterizations.  The

 Chapter 4
 resolution should be increased at least to the point
 at  which major  geographic  features  are  well
 represented hi the models.

      It  is also important that the estimates  of
 physical  processes in the models be improved to
 increase  the  confidence  about estimates  of >the
 magnitude and timing of changes. Three areas need
 the most attention: oceans, clouds, and hydrology.
 The  oceans  play an important role in delaying
 climate  change and  have a large  influence on
 regional  climates.   However,  the  ocean models
 currently used  in  GCMs are  relatively simple.
 Ocean models that better simulate the absorption
 and  transport  of  heat  and  gases  would  give
 improved estimates of transient and regional climate
 change.   Clouds are a major feedback to global
 warming  and influence  regional climate.  More
 realistic  modeling  of clouds  by GCMs  would
 improve  the  estimates of the magnitude  of global
 warming  and  regional  change.   Finally,  more
 sophisticated hydrology in GCMs will yield better
 estimates of soil moisture and runoff, which will also
 improve  estimates of regional climate changes.


      The scenarios hi this report were based on
 changes hi average conditions, either at equilibrium
 (doubled  CO2)  or due  to a gradual change  hi
 average  underlying  conditions  (transient).   As
 pointed out hi Chapter 3: Variability, many systems
 are quite sensitive to changes in the frequency and
 intensity of extreme events. In the future, scenarios
 should incorporate change hi  variability to help
 identify  sensitivities  to  variability.    Transient
 scenarios can also be  improved.  Such scenarios
 should be useful for testing sensitivities to changes
 in long-term  climate trends as well as year-to-year
 variations. At the same time, it is  important to
 keep  scenarios simple.   More  detailed scenarios,
 involving  a lot of data (such as  daily data from
 GCMs) may be difficult to use.  The more detailed
 the scenario,  the more  likely  it will be applied
 incorrectly, which limits the  ability to  compare
results by different  researchers.    In addition,
scenarios should be simple, so the assumptions used
in  creating  them  can  be   easily understood.
Designers of scenarios will have to wrestle with the
competing desires  of  being  more  detailed  and
maintaining simplicity.

 Broecker, W.S.  1987.  Unpleasant surprises in the
 greenhouse? Nature 328:123-126.

 Dean, R.G., RA. Darylrumple, R.W. Fairbridge,
 S.P. Leatherman,  D.  Nummendal, M.P. O'Brien,
 O.H.  Pilkey, W.  Sturges,  R.L.  Wiegel.   1987.
 Responding to Changes hi Sea Level:  Engineering
 Implications.      National   Research  Council.
 Washington, DC:  National Academy Press.

 Dickinson, R.E. 1986.  How will climate change:
 The climate system and modelling of future climate.
 In:  Bolin, B., B.R.  Doos,  J. Jager,  and  RA.
 Warrick, eds. Scope  29: The Greenhouse Effect,
 Climatic Change and Ecosystems.  New York: John
 Wiley and Sons. pp. 221-231.

 Gates, W.L. 1985.  Modeling as a means of studying
 the  climate system.   In: MacCracken, M.C., and
 F.M. Luther, eds. Projecting the Climatic Effects of
 Increasing Carbon Dioxide. Washington, DC: U.S.
 Department of Energy.  DOE/ER-0237.

 Grotch, S.L. 1988.  Regional Intercomparisons of
 General  Circulation   Model  Predictions   and
 Historical  Climate  Data.   Prepared  for  U.S.
 Department of Energy. TR041.

 Hansen, J., I. Fung, A. Lacis, D. Rind, G. Russell,
 S. Lebedeff, R. Ruedy, and P. Stone. 1988. Global
 climate changes as  forecast by the GISS 3-D model.
 Journal of Geophysical Research 93:9341-9364.

 Hansen, J., G. Russell, D. Rind, P. Stone, A. Lacis,
 S. Lebedeff,  R.  Ruedy,  and  L.  Travis.  1983.
 Efficient  three-dimensional  global  models  for
 climate studies: Models I and II. Monthly Weather
 Review 3(4):609-622.

 Hoffman, J.S., D.  Keyes,  and J.G. Titus.   1983.
 Projecting future sea level rise.  Washington, DC:
 U.S. Government Printing Office.

 Hoffman, J.S., J.  Wells, and J.G.  Titus.   1986.
 Future  global warming and sea  level rise.   In:
 Sigbjarnarson G., ed.   Iceland Coastal and River
 Symposium. Reykjavik, Iceland:  National Energy

Manabe, S., and R.T. Wetherald. 1987.  Large scale
changes in soil wetness induced by an increase in
carbon dioxide.  Journal of Atmospheric Sciences

Meier, M.F. et al.  1985.  Glaciers, ice sheets, and
sea level.  Washington, DC:   National Academy

Mitchell, J.F.B. 1988.  Local effects of greenhouse
gases.  Nature 332:399-400.

Parry, M., T. Carter, N. Konijin, and J. Lockwood.
1987.   The  Impact  of  Climatic  Variations  on
Agriculture: Introduction to the IIASA/UNEP Case
Studies in Semi-Arid Regions. Laxenburg, Austria:
International Institute for Applied Systems Analysis.

Revelle, R.  1983.  Probable future changes in sea
level resulting from increased atmospheric carbon
dioxide. In: Changing Climate. Washington, DC:
National Academy Press.

Schlesinger,   M.E.,  and J.F.B. Mitchell.   1985.
Model  projections of  the  equilibrium  climatic
response  to  increased  carbon  dioxide.     In:
MacCracken,  M.D.,   and  P.M.   Luther,  eds.
Projecting the Climatic Effects of Increasing Carbon
Dioxide.  Washington, DC:   U.S. Department of
Energy. DOE/ER-0237.
Schlesinger,  M., and Z. Zhao.  1988.   Seasonal
Climate Changes Induced  by Doubled CO^  or
Simulated by the OSU Atmospheric GCM/Mrxed-
Layer Ocean Model. Corvallis, OR: Oregon State
University, Climate Research Institute.

Thomas, R.H.  1986. In: Titus, J.G., ed. Effects of
Changes  hi  Stratospheric  Ozone   and  Global
Climate.  Washington, DC:   U.S. Environmental
Protection Agency and UNEP.

Vinnikov,  K.Y., and NA. Lemeshko.  1987.  Soil
moisture content and runoff hi the USSR territory
with global warming. Journal of Meteorology and
Hydrology No. 12.

Wigley, T.M.L. 1987. Climate Scenarios.  Prepared
for  the  European  Workshop  in  Interrelated
Bioclimate and  Land-Use Changes. Boulder, CO:
National Center for Atmospheric Research.  NCAR


                                        CHAPTER 5


Global warming could significantly affect the forests
of the United States. Changes could be apparent in
30 to 80 years, depending upon the region,  the
quality of a site, and the rate of climate change.
There may be  northward shifts in species ranges,
dieback along  the southern reaches of species
ranges, and changes in forest productivity.  Other
stresses hi combination with climate change may
exacerbate these impacts. Different migration rates
and  climate sensitivities may result in changes in
forest  composition.      Without   large-scale
reforestation, large reductions in the land area of
healthy forests are possible during this century of
adjustment to climate changes.  Although climate
fluctuations,  timber harvests, disease outbreaks,
wildfires,  and other factors have affected forests
during the  last century, the magnitude of these
changes is substantially less than those projected in
response  to climate changes considered in  this

Range Shifts

•    The southern ranges of many forest species in
     the eastern United States could die back  as a
     result of higher temperatures and drier soils.
     The southern boundary could  move  several
     hundred to 1,000 kilometers (up to 600 miles)
     in a generally northward  direction for  the
     scenarios studied.

     The potential northern range of forest species
     hi  the  eastern United  States  could  shift
     northward as much as 600 to  700 kilometers
     (370  to 430 miles) over  the  next century.
     Actual northward migration could be limited to
     as little as 100 kilometers (60 miles) owing to
     the slow rates of migration of  forest species.
     Without reforestation, full migration of eastern
     forests to potential northern distributions could
     take  centuries. If climate change occurs too
     rapidly, some tree species may  not be able to
     form  healthy seeds, thus  halting migration.
    Reforestation  along  northern  portions  of
    potential forest ranges could mitigate some of
    these impacts.

•   If elevated CO2  concentrations substantially
    increase the water-use efficiency of tree species,
    the southern declines could be alleviated.

*   If climate stabilizes, forests might eventually
    regain a generally healthy status (over a period
    of several centuries).    In the  meantime,
    declining forests could be subject to increased
    fires, pest attacks, and replacement with low-
    value  trees,  grasslands,  and  shrubs.    A
    continually changing climate could result  in
    even greater dislocations among forests.

Productivity Changes

•   Dieback  along   the   southern  limits   of
    distribution of  many species  could  result  in
    productivity declines of 40 to 100%, depending
    on how dry soils become.

•   Productivity could increase along the northern
    limits of some eastern tree species, particularly
    as slow-growing conifers are replaced by more
    rapidly growing hardwoods.

Combined Impacts With Other Stresses

 •   Large  regions  of severely  stressed forests,
    combined with possible increases in fires, pests,
    disease  outbreaks,  wind  damage,  and  air
    pollution,  could   produce  major  regional
    disturbances.     These   factors  were   not
    considered for this report.

 ซ   Additional impacts of changes in forests could
    include reductions in biotic diversity, increased
    soil runoff and soil erosion,  reduced  aquifer
    recharge, changes in recreation, and changes in
    wildlife habitat.

 Chapter 5
Policy Implications

•   Institutions such as  the  U.S. Forest Service,
    state forest  agencies, and private companies
    should begin to consider how to factor climate
    change in their long-term planning.   Global
    climate change may need to be a factor in the
    Forest Service's 50-year planning horizon.

•   Where  TLS. forests are  clearly reduced  by
    climate change, forest agencies will  have to
    consider  intensive   strategies   to   maintain
    productivity.     For  example,  they  could
    undertake reforestation on a more massive scale
    than  now  practiced and possibly  introduce
    subtropical species into the Southeast.

•    A coordinated public and private reforestation
    effort, together with development of new and
    adapted silvicultural practices, would  also be
    required. Forests are major carbon sinks, so a
    large reforestation program would also reduce
    atmospheric CO2 concentrations, slowing the
 rate of global warming. This study did not evaluate
 the effectiveness of reforestation efforts.


    Forests occupy 33% of the U.S. land area and
 exist on some lands in all 50 states.  In total, they
 occupy 298 million hectares (738 million acres) and
 are rich in such resources as water and wildlife.

    Many biotic and abiotic factors  influence the
 condition of forests,  but climate  is the dominant
 factor  (Spurr and Barnes, 1980).   This  chapter
 summarizes the current knowledge and predictions
 concerning the effects of rapid climate change on
 U.S. forests.

 Distribution and  Ownership

    Eight major forest regions of the conterminous
48 states contain 84% of the forested ecosystems of
the United States (Figure 5-1).  The forested areas
                                         STUDY PLOT SITES
                                         •  Botkin et  ol.
                                         ป  0. Dovis
                                         A  Urban and Shugart
    HI Spruce/hemlock
 Pacific Northwest
  |\^\| Douglas fir/hemlock/fir
 Northern  Rockies
  HH Pine/fir/birch
 Southern  Rockies

  jffiQI Spruce/fir
  gg] Oak/hickory
     Southern pine

 Lake States
  K3 Spruce/fir
           Figure 5-1. Major forest regions of the United States and their primary tree groups.

Table 5-1.  Area of U.S. Forest Lands in 1977 by Federal, State, Private, and Other Ownerships (millions
           of hectares)3	
                                                         Commercial Forests
                                    Primary Tree Species     Federal  State   Industry  Indus   Other0   Total   Total



Northeast -
Lake States -
Central -
Southeast -
Northern Rockies -
Southern Rockies -
Pacific Northwest -
California - CA
Alaska - AK
Hawaii - HI

maple-beech-bi rch

loblolly, shortleaf
slash pine


pinyon- juniper-pine

D. fir-hemlock-fir

pi ne- f i r- redwood







0.4 '



































4.4 I







TOTAL 40.2 9.5 34.8 105.5 108.3 298.3 100
% TOTAL, 13.5 3.1 11.7 35.4 36.3 100
 a Hectare x 2.47 = acres.
 b Commercial forests are those capable of growing at least 1.4 cubic meters per hectare per year (20 cubic teet
   per acre per year) of industrial wood materials.
 c Other forests include county and municipal forests and those federal lands withdrawn from industrial and wood
   production for use  as parks, preserves, and wilderness.
 Source: USDA (1982).
 of Alaska and Hawaii represent the remaining 16%
 (Table 5-1).  Each forest region includes one or
 more forest types distinguished by the major tree
 species present.  As a general rule,  some types in
 each  region  have predominantly coniferous tree
 species    (i.e.,   evergreen,   needle-leaved,    and
 softwoods); other forest types are composed mostly
 of deciduous trees (i.e., tree species that are broad-
 leaved, have no winter fpliage, and are hardwoods).
 Forest types with a mix of coniferous and deciduous
 trees, however, are not uncommon.
    Superimposed over the natural distribution of
trees, forests, and ecosystems in the United States
is the human infrastructure.  Ownerships include
federal, state, and private lands (Table 5-1). Within
the forests classified as "commercial" (64% of 298
million hectares), the federal government ownership
of 40 million hectares (99 million acres) is primarily
in the national forest system managed by the U.S.
Department  of Agriculture's Forest  Service  (36
million  hectares or 91 million acres);  most of the
remainder is  managed by  the   Department  of

 Chapter 5
 Interior's Park Service, Fish and Wildlife Service,
 or  Bureaus  of Land  Management  and 'Indian
 Affairs.  State ownerships total 9 million hectares
 (23  million  acres).    Private lands  are  divided
 between those  of industrial forest companies (35
 million hectares or 86 million acres) and those of
 small, private landowners, who collectively have 106
 million hectares (262 million acres) (USDA, 1982).

     Another significant segment of American forests
 consists  of  those  maintained within urban  and
 suburban areas. Examples are community parks,
 greenbelts, roadside forests, and wooded residential
 and industrial zones (USDA, 1981).  These forest
 areas are important sources of outdoor recreation,
 wildlife habitat, and real estate values.  In total, the
 urban/suburban forests of the United States occupy
 approximately 28 million hectares (69 million acres)
 (Grey and Deneke, 1978).

     To the degree that all forest lands are owned by
 some individual or organization, all forest lands are
 under some form of management. A continuum of
 management  policies  exists, ranging  from lands
 intended to have minimal human intervention except
 for protection from catastrophic wildfire (e.g., some
 parks and most wilderness areas) to lands where
 silvicultural practices are intensively applied (e.g.,
 the  most productive federal, state, and  industrial
 forest  lands  dedicated  to growing  tree crops);
 (Table  5-2).  These forests under government and
 industrial management constitute roughly one-fourth
 of the  total and might be the easiest to manage
 under climatic impacts  simply because  they  are
 larger  blocks of  lands  already under  strong
 management commitments.
                     Value of U.S. Forests

                         Most populated regions in the United States are
                     located close to or within a forested region. For
                     instance, the Boston-Washington corridor is within
                     the eastern hardwoods. The populations of Atlanta
                     and the Southeast are  interspersed among  the
                     southern pine forests.  Chicago and  nearby  Great
                     Lakes  communities are surrounded  by the mixed
                     conifer-hardwood forests of that region, and the Los
                     Angeles to San Francisco populations parallel the
                     Sierra   Nevadas  to  the  east.     In   addition,
                     urban/suburban forests exist in or near most  of the
                     nation's cities.  Forests, therefore,  are part of the
                     environmental fabric and general habitability for the
                     majority of U.S. citizens.

                         All forests shed water to some degree, and two-
                     thirds of the water runoff in the contiguous 48  states
                     comes  from  forested ecosystems.   Precipitation
                     passes  through  forested  ecosystems as  canopy
                     throughfall or flows along tree stems, and then flows
                     along the ground surface or into the soil; eventually,
                     some of the water flows into streams.  Water yields
                     from U.S. forests provide about 750 billion  liters
                     (200 billion gallons) of water each day for major
                     uses such  as  irrigation, electricity production,
                     manufacturing, and domestic consumption. These
                     levels of demand are projected to continue to the
                     year 2030 (USDA, 1981).

                         A favorite use of forests is outdoor recreation.
                     Activities  include  hiking,   camping,   hunting,
                     sightseeing,  boating,   swimming,  fishing,  skiing,
                     sledding, and snowmobiling. A 1977 survey of U.S.
Table 5-2.  Percentage of Forest Lands by Level of Management within Four U.S. Regions (estimates for 1977)
                                  Forest                  Other                Reserved/
    U.S. regions	plantations3	  commercial	deferred0

        Rocky Mountains
        Pacific Coast
I* Intensively managed plantations.
  Moderately managed forests.
c Recreational and protected forests.
Source: USDA (1982).

households indicated  that  a majority of people
participated in outdoor recreation four or more
times each year (USDA, 1981).

    About 190 million hectares (470 million acres),
or 64% of the total U.S. forested ecosystems, are
highly productive commercial forest lands.  These
lands represent  about 10% of the world's forest
area, but they supplied  nearly a  quarter of the
world's industrial forest products in the late 1970s
(USDA,  1982).  In 1980, 1.7 million people were
employed in timber-based  occupations across the
United States.   Such employment is basic to the
economic well-being  of many small towns  and
communities (Schallau, 1988).  The total value of
timber products harvested  in 1972 was about $6.4
billion, and the total value  after such processes as
manufacturing,    marketing,   transport,   and
construction amounted to $48 billion, or 4% of the
nation's  gross national product.  In 1979, timber
product  exports  and imports were valued at $7
billion and $9 billion, respectively.  Looking ahead,
the consumption of wood  products in the United
States is projected  to increase between current
levels and the year 2030 (USDA, 1982).

    Scientific understanding of forest ecosystems
has greatly advanced  with  each decade  of this
century.     Yet  the  literature  contains  little
information concerning the direct or indirect effects
of climate  change on the complex biological and
physical  processes in  forest ecosystems.   Some
insights are gained from paleobotanical studies of
past rates  and magnitudes  of ecological change
during glacial-interglacial cycles, as well as changes
in the  species composition of forested ecosystems.
Similarly,  observations  of  forest  responses to
unusual drought or other weather extremes provide
some knowledge.  Estimates of rate, magnitude, and
quality of  change have  also been  derived  using
computer models developed by plant ecologists or
forest  management scientists for  other objectives.
Their validation for understanding how a forest can
adapt to climate change is only in the initial stages.

    Climate  is  a primary determinant of existing
forests. The ranges of annual average temperature
and rainfall variation determine  global forest
distributions relative to different  biotic regions
(Figure 5-2). Substantial increases in temperature
                                 100      '     200          300

                                        MEAN ANNUAL PRECIPITATION (cm)
  Figure  5-2.  Approximate distributions of the  major groups of world biomass based  upon mean annual
  temperatures and precipitation (Hammond, 1972).

 Chapter 5
 or decreases in rainfall could, for example, produce
 a shift from  a forest to a grassland  type.  Thus,
 accelerated climate change resulting from human
 activities and related effects  on U.S.  forestsds of
 high concern to citizens and policymakers alike.


     Vegetation has been in an almost constant state
 of distributional change and adjustment due to an
 almost constantly changing climate over the past
 10,000 years and even over the past several hundred
 years (Spurr and Barnes, 1980).  Lines of evidence
 come from studies of fossils,  tree rings, carbon-14
 dating, plus peat and pollen analyses (Webb, 1987).

     Historical climate changes appear to have been
 associated with such phenomena as fluctuations in
 solar radiation, earth orbit variations, and volcanic
 activity.   Evidence of repeated continental glacial
 advances  and contractions   in  the  Northern
 Hemisphere dramatically illustrates the large-scale
 effects of global climate change.

     In response to  the glaciation, species shifted
 south.  Evidence from fossil  pollen, for example,
 indicates a southward shift of spruce into Georgia
 and east Texas during the last glacial advance and
 treeless  tundra in the Great  Lakes region (Spurr
 and  Barnes,   1980).    During  the   maximum
 interglacial warmth of 6,000  to 9,000 years  ago,
 which was 1.5ฐC  (2.7ฐF) warmer than the present
 temperature level, plant zones were one to several
 hundred kilometers  (60  to 250 miles) north  of
 present distributions.


    All forested ecosystems experience change on
 both spatial and temporal scales; each biological and
 physical forest component may respond to climatic
 variation on different spatial and temporal scales.
 For example,  microorganisms, insects, and birds
 come  and go with  relatively short-term climatic
 variation; shrub species' abundances vary within the
 timespan of decades; trees, once established, could
 persist  for centuries.    This  understanding  is
 important from the perspective of climate  change,
since it implies that forested ecosystems  do not
respond as a unit, but in terms of parts.  Different
parts  respond differently;  consequently,  future
forested  ecosystems under   a  rapidly  changing
climate could be quite different from those existing
     At the expected rapid rate of climate change,
 the  potential rates of  forest  migration would
 become a major concern.  Migration rates vary by
 species. Paleorecords of the Holocene (10,000 years
 ago to present) show that extension of ranges for
 tree species of eastern North American (in response
 to glacial retreat) varied from 10 to 20 kilometers (6
 to 12 miles) per century for chestnut, beech, maple,
 and balsam fir  (Zabinski and Davis, Volume D).
 Other  species within  the  oak  and pine  groups
 extended at faster rates, i.e., 30 to 40 kilometers (19
 to 25 miles) per century.  It should be noted that
 there is  some  uncertainty  as to whether these
 migration rates were in response to glacial retreat
 plus climate warming or primarily warming alone.


    Knowledge  of causal  links  between weather
 patterns  and forest  response is fundamental  to
 projecting growth and composition effects resulting
 from climate  change.  Another requirement is to
 understand the  climatic influences on processes
 influencing populations of forest plants and animals.
 These include such phenomena as fires, windstorms,
 landslides, pest outbreaks, and other disturbances
 that affect survival and  subsequent colonization by
 different species.  Furthermore, the processes that
 control the dispersal  of seeds through a mosaic of
 different  ecosystem  types (such  as forest patches
 interspersed   with agricultural  lands,  wetlands,
 grasslands, and  other land-use groups) must be
 clearly defined.

    Among the  important factors  now known to
 influence  the growth and distribution of forests are
 the following.


    The optimum  temperature for growth depends
 upon the tree  species  and other  conditions.
 Warmer temperatures usually increase the growth of
 plants.  However,  high temperatures can decrease
 the growth of plants or cause  mortality  where
 temperatures  greatly exceed optimum ranges for
 growth.    Cold  temperatures  can  limit  plant
 distributions by  simply  limiting growth at critical
 stages or by dkectly killing plants.


    Too much or  too little precipitation can limit
forest production and survival. Too much rainfall in

some areas can cause flooding or raise the water
table, thus drowning roots by reducing soil air that
contains oxygen required for respiration or by
promoting fungal attack.  Too little  rainfall can
reduce  growth,  cause  susceptibility  to  fire or
pestilence, and possibly  kill plants.  The seasonal
timing  of  rainfall is more important than total
annual rainfall, although forests also require some
minimum total annual rainfall (see Figure 5-2).

CCX. Concentration

    High CO, concentrations could increase tree
growth  through increases  in photosynthesis rates
and  water-use  efficiency  (primarily hardwood
species) when water and other nutrients  are not
limited  (Strain and Cure, 1985).  Plant responses to
CO2  have been  investigated largely in growth
chambers and are difficult to extrapolate to the real
world.  Responses are varied and do indicate some
measure of adaptive capability most likely imparted
from ancestral exposure to much higher and lower
levels in the geologic past.  However, in natural
situations, water nutrients or temperature usually
are limiting factors  in forest growth, thus making
the impacts of CO2 enrichment uncertain. If water-
use efficiency increases, then tolerance to drought
might increase, ameliorating declines in southern
parts of ranges. Unfortunately, the current state of
knowledge does not allow generalizations on this

    Another important relationship between forests
and CO2 is the role forests play as carbon sinks.
Globally,  forest vegetation and supporting soil
contain about 60% of the organic carbon stored on
world land surfaces.  This organic carbon is largely
cycled   between  forest   ecosystems  and  the
atmosphere by photosynthesis (uptake of CO2) and
respiration (COu release) in the plants (Waring and
Schlesinger,  1985).    Anthropogenically  caused
reductions  of  forests  either   directly  (e.g.,
urbanization, mismanagement)  or  indirectly (as a
response to CO2-induced global warming)  would
tend to increase the "greenhouse effect."


     The amount of sunlight bathing an ecosystem
sets the upper limit on net primary productivity.
Thus, the tropics exhibit higher productivity than do
the boreal regions.  This potential productivity
would,  of course, be limited by other climatic effects
such   as  drought,  cold,  heat,  and   natural
disturbances, and by the time required for forests to
shift into new ranges.  The length of day exerts
considerable control on physiological processes such
as release from and onset of dormancy.  Significant
northward shifts of forests would alter their day-
length regime, producing uncertain results.

Nutrient Status

    In addition  to  climate,  most forest growth is
strongly influenced  by availability of soil nutrients.
Disturbances over  vast regions,  such as drought
followed by fire, can release  large quantities of
essential  nutrients  into  the atmosphere or into
surface  waters. This leaves soils nutrient deficient.
Lengthy periods of soil development are usually
required to replenish the soil nutrients before a
large, mature forest stand can be supported.  In
turn, soils reflect properties of the forests that they
support.  This  results  from decades of nutrient
uptake,   litterfall,   decomposition,   and   other

Atmospheric Chemistry

    Much of the nutrient budget of forests involves
deposition of chemicals from  the atmosphere as
gases, aerosols, or particles, or in solution with
water as precipitation. Although most of these act
as nutrients, some produce acid deposition that can
leach important soil nutrients (e.g., SO4=), produce
a fertilizing effect (e.g., NO3"), or damage leaf tissue
(e.g., O3). Climate  change will alter transport paths
of air pollutants, and increased temperature could
increase  the  rates at  which  gases  convert  to
deleterious forms.


    Almost continually,  forests experience natural
disturbances or stresses from biotic or abiotic agents
alone or in combination. Examples are insect and
disease  outbreaks, plant  competition,  wildfire,
drought, cold extremes, and windstorms.

    These  disturbances, which  are  among  the
primary  factors   controlling  the   successional
processes in forests (Pickett and White, 1985), may
range from an opening of small gaps in the canopy
as the result of single  tree death or of windthrow
occurring when trees are blown  down by  heavy
winds  (predominant successional  mechanisms in
eastern hardwoods) to  large  clearings from fire,

 Chapter 5
 windthrow, or pestilence (predominant successional
 mechanisms in western forests).

 Landscape Processes

     The horizontal movements of materials such as
 soil and biological organisms, together with human
 disturbances across  the landscape,  are  critical  to
 processes  controlling  tree  migration,  species
 diversity  in forests,  and  the spread of  fire,
 windthrow, and pestilence effects. These processes
 are very poorly understood; quantification  in the
 emerging  field  of  landscape  ecology is  just

 Multiple Stresses

     In  general,  trees or  forests stressed by one
 factor,  e.g., accelerated climate change,  are more
 susceptible  to   natural   stresses  (secondary
 disturbances) such as insects, disease, or invading
 weed species.  The concept of multiple stresses
 leading to forest declines is becoming more widely
 recognized  (Manion, 1981).   Regional climate
 changes, even if temporary, frequently predispose
 forests to damage by other natural or anthropogenic

    Concern regarding effects of climate change on
U.S. forests has prompted several excellent reviews.
One  of the most comprehensive (Shands  and
Hoffman, 1987) was  the  result  of a conference
sponsored by EPA, the National Forest Products
Association, and the Society of American Foresters.
While pointing  out the high uncertainty associated
with current predictions of climate change, several
authors suggested  that if predictions are true,
distributions of key forest species  in the  United
States will change significantly.

    Other recently produced compilations broadly
consider forest effects along with  other  impacts
(e.g.,  those  on  agriculture, prairie  land, and  the
Great Lakes) (White, 1985; Titus, 1986; Meo, 1987;
Tirpak, 1987). These reviews are largely pioneering
efforts and some overlap occurs, but each presents
some key points.
     The methods used in the previous studies are
 quite similar to those used in this report.  They
 include  computer modeling of forest processes,
 literature surveys, studies of fossil evidence, and
 empirical relationships constructed by experts. The
 estimates of future  change  produced from  these
 studies are generally based on the output of one or
 more of the general circulation models  (GCMs)
 used for this report.   Thus, the results of the
 previous studies are consistent with those reported

    Six studies on forest effects contributed to the
 regional case studies reported in this volume. The
 purpose was to use existing data bases analyzed in
 new ways to estimate effects on U.S. forests from
 climate change scenarios.  The selection of the six
 studies was based upon three criteria:  use  of
 established  statistical methods;  hypotheses testing
 concerning causal mechanisms;  and selection of a
 mix of studies that complemented each other, such
 that the strengths in one approach might overcome
 the weaknesses of another.

    This report focuses primarily on forests within
 the contiguous  48 states.   It is  worth  noting,
 however, that the largest magnitude of warming is
 expected in the northern latitudes encompassing the
 boreal forest and other forest types in Alaska and
 Canada.   Thus, these large forests could also be
 under significant risk from climate warming.

Design of the Studies

    Characteristics of the six  studies are briefly
listed in Table 5-3.  With the exceptions of the
Overpeck  and Bartlein study and the Woodman
study, the methods are discussed in  the regional
case study chapters and  will be mentioned  only
briefly here. All of the forest studies are in Volume

    Two studies used correlations between  tree
distributions and climate  (Overpeck and Bartlein;
Zabinski and Davis).   Overpeck and  Bartlein's
approach consisted of correlating the modern pollen
distributions of important tree species with January
and July mean temperature and annual rainfall.

Table 5-3.  Principal Investigators, Regional Focus, and Method of Approach for the Regional Forested
           Ecosystem Studies
    Principal investigator
    Overpeck and Bartlein

    Urban and Shugart

    Botkin et al.

    Zabinski and M. Davis


    Woodman et al.
Eastern North America

Southeast Uplands

Great Lakes

Great Lakes


Southeast, California,
    and National
Correlation and fossil studies

Forest dynamics model

Forest dynamics model


Fossil studies

Literature review
The correlation was then tested by reconstructing
past pollen distributions from general circulation
model simulations of past climates (during the most
recent glacial-interglacial cycle) for each species and
comparing them to observed pollen distributions
from those periods.   Future pollen distributions
were then constructed from the expected doubled
CO2 climate projected from  the different model
climate   scenarios.    The   correlations  were
constructed on modern pollen distributions, rather
than  tree  distributions,   to  allow  the  direct
comparison to fossil pollen data.   Modern pollen
distributions are similar to, but not exactly the same
as, modern  tree  distributions.   The verification
studies  indicated  that   the  approach   works
reasonably well at a coarse spatial resolution. That
is,  northern trees are in the north and southern
trees are in the south, with the regional patterns
being reasonably well represented.

     The  approach  of Zabinski and Davis was
essentially the  same  as  that of  Overpeck and
Bartlein,   except   that  the  correlations  were
constructed   from   the    actual   modern  tree
distributions rather than from the  modern pollen
distributions (see Chapter  15: Great Lakes).

     Two  of the studies used  computer models of
forest dynamics (Botkin et  al.; Urban and Shugart).
Growth  characteristics  of  each tree  species
occurring in the study region are used by the models
to   determine  the   growth  and  development
ofindividual  trees on a  site.    These   growth
characteristics include sUch attributes as maximum
age,  maximum height,  maximum  diameter, and
                ranges in tolerance to stresses  of temperature,
                moisture, and shade. Both studies explored forest
                response starting with bare ground on a range of
                soil types from well drained to poorly drained.
                Forest  growth  simulations  from  bare  ground
                represent conditions after a fire, logging, or similar
                disturbance.  Mature stand simulations are  useful
                for investigating the potential response of present
                forests to gradual climate change in the immediate

                    For   the    California   case   study,   Davis
                reconstructed  vegetation  patterns  in the  Sierra
                Nevadas  from  fossil  pollen  studies  for  the
                interglacial warm  periods that occurred between
                about  6,000  and  9,000  years   ago.     These
                reconstructions  represent possible analogs  of  a
                future warm period at the lower magnitude of the
                predicted future warming.

                    Woodman conducted a literature review for the
                Southeast  and  California forested  regions and
                peripherally for the entire nation.  The purpose was
                to ascertain the attributes of  the forest resource in
                terms   of  extent,  ownership,   economic and
                recreational value, and policy considerations.


                    Although  predicted  effects  vary, these  six
                analytical studies have results that  are collectively
                consistent enough to advance our knowledge and
                justify concern regarding the  future of U.S. forests
                under  rapid   climate  change.   The range of

Chapter 5
predicted effects is large; however, uncertainties
exist regarding (1) the climate scenarios; (2) the
kind and rates  of responses of  individual tree
species; and (3) changes in forested ecosystems as
a whole resulting from environmental change.  All
of these factors significantly influence the precision
and accuracy of the results.

    A  major uncertainty in the simulation model
approach involves the rates of species dispersal into
a region. The current generation of models has no
dispersal mechanisms. A species is simply present
or it is not present.  For example,  Botkin et al.
excluded most southern tree species so that their
dispersal was unrealistically nonexistent, and these
southern species could never enter the Great Lakes
region.  But  if  they had been included  in the
simulations, these species would have entered the
northern forests  at the same rate as the climate
change.  This would have assumed dispersal rates
far in excess of reality. This limitation can, in part,
be overcome by studies, such as those of Zabinski
and Davis, that provide some insight  into  actual
dispersal  rates   and  species   migration.    The
simulations  did  not  consider the  impact  of
transplanting southern species in these areas.

    The  timing of  forest  declines as estimated by
the models should be interpreted with caution.
Declines  are  triggered  by   periods  of  high
environmental stress. Forest models are usually not
operated far beyond current conditions, such as for
extremely dry soils. Therefore, the extreme climate
simulated by these models may not estimate the
timing and behavior of forest declines as accurately
as desired. It should also be remembered that there
is much uncertainty concerning the  rate and timing
of the  climate change itself.

    A  further cautionary point is that although the
models considered temperature limitations, nutrient
deficiencies,  and  soil   moisture   stress,   other
important factors  might  affect the  timing and
magnitude of tree responses.  Examples of factors
in need of consideration include disturbance effects
(e.g., impacts from wildfires, pests, and pathogens),
age-dependent differences in tree sensitivities to
stress (e.g., older trees are often more susceptible),
and potential CO2-induced increases in water-use

    The models also carry assumptions about the
environmental controls of species limits.  In most
cases these assumptions are reasonable, given that
indices  of  environmental  stress,  such  as  July
temperature or annual rainfall, are usually related to
factors that more directly affect plant growth,  such
as  accumulated warmth  or summer  drought.
However, large uncertainties exist in some instances.
This is particularly true with regard to the climatic
controls  of the southern limits of  southeastern
forests, simply because of their association with the
continental  margin.   Does the climate  at  that
latitude represent the actual climatic limitation to
the distributions, or are the species simply stopped
by  a geographic barrier?   No one  really knows.
These uncertainties were partially  addressed by
Overpeck and Bartlein, who compared their fossil
pollen approach to the modeling approach.  The
two approaches use similar relations to climate, and
both can be used reasonably well to simulate forest
distributions in the geologic past.

    Several uncertainties with the pollen-climate
correlation  approach  limit   its  precision  and
accuracy.  First, many of the plant taxa used in the
study are plant genera (e.g., pine, oak) rather  than
species, and thus the  simulated  results  are  not
taxonomically precise.   Second,  the results  are
applicable  only on  a regional scale; local scale
predictions  are  not  made.    Third, and  very
significant, the simulated results assume that all the
plants are in equilibrium with the new climate.
Rates of dispersal vary between species, and several
hundred years may pass before plant communities
are again hi equilibrium with climate.  How this lag
would affect  plant  community  dynamics is  not
addressed hi this study and is an important research

    The   paleoecological  analysis  of the  past
vegetation hi the Sierra Nevadas (O. Davis) presents
several uncertainties.  First, differences with respect
to weather variations (i.e., season to season and year
to year) could produce strikingly different types of
vegetation.  Also, there is much uncertainty about
what the most appropriate analog period might be
— or if one even exists.  Furthermore, the rate of
climate change in the future is predicted to be much
faster than the rate of climate change during the
past 20,000 years.  Lags in the response of species
to the future climate could strongly affect the type
of forest at any one location,  whereas in  the past,
with a more slowly varying climate, lags in species
response were not  as  important hi  determining
forest composition.

    All  of the  studies are deficient  in some  very
important processes controlling forest responses to

climate, particularly disturbance regimes  such as
fires, windstorms,  hurricanes, landslides, and pest
outbreaks. Over some forest areas, periods of cloud
cover  could  change.    This  is  an  important
uncertainty, for  if the annual total is significantly
increased, reductions in solar radiation could mean
reduced photosynthesis and thus less forest growth.

    In addition, the responses of mature trees to
elevated   CO2  under  conditions  of   moisture,
temperature, or other nutrient limitations remain
largely unexplored. Most research on elevated CO2
on trees has been performed in controlled chambers
using seedlings, and  results  show an increase in
photosynthesis and improved water-use efficiency in
some cases (Strain and Cure, 1985). However,  the
seedlings   were  not  previously  grown   in  or
acclimatized to high CO2 environments.  Evidence
has shown that plants grown under high CO2 will
respond differently to changes in temperature, light,
and moisture conditions (Strain and Cure, 1985).

    Another shortcoming  is  that  methods  to
extrapolate CO, fertilization results from laboratory
experiments to the natural world are limited, and an
understanding of  regional  changes in  water-use
efficiency  is  even more  limited.   Furthermore,
complex  interactions  between fertilization  effects
and changes in water-use efficiency  can produce
unexpected problems such as increased heat loads
due  to effects  on evaporation cooling.   These
interactions are not  well understood  but could
produce major regional changes in forest responses.
Therefore, it is not yet possible to quantitatively
incorporate the direct effects of CO2 on forests into
studies such as these.  Further, if water or nutrients
are limiting  to  forest growth,  they  would likely
exceed the fertilization effects of elevated CO2.
Also, forest canopies at optimum development have
multilayered leaf areas so that light limitations exist
for the lower portion of the foliage in addition to
frequent  water and nutrient limitations.   This adds
further weight to  the belief that CO2 enrichment
may not significantly affect forest productivity.


    The six studies conducted for EPA consistently
indicate that climate changes would significantly
affect the natural forests of the United States. The
distribution of healthy forests in the eastern United
States appears to become greatly reduced from their
present areas during the next century (Figures  5-3
and 5-4).  This results from a very slow northward
migration coupled with a fairly rapid decline in the
southern and western parts of species ranges. Drier
forest conditions in the United States, induced as
much by increased temperature as by changes in
rainfall, would mean less tree growth and therefore
reduced forest productivity in general.

    The  forest  simulation  models provide  an
indication  of  the  importance  of uncertainties
imparted  by the climate  scenarios.  The  climate
scenarios differ primarily in their representation of
regional rainfall  patterns.    The  model  results
indicate that temperature has a  large  effect on
forest health, either directly through cold and heat
stress or indirectly through exaggeratedg drought
effects.  Thus, the overall characteristics of forest
responses are remarkably similar among the  three
climate scenarios. However, this generalization is
uncertain because models usually do not incorporate
all possible mechanisms of impact.


Eastern Forests - Northern Limits

    All of the study results  suggest a northward
expansion of most eastern tree species (Figure 5-3
displays results from Overpeck and Bartlein).  That
is, spruce, northern  pine, and northern hardwood
species  would  move north by   about 600-700
kilometers (375-440 miles) into the Hudson Bay
region of the Canadian boreal forest (Overpeck and
Bartlein;  Zabinski and  Davis).    New England
coniferous forests would be replaced  by  more
hardwood forests and especially by the oak species
from the eastern mid-United States (Botkin et al.;
Overpeck and Bartlein; Zabinski and Davis). As
the northern mixed forests shift from spruce-fir to
sugar maple, some sites could actually triple their
present productivity (Botkin et al.).

    Additionally, southern pine species could shift
about 500 kilometers (310 miles) into the  present
hardwood forest lands of eastern Pennsylvania and
New Jersey  (Overpeck and  Bartlein;  Urban and
Shugart; Solomon and West, 1986; Miller et al.,
1987).   Depending  upon the severity  of  climate
change, Urban and Shugart estimated that near the
northern limits  of slash pine in East Tennessee,
aboveground woody biomass in 100 years  could
range  from  little change to  an  extremely low
biomass with almost no  trees  (i.e., a grassland,
savanna,  or  scrub).   However, even  with  little
decrease in productivity, species shifts would alter

Chapter 5
           Current Climate
              Spruce      Birch
N. Pines       Oak
           S. Pines    Prairie  Forbs
           GISS Model Output
              Spruce      Birch
N. Pines
S. Pines    Prairie Forbs
Figure 5-3. Maps of eastern North America depicting present distributions of major forest genera and herbacious
vegetation compared with potential future distributions after reaching equilibrium with the climate predicted for
doubled CO^.  The comparison is based upon (A) simulations using modern pollen data and simulated future
pollen abundances for each of the three doubled CO2 scenarios: (B) GISS; (C) GFDL; and (D) OSU.  The
three levels of shading in each scenario map indicate estimated future pollen abundances ranging from 20%
(darkest or  strongest chance of future distributions) to 5% and 1% (lightest or least  chance  of future
distributions) (Overpeck and Bartlein, Volume D).

              Sugar Maple
                     Present Range

                       JH] Potential Range

                       • Inhabited Range
                                            Range After 2050: GISS
              Range After 2050: GFDL

                      Scale  0 400Km
Figure 5-4.  Present and future geographical range for sugar maple (Zabinski and Davis, Volume D).
the forest  composition from shortleaf to loblolly
pine, a more commercially valuable tree species.

Eastern Forests - Southern Limits

    Ultimately, forest  decline and mortality could
truncate southern distributions of tree species by as
much  as  1,000  kilometers  (625 miles) in many
northern hardwood  species (Zabinski and Davis;
Overpeck and Bartlein) or by as  little  as  a few
hundred kilometers (about 120 miles) in southern
pines and hardwoods (Urban and Shugart; Solomon
and West,  1986).    Under the  driest scenario
(GFDL),   Zabinski  and  Davis  estimate  local
extinction  in the  Great Lakes  region  of many
eastern tree  species  such as eastern hemlock and
sugar maple (Figures 5-3 and 5-4). These estimates
bear considerable uncertainty for all species.

    These uncertainties are particularly true for the
southern limits of southeastern species that border
the  continental margin.   The  actual  southern
climatic limitations of these species  are not  well
known (Urban and Shugart).  Nevertheless, under
the most severe climate scenario in the Southeast
with increased temperatures and decreased growing-
season precipitation, Urban and Shugart's results
suggest that  the 18  tree species they considered
would  no longer grow in the southern half of the
region. Present forest lands in the region would be
replaced by   scrub, savanna,   or  sparse   forest
conditions.  This estimation results from scenario
conditions of heat that would exceed the tolerance
limits for most  tree species.  Under the mildest
scenario (OSU), even forest areas in South Carolina
and southward would be marginal, supporting about
half their current biomass.

    Biomass accumulations in 100 years for mature
natural  forests in productive  sites in  the  Great
Lakes region could be reduced to 23-54% of their
present  values (Botkin et al.; Solomon  and West,
1986).  On poor sites, forests could be converted to
grassland  or savanna with very low  productivity,
ranging from 0.4 to 28%  of their present values.

Western Forests

    Similar projections were made for six western
coniferous species: ponderosa and lodgepole pine,
Douglas-fir, western hemlock, western  larch, and
Englemann  spruce  (Leverenz  and  Lev,   1987).
Estimations are mixed for the West.  Because of
the mountainous conditions in the West,  upslope
shifts are possible for Douglas-fir, ponderosa pine,
and  western  hemlock  in  the  northern  Rocky
Mountains.  In the coastal mountains of California
and  Oregon, Douglas-fir could  shrink  in the
lowlands and be replaced by western pine species
(O. Davis; Leverenz  and Lev,  1987).  Overall, the
western forest lands are estimated to favor more
drought-tolerant tree species, such as the hard pine

Chapter 5
                                                               B. SOUTH CAROLINA FORESTS
   ฃ  140








                                                            —— No Climate Change
                                                             	GISS A
Figure 5-5.  Estimated changes in biomass of mature forests in Mississippi (A) and South Carolina (B) under
the GISS transient climate change scenario (Urban and Shugart, Volume B).
group, at the expense of fir, hemlock, larch, and
spruce species.

    If regional drought persisted, the frequency of
fires  could increase,  significantly  reducing total
forested   area.    Also,  with  massive  upslope
movement, some species could be  pushed off the
tops of mountains into local extinction.

    No quantitative estimates have been derived for
productivity for the western forests  under potential
warming conditions.   However, using the analog
approach of Davis, under the most severe conditions
projected for California, the species composition of
the west-side Sierra Nevada forests would become
more similar to that of the east-side forests. This
could reduce the standing biomass to about 60% of
current levels.

Rates of Decline and Migration

    In the Great Lakes region,  significant forest
decline and forest  compositional change could
become evident within 30 to 60 years (Figure 5-5A;
Botkin et  al.).   In the Southeast region, forest
declines  could become most evident in 60 to 80
years with declines in the drier  western portions
occurring even earlier, perhaps in  about 30 years
(Figure 5-5B  and C);  Urban and Shugart).   As
previously discussed in this chapter (see Limitations)
there  is  considerable uncertainty  about these
   .   These rapid declines, coupled with the expected
  magnitude of climate change, raise the question of
  how fast  forests can migrate.  Based upon fossil
  records, Zabinski and Davis have estimated that the
  maximum dispersal rate of several tree  species in
  response  to the last glacial retreat was roughly 50
  kilometers (30  miles)  per century.   Under  the
  expected  rapid  warming,  they estimated that a
  dispersal rate of about 1,000 kilometers (600 miles)
  per century would be required to maintain species
  distributions  near  their  current extent.   Such
  migration rates are  doubtful, suggesting greater
  reductions in species ranges under  rapid climate
  change, with  declines in the drier western portions.

  Mechanisms  of Migration

      Distribution changes (i.e., migrations) suggested
  by  these  studies  must 'be  considered  carefully.
  Reproductive processes  are  essential  for  the
  migration of tree species across the landscape. For
  many tree species,  climate  change  could reduce
  natural regeneration in an  existing  location and
  introduce the  species  at different  latitudes  or
  altitudes.  Reproductive processes in trees, such as
  flowering, pollination, seed set, seed germination,
  and  seedling competitive  success, are particularly
  sensitive to climate.

      Specific  regional climate scenarios  vary as a
  function  of  the  GCM.    All  scenarios  estimate
  increases  in  temperature; however, some include

increases in rainfall, and others have decreases. The
northward shifts of species appear to result from a
release  from  cold   temperature  stress,  which
normally freezes flowers, seedlings, and even adult
trees. However, the western and southern limits of
eastern   tree  species  appear  to  result  from
insufficient moisture and excessive heat stress, which
primarily affect sensitive life history stages but can
also affect mortality rates of adult trees. Though
difficult to  detect in  the  early phases of rapid
climate change, tree mortality is sensitive to chronic
moisture stress and mortality rates would likely
increase  among the major  forest regions  of the
United States.

    Two points  are  important about  regional
uncertainties of future rainfall distribution.  First,
changes in the seasonal distribution of rainfall are as
or more important than relatively small changes in
the annual total. If summer rainfall decreases while
winter  rainfall  increases,  the   trees  may still
experience   summer  drought  stress.   Second,
evapotranspiration is a log function of temperature.
Therefore, as temperature goes up, water loss from
trees and soils can increase tremendously. If minor
increases in rainfall are not sufficient to override the
evapotranspirational  losses  of  water,  drought
impacts will pervade.  Both of these  mechanisms
appear to dominate the forest impacts in this study.

    All of the study approaches used under all of
the climatic scenarios estimate major forest declines
in  the  southern  parts  of  species  ranges and
expansions to the north.  These  declines, resulting
primarily from drought stress, would occur despite
the differing rainfall predictions among the climate
scenarios used in this study.  Global precipitation is
generally projected to increase slightly with global
warming (see Chapter 4: Methodology), but it is not
known whether this increase would be sufficient to
compensate for potential increases in plant moisture
stress caused by higher temperatures. Precipitation
in  some regions may decline.   Droughts would
become  more common.   The  western limits  of
eastern forests could similarly retract as the climate

    Existing forests probably would not shift intact,
but would change in composition.  Variations in
migration rates and sensitivities to weather variables
produce individual responses to climate  change.
These changes are consistent with the well-known
dynamic nature of ecosystems and were projected
for the forests of all regions. In the Great Lakes
region,  for  example,  beech  could  decrease  in
abundance (Zabinski and Davis), and  birch and
maple  could  increase (Botkin et al.).   On some
lands, forest  productivity could remain  about the
same as today, but changes to less  economically
important species could be significant.

    Not  considered  quantitatively in any  of the
studies are changes in forest disturbance regimes.
These  changes should not be considered  lightly.
Extreme and more frequent climatic variations (see
Chapter  3: Variability) could  cause  much  higher
mortality in  U.S.   forests   than  the  current
experience. Although little is known as yet, some
locations  may experience  an  increase  in the
frequency of extreme weather events, for example,
wind, ice, or  snow storms, droughts,  and flooding.
Besides the direct damage these events can cause,
they can predispose  forests  to damage  from
secondary stresses such  as insects,  disease, and

    The effects of doubled CO2 climate changes
may  be   considered   from  two  perspectives:
ecological  and  socioeconomic.    Evidence  for
significant national implications is strong from both

Ecological Implications

    Ecological implications for forests commonly
start with tree response.  But strong implications
also exist for other ecosystem  components, e.g.,
animals, soils, water, secondary impacts, and  as
noted,  the  atmosphere through  which  climate
change is mediated. Forest effects are described in
terms of tree distribution  changes and biomass
production  changes,  but  many  other processes
interact among the other major components.  Thus,
significant  changes in  tree response would be
accompanied  by   ecological   reverberations
throughout  all the forested areas  of major U.S.

Tree Distributions and Biomass Productivity

    As  discussed, migrations of forest tree species
to the North in response to rapid warming in North
America during  the  next century will be likely.

Chapter 5
However, significant lag is possible. Even under the
maximum  rates of species dispersal estimated  by
Zabinski and Davis, healthy forest areas may not
redevelop  for several centuries.  Furthermore, if
climate  continues to  change  beyond the  next
century, then healthy forests may never redevelop.
Meanwhile, distribution ranges  may not be under
such constraints, so the extent of healthy forested
regions in the United States probably would  be
greatly reduced.  Though some locations may have
increased productive potential from a biomass per
hectare standpoint,  the large reductions in areas
with healthy forests  would likely create a net
reduction in forest productivity for the United States
for several centuries or longer.

     Even if a massive reforestation effort  were
undertaken, the new forests resulting from species
shifts  might or  might  not be  as  productive  as
existing forests.  More northern latitudes or higher
elevations raise other considerations. Farther north,
days are longer in the summer  and shorter in the
winter. At higher elevations, damaging ultraviolet
light intensity is  greater.  All of these conditions
could lower forest productivity below present levels.
Furthermore, it is not clear that reforestation would
be successful. A major intent of reforestation would
be to artificially speed up northward migration of
tree species.  However, seedlings that would appear
to  be favored  on  some northern  sites  several
decades in the future may not  survive there now
because of constraints imposed by temperature, day
length, or soil conditions.  Similarly, seedlings that
could not survive  on those sites now might not  be
the best adapted species for those same sites several
decades in the future.


     A change in the size and relative homogeneity
of forests could influence whether some animals can
continue to live in their present locations.  Often,
animals are finely adapted to habitats specific to a
certain location.  For some animals, migration can
be hindered by  boundaries between forests and
other  land  types  or facilitated  as  animals  move
along  edges.   Furthermore, some  animals  (e.g.,
many game species) prefer young forests, and others
(e.g., many rare and endangered species) prefer old
forests.  In  turn, animals  can exert a profound
influence on forest structure   and composition
through selective browsing of seedlings, insect attack
of different tree species, seed dispersal, and other
effects. All of these factors illustrate that climate
change could influence the regional patterns  of
biotic  diversity in both plants  and  animals  (see
Chapter 8: Biodiversity).


    Soils under warmer climates also would change,
although at a much slower rate than shifts in species
distribution.  Increased soil temperatures, however,
would affect the entire range of physical, chemical,
and biological soil processes and interactions.  For
example, populations of bacteria, fungi, and animals
could  increase in  a way  that would accelerate
decomposition of litter and thereby reduce the
availability of nutrients essential for forest growth
(Spurr and Barnes, 1980).

    Considerable time may be required to  develop
optimum soil conditions for high forest productivity
supporting species  at more northern latitudes  or
higher elevations. Furthermore, it is not at all clear
how well some northern soils could support more
southern species.  The soils of the  boreal forest
differ from those under the deciduous forests to the


    Where forests give way to drier conditions (e.g.,
in the Great Lakes  region and California), many
lands now  serving as watersheds might be used for
different purposes.   Furthermore,  regional-scale
disturbances  (such  as  fire) and  applications  of
chemicals (such as fertilizers and pesticides) could
degrade regional water quality and increase airborne
toxic chemicals (see  Chapter 9: Water Resources).

    Sea level rise may impact some coastal forests.
Many forest  lands  of  high  value  for   timber
production (e.g., in the Southeast) or recreation (in
the Northeast, Northwest, and California) are close
to ocean coasts. Inundations, decreases in depth to
the water  tables,  and  saltwater intrusions could
trigger rapid  forest declines near these areas.

Secondary Impacts

    As the southern bounds of forests tend to shift
north, forest decline (sick and  dying trees) could
become  extreme  over  large  areas that  would
become highly susceptible to weed competition, pest

outbreaks, or wildfire. As forests decline, species of
lower economic value, as well as weedy shrubs and
herbs, could invade via wind dispersion.   Under
stressful  environments,  such species are  severe
competitors with most commercial tree species.

    Trees  experiencing  less   favorable  growth
conditions are more stressed and will be vulnerable
to insect and disease attack.  These secondary pest
impacts could last "until the most vulnerable forest
stands or tree  species  are eliminated"  (Hedden,
1987). In addition, it is estimated that the incidence
of catastrophic wildfires will increase in U.S. forests
with higher temperatures. Simand and Main (1987)
estimated that fire occurrence and fire-suppression
costs would increase 8 and 20%, respectively.

Socioeconomic Implications

    The United States enjoys substantial economic
and cultural benefits from its forests.  Until recently,
the nation's forest managers assumed that these
benefits could be sustained by maintaining forests in
a  healthy condition (Fosberg,  1988).   This  was
achieved,  for example, by preventing fires  or pest
invasions,  avoiding careless use, and enhancing
productivity through good silviculture.

    Beginning with the possibility of regional  air
pollution damage to forests, suspected in the 1980s,
alterations of the environment external to forests
presented a new  concern.   Research and policy
discussions to deal with this issue are ongoing.

    If climate changes as rapidly as predicted, this
additional external influence with its more global
dimensions looms as  a possible hazard to forests
and their  use.   As  can be  imagined,  a  list  of
potential  socioeconomic  concerns would be large.
To provide a brief perspective, three issues are

Quality of the Human Environment

    The forest  amenities  enjoyed  by most U.S.
citizens will be affected according to different forest
responses. In the Boston-Washington corridor, a
composition change from predominantly hardwood
to predominantly pine forests, though ecologically
significant, may not be noticed by most people if it
occurs gradually.  However, a  delay of years  or
decades between the decline of existing forests and
replacement by migrating tree species would likely
 elicit a strong concern.  In the Atlanta-Southeast
 region, the southern pine forests, while undergoing
 a gradual expansion of their northern boundaries,
 would have less vigor in the remaining stands.  This
 could  raise  their vulnerability to  damage from
 insects  and disease, reducing  esthetic  values —
 atleast an intermediate impact for most of the local
 citizens.   In contrast, within  some portions of the
 Southeast, the Great Lake region,  and California,
 drier climates may cause the loss  of some forest
 lands  to prairie  or  desert conditions — a severe
 change for the people there, not only in their living
 environment but also in the whole spectrum  of
 forest  land use.


    Forests must be in a relatively healthy condition
 to support quality recreational use (Clawson, 1975).
 Forests undergoing  gradual  composition changes
 might remain healthy, but rapid changes would most
 likely  cause  stressed or declining  forests.   Such
 forest conditions would have less recreational appeal
 because of such factors as less pleasing appearance,
 greater threat of wildfire,  and  reduced hunting
 quality when  game  populations change  or  are
 diminished.  Furthermore, drier conditions in U.S.
 forests would harm recreational opportunities that
 depend on abundant water or snow.

 Wood  Products

    Altered U.S. forest productivity resulting from
 climate change would have obvious major economic
 impacts.  Significant yield reductions could lead to
 unemployment, community  instability, industrial
 dislocation,  and  increased net  imports  of wood

    Reforestation projects could make up for some
 losses in forest productivity and artificially advance
 migrations forced by climate change. Reforestation
 technology has greatly improved in  recent decades
 so that success rates also have  increased greatly.
 Examples  are  high-vigor  seedlings  developed
 through   improved  nursery  practices,   genetic
 selection,   and  vegetative  propagation.
Improvements in the field include machine planting,
fertilization,  and weed  control  on  selected sites.
Results are  evident from  the large acreages of
plantations established  in  the  United States in
recent  decades, particularly with loblolly pine in the
Southeast and Douglas-fir in the Pacific Northwest

Chapter 5
(Table 5-2). Large-scale reforestation in the United
States and elsewhere could significantly add to the
total carbon sink provided by world forests, thereby
offsetting some of the buildup of atmospheric CO2-
Although this was not studied, attempts to reforest
some very dry sites may be unsuccessful.

    Innovative manufacturing trends should prove to
be  timely  during times  of rapid forest change.
High-strength    and   durable   products   from
reconstituted  wood  (e.g., new  particle  board
concepts,  warp-proof  hardwood  lumber, paper
products of fiber from multispecies) are now in use
or well along in development.  These new methods
will lessen  the present overdependency  on a  few
commercial conifer species from stands above
minimum size and quantity (Ince, 1987). The result
will be an ability to use the timber resources of the
future, however they change in composition.

    Historically, U.S. forest policies have undergone
continued development  to  meet national  change
(Young, 1982).  The earliest policies were adapted
by  the  New England colonies in  the  1600s  to
regulate overcutting near settlements.  Wood was
needed for fuel and buildings, but existing methods
were not capable of long distance log transportation.
Development of U.S. forest policies has continued
and has been particularly intense this century, as the
national forests, national parks, and wilderness areas
have been established.

    At present, forest managers are dealing with
many additional  policy  issues.   Five  of these
(Clawson,   1975)   are   important  to   climate
change/forest response:

    •   How much U.S. land should be devoted to

    •   How much forest land should be withdrawn
        from timber production and harvest?

    •   How should the federal forest  lands be
        managed?  (That is, the lands under the
        USDA Forest Service, USDI Park Service,
        Bureaus of Land Management and Indian
        Affairs,  and other federal  agencies that
        manage forest lands.)
    •   What  constraints (e.g., mandatory forest
        practices)  should be  placed  on forest
        managers to ensure national environmental

    •   Who  should  pay  the  additional  costs
        incurred in implementing new policies?

    The large array of forest  ownerships in  the
United   States,   public  and   private,   makes
development and implementation of  forest policy
more complicated than in most countries.  Around
the world, about 77% of all forests  are in some
form of public ownership (Hummel,  1984).  The
diversity of owners and managers results in widely
divergent goals and objectives.

How Much Land Should Be Forested?

    Changes  in  forest  composition  or regional
boundaries induced by rapid climate change would
magnify the complexity of national forest policy even
further. Lands in forests now would require review
relative to such competing needs as agriculture  and
residential use, which  would also be  adjusting to
climate change.

How Much  Should  Be Withdrawn From
Timber Production?

    Where the productivity of wood is significantly
reduced, increased, or shifted, a policy question that
would  surely arise concerns whether  forest lands
should  be   reallocated  to  maintain  timber
production.  If so, how should competing forest
uses, such as watersheds, parks, and wilderness, be
treated? How much of each can the United States
afford under changed climatic conditions?  Should
the federal government purchase more forest lands
to support all public needs?

    In  the  short  term, forest  managers  could
compensate  for   some  loss of  productivity by
improved technology, although  at  increased costs.
An example would  be  establishment of  more
drought-tolerant   plantations   through  genetic
selections,  improved  nursery  stock,  and  more
intensive silvicultural practices (e.g., weed control
and thinning).  Introducing new species adapted to
warmer  climates  might be  possible  in  some
locations, but this  would call for  development of
new silvicultural regimes and utilization methods —
possible, but time consuming and costly. In the long

term,  if  growing  conditions become  extremely
difficult  on some  U.S.  forest lands because of
climate  changes,   establishing  trees  for  wood
production on such sites may not be economically

How Should We Manage Federal Forests?

    The national forests under the USDA Forest
Service are managed according  to  a series of
complex  legal   directives  and   administrative
procedures, beginning with the Organic Act of 1897
(Woodman and Furiness, Volume D).  Ultimately,
the objective became to manage the national forests
for multiple  uses,  with  timber  and other  forest
resources on  a  sustained-yield basis and certain
lands set aside as wilderness areas. The National
Forest Management Act  of the mid-1970s  requires
management plans for each national forest subject
to public review.  The plans look ahead 50 years
and are to be updated every 10 years.

    Lands managed by the Department of Interior
are  under similar mandates.   For  example,  a
congressional act passed in 1976 charged the Bureau
of Land Management to manage its 2.3 million
hectares (5.1 million acres) of forest and range land
according  to  multiple-use  and  sustained-yield
principles.  Similarly, the National Park Service is
mandated to  manage national parks, monuments,
historic sites,  and  so forth, for  the  recreational
enjoyment  of  people.   Such activities as .timber
harvesting, hunting, mining,  and grazing  are not
permitted.  In addition to the federal government,
most states, many counties, and some municipalities
own forest lands.               ,

    The  Forest   and    Rangeland  Renewable
Resources  Planning  Act ,of 1974  requires  the
Secretary of Agriculture to make periodic reviews of
the nation's forest and rangeland resources. In the
future,  these  assessments  and planning  efforts
should include consideration of the possible effects
of predicted climate changes.

    A key issue is the  level of priority given to
maintaining forest  health under changed climate
conditions.  For  instance,  under more  adverse
environments,  should national forests be left to
decline as a natural process, thereby losing esthetic
values in parks, water yields from watersheds, and
highly productive  timber  crops?    Or  should
silvicultural forest techniques such as thinning, weed
control, fertilization, and reforestation be employed
in an attempt to preserve them?  This question and
others will challenge the fundamental concepts of
the benefits of multiple use and sustained yield of
U.S. forests.

How Can We Ensure National Goals?

    At the minimum,  federal agencies  must plan
and act in concert with the state and private forest
organizations.   In the first half of this century, the
federal government attempted to regulate  forest
harvests  on all federal,  state, and private  lands.
Development of this policy did not survive strong
public concern and  intense political debate against
such policy (Worrell,  1970); the same sentiment
would likely exist  today.   However, under  the
influence of climate change, the nation may once
again have to  face  the  touchy  issue of what
restraints or forest practices must be regulated for
all public and private lands.

    Solomon and West (1985) point out that while
climate change might disrupt forest ecosystems in
the, future,  it is uncertain whether forest managers
could or would be  able  to apply silvicultural
practices on a scale large enough to maintain the
net productivity of commercial forest lands in the
United  States.  Some states (e.g., Washington,
Oregon,  and California) have laws specifying  fire
protection requirements, control burn practices, and
reforestation minimums following timber harvests.
Zoning,  permits, licenses,  and various  taxation
measures also  have been  attempted with mixed
results.  It  is much  easier to prevent owners from
destroying  forests  than  to  compel  them  to
implement  silvicultural practices.


    To keep pace with the global climate changes
estimated, the  U.S. reforestation effort conceivably
would need to be doubled or tripled in size.   In
recent years,  about 800,000 hectares (2 million
acres)  per year  (approximately  700+  million
seedlings) have been reforested in the United States
(USDA,  1982). Costs range from $200 to $700 per
hectare  ($80 to $280  per  acre) depending upon
species,  site preparation,  plantation density, and
planting method. Using $500 per hectare ($200 per
acre)  as  a  mode, the  total annual  expenditure is
near $400 million.  About 0.4% of the commercial
land base is reforested annually.   At this rate,  it

Chapter 5
would take 100 years to reforest 40% of the U.S.
forest lands, assuming no repeat hectares to cover
failures or harvests of the first plantations.

    An expansion on the  scale suggested above
would  require  large   investments   in   seed
procurement,  tissue  culture  capability,  nursery
capacity, and research to improve knowledge about
the  establishment  and  silviculture  of  drought-
resistant  plantations.    Even  if   the  dollar
commitments were made, reforestation at this scale
might be  possible  only if all forest lands were
managed by one organization.  The complex forest
ownership pattern in the United States, therefore,
would be an issue to overcome in  a national
reforestation program.

Who Should Pay?

     Adjusting forest policies to address the issues
arising from climate change will most  likely raise
the costs of using the nation's forests — whether for
water, recreation, esthetics, or timber.  Additional
research to answer many new questions will also
require more funds. A major question will be who
should pay for these costs.  Land owners?  Forest
users?  Consumers? All taxpayers?  The answers
will come when better information is available on
resulting forest effects, followed by public  debate
establishing new priorities  for  forest  use in  a
changed climate.

    The forest effects resulting from rapid climate
change are at present hypothetical.  The change has
not  yet occurred,  and  many  uncertainties  are
associated with the predictions. Effective policies to
deal with  new  forest effects will require  more
information and fewer uncertainties that must come
through forest ecosystem research.   Four broad
questions   concerning U.S.   forests frame  the
research needs for the 1990s:  What will the effects
be?  How can they be measured  reliably?  How
should they be managed? How can we ensure that
research will be conducted in a timely fashion?

Effects of Climate Change

    What will be the  effect on the nation's forest
ecosystems if climate changes occur as predicted by
the middle of the 21st century?  While subsets of
this question must include extent, magnitude, and
risk considerations, additional knowledge is needed
concerning the following:

    1.   Forest  migration  processes  and rates,
        including the landscape processes that
        control the horizontal movements of forests,
        animals, and disturbances;

    2.   Interactions among the different landscape
        components  and land-use practices  that
        affect biodiversity, and water quantity and

    3.   The impact of climate  change alone and in
        combination   with   other  natural  or
        anthropogenic influences, such as insects,
        pathogens,   COj   enhancement,   air
        pollutants,  UV-B   radiation,  and  acid
        deposition on U.S. forests;  and

    4.   The processes and mechanisms that  play
        key roles in forest ecosystem effects — both
        biologically  as  in  photosynthesis  and
        respiration, and physically as in flows of
        energy,  carbon,  water,  and  nutrients
        through ecosystems.


    How  can forest ecosystems be measured to
reliably detect the effects of rapid climate change?
Today, the response of ecosystems to environmental
change  is largely based upon extrapolating from
field observations, from knowledge  about seedlings
or  individual  trees   of  a   small  number  of
commercially valuable species,  and  from computer
models. The following must be accomplished:

    1.   A determination  of  the most  useful
        integrating variables for forest ecosystems
        that indicate the effects of  climate change
        —  particularly variables that  are early-
        warning indicators of ecosystem response;

    2.   Effective sampling designs developed for
        experiments  and long-term monitoring at
        the forest ecosystem scale;  and

    3.   Improved  models  capable of projecting
        regional effects on forests  across multiple
        spatial and temporal scales.

 Forest Management

    What options are available to the public and
 private forest managers and owners in the United
 States to address the changes in the nation's forests
 that might occur in the next century?  Research is
 needed to accomplish the following:

    1.  Understand the socioeconomic impacts of
        all  forest  ecosystem effects  to clarify
        economic risks and alternatives; and

    2.  Develop technology to mitigate the adverse
        effects or to exploit the benefits of forest
        change, such as breeding, bioengineering,
        transplanting, fertilization, irrigation,  and
        other management approaches.

 Timing of Research

    The  timing of the  research is  critical.  The
 effects of climate change may be some  decades
 away,  but this  should not lessen the urgency to
 begin  research toward better  information  and
 methods.  The complexities of the science  are very
 large.  Developing a base of knowledge to identify
 potential forest changes before they are upon the
 nation will require significant time and resources.

Harbour, M.G., J.H. Burk, and W.D. Pitts.  1987.
Terrestrial Plant Ecology, 2nd Ed.  Menlo Park,
CA: Benjamin/Cummings Publishers.

Botkin,D.B.  1979.  A grandfather clock down the
staircase:  stability  and  disturbance  in  natural
ecosystems.  In: Waring, R.H., ed. Proceedings of
the 40th Annual Biological Colloquium. Forests:
Fresh  Perspectives  From  Ecosystem Analysis.
Corvallis, OR: Oregon State University Press, pp. 1-

Clawson, M.   1975.  Forests for Whom and for
What? Resources for the Future. Baltimore, MD:
Johns Hopkins University Press.

Cubbage,  F.W., D.G. Hodges, and J.L. Regens.
1987.   Economic implications  of climate change
impacts on forestry in the South. In: Meo, M., ed.
Proceedings of the Symposium on Climate Change
in the Southern U.S.: Future Impacts and Present
 Policy   Issues.      New  Orleans,
 Environmental Protection Agency.
LA:   U.S.
 Davis, M.D., and D.B. Botkin. 1985.  Sensitivity of
 the cool-temperate forests and their pollen to rapid
 climatic change.  Quaternary Research 23:327-340.

 Fosberg, MA.  1988. Forest productivity and health
 in a  changing  atmospheric environment.    In:
 Berger, A., et al., eds. Climate and Geosciences: A
 Challenge  for  Science  and  Society in the 21st
 Century.     NATO  ASI   Series.    Series   C;
 Mathematical  and  Physical  Sciences,  Vol. 285.
 Dordrecht, The Netherlands:   Kluwer Academic
 Publishers, pp. 681-688.

 Grey,  G.W., and FJ.  Deneke.   1978.   Urban
 Forestry. New York: John Wiley and Sons.

 Hammond, A.L.  1972. Ecosystem analysis: biome
 approach to environmental science. Science 175:46-

 Hedden, R.  1987.  Impact of climate change  on
 forest insect pests in the southern U.S.  In:  Meo,
 M., ed. Proceedings of the Symposium on Climate
 Change in the Southern U.S.: Future Impacts and
 Present Policy  Issues. New Orleans, LA:  U.S.
 Environmental  Protection Agency.

 Hummel, F.C.,  ed.   1984.   Forest  Policy, a
 Contribution  to  Resource  Development.   The
 Hague: Martinus Nijhoff/Dr. W. Junk, Publishers.

 Ince, P J. 1987.  Technology, timber  demand and
 timberland investment.   In:   A Clear  Look  at
 Timberland Investment, Milwaukee, WI; April 27-
 29.  Conference proceedings.  Forest  Products
 Research Society.

 Lavdas, L.G.  1987.  The impact of climate change
 on forest  productivity.    In:   Meo,  M.,  ed.
 Proceedings of the Symposium on Climate Change
 in the Southern U.S.:  Future Impacts and Present
 Policy   Issues.     New  Orleans,  LA:    U.S.
 Environmental Protection Agency.

 Leverenz, J.W.,  and DJ. Lev.  1987.  Effects  of
 carbon dioxide-induced  climate changes on the
 natural ranges of six major commercial tree species
 in the western United States. In: Shands, W.E., and
J.S. Hoffman, eds. The Greenhouse Effect, Climate
 Change,  and U.S.  Forests.   Washington, DC:
 Conservation Foundation, pp. 123-156.

Chapter 5
Manion, P.D. 1981. Tree Disease Concepts.  New
Jersey: Prentice Hall.

Meo, M., ed.  1987. Proceedings of the Symposium
on Climate Change in the Southern United States:
Future Impacts and Present Policy Issues; May 28-
29.  New Orleans, LA: University of Oklahoma and
U.S. Environmental Protection Agency.

Miller, F.W., P.M. Dougherty, and G.L. Switzer.
1987. Effect of rising carbon dioxide and potential
climate change on loblolly pine distribution, growth,
survival and productivity.  In:  Shands, W.E., and
J.S. Hoffman, eds. The Greenhouse Effect, Climate
Change,  and  U.S.  Forests.   Washington,  DC:
Conservation Foundation,  pp. 157-188.

Pickett, S.TA., and P.S. White.  1985. The Ecology
of  Natural  Disturbance  and  Patch Dynamics.
Academic Press, Inc.  Harcourt Brace Jovanovich.

Schallau, C.H.  1988. The forest products industry
and community stability: the evolution of the issue.
Montana Business Quarterly Summer: 1-8.

Shands, W.E., and J.S. Hoffman, eds.  1987.  The
Greenhouse  Effect,  Climate Change,  and  U.S.
Forests.     Washington  DC:     Conservation

Simand.AJ. and WA. Main. 1987. Global climate
change: the potential for changes in wildland fire
activity hi the Southeast.   In:    Meo, M., ed.
Proceedings of the Symposium on Climate Change
in the Southern U.S.:  Future Impacts and Present
Policy   Issues.     New   Orleans,  LA:   U.S.
Environmental Protection Agency.

Solomon, A.M.,  and D.C.  West.  1985.  Potential
responses of forests to CO2 induced climate change.
In:    White,  M.R.,  ed.    Characterization  of
Information Requirements for Studies  of  CO2
Effects: Water Resources, Agriculture, Fisheries,
Forests and Human Health. Washington, DC: U.S.
Department of Energy.  DOE/ER-0236.  pp. 145-

Solomon, A.M., and D.C. West.  1986. Atmospheric
carbon dioxide change:   agent  of future forest
growth or decline? In:  Titus J.G., ed.  Effects of
Changes   in  Stratospheric  Ozone and   Global
Climate. Vol. 3: Climate Change. Washington, DC:
U.S. Environmental Protection Agency.
Spurr,  S.H.,  and B.V.  Barnes.   1980.   Forest
Ecology, 3rd Ed. New York:  John Wiley and Sons.

Strain, B.R.,  and J.D. Cure, eds.  1985.   Direct
Effect of Increasing Carbon Dioxide on Vegetation.
Washington, DC:  U.S.  Department  of Energy.

Tirpak, D.A., ed. 1987.  Potential Effects of Future
Climate  Changes  on   Forest  and  Vegetation,
Agriculture, Water Resources and Human Health,
Vol. V.  Assessing the Risks of Trace Gases That
Can Modify the Stratosphere. Washington, DC:
U.S. Environmental Protection Agency. EPA 400/1
- 87/001E.

Titus, J.G., ed.  1986.   Climate  Change, Vol. 3.
Effects  of  Changes in Stratospheric Ozone and
Global  Climate.    Washington,  DC:    U.N.
Environmental Program and U.S. Environmental
Protection Agency.

USDA.   1981.   U.S. Department of Agriculture,
Forest Service.  An Assessment of the Forest and
Range Land Situation in the U.S.  Forest Resource
Report No. 22.

USDA.   1982.   U.S. Department of Agriculture,
Forest Service. An Analysis of the Timber Situation
in the U.S. 1952-2030. Forest Resource Report No.

Waring, R.H., and W.H. Schlesinger.  1985.  Forest
Ecosystems, Concepts and Management. Orlando,
FL: Academic Press, Inc.

Webb, T. 1987.  The appearance and disappearance
of  major  vegetational  assemblages:   long-term
vegetational dynamics in eastern North America.
Vegetation 69:177-187.

White,  M.R., ed.   1985.    Characterization  of
Information Requirements for Studies of CO2
Effects:  Water Resources, Agriculture, Fisheries,
Forests and Human Health. Washington, DC: U.S.
Department of Energy.  DOE/ER-0236.

Worrell, A.C.  1970.  Principles of Forest  Policy.
New York: McGraw-Hill.

Young,  RA., ed.  1982.   Introduction  to  Forest
Science. New York: John Wiley  and Sons.

                                       CHAPTER 6

 Climate change would affect crop yields and result
 in  northward  shifts  in cultivated land,  causing
 significant regional dislocations in agriculture with
 associated impacts on regional economies. It would
 expand crop irrigation requirements, stress livestock
 production, and increase infestations of agricultural
 pests and diseases.  Preliminary results suggest that
 although  U.S.  crop production  could  decline,
 supplies would be adequate to meet domestic needs.
 The  potential  for reduction  of the  national
 agricultural  capacity and the many uncertainties
 surrounding   the  interactive   effects   on   the
 agricultural system  create the necessity to respond
 to the climate change issue.

 Crop Yields

 •    The effects of  climate  change alone may
     reduce average yields of corn, soybeans, and
     wheat, both rainfed and irrigated, except in the
     northernmost   latitudes   where   warmer
     conditions provide a longer frost-free growing
     season.  Decreases in modeled yields result
     primarily  from higher temperatures, which
     shorten a crop's life cycle.

 •    When  the direct  effects  of CO2 on crop
     photosynthesis  and   transpiration   are
     approximated along with the effects of climate
     change, average rainfed and irrigated corn,
     soybean, and wheat yields could overcome the
     negative effects of climate change in some
     locations.  If climate changes are severe, yields
     could still decline.  The extent to which the
     beneficial  direct effects of CO2 will be seen
     under field conditions with changed climate is

•    Even if the patterns of climate variability are
     unchanged,  yield  stability  may decrease,
     particularly under rainfed conditions.  This
     may occur because there would be more days
 above temperature thresholds for particular crops in
 some locations. The exact magnitude of change will
 be sensitive to  changes in  climatic variability,
 particularly the frequency of droughts.

 Economic Impacts

 •    Under three out of four scenarios, a small to
     moderate aggregate reduction in the nation's
     agricultural  output  was  estimated.    The
     estimated production levels  appeared to be
     adequate  to  meet  domestic  consumption
     needs.  If droughts occur more frequently
     under changing climate, effects on agriculture
     may be more severe.

 •    Assuming no change  in  export  demand,
     reduced outputs would decrease exports, which
     could negatively affect global food supplies and
     the U.S. trade balance.  This report  did not
     analyze global changes in agriculture, which
     could have a major effect on demand for  U.S.

 •    Under the  most  severe  climate  change
     scenarios,   continued   technological
     improvements, similar to those in recent years,
     would have  to be sustained to offset losses.
     Increasing food demand from higher U.S. and
     world  population  would  aggravate   the
     economic losses due to climate change.

•    The  economic  response of  agriculture to
     changes in regional productivity may be to
     shift  crop   production  and  associated
     infrastructure in a northward direction.  This
     is because yields in northern areas generally
     increase relative to yields in southern areas.
     Although availability of agricultural soils  was
     included in the economic analysis, neither the
     sustainability of crop production hi northern
     areas nor the introduction of new crops  into
     southern areas was studied.

Chapter 6
Irrigation Demand

•    The demand for irrigated acreage is likely to
     increase in all regions.  This is due to the
     reliability of irrigated yields relative to dryland
     yields and to higher commodity prices that
     make expansion of irrigated production more
     economically feasible.   Actual  increases in
     irrigated  acreage  would   depend  on  the
     adequacy of water supply and on whether the
     cost of water to farmers increases.

•    Demand for more irrigation would increase
     stress on and competition for regional water
     supplies.  If irrigation does  increase, it could
     increase surface and  groundwater pollution
     and other forms of environmental degradation.

Agricultural Pests

•    Climate warming could change the ranges and
     populations of agricultural pests. Temperature
     increases may enhance the  survival of insect
     pests  in the winter, extend  their northward
     ranges, increase pest species with more than
     one  generation  per year,   and allow pest
     establishment earlier in the growing season.
     These effects could result in a substantial rise
     hi  pesticide   use,   with  accompanying
     environmental hazards. Changes in pests will
     also   depend  on  regional  shifts  in  crop

Farm-Level Adjustments

•    Farmers may adjust to climate change by using
     full-season and heat-resistant crop species or
     varieties, by altering planting dates, by planting
     two crops  during  one  growing season,  by
     increasing or  altering  their scheduling of
     irrigation, by using more pesticides, and by
     harvesting earlier.  If climate change is not
     severe, these adjustments may mitigate losses
     in crop yields; more severe climate change is
     likely to make major adaptation necessary.

Livestock Effects

•    Higher temperatures may increase disease and
     heat   stress  on  livestock in some regions.
     Existing livestock diseases  may shift north,
     while tropical diseases may extend their ranges
                                                     into southern regions of the United States.  Cold
                                                     stress conditions may be reduced in the winter, but
                                                     heat stress is likely to increase in the summer.
                                                     Reproductive capabilities may also decrease.

                                                     Policy Implications

                                                     •    Global   climate   change  has   important
                                                          implications for all parts of the  agricultural
                                                          system.  The agricultural research structure,
                                                          which is dedicated to maintaining U.S. farm
                                                          productivity, should  expand climate  change
                                                          research in activities ranging from the field
                                                          level to  the national policy level.

                                                     •    Current  U.S.  Department of  Agriculture
                                                          (USDA) research on heat-  and drought-
                                                          tolerant crops and practices and maintenance
                                                          of crop  germ plasm should be sustained and
                                                          enhanced  to  limit,  vulnerability to future
                                                          climate  change.

                                                     •    The USDA should evaluate current legislation
                                                          in regard to its ability to allow adaptation to
                                                          global warming.  Flexibility in shifting crop
                                                          types and farm practices will speed adjustment.
                                                          Such adaptation strategies should  consider the
                                                          impacts on soil erosion and water quality.

                                                     •    The USDA, the Department of  Commerce,
                                                          the U.S. Trade Representative, and the State
                                                          Department should consider the  implications
                                                          of potential long-term changes in the  level of
                                                          U.S. crop exports for the U.S. balance of trade
                                                          and strategic interests.

                                                     •    A national drought policy is strongly needed to
                                                          coordinate  federal response to the possibility
                                                          of increased droughts due  to climate  change.
                                                          Even without climate change, such a policy is
                                                          necessary not only for the agricultural sector
                                                          but also for other sectors.
                                                     SENSITIVITY OF AGRICULTURE
                                                     TO CHANGES IN CLIMATE

                                                          Agriculture is a critical American industry,
                                                     providing food for the nation's  population and as
                                                     much as $42.6 billion in exports for the nation's
                                                     trade balance (Figure 6-1). Agriculture employs 21
                                                     million  people ~ more  than any  other industry,

                         '  ' Fruits, nuts, and vegetables
                         ESJ Cotton
                            Livestock and by-products
                            Oilseeds and by-products
                            Grains and preparations
                          72  73  74   75  76  77  78  79  80  81   82  83  84  85  86
Figure 6-1.   Value of U.S. agricultural exports by commodity, 1972-86 (not adjusted for inflation). Livestock
excludes poultry and dairy products (The World Food Institute, 1987; U.S. Department of Agriculture, Economic
Research Service, Foreign Agricultural Trade of the United States, Washington, DC, January-February 1987,
and various other issues).
when taking into account workers on farms and in
meat, poultry,  dairy, baking,  and food-processing
activities  (Council for Agricultural  Science and
Technology,  1988).     The   U.S.   agricultural
production  system  includes  farm   equipment
manufacture,  fertilizer and seed  supplies, rural
banking, and shipping. Total farm assets were $771
billion in  1985; food and fiber were 17.5% of the
total  gross  national product in  the  same year.
Wheat, corn, soybeans, cotton, fruits and vegetables,
and livestock are among  the most important U.S.
agricultural  commodities.

     Worldwide, agricultural products must provide
sustenance for the world's growing population, now
estimated at about 5 billion and projected to rise to
8.2 billion  by 2025  (Zachariah and Vu,  1988).
Global production and consumption of grain have
grown steadily since  1960, although regional food
shortages  continue  to occur  owing to  climate
variability and socioeconomic factors. Technological
advances,  such as improved hybrids and irrigation
systems, have reduced the dependence of crop yields
on local environmental conditions, but weather is
still an important factor in agricultural productivity.
     For example, failure of the monsoon season
caused shortfalls in  crop  production  in India,
Bangladesh, and Pakistan in 1987. The 1980s have
also seen the continued  deterioration of food
production in Africa, despite adequate world food
supplied elsewhere, because of persistent drought,
internal wars, poor distribution, weak infrastructure,
and a deteriorating environment.  Climate extremes
have had large effects on U.S. agriculture.  During
the Dust Bowl years of the 1930s, U.S. wheat and
corn yields dropped by  up to 50%.   Midsummer
1983 saw an unpredicted drought in the U.S. Corn
Belt and in the southeastern United States, causing
U.S. corn yields to fall by about a third,  from over
7,000  kilograms   per  hectare   to   about 5,000
kilograms per hectare (from about 110 to 80 bushels
per acre).

     The 1988 drought recently demonstrated the
impact  that  climate  variability  can  have  on
agricultural productivity.  This drought  decreased
U.S. corn yields by almost 40%, and the cost of the
1988 Drought Relief Bill is estimated to  be $3.9
billion  (Schneider,  1988).    The  1988 drought
emphasizes anew the close link between agriculture
and climate.

Chapter 6
     Light from the sun, frost-free growing seasons,
and  the  hydrologic  cycle  largely  govern  the
suitability of geographic areas for crop production
and affect crop productivity.  Livestock production
is responsive to climate through differing levels of
heat and cold stress and altered ranges of disease-
carrying vectors such as mosquitoes and ticks.

     Higher levels  of  COp in the air would also
affect crops.  Increased CO2 has enhanced crop
photosynthesis and has improved crops' use of water
in experimental settings.   Because experimental
research has rarely simultaneously investigated both
the climatic  and the direct effects of CO2 on plants,
it is difficult to assess the relative contributions of
CO2 and increased temperature to plant responses.
This remains one of the most crucial questions in
the analysis of impacts of  climate  change and
increased CO2 on agriculture.

     The presence and abundance of pests affecting
both crops and livestock are highly dependent on
climate.  The severity  of the winter season, wind
patterns, and moisture conditions determine in large
part   where pests  will   be  prevalent.    The
geographical distribution of pests also depends on
locations of  crop types.

     Much  of U.S. agricultural production takes
place under technologically  advanced  cropping
systems that are primarily monocultural. Likewise,
livestock production is  highly specialized,  both
technically and geographically, and a high degree of
integration  exists  between  grain  and  livestock
production.   Any  significant level  of economic
robustness  associated  with  general,  multiple-
enterprise farms has long  since passed from the
scene.   The ability of our agricultural system  to
adapt to  climate change may be more limited now
in some ways than it was in the past.

     Agriculture  strongly  affects  the  natural
environment.  It  often  increases  soil  erosion,
intensifies demand for water,   degrades  water
quality, reduces forested land, and destroys wildlife
habitats.  Many agricultural practices contribute to
soil degradation, groundwater overdraft, loss  of
plant  and  aquatic  communities,  and generally
reduced  resilience in  environmental  and genetic
resources.    Therefore,  climate-driven shifts  in
agricultural   regions   have   implications   for
environmental quality.
     Thus,  climate  plays  a  major  role   in
determining  crop   and  livestock  productivity.
Agricultural productivity determines profitability and
decisionmaking at the farm  level, which in turn
define farming systems at  the regional level and
import-export  supply and demand at the national
and   international   levels.     These  complex
interrelationships necessitate a broad consideration
of the impacts of potential climate change on U.S.

     Relationships between climate and agriculture
have  been  studied  intensively  for  many years.
However, relatively  few studies have  specifically
addressed both the climatic and the  direct effects
that the  growth  in trace  gases will have  on
agriculture.   Even  fewer studies have  addressed
these potential effects in an integrated approach
that links both biophysical and economic spheres of

     Most research attention in the United States,
supported primarily by the U.S. Department  of
Energy, has focused on the direct effects of CO2 on
crops.  These studies are reviewed by Acock and
Allen  (1985)  and  Cure  (1985), who  found  an
average  increase in yields of  about  30% .and
increases in water-use efficiency for crops growing
in air with doubled CO2 (660 ppm) and favorable,
current climate  conditions.  Kimball (1985) and
Decker et al.  (1985) suggested that  the potential
effects  of  CO2  and/or  climate   change  on
agricultural production systems may include shifts in
production areas and changes in levels of livestock
stresses,   water  availability,  and  pest  control

     Integrated  approaches  to  the  impacts   of
climate  change  on  agriculture  involving both
biophysical and  economic processes have been
considered in studies by Callaway et al. (1982), the
Carbon Dioxide Assessment  Committee  (1983),
Warrick  et al. (1986),  and  the  Land  Evaluation
Group (1987). A benchmark international study on
both the agronomic and economic effects of climate
change on  agriculture  was  conducted  by the
International Institute for Applied Systems Analysis
(Parry  et  al.,  1988).   No  study  has   as  yet

comprehensively examined the combined effects of
climate change and the direct effects of CO2 on
U.S. agriculture.

Structure of and Rationale for the Studies

     The  regions  studied  for this  report  are
important agricultural production areas (see Table
6-1). The Great Lakes and Southeastern States are
major corn and soybean producers, and the Great
Plains  States  grow   mainly wheat   and  corn.
California  annually produces about 10%  of  U.S.
cash farm receipts from cotton, grapes, tomatoes,
lettuce, and many other crops.

     The agricultural studies involve the following
research topics (see Table 6-2): (1) crop growth and
yield,  (2)  regional   and  national  agricultural
economics, (3) demand for water for irrigation, (4)
water quality,  (5) pest-plant interactions, (6) direct
effects of  CO2  on  crop growth  and  yield,  (7)
impacts of extreme events, (8) potential farm-level
adjustments,   (9)  livestock  diseases,   and  (10)
agricultural policy.

     Production of corn,  wheat, and  soybeans is
critical to the  economic well-being of the  nation's
farmers and the  national trade balance.   These
crops make up about two-thirds of the  total  U.S.
agricultural acreage, and  their  economic value is
equal to that of all other crops combined.   These
                      three crops were selected for the modeling studies
                      on the effects of climate change on yields.

                           The results from the regional studies of crop
                      production (not including California), hydrological
                      predictions from  the  climate models,  and  an
                      agricultural economics model  were linked  in an
                      integrated  approach to  enable  investigators to
                      translate the estimated yield changes from the crop
                      modeling studies and predicted changes in water
                      availability into economic consequences (see Figure
                      6-2). Such a coordinated analytical framework is
                      necessary to account for the effects of market forces
                      on the total agricultural sector, including livestock,
                      and  to  evaluate  the  adequacy  of the  nation's
                      resource base for  agricultural production  under
                      climate change.  Economic forces may lead farmers
                      to  grow more crops in areas with relatively high
                      productivity and fewer crops in  areas with relatively
                      low productivity.

                           The studies of demand for irrigation  water,
                      water quality,  and farm-level adjustment were  also
                      linked with the integrated modeling  studies by
                      common assumptions, sites, or outputs.   Because
                      California grows a large and diverse number of crop
                      commodities,  a  simple  approach  was  used to
                      estimate crop  yield changes for the California case
                      study based on heat, sunlight,  and photosynthetic
                      response to CO2.  These yield changes were then
                      used in a model of agricultural land and water use
                      in  California.   Adjustment  experiments  were
                      included  in  several   studies   to  test   possible
                      adaptation mechanisms, such as changes in planting
                      dates and crop varieties.
                                Table 6-1.  Crop Production by Region
EPA study areas
(thousands of bushels)
 Great Lakes
 Great Plains

 Total (48 states)



Source: U.S. Department of Commerce (1983).

Chapter 6
       Table 6-2.  Agriculture Projects for EPA Report to Congress on the Effects of Climate Change

Regional Studies

•    Effects of Projected CO2-Induced Climate Changes on Irrigation Water Requirements in the Great Plains
     States - Allen and Gichuki, Utah State University (Volume C)

•    Climate Change  Impacts  upon Agriculture and Resources:   A Case  Study of California  - Dudek,
     Environmental Defense Fund (Volume C)

•    Farm-Level Adjustments by Illinois Corn Producers to Climate Change - Easterling, Illinois State  Water
     Survey (Volume C)

•    Impacts of Climate Change on the Fate of Agricultural Chemicals Across the USA Great Plains and Central
     Prairie - Johnson, Cooter, and Sladewski, Oklahoma Climatological Survey (Volume C)

•    Impact of Climate Change on Crop Yield in the Southeastern U.S A.: A Simulation Study - Peart,  Jones,
     Curry, Boote, and Allen, University of Florida (Volume C)

•    Effects of Global Climate Change on Agriculture: Great Lakes Region - Ritchie, Baer, and Chou, Michigan
     State University (Volume C)

•    Potential Effects of Climate Change on Agricultural Production in the Great Plains:  A Simulation Study -
     Rosenzweig, Columbia University/NASA Goddard Institute for Space Studies (Volume C)

National Studies

•    The Economic Effects of Climate Change on U.S. Agriculture: A Preliminary Assessment - Adams,  Glyer,
     and McCarl, Oregon State University and Texas A&M University (Volume C)

•    Analysis of Climate Variability in General Circulation Models - Mearns, Schneider, Thompson, and
     McDaniel, National Center for Atmospheric Research (Volume  I)

•    Direct Effects of Increasing CO,-, on Plants and Their Interactions with Indirect (Climatic) Effects - Rose,
     Consultant  (Volume C)

•    Potential Effects of Climatic Change on Plant-Pest Interactions - Stinner, Rodenhouse, Taylor, Hammond,
     Purrington, McCartney, and Barrett, Ohio Agricultural Research and Development Center and Miami
     University (Volume C)

•    Agricultural Policies for Climate Changes Induced by Greenhouse Gases - Schuh, University of Minnesota
     (Volume C)

•    Changing Animal  Disease Patterns Induced by the Greenhouse Effect - Stem, Mertz, Stryker, and Huppi,
     Tufts University (Volume C)

•    Effect of Climatic Warming on Populations of the Horn Fly - Schmidtmann and Miller, USDA, Agricultural
     Research Service (Volume C)

f                Economic Consequences
                 Land Use and Irrigated
                   Acreage Changes  /
Figure 6-2.  Flow chart of model interactions in
EPA studies of the effects of global climate change
on U.S. agriculture (Dudek, 1987).
     The agricultural studies performed for  this
EPA report explore the sensitivities of the different
parts of the agricultural system (shown in Table 6-
2) to climate change scenarios. They are not meant
to be predictions of what will happen; rather, they
aim  to  define  ranges and  magnitudes of  the
potential  responses  as the  system is  currently
understood. Regional results were extrapolated to
other areas to give estimates of changes in national

     All of the modeling studies used the doubled
CO2 climate change scenarios developed for the
report  (see  Chapter  4:  Methodology).   These
scenarios were developed from estimated changes in
monthly mean  climate variables  from  general
circulation models (GCMs), without alterations in
climate variability.  For example, the number of
days of precipitation  remains the  same  in the
baseline and climate  change scenarios, and the
amount of precipitation on each of those days is
adjusted by the GCM ratio for climate change.
Extreme events, such  as maximum temperature,
vary in the climate change scenarios according to
the ratios, but the daily and interannual patterns of
warm  episodes are determined by  the  observed
baseline climate.

     The lack  of  changes  in  the  daily   and
interannual patterns of extreme events may result in
underestimation of impacts of climate change.  This
is because runs of extreme climate variables  (for
example, prolonged heat spells during grain filling
and drought) can decrease crop  productivity.   For
rainfed crops,  yields  may change  considerably,
depending on whether  a change  in precipitation is
caused by more or fewer events or by higher or
lower  precipitation per event.   The frequency,
intensity, and/or duration of extreme climatic events
can be much more consequential to crop yields than
are simple changes in means.

Timing of Effects

     The timing of climate  change is uncertain —
rates of future emissions of trace gases, as well as
when the full magnitude of their effects will be
realized, are unknown.  CO2 concentrations are
estimated to  be about 450  ppm in 2030 and 555
ppm in 2060 if current emission trends continue
(Hansen et  al., 1988).   Other greenhouse gases
besides CO2 (e.g., methane (CH4),  nitrous oxide
(N2O), and chlorofluorocarbons (CFCs)) are also
increasing. The effective  doubling of CO2 means
that  the  combined   radiative   forcing  of  all
greenhouse gases has the same radiative forcing as
doubled CO2 (usually  defined as 600 ppm).  The
effective doubling of CO, concentrations will occur
around the year 2030,  u  current emission trends
continue. The climate change caused by an effective
doubling of CO2 may be delayed by 30 to 40 years
or longer.

 Chapter 6

 Regional Crop Modeling Studies

 Design of the Studies

      Widely validated  crop growth models -
 CERES-Wheat  and CERES-Maize (Ritchie  and
 Otter, 1985; Jones and Kiniry, 1986) and SOYGRO
 (Jones et al., 1988) - were used to simulate wheat,
 corn, and soybean yields at selected geographically
 distributed locations within the  Great Lakes, the
 Southeast, and the Great  Plains.  Representative
 agricultural  soils were modeled  at each  site.
 California  crop  yield   changes  were  predicted
 separately by using an agroclimatic index. (See the
 regional chapters, Chapters 14 through 17 of this
 report,  for  descriptions  of  individual  studies.)
 Changes in temperature, precipitation,  and solar
 radiation were  included  in  the  crop  modeling
 studies.  The crop models simulated both rainfed
 and irrigated  production systems.   The  crop
 modeling  approach  allowed  for   analysis  of
 latitudinal  gradients in changes in crop yields and
 provided compatible results for each climate change
 scenario to be used as Inputs hi  the agricultural
 economics study.  (See Ritchie et al., Peart et al.,
 and Rosenzweig, Volume C.)

     The  direct  effects  of CO2 — i.e., increased
 photosynthesis  and improved  water-use efficiency
 -- were  also  included  with  the climate change
 scenarios in some model runs to  evaluate  the
 combined  effects.    The  direct  effects  were
 approximated by computing ratios of elevated CO2
 (660 ppm) to ambient CO2 (330 ppm) values for
 daily   photosynthesis   (Table   6-3)   and
 evapotranspiration rates (see  Peart et al., Volume
 C, for detailed description of method).


     Uncertainties in the  crop modeling studies
 reside in climate model predictions, locations of the
 climate stations (not always in production centers),
 crop growth models, and estimates of the direct
 effects  of CO,.  In particular, the  climate change
 scenarios did*^ not  include changes  in  climate
variability, even though changes in the frequencies
 of extreme events  may considerably affect crop
 Table 6-3. Increase in Daily Canopy Photosynthesis
           Rates Used hi Crop Modeling Studies
                 Soybean      Wheat    Corn
Increase in
photosynthesis      35
 Source:  Peart  et  al.  (Volume C); Ritchie  et al.
 (Volume C); Rosenzweig (Volume C).
 yields. Technology and cultivars were assumed not
 to change from present conditions.

     The CERES and SOYGRO models describe
 relationships between plant processes and current
 climate.  These relationships may or may not hold
 under differing climatic conditions, particularly the
 high temperatures estimated for the greenhouse
 warming. Lack of analysis of the nature and extent
 of agricultural soils  at each modeling  site adds
 uncertainty to the results.

     The direct effects of CO2 in the crop modeling
 results maybe overestimated for two reasons.  First,
 experimental results from controlled environments
 may show more positive effects of CO2 than would
 actually occur in variable, windy, and pest-infested
 (weeds,  insects,  and  diseases) field conditions.
 Second, since the study assumed higher CO2 levels
 (660 ppm)  in 2060 than  will occur if  current
 emission trends continue (555 ppm), the simulated
 beneficial effects of CO2 may be greater than what
 will actually occur.
     Under climate change scenarios alone, without
the direct effects of CO2, yields of corn, soybeans,
and wheat were generally estimated to decrease in
the Great Lakes, Southeast,  arid  Great Plains
regions, except in the northernmost latitudes, where
warmer conditions provided a longer frost-free
growing season.  Figures 6-3 and 6-4 show change
in modeled rainfed corn and soybean yields for the
GISS and GFDL scenarios.  The northern locations
where yields increased included sites in Minnesota.

                                              GISS +Direct Effects of COa

                                              GFDL + Direct Effects of CO2
Figure 6-3.  Percent change in rainfed corn yields simulated by the CERES-Maize model for baseline (1951-
80) and GISS and GFDL climate change scenarios with and without the direct effects of CO2 for selected
locations (Peart et al., Volume C; Ritchie et al., Volume C; Rosenzweig, Volume C).
Decreases in modeled yields resulted primarily from
higher temperatures, which would shorten the crop
life cycle thus curtailing the production of usable
biomass. In the Southeast, rainfall reductions were
a major factor in  the  GFDL results.   Modeled
rainfed yields were estimated to decrease more than
irrigated yields.

      When increased photosynthesis and improved
water-use efficiency were included in  the  crop
models along with  the  climate  change scenarios,
yields increased over the baseline in some locations
but  not in  others  (see Figures 6-3  and 6-4).
Particularly  when combined with the  hotter  and
drier GFDL  climate   change  scenario in   the
Southeast, the direct effects of CO2 would not fully
compensate for changes in climate variables — net
yields were estimated to decrease significantly from
the base case.  Elsewhere,  yields were generally
estimated  to increase,  with relatively greater
increases at the northern locations.

     The crop models were also used to test several
possible adaptations by farmers to  the  predicted
climate changes. For example, a corn variety that is
better adapted to longer growing seasons was tested
in Indiana.  Use of this later maturing variety would
not compensate entirely for  the yield decreases
caused by the warmer climate change scenarios.

Chapter 6
                                          |    | GISS


                                               GISS +Direct Effects of CC>2

                                               GFDL + Direct Effects of CO2
Figure 6-4. Percent change in rainfed soybean yields simulated by the SOYGRO model for baseline (1951-80)
and GISS and GFDL climate change scenarios with and without the direct effects of CO2 for selected.locations
(Peart et al., Volume C; Ritchie et al., Volume C).

      The  potential for  climate  change-induced
decreases in crop yields exists in many agricultural
regions of the United States.  In some northern
areas, crop yields may increase.   Farmers  would
need varieties of corn, soybeans, and wheat that are
better acclimated  to  hotter and  possibly  drier
conditions to substitute for present varieties.

      If the major agricultural areas are to continue
to provide a stable supply  of  food under the
predicted  changes  in  climate,   supplemental
irrigation may be required for many soils. Pressure
for increased irrigation may grow in these regions.
This could further tighten water supply problems in
some areas  and increase pollution from nonpoint
sources (i.e., pollution that is not traceable to any
one distinct  source, such as agricultural chemicals
from  farmers'  fields).  Considerable uncertainty
exists regarding the future  availability  of surface
water  and  groundwater  supplies  with climate
change, and  concerning the competing demands for
and costs of using or extracting the water  (see
Chapter 9: Water Resources).

Regional and National Economics Study

     The  estimated yield changes from the crop
modeling  studies  (not including  California) and

projected changes in irrigation water demand and
availability were  introduced into  an agricultural
economic model to translate the physical effects of
climate  change  into  economic  consequences.
Adams et al. (see Volume C) estimated the regional
and national economic implications of changes in
yields of wheat, corn, soybeans, and other crops and
in the demand for  and  availability  of  water
associated with alternative global climate change

Study Design

      A  spatial  equilibrium  agricultural  model
developed by Adams et  al. (1984) was used to
represent production and consumption of numerous
agricultural   commodities  for  the  U.S.   farm
production regions as designated by the USDA
(Figure 6-5). The model has been used to estimate
agricultural  losses due to increased ultraviolet-B
(UV-B) radiation caused by stratospheric  ozone
depletion (Adams et al., 1984). It consists of farm-
level models for production regions, integrated with
a national-level model of the agricultural sector.
Acreage available for production is based on current
definition of agricultural land classes. Both irrigated
and nonirrigated crop production and water supply
relationships  are included for most regions.  The
model simulates a long-run,  perfectly competitive
equilibrium  and  was developed using  1980-83
economic and environmental parameters.

     A set of model runs was conducted, using the
GISS and GFDL climate change scenarios, with and
without the direct effects on crop yields.  Potential
changes in technology and in future U.S. and world
food demand due to population growth were also
introduced into the climate change analysis.


     The economic approach used in this study has
several limitations. The economic model is static in
the sense that it simulates an equilibrium response
to climate change,  rather than a path  of  future
changes. Substitution of crop varieties, new crops,
and adjustments hi  farm management techniques
were not  included; thus,  the negative  effects of
                  Figure 6-5. Farm production regions in the United States (USDA, 1976).

 Chapter 6
 climate change were possibly overestimated.  Since
 CO2 levels were assumed to be high in the crop
 modeling study,  estimates of the beneficial  direct
 effects of CO2 on crop yields may have biased the
 economic results in the positive direction in some

      Furthermore, changes in yields used as inputs
 to the  economic model were  modeled  for only
 wheat, corn, and soybeans for a limited number of
 sites and regions. The regional crop yield analyses
 cover 72% of current U.S. corn production, 33% of
 wheat production, and 57% of the soybean output.
 National estimates were extrapolated from these for
 all other crop commodities in the model.  Changes
 in risk, where risk is defined as increases in variance
 of crop yields, were not explicitly included in the
 economic analysis. The accuracy of the estimates of
 changes   in   water   supply  and   crop  water
 requirements derived  from the GCMs cannot be
 ascertained.  Potential increases in the demand for
 water by nonagricultural users, which would reduce
 water available for irrigation, were not included. All
 of these assumptions  introduce uncertainties into
 the results.

      Potential changes in international agricultural
 supply, demand,  and prices due to climate change
 are  not explicitly included  hi  the  model.  Such
 changes  could  have  major  impacts   on  U.S.
 agriculture.  For example, warming may enhance
 the agricultural capabilities of high-latitude countries
 such as Canada and the U.S.S.R.  While the net
 effect of climate change on the rest  of the world is
 uncertain,  global changes could  overwhelm  U.S.
 national impacts.   A  net negative effect  on
 agriculture abroad would improve the position of
 U.S.  agricultural  producers  through  enhanced
 exports, but could increase the negative impacts on
 U.S.  consumers  through  increases  in  global
 commodity prices.


     It is important to note  that the results of the
 economic study are not predictions.  Rather,  they
 are initial estimates of how the current agricultural
 system would respond  to the  projected climate
 change scenarios.

     The  economic  model  showed  a  small to
moderate  aggregate  loss in economic welfare
associated  with  the  estimated crop  yield  and
hydrologic changes derived from the climate change
scenarios (see Table 6-4).  For the moderate GISS
climate change scenario, net  losses were small; for
        Table 6-4.  Aggregate Economic Effects of GISS and GFDL Doubled CO2 Climate Change on U.S.
                   Agriculture with and without the Direct Effects of CO2 on Crop Yields
                                                         Economic effects
                                                     (billions of 1982 dollars')
GISS Analysis 4a:
without CO2
GISS Analysis 4:
with CO2
GFDL Analysis 4:
without CO2
GFDL Analysis 4:
 Analysis 4 includes the crop yield and irrigation water supply and demand consequences of climate change
 throughout the United States.
Source: Adams et al. (Volume C).

the more  extreme  GFDL  scenario, they  were
greater.  The magnitudes of these changes, which
are annual, may be  compared with the estimated
$2.5 billion (in 1982 dollars) in agricultural losses
due to  increased  UV-B  radiation  caused  by
stratospheric ozone depletion of 15% (Adams et al.,
1984).  In  general, consumers  lose and producers
gain because of the increased prices of agricultural
commodities and inelastic demand (i.e., insensitivity
to price changes) for agricultural crops.

     Higher  CO2  levels  could reduce  negative
economic  impacts (Table 6-4).   Under  the less
severe GISS climate scenario, the CO2 direct effects
were estimated to sufficiently counter the climatic
effects in most regions, so that both producers and
consumers  gain.  With the more severe GFDL
climate change scenario combined with the direct
effects of CO2, lower yields led to higher prices, but
not by as much as occurred with the climate change
scenarios alone.  However,  significant changes in
regional agricultural land use occurred even when
the beneficial direct  effects of CO2 were taken into
     Production of most crops was reduced because
of yield declines and limited availability of land and
resources.    With  climate  change  alone,  corn
production decreased 12 and  47% in the GISS and
GFDL  scenarios,  respectively,   while   soybean
production was estimated to be reduced by 12 and
53% for the same scenarios.  In all scenarios, land
under production in Appalachia, the Southeast, the
Mississippi Delta, and the Southern Plains could
decrease on  average  by 11 to 37%, while in the
Lake States, the Northern Plains, and the Pacific it
could increase by small amounts (see Figure 6-6).
While availability of agricultural soils was included
in the economic analysis, the sustainability of crop
production in northern areas was not studied.

     Irrigated acreage was estimated to increase in
all  areas,  primarily  because irrigation becomes
economically feasible as agricultural prices rise (see
Figure 6-7).  These changes reflect both increased
demand by farmers for irrigation water and changes
in water  availability as estimated  by the GCM
scenarios, but do not take  into account changes in
competition with industrial or municipal  users.
                  I   | GISS

                  JIM GFDL

                      GISS +Direct Effects of OO2

                      GFDL + Direct Effects of CO2
 Figure 6-6.   Percent change in regional agricultural acreage simulated by an economic model of the U.S.
 agricultural sector for the GISS and GFDL climate change scenarios with and without the direct effects of CO2
 on crop yields (Adams et al., Volume C).

 Chapter 6
P.cilto  I   I GISS


      [552 Glss +Dlrect Effects of COz

      HI GFDL+Direct Effects of C02
                                             Southern Plains
 Figure 6-7.  Change (100,000s of acres) in regional irrigation acreage simulated by an economic model of the
 U.S. agricultural sector for the GISS and GFDL climate change scenarios with and without the direct effects of
 CO2 on crop yields.  Changes are not shown in the Great Lakes, Corn Belt, Appalachia, and Northeast because
 currently irrigated acreage is  small  (2% of total U.S. irrigated acreage)  in these regions (Adams et al.,
 Volume C).
      Technological changes, such as higher yielding
crop varieties, chemicals, fertilizers, and mechanical
power,  have  historically enabled  agriculture  to
increase production with the same amount of, or
less, land, labor, and other resources.  When the
effect of future technological change (based on yield
increases from 1955 to 1987) was modeled along
with the less severe GISS climate change  (without
the direct  effects of CO2), most of the adverse
climate effects were estimated to be offset. Under
the  severe   GFDL  climate  change   scenario,
continued and substantial improvements in yields
would be required to overcome the climate change
effects.  Stated another way, the adverse effects of
climate change  could  negate  most of the higher
output attributable to improved technology over the
next 50 years.  It is important to note, however, that
the rate of future  technological advances is very
difficult to predict.  Increasing food demand from
higher TLS. and world population aggravated the
estimated economic losses from the climate change

                                       Food Supply and Exports

                                            The economic analysis implies that although
                                       climate change could reduce the productive capacity
                                       of U.S. agriculture, major disruption in the supply of
                                       basic commodities for American consumers would
                                       not occur.  Domestic consumers would face slightly
                                       to moderately higher prices under some analyses,
                                       but supplies could be adequate to meet current and
                                       projected domestic demand. However, if droughts
                                       occur  more  frequently  under changed climate,
                                       effects on agriculture may be more severe.

                                            Exported  commodities in  some scenarios
                                       decline by up to 70%, assuming the demand for
                                       exports remains constant.  Thus,  climate change
                                       could  affect  the  United States in its role  as a
                                       reliable supplier of agricultural export commodities.
                                       It is  likely  that  supply  of  and  demand for

agricultural   commodities  could   shift   among
international  regions,  and  responses  of  U.S.
agriculture will take place in this global context.
There is a great need to  determine the nature of
these changes in global agriculture by analyzing the
potential impacts of climate change on both major
world agricultural production regions and potentially
vulnerable food deficit regions.

Regional Economics and Land Use

     Regional shifts in U.S. agricultural production
patterns (not only grain crops but also vegetables
and fruits) are highly likely, as all climate change
scenarios tested show that the southern areas of the
United States become less productive relative to the
northern areas.  This is primarily because the high
temperatures estimated for climate  change would
stress crop production more in southern areas than
in northern areas where crops are currently limited
by lower temperatures and shorter growing seasons.
However, increased agricultural production may be
difficult to sustain in the North, because some soils
may be less fertile and  may have lower water-
holding capacity.  Crops grown in soils with lower
water-holding   capacity   require   more   evenly
distributed rainfall to produce comparable yields.

      Regional changes in agriculture would have
important implications for rural communities. As
production areas shift, climate change effects would
reverberate  through these communities and are
likely to  result in  structural  changes  in  local
economies,  such as  relocation of markets and
transportation networks.   At its  most  extreme,
climate  change could  cause  dislocation of  rural
communities through farm abandonment.

Environmental Concerns

      Regional agricultural adjustments could place
 environmental resources at risk.  Where agricultural
 acreage  would  increase,  demands  for natural
 resources, such as soil and water, might intensify
 current pressures on environmental elements, such
 as  rivers, lakes,  aquifers, wetlands,  and wildlife
 habitats. Northern States, such as Minnesota and
 North Dakota, could become more productive for
 annual  crops like corn and  soybeans because of
 warmer  temperatures  and  a  longer  frost-free
 growing season.  Given the presence of forests and
 wetlands in these regions, increased agricultural
 production  in  the area might threaten natural
 ecosystems, including  wildlife  habitats  such  as
prairie potholes for ducks  and flyways for bird

     In addition, many of the glacial till soils in the
northern latitudes are not as productive as Corn
Belt soils.  Thus, large increases in production of
crops would most likely require greater applications
of chemical fertilizers. The use of these fertilizers
in humid regions on glacial till and sandy soils is
now creating an environmental  hazard  to  the
underlying groundwater, receiving  waters, and
aquatic habitats  in  many  areas.   With  climate
change, water and fertilizer use would have to be
carefully managed to minimize still more leaching of
water-soluble nutrients such as nitrogen and potash.

Demand for Water for Irrigation

     Water  is the single  most critical factor  in
determining   the  development,   survival,   and
productivity  of crops.   The amount of water that
crops use and thus the demand for irritation water
are governed largely by the evaporation process.
Higher air temperatures due  to  increasing trace
gases hi the atmosphere could heighten evaporative
demands.    Increased  irrigation to  satisfy  these
higher  demands  could  accelerate  depletion  of
groundwater and surface water resources. Also, the
rate of evaporation might outstrip precipitation, thus
decreasing crop yields.

     Studies reported  in the California and  the
Great Plains case studies (see Chapters 14 and 15)
explicitly examined the potential changes in demand
for  water for irrigation.    The  studies  did  not
consider changes hi competing demands for water
such as  industrial and residential use, which also
may change  in a warmer climate.  The California
study, however, considered changes hi supply due to
earlier snowmelt and sea  level  rise.  In these
regions, water is a critical resource for agriculture;
 California and the parts of the Great Plains fed by
the Ogallala Aquifer,  in particular,  depend very
 heavily on irrigation for crop production.

 Irrigation Requirements in the Great Plains

      Allen and Gichuki (see Volume C) computed
 irrigation water requirements for sites in the Great
 Plains for the baseline climate and the GISS  and
 GFDL climate change scenarios.  The direct effect
 of CO2  on water use was also included.  (For study
 design  and  limitations,  see  Chapter  17:  Great
 Plains.)  Major   changes   in  irrigation water

 Chapter 6
 requirements were estimated for all locations in the
 Great Plains and for all crops (see Figure 6-8).  The
 most significant would be the persistent increases in
 seasonal net  irrigation  water  requirements  for
 alfalfa,  which would be driven by the climate
 changes in temperature, wind, humidity, and solar
 radiation, and by the lengthening of the  growing
 season.  Decreases in irrigation requirements were
 estimated for winter wheat in most regions. These
 decreases would  be the result of earlier  planting
 dates and  shorter  crop  life  cycle  due  to  high
 temperatures.  When crop varieties appropriate to
 the longer growing season were modeled, irrigation
 water requirements for winter wheat were estimated
 to   increase.       Simulated    irrigation   water
 requirements during  peak  periods  increased in
 almost all areas (see Figure 6-9).
| "
i '"
1 K
1 ซ
| <0
i 20
* o

~ ^






Nebraska Kansas Oklahoma Texas
I 20
* n
1 'I0

i — |


1 — 1 "


Nebraska Kansas Oklahoma Texas
i o
i "*
f '"
ฐ .18
a .ซ

1 	 1

Nooraska Kansas Oklahoma Texas
Figure  6-8.   Percent  change  in  net  seasonal
irrigation requirements for GISS and GFDL climate
change scenarios with direct effect of CO, on crop
water use included (Allen and Gichuki, Volume C).
                                                       Figure 6-9.   Percent  change  in  peak irrigation
                                                       requirements of corn for GISS and GFDL climate
                                                       change scenarios with direct effect  of CO, on crop
                                                       water use included (Allen and Gichuki, Volume C).
      While farmers  in the Great  Plains  would
 probably shift  to  longer season crops,  climate
 change conditions (warmer temperatures and drying
 in some areas) during the later summer months
 could increase irrigation requirements and elevate
 leaf temperatures to a point that exceeds optimum
 temperatures  required for high productivity.  This
 might make it uneconomical to take full advantage
 of the longer growing season, especially if the higher
 CO2 levels increase photosynthesis and offset the
 effects of a shorter season to some degree.

 Water Resources for Agriculture in California

      In the California regional case  study, Dudek
 (see Volume C) characterized the potential shifts in
 demand  for water for agricultural production that
 would accompany shifts in cropping patterns driven
 by  changing  climate.    Changes  in  competing
 demands for  water from industrial  or municipal
 users were not considered.  (For description of
 study  design  and  limitations, see  Chapter 14:
 California.)  When climate change was considered
 alone, groundwater extraction and surface water use
 were estimated to decline in California as a result of
 changes  in both supply of (derived  from  GCM
 climate change scenarios) and agricultural demand
 for water. When the direct effects of CO2 on crop
yields were included, groundwater extraction would
increase  because of improved  yields of all crops

except corn and because  of enhanced  economic
welfare.   Institutional  responses to  changes  in
surface and groundwater use could  include water
transfers, which could improve irrigation efficiency.
When  water   markets  were  included  in  the
simulations, economic welfare was improved by 6 to
15% over the base, while  crop acreage  increased
and groundwater extraction decreased.

Implications for Demand for Irrigation Water
     Expanded use of irrigation is implied from the
regional crop modeling studies for the Great Lakes,
the Southeast, and the Great Plains (see Chapters
15, 16, and 17, respectively).  Increases in irrigated
acreage are also estimated for most regions when
the economics  of  crop production  are factored in
(see Adams et al., Volume C).  When these results
are considered along with the irrigation studies, it
appears that climate change is likely to increase the
demand for water from the  agricultural sector in
many regions.

     In the Great Plains, heightened evaporative
demand and variability of rainfall may increase the
need for irrigation in dryland farming regions. The
simulated changes in irrigation water requirements
are varied, and specific crops and locations probably
would  be  affected  differently.    Higher  peak
irrigation water requirements for some crops may
require larger capacity irrigation systems and may
enlarge energy demands.

     Intensified extraction of water poses serious
environmental and economic problems, especially in
areas  where  groundwater  is being overdrawn.
Streamflows also may slacken if more surface water
is used for irrigation,  thereby aggravating  water
quality problems.   This in turn would harm fish,
wildlife, and recreational activities.

     Regional  changes in cropping locations  and
patterns  of water  use  also  could  exacerbate
agricultural, nonpoint source pollution, and could
further deplete groundwater resources, institutional
responses, such  as markets for water transfers,
could  help  improve irrigation  water management
and alleviate some of these negative effects.

     The economic and social  costs of shifting the
location  of   irrigated  agriculture   could   be
considerable. The construction of irrigation systems
consisting of reservoirs, wells, ditches, pipes, pumps,
and sprinklers currently requires about  $1,500 to
$5,000 per hectare in capital investment (Postel,

Direct Effects of CO2 on Crops

     Global increases in CO, are likely to influence
crop metabolism, growth, aria development directly
through  physiological  processes  and  indirectly
through climate.  Rose (see Volume C) reviewed
recent experimental work performed on  the direct
effects of CO2 on crops, with emphasis  on wheat,
corn,  soybeans, and cotton.

     Elevated concentrations of CO2 directly affect
plant  processes  such  as  photosynthesis  and
transpiration.  Higher CO2 concentrations are also
expected  to  influence  these  processes  indirectly
through  predicted increases in temperature  and
other  changes   in  climate  variables  such  as
precipitation.  Because experimental research has
rarely simultaneously studied both the direct and
indirect effects of plant responses, it is difficult to
assess the relative contributions of  elevated CO2
and   climate  changes to  predictions  of  crop

     Research on  the physiological  effects  has
focused  primarily   on  responses  of  rates  of
photosynthesis   and  transpiration   to  increasing
concentrations of atmospheric CO2. Photosynthesis
rates have increased in these crops in relatively ideal
experimental   environments.     At   moderate
temperatures, most  crops  will  probably show
increases  in  size  and possibly  yield  as  CO2
concentrations  rise.    However, plants  also have
internal  regulation mechanisms that may  lessen
these effects under field conditions.

     Transpiration rates per unit leaf area decrease,
while  total transpiration from the  entire plant
sometimes increases because of greater  leaf area.
Drought-stressed plants exposed  to high partial
pressures of CC^ should be better able to cope with
water deficits. Leaf temperatures in all species are
expected to rise even more than air temperatures;
this may inhibit plant processes that are sensitive to
high temperature.

      Few  studies have  examined  the  interactive
effects of CO2, water, nutrients,  light, temperature,
pollutants,  and  sensitivity  to   daylength   on
photosynthesis  and  transpiration.   Even  fewer

 Chapter 6
studies  have  examined  the   effects  of  these
interactions on the growth and development of the
whole plant.  Therefore, considerable uncertainty
exists concerning the extent to which the beneficial
effects of increasing CO2 will  be seen in crops
growing  in the  field  under  normal  farming
conditions with climate change.

Climate    Impacts   on   Pest-Plant

     Compared with the existing information on the
potential effects  of  climate   change on  crop
production,  relatively little effort has been  directed
toward assessing the influence of climate change on
plant-pest interactions.  Atmospheric increases in
temperature and  CO,, and changes in moisture
regimes,  all  can directly or  indirectly affect
interactions between pests and crops.  Changes in
pests will also depend on regional shifts  in  crop
production.  Although crop pests may be defined as
weeds, insects, or disease pathogens, the EPA work
on this subject focused on insects.

Study Design and Results

     Stinner et al. (see  Volume C) conducted  a
literature survey and modeling experiments on the
major  mechanisms through which climate change
may affect  pest-plant interactions.    This  study
emphasized the major insect pest and pathogen
species of corn and soybeans. The survey indicates
that temperature and precipitation patterns are the
key variables that affect crop-pest interactions. The
temperature increases  associated with the climate
change scenarios  would bring about the following
trends: (1)  increased  survival for migratory and
nonmigratory insect pest  species in the whiter; (2)
northern range extensions  of current pests hi the
higher latitudes and migration of southern species
into the northern Grain Belt regions; (3) an increase
in pest species with more than one generation per
year  in  the  northern  Grain   Belt;  (4)   earlier
establishment  of pest populations  in the  growing
season; and (5) increased abundance of pests during
more susceptible crop growth stages.

     The potential changes in  the overwintering
ranges of four major pests were mapped  for the
GISS and GFDL climate change scenarios and were
compared to present ranges.  The overwintering
capability of the  four major pests may  extend
northward  with both  climate change scenarios.
For example, the potato leafhopper, a serious pest
on soybeans and other crops, at present overwinters
only hi a narrow band along the coast of the Gulf of
Mexico (Figure 6-10). Warmer winter temperatures
in the GFDL and GISS scenarios could cause  a
doubling or tripling  of the overwintering range hi
the United States, respectively. This would increase
the invasion populations hi the  northern states by
similar factors.  The invasions also would be earlier
hi the growing  season, assuming planting dates do
not change.  Both features are likely to lead to
greater insect density and damage. This pattern is
repeated with the other three  pests  studied  and
indicates that these pests, and possibly others, may
move northward and invade cropping systems earlier
in  the  growing season  under  climate change
              Potato leafhopper
Figure 6-10.   Present and potential (GISS and
GFDL  climate  change scenarios)  overwintering
range of the potato leafhopper, Empoasca fabae. a
major pest of soybeans (Stinner et al., Volume C).
     The  Soybean Integrated Crop Management
(SICM) model (Jones et al., 1986) was run with the
GISS and GFDL  climate  change scenarios to
estimate changes  hi damages  caused by corn
earworm.   Modeling results show that earworm
damage to soybeans would increase hi severity in
the Grain  Belt under a warmer climate.   Such
damage could cause grain farmers hi the Midwest to
suffer significant economic losses.   These results
were particularly marked with the warmer and drier
GFDL scenario.


     Lack  of knowledge  about the physiological
effects  of CO,  on  crop  plants  and  lack  of
experimental evidence of direct CO2  effects  on
insect-plant  interactions make  the study of pest-
plant interactions particularly difficult.  Only one
cultivar was used in the modeling study under both
the baseline and the climate change scenarios, and
planting dates remained  the  same.   In reality,
farmers  would  probably  switch  to  a  more
climatically adapted cultivar as climate changed, and
they would advance planting dates in response to
longer growing seasons.


     Increased pest-related  crop  damage could
intensify  pesticide   use.    The  economic  and
environmental  ramifications of such an  increase
could be substantial, not  only in  current farming
regions but also in new areas if agriculture shifts to
the more northern regions  such as the northern
Plains,  the  Great Lakes  States, and  the Pacific
Northwest (see Figure 6-6).

     Increased use of pesticides  would create
additional threats to the integrity of ecosystems
through soil  and water contamination  and could
increase risks  to public  health.   If  agricultural
production is not to rely increasingly on chemicals
that are potentially harmful to the environment, an
increased need  will  exist  for  alternative  pest
management strategies such as biological control,
genetic resistance, and innovative cropping systems.

Effects  of  Climate  Change on  Water

     Agricultural pesticides  are ranked as a high-
priority pollution problem in  many rural regions.
Potentially toxic agricultural  chemicals can  be
transported away from fields via runoff of surface
soils and via  downward leaching  and percolation
through the  soil.   An  understanding of these
processes is needed to evaluate potential threats to
drinking water quality caused by climate change.

Study Design

     Johnson et al.  (see Volume  C) modeled the
partitioning of agricultural pesticides among uptake,
degradation,  surface runoff, and  soil leaching for
wheat, corn, and cotton production regions in the
Great Plains and the Corn Belt. (For details of the
study, see Chapter  17: Great Plains.)  They used
the Pesticide Root Zone Model (PRZM) (Carsel et
al., 1984), which simulates the vertical movement of
pesticides in  the soil.   The model  consists  of
hydrological and chemical transport  components
that simulate runoff, erosion, plant uptake, leaching,
decay,   foliar  washoff,   and  volatilization of  a
pesticide.  The  interactions  among  soil,  tillage,
management   systems,  pesticide  transport,  and
climate change were studied.


     The frequency and duration of precipitation
remain the same in the climate change scenarios,
even though these storm characteristics are critical
factors in determining the transport of agricultural
chemicals and may change.  The scenarios  assume
that  the number of days  with  rainfall  does not
change, but the  intensity of rainfall increases or
decreases.  Runoff and leaching estimates would
most likely be different if  the number of  days of
rainfall changed and daily rainfall amounts  were
held constant.

     The PRZM is a one-dimensional, point model
that does not simulate the transport of water below
the root zone.  Thus, results on a regional  basis
must be extrapolated with care.  The direct effects
of CO2 on crop growth, which may increase  the size
of the plants and the extent  to which  crops cover
the soil, are not included.


     Regional changes in chemical  loadings  of
water and sediment are likely due to climate change
but probably will not be uniform. There appears to
be some consensus  between the GCM  scenarios
concerning the estimated regional changes (Table
6-5).  Modeled pesticides in runoff increase in the
cotton production area, and  pesticides carried by
sediments decrease in the spring wheat  and corn
regions. Leaching of pesticides  tends to  be less
everywhere   owing   to   changes  in  seasonal
precipitation and increased evaporation.


     When the  changes in water quality from the
predicted climate change scenarios are considered

 Chapter 6

 Table 6-5. Summary of GISS and GFDL GCM Model Consensus of PRZM Pesticide Transport by Cropping
           Region and Pesticide3
             Crop and
          pesticide type
runoff losses
erosion losses
Spring wheat
   Highly soluble/short-livedb
   Highly soluble/long-lived
   Slightly soluble/long-lived

Winter wheat
   Highly soluble/short-lived
   Highly soluble/long-lived
   Slightly soluble/long-lived

   Highly soluble/short-lived
   Highly soluble/long-lived
   Slightly soluble/long-lived

   Highly soluble/short-lived
   Highly soluble/long-lived
   Slightly soluble/long-lived
  + indicates that median values increase under climate change; - indicates that median values decrease under
  climate change; blank indicates no consensus among median values.
  Example: median value of all tillage, soil, weather site scenarios for highly soluble/short-lived pesticides in the
  spring wheat crop area.
Source: Johnson et al. (Volume C).
in conjunction with the estimated increases in pests
and  implied  higher  applications  of pesticides
described in the study on pest-plant interactions, the
potential for changes in the nation's water quality
becomes apparent.   Any deterioration in water
quality could adversely affect public drinking water
supplies and human health.

Climate Variability

     The impacts of climate change result not only
from  a slow  change in  the  mean of a climate
variable but often from shifts in the frequency of
extreme events.  Droughts, freezes, and prolonged
periods of  hot  weather  have strong  effects  on
agricultural production. Although the agricultural
modeling  studies  did not include the effects of
         potential changes in climate variability, a review of
         literature  on agriculture and extreme events that
         focuses on the nature and magnitudes of significant
         impacts  is   included   in  Chapter  3:   Climate

              Corn,  soybeans,  wheat, and  sorghum  are
         sensitive  to  high maximum  temperatures during
         blooming.   Lower yields of  corn, wheat, and
         soybeans  have  been   correlated   with   high
         temperatures. The damaging effect of runs of hot
         days on corn yields was particularly evident in the
         U.S. Corn Belt in 1983.

              Although the problems associated with low
         temperatures may diminish with climate  change,
         risks  of frost  damage to crops may  change in the

growing areas of certain crops. Citrus trees are very
vulnerable to low minimum temperatures.  Winter
wheat is often damaged by low temperatures known
as winter kill, especially in the absence of snow.
Even with warmer winters and fewer frosts, more
damage may occur at less extreme temperatures.
For example, the effect of freezing temperatures is
exacerbated if crops have not yet been hardened by
cold temperatures or  if the crops are no longer
dormant and a cold snap occurs.

     Drought is  a  major  cause  of year-to-year
variability in crop production.  In the Dust Bowl
years of the 1930s, yields of wheat and corn in the
Great Plains dropped to  as much as 50% below
normal. In 1988, agricultural disaster in areas of the
northern  Great   Plains   demonstrated  a  high
vulnerability to drought, and nationwide corn yields
decreased by nearly 40%.  Reduction in vegetative
cover associated with drought also brings about
severe wind erosion of soils, which will affect future
crop productivity. Low yields of forage crops during
droughts result in food shortages for livestock and
premature selling of livestock.  If frequency of
drought increases with climate  change, impacts on
agriculture can be severe.

Farm-Level Management and Adjustments
to Climate Change

     Adjustments to existing production practices
would be the first course of action in the  face of
climate change.  The net effect of climate  change
with adjustment by  farmers may be  significantly
different  from the  estimated  effects of  climate
change alone.

Study Design

     Several studies addressed possible adjustments
that could modify the effects  of  climate change.
These adjustments include changes in planting and
harvesting dates, tillage practices, crop varieties,
application  of agricultural  chemicals,  irrigation
technology,  and institutional responses for water
resource management.


     Ritchie et al.  demonstrated that  the  yield
reduction in corn in the Great Lakes could be partly
overcome with selection of new varieties that have
a longer growing season (see Chapter 15:  Great
Lakes). Rosenzweig (see Chapter 17: Great Plains)
showed that adjusting the planting date of winter
wheat to later in the fall would not ameliorate the
effects of climate change,  but  that  changing to
varieties more suited to the predicted climate could
overcome yield decreases at some locations.

     Dudek's  California study found that flexible
institutional responses to climate change would help
to compensate partly for negative climate change
effects (see  Chapter 14: California).  By allowing
movement of water around the state by transferral
of  water  rights,  California's   water  resource
managers  could  alleviate  some   groundwater
extraction  and compensate,  for  surface  water

     Easterling (see  Chapter  15:  Great  Lakes)
found that potential farmer adjustments to climate
change   include  changes  in   tillage   practices,
increased application of fertilizers, selection of more
full-season and heat-resistant varieties, changes in
planting densities, higher use of pesticides, earlier
harvest, and reduced artificial  drying.   Different
adjustments could occur at  different  times in the
cropping season. With the hotter and drier GFDL
scenario, farmers may  have to  adopt production
practices different from those in use today. Climate
changes that leave soils drier during summer than
they are at present will most  likely lead  to  an
increased use  of irrigation in the Corn Belt.  This
increased irrigation  is  also  supported by the
projected price increases for all crops  grown in


     Although detrimental climate change effects on
agriculture  may be  partly offset  naturally  by
Increased photosynthesis and water-use  efficiency
caused by higher levels of atmospheric  carbon
dioxide, farmers themselves  would use a variety of
adjustments to adapt to climate  change.  Market
forces also would aid adaptation to climate change
because they help to allocate resources efficiently.
Each crop and region would respond differently to
climate change, and  adjustment strategies would
need to be tailored to each situation.

     Costs  of  adjustments are  likely to  vary
considerably from region to region. Costs would be

Chapter 6
relatively small in regions where fanners can switch
from one variety to another or from one grain crop
to another, thus enabling continued use of existing
farm machinery and marketing outlets. However, at
locations near the present limit of major agricultural
regions (e.g., the boundary between wheat farming
and  ranching), relatively small changes in climate
may require a substantial switch in type of farming.
This may require substantial costs in new equipment
and  other changes in  agricultural infrastructure.
Severe  climate  change  may  necessitate farm
abandonment in some regions.

     Improvements in agricultural technology also
may be  expected to  ease  adjustment  through
development of appropriate farming practices, crop
varieties,  and livestock species.   Adjustment and
adaptation to climate  change should be included in
agricultural  research  programs   to  enable this
process to occur.


     Animal products  are  a critical  source  of
protein,  energy,  vitamins, and  minerals.   U.S.
livestock production,  mainly  from cattle, swine,
sheep, and poultry, was estimated to be worth over
$31 billion in 1986 (USDA, 1987).

     Climate is known to significantly affect many
aspects of animal health and production. The direct
effects of climate warming on animal health include
differences in incidence of heat and cold stress,
changes in weight gain, and decline in reproductive
capabilities. Indirect  effects may involve trends in
the availability and prices of animal feeds and the
expanded geographic distribution and  activity of
disease-carrying vectors.

     Higher  winter temperatures may lower  the
incidence of respiratory  diseases   in  livestock
(Webster, 1981). Conversely, warmer summers may
necessitate more hours of indoor cooling during
which pathogens are confined to housing structures.
Climate warming may significantly increase the costs
of air-conditioning in  poultry housing.  Changes in
reproductive capabilities such as decreased ovulation
rates, shortened intensity and duration of estrus,
decreased  fertility   of males,   and  increased
embryonic mortality also have been shown to occur
with high temperatures (Ames, 1981).

     Climate   change   may   also  affect   the
survivability, activity, and geographic distribution of
vectors responsible for the transmission of infectious
diseases in livestock. The activity and reproduction
of  disease-carrying vectors  infecting livestock,
humans,   and  crops  are  driven primarily  by
temperature, humidity,  and precipitation.  These
impacts are likely to be similar to those on mortality
and morbidity of disease in humans (see Chapter
12: Human  Health), and they also are similar to
changes predicted for crop pests.

Design of Studies

     Stem et al.  (see  Volume C)  studied  the
available literature on  four  livestock  diseases to
evaluate the range of potential changes in disease
distribution  and occurrence under climate change
conditions.  Schmidtmann and Miller (see Volume
C) used a population dynamics simulation model to
estimate the effects of the GFDL climate change
scenario  on the life  cycle  of the  horn fly,  a
ubiquitous pest of pastured cattle throughout the
United States.


     The  horn fly model is  based on population
counts taken  at various  times under  different
weather and management conditions. However, the
prediction of current horn fly populations appears to
be well correlated with observations. The model is
not validated for the high temperatures predicted
for the climate change.  Schmidtmann and Miller
used  only the hottest  cliniate  change scenario,
GFDL; the other scenarios may have resulted in a
smaller geographic shift in the range of the horn fly.
It should also be noted that the horn fly analysis is
based on current livestock management, breeds, and
distribution. Possible changes in these factors are
beyond the  scope  of this study.   For example,
changes in location and  extent of grassland regions
and forage production caused by climate warming
would  affect livestock  production and horn  fly
     Stem  et  al.  found  that  under  warmer
conditions,  livestock  diseases  currently  causing
serious economic losses in tropical countries could
spread into the United States. Rift Valley fever is
transmitted  principally by  mosquitoes,  and  the
disease may spread as rising winter  temperatures
become able to support an increase in the mosquito

population (see Figure 6-11).  African swine fever
also may become a greater threat.

     The ranges  and activities of disease-carrying
agents of blue tongue and anaplasmosis, diseases
currently causing severe losses in cattle and sheep
production in the United States, may expand.  If
disease-carrying insects increase their winter survival
and reproduce  year-round in  more states, the
geographical distribution of blue tongue, which is
caused  by a  virus, may expand northward and
eastward.  Anaplasmosis, a rickettsial infection of
ruminants, is the second most  important disease of
cattle in the United States.  Distribution of the
insect carrier's habitat could  expand to northern
states with climate change, and the insects' day-to-
day activity may increase; this  process  may also
cause an increase in disease transmission.

     The horn fly causes annual losses  of $730.3
million  in the beef and  dairy cattle  industries
(Drummond, 1987). Schmidtmann and Miller found
that with the very warm  GFDL climate change
scenario, the horn fly season throughout most of the
United States could be extended by 8 to 10 weeks.
The  increase  in  horn  fly  populations  could
substantially  reduce  the  average  daily  gain of
growing  beef  cattle.   Also  under  the  GFDL
simulation, increased pest activity was estimated in
dairy cattle in the North and Northwest -- a result
that could significantly decrease  milk production.
Conversely,   under   the  same   scenario,  the
summertime activity of the horn fly could decrease
in the South because the warmer climate  would
exceed the horn fly's tolerance to high temperatures.


     With climate  change,  patterns of livestock
diseases  and pests may  also  change.  Tropical
livestock diseases may become an increased threat,
because  more  geographical areas  are potential
ranges for the insect  carriers of the diseases.
Temperature conditions may improve in the winter
but   may be  exacerbated   in   the  summer.
Reproductive capabilities may be lower. Livestock
production would also be affected if rangeland areas
shift and forage production levels change.
            | CURRENT & DOUBLED CO2

            ! DOUBLED CO2
Figure 6-11.  States where significant Culex spp. activity permits establishment of Rift Valley fever for current
and doubled CO2 levels (Stem et al., Volume C).

 Chapter 6

     The U.S. agricultural system has historically
been able to  adopt new technologies rapidly and
may be less  vulnerable to climate change than
natural ecosystems.  In fact, global warming may
cause a number of benefits.  Potential benefits of
CO2-induced  climate change include  increases in
length of growing season and hi air temperatures,
which would benefit regions where crop growth is
constrained    by  short   summers   and   low
temperatures.  Longer growing seasons would likely
lead to increased yields of hay and other perennial
crops.   Energy  costs for grain drying may be
reduced, since annual crops would reach maturity
earlier and would have more opportunity to dry in
the  fields.     Furthermore,  in  places  where
precipitation increases during the growing season,
irrigation requirements  could  be  reduced.   If
irrigation requirements  are lessened,  demand on
regional water resources and associated costs  to
farmers may fall.

     However, many reasons to avoid complacency
about  the  predicted  climate  change  remain.
Concern for our major  resources (especially land
and water), rural communities, and the environment
is justified.  While many critical uncertainties exist
regarding the  magnitude and timing of impacts, it
appears that climate change is likely to affect U.S.
agriculture significantly in the coming century.

Costs and Timing of Adjustment

     Since  our  agricultural production system
primarily consists of specialized farms producing
commodities   in   geographically   specialized
production  patterns,  the  costs  of  adjusting   to
changed comparative advantage among agricultural
regions, with  ensuing changed  resource  use and
changed agricultural infrastructure,  may  be  quite
high in some regions. These shifts would also entail
involvement of and costs to the federal government.

     If warming occurs rapidly, U.S. agriculture will
have less tune to adjust  and costs may be greater.
As climate continues to warm, costs may rise at an
increasing rate. Finally, unless CO2 and other trace
gas emissions  are limited, we may be  facing a
continual  and   possibly  accelerating   rate  of
atmospheric accumulations and climate change.  As
the agricultural system strives to adapt to a changing
climate, there may be no chance of optimizing for
static conditions. Rather, the system maybe caught
in forever playing catch-up.

Effects of CO2

     It is  also important to note that the crop
modeling studies showed that the direct CCueffects
on crop photosynthesis and  water-use  efficiency
ameliorate the negative effects of climate change in
some locations under  certain climate conditions;
however, such, effects do not  occur uniformly, and
they do not occur everywhere. Regional changes in
U.S. agriculture occurred with the GISS and GFDL
climate change scenarios both with and without the
direct effects of CO2-  While much work must be
done to  improve both climate and crop models,
policy analysis should consider that the  beneficial
direct effects of CO2 may not offset the negative
effects of climate change.

Environmental Quality

     Changes in the agricultural production system
are likely to  have significant  impacts on resource
use and the environment. Many of the agricultural
studies suggest that climate warming could result in
accelerated rates of demand for water for irrigation
(see Chapter 9: Water Resources),  increases in
pesticide usage to control  changes in pest vectors,
and  changes in water  quality from  agricultural
chemicals.  Decreases  in biological diversity may
limit the  adaptive capacity of agriculture, which
requires a broad base of germ plasm for modifying
current crops and developing new ones (see Chapter
8: Biodiversity).

     A northward migration  of agriculture would
increase the use of irrigation and fertilizers on sandy
soils,  thus endangering  underlying groundwater
quality.  From South Dakota  to southern Canada,
critical prairie wetlands may be lost to drainage and
conversion to cropland.  Many of these  areas are
important wildlife habitats.  Shifts in  agricultural
activities may increase the susceptibility of soils to
wind and water erosion. Climate change could thus
exacerbate  many   of   the   current  trends  in
environmental pollution and resource use associated
with agriculture as well as initiate new ones.

     Sea level rise; an associated impact of climate
change, will threaten low-lying coastal  agricultural
regions  with  seasonal  -- and in  some  instances

permanent  ~  flooding,  saltwater  intrusion  of
freshwater  aquifers   and   rivers,   and   salt
contamination of soils. Agricultural lands in coastal
regions may be  lost.    (See Chapter  9:  Water
Resources, and Chapter 7: Sea Level Rise, for
linkages with agriculture.)

     Furthermore,  climate  change  will  act on
agriculture simultaneously with other environmental
stresses.   Levels of UV-B  radiation caused by
depletion of stratospheric  ozone  are likely  to
increase in the future, as are levels of tropospheric
ozone  and acid  precipitation.   The interactions
among these multiple stresses and climate change
need to be studied in agricultural settings.

Global  Agriculture

     Finally, U.S. agriculture is an integral part of
the  global,  international  agricultural   system.
Consequently, the adjustment of U.S.  agriculture to
climate change cannot be considered in isolation
from   the  rest of  the  world.   The   optimal
configuration of U.S. adjustments will depend very
much on how  simultaneous changes in regional
climates  affect  global agriculture and how  other
countries, in turn, respond to those changes.

     Since  climate   change  appears  likely  to
reconfigure   the   agricultural   activities   and
demographics of rural America, policies should be
examined in  light  of  these potential  effects.
Agricultural policies should be designed  to  ease
adjustments to  climate change and to ensure the
sustainability of our natural and human resources
(see Schuh, Volume  C, and Dudek,  Volume C).
Following are specific policy areas that policymakers
could investigate  to respond appropriately to the
projected climate change.

Commodity Policies

     Agricultural pricing  and production  policies
should promote efficient adjustment to the changing
conditions of global supply and demand induced by
the greenhouse effect, which may include  shifts in
comparative advantage among regions and increased
likelihood of droughts in some regions.  Although
these shifts may be slow, the cumulative effects may
be large and they deserve close monitoring.  Market
forces as well as government programs would play
a crucial role in creating the flexibility to respond to
climate  changes by sending signals on the efficient
use of resources, and in mitigating their ultimate
impact as they have done in the past.  Agricultural
policies  should be evaluated to ensure that they are
appropriate  to  both current  and possible future
conditions hi regard to  their ability  to facilitate
adaptation  to  climate  change.   For  example,
flexibility hi shifting crop types and farm practices
will speed adjustments.

Land-Use Programs

     Federal legislation aimed at reducing the use
of newly plowed grasslands, e.g.,  the  "Sod-Buster
Bill," and the related "Swamp-Buster  Bill," which
restricts agricultural encroachment  into wetlands
subject to flooding and water-logging, are examples
of new  policies meant to protect marginal lands.
The basic goals of these new laws, which are part of
the 1985 Farm Bill, are to protect the most erodible
farmland by removing it from crop production and
to  use   conservation  as  a   tool  for  reducing
overproduction.  Nearly  80 million acres  of  U.S.
cropland were retired under these and other farm
programs in 1988.  Policy research should  address
how these  programs may fare  under changing
climate  conditions.

     Another program established in the 1985 Farm
Bill that may help alleviate the negative effects of
climate   change  is  the  Conservation  Reserve
Program. This program is aimed at removing from
crop production the cropland classified as "highly
erodible" by the Soil Conservation Service. The bill
created  a new form of long-term contract of up to
10 years and provides payments  to farmers  who
apply conservation practices, such  as maintaining a
grass  cover, on those acres.   If successful,  the
Conservation Reserve  Program may reduce  the
impact  of  climate fluctuations  on  total  grain
production by taking the most  sensitive lands out of

     The 1988 drought, however, demonstrated that
the Conservation Reserve Program may be  difficult
to maintain in the face of climate stress.  As the
drought worsened during the summer, use of the
set-aside lands was requested so that  badly hit
farmers  could salvage some economic benefits from
these acres.  Such conflicts may be more common
hi the future, and land retirement strategies must be

Chapter 6
weighed against  possible  needed  increases  in

     Awareness of potential changes in agricultural
land use due to regional climate change should be
built into land-use planning programs, especially in
regions where agricultural activities may expand into
natural,  unmanaged  ecosystems.    Large-scale
drainage  and   water  projects  would  need
environmental impact studies to carefully assess this
potential  expansion  of  agricultural  land  (see
Baldwin, Volume J).

Water-Resource Management Programs

     Current water supply policies do not generally
encourage optimum water-use efficiency. A greater
degree of water efficiency should promote flexibility
in light  of the  potential for  increased  irrigation
demands with  climate  change.   Policies such as
water transfers and markets should be considered
for irrigated areas.

Water Quality  Policy

     The increased use of agricultural  chemicals,
along  with  changes in  the  hydrological  cycle,
potentially threaten both soil  and water supplies,
and   eventually,   public   health.      Negative
consequences  could be avoided or  lessened by
including potential climate change effects in water
quality planning and by supporting alternative pest
management strategies that use such techniques as
biological control, genetic resistance, and innovative
cropping systems.

Risk Management and Drought Policy

     Changes  in  the  frequency, intensity,  and
location of extreme  events  are important  for
agriculture and the regional income that it produces.
The  adequacy of  the private  crop insurance  and
federal  disaster payment  programs  should be
assessed in the  face of climatic uncertainty.  For
example, only about  20 to  25%  of potentially
insurable acreage is currently covered by crop
insurance. Farmers tend to rely on federal disaster
relief programs to bail them out of such disasters as
droughts, floods, hail,  and windstorms.   Financial
risk is also part of the credit structure that covers
land, equipment, and production in modern farming.
     The  frequency and magnitude of  climate
extremes  may  be altered with climate  change.
Responding to  the changes may be costly for the
government if crops fail frequently.  The Drought
Relief Bill for the drought of 1988 is scheduled to
cost $3.9 billion to cover just 1 year of  a  climatic
extreme.  On  the other  hand,  some  areas that
currently suffer from climate extremes may benefit
from climate change.  Risk policy mechanisms for
relief, recovery, and  mitigation of climate change
should be examined  so that they will be ready to
help farmers adjust.

     A national drought policy is strongly needed to
coordinate federal response to the possibility  of
increased frequency and duration of future droughts
due to  climate change.    Even without  climate
change, such a policy is  needed not only for the
agricultural sector but also for other sectors.

International Trade Agreements

     Policies designed to ease the  adjustment  to
greenhouse effects must be global in scope because
the effects, although varied, are global  in nature.
Comparative advantage will likely shift significantly
both within the  United  States  and  in  other
countries. Population and economic activities also
would change geographically with climate  change,
thus  affecting   the   location  of  demand  for
agricultural products. It  is already a goal of U.S.
agricultural policy to incorporate global  conditions
of supply  and demand into the agricultural sector.
The potential seriousness of the impacts on the
agricultural production system of the greenhouse
effect may provide added incentive to establish such
policies both nationally and internationally.  The
vulnerability  of current and potential food-deficit
regions to climate change should also be considered.

Agricultural  Contributions to the
Greenhouse Effect

     Agriculture itself is an active contributor to the
greenhouse effect.  Clearing of forested land for
agriculture often involves  burning  of  trees and
shrubs that release CO2.  The biomass that is not
burned tends to decay gradually, also emitting CO2.
Agricultural activities release other radiatively active
trace gases.  Flooded rice fields emit methane
(CH4) as a product of the anaerobic decomposition
of organic matter. Ruminants also release methane

as a consequence of their digestive processes. In
addition,   soils  may  volatilize   some   of   the
nitrogenous fertilizer applied to them in the form of
nitrous oxide (N2O).  Finding effective ways to
reduce these emissions presents a major challenge
to the agricultural research community.  In  this
regard, the  Conservation Reserve Program   and
forestation efforts could provide a partial solution,
since vegetation fixes CO2 from  the  air.   (See
Lashof and Tirpak, 1989, for further discussion of
agriculture's contribution to the greenhouse effect.)

Agricultural Research

     The agricultural research community should
enhance climate change research from the field level
to the national policy level. It should continue to
breed heat- and drought-resistant crop varieties and
new crop species in preparation for global warming.
Research in biotechnology may  also be directed
toward alleviating the negative effects  of climate
change.    Improved  water-use   and  irrigation
efficiency also take on renewed importance in the
light   of  potential   climate   change.     Energy
requirements  of  the  agricultural system under
climate  change  should   be  defined,  given   the
potential for increases in energy-intensive activities
such as irrigation and application of agricultural
chemicals.   Research attention  also should be
directed toward reducing  agricultural emissions of
trace gases.

     1.  International  agriculture  —  Study  the
        potential shifts in international comparative
        advantage and the vulnerability of food-
        deficit  regions,   and    evaluate   the
        implications of such shifts for the United

        One of the most crucial  areas for further
        research  is  the projection of  potential
        climate change effects at the international
        level.  Potential changes  in  agricultural
        yields and production of major crops, and
        impacts on regions that are food-deficient
        now or that may become food-deficient in
        the   future,  all  need  to be   studied.
        Economics and policy  research  should
        consider the implications of shifts in global
        agriculture for the  levels  of U.S.  crop
   exports and the role of the United States as
   a reliable supplier  of agricultural export

2.  Crop and livestock productivity — Study the
   interactive effects of climate variability and
   change, CO2,  tropospheric ozone, UV-B
   from  stratospheric  ozone  depletion,  and
   other  environmental and societal variables
   on agricultural productivity.   Determine
   how   changed  climatic  variability  may
   amplify or lessen  the  preliminary EPA

   Because  of  the  significant  production
   changes indicated by these studies, the need
   for better simulation of the direct effects of
   CO2 in the crop  models, and the limited
   adjustment studies performed, further crop
   research should be conducted on a longer
   term  basis.    Necessary work includes
   resolving the differences in forecasts of the
   GCMs, and designing more appropriate
   scenarios including transient climate change
   and   changes   in  climatic  variability.
   Physiologicallybasedsubmodels are needed
   for the effects of increased CO2 on various
   crops.   The effects  on other  major crops
   such as cotton also should be studied. Crop
   models  should  be improved  in  their
   simulation of  the  effects  of  increasing

   Research on  the direct  CO2 effects  on
   crops to this point has provided windows of
   knowledge  concerning  certain  crops  at
   specific stages of their life cycles.  Both the
   direct  and the climate change effects of
   high CO2 are probably quite different at
   different stages of development. Research
   should evaluate the interactive effects of
   CO, and temperature over the whole life
   cycte of the plant, with varying conditions of
   water and nutrition, rather than with plants
   under  optimal  conditions.    Then crop
   response to  the  combined climatic and
   physiological  effects  of CO,  may  be
   predicted more realistically. Much more
   research on climate change and livestock
   production is needed.  Important research
   areas  include crop-livestock interactions,
   reproduction, and  diseases.

Chapter 6
     3. Adaptation strategies — Study the dynamic
        nature of climate change:  What is the rate
        of  adaptation  of  regional  agricultural
        systems compared with the rate of climate
        change?    Evaluate  the  thresholds  of
        sensitivity of U.S.  agriculture.    Studies
        should analyze the ability of various aspects
        of the agricultural production systems to
        adapt to various  rates and degrees  of
        climate   change   to   determine  these
        thresholds of sensitivity. It would also be
        useful to identify the costs of different types
        of adjustments and the regions most likely
        to experience greater costs.

     4. Agricultural  economics —  Expand  the
        national  analysis  to  include crops and
        regions not now  included (for example,
        cotton and  grasslands, and  the western
        regions of  the United States).   Conduct
        further  analyses  of  regional  shifts  in
        agriculture.    Studies that  link  water
        resource and agriculture models are needed
        to estimate  changes  in  water  demand
        among agriculture and competing users.
        Thus, estimates  of  actual  changes  in
        irrigated acreage could be made.

     5. Environmental impacts - Elucidate the
        impacts   of  climate  change  on  water
        quantity,    water   quality,   and  other
        components of the environment caused by
        shifts in crop and livestock production and
        related industries.

     6. Agricultural emissions of trace  gases —
        Discover effective ways to reduce emissions
        of methane from livestock, nitrous oxide
        from  fertilizer  application,  and  other
        agricultural sources of trace gases.
Acock,  B.,  and L.H.  Allen,  Jr.   1985.   Crop
responses to elevated carbon dioxide concentrations.
In: Strain, B.R., and J.D. Cure, eds. Direct Effects
of Increasing Carbon  Dioxide  on  Vegetation.
Washington,  DC:  U.S. Department  of Energy.
DOE/ER-0238.  pp. 33-97.
Adams, R.M., SA. Hamilton, and BA. McCarl.
1984.   The  Economic Effects  of  Ozone on
Agriculture. Corvallis,  OR: U.S. Environmental
Protection Agency.  EPA-600/3-84-090.

Ames, David  R.   1981.   Effect of  climate on
livestock production data.  In: Knapp, F.W., ed.
Systems  Approach  to  Animal   Health  and
Production.    Lexington,   KY:    University  of
Kentucky, pp. 148-148.

Callaway, J.M., FJ. Cronin, J.W. Currie,  and J.
Tawil. 1982. An analysis of methods  and models
for assessing the direct and indirect impacts of CO2-
induced environmental changes in the  agricultural
sector of the U.S. economy.  Richland, WA: Pacific
Northwest Laboratory, Battelle Memorial Institute.

Carbon Dioxide Assessment Committee.   1983.
Changing  Climate.  Washington,  DC: National
Academy of Sciences.

Carsel, R.F., C.N. Smith, LA. Mulkey, J.D. Dean,
and  P.  Jowise.   1984.   Users  Manual for the
Pesticide Root Zone Model (PRZM). Athens, GA:
U.S.  Environmental Protection Agency.  EPA-

Council for Agricultural Science and Technology.
1988.  Long-Term  Viability of U.S. Agriculture.
Ames, IA:  Council for  Agricultural Science and
Technology. Report No. 114.

Cure, J.D.    1985.   Carbon  dioxide  doubling
responses: A crop survey. In: Strain, B.R., and J.D.
Cure, eds.  Direct  Effects  of Increasing Carbon
Dioxide on Vegetation.  Washington, DC:  U.S.
Department of Energy.  DOE/ER-0238.  pp. 99-
Decker, W.L., V. Jones, and R. Achutuni.  1985.
The impact of COg-induced climate change on U.S.
agriculture. In: White, M.R., ed. Characterization of
Information Requirements for Studies of  CO2
Effects: Water Resources, Agriculture, Fisheries,
Forests and Human Health. Washington, DC: U.S.
Department of Energy. DOE/ER-0236. pp. 69-93.

 Dudek, D.J.  1987.   Economic implications of
 climate change impacts on southern agriculture. In:
 Meo, M., ed.  Proceedings of the Symposium on
 Climate  Change in the  Southern United States:
 Future Impacts and Present Policy Issues. Norman,
 OK: University of Oklahoma,  Science and Public
 Policy Program,  pp. 44-72.

 Drummond, R.O.   1987.  Economic aspects of
 ectoparasites of cattle in North America.   In:
 Leaning, W.H.D., and J. Guerrero, eds. Proceedings
 of the MSB AGVET Symposium, The Economic
 Impact of Parasitism  in  Cattle.   XXIII.  World
 Veterinary Congress.  Montreal, Quebec, pp. 9-24.

 Hansen, J., I. Fung, A. Lacis, D. Rind, G. Russell,
 S. Lebedeff, R. Ruedy, and P. Stone.  1988. Global
 climate changes as forecast by the  GISS 3-D model.
 Journal of Geophysical Research 93(D8):9341-9364.

 Jones, CA.,  and  J.R. Kiniry, eds.   1986.  CERES-
 Maize: A Simulation Model of Maize Growth and
 Development.  College Station, TX: Texas A&M
 University Press.

 Jones,  J.W.,  K.J.   Boote,   S.S.  Jagtap,   G.
 Hoogenboom,  and  G.G.  Wilkerson.     1988.
 SOYGRO V5.41: Soybean Crop Growth Simulation
 Model.   User's  Guide.   Florida Agr. Exp.  Sta.
 Journal No. 8304, IFAS. Gainesville, FL: University
 of Florida.

 Jones, J.W., J.W. Mishoe, G. Wilkerson, J.L. Stimac,
 and W.G. Boggess. 1986.  Integration of soybean
 crop and pest models. In: Frisbie,  P.E., and P.
 Adisson,  eds.  Integrated Pest Management  on
 Major Agriculture  Systems.    Texas Agriculture
 Experiment Station.   Publication No.  MP-1616.
 College Station, TX: Texas A&M  University.

 Kimball, B A. 1985.  Adaptation of vegetation and
 management practices  to a higher carbon dioxide
 world. In: Strain, B.R., and J.D. Cure, eds. Direct
 Effects of Increasing Carbon Dioxide on Vegetation.
 Washington,  DC: U.S. Department  of Energy.
 DOE/ER-0238.  pp. 185-204.

Land Evaluation  Group.  1987.  Implications of
 climatic warming for Canada's comparative position
in agricultural  production and trade.   Guelph,
Ontario: University of Guelph, University School of
Rural Planning and Development.

 Lashof, D., and D. Tirpak,  eds.   1989.   Policy
 Options  for  Stabilizing Global  Climate.   Draft
 report.   Washington,  DC: U.S.  Environmental
 Protection Agency.

 Parry, M.L., T.R. Carter, N.T. Konijn, eds.  1988.
 The Impact of Climatic Variations on Agriculture.
 Vol. 1. Assessments in Cool Temperate and Cold
 Regions.  Dordrecht: Kluwer.

 Postel, S.  1986.  Altering the Earth's Chemistry.
 Assessing  the  Risks.  WorldWatch  Paper  71.
 Washington, DC: Worldwatch Institute.

 Ritchie, J.T., and S. Otter.  1985. Description and
 performance  of CERES-Wheat: A user-oriented
 wheat yield model.  In: Willis, W.O.,  ed.  ARS
 Wheat Yield Project.    Washington,   DC: U.S.
 Department of Agriculture, Agricultural Research
 Service. ARS-38.  pp. 159-175.

 Schneider, K.  1988. Drought cutting U.S. grain
 crop 31% this year.  The New York Times August

 USDA.   1976.  U.S. Department of Agriculture.
 Handbook  of Agricultural  Charts.  Agricultural
 Handbook 504. Washington, DC: U.S. Government
 Printing Office.

 USDA.   1987.  U.S. Department of Agriculture.
 Agricultural Statistics.   Washington,  DC:  U.S.
 Government Printing Office.

 U.S.  Department  of Commerce.   1983.   U.S.
 Department of Commerce, Bureau of the Census.
 1982 Census of Agriculture, Vol. 1.  Geographical
 Area Series", Part 51, United States  Summary and
 State Data.   Washington, DC: U.S. Government
 Printing Office.                         s

 Warrick, R.A., R.M. Gifford,  M.L. Parry.   1986.
 CO2, climatic change and agriculture. Assessing the
 response  of food  crops  to the direct  effects  of
 increased  CO2 and climatic change.  In:  Bolin, B.,
 B.R. Doos, J. Jager, and  RA.  Warrick, eds.  The
 Greenhouse   Effect,  Climatic   Change   and
 Ecosystems. A Synthesis of the Present Knowledge,
 SCOPE 29.  New York: John Wiley and Sons. pp.

Webster, A.J.F.   1981.   Weather and infectious
 disease in  cattle. The Veterinary Record 108:183-

Chapter 6
World Food Institute.  1987.  World Food Trade
and UJS. Agriculture, 1960-1986.  Ames, IA: Iowa
State University.
Zachariah, K.C., and Vu, M.T. World Bank.  1988.
World Population Projections. 1987-1988 Ed. Short-
and Long-Term Estimates. Baltimore, MD: Johns
Hopkins University Press.

                                         CHAPTER 7
                                    SEA LEVEL RISE

Global warming could cause sea level to rise 0.5 to
2 meters  by 2100.   Such a rise would inundate
wetlands and lowlands, erode beaches, exacerbate
coastal  flooding,  and  increase  the  salinity  of
estuaries and aquifers.

•   A 1-meter rise could drown approximately 25 to
    80% of the U.S. coastal wetlands; ability  to
    survive would depend largely on whether  they
    could migrate inland or whether  levees  and
    bulkheads  blocked  their migration.   Even
    current sea level trends threaten the wetlands of

•   A 1-meter rise could inundate 5,000 to 10,000
    square miles  of dryland if shores were not
    protected and 4,000 to 9,000 square miles  of
    dryland if only developed areas were protected.

*   Most coastal barrier island communities would
    probably respond to sea level  rise by raising
    land with sand pumped  from offshore.  Wide
    and heavily urbanized islands may use levees,
    while communities on lightly developed islands
    may adjust to a gradual landward migration  of
    the islands.

•   Protecting  developed   areas   against  such
    inundation and erosion by building bulkheads
    and levees, pumping sand, and raising barrier
    islands  could  cost   $73  to   $111  billion
    (cumulative capital costs in 1985 dollars) for a
    1-meter  rise by the year 2100 (compared with
    $6 to $11 billion under current sea level trends).
    Of this total, $50 to $75 billion would be spent
    (cumulative capital costs in 1985 dollars) to
    elevate beaches, houses, land, and roadways by
    the  year  2100  to   protect barrier  islands
    (compared with $4 billion under current trends).
    Developed  barrier  islands would likely be
    protected from sea level rise because of their
    high property values.
 •   The Southeast would bear approximately 90% of
    the land loss and 66% of the shore protection

 Policy Implications

 •   Many of the necessary responses to  sea level
    rise, such  as rebuilding  ports,  constructing
    levees, and pumping sand onto beaches, need
    not be implemented until the rise is imminent.
    On the other hand, the cost of incorporating sea
    level rise into a wide variety of engineering and
    land use decisions would be negligible compared
    with the costs of not responding until sea level

 •   Many wetland ecosystems are likely to survive
    sea level rise only if appropriate measures  are
    implemented in the near future.  At  the state
    and local levels, these measures include land  use
    planning,  regulation,  and  redefinitions  of
    property rights. The State of Maine has already
    issued regulations to enable wetlands to migrate
    landward  by  requiring  that  structures  be
    removed as sea level rises.

•   The coastal  wetlands protected under Section
    404 of the Clean Water Act will gradually be
    inundated. The act does not authorize measures
    to ensure survival of wetland ecosystems as  sea
    level rises.

•   The  National Flood  Insurance Program may
    wish to consider the  implications of sea level
    rise on its future liabilities.  A recent HUD
    authorization act requires this  program  to
    purchase property threatened with erosion. The
    act may imply a commitment by the federal
    government to compensate property owners  for
    losses due to sea level rise.

•   The need to take action is particularly urgent hi
    coastal Louisiana, which is already losing 100
    square kilometers per year.

Chapter 7

    Global warming  from the greenhouse effect
could raise sea level approximately 1 meter  by
expanding ocean water, melting mountain glaciers,
and causing ice sheets hi Greenland to melt or slide
into the oceans. Such a rise would inundate coastal
wetlands and lowlands, erode beaches, increase the
risk of  flooding,  and increase  the  salinity  of
estuaries, aquifers, and wetlands.

    In the last 5 years, many coastal communities
throughout the world have started to prepare for the
possibility of such a  rise.  In the United States,
Maine has enacted a policy declaring that shorefront
buildings will have to  be moved to enable beaches
and wetlands to migrate inland to higher ground.
Maryland has shifted  its shore-protection strategy
from a technology that can not accommodate sea
level rise to one that can. Seven coastal states have
held large public meetings on how to prepare for a
rising sea.   Australia,  the Netherlands, and the
Republic of Maldives are beginning  to undergo a
similar process.


    Ocean  levels  have always  fluctuated  with
changes in global temperatures. During the ice ages
when the earth was 5ฐC (9ฐF) colder than today,
much of the  ocean's water was frozen in glaciers
and sea level often was more  than 100 meters (300
feet) below the present level (Dorm et al., 1962;
Kennett, 1982; Oldale,  1985). Conversely, during
the last interglacial period (100,000 years ago) when
the average  temperature  was about 1ฐC  (2ฐF)
warmer than  today, sea level was approximately 20
feet higher than the current sea level  (Mercer,

    When  considering  shorter periods  of time,
worldwide sea level rise must be distinguished from
relative sea level rise.  Although climate change
alters worldwide sea level, the rate of sea level rise
relative to a particular  coast  has  greater practical
importance and is  all  that monitoring stations can
measure. Because most coasts are sinking (and a
few are rising), the range  of relative sea level rise
varies  from  more than  3 feet  per  century  in
Louisiana and parts of California and Texas to 1
foot per century along most of the  Atlantic and
gulf coasts, to a slight drop in much of the Pacific
Northwest (Figure 7-1).  Areas such as Louisiana
provide  natural  laboratories for  assessing  the
possible effects of future sea level rise (Lyle et al.,
    1850 1865
            1880 1895  1910  1925  1940  1955  1970 1985
                      SITKA, AK
             CHARLESTON, SC

                MIAMI BEACH, FL
          GALVESTON, TX
Figure 7-1.  Tune series graph of sea level trends
for New York, Charleston, Miami, Galveston, and
Sitka (Lyle et al., 1987).
    Global sea level trends have  generally been
estimated by combining the trends at tidal stations
around  the  world.    Studies  combining  these
measurements suggest that  during the last century,
worldwide sea level has risen 10 to 15 centimeters
(4 to 6 inches) (Barnett,  1984;  Fairbridge and
Krebs, 1962). Much of this rise has been attributed
to the global warming that has occurred during the
last century (Meier, 1984;  Gornitz et al., 1982).
Hughes (1983) and Bentley  (1983) estimated that a
complete  disintegration  of West Antarctica  hi
response to global warming  would require a 200- to
500-year period,  and  that  such a disintegration
would  raise sea  level 20 feet.    Most  recent
assessments, however,  have focused  on the likely
rise by the year 2100.  Figure 7-2 illustrates recent
estimates of sea level rise, which generally fall into
the range of 50 to 200 centimeters.

                                                                                          Sea Level Rise

f 3.0
r* ro
= b

• Hoffman (1983) High

* Olacler Volume Estimate of Polar • Hoffman (1983) Mid-High
_ Board Augmented With Thermal
Expansion Estimates by NRC • Meier (198Sb) High
• WMO(1986) High
> Hoffman (1983) Mid-Lorn
/ / • Reve!le(1983)
/ / •!• • Hoffman (1983) Low
^ Past Century ( //^S^'' Meler " 9B5b) Low"
Estimated I ^x^^iwiMO (1986) Lovซ
0.12m Rise \^P^*^ | |
~^~ 2000 2050 2100
 Figure 7-2.   Estimates  of future  sea level  rise
 (derived from Hoffman,  1983, 1986; Meier, 1985:
 Revelle, 1983).

    Although most studies have focused on the
 impact of global warming on global sea level, the
 greenhouse effect would  not necessarily raise sea
 level by the same amount everywhere.  Removal of
 water from the world's ice sheets would move the
 earth's  center of gravity away from Greenland  and
 Antarctica and would thus redistribute the oceans'
 water toward the new center of gravity. Along the
 U.S. coast, this effect would generally increase sea
 level rise by less than 10%.  Sea level could actually
 drop, however, at Cape Horn and along the coast of
 Iceland. Climate change could also affect local sea
 level by  changing  ocean currents, winds,  and
 atmospheric pressure; no one has estimated these


    In this section and in the following sections, the
 effects  of  and  responses to  sea level rise  are
presented separately. However,  the distinction is
largely  academic and is  solely  for presentation
 purposes. In many cases, the responses to sea level
 rise  are  sufficiently  well  established  and  the
 probability of no response is sufficiently low that it
 would be misleading to discuss the potential effects
 without also discussing responses.  For  example,
 much of Manhattan Island is less than 2 meters
 above high tide; the effect of sea level rise would
 almost certainly be the increased use of coastal
 engineering structures and not the  inundation of
 downtown New York.

     A rise in sea level would inundate wetlands and
 lowlands,   accelerate  coastal erosion,  exacerbate
 coastal flooding, threaten coastal structures, raise
 water  tables, and  increase the salinity  of rivers,
 bays, and aquifers (Earth and Titus, 1984). Most of
 the wetlands and lowlands are found along the gulf
 coast  and along the Atlantic coast south of central
 New Jersey, although a large area also exists around
 San Francisco Bay. Similarly, the areas vulnerable
 to erosion and  flooding are also predominately in
 the Southeast; potential salinity problems are spread
 more evenly along the U.S. Atlantic coast. We now
 discuss some of the impacts that would result if no
 responses were  initiated to address sea level rise.

 Destruction of Coastal Wetlands

     Coastal wetlands are generally found between
 the highest tide of the year  and mean sea level.
 Wetlands have kept pace with the past rate of sea
 level rise because they collect sediment and produce
 peat upon which they  can build; meanwhile, they
 expanded inland as lowlands were inundated (Figure
 7-3). Wetlands accrete vertically and expand inland.
 Thus,  as Figure 7-3 illustrates, the  present area of
 wetlands is generally far greater than the area that
 would be available for new wetlands as sea level
 rises (Titus  et  al., 1984b;   Titus,  1986).   The
 potential loss would be the greatest in Louisiana
 (see Chapter 16: Southeast).

    In many areas, people have built bulkheads just
 above the marsh. If sea level rises, the wetlands will
 be squeezed between the sea and the bulkheads (see
 Figure 7-3). Previous studies have estimated that if
 the development in coastal areas were removed to
 allow new wetlands to form inland, a 1.5- to 2-meter
rise would  destroy  30 to 70%  of the U.S. coastal
wetlands.  If levees  and bulkheads were erected to
protect today's dryland, the loss could be 50 to 80%
(Titus, 1988; Armentano et al., 1988).

Chapter 7
                       5000 YEARS AGO
                                                    SEDIMENTATION AND
                                                    PEAT FORMATION
                                       3— SEA LEVEL
                                                              COMPLETE WETLAND LOSS WHERE HOUSE IS PROTECTED
                                                              IN RESPONSE TO RISE IN SEA LEVEL .
                                                                                . SEA LEVEL
                                                                                - CURRENT
                                                                                 SEA LEVEL
 Figure 7-3.  Evolution of marsh as sea rises. Coastal marshes have kept pace with the slow rate of sea level rise
that has characterized the last several thousand years. Thus, the area of marsh has expanded over time as new
lands have been inundated.  If in the future, sea level rises faster than the ability of the marsh to keep pace, the
marsh area will contract.  Construction of bulkheads to protect economic development may prevent new marsh
from forming and result in a total loss of marsh in some areas.
Such a loss would reduce the available habitat for
birds  and juvenile fish and  would reduce  the
production of organic materials on which estuarine
fish rely.

    The  dryland  within 2  meters  of  high tide
includes forests, farms, low parts of some port cities,
cities that sank after they were built and are now
protected with levees, and the bay sides of barrier
islands. The low forests and farms are generally in
the mid-Atlantic and Southeast regions; these would
provide potential areas for new wetland formation.
Major port cities with low  areas include Boston,
New York, Charleston, and Miami. New Orleans is
generally 8  feet  below sea level, and parts of
Galveston, Texas City, and  areas around the  San
Francisco  Bay are  also  well  below  sea level.
Because they are already protected by levees, these
cities are more concerned with flooding  than with

Inundation  and  Erosion of Beaches and  Barrier

    Some of the most important vulnerable areas
are  the  recreational  barrier  islands  and  spits
(peninsulas) of the Atlantic and gulf coasts.  Coastal
                                            barriers are generally long narrow islands and spits
                                            with the ocean on one side and a bay on the other.
                                            Typically, the oceanfront block of an island ranges
                                            from 5 to 10 feet above high tide, and the bay side
                                            is 2 to 3 feet above high  water.  Thus, even a 1-
                                            meter sea level rise would threaten much of this
                                            valuable land with inundation.

                                                Erosion threatens the high part of these islands
                                            and  is generally viewed  as  a  more immediate
                                            problem than the inundation of the bay sides. As
                                            Figure 7-4 shows, a rise in sea level can cause an
                                            ocean beach to retreat considerably more than it
                                            would from the effects of inundation alone. The
                                            visible part of the beach is much steeper than the
                                            underwater portion, which comprises most  of the
                                            active  "surf zone."    While  inundation alone  is
                                            determined by the slope of the land just above the
                                            water, Bruun (1962) and others have shown that the
                                            total shoreline retreat from a sea level rise depends
                                            on the average slope of the entire beach profile.

                                                Previous studies suggest that a 1-foot rise in sea
                                            level would generally cause beaches to erode 50 to
                                            100 feet from the  Northeast to Maryland (e.g.,
                                            Kyper and Sorensen, 1985; Everts,  1985); 200 feet
                                            along the Carolinas (Kana et al., 1984); 100 to 1,000

                                                                                        Sea Level Rise
Figure 7-4.  The Bruun Rule: (A) initial condition;
(B) immediate inundation when sea level rises; (C)
subsequent erosion due to sea level rise.  A rise in
sea level  immediately results  in shoreline retreat
due to inundation, shown in the first two examples.
However, a 1-meter rise in sea level implies that
the offshore bottom must also rise 1 meter.  The
sand  required to  raise the bottom  (X1)  can be
supplied by beach nourishment.  Otherwise, waves
will erode the necessary sand (X) from upper part
of the beach as shown in (C).
feet along the Florida coast (Bruun, 1962); 200 to
400 feet along the California coast (Wilcoxen, 1986);
and perhaps several miles in Louisiana.  Because
most  U.S. recreational beaches are  less  than 100
feet wide at  high tide, even a 1-foot rise in sea level
would  require  a  response.    In  many areas,
undeveloped barrier  islands  could keep up  with
rising sea level by  "overwashing" landward.   In
Louisiana, however, barrier islands are breaking up
and exposing the  wetlands behind  them to gulf
waves; consequently, the Louisiana barrier islands
have rapidly eroded.


    If sea level rises, flooding would increase along
the coast for four reasons: (1) A higher sea  level
provides a higher base for storm surges  to build
 upon.  A 1-meter sea level rise would enable a
 15-year storm to flood many areas that today are
 flooded only by a 100-year storm (e.g., Kana et al.,
 1984; Leatherman, 1984).  (2)  Beach erosion also
 would leave oceanfront properties more vulnerable
 to  storm waves.  (3) Higher  water  levels would
 reduce coastal drainage and thus would increase
 flooding attributable  to rainstorms.  In artificially
 drained areas such as New Orleans, the increased
 need for pumping could exceed current capacities.
 (4) Finally, a rise in sea level would raise water
 tables and would flood basements, and  in cases
 where the groundwater is just below the surface,
 perhaps raise it  above the surface.

 Saltwater Intrusion

    A rise in sea level would  enable saltwater to
 penetrate farther inland and upstream into rivers,
 bays, wetlands,  and  aquifers.   Salinity increases
 would be harmful to  some  aquatic plants  and
 animals, and would threaten human uses of water.
 For example, increased salinity already has been
 cited as a factor contributing to reduced  oyster
 harvests in the Delaware and Chesapeake Bays, and
 to conversion of cypress swamps to open lakes in
 Louisiana. Moreover, New York, Philadelphia, and
 much of  California's Central Valley obtain their
 water from areas located just upstream from areas
 where the water is salty during  droughts. Farmers
 in central New Jersey and the city of Camden rely
 on  the Potomac-Raritan-Magothy Aquifer,  which
 could become salty if sea level rises (Hull and Titus,
 1986).  The South Florida  Water  Management
 District already spends millions of dollars every year
 to prevent Miami's Biscayne Aquifer from becoming
 contaminated with seawater.


    The possible responses to inundation, erosion,
 and flooding fall broadly into three categories:
 erecting walls to hold back the sea, allowing the sea
 to advance and adapting to the advance, and raising
 the land.  Both the slow rise in sea level over the
 last thousand years and the areas where land has
been sinking more rapidly offer numerous historical
 examples of all three responses.

    For over  five centuries, the Dutch and others
have used dikes and windmills to prevent inundation
from the North Sea.  By contrast, many cities have
been rebuilt landward as structures have eroded; the

Chapter 7
town of Dunwich, England, has rebuilt its church
seven times  in  the  last seven centuries.  More
recently,  rapidly  subsiding  communities  (e.g.,
Galveston, Texas) have used fill  to  raise land
elevations; the U.S. Army Corps of Engineers and
coastal states regularly pump sand from offshore
locations to counteract beach erosion.  Venice, a
hybrid of all three responses, has allowed the sea to
advance into the canals, has raised some lowlands,
and has erected  storm protection barriers.

    Most assessments  in  the  United States have
concluded that  low-lying  coastal cities would be
protected with  bulkheads,  levees,  and  pumping
systems,  and that sparsely developed areas would
adapt to a naturally retreating shoreline (e.g., Dean
et al.,  1987;  Gibbs,  1984;  Schelling, 1983).  This
conclusion has generally been based on estimates
that  the cost of  structural protection would be far
less  than the value of the  urban areas being
protected but would be greater than the value of
undeveloped land.

    Studies  on the possible responses of barrier
islands  and     moderately developed  mainland
communities show less agreement but generally
suggest that environmental factors  would be as
important as economics. Some have suggested that
barrier islands should use seawalls and other "hard"
engineering approaches (e.g., Kyper and Sorensen,
1985; Sorensen et al., 1984). Others have pointed to
the esthetic problems associated with losing beaches
and have advocated a gradual retreat from the shore
(Howard et  al., 1985).  Noting that new houses on
barrier islands are generally  elevated  on pilings,
Titus (1986) suggested that communities could hold
back the sea but  keep a natural beach by extending
the current practice of pumping sand onto beaches
to raising entire  islands in place.

    Responses to erosion are  more likely to have
adverse  environmental impacts  along  sheltered
water  than   on  the  open coast (Titus,  1986).
Because  the  beach generally is a barrier island's
most important  asset,  economics would tend to
encourage these communities to preserve  their
natural shorelines; actions  that would prevent the
island from  breaking up  also would protect the
adjacent wetlands. However, along most mainland
shorelines, economic self-interest would encourage
property owners to erect  bulkheads; these would
prevent new wetland  formation from offsetting the
loss of wetlands  that  were inundated.
    Most  of  the  measures  for  counteracting
saltwater intrusion attributable to sea level rise have
also been employed to address current problems.
For example, the Delaware River Basin Commission
protects  Philadelphia's freshwater intake  on the
river and New Jersey aquifers recharged by the river
by storing water in reservoirs during the wet season
and releasing it during droughts, thereby forcing the
saltwater back toward the sea.  Other communities
have  protected  coastal  aquifers   by  erecting
underground barriers and by maintaining freshwater
pressure through the use of impoundments  and
injection wells.

    The studies referenced in the previous section
have illustrated a wide variety of possible effects
from and responses to a rise in sea level from the
greenhouse effect.   Although they have identified
the implications of the risk of sea level rise for
specific locations and decisions, these studies have
not estimated the  nationwide magnitude of the
impacts.  This report seeks to fill that void.

    It was not possible to estimate the nationwide
value of every impact of sea level rise. The studies
thus far conducted suggest that the majority of the
environmental  and  economic   costs  would  be
associated with shoreline retreat and measures to
hold back  the sea, which  can be  more easily
assessed  on  a  nationwide  basis.  Because  the
eventual impact will depend on what people actually
do, a  number of  important questions can  be
addressed within this context:

    •   Would  a   gradual   abandonment   of
       moderately  developed  mainland  areas
       significantly increase  the  amount  of
       wetlands that survived a rise in sea level?

    •   Would the concave profiles of coastal areas
       ensure that more wetlands would be lost
       than   gained,   regardless  of   land-use

    •   Should barrier islands be raised in place by
       pumping sand and elevating structures and

                                                                          Sea Level Rise
                                        Decision to Use
                                         Island Raising

               Figure 7-5. Overview of sea level rise studies and authors.
Would a landward migration of developed
barrier islands or encircling them with dikes
and levees be feasible alternatives?

How much property would be lost if barrier
islands were abandoned?

   A central theme underlying these questions is
that  the   implications of sea  level  rise for a
community depend greatly on  whether people
adjust to the natural impact of shoreline retreat or

 Chapter 7
 undertake efforts to hold back the sea. Because no
 one knows  the  extent to which each of these
 approaches  would  be  applied,  this study  was
 designed to estimate the impacts of sea level rise for
 (1) holding back the sea, and (2) natural shoreline

    The tasks were split into five discrete projects:

    1.  Park et al. estimated  the  loss of coastal
        wetlands and dryland.

    2.  Leatherman estimated the cost of pumping
        sand onto open coastal beaches and barrier

    3.  Weggel  et al. estimated  the  cost  of
        protecting sheltered shores with levees and

    4.  Yohe   began   a   national   economic
        assessment  by estimating the  value  of
        threatened property.

    5.  Titus and Greene synthesized the results of
        other studies  to  estimate  ranges  of the
        nationwide impacts.

    Figure 7-5 illustrates the relationships between
 the various reports. (All of the sea level rise studies
 are in Volume B of the Appendices to this report.)
As the top portion shows, the  assessment began
with a case study of Long Beach Island, New Jersey,
which was necessary for evaluating methods and
providing data for purposes of extrapolation.  The
Park and Leatherman studies performed the same
calculations for the case study site that they would
subsequently perform  for  the  other  sites  in the
nationwide analysis. However, Weggel and Yohe
conducted more detailed assessments of the case
study whose  results were used in the Leatherman
and Titus studies.

    Because   it  would  not   be  feasible  for
Leatherman to examine more than one option for
the cost  of  protecting the open coast, Weggel
estimated the cost of protecting Long Beach Island
by three approaches: (1) raising the island in place;
(2) gradually rebuilding the island landward; and (3)
encircling the island with dikes and levees.  Yohe
estimated the value of threatened structures. Titus
analyzed Weggel's and Yohe's results and concluded
that raising  barrier islands would be  the most
reasonable option for the Leatherman study and
 noted that the  cost  of  this option  would  be
 considerably less than the resources that would be
 lost if the islands were not protected as shown in
 Figure 7-6.

    Once  the  case study was  complete,  Park,
 Leatherman, and Weggel proceeded independently
 with   their  studies   (although   Park  provided
 Weggel with elevation data).  When those studies
 were complete,  Titus  synthesized  their results,
 developing a nationwide  estimate  of the cost of
 holding back the sea  and interpolating Weggel's
 200-centimeter  results for the  50-  and  100-
 centimeter scenarios.

    In presenting results from the Park and Weggel
 studies, the sites were grouped into seven coastal
 regions, four of which are in the  Southeast:  New
 England,   mid-Atlantic,   south  Atlantic,   south
 Florida/gulf coast peninsula, Louisiana, other gulf
 (Texas, Mississippi, Alabama, Florida Panhandle),
 and the Pacific coast.  Figure 7-7 illustrates these
                                       ™ Lost Rent
                                       - From Not
                                   _ Raising the Island
                                  Costs of Raising
               2000    8020   2040
Figure 7-6.  Annual cost of raising island versus
annual costs (lost rent)  from not  protecting  the
island (in 1986 dollars) (Titus and Greene, Volume
   Although the researchers considered a variety of
scenarios of future sea level rise, this report focuses

                                                                                      Sea Level Rise
                 WEST COAST
                                                              ,'   SOUTHERN
                                                                 WEST FLORIDA
                             Figure 7-7.  Coastal regions used in this study.
on the impacts of three scenarios: rises of 50, 100,
and 200 centimeters by the year 2100. All three of
these scenarios are based on quantitative estimates
of sea level rise. No probabilities were associated
with these scenarios.  Following the convention of a
recent National Research Council report (Dean et
al., 1987), the rise was interpolated throughout the
21st century using a quadratic (parabola). For each
site,  local subsidence  was  added  to  determine
relative sea level  rise.    Figure 7-8 shows  the
scenarios for the coast of Florida where relative sea
level rise will be typical of most of the U.S. coast.
Sea level would rise 1 foot by 2025, 2040, and 2060
for the three scenarios and 2 feet by 2045, 2065, and

Loss of Coastal Wetlands and Dryland

    Park (Volume B) sought to test a number of
hypotheses  presented in previous   publications:
    •    A rise in sea level greater than the rate of
        vertical  wetland accretion would result in
        a net loss of coastal wetlands.

    •    The loss of wetlands would be greatest if all
        developed areas were  protected,  less  if
        shorelines retreated naturally, and least if
        barrier  islands  were   protected  while
        mainland shores retreated naturally.

    •    The loss of coastal  wetlands would  be
        greatest  in  the  Southeast,  particularly

Study Design

    Park's study was  based on a sample  of  46
coastal sites  that were selected at regular intervals.
This guaranteed that particular regions would  be
represented in proportion to their total area in the
coastal zone. The sites chosen accounted for 10%
of the U.S. coastal zone  excluding Alaska and
Hawaii. To  estimate the potential loss of wet and
dry land, Park first  had to  characterize their
elevations. For wetlands, he used satellite imagery
to determine plant species for 60- by 80-meter
parcels. Using estimates from the literature on the

 Chapter 7
       19M  2000
  Figure 7-8. Sea level scenarios (Miami Beach).
frequency of flooding that can be tolerated  by
various  wetland  plants,  Park  determined  the
percentage  of time that particular  parcels  are
currently under water.  From this, Park inferred
wetland elevation based on the known tidal range.
For dryland, he used spot elevation measurements
to interpolate between contours on U.S. Geological
Survey topographic maps.

    Park  estimated  the net loss of wetlands and
dryland for no protection, protection of developed
areas, and protection of  all shores.   For  the
no-protection  scenario,  estimating  the loss  of
dryland is straightforward. However, for calculating
net wetland loss, Park had to estimate the loss of
existing wetlands as well as the creation of new
wetlands.    For  calculating losses,  Park  used
published vertical accretion rates (see Armentano et
al.,   1988),  although   he  allowed   for   some
acceleration of vertical accretion  in areas with
ample supplies of sediment, such  as  tidal deltas.
Park  assumed  that dryland would  convert  to
wetlands within 5 years of being inundated.

    For sites in the Southeast, Park also allowed for
the  gradual  replacement  of  salt  marshes  by
mangrove swamps.  The upper limit for mangroves
is  around Fort Lauderdale.  Park  used the  GISS
transient scenario to determine the year particular
sites would be as warm as Fort Lauderdale is  today
 and assumed that mangroves would begin to replace
 marsh after that year.


    The greatest uncertainty in Park's analysis is a
 poor understanding of the potential rates of vertical
 accretion.  Although this could substantially affect
 the  results for low  sea  level rise  scenarios, the
 practical significance is small for a rise of 1 meter
 because it is generally recognized  that  wetlands
 could  not keep pace  with the  rise  of 1  to  2
 centimeters per year that such a scenario implies for
 the second half of the 21st century.

    Errors  can  be  made  when  determining
 vegetation type  based  on  the  use  of  infrared
 "signatures" that satellites receive. Park noted, for
 example, that in California  the redwoods have  a
 signature similar to that of marsh grass. For only a
 few sites, Park was able to corroborate his estimates
 of vegetation type.

    Park's study did not consider  the  potential
 implications  of alternative methods of managing
 riverflow.  This limitation  is particularly serious
 regarding application to  Louisiana,  where widely
 varying measures have  been proposed  to increase
 the amount of water and sediment delivered to the
 wetlands.  Finally, the study makes  no attempt to
 predict  which   undeveloped  areas   might  be
 developed in the next century.

   At the coarse (500-meter) scale Park used, the
 assumption of  protecting only  developed  areas
 amounts to not protecting a number of mainland
 areas where  the shoreline is developed but  areas
 behind  the shoreline are not.   Therefore, Park's
 estimates for protecting developed areas should be
 interpreted  as  applying to the case  where  only
 densely  developed  areas  are protected.   Finally,
 Park's assumption that dryland would convert to
 vegetated  wetlands  within  5   years  of being
 inundated probably led him  to underestimate the
 net loss of wetlands due to sea level  rise.
    Park's   results  supported   the  hypotheses
suggested by previous studies.  Figure 7-9 shows
nationwide wetlands loss for various (0- to 3-meter)
sea  level  rises  for  the three  policy  options
investigated.  For a 1-meter rise, 66% of all coastal
wetlands would  be lost if  all shorelines  were
protected,  49% would be lost if  only developed

                                                                                         Sea Level Rise
areas were  protected, and  46% would be lost if
shorelines retreated naturally.

    As  expected, the  greatest losses  of wetlands
would be hi the Southeast, which currently contains
85% of U.S. coastal wetlands (Figure 7-9). For a 1-
meter sea level rise, 6,000  to 8,600 square miles
(depending on which policy is implemented) of U.S.
wetlands would be lost; 90 to 95% of this area
would be in the Southeast, and 40 to 50% would be
in Louisiana alone.    By  contrast,  neither  the
Northeast nor the West would lose more than 10%
of its wetlands if only currently developed areas are

          0.0  0.1 0.3  0.6  1.0 1.5  2.2 3.0
               SEA LEVEL RISE {Meters)

          0.0 0.1 0.3  0.6  1.0 1.5  2.2 3.0
               SEA LEVEL RISE (Motors)
                NO PROTECTION
          0.0  0.1  0.3  0.6  1.0 1.5  2.2 3.0

               SEA LEVEL RISE (Meters)
Figure 7-9.  Nationwide wetlands loss  for  three
shoreline-protection options. Note: These wetlands
include beaches and  flats that  are  not  vegetated
wetlands; however, results cited in the text refer to
vegetated wetlands (Park, Volume B).
    Figure 7-10 illustrates Park's estimates of the
inundation of dryland for the seven coastal regions.
If shorelines retreated naturally, a  1-meter rise
would inundate 7,700 square miles of dryland, an
area the size of Massachusetts.  Rises of 50 and 200
centimeters would result in losses  of 5,000 and
12,000 square miles, respectively. Approximately
70%  of the  dryland losses would  occur  in  the
Southeast, particularly Florida, Louisiana, and North
Carolina. The eastern shores of the Chesapeake and
Delaware  Bays  also  would  lose   considerable

Costs of Defending Sheltered Shorelines

Study Design

    This study began  by examining  Long  Beach
Island in depth. This site and five other sites were
used to develop engineering rules of thumb for the
cost of protecting coastal lowlands from inundation.
Examining the costs  of raising barrier islands
required an assessment of  two alternatives:   (1)
building a levee around the island; and (2) allowing
the island to migrate landward.

    After  visiting Long Beach  Island   and  the
adjacent mainland, Weggel  (Volume B)  designed
and estimated  costs for an  encirclement scheme
consisting  of  a levee  around the island  and a
drainage  system  that  included pumping  and
underground retention of stormwater. For island
migration, he used  the Bruun Rule to  estimate
oceanside erosion and navigation charts to calculate
the amount of sand necessary to fill the bay an
equivalent distance landward. For island raising and
island  migration,  Weggel used the  literature to
estimate the costs of elevating and moving houses
and of rebuilding roads and utilities.

    Weggel's approach for estimating the nationwide
costs was to  examine a number of index sites in
depth   and  thereby  develop  generalized  cost
estimates for protecting different types of shorelines.
He used the topographic information collected by
Park for a sample of 95 sites to determine the area
and shoreline length that had to be protected.  He
then applied the cost estimation factors to each site
and extrapolated the sample to the entire coast.

    After  assessing  Long Beach Island, Weggel
conducted less  detailed studies  of the  following
areas:  metropolitan  New York; Dividing Creek,

 Chapter 7
                            A.  DRYLAND LOSS BY 2100 WITHOUT SHORE PROTECTION
                              Northeast    Mtd-
                                                             Louisiana   Other Gulf
                        g  1.0
                              Northeast    Mid-     South
                                      Atlantic    Atlantic
                            SEA LEVEL RISE I "  I BASELINE
                            SCENARIO:    I	J BASELINE
Figure 7-10.  Loss of dryland by 2100: (A) if no areas are protected, and (B) if developed areas are protected
with levees (derived from Park, Volume B; see also Titus and Greene, Volume B).
New  Jersey; Miami  and Miami Beach; the area
around Corpus Christi, Texas; and parts  of San
Francisco Bay.


    The most serious limitation of the Weggel study
is  that cruder  methods are used for the national
assessment than for the index sites.  Even  for the
index sites, the cost estimates are based  on the
literature, not on site-specific designs that take into
consideration wave data for bulkheads and potential
savings from tolerating substandard roads. Weggel
did not estimate the cost of pumping rainwater out
of areas protected by levees.
              Finally, Weggel was able to examine only one
          scenario: a 2-meter rise by 2100. This scenario was
          chosen  over  the  more  likely  1-meter  scenario
          because an interpolation from 2 meters to 1 meter
          would be more reliable than an extrapolation from
          1 meter to 2 meters.  (See the discussion of Titus
          and Greene for results of the interpolation.)


          Case Study of Long Beach Island

              Weggel's  cumulative  cost  estimates  clearly
          indicate that raising Long Beach Island would be
          much less expensive ($1.7 billion) than allowing it to

                                                                                       Sea Level Rise
migrate landward ($7.7 billion). Although the cost
of building a levee around the island ($800 million)
would be less, the "present value" would be greater.
Weggel concluded that the levee would have to be
built  in the  2020s, whereas the island  could be
raised gradually between 2020 and 2100. Thus, the
(discounted) present value of the levee cost would
be greater, and raising the necessary capital for a
levee at any one time could be more difficult than
gradually rebuilding the roads and elevating houses
as the island was raised.  Moreover, a levee would
eliminate the waterfront view. A final disadvantage
of building a levee is that one must design for a
specific magnitude of sea level rise; by contrast, an
island could be raised incrementally.

    The Weggel  analysis  shows that  landward
migration is more  expensive  than  island raising,
primarily  because   of  the increased  costs  of
rebuilding infrastructure. Thus, migration might be
less expensive in the case of a very lightly developed
island.  Levees might be  more practical for wide
barrier  islands  where most people do  not have a
waterfront view.
                  Nationwide Costs

                     Table  7-2  shows WeggePs  estimates  for the
                  index sites and his nationwide estimate. The index
                  sites represent two distinct patterns. Because urban
                  areas such as New  York  and Miami would be
                  entirely protected by levees, the cost of moving
                  buildings and rebuilding roads and utilities would be
                  relatively  small.   On the  other  hand,  Weggel
                  concluded that in more rural areas such as Dividing
                  Creek, New Jersey, only the pockets of development
                  would be protected. The roads that connected them
                  would have to be elevated or replaced with bridges,
                  and the small number of isolated buildings would
                  have to be moved.

                     Weggel estimates that the nationwide cost  of
                  protecting   developed shorelines  would  be $25
                  billion, assuming bulkheads are built, and $80 billion
                  assuming levees are built. Unlike wetlands loss, the
                  cost  of  protecting developed areas from  the sea
                  would be concentrated more in the Northeast than
                  in the Southeast because a much greater portion of
                  the southeastern coast is undeveloped.
Table 7-1.  Total Cost of Protecting Long Beach Island from a 2-Meter Rise in Sea Level (millions of 1986
     Protective measure
Sand costs:


   Land creation/maintenance

Moving/elevating houses


Levee and drainage














 NA = Not applicable.
 Source:  Leatherman (Volume B); Weggel (Volume B)

 Chapter 7
 Table 7-2.  Cumulative Cost of Protecting Sheltered Waters for a 2-Meter Rise in Sea Level (millions of 1986
Raise old
 Index sites

  New York
  Long Beach Island
  Dividing Creek
  Miami area
  Corpus Christi
  San Francisco Bay3

 Nationwide estimate
.1 	 o —
  Site names refer to the name of U.S. Geological Survey quadrant, not to the geographical area of the same
 Source:  Weggel et al. (Volume B).
 Case Study of the Value  of Threatened
 Coastal Property

 Study Design

    Yohe's (Volume B) objective was to estimate
 the loss of property  that would result from not
 holding back the sea.  Using estimates of erosion
 and   inundation  for  Long Beach  Island  from
 Leatherman and Park et  al.,  Yohe determined
 which land would be lost from sea level rise for a
 sample of strips spanning the island from the ocean
 to the bay. He then used the Ocean County, New
 Jersey, tax assessor's estimates of the value of the
 land and  structures that would be lost, assuming
 that the premium associated with a view of the bay
 or ocean would be transferred to another property
 owner and not lost to the community. He estimated
 the annual stream of  rents that  would be lost by
 assuming that  the required return on real estate is
 10% after tax.  Yohe  assumed that a property on
 the bay  side was "lost" whenever it was flooded at
high tide, and that property on the ocean side was
                  "lost" when the house was within 40 feet  of the
                  spring high tide  mark.  (See Titus and Greene,
                  Volume B, for discussion.)


                     Yohe's results for a sea level rise of less than 18
                  inches are sensitive  to  the assumption regarding
                  when a property  would be lost.  On the bay side,
                  people  might learn  to  tolerate tidal inundation.
                  Unless  a major  storm occurred,  people  'could
                  probably occupy oceanfront houses until they were
                  flooded at high tide.  However, the resulting loss of
                  recreational use of the beach probably would have
                  a  greater impact than abandoning the  structure.
                  Tax maps  do  not  always  provide  up-to-date
                  estimates  of  property values.    However,  the
                  distinction between the tax assessor's most recent
                  estimate of market value and the current market
                  value is small compared with the possible changes
                  in property  values that will occur over the next
                  century; hence, Titus and Greene used tax assessors
                  estimates of market values.

                                                                                       Sea Level Rise

    Yohe's results suggest that the cost of gradually
raising  Long Beach Island would be far less than
the value of the resources that would be protected.
Figure 7-6 compares Yohe's estimates of the annual
loss in rents resulting from not holding back the sea
with Weggel's estimates of the annual cost of raising
the island  for  the 2-meter scenario.   With  the
exception of the 2020s,  the  annual loss in rents
resulting from not holding back the sea would be
far less than the annual costs of pumping sand  and
elevating structures. Titus and  Greene point  out
that the cost would be approximately $1,000  per
year per house,  equivalent to 1 week's rent  (peak

Nationwide Cost of Pumping Sand Onto
Recreational  Beaches

    Leatherman's goal (Volume B) was to estimate
the cost of defending the U.S. ocean coast from a
rise in sea level.

Study Design

    Owing to time constraints, it was  possible to
consider only one technology. Based on the Long
Beach Island results, Leatherman assumed that the
cost of elevating recreational beaches  and coastal
barrier islands by pumping in offshore  sand would
provide a more representative cost estimate than
assuming that barrier islands would be  abandoned,
would migrate landward,  or would be encircled with
dikes and levees.

    The first step in Leatherman's  analysis was to
estimate the area of (1)  the beach  system, (2) the
low bayside, and (3) the slightly elevated oceanside
of the island.  Given the  areas, the volume of sand
was estimated by assuming  that the beach system
would be raised by the  amount of sea level  rise.
The bay and ocean sides  of the island would not be
raised until after a sea level rise of 1 and 3 feet,
respectively.   Cost estimates for  the  sand were
derived from inventories conducted by the U.S.
Army Corps of Engineers.

    Leatherman  applied   this  method   to   all
recreational beaches from  Delaware  Bay to the
mouth of the Rio  Grande,  as well as   California,
which accounts for 80% of the nation's beaches. He
also examined one representative site in each of the
remaining states.


   Although the  samples of sites hi the  Northeast
and   Northwest  are   representative,  complete
coverage  would   have  been  more  accurate.
Furthermore,   Leatherman   used   conservative
assumptions in estimating the unit costs of sand.
Generally, a fraction of the sand placed on a beach
washes away because the sand's grain is  too small.
Moreover, as dredges have to move farther offshore
to find sand, costs will increase.

   For  Florida,  Leatherman  used  published
estimates of the percentage of fine-grain sand  and
assumed that the dredging  cost would rise $1  per
cubic yard for  every additional mile  offshore  the
dredge had to move.  For the other states, however,
he assumed that the deposits mined would have no
fine-grain sand and that dredging  costs  would  not
increase. (To test the sensitivity of  this assumption,
Titus and  Greene developed  an increasing-cost
scenario.)  Leatherman assumed no  storm worse
than the 1-year storm, which underestimates  the
sand volumes required.

   A final limitation of the  Leatherman study is
that  it  represents the  cost  of applying a single
technology throughout  the ocean coasts  of  the
United  States.   Undoubtedly, some  communities
(particularly  Galveston and other  wide  barrier
islands  in Texas) would find it less  expensive to
erect levees and  seawalls or to accept a  natural
shoreline retreat.


    Table 7-3 illustrates Leatherman's estimates. A
total of 1,900 miles of shoreline would be nourished.
Of 746  square  miles of coastal barrier islands  that
would be raised  for a 4-foot sea level rise,  208
square  miles would be for a 2-foot rise.   As the
table shows, two-thirds of the nationwide costs
would be borne by four southeastern states: Texas,
Louisiana, Florida, and South Carolina.

    Figure 7-11 illustrates the cumulative nationwide
costs over time.  For  the 50- and 200-centimeter
scenarios, the cumulative cost would be $2.3 to  $4.4
billion  through 2020,  $11  to $20 billion through
2060, and $14 to $58 billion through  2100.  By

 Chapter 7

    Table 7-3. Cost of Placing Sand on U.S. Recreational Beaches and Coastal Barrier Islands and Spits
              (millions of 1986 dollars)
                                                      Sea level rise by 2100
   100 cm
New Hampshire3
Rhode Island3
New York3
New Jersey3
North Carolina
South Carolina
(Atlantic coast)
(gulf coast)
Washington State3




Indicates states where estimate was based on extrapolating a representative site to the entke state.  All other
  states have 100% coverage.
 Florida estimates account for the percentage of fine-grain sediment, which generally washes away, and for cost
  escalation as least expensive sand deposits are exhausted.  All other estimates conservatively ignore this
Source: Leatherman (Volume B) (baseline derived from Leatherman).
contrast, if current trends continue, the total cost of
sea level rise for beach nourishment would be about
$35 million per year.

Synthesis of the Three National Studies

Study Design

    Although Weggel used Park's topographic data,
the  analysis  in  the three  nationwide  studies
                            proceeded  independently.   Titus and Greene's
                            primary objectives (Volume B) were to combine
                            various results to estimate the nationwide cost of
                            holding  back the sea  for various sea level rise
                            scenarios and to derive ranges for the  specific
                            impacts.  Their objectives were as follows:

                            1.  Use Park's results to weight WeggePs high and
                               low  scenarios according  to whether levees or
                               bulkheads would be necessary, and interpolate

                                                                                       Sea Level Rise

$ Billons
0 V- 001







200 Cm
100 Cm

1B80 2000 2020 2040 3080 2090 2100
Figure 7-11. Nationwide cost of sand for protecting
ocean coast (in 1986 dollars) (Leatherman, Volume

    Weggel's cost estimate for the 2-meter rise to
    rises of 50 and 100 centimeters;
    Use results from Leatherman  and  Weggel,
    along  with census   data,  to  estimate  the
    nationwide cost (other than pumping  sand) of
    raising barrier islands;

    Develop an increasing-cost scenario for the cost
    of protecting the open ocean coast; and

    Develop statistical  confidence  intervals for
    wetland loss,  impacts of  the various  policy
    options,  and  costs  of protecting  developed
Cost of Protecting Sheltered Shores

    Titus and Greene developed a single estimate
for protecting each site with bulkheads and levees
by assuming that the portion of developed areas
protected with levees would be equal to the portion
of the  lowlands  that Park estimated would be
inundated. They interpolated the resulting 2-meter
estimate to 50- and 100-centimeter estimates, based
on Weggel's assumption that the cost of building
bulkheads and levees  rises as a function of the
structure's height.

Cost of Raising Barrier Islands Other Than Dredging

    Weggel's  case  study of Long Beach Island
provided cost estimates for elevating structures and
rebuilding roads, while Leatherman estimated the
area that would have to be raised. Many barrier
islands have development densities different  from
those of Long Beach Island because they have large
tracts  of undeveloped land or  larger  lot  sizes.
Therefore, Titus and Greene used census data to
estimate a  confidence interval  for  the average
building density of barrier islands, and they applied
Weggel's cost factors.

Sensitivity of Sand Costs to Increasing Scarcity of

    Titus and  Greene used Leathennan's escalating
cost  assumptions for Florida to  estimate  sand
pumping costs for the rest of the nation.

Confidence Intervals

    The Park and Weggel studies involved sampling,
but  the  researchers  did not  calculate  statistical
confidence intervals.  Therefore, Titus and Greene
developed 95% confidence intervals for the cost of
protecting sheltered coasts, the area of wetlands loss
for various scenarios.


    Besides all of the  limitations that apply to the
Park, Leatherman, and Weggel studies, a number of
others apply to Titus and Greene.

Cost of Protecting Sheltered Shores

    Titus and Greene assumed that the portion of
the coast requiring levees (instead of bulkheads)
would be  equal to the portion of lowlands that
otherwise would be inundated.  This assumption
tends to understate  the  need  for  levees.   For
example, a community  that is 75%  high ground
often would still have very low land along all of its
shoreline and hence would require a levee  along
 100% of the shore.  But Titus and Greene assume
that only 25% would be protected by levees.
                                                :  139

 Chapter 7
 Cost of Raising Barrier Islands

     The data provided by Weggel focused only on
 elevating roads, buildings, and bulkheads.  Thus,
 Titus and  Greene do  not consider the cost of
 replacing sewers, water mains, or buried cables. On
 the other hand, Weggel's cost factors assume that
 rebuilt roads would be up to engineering standards;
 it  is  possible  that  communities would tolerate
 substandard  roads.  In addition, the census data
 Titus and  Greene used were only available for
 incorporated communities, many of which are part
 barrier island and part  mainland; thus, the data
 provide  only a rough  measure of typical  road

 Sensitivity of Sand Costs to Increased Scarcity of

     Finally, Titus and Greene made no attempt to
 determine how realistic their assumption was that
                          sand costs would increase  by the same pattern
                          nationwide as they would in Florida.


                          Loss of Wetlands and Dryland

                             Table  7-4 illustrates 95% confidence intervals
                          for the nationwide losses of wetlands and dryland.
                          If all shorelines  were protected,  a  1-meter rise
                          would result in a loss of 50 to 82% of U.S. coastal
                          wetlands, and a 2-meter rise would result in a loss
                          of 66 to  90%.  If only the densely developed areas
                          were protected, the losses would be 29 to 69% and
                          61  to 80% for  the  1- and  2-meter  scenarios,
                          respectively.    Except  for the   Northeast,  no
                          protection results hi only slightly lower wetland loss
                          than  protecting  only  densely  developed  areas.
                          Although  the estimates for the Northeast,  mid-
                          Atlantic, the gulf regions outside Louisiana, and the
                          Florida peninsula are not statistically significant (at
                          the 95% confidence levels), results  suggest  that
                          wetlands loss would be least in the Northeast and
             Table 7-4. Nationwide Loss of Wetlands and Dryland3 (95% confidence intervals)

                                                       'Square milesp
                 50-cm rise
                      100-cm rise
                     200-cm rise

  Total protection


  No protection


  Total protection


  No protection









^Wetlands loss refers to vegetative wetlands only.
 Numbers in parentheses are percentages.
 NC = Not calculated.
Source: Titus and Greene (Volume B).

                                                                                       Sea Level Rise
Table 7-5.  Cumulative Nationwide Cost of Protecting Barrier Islands and Developed Mainland Through the
          Year 2100 (billions of 1986 dollars)3

                                                   Sea level scenario	
50-cm rise
100-cm rise
200-cm rise
Open coast
Raise houses,
roads, utilities
Sheltered shores
aCosts due to sea level rise only.
bRanges for totals are based on the square root of the sum of squared ranges.
Source:  Titus and Greene (Volume B).
Costs of Holding Back the Sea

    Table  7-5  illustrates  the Titus and  Greene
estimates of the costs of holding back the sea. The
low   range for  the  sand  costs  is  based  on
Leatherman's study, and the high range is based on
the  increasing cost  scenario Titus and  Greene
developed. The uncertainty range for the costs of
elevating structures reflects the uncertainty in census
data regarding the current density of development.
High and low estimates for the cost of protecting
sheltered shorelines are based  on the  sampling
errors of the estimates for the 46 sites that both
Park et al. and Weggel et al. examined.

    Titus and Greene estimated that the cumulative
nationwide cost of protecting currently developed
areas in the face of a  1-meter rise would be from
$73 to 111 billion, with costs for  the 50- and 200-
centimeter scenarios  ranging from  $32 to  309
billion.  These costs  would  imply a severalfold
increase in annual expenditures for coastal defense.
Nevertheless,  compared with the value of coastal
property, the costs are small.
Wetlands Protection

    The nationwide analysis showed that a 50- to
200-centimeter rise in  sea level could reduce the
        coastal wetlands acreage (outside Louisiana) by 17
        to 76% if no mainland areas were protected, by 20
        to 80% if only currently  developed areas were
        protected, and by 38 to 90% if all mainland areas
        were protected. These estimates of the areal losses
        understate the differences in impacts for the various
        land-use options. Although a substantial loss would
        occur even with  no protection, most of today's
        wetland shorelines would still have wetlands;  the
        strip simply  would  be  narrower.   By  contrast,
        protecting  all  mainland  areas  would  generally
        replace  natural shorelines with bulkheads  and
        levees.  This distinction is important because for
        many species of fish, the length  of a wetland
        shoreline is more critical than the total area.

        Options for State and Local Governments

            Titus (1986)  examined three approaches  for
        maintaining wetland shorelines in the face of a
        rising sea: (1) no further development in lowlands;
        (2) no action now but a gradual  abandonment of
        lowlands as sea level rises; and (3) allowing future
        development only with a binding agreement to allow
        such  development  to  revert  to nature if it is
        threatened by inundation.

            The first option would  encounter  legal or
        financial hurdles.   The  extent  to which  the
        "due-process" clause of the Constitution would allow
        governments to prevent development in anticipation
        of sea level rise has not been specifically addressed

 Chapter 7
by the courts. Although purchases of land would be
feasible for parks and refuges, the cost of buying
the  majority of lowlands  would be prohibitive.
Moreover, this approach requires preparation for a
rise in sea level of a given magnitude; if and when
the sea rises beyond that point, the wetlands would
be lost.   Finally, preventing future  development
would not solve the loss of wetlands resulting from
areas that have already been developed.

    Enacting no policy today and addressing  the
issue as  sea level rises would avoid the costs of
planning for the wrong amount of sea level rise  but
would probably result in less wetlands protection.
People are  developing  coastal  properly  on  the
assumption that they can use the land indefinitely.
It would be difficult for any level of government to
tell property owners that they must abandon their
land with only a few years' notice, and the cost of
purchasing developed areas would be even greater
than the cost of buying undeveloped areas.

    Economic theory suggests that under the third
alternative, people would develop a properly only if
the temporary use provided benefits greater than
the costs of writing it  off early. This  approach
would result in the  greatest degree  of  flexibility,
because  it  would allow  real  estate markets   to
incorporate sea level rise and to determine the most
efficient use of land as long as it remains dry.

    This  approach  could  be   implemented  by
regulations that prohibit construction of bulkheads
as sea level rises or by the  use of conditional long-
term leases  that expire when high tide falls above a
property's elevation.

    The  State  of Maine (1987) has implemented
this  third  approach  through  its  coastal dune
regulations, which state that people building houses
along the shore should assume that they will have to
move their  houses if their  presence  prevents  the
natural migration of coastal wetlands,  dunes,  or
other natural shorelines. A number of states also
have regulations that discourage bulkheads, although
they do not specifically address sea level rise. The
option can  be implemented through cooperative
arrangements  between  developers,  conservancy
groups, and local governments.  (See  Titus and
Greene, Volume B, for additional details.)
The Federal Role

    Section 404 of the Clean Water Act discourages
development of existing wetlands, but it does not
address development of areas that might one day be
necessary for wetland migration. This program will
provide  lasting  benefits,  even  if  most  coastal
wetlands  are inundated.  Although marshes and
swamps would be inundated, the shallow waters that
formed  could  provide  habitat   for  fish  and
submerged aquatic vegetation. No one has assessed
the need for a federal program to protect wetlands
in the face of rising sea level.

Coastal Protection

State and Local Efforts

    State  and local governments currently decide
which areas would be protected and which would
be  allowed to erode.   Currently,  few localities
contribute more  than 10% of the cost of beach
nourishment, with the states taking on an increasing
share from the federal government. However, many
coastal officials doubt that their states could raise
the necessary funds if global warming increased the
costs of coastal protection over the next century by
$50  to $300 billion.   If state funds could not  be
found, the communities themselves would have to
take  on the necessary expenditures or adapt to

    Long Beach Island, New Jersey, illustrates the
potential difficulties.  The annual cost of raising the
island would average $200 to $1,000 per house over
the next century (Titus and Greene, Volume B).
Although this amount is less than one week's rent
during the summer,  it would  more than double
property taxes, an action that is difficult for local
governments to contemplate. Moreover, the island
is divided into six jurisdictions, all of which would
have to participate.

    More lightly developed communities may decide
that the benefits  of holding back the sea are not
worthwhile.  Sand costs would be much less for an
island that migrated.  Although Weggel estimated
that higher costs would be associated with allowing
Long Beach Island to migrate landward than with
raising the island in place, this conclusion resulted
largely from the cost of rebuilding sewers and other
utilities that would still be useful if the island were

                                                                                         Sea Level Rise
    Regardless of how a barrier island community
intends to respond to sea level rise, the eventual
costs can be reduced by deciding on a response well
in advance.  The cost of raising an island can be
reduced if roads and utilities are routinely elevated
or if they have to be rebuilt for other reasons (e.g.,
Titus  et  al.,  1987).   The cost of  a  landward
migration  also can be  reduced by discouraging
reconstruction of oceanfront houses destroyed by
storms (Titus et al., 1984a). The ability to fund the
required  measures  also  would  be increased by
fostering  the necessary public debate well before
the funds are needed.

Federal Efforts

    While  state   governments   generally  are
responsible for protecting recreational beaches, the
U.S. Army Corps of Engineers is  responsible for
several major federal projects to rebuild beaches
and for efforts to curtail land loss hi Louisiana. The
long-term  success  of  these  efforts  would  be
improved if the corps were authorized to develop
comprehensive long-term  plans  to address the
impacts of sea level rise.

Beach Erosion

    In  its erosion-control  efforts, the corps has
recently shifted its focus from hard structures (e.g.,
seawalls, bulkheads, and groins) to soft approaches,
such as pumping sand onto beaches.  This shift is
consistent with the implications  of sea level rise:
groins and seawalls will not prevent loss of beaches
due to sea level rise. Although more sand will have
to be  pumped than current analyses suggest, this
approach could ensure the survival of the nation's

    Nevertheless, consideration of accelerated sea
level rise would change  the cost-benefit ratios  of
many corps erosion control projects.  As with the
operations of reservoirs (discussed in  Chapter 16:
Southeast), the corps is authorized to consider flood
protection but not recreation. When they evaluated
the  benefits of erosion  control  at Ocean City,
Maryland, the corps concluded that less than 10%
of the benefits would be  for flood control (most
were related to recreation). Had they considered
accelerated sea level rise, however, the estimated
flood   protection   benefits  from  having  a
protective   beach  would  have   constituted  a
considerably higher fraction of the total benefits
(Titus, 1985).

Wetlands Loss in Louisiana

    By preventing freshwater and  sediment from
reaching the coastal wetlands, federal management
of  the  Mississippi   River  is  increasing   the
vulnerability of coastal Louisiana to a sea level rise
(e.g.,   Houck,   1983).    For   example,  current
navigation routes require the U.S. Army Corps of
Engineers to limit the amount of water  flowing
through the Atchafalaya River and to close natural
breaches in the main channel of the Mississippi;
these actions limit the amount  of freshwater and
sediment reaching the wetlands.  Alternative routes
have been proposed that would enable water and
sediment to reach the wetlands (Louisiana Wetland
Protection Panel,  1987).  These include dredging
additional canals parallel to the existing Mississippi
River gulf outlet or constructing a deepwater port
east of the city.

    Either of these options would cost a few billion
dollars. By contrast, annual resources for correcting
land loss in Louisiana have been  in the  tens of
millions of dollars. As a result, mitigation activities
have focused on freshwater diversion structures and
on other strategies that can reduce current wetland
loss attributable to high salinities but that would not
substantially reduce wetlands loss if sea  level rises
50  to   200  centimeters   (Louisiana  Wetland
Protection Panel, 1987).

    The prospect of even a 50-centimeter rise in sea
level suggests that solving the Louisiana wetlands
loss problem is much more urgent than is commonly
assumed.   Because federal  activities are now a
major  cause of land  loss, and would have to be
modified to  enable wetlands to survive a rising sea,
the problem is unlikely  to  be solved without a
congressional mandate. A recent interagency report
concluded   that  "no   one  has   systematically
determined what must be done to save 10, 25, or 50
percent of Louisiana's coastal ecosystem" (Louisiana
Wetland Protection Panel, 1987).  Until someone
estimates the costs and likely results of strategies
with a chance of protecting a significant fraction of
the wetlands in face of rising sea level, it will be
difficult for Congress to devise a long-term solution.

 Chapter 7

Flood Insurance

    In 1968, Congress created the National Flood
 Insurance Program with the objective of reducing
 federal disaster relief resulting from floods.  The
 Federal Emergency Management Agency (FEMA),
which already had responsibility for administering
 disaster relief, was placed in charge of this program
as well.

    The National Flood Insurance Program sought
to  offer  localities  an  incentive   to   prevent
flood-prone construction.  In return for requiring
that any construction in a floodplain be designed to
withstand a 100-year flood, the federal government
would provide  subsidized  insurance  to  existing
homes  and a fair-market  rate  for  any  new
construction (which was itself a benefit, since private
insurers  generally did not offer flood insurance).
Moreover,  as  long  as a  community joined  the
program, it would continue to be eligible for federal
disaster relief; if it did not join, it would no longer
be eligible. As a result of this program, new coastal
houses are generally elevated on pilings.

    Although Congress intended to prevent coastal
disasters, the  National Flood Insurance Act does
not require strategic assessments of long-term issues
(see Riebsame, Volume J). Thus, FEMA has  not
conducted strategic assessments of how the program
could be managed to minimize flood damage from
shoreline retreat caused by both present and future
rates of sea level rise.

    Congress  recently enacted the Upton-Jones
Amendment (Public Housing Act of 1988), which
commits  the  federal  government  to   pay  for
rebuilding or relocating houses that are  about to
erode  into  the  sea.   Although  the cost of this
provision is modest today,  a sea  level rise could
commit the federal government  to purchase  the
houses on all barrier islands that did not choose to
hold back the sea. Furthermore, this commitment
could increase the number  of communities  that
decided not to hold back the sea.

    The  planned  implementation of  actuarially
sound insurance rates would ensure that as sea level
rise increased property risk, insurance  rates would
rise to  reflect  the risk.  This would discourage
construction of vulnerable houses, unless their value
was great enough to outweigh the likely damages
from floods.  However, statutes limiting the rate at
which flood insurance rates can increase could keep
rates from rising as rapidly as the risk of flooding,
thereby increasing the federal subsidy.

    No assessment of the impacts of sea level rise
on the federal flood insurance program has been

Sewers and Drains

    Sea level rise also would have important impacts
on coastal sewage and drainage systems. Wilcoxen
(1986) examined the  implications of the failure to
consider accelerated sea level rise in the design of
San Francisco's West Side (sewerage) Transport,
which is  a large, steel-reiriforced  concrete  box
buried under the city's ocean beach.  He found that
beach erosion will gradually expose the transport to
the  ocean,  leaving  the  system  vulnerable   to
undermining and eventual collapse. Protection costs
for the $100 million project would likely amount to
an additional $70 million. Wilcoxen concludes that
had sea level  rise been considered,  the  project
probably would have been sited elsewhere.

    The impacts of sea level rise on the construction
grants program probably would be less in  most
other cases.  As sea level rises, larger pumps will
be necessary to transport effluents from  settling
ponds to the adjacent body of water.  However, sea
level rise would not necessarily require alternative
siting. The projects serving barrier islands often are
located on the  mainland, and projects located on
barrier islands  are generally   elevated well above
flood levels.    If  barrier islands  are raised    in
response to sea level as the nationwide analysis
suggests, sewerage treatment plants will be a small
part of the infrastructure that has to be modified.

    Engineering assessments have concluded that it
is  already cost-beneficial to consider sea level rise
in the construction of coastal drainage systems  in
urban areas.   For  example,  the extra cost  of
installing the larger pipes necessary to accommodate
a 1-foot rise in sea level would add less than 10% to
the  cost  of  rebuilding  a  drainage   system   in
Charleston,  South Carolina;  however, failure to
consider  sea  level rise would require  premature
rebuilding of the $4  million  system (Titus et al.,

                                                                                      Sea Level Rise

    A  much  better  understanding  of  erosion
processes is needed to (1) understand how much
erosion will take place if no action is taken; and (2)
help identify the most cost-effective means for
protecting   sandy   shores.      An  improved
understanding of how wetland accretion responds
to   different   temperatures,    higher   CO2
concentrations,  changing mineral content, and the
drowning of adjacent wetlands is needed. This will
refine our ability to project future wetlands loss and,
perhaps, devise measures for artificially enhancing
their vertical growth.

    This report did not  examine the impacts  of
increased flooding because flood models have not
been applied to the large numbers of coastal sites
that would be necessary to conduct a nationwide
assessment.   Time-dependent  estuarine salinity
models, such as that of the Delaware River Basin
Commission, should be applied to major estuaries
to examine  impacts on ecosystems and  drinking
water supplies.

    Assessments of the impacts of global warming
on coastal environments would be greatly improved
by better  estimates of future sea level rise.  In
addition to the  improved  ocean modeling that will
be necessary for better projections of surface air
temperatures (see Chapter 2: Climate Change), this
will  also  require  a substantial  increase  in the
resources allocated for monitoring and modeling
glacial processes. Finally, this report assumed that
winds, waves, and storms remained constant; future
studies will need estimates of the changes in these
climatic variables.

Armentano, T.V.,  R.A. Park,  and C.L. Cloonan.
1988. Impacts on coastal wetlands throughout the
United States.  In: Titus, J.G. ed.   Greenhouse
Effect, Sea Level  Rise, and  Coastal Wetlands.
Washington, DC:  U.S. Environmental Protection

Barnett, T.P. 1984. The estimation of "global" sea
level change: a problem of uniqueness. Journal of
Geophysical Research  89(C5):7980-7988.
Earth,   M.C.,  and  J.G.  Titus,  eds.    1984.
Greenhouse  Effect  and  Sea Level  Rise:    A
Challenge for This Generation.  New York:  Van
Nostrand Reinhold Company.

Bentley, C.R.   1983.  West Antarctic ice sheets:
diagnosis and prognosis.   In: Proceedings of the
Carbon Dioxide Research Conference:   Carbon
Dioxide,  Science,  and  Consensus.  Conference
820970. Washington, DC:  Department of Energy.

Bruun, P.  1962. Sea level rise as a cause of shore
erosion.   Journal of  Waterways and Harbors
Division (ASCE) 1:116-130.

Dean, R.G. et al. 1987.  Responding to Changes in
Sea Level.   Washington, DC: National Academy

Dony, W.L., W.R. Farrand, and M. Ewing.  1962.
Pleistocene ice volumes  and sea level  lowering.
Journal of Geology 70:206-214.

Everts, C.H.  1985. Effect of sea level rise and net
sand volume change on shoreline position at Ocean
City, Maryland.  In: Titus, J.G.,  ed.   Potential
Impact of Sea Level Rise  on the Beach at Ocean
City,   Maryland.     Washington,  DC:    U.S.
Environmental Protection Agency.

Fairbridge, R.W., and W.S. Krebs, Jr.  1962. Sea
level and  the southern oscillation.   Geophysical
Journal 6:532-545.

Gibbs, M.  1984. Economic analysis of sea level
rise:  methods  and results.  In: Barth,  M.C., and
J.G. Titus, eds.  Greenhouse Effect and Sea Level
Rise: A Challenge for This Generation. New York:
Van Nostrand Reinhold Company.

Gornitz, V.S., S. Lebedeff, and J. Hansen.  1982.
Global sea level trend in the past century. Science

Hoffman,  J.S.,  D.  Keyes,  and J.G. Titus.  1983.
Projecting Future Sea Level Rise. Washington, DC:
U.S. Environmental Protection Agency.

Hoffman,  J.S.,  J. Wells, and J.G. Titus.  1986.
Future global warming and  sea  level  rise.   In:
Sigbjarnarson, G., ed.  Iceland Coastal  and River
Symposium.  Reykjavik, Iceland:  National Energy

Chapter 7
Houck, O A. 1983. Land loss in coastal Louisiana-
causes, consequences, and remedies. Tulane Law
Review 58(1):3-168.

Howard, J.D., O.H. Pilkey, and A. Kaufman. 1985.
Strategy for beach preservation proposed. Geotimes

Hughes,  T.  1983.   The  stability of the West
Antarctic ice sheet: what has happened and what
will  happen.  In:   Proceedings of the  Carbon
Dioxide Research  Conference:   Carbon Dioxide,
Science,  and  Consensus.  Conference  820970.
Washington, DC: Department of Energy.

Hull, C.HJ., and J.G. Titus.  1986.  Responses to
salinity increases. In:  Hull, C.H J., and J.G. Titus.
1986.  Greenhouse Effect,  Sea Level Rise,  and
Salinity in the Delaware Estuary. Washington, DC:
U.S.   Environmental  Protection  Agency   and
Delaware River Basin Commission.

Hull, C.HJ., and J.G. Titus.  1986.  Greenhouse
Effect, Sea Level Rise, and Salinity in the Delaware
Estuary.  Washington, DC:   U.S. Environmental
Protection  Agency and Delaware River  Basin

Kana, T.W., J.  Michel, M.O. Hayes, and J.R.
Jensen. 1984. The physical impact of sea level rise
in the area of Charleston,  South Carolina  In:
Barth, M.C., and  J.G. Titus, eds.   Greenhouse
Effect and  Sea Level Rise:   A Challenge for This
Generation.  New York:  Van Nostrand Reinhold

Kennett,  J. 1982.  Marine  Geology.  Englewood
Cliffs, NJ:  Prentice-Hall.

Kyper, T.,  and  R. Sorenson.   1985.  Potential
impacts of  sea level rise on the beach and coastal
structures at Sea Bright, New Jersey.  In: O.T.
Magson,  ed.   Coastal Zone '85.   New York:
American Society of Civil Engineers.

Leathennan,  S.P.   1984.   Coastal geomorphic
responses to sea level rise:  Galveston Bay, Texas.
In: Barth, M.C.,  and J.G. Titus, eds.  Greenhouse
Effect and  Sea Level Rise:   A Challenge for This
Generation.  New York:  Van Nostrand Reinhold
Louisiana Wetland Protection Panel. 1987. Saving
Louisiana's Coastal Wetlands and the Need for a
Long-Term Plan of Action. Washington, DC:  U.S.
Environmental   Protection   Agency/Louisiana
Geological Survey. EPA-230-02-87-026.

Lyle, S.D., L.E. Hickman, and HA. Debaugh. 1987.
Sea Level Variations in the United States. Rockville,
MD:  National Ocean Service.

Meier, M.F. et al.  1985.  Glaciers, Ice Sheets, and
Sea Level.  Washington, DC:  National Academy

Meier, M.F. 1984.  Contribution of small glaciers to
global sea level.  Science  226(4681):1418-1421.

Mercer, J.H. 1968.  Antarctic ice and Sangamon
sea level.  Geological Society of America Bulletin

Oldale, R.  1985.  Late quaternary sea level history
of New England:  a review of published sea level
data.  Northeastern Geology 7:192-200.

Revelle, R.  1983.  Probable future changes in sea
level resulting from increased atmospheric carbon
dioxide. In: Changing Climate. Washington, DC:
National Academy Press.

Schelling, T.  1983.  Climatic change:  implications
for welfare  and policy.   In:  Changing Climate.
Washington, DC: National Academy Press.

Sorensen, R.M., R.N. Weisman, and G.P. Lennon.
1984.   Control of erosion, inundation,  and salinity
intrusion.   In: Barth, M.C., and J.G. Titus, eds.
Greenhouse  Effect  and  Sea Level  Rise:    A
Challenge for This Generation. New York:  Van
Nostrand Reinhold Company.

State of Maine.  1987.  Dune Rule 355. Augusta,
ME:   Maine   Department   of  Environmental

Thomas, R.H. 1986. Future sea level rise and its
early detection by satelite remote sensing. In: Titus,
J.G., ed. Effects of Changes in Stratosphere Ozone
and Global Climate.  Washington, DC:  U.S. EPA
and United Nations Environment Programme.

                                                                                      Sea Level Rise
Thomas, R.H.  1985.  Responses of the polar ice
sheets to climatic warming.  In:  Meier, M.F., D.G.
Aubrey, C.R. Bentley, W.S. Broecker, J.E. Hansen,
W.R. Peltier, and RJ.C. Somerville, eds. Glaciers,
Ice Sheets, and  Sea Level.   Washington, DC:
National Academy Press.

Titus, J.G.  1988. Sea level rise and wetlands loss:
an  overview.   In: Titus, J.G., ed.  Greenhouse
Effect,  Sea Level Rise, and Coastal Wetlands.
Washington, DC: U.S.  Environmental Protection

Titus, J.G., ed. 1988.  Greenhouse Effect, Sea Level
Rise, and Coastal Wetlands. Washington, DC: U.S.
Environmental Protection Agency.

Titus, J.G.  1987. The greenhouse effects, rising sea
level, and society's response.  In:   Devoy,  RJ.N.
Sea Surface Studies.  New York: Croom Helm.

Titus, J.G.  1986. Greenhouse effect, sea level rise,
and coastal  zone  management.    Coastal Zone
Management Journal 14(3):147-171.

Titus, J.G.  1985. Sea level rise and the Maryland
coast.  In: Potential Impacts of Sea Level Rise on
the Beach at Ocean City, Maryland. Washington,
DC: U.S. Environmental Protection Agency.

Titus, J.G.  1984a. Planning for sea level rise before
and after a coastal disaster.  In: Earth, M.C., and
J.G. Titus,  eds.  Greenhouse Effect and Sea Level
Rise: A Challenge for This Generation.  New York:
Van Nostrand Reinhold Company.
Titus, J.G., M.C. Barth, J.S. Hoffman, M. Gibbs,
and M. Kenney.  1984b. An overview of the causes
and effects of sea level rise.  In: Barth, M.C., and
J.G. Titus, eds.  Greenhouse Effect and Sea Level
Rise:  A Challenge for This Generation. New York:
Van Nostrand Reinhold Company.

Titus, J.G., T. Henderson, and J.M. Teal.  1984.
Sea level rise and wetlands loss in the United States.
National Wetlands Newsletter 6:4.

Titus, J.G., C.Y. Kuo, MJ. Gibbs, T.B. LaRoche,
M.K. Webb, and J.O. Waddell.  1987. Greenhouse
effect, sea level rise, and coastal drainage systems.
Journal  of  Water  Resources  Planning   and
Management. American Society of Civil Engineers

Wilcoxen, PJ. 1986.  Coastal erosion and sea level
rise:   implications  for  ocean beach  and  San
Francisco's West Side Transport Project.  Coastal
Zone Management Journal 14:3.

WMO.  1986. World Meteorological Organization.
Atmospheric  ozone  1985.  Assessment  of  our
understanding of  the processes controlling  its
present distribution and change.   Global Ozone
Research and Monitoring  Project, Report No. 16.
Geneva,   Switzerland:   World  Meteorological


                                        CHAPTER 8
                           BIOLOGICAL DIVERSITY

Unlike most other impacts, loss of species and
reduced biological diversity are irreversible.  The
ability of a natural community to adapt to changing
climate conditions will depend on the rate of climate
change,  the size  of species  ranges, the  dispersal
rates of the individual species, and whether or not
barriers to species migration are present. If climate
changes rapidly, many species will be lost.

Species Diversity

•    The effect of climate change on species and
     ecosystems will most likely vary, with some
     species benefiting and others facing extinction.
     The  uncertainties  surrounding  the rate  of
     warming,  individual  species  response, and
     interspecies dynamics make impacts difficult to
     assess. However, climate change would alter
     competitive outcomes and destabilize natural
     ecosystems in unpredictable ways.

•    In many cases, the indirect effects of climate
     change on a population, such as changes in
     habitat,  in   food  availability,   and   in
     predator/prey  relationships,  may  have  a
     greater impact  than the direct physiological
     effects of climate change.

•    Natural and manmade  barriers,  including
     roads, cities, mountains,  bodies  of water,
     agricultural  land, unsuitable soil types, and
     habitat fragmentation, may block migration of
     species in response to climate change and
     exacerbate losses.

•    The areas within the United States that appear
     to be most sensitive to changes in climate are
     those that have a number of threatened and
     endangered species, species especially sensitive
     to heat  or drought  stress,  and species
     inhabiting coastal areas.
•    Rapid climate change would add to the already
     existing  threats  biodiversity  faces  from
     anthropogenic activities, such as deforestation
     and habitat fragmentation.

Marine Ecosystems

•    The  loss  of coastal  wetlands  and coastal
     habitat  resulting  from  sea  level  rise  and
     saltwater intrusion may profoundly affect the
     populations  of  all   inhabitants  of  these
     ecosystems,  including  mollusks,  shellfish,
     finfish, and waterfowl. However, there is no
     evidence to  indicate  these  species  would
     become extinct.

Freshwater Ecosystems

•    Freshwater fish in large bodies of water, such
     as  the  Great  Lakes,  may  increase  in
     productivity, but some significant species could
     decline. Fish in smaller bodies of water may
     be more constrained in their ability to respond
     to climate change. They also may be harmed
     by reductions in water quality.

Migratory Birds

•    Migratory birds are likely to experience  mixed
     effects from climate change, with some arctic
     nesting herbivores benefiting and continental
     nesters and shorebirds suffering.  The loss of
     wintering grounds due to sea level rise  and
     changing climate could harm many species, as
     would the loss of inland prairie potholes  due to
     potentially increased continental dryness.

Policy Implications

•    Existing refuges, sited  to protect a species or
     ecosystem under current climate,  may not be
     properly located for this purpose if climate
     changes or as species migrate.

Chapter 8
     Wildlife agencies such as the Department of
     the Interior, state government agencies, and
     conservation organizations may wish to assess
     the  feasibility  of  establishing   migratory
     corridors to facilitate species migration.

     Areas that may become suitable future habitat
     for threatened and endangered species, such as
     lowland areas adjacent to current wetlands,
     need to be identified and protected.

     The practice of restoration ecology may need
     to be broadened to rebuild parts of ecosystems
     in new areas as climates shift.

     The increase in the number of species  at risk
     as a result of climate change may require new
     strategies   for  balancing  ecosystem  level
     concerns with single species concerns. Agency
     programs  such  as   the  Fish and Wildlife
     Service's Endangered Species Program,  may
     wish to assess the relative  risk  of climate
     change and more current stresses on ecological

    Maintaining the  biological  diversity  of our
natural  resources  is  an important goal for the
nation.  The preamble to the Endangered Species
Act  of  1973  emphasizes the  value of  individual
species, stating that  endangered and threatened
species of fish, wildlife, and plants "are of aesthetic,
ecological, educational, historical, recreational and
scientific value to the Nation and its people." We
depend  upon our nation's biological resources for
food, medicine, energy, shelter, and other important
products.   In  addition to  species diversity, the
genetic  variability  within a  species and the wide
variety of ecosystems add  to  biological diversity.
Reduced  biological  diversity  could have  serious
implications for mankind as untapped resources for
research in agriculture, medicine, and industry are
irretrievably lost.

     The evolving biological diversity of this planet
is inevitably affected by climate  change.  Historic
climate changes have resulted  in major changes  in
species diversity. This has been true for the millions
of years life has existed on Earth. Now our planet
may face a more rapid change in climate that may
have important consequences for biological diversity.

The National Resource

     Public and private lands  in the United States
provide sanctuary for an abundant diversity of plants
and animals.  About 650 species of birds reside in
or pass through the United States annually. Over
400  species   of  mammals,  460  reptiles,  660
freshwater  fishes,  and  tens of  thousands  of
invertebrates  can be found  in  this  country, in
addition  to some 22,000  plants  (U.S.  Fish  and
Wildlife Service, 1981).  These species compose a
wide variety of ecosystem types within the United
States, including coniferous and broad-leaf forest,
grassland,  desert, freshwater, marine, estuarine,
inland wetland, and agricultural ecosystems. Figure
8-1 shows the major ranges of natural vegetation in
the United States.

     The U.S. national parks, forests, wilderness
areas, and fish and wildlife refuges are among the
public lands  that provide  sanctuary  for  wildlife
resources, including many endangered species. U.S.
public lands,  which encompass over  700  million
acres (about 32% of the land area  of the United
States),  support  about  700  rare species  and
communities (Roush, 1986). Over 45%  of the lands
held by  the  Forest  Service,  Fish and  Wildlife
Service, National Park Service, and Bureau of Land
Management  are in Alaska,  and  over  48%  are
located in the 11  most  western states (U.S.
Department of the Interior, 1987). However, much
of the nation's biological diversity lies outside these

     Private land holdings also account for a great
deal of this nation's biological endowment.  Private
groups, such  as the Nature Conservancy and the
Audubon Society, manage 500,000 acres and 86,000
acres, respectively, for biological diversity.

     Biological diversity can be broadly defined as
the full range of variety and variability within and
among  living  organisms.    It  includes  species
diversity,  genetic  diversity, and  ecosystemic or
community diversity.  This report concentrates on

                                                                                            Biological Diversity
                Sf 3 Spruce-fir forest

               r * va Transition pine-aspen forest

                Ne  Northeast hardwood forest

                phu Oak-hickory forest

                   Oak-pine forest

                   Southeast pine forest

                                            hSWj Southwest broadleaf woodland
p SgJ Short grassland
                                  Mesqulta and desert grassland
rซ*TB'd Tall grassland
                                c  Creosote bush shrubland
    Pacific coast forest
    north: spruce, hemlock
    south: douglas fir

   I Coast Range-Rocky
   •] Mountain conifer forest
   * lower: pine, douglas fir
    hlghen spruce-fir
    summits: alpine meadow
IVPJ Rlverbottom cypress-tupelo-sweetgum

feM a Mangrove swampland
                       Figure 8-1.  Natural vegetation in the United States (Hunt, 1972).
species  diversity,  but  only  because it is better
understood.  Genetic and ecosystemic diversity are
equally important.

Species Diversity

     Each species occurs in a characteristic range or
geographical area.  The factors controlling species
ranges are critical constraints on biological diversity.
The presence of a species in an area suggests that
the  species must have successfully achieved  the
following: (1) dispersal into an area (no barriers to
dispersal, such as the presense of bodies of water or
unsuitable soil types); (2) survival in that area (the
physical characteristics  of the area were suited to
the species' physiology,  and food was available); and
(3) establishment in the area (the organism found
an appropriate place in the food web in the absence
of excessive  competition and predation, and was
able to reproduce).
                    The stresses brought about by development,
               overuse, and alteration of habitat have fragmented
               much of  the  world's natural  habitat and  have
               created many new barriers. Consequently, for many
               species,  dispersal has become much more difficult
               than it was in the past.  For other species, humans
               have inadvertently aided dispersal and have caused
               rapid  spread  in  recent years.   Such practices as
               clearcut logging  prevent the  dispersal of species
               adapted to  dense  forest  conditions   (e.g.,  flying
               squirrels)  and promote  the dispersal  of species
               suited to open areas  (e.g., deer).

                    Currently, 495 species are listed as endangered
               within the United  States, and over 2,500 species
               await consideration for that status by the  Fish and
               Wildlife Service.  The list of endangered species is
               dominated by plants, birds, fishes, and mammals but
               also includes insects, amphibians, reptiles, mollusks,
               and crustaceans  (U.S. Fish and  Wildlife Service,

Chapter 8
     New  species  are  created  through  the
evolutionary process of speciation, whereas existing
species  are lost  through extinction.   Speciation
generally requires at least hundreds of thousands of
years.  However, extinction as a result of human
activities, even without climate change, is occurring
rapidly and at  an increasing rate.  Owing to its
slowness, the process of speciation does little to
offset species' loss to extinction.

Stressed Biological Diversity

     Biological diversity continues to erode steadily
around the globe  as a result of human activities.
Habitat destruction, degradation, and fragmentation
have resulted in the loss of many species and have
reduced  the  ranges  and populations  of others.
These impacts  affect  all  three levels of biological
diversity. Through providing an additional pressure
on ecological systems, climate change will further
reduce the biological diversity in this nation and
around the globe.

     It is difficult to  determine the exact rate of
species extinction because the number of species on
the Earth is known only to an order of magnitude.
A recent estimate by Wilson (1988) places the total
number  of species  between  5  and  30 million.
Assuming  10  million species, Wilson made the
rough calculation that one in every 1,000 species is
lost  each year.  Wilson then compared this to
estimates of extinction rates over geologic  time,
which ranged between 1 in every 1 million and 1 in
every 10 million each year.  Thus, human activities
may be eliminating  species at least  1,000 times
faster than natural forces.

     The significance of rare species should not be
underestimated. A narrowly or sparsely distributed
species  may be  a  keystone in  an  ecosystem,
controlling  the structure and functioning of the
community, or it may be a species of great and yet
unknown value to humans.

Genetic Diversity

     Each species that persists has a characteristic
genetic  diversity.  The pool of genetic  diversity
within a species  constitutes an adaptation to its
present environment as well as a store of adaptive
options  for   some   possible  changes  in  the
environment.  The loss  of genetic  diversity can
contribute to the extinction of a species by reducing
its  ability  to  adapt to  changing environmental

     Generally, species with larger populations have
greater genetic diversity.   Species near extinction
represented by few individuals in  few populations
have lower genetic diversity, a situation exacerbated
by inbreeding.  Additionally, extreme climatic events
may cause bouts of natural  selection that reduce
genetic variability (Mayr, 1963).

Community and Ecosystemic Diversity

     Ecosystemic  diversity  is  the  number,  of
distinctive  assemblages  of  species  and  biotic
processes that occur in different physical settings.
A long-leaf pine forest, a sand dune, and a small
pond are all  part of our diversity at this level.
Ecosystems  come into existence through complex
physical  and biological processes not now well
understood.   They may  be  lost  by  outright
replacement  of  one  by  another  (as  in  the
desertification of a grassland) or by  the gradual
merging of two formerly separate ecosystems (as in
the loss  of some  estuarine systems when they
become  saltier   and  take   on  more  of  the
characteristics of a purely  saltwater  ecosystem).
Ecosystems  can also be  eliminated  because  of
human activities  (as in the filling in of a wetland).

     Species respond to environmental change on a
hierarchy  of time  scales.   For  relatively small
changes  occurring  within  the   lifetime  of  an
individual, each member of the species can respond
through a variety  of physiological  adjustments.
Individual species differ in their ability to adjust to
change. Some can withstand a great deal of climate
change, whereas others are restricted to a narrow
range.  Over several generations, natural selection
can  cause genetic  adaptation  and  evolution hi
response to the change. Alternatively, a species can
respond to climate  change  by moving into a new
area through migration and dispersal.  This can
occur over a relatively short period of time if the

                                                                                      Biological Diversity
 species has the biological ability to move quickly.
 The discussion of  response to  climate change
 centers on  migration  as the  response that  could
 occur over a relatively short period of time.

      The distributions of species are significant
 indicators of climate change. Local climate appears
 to be the primary factor defining an environmental
 setting and determining the species composition and
 spatial patterns of communities in terrestrial zones
 (Bolin  et  al.,  1986).    Temperature  means,
 temperature extremes, and precipitation are the
 factors most often affecting the potential natural
 distributional limits of a species (Ford, 1982),  while
 the actual distribution  of a species is also affected
 by soil type, soil moisture, ecological dynamics, and
 regional isolation.

 Rate of Climate Change

      Predicting how a  species of ecosystem might
 respond to a given environmental change is difficult.
 Adaptation  to climate  change  will  inextricably
 depend on the rate of climate change.  For some
 species, migration rates may be inadequate to keep

      The large number of combinations of dispersal
 range and age to reproduction make the potential
 rate  of  migration  different  for  every  species.
 Paleorecords suggest migration  rates between 10
 and 20 kilometers per century for chestnut, maple,
 and balsam fir, and between 30 and 40 kilometers
 per century for some  oak and  pine  species (see
 Chapter 5: Forests).   On the other  hand, cattle
 egrets have shown a much quicker migration rate by
 colonizing all of the North American tropics within
 approximately  40 years.

     As species .shift at different rates in response
,to  climate change, communities may disassociate
 into new arrangements  of species. Local extinction
 can  result  either   directly   from  physiological
 pressures or indirectly from changes in interspecies
 dynamics.  Hence, the effect of climate change on
 an area will  be to cause sorting and. separation of
 species  as  a  result of the  differential  rate  of
 migration  and  species retreat  (Ford,  1982).
 Ecosystems, therefore, will not migrate as a unit.

     Species  do  not  immediately respond to
 changed and changing environmental conditions, A
 negative response,  such as local extinction in an
 area, is usually quicker than the positive response of
 new species' colonization of a region (see Chapter
 5: Forests).  In the Arctics, the lag period between
 climate  change and species response by migration
 and colonization may be  several  hundred years
 (Edlund, 1986).  This lag  period will  leave areas
 open  for weedy, opportunistic  species that  can
 quickly migrate and propagate in a region.

     The rate of climate change will be crucial to
 the survival of the species hi an ecosystem. A 3ฐC
 (5ฐF)  increase in temperature, for example, would
 effect  a several hundred kilometer poleward shift in
 the temperate vegetation belts (Frye, 1983). If this
 change took place within a century,  species would
 need to migrate several kilometers each year to
 adapt  to this warming.  Plants have a wide  range of
 migration rates, and  only  some  may  be  able to
 achieve this rate. Failure of a species to "keep up"
 with   suitable   environmental  conditions would
 eventually result in extinction.

     Many factors make evaluating the impact of
 climate change  on ecosystems difficult.  The great
 inter dependencies among species in an ecosystem
 add considerable uncertainty to the  effect  that the
 various responses of individual species will have on
 the system. An impact upon a single species could
 profoundly affect the entire  ecosystem.  Certain
 species are vital to the workings of their ecosystems.
 Among  them are large  carnivores  that  regulate
 predator-prey relationships,  large herbivores that
 significantly change vegetation, and organisms that
 pollinate plants (WEI, 1988). Plants can also be key
 species  within   an  ecosystem.    For example,
 elimination of a tree species in a region could have
 a significant effect on the whole  forest  ecosystem,
 including birds, insects, and mammals.

     Animal populations are generally much more
 mobile than plants. But animal distributions heavily
 depend on vegetation  for  food,  protection,  and
 nesting habitat.  Species not directly dependent on
 vegetation ultimately depend on some other species
 that  is.  The ranges of the fig wasp and the fig
 depend entirely upon one another. In this case,  the
 plant species depends  on a single pollinator, and
the insect species relies upon a  single  species of
plant for food (Kiester et al., 1984).

Chapter 8

Effect on Genetic Diversity

     With regard to genetic diversity, rapid climate
change   would  select  for   those   genotypes
(combinations of genes) that were best suited to the
new climate regime and would tend to eliminate
others.   This process of natural selection would
usually  decrease the genetic variability within a
population.  In the long term (evolutionary tune), it
is possible  that greater climatic variability could
select for greater genetic variability.

Barriers to Response

     The rate of species migration is also affected
by  natural and  manmade  barriers and  by
competition. Peters and Darling (1985) examined
the potential responses of species to climate change,
ecological interactions, and barriers to  adaptation.
Physical barriers include mountains, bodies of water,
roads, cities, agricultural land, inappropriate soil
type,  and   habitat   heterogeneity   (landscape
patchiness).  A  species whose migration rate is
sufficient to keep up with changing conditions could
become constrained by a physical barrier. Inability
to cross the barrier could result in a reduction of
the range of the species and its eventual extinction.

Reserve and Island Species

     Additional  constraints  on  the  ability  of
populations living on reserves to respond to climate
change frequently result from  insufficient habitat
area or  isolation  from other  populations.  The
problem  of isolation is similar to that  of island
species  and  has  become known  as the  island
dilemma. Species on reserves are often remnants of
larger  populations and are  more susceptible to
environmental stress and extinction.

     Species on reserves are likely to be pressured
from two directions as a result of climate change.
A population isolated on  a reserve surrounded by
altered   or  unsuitable  habitat  receives  little
immigration from  populations outside the reserve.
Also, that population may not be able to colonize
areas outside  the  reserve as these areas become
suitable because of development or other alterations
of habitat.

     Even without the  added pressure of climate
change, reserve populations are vulnerable because
many reserves  are not large enough  to support a
self-sustaining  population (Lovejoy,  1979).   The
predictive theory  of island  biogeography showed
that, other factors being  equal, small islands
accommodate smaller numbers of species than do
large islands (MacArthur  and Wilson, 1967). This
held true for  other ecological "islands," such as
mountaintops,  woodlots,  and lakes.  Also, when
large  ecosystems   become   smaller   through
fragmentation, the number   of  species  always
declines.    Figure  8-2  shows  how mammalian
extinctions have been inversely related  to  refuge
area in North American parks.
3 35-
u_ 30-
S 25-
ฃ> "
ซ 20-

g 15-
a 10-
v* S-

1 3 5

4 6
7 B

9 .


6 ' 8 ' 10

Bryce Canyon
Lassen Volcano
Crater Lake
Mount Rainier
Rocky Mountain
Sequoia-Kings Canyon
Grand Teton-Yellowstone

AREA (km2)'

31 •

 Figure 8-2. Habitat area and loss of large animal species in North American parks (1986) (Newmark, 1987)

                                                                                   Biological Diversity
     Reserves that originally may have been well
sited to protect  a vulnerable population and  its
habitat may, after climate change and population
response, exist  outside  the  now suitable range.
Figure 8-3 illustrates this problem.  Large reserves
and buffer zones around reserves  help  to lessen
these problems.  Corridors between reserves lessen
the  problem of  spatial  isolation by  allowing for
some migration between reserves.
Figure 8-3.  Effect of climate change on biologica
reserves.   Hatching   indicates   the  following:
(A) species distribution before human habitation
(SL  indicates southern limit of  species range);
(B) fragmented species distribution after human
habitation; (C) species distribution after warming
(Peters and Darling, 1985).

Mountain Species

    Just as species can migrate latitudinally, they
can respond altitudinally to  climate  change by
moving up or down a mountain slope. Species can
often respond more easily to changing conditions on
a slope because a shorter distance is required to
migrate to achieve the same temperature change.

     Among the problems associated with altitudinal
migration are displacement of the species at the top
(Peters and Darling, 1985). Also, with the increase
in  altitude,  the  area available for  colonization
usually becomes  smaller,  communities  become
isolated,  and these smaller populations are more
prone to extinction.

     This section reviews some previous studies of
ecological  response to past changes  in  climate,
recent studies  of potential  response  to climate
change, and studies done for this report, which use
climate change scenarios from general circulation
models for a doubled CO9 environment (see Table

Forest Ecosystems

     The tree species that make up any forest are
major factors in determining the biological diversity
found there. Trees provide a multitude of habitats
and  are the basis of much of the food web in a

     Changes in forest composition resulting from
climate change (see Chapter 5: Forests) would have
significant  implications  for  biological  diversity.
Potential northerly range shifts of several hundred
to a thousand kilometers may be limited by the tree
species' ability to disperse.  One possibility is that
southern pine forests will move farther north into
the regions currently occupied by mixed hardwood
species.  Some  of these hardwood forests contain
the highest tree species diversity found anywhere in
the United States (Braun, 1950).  If they migrated
north, species would inevitably be lost,  and overall
biological diversity would substantially decrease.

     If forests were disrupted by the extinction of
the dominant tree  species,  the land would be
invaded by weedy, opportunistic species.  This would
create a system with very low diversity, similar to
that  following  logging.   Ultimately,  these  new
systems would not persist as succession took place,
but the pattern of succession following the removal
of a forest by rapid climate change is unknown.

Chapter 8
Table 8-1.  Studies Conducted for This Report and
           Cited in This Chapter
     Potential Responses of Great Lakes Fishes and
     Their Habitat to Global Climate  Warming -
     Magnuson, Regier, Shuter, Hill, Holmer, and
     Meisner, University of Wisconsin (Volume E)

     The Effects of Global Climate Change on the
     Water Quality of Mountain Lakes and Streams
     - Byron, Jassby, and Goldman,  University of
     California at Davis (Volume E)

     The Effects of Climate Warming on Lake Erie
     Water  Quality -  Blumberg  and  DiToro,
     HydroQual, Inc. (Volume A)

     Ecological Effects of Global Climate Change:
     Wetland Resources of San Francisco Bay -
     Josselyn and Callaway, San Francisco State
     University (Volume E)

    , Projected  Changes  in  Estuarine  Conditions
     Based on Models of Long-Term Atmospheric
     Alteration - Livingston, Florida State University
     (Volume E)
Tropical Forest Ecosystems

     The   greatest   concentration  of  biological
diversity in the world is in the rain forests of the
Tropics (Wilson, 1988). Besides reducing diversity,
deforestation contributes to disruption of the global
carbon cycle by releasing CO2 into the atmosphere
and will directly affect the rate of climate change
(Prance, 1986).   Indeed,  on  a global  scale, the
problems  of  tropical deforestation, rapid  climate
change  through (among other factors)  increased
CO, production, and the loss of biological diversity
can be seen as aspects of the same problem.

     Tropical forests are also important as wintering
grounds for migratory birds coming from the United
States and as sources of new knowledge, because
the patterns  of interactions between species and
climate  are at  their most sensitive  and complex
there (Robinson, 1978; Janzen, 1986). The  Tropics
may provide  important leading indicators of the
ecological effects of climate change.
Freshwater Ecosystems

     A  study  conducted  by  Magnuson  et  al.
(Volume E) concludes that in most areas of the
Great Lakes, climate warming would increase the
amount of optimal thermal habitat for warm-, cool-
and coldwater fishes (see Chapter 15: Great Lakes).
Although overall productivity would increase, overall
biological   diversity  could   decrease  through
intensified species interactions.

     A  study by Byron et  al. (Volume E)  on
mountain lakes suggests that climate change would
cause  a  range  of  impacts,  including  higher
productivity,  changes in species composition, and
decreased water quality resulting from an increase
in  algal growth  (see  Chapter  14:  California).
Blumberg  (Volume  A)  found  that  thermal
stratification in Lake Erie could decrease dissolved
oxygen levels.

     The combined pressures of warmer waters,
saltwater intrusion,  and a  rising  sea level would
significantly affect estuaries.  The regional studies
suggest that coastal estuaries would see a growth in
marine species and a loss of some estuarine species.
A study by Josselyn (Volume E)  on  the  San
Francisco Bay estuary suggests a decline in species
that use the  delta for spawning (see Chapter 14:
California).  Livingston (Volume E) concluded that
crabs,  shrimp,  oysters,  and  flounder   in  the
Apalachicola estuary could not survive the  warming
in the GISS and GFDL scenarios (see Chapter 16:

Saltwater Ecosystems

     In  general, a warmer global climate  would
increase  productivity in ocean fisheries,  but the
location and relative abundance of species are likely
to change (Sibley and  Strickland,  1985).  Up to
some threshold temperatures, such as  2ฐC (4 F),
warmer ocean temperatures would increase ocean
productivity in many species,  but beyond that
threshold,  productivity  could  decline  (Glantz,
Volume J). It is likely that as productivity decreases,
biological diversity would decrease as well.  Warmer
temperatures would most likely  cause  fish  to
migrate poleward,  although many other factors, such
as shifts in upwelling, may affect this.

                                                                                     Biological Diversity
 Coral Reef Ecosystems

      Coral reefs provide the structural base for the
 very biologically diverse reef ecosystems.   Coral
 reefs  in  the  Caribbean  and the Pacific may  be
 severely  stressed as  a  result  of warmer  water
 temperatures and the rising sea level associated with
 climate change.  Extensive bleaching of coral (the
 expelling  of  symbiotic   algae  in  response  to
 environmental stress)  occurred in the Pacific after
 the 1982-83 El  Nino  (Glynn,  1984) and  in  the
 Caribbean following a summer of elevated water
 temperatures in 1987 (Roberts, 1987). Loss of the
 algae, the primary  food  source  of  the coral, is
 thought  to kill coral,  making the reef  ecosystem
 vulnerable to erosion and physical devastation.

     Coral reefs also will very likely be affected by
 sea level  rise.  Studies by Buddemeier and Smith
 (1988) and  Cubit  (1985)  suggest  that  vertical
 accretion of reef flats eventually may be unable to
 keep up with an accelerating rise in sea level. Reef
 flats also maybe subject to the stress of increasingly
 large waves, erosion, and sedimentation,  which can
 inhibit coral growth  (Buddemeier, 1988).

 Arctic Ecosystems

     Within the North American Arctics,  plant size,
 vigor,  and  reproduction  could  be  expected  to
 increase with higher  temperatures in the  near term
 (years to  decades).  Some low-lying plants would
 most likely become upright, and there would be a
 northerly movement  of  the  tree  line and  all
 vegetative zones (Edlund, 1986).

     Over   the  longer  term,   however,  rising
 temperatures may be   a  mixed blessing. Overall
 biological productivity is likely to increase, and some
 species may be able  to increase  their  range.
 However, some arctic plant species are likely to be
 out-competed by invading species, and many others
 would face  the  same  type  of  problem that
 mountaintop species face:  they would have nowhere
 to go  once they reach the Arctic Ocean.  Thus,
native arctic species may be especially at risk.
 Other arctic species may face their own problems.
For example, caribou would be severely harmed if
rivers  do  not freeze for  periods long enough to
allow for migration.
 Migratory Birds

      Migratory waterfowl are likely to experience
 very  mixed   effects   as  a  result  of  warmer
 temperatures  (Boyd,  1988).   Herbivorous, arctic
 nesting species, such as geese, could benefit from
 the shortened winter season and from the increases
 in vegetation, in nesting habitat, and in ecosystem
 productivity (Harington,  1987).   Smaller  arctic
 nesting shorebirds, on the other hand, would be
 harmed by the encroachment of taller vegetation,
 potentially eliminating  the  preferred  low-lying
 tundra breeding  ground.    Other   effects   on
 shorebirds could result from changes in ecosystem
 predator-competitor relationships and  changes in
 the seasonal timing of such events as larval blooms,
 upon which these birds  depend for nourishment
 while they are in a  flightless stage  and during
 migration (Myers, 1988).

      Waterfowl that breed in the continental interior
 may  suffer more than arctic nesters. Over half of
 all waterfowl in North America originate hi  the
 prairie  pothole region, a large  agricultural area
 riddled with ecologically productive permanent and
 semipermanent wetlands.  Increased temperature
 and changes in seasonal precipitation could reduce
 the highly variable number of potholes (wetlands)
 in  the  area and could  significantly  impair  the
 productivity of breeding ducks.

     Because of the drought of 1988, over 35% of
 the seasonal wetlands  within the prairie  pothole
 region were dry during the breeding season (U.S.
 Fish  and  Wildlife Service,  1988).  The Fish and
 Wildlife Service forecast that only 66 million ducks
 would migrate  during the fall of 1988,  a total of 8
 million  fewer than in 1987 and the  second-lowest
 migration on record (Irion, 1988). The productivity
 index for mallards (number of young per adult) was
 0.8, which was  down by over 20%  from the
 historical  average (U.S. Fish and Wildlife Service,

    Waterfowl and other migratory birds are likely
 to  be affected on both ends  of their migratory
journey and at staging areas along  the  way. The
 loss of  coastal  wetlands, already an area of great
 concern in the United States, reduces the amount of
 habitat  available to waterfowl, creating population
 pressures on a limited resource. Of the 215 million

Chapter 8
acres of wetlands in the coterminous United States
at the  tune of settlement, fewer than 99 million
acres (46%) remain (U.S. Fish and Wildlife Service,
1988).  Loss of an additional 26 to 82% of existing
coastal wetlands could  occur over the next century
as a result of a 1-meter rise in sea level, saltwater
intrusion, and human development (see Chapter 7:
Sea Level Rise).  Loss of wintering habitat along the
Gulf of Mexico would  affect  many waterfowl,
including mallards, pintails, and snow geese.

     The Tropics, the winter home for many species
of migratory birds, may be significantly altered by
rapid climate change. The need to protect a species
in all parts of its range  underscores the truly global
nature of the effects of  rapid climate change on
biological diversity (Terborgh, 1974).

Endangered Species

     Hundreds of species  are currently listed  as
endangered in  the  United  States,   and  several
thousand await consideration for that status.  These
species are likely to be stressed further as a result
of climate change.

     Threatened and  endangered species of the
Southeast would be very susceptible to the impacts
of sea level rise.  Some species potentially at risk in
that region include the Key deer, manatee, Florida
panther, and Everglades kite (Breckenridge, 1988).
Climate  change could also greatly  increase the
number of rare, threatened, and endangered species
in the United States.

Other Direct  and Indirect  Stresses

     As  plant  and animal species  experience
increasing pressures  from changes in temperature,
precipitation,  and   soil   moisture,  so  too  will
agriculture and urban water supplies.  The changes
that result from the human response to climate
change may have the greatest impact on biological
diversity.   If the continental  interior  of North
America dries, for example, wetlands that dry out
may be cultivated, and our current uses of water
resources may change.  These secondary effects may
significantly  compound  the  loss  of  biological

     Climate change presents new challenges for
policymakers, regulators, and resource managers.
Planning for climate change may help to minimize
the disruption  to  natural systems  and facilitate
adaptation under  changing conditions.  Decisions
will  need  to  be  made in  an environment  of
increased pressure on many other resources.

     Policies regarding rare and endangered species
are likely to change as the number of species at risk
greatly increases. As more species become stressed
and  potentially threatened   by climate change,
reevaluation of protection policies may be required.
The  tradeoffs  between protection  of individual
species  and  species'  habitats  and  the  broader
protection of biodiversity at the level of ecosystems
may need  to be reexamined.   As a part of this
question, decisions concerning whether to protect
existing communities or to foster establishment of
new communities may need to be made.

Management Options to  Maintain
Biological Diversity

    • Only  a limited number of techniques are
available  for  maintaining   biological  diversity.
However, these techniques  can be  adapted and
intensified to meet the potentially great impacts of
rapid climate change.

Maintenance of Native Habitats

     The  most  direct way to maintain  biological
diversity is to manage land  to retain ecosystems,
communities, and habitats.  This already has been
successfully undertaken on a broad scale by federal
and state governments and by private organizations.
Ecosystem conservation, especially as represented by
the  national  parks and other large   reserves,
maintains much of our national biological diversity.
These ongoing  efforts will be the crucial first step
for maintaining biological diversity in the face of
climate change.

                                                                                     Biological Diversity
     Land acquisition  and management  policies
should take climate change into account.  Climate
change  and the  future  requirements  of whole
ecosystems  should be considered  in  siting and
managing  reserves.    To preserve  functioning
ecosystems, large areas of land  will be required.
Preserves would need to be at least large enough to
support  self-sustaining populations.   Lands that
could be more important as future plant and animal
habitats need to be identified and evaluated.  Land
managers should consider whether these  lands
should be set  aside.   Although identification  of
appropriate future habitats is  difficult  and highly
dependent on the future rate and extent of climate
change, some areas, such as lowland areas adjacent
to current wetlands, hold good potential for habitat

     To protect a species, alternative sites should
be considered with regard to the ecological needs of
target species under  changing conditions.  Siting
reserves in mountainous areas is beneficial because
it allows for the shorter-distance altitudinal shifts of
adjustment to changing climate. Stream corridors,
which can be  effective avenues of dispersal for
terrestrial as well as aquatic organisms, should be
protected wherever possible.  Providing corridors
for migration between reserves also should enhance
the ability of wildlife to adapt to climate change.
Ideally,  these corridors should  be wide enough to
maintain the ecosystem characteristics of the reserve
in their  center.   Some species  do not  find the
habitat conditions of narrow corridors suitable for

     The pressures caused by changing climate are
likely to exacerbate competing land-use demands.
Acquisition   of  land   for  preserving   biological
diversity will often be difficult, especially in  areas
where agriculture or forestry may  be  expanding.
Flexible  management strategies  that reserve the
possibility of  land  management  for  biological
diversity in the future, while allowing for other use
in the interim, hold potential for reducing resource
conflicts  and  maintaining  biological  diversity.
Creative   approaches   such    as   encouraging
hedgerows, which may serve as migratory corridors,
should also be considered.

Maintenance of Species in Artificial  Conditions

     When individual species are threatened with
extinction, a possible option is to ensure that the
species is propagated in  captivity.  Indeed, some
rare species, such as  the Pere David deer and the
California condor, now exist only in captivity.  This
technique can be made to work for a variety of
species, depending on their biology and the degree
to  which   they   successfully  adapt to  captive
conditions.  As more species become  threatened
with extinction due to climate change,  the effort
applied  in  this  area  may  have  to  increase
dramatically. However, only a tiny fraction of the
nation's  species  can  be  maintained in this  way.
Existing seed  bank  programs  also provide  an
important  method for conserving  plant genetic

Restoration of Habitat

     Restoration  ecology is a new discipline whose
goal is to  develop methods to  restore  damaged
ecological  communities  to their prior  unaltered
state. Except in forestry, where reforestation has a
longer tradition,  restoration ecology has been in
existence for only a few years. Nonetheless, it offers
some real promise for ameliorating the effects of
rapid climate change.

     Normally, restoration is done at the site where
the community previously existed and was altered or
damaged. Historical  and baseline information is
used to manage  the  species in such a way as to
eliminate unwanted new species and to encourage
and possibly reintroduce native species.

     Perhaps the  theory and practice of restoration
ecology could be expanded to include rebuilding
natural communities on sites where  they have not
previously existed.  This activity  has not yet been
attempted   but  may  be  necessary   to   save
communities displaced by climate change.  If the
climate changes so that many of the  key species of
a community can no longer survive in their original
range, and if the species are incapable of dispersing
and establishing  themselves elsewhere,  then  the
artificial transplantation of components  of entire
communities  may  become  necessary.     This
transplantation   of  communities  would  be   a
monumental task and could  help to save much
biological  diversity,   but  it  cannot possibly be
undertaken on  the scale necessary to preserve all
species threatened by climate change. Restoration
ecology  can be  useful   for  extending  reserve
boundaries and for providing migratory corridors.

 Chapter 8
 Planning Options

      While  there  are only  a few management
 techniques to maintain biological  diversity,  many
 different groups in our society can implement them.
 These groups can be divided into  the private and
 public sectors.

      Many different  groups in the private sector,
 ranging   from  private  individuals  to   large
 conservation organizations, will have an interest in
 maintaining biological diversity. However, all would
 need information about the current and probable
 future state of biological diversity.  The federal
 government may be  able to  play  a role here by
 providing  information  on the state of  biological
 diversity, including the systematics and distribution
 of species; on the genetic variability of species; and
 on the distribution of communities and ecosystems.

     The  four major  federal land management
 agencies  develop plans  intended   to lay  out a
 comprehensive   framework   and   direction  for
 managing   federal   land.     Land   Resource
 Management Plans,  required for  each  national
 forest, define the direction of management in the
 forest for the next 10 to 15 years.  In addition, the
 Forest Service prepares 50-year plans, as required
 by the Resource  Conservation Act.  The National
 Park Service prepares a General Management Plan
 for each unit in the system that defines a strategy
 for achieving  management objectives within  a 10-
 year time frame.  A Statement for Management is
 also  prepared for  each  national  park  and  is
 evaluated   every  2   years;  this  includes   a
 determination of information needs.  The Bureau
 of  Land   Management's   (BLM)   Resource
 Management  Plans and  the  Fish and  Wildlife
 Service's Refuge  Master  Plans are prepared and
 revised  as needed  for  BLM  resource areas and
 wildlife  refuges (U.S. Department of the Interior,
 1987). These periodic reviews of the management
 plans for public lands should include consideration
 of the  possible   effect  of  climate change  on

     Some federal land management agencies are
 beginning to devote resources to the climate change
 issue. The Forest Service, for example, has begun
 planning the Forest Atmosphere Interaction (FAI),
which will  be concerned with  the relationship
between the atmosphere and our national forests.
The FAI has been designated  a priority research
program for the Forest Service.
     The federal government manages an enormous
 amount of land and should consider management
 options to preserve biological diversity on much of
 that land.  The major management techniques of
 habitat maintenance and restoration ecology could
 be applied by the agencies actively responsible for
 managing the nation's public lands.

     The ability to  protect biological diversity is
 severely restricted by a lack of knowledge regarding
 the rate of climate change, the precise nature of the
 change, how individual species will respond, and
 how ecological balances will shift. Research should
 be  expanded  in  two  areas: identification  of
 biological  diversity,  and species interactions and
 biological diversity.  New management options for
 biological diversity  should  be derived from  these

 Identification of Biological Diversity

     First and most important, an intensified, better
 coordinated   research   effort,  involving   both
 systematics (organism classification) and ecology, is
 required  to  identify  the  biologically   diverse
 resources of our country.  There should be more
 coordination to  identify U.S. plants and animals,
 range maps, and habitat requirement information
 for those species.

     The apparently simple task of identifying the
 species  of plants  and animals that  exist in a given
 area  is actually a  major  barrier  to   further
 understanding.    Although common species  are
 usually easy to identify, serious problems are often
 encountered in attempts to determine whether a
widespread  group is,  for  example, one  or two
species. For example, there is currently no federally
sponsored Flora (listing of all known plants) of the
United States.

    Although it is necessary to describe the genetic
diversity of our nation's  species, it is difficult to do
so in a direct fashion. What may be feasible is the
further development of population genetic  theory
and of data that would predict the genetic diversity
of a species based on species' properties, such as
population size and habitat range variability.

                                                                                  Biological Diversity
    The challenge in describing ecosystem diversity
is to find the system of classification that best helps
make decisions intended to minimize the loss of
biological diversity.  Such a system will most likely
only be found through experience. For  now, we
should continue with the many different approaches
of ecosystem classification, and we should look for
the strengths of each.

Species Interactions and Biological

     The second area to which research should be
devoted is the direct effects of climate on species
and the indirect interactions of species with other
species dependent  on climate.    Comprehensive
mapping of species' ranges along temperature and
moisture  gradients  would  provide   valuable
information. The direct effect of climate change on
vegetation needs  to be better assessed, and more
estimates   of  species'   dispersal  rates  would
significantly improve our ability to identify species
at greatest risk.

     A variety of ecosystems within a diversity of
climatic regions and terrains should be intensively
studied using analog climate regions under changed
climate conditions.    Although   an  ecosystem's
response under changing climate conditions will not
be  wholly  predictable,   modeling   individual
ecosystemic responses would enhance knowledge of
the likely effects.  Further research on how species
interact and how trophic structures might change
with climate would help predictive capabilities.

     There should be further study on the question
of the relationship between ecosystem  function and
species diversity to resolve the uncertainty in this
area  Modeling  the effect of climate change on
ecosystem function and its relationship to diversity
would help with predictive capabilities.

     It will be impossible to study in detail even  a
fraction of the nation's species. The groups chosen
for study either  must be representative of many
species or must  possess some special properties
(such as extreme sensitivity to climate change). The
method of deciding which group to study is itself a
major outstanding research question.

Bolin, B., B. Doos, J. Jager, and R. Warrick, eds.
1986.  The Greenhouse Effect, Climatic Change,
and Ecosystems (SCOPE 29).  Chichester, England:
John Wiley and Sons.

Boyd, H.  1988.  Impact of Climate Change on
Arctic-Nesting Geese. Paper  presented at Climate
Institute Wildlife Symposium,  Washington,  DC;
January 21.

Braun, E.L. 1950.  Deciduous Forests of Eastern
North America.  New York: Hafner Publishing

Breckenridge,  R.  1988.   Challenge of Assessing
Climate  Change Effects  on Fish and Wildlife
Resources in the Western United States.  Paper
presented at the Western Association of Fish and
Wildlife  Agencies,  Western  Division of the
American Fisheries Society, Albuquerque, MM; July

Brown, L., W. Chandler, A. During, C. Flavin, L.
Heise, J. Jacobson, S. Postel, C. Shea,  L. Starke,
and  E. Wolf.   1988.  State  of the World 1988.
WorldWatch Institute.  New York:  W.W. Norton
and  Co., Inc.

Buddemeier, R. 1988. Impacts of Climate Change
on Coral Reefs, Islands and Tropical Coasts. Paper
presented at the Climate Institute Workshop on the
Impact of Climate  Change on the Third World,
Washington, DC; March 24-25.

Buddemeier, R., and D. Hopley. 1988.  Turn-Ons
and  Turn-Offs:   Causes and Mechanisms  of the
Initiation and Termination of Coral Reef Growth.
Paper  presented for submittal to the Proceedings
of the Sixth International Coral Reef Symposium,
James  Cook  University,  Townsville,  Australia;
August 8-12.

Buddemeier, R., and S. Smith. 1988. Coral Reef
Growth  in an  Era of Rapidly Rising Sea Level:
Predictions   and  Suggestions  for  Long-Term
Research.  Coral Reefs 7:51-56.

 Chapter 8
 CEQ.  1980.  Council on Environmental Quality.
 Environmental Quality:  Eleventh Annual Report.
 Washington, DC: U.S. Government Printing Office.

 Cubit, J.  1985. Possible Effects of Recent Changes
 in Sea Level on the Biota of a Caribbean Reef Flat
 and  Predicted Effects of  Rising  Sea Levels.
 Preprint  from the  Proceedings  of  the  Fifth
 International Coral Reef Congress, Tahiti.

 Edlund, S.   1986.   Modern Arctic Vegetation
 Distribution and  Its Congruence With Summer
 Climate   Patterns.     Environment   Canada,
 Proceedings: Impact  of Climatic Change on the
 Canadian Arctic.

 Ford, Michael.  1982.  The Changing Climate.
 London: George Allen and Unwin.

 Frye, R.   1983.   Climatic Change  and  Fisheries
 Management. Natural Resources Journal 23:77-96.

 Glynn, P. 1984. Widespread Coral Mortality and the
 1982-83 El Nino Warming Event.  Environmental
 Conservation 11(2):133-140.

 Harington, C.   1987.   The Impact of Changing
 Climate  on Some  Vertebrates  in  the  Canadian
 Arctic.  Climate Institute, Proceedings of the First
 North American  Conference on Preparing  for
 Climate Change.

 Hunt, C.B. 1972. Geology of Soils. San Francisco,
 CA:  W.H. Freeman.

 Hunt, C.B.  1974.  Natural Regions  of the United
 States and Canada.   San Francisco, CA:  W.H.

 Irion, R.  1988.  Drought Helps  Two Endangered
 Species Rebound, but It's a Dismal Year for Ducks.
 The Washington Post, August 1.

 Janzen, D. 1986. The Future of Tropical Ecology.
Annual Review of Ecology and Systematics 17:305-

Jordan,  W.,  R. Peters, and  E. Allen.   1988.
Ecological Restoration as a Strategy for Conserving
Biological Diversity.  Environmental Management
 Kiester, A.R., R. Lande, D.W.  Schemske.  1984.
 Models of Coevolution and Speciation in Plants and
 Their Pollinators. American Naturalist 124(2):220-

 Leatherman, S. 1987. Effects of Sea Level Rise on
 Beaches and Coastal Wetlands.  Climate Institute,
 Proceedings  of  the  First   North   American
 Conference on Climate Change, Washington, DC,
 pp. 140-146.

 Lovejoy, T. 1979. Refugia, refuges and minimum
 critical size: problems in the conservation of the
 neotropical herpetofauna. In: Duellman, W., ed.
 The South American Herpetofauna:  Its Origin,
 Evolution  and Dispersal.   University of Kansas
 Museum Natural Hist. Monograph 7:1-485.

 MacArthur, R., and E. Wilson. 1967.  The Theory
 of Island Biogeography. Princeton, NJ:  Princeton
 University Press.

 Mayr, E.  1963.  Animal Species and Evolution.
 Cambridge, MA: The Becknap Press of Harvard
 University Press.

 Myers,  J.P.   1988.  Remarks to  Climate Institute
 Wildlife Symposium.  Washington, DC; January 21.
Myers, N.   1979.
Pergamon Press.
The Sinking Ark.  New York:
Naveh, Z., and A. Lieberman.  1984.  Landscape
Ecology. New York: Springer-Verlag.

Newmark,  W.D.    1987.   A  land-bridge  island
perspective on mammalian extinctions in western
North American parks. Nature 325:430-432.
Peters, R., and J. Darling.
effect and nature reserves.
       1985. The greenhouse
       Bioscience 35(11):707-
Prance, G.T.  1986.  Tropical Rain Forests and the
World Atmosphere.  AAAS Selected Symposium
No. 101. Boulder, CO: Westview Press, pp. xxi-105.

Regier, H., J. Holmes, J. Meisner.  1987.  Likely
Effects  of Climate  Change  on Fisheries  and
Wetlands,  With Emphasis on  the Great  Lakes.

                                                                                   Biological Diversity
Climate  Institute,  Report  for  the First North
American Conference on Preparing for  Climate

Roberts, L.   1987.   Coral bleaching threatens
Atlantic reefs. Science 238:1228-9.

Robinson, M.H.  1978.  Is  tropical biology real?
Tropical Ecology 19(1):30-50.

Roush, J.  1986.  Private action for public lands.
The Nature Conservancy News 36(4):4-7.

Sibley, T.,  and R.  Strickland.   1985.   Fisheries:
some relationships  to climate change and marine
environmental factors. Chapter 5. In: White, M.R.,
ed. Characterization of Information Requirements
for Studies of CO2 Effects. Washington, DC:  U.S.
Department of Energy.

Strain, B.   1986.  The biosphere and links  to
climate. In: Rosenzweig, C., and R. Dickinson, eds.
Climate- Vegetation Interactions. Proceedings of a
NASA/Goddard  Space Flight Center  Workshop;
January  27-29.     Boulder,  CO:  Office   for
Interdisciplinary   Earth   Studies,   University
Corporation for Atmospheric Research.

Strain, B.   1987.   Direct  effects of  increasing
atmospheric CO2 on plants and ecosystems.  Tree
Terborgh,  J.    1974.    Preservation  of natural
diversity:  the problem of extinction prone species.
BioScience 24(12):715-722.
Tiner, R.Jr.  1984.  Wetlands of the United States:
Current Status and  Recent Trends.  U.S. Fish and
Wildlife Service.  Newton Corner,  MA:  Habitat

Topping, J.C., and J. Bond.   1988.  The Potential
Impact of Climate Change on Fisheries and Wildlife
in North America. Report of the Climate Institute
to the U.S. EPA.

U.S. Department of the Interior. 1987.  National
Park  Service and National  Recreation and  Park
Association.  Trends 24(2):2-43.

U.S. Fish and Wildlife Service.  1981.  Endangered
Means  There's  Still  Time.   Washington,  DC:
Department of the Interior, U.S. Fish and Wildlife

U.S. Fish and Wildlife Service.  1988.  Endangered
Species Technical Bulletin; May.

Wilcox, B.   1982.  Biosphere Reserves and the
Preservation  of Biological Diversity. Towards the
Biosphere  Reserve:    Exploring  Relationships
Between Parks and Adjacent Lands.  Kalispell, MT:
Parks Canada and U.S. National Park Service; June.

Wilson, E.O.  1988.  Biodiversity. Washington, DC:
National Academy Press.

WRI.  1988.  World Resources Institute.  World
Resources Report.  New York: Basic Books.


                                       CHAPTER 9
                              WATER RESOURCES

Higher  temperatures will most  likely result  in
greater   evaporation  and  precipitation;   earlier
snowmelt and reduced water availability in summer;
and, during dry periods, more rapid declines in soil
moisture and water  levels, volumes,  and flows.
Although a general warming and global increase in
precipitation  are  likely,   the  distribution   of
precipitation is highly uncertain and may change in
unexpected  ways.  As  a  result, the frequency,
seasonality,  variability, and spatial distribution  of
droughts, water availability constraints, floods, and
water quality problems  will very likely  change.
Some  regions   could  benefit  from  changing
precipitation patterns, while others could experience
great losses.

Although great uncertainty is  associated with the
projection of future hydrologjc conditions and their
water-use implications, we must be most concerned
about current vulnerabilities to climate extremes
that could  become  exacerbated  under  climate
change.  For  instance, certain dry  regions could
become more vulnerable to drought as a result  of
higher temperatures, earlier snowmelt, and/or shifts
in precipitation.

Impacts on Water Uses

•    If climate in a given region were to  become
     warmer  and drier, water availability would
     decrease and water demand would increase,
     especially demand for irrigation and electric
     power production.

•    Lower   riverflows   resulting   from  drier
     conditions could adversely affect instream uses
     such as  hydropower production, navigation,
     aquatic  ecosystems,  wildlife  habitat,  and

•    Lower streamflow and lower lake levels could
     cause powerplants to shift from once-through
to evaporative cooling.  New plants may also locate
in coastal areas to  obtain a water source that is
reliable and that may be used without violation of
thermal restrictions, although sea level rise could be
a problem. This would have important implications
for land use, transmission lines, and the costs of

•    Where water availability is reduced, conflicts
     among users could increase.  These include
     conflicts over the use of reservoir systems for
     flood control storage, water supply, or flow
     regulation; and conflicts over  water rights
     among agricultural, municipal, and industrial
     users of water  supply.

•    Should extreme flood events become more
     frequent in a river basin as a result of earlier
     snowmelt and increased precipitation, activities
     located in  the  floodplain would endure more
     damages  or  could  require  more  storage
     capacity (whether by construction, reallocation,
     or changes in operating procedures), often at
     the expense of other water uses.

 Policy Implications

     Water management  responses  to  current
vulnerabilities are available  and in use,  and can
appropriately be brought into  play to respond to
changing hydrologic conditions.  These responses
include the following:

•    Build new storage capacity, provided that  the
     structures  show positive net benefits under a
     variety of possible climatic conditions;

•    Modify water  system operations to improve
     performance under extreme conditions,  to
     enhance recovery from  extreme conditions,
     and to accept greater risk to low-valued uses to
     protect high-valued uses; and

•    Encourage a reduction in water demand and
     an increase  in water-use efficiency through

 Chapter 9
      conservation,  water markets, water quality
      control, drought contingency planning, and
      coordinated uses of regional and interstate
      water resources, provided that such measures
      do not reduce the performance and recovery
      capabilities of supply systems.

 Current Status of Water Resources

      The potential  effects of climate  change on
 water resources  must  be examined  within the
 context of the existing and projected supply of, and
 demands for, water.

      The United States is endowed with a bountiful
 supply of water, but the water is not always in the
 right place at the right time, or of the right quality.
 On the average, 4,200 billion gallons per day (bgd)
 of  precipitation  fall on  the  lower  48  states.
 However, a  large  portion of  this water  (66%)
 evaporates, leaving 1,435  bgd (34%) for surface
 water runoff and groundwater recharge. Largely
owing to weather variability, 675 bgd of the 1,435
bgd of runoff water  in  the  coterminous  United
States is considered to be available for  use in 95
years out of 100 (Figure 9-1).

     Surface and groundwaters  are managed by
controlling   and  diverting   flows   through
impoundments and aqueducts; by withdrawing water
for such "offstream" applications  as irrigation  and
municipal use; by regulating flows  to  maintain
"instream" water quality and such uses as navigation,
hydropower,  and recreation;  and  by controlling
flows under flood conditions to avoid loss  of  life,
damage to property, or inconvenience to the public.
Water may  be "withdrawn" and returned  to  the
source more than once,  or "consumed" and  not
returned to the source.

     In 1985, freshwater withdrawals for offstream
uses totaled 338 bgd.  Of the  withdrawals,  92  bgd
were consumed, mostly for irrigation. Withdrawals
and consumption of freshwater by major offstream
uses in 1985 are summarized in Figure 9-1.

     Our  investment  in  water  infrastructure is
substantial.   Water   supply  for  municipal   and
industrial use represented a $108 billion national
investment  in infrastructure  in  1984  (National
Council on  Public Works Improvement,  1988).
Government  agencies  and industries spent $336
                                               IRRIGATION/LIVESTOCK 140
                                              THERMOELECTRIC POWER  131
                                               DOMESTIC/COMMERCIAL  36
                                                  INDUSTRIAL/MINING  31
                                                               CONSUMPTION I  RETURN FLOWS
                       INSTREAM/SUBSURFACE USE
Figure 9-1. Water withdrawals and consumption by offstream uses, coterminous United States. 1985 (Solley et
al., 1988).

                                                                                        Water Resources
billion (in constant 1982 dollars) from 1972 to 1985
(Farber  and Rutledge, 1987)  on water pollution
abatement and control activities.  In other  areas,
excess water  periodically floods  agricultural  and
urban areas, causing annual average damages valued
at $3 billion (in constant 1984 dollars) during the
past decade  (National Council on  Public Works
Improvement, 1988).

      On a  national  scale,  water  supplies  are
adequate, and water availability exceeds withdrawals
and consumption. However, in some regions,  the
gap between demand for water and available supply
is narrow, or the variability in water supply is high,
or both.   For  example,  average surface water
withdrawal exceeds average streamflow in the Great
Basin, Rio Grande, and Colorado River Basins.  In
these water-short basins, offstream water uses often
conflict with instream uses, such as recreation and
maintenance of environmental quality.  Degraded
water quality  further  limits  water availability  in
many regions.  Table 9-1 summarizes the current
status of water supply by major river basin.  The
regions are delineated in Figure 9-2.
                                 Table 9-1.  Current Status of Water Supply

River basin



95% of time0
G round -
New England                 78.4         11.7
Mid-Atlantic                80.7         29.5
South Atlantic-Gulf         233.5         13.5
Great Lakes                 74.3         42.9
Ohio (exclusive of          139.5         22.3
  Tennessee region)
Tennessee                   41.2         22.3

Upper Mississippi (exclusive



of Missouri region)
Mississippi (entire basin)
Souris-Red Rainy
Texas -Gulf
Rio Grande
Upper Colorado
Colorado (entire basin)
Great Basin
Pacific Northwest
"Average renewable supply is defined as the average  flow potentially or theoretically available for use in the
 region; units = billion gallons per day.
bWithdrawals,  consumption,  and reservoir storage are expressed as a percentage of the average renewable supply.
cAs a percentage of average streamflow.
Source: U.S. Water  Resources Council (1978);  U.S. Geological Survey (1984); Solley et al.  (1988).

 Chapter 9
                                                          MISSISSIPPI/ I  n,.TH ATLANTIC-GULF
                                                          -REGION/ | SOU™ซGION 3
                   Figure 9-2. Water resources regions (U.S. Geological Survey, 1985).
     Water   supply  and   use  have   changed
significantly during the past decade.  For the first
time since 1950, when the United States Geological
Survey began recording water withdrawals, national
total fresh and saline  water withdrawals dropped
10% from 1980 to 1985 (from 443 billion gallons to
399 billion gallons, of which 338 billion gallons were
freshwater)  (Solley  et al.,  1988).     Increased
conservation and water recycling  in  agriculture,
industry, and energy production, slower  growth in
energy demand, and decline in availability of new
water supply reduced or tempered water use in all
sectors (Solley et al., 1988).  Withdrawals declined
by 7% in irrigation, by 33% in industry, and by 13%
in thermal power during the same period.  Of the
major users, only municipal/domestic water supply
increased (by 7%).

     The value of instream  uses has risen relative
to  that   of  offstream  uses.    Navigation  and
hydropower have retained  their  importance as
society  has  begun  to  place  greater  value  on
wastewater dilution, ambient water quality, fish and
wildlife  habitats,  and  recreation.  Higher values
on instream uses have made diversion of water for
such applications as agriculture in the West and for
powerplant cooling in the East more difficult.

Climate Change,  Hydrologic Conditions,
and Water Resources

     As  shown in Figure 9-3,  weather  controls
hydrologic conditions through precipitation (mean
and frequency), runoff, snowmelt, transpiration and
evaporation, soil moisture, and the variability of
storms and  drought.  In turn, the ability to use
water resources is greatly influenced by variability in
hydrologic conditions.

     Climate change will affect both the supply of
and demand for water.   Figure 9-4 outlines the
major  potential impacts of global warming  and
changes in precipitation on water resources.

     If climate warms in the United States, there
will likely be greater evaporation and,  in  turn,
greater precipitation; earlier snowmelt and, in turn,
reduced water  availability in summer; and, during

                                                                                       Water Resources
                AND VEGETATION >    l
                  2.800 bgd
I'J 4 200 Dgd /, i j '

'. ijt&dfTjIti'  it


            ^1SB*iaS&Vฐ.!*ซ ปSAU* n6RซUMOJ(AJ
                                                                       bgd=bi)lion gallons per day
Figure 9-3. Hydrologic cycle showing the gross water budget of the coterminous United States (Langbein et al.,
1949; Solley et al, 1983).
dry periods,  more rapid declines in soil moisture
and water levels, volumes, and flows.  Over the very
long term, groundwater availability may be affected
by altered recharge rates.  Transpiration may not
increase as much because increased levels of carbon
dioxide may shrink  the stoma or pores of plants
(Rosenberg,  1988).  Although general warming is
likely to occur, the  distribution of precipitation is
highly uncertain and  may change  in unexpected

     Earlier studies have shown that small changes
in   regional  temperature,   precipitation,   and
evaporation patterns can cause significant changes
in water  availability, especially in arid areas (see
Nemec and Shaake, 1982; Klemes and Nemec, 1985;
Beran, 1986). Precipitation is more variable in arid
                than in humid areas.  In addition, each degree of
                temperature increase causes a  relatively greater
                decline in  runoff  and water availability in arid
                regions as  compared  with  humid regions.    If
                regional climate becomes warmer and drier, more
                vulnerability to interruptions hi water availability
                may be observed.

                     As  a  result,  the  frequency,  seasonally,
                variability,  and spatial  distribution of  droughts,
                water  availability  constraints, flodds,  and  water
                quality problems would probably change. In many
                locations, extreme  events  of dryness and flooding
                could become more  frequent. Some regions may
                experience  more drought  conditions, others more
                flooding, others degraded water quality, and others
                a combination.

Chapter 9
                                                     Climate Change
                          Temperature Increase in
                                All Regions
                                  Regional Weather Variability
               Increased demand
               for air conditioning
Greater evapotranspiration
    Soil moisture loss
    Earlier snowmelt
Less precipitation
 Less runoff and
                                                               Reduced water supply
                                                                  in hotter, drier
              Increased demand for
                 cooling water for
                  electric power
               Increased demand
                  for irrigation
                Increased surface
               water withdrawals
                Increased water
               consumption and

e effects

Conflicts between
off-stream and
in-stream uses

Conflicts b
More precipitation
 More runoff and
                                                  Increased flooding
                                                    in hotter, wetter
                         Increased demand
                             for flood
                                                                                         Conflicts between
                                                                                         flood control and
                                                                                           all other uses
                             policy alternatives
                                        policy alternatives
                     NOTE:  Figure does not trace the impacts of reduced flows and increased evapotranspiration
                              on navigation, hydropower, and the Great Lakes.
                Figure 9-4.  National impacts of climate change on water supply and demand.

     Global warming may have a significant impact
on the demand for water in some regions. Warmer
temperatures may  raise  the  demand for  air-
conditioning in  the South without a proportionate
decrease in demand for electric heat.  Increased
demand for cooling water for electricity powerplants
would result (see Chapter 10: Electricity Demand).
Warmer temperatures  may  also prompt more
farmers  to  irrigate  crops  (see  Chapter   6:

Impacts of Climate Change on Water

      Models of global climate  change do not yet
provide reliable data to predict regional changes in
the  water  supplies;  however, we  can indicate
possible directions of impacts and the water uses
and sectors affected.

      The following sections outline the potential
impacts  of climate  change  on  offstream  and
instream water uses.  The uses most likely to  be
affected are those  currently vulnerable to water
quantity and quality constraints:

    •    irrigation, the major source of withdrawals
         and consumption in the West;

    •   thermal power production, a major source
         of  heat   effluent   and   evaporative
         consumption, especially in the East;
                              Water Resources

    •    instream uses that depend on levels  and
        flows; and

    •    domestic supplies that  are vulnerable to
        hazardous  and toxic substances in ground
        and surface water.

    Table 9-2 highlights the vulnerability of major
water uses in each  region to climate change.


    Irrigation  accounts for 42%  of  freshwater
withdrawals and 82% of freshwater consumption in
the  United  States.   Although irrigated land
comprises about 10% of harvested cropland acreage
nationwide,  it  contributes  30% of the value of
cropland production.   Many of these crops are
fruits, vegetables, and specialty  crops (U.S. Water
Resources Council, 1978; Bajwa et al., 1987).  The
17 western states account for 85% of the irrigated
lands in the country (Bajwa et al., 1987).

    Water-short western  states  are  exploring
numerous    options   for   minimizing   water
requirements.  Because of depleted groundwater
supplies, the rising cost of obtaining groundwater,
and the high cost and limited availability of sites for
new surface water developments, irrigated acreage
has stabilized or is declining in some  areas of the
West (Solley et al., 1988).  Groundwater pumping
for irrigation has  already started to decline in the
      Table 9-2. Potential Regional Impacts of Climate Change on Water Uses:  Areas of Vulnerability

Use Pacific
Irrigation X
Water quality
Flood control X
Hydropower X





Basins Great Plains Great Lakes Mississippi
X X 	 	


x • x







 Chapter 9
 southern  Great  Plains  States  and  in Arizona,
 although  the  impacts on production have  been
 mitigated  by  the  adoption of  more  efficient
 irrigation systems and by a switch to crops offering
 higher returns to water  (Frederick and Kneese,
 1989). In contrast, supplemental irrigation is rising
 in the Southeast, largely because of expansion in
 Georgia (Bajwa et al., 1987).

       Climate   change  may  significantly  affect
 agriculture. Summer drought and earlier runoff are
 likely to change agricultural practices and increase
 demands  for irrigation in most  areas  east of the
 Rocky Mountains.

 Thermal Power Generation

      Steam electric powerplants withdraw almost as
 much freshwater  as  irrigation but consume much
 less  than  irrigation.  Although the  freshwater
 withdrawn to produce the nation's electricity totals
 131 bgd,  only 4.35  bgd  are actually consumed
 (Solley et  al., 1988).

      Future   demand  for  water  for  power
 production  will  depend  on   energy  demand,
 technology, and on federal and state  regulations
 governing instream water quality, instream flow, and
 thermal pollution.  Although a  large  amount of
 installed   capacity exists   along  eastern  rivers,
 freshwater withdrawals by powerplants hi the  East
 have decreased, and siting of plants hi coastal areas
 has increased,  so that by 1987,  30%  of installed
 capacity in coastal areas used saline surface water
 (Solley et  al.,   1988).  In  addition, the thermal
 regulations have caused a shift in the design of new
 cooling systems from once-through cooling, which
 discharges heat back into  the water  sources, to
 evaporative  cooling   with  towers  and  ponds
 (Breitstein  and   Tucker,   1986).     Although
 evaporative cooling alleviates thermal pollution, it
 increases water consumption.

     During droughts, federal and state regulations
 protecting  instream  uses   and limiting  thermal
 discharges   may   constrain  withdrawals    for
 powerplant cooling. In addition, powerplant water
 needs  on  some eastern rivers are  so  large  that
 insufficient water may be available to dissipate heat
 during  low-flow conditions  (Hobbs and Meier,

     Demand for electric power and construction of
new generating capacity may increase  as  warmer
 temperatures raise air-conditioning use (see Chapter
 10:  Electricity  Demand).   If  streamflows  are
 reduced as a result of climate change, powerplants
 using  once-through  cooling could  be adversely
 affected.    Increased  demand  for  power  may
 reinforce  existing trends  in  powerplant  design
 toward evaporative cooling, and hi powerplant siting
 toward coastal locations.  With less water available,
 low-flow conditions may interrupt power production
 and  may increase  power production  costs and
 consumer electricity prices.

 Industrial Uses

      Since 1954, self-supplied industry steadily used
 less and less water per unit of production (Solley et
 al.,  1988).   This  decline was partly  due  to
 efficiencies achieved to comply with federal and
 state water pollution legislation that restricts the
 discharge  of untreated water.   The trend toward
 more efficient  industrial water  uses  is likely  to

      In regions  where flows are reduced,  there
 could be  a reduction in both the quantity and the
 quality of water available  for industrial production.
 In addition, if  the climate becomes  drier, the
 potential for interruption of industrial supply will be

 Domestic Water Uses

      Domestic uses account for 10% of total water
 withdrawn and 11% of consumption.  Over the past
 20 years, domestic water use has increased from 16
 to  25 bgd owing to growth hi  the  number of
 households,  with little  change  hi  usage  per
 household (Solley et al., 1988).

     Most municipal water supply  systems  are
 designed to provide reliable water at all tunes (safe
 yield).   However,  urban  growth depends upon
 developed  water  supply,  which  is  approaching
 exhaustion hi some  areas.   For instance, hi the
 Southeast and parts of the  West, a large percentage
 of municipal water supply comes from groundwater
 (U.S. Water Resources Council, 1978; Solley et al.,
 1988). These regions withdraw more groundwater
 than can be recharged; consequently,  any increased
 drought caused by climate change could accelerate
groundwater mining (see Chapter 14: California and
Chapter 16: Southeast).

                                                                                      Water Resources
      Municipalities in the  West  are purchasing
 irrigators' water rights to ensure  adequate water
 supplies for urban growth. If climate change results
 in reduced municipal supply, this trend will continue
 or  accelerate,  leading  to  the  loss  of irrigated

      In   the   East,  Midwest,   and  Southeast,
 municipalities may be able to increase safe yield by
 repairing  and  replacing existing leaking  water
 delivery systems and  by consolidating fragmented
 water supply districts. These actions could provide
 the margins of safety necessary to  accommodate
 climate change.


      If riverflow  and lake levels became  lower,
 navigation would be  impeded.  Systems that are
 particularly vulnerable are those with unregulated
 flows  or levels  and  high  traffic,  such  as the
 Mississippi River and the Great Lakes. The effects
 of dry conditions and reduced water levels on barge
 traffic on the  Mississippi  in 1988  illustrate the
 potential impacts of climate change.


      Because  of the decline hi water availability
 that could result from climate change, hydropower
 output and reliability, which depend on flows, could
 decline in the  West and the Great  Lakes.  If the
 Southeast became drier, it could face  the same
 problems unless it sacrificed water supply reliability
 to maintain hydropower  production.


       If the Southeast becomes drier, there may be
 an increase in the  conflict among  water uses,
 especially over reservoir releases and levels in the
 Tennessee Valley and the  Lake Lanier, Georgia,
 system.   The  conflicts  are  among  flood  control,
 which relies on storage;  recreation, which depends
' on stable reservoir pool elevations; and downstream
 uses and water supply, which depend on flows.

 Climate Change and Water Quality

       Water quality directly affects the availability of
 water for human  and  environmental  uses, since
 water of unsuitable quality  is not really "available."
 Likewise, water quality in the nation's rivers, lakes,
and  streams depends in part on water quantity.
Water supply is needed for dilution of wastewaters
that  flow into surface and groundwater  sources.
Freshwater inflows are needed to repel saline waters
hi estuaries and  to regulate water temperatures in
order  to   forestall  changes   in   the  thermal
stratification, aquatic biota, and ecosystems of lakes,
streams,  and rivers.

     The Federal Clean Water Act of 1972 and
subsequent  amendments ushered in a new era of
water pollution control.  Massive expenditures for
treatment  facilities  and  changes  in water-use
practices  by government  and  industry  have
decreased  the  amount  of  "conventional" water
pollutants, such  as organic  waste, sediment, oil,
grease, and heat, that enters water supplies. Total
public and private, point and nonpoint, and capital
and   operating  water  pollution abatement  and
control expenditures from 1972 to 1985  totaled $336
billion in 1982 dollars (Farber and Rutledge, 1987).

     Nevertheless,  serious surface  water  quality
problems remain. Groundwater pollution problems,
especially toxic contamination and nonpoint source
pollution, are receiving increased recognition (U.S.
EPA, 1987b).

     One-third  of municipal  sewage treatment
plants have yet  to complete actions to be hi full
compliance with the provisions  of the Clean Water
Act   (U.S.  EPA,  1987a).   Federal and state
regulation  of previously  unregulated toxic  and
hazardous water pollutants has just begun.  In the
West, irrigation has increased the salinity levels in
the  return water and soils of several  river basins
(the lower Colorado, the Rio Grande,  and the San
Joaquin) to an extent that threatens the viability of
irrigation (Frederick and Kneese, 1989).

      Should climate change involve reduced flows,
less freshwater maybe available in some regions for
diluting wastewater salt and heat, especially in low-
flow periods (Jacoby, 1989). Dissolved oxygen levels
in the water would decline while temperature and
salinity levels would increase, affecting the viability
of existing fish and  wildlife.  Increased  thermal
stratification and enhanced algal production due to
higher temperatures may degrade the water quality
of many lakes  (see  Chapter  15:  Great  Lakes;
Blumberg and DiToro,  Volume A).  Finally, the
combination of declining  freshwater availability

 Chapter 9
 and rising sea level would move salt wedges  up
 estuaries,   changing   estuarine   ecology  and
 threatening municipal and industrial water supplies.
 On the other hand, should climate change involve
 increased flows, greater dilution of pollutants would
 be possible in some regions.

      Groundwater is the source for over 63% of
 domestic and commercial use (Solley et al., 1988).
 Although only  a small portion of the  nation's
 groundwater is  thought  to  be contaminated, the
 potential consequences may be significant and may
 include cancer, damage to human organs, and other
 health effects (U.S. Congress, 1984).

      Adequate recharge of aquifers is needed not
 only  to perpetuate supplies but  also  to flush
 contaminants.   Should climate change  result  in
 reduced flows and reduced recharge, the quality as
 well as the available quantity of groundwater could
 be adversely affected.

 Climate Change and Flood Hazards

      Because of  the  buffering  and redundancy
 designed into large structures, major federal flood
 control projects may be able to contain or mitigate
 the  impacts   of  more  frequent  severe  floods.
 However, continued performance for flood control
 may  come at the  expense of other uses.  For
 example, drawing down the levels of reservoirs  to
 contain floodwaters from anticipated increases  in
 precipitation or earlier snowmelt may curtail water
 availability for water  supply.   (This aggravated
 conflict is a  distinct possibility in California, for
 example; see  Chapter 14: California.)

     The major concern with existing dams and
 levees is the consequence of failure under extreme
 conditions. For instance, an increased probability of
 great floods,  whether  due  to  urbanization  of
 upstream watersheds or to climate change, would
 cause  dams  with  inadequate spillways  to  fail.
 (Spillways are  designed  to prevent dam  failure
 through overtopping.)

     The majority of large  dams that provide
 substantial flood storage are in good condition. The
 National Dam Safety  Inventory  shows that the
 overall  condition  of  the U.S.  Army Corps  of
Engineers' more than 300 flood control reservoirs is
sound   (National   Council   on  Public  Works
Improvement, 1988).  In addition, the spillways of
 many large dams are designed to pass a "probable
 maximum flood" (an extreme  flood  event  much
 greater than the 100-year flood).

      Smaller  structures, such as urban drainage
 culverts and  sewers and  local  flood protection
 projects, are currently more susceptible to failure
 and are in poorer condition than large structures
 (National Council on Public Works Improvement,
 1988).  One-third of the non-federal flood control
 dams inspected under the national non-federal dam
 program were found to be unsafe, mostly owing to
 inadequate spillways (National  Council on Public
 Works Improvement, 1988). The capacity of these
 non-federal, smaller, mostly urban flood control and
 stormwater structures is more likely to be exceeded.
 Urbanization upstream from many dams and water
 control structures is already resulting in increased
 impervious surfaces  (such  as  pavement)  and
 increased  peak runoff,  making  some structures
 increasingly vulnerable to failure.

 Climate  Change  and  Conflicts  Among
 Water Uses

     There is no doubt that climate change has the
 potential to exacerbate water availability and quality
 problems and to increase conflicts between regional
 water uses as a result. The foregoing discussion has
 highlighted a number of such conflicts:

    •   conflicts between instream and offstream

    •   conflicts among offstream uses, such as
        agriculture,  domestic  use, and  thermal
        power production;

    •   conflicts between water supply and flood
        control in the West;

    •   conflicts between all uses and recreation in
        the Southeast; and

    •   conflicts between thermal power production
        and instream uses, especially in the East.

    In some areas, increased precipitation due to
climate change could alleviate water quality/quantity
problems   and  conflicts, but  only  after  water
infrastructure  is modified  to  accommodate  the
increased probability of extreme events.

                                                                                Water Resources

    Water resources supply and management occurs
at the regional, river basin,  state, and local levels.
To  be of use to water resources decisionmakers,
climate  change  models  and forecasts need  to
address regional impacts.

   The  regional studies  conducted by the U.S.
Environmental Protection Agency for this document
(see  Table 9-3)  examine the potential  regional
impacts of climate change. (With the exception of
                            Table 9-3.  Regional Water Resource Studies

ซ   Interpretation of Hvdrologic Effects of Climate Change in the Sacramento-San Joaquin River Basin.
    California - Lettenmaier, University of Washington (Volume A)

.   Methods for Evaluating the Potential Impact of Global Climate Change  - Sheer and Randall, Water
    Resources Management, Inc. (Volume A)

.   The Impacts of Climate Change on the Salinity of San Francisco Bay -  Williams,  Philip Williams &
    Associates (Volume A)

Great Lakes

.   Effects of Climate Changes on the Laurentian Great Lakes Levels - Croley, Great Lakes Environment
    Research Laboratory (Volume A)

.   Impact of Global Warming on Great  Lakes Ice Cycles - Assel,  Great Lakes  Environment Research
    Laboratory (Volume A)

•   The Effects  of Climate Warming on Lake Erie Water Quality - Blumberg and DiToro, HydroQual, Inc.
    (Volume A)

.   Potential Climatic Changes to the Lake Michigan Thermal Structure - McCormick. Great Lakes Environment
    Research Laboratory (Volume A)

Great Plains

    Effects of Projected CO^-Induced Climate Changes on Irrigation Water Requirements  in the Great Plains
    States - Allen and Gichuki, Utah State University (Volume C)


•   Potential Impacts of Climatic Change on the Tennessee  Valley Authority Reservoir System -  Miller and
    Brock, Tennessee Valley Authority (Volume A)

•   Impacts on Runoff in the Upper Chattahoochee River Basin - Mains, C.F. Hydrologist, Inc. (Volume A)

•   Methods for Evaluating the Potential Impact of Global Climate Change  - Sheer and Randall, Water
    Resources Management, Inc. (Volume A)

 Chapter 9
 Allan and Gichuki (Volume C), all studies listed in
 Table 9-3 are found in Volume A.) The studies use
 scenarios  generated  from  up to  four  global
 circulation models (GCMs) as thek starting points
 (see Chapter 4: Methodology) and match them with
 regional or  subregional  water resource models.
 This section reviews the findings from the studies on
 California, the Great Plains, the Great Lakes, and
 the Southeast; from previous studies of the impacts
 of climate change on these and other regions; and
 from previous hydrologic studies and  models  of
 individual river basins.

    The  GCMs do not yet  provide  definitive
 forecasts  concerning the  frequency, amount, and
 seasonality  of  precipitation  and  the regional
 distribution of these hydrologic effects (see Chapter
 2: Climate Change; Chapter 3: Variability; Chapter
 4: Methodology; Rind and Lebedeff, 1984; Hansen
 et al., 1986; Gleick, 1987; Rosenberg, 1988).  The
 uncertainty of the forecasts is partially due  to the
 limitations and simplifications inherent in modeling
 complex  natural  and   mamnade   phenomena.
 Modeling efforts are made more difficult by the
 feedbacks and interconnections between changes  in
 temperature; and the amount  and  frequency  of
 precipitation, runoff, carbon dioxide, growth and
 transpiration  of  foliage,  cloud  cover,   ocean
 circulation, and windspeed.

    However, the regional  studies commissioned
 for this  report are a significant step in the effort to
 bring GCM and  regional water resources models
 together to examine the regional impacts of climate

 The West

    The arid and semiarid river basins west  of the
 Mississippi  River  have   significant  surface  and
groundwater quantity and quality problems and are
vulnerable to  restricted water availability.   Total
water use exceeds average streamflow in 24 of 53
western  water  resource  regions  (U.S.  Water
Resources Council,  1978), with the majority  of the
West's  water  withdrawals   going  to  irrigation.
Surface  and groundwater quality in the West have
 deteriorated  as  a  result   of  low  flow,   salts
concentrated by irrigation, and pesticide use. The
West also depends uponnonrenewable groundwater
supplies for irrigation (Solley et al., 1988).
    Climate change may exacerbate water shortage
 and  quality  problems  in the  West.    Higher
 temperatures could cause  earlier  snowmelt  and
 runoff, resulting in lower water availability in the
 summer. Some GCM scenarios predict midsummer
 drought and heat, less groundwater recharge, and
 less groundwater and surface water availability for
 irrigation in the  middle latitudes of the country.
 The sensitivity analyses conducted by Stockton and
 Boggess (1979) indicated that a warmer and drier
 climate would severely  reduce the quantity and
 quality of water in arid  western river basins (Rio
 Grande,  Colorado,  Missouri,  California)   by
 increasing water shortages. Water shortages and
 associated conflicts between instream and offstream
 uses,  between agricultural and urban/industrial
 water uses,  and between flood control and other
 water  uses  of reservoirs may  be expected under
 these scenarios.  Hydropower output also would
 decline as a result of lower riverflow.

 Pacific Northwest

    The competition  for   water  for  irrigation,
 hydropower, and fisheries  habitat is increasing in
 the Pacific  Northwest  (Butcher and  Whittlesey,
 1986).  Climate change  may alter the seasonality
 and  volume of   precipitation  and   snowmelt,
 increasing the risk of flooding, changing reservoir
 management practices, and  affecting the output and
 reliability of hydroelectric power production and the
 availability of water for irrigation.


    The diversion of water from water-rich northern
 California and from the Colorado River to southern
 California via federal and  state systems of dams,
 aqueducts, and pumping stations has transformed
 California into the nation's leading agricultural state
 and has made possible the urbanization of southern
 California.   Irrigation accounted for 83% of the
 total value of California's agricultural output in 1982
 (Bajwa et al., 1987). Because of this high economic
 dependence  on water in an arid area, southern
 California is vulnerable to droughts and any altered
 temporal pattern of runoff  that may be  caused by
 atmospheric warming.

    Total  annual  runoff  from  the  mountains
surrounding  the Central Valley is  estimated  to
increase slightly under GCM scenarios, but runoff
in the late  spring and summer may be much less

                                                                                    Water Resources
than today  because  higher  temperatures cause
earlier snowmelt (Lettenmaier, Volume A).  The
volume of water from the State Water Project may
decrease by 7 to 16% (see Chapter 14: California;
Sheer,  Volume A). Existing reservoirs do not have
the capacity to increase storage of winter runoff and
at the same tune to retain flood control capabilities.
In addition,  flows  required to  repel saline water
near the major freshwater pumping facilities in the
upper  Sacramento-San Joaquin River Delta  may
have to be doubled  as a result of sea  level  rise,
further reducing  water  available   to  southern
California (Williams, Volume A).

    Decreases in water availability may also reduce
hydroelectric power produced in California. In the
1976-77  drought,   hydroelectric  production  in
northern California dropped  to less  than  50%  of
normal, a deficiency relieved  by importing surplus
power  from  the Pacific Northwest and by  burning
additional fossil fuels at an approximate cost of $500
million (Gleick, 1989).

Colorado, Rio Grande, and  Great Basins

    Total consumption  is more  than 40%  of
renewable  supply  in these  river basins.   The
Colorado River Basin has huge reservoir  storage,
but demand exceeds supply in the lower half of the
basin.  Ordinarily all of the Colorado  River's water
is consumed before it reaches the Gulf of California
in Mexico. The Colorado River Compact of 1922,
the 1963  Supreme Court decision  in Arizona  v.
California, the  treaties with  Mexico  of 1944 and
1973, and other agreements allocate Colorado River
water to seven  states and Mexico (Dracup, 1977).
Some studies show that the Upper Colorado region
will  use all of its allocation by the year 2000,
reducing water hitherto available to lower Colorado
and California (Kneese and Bonem, 1986).

    Climate  change may   further   reduce  the
availability of water  in these  basins.  A model  by
Stockton and Boggess (1979)  of a 2ฐC temperature
increase and a 10% precipitation decrease shows
decreases in the water supply in the upper Colorado
and the Rio Grande  of 40 and 76%, respectively.

Great Plains

    The southern  Great Plains States of  Kansas,
Nebraska, Oklahoma, and  Texas produce almost
40% of the nation's wheat, 15%  of  its corn, and
50%  of its  fattened cattle  (see Chapter 17: Great
Plains).    The  region  heavily  depends   on
groundwater mining (when pumping exceeds aquifer
recharge) for irrigation.  The region was severely
affected during the "Dust Bowl" years of the 1930s
and suffered from severe drought in 1988.
    Because of the greater reliability in irrigated
yields relative to dryland yields, the demand  for
irrigation could rise (Allen and Gichuki, Volume C;
Adams  et  al., Volume  C).   Thus,  while total
agricultural  acreage  could  decrease,  irrigated
acreage and groundwater mining may increase in
the southern Great Plains. Greater demand may be
placed on the Ogallala Aquifer, which underlies
much of the region, causing further mining of  the

Great  Lakes

    Based on analyses for this report (Croley and
Hartmann, Volume A), higher temperatures may
overwhelm any increase  hi precipitation and may
evaporate lakes to below the  lowest  levels  on
record.    However,  changes  in  Great  Lakes
evaporation under climate  change  are  highly
uncertain  and  depend   on   such variables  as
basinwide precipitation, humidity, cloud cover, and
windspeed. Under a possible set of conditions, lake
levels could rise.  The winter ice cover would be
reduced but would still  be present, especially in
shallow areas  and northern lakes (Assel, Volume
A).  Navigation depths,  hydropower  output, and
water quality all would be adversely affected,  but
losses of existing shorelands from erosion would be
reduced as a result of lower lake levels (see Chapter
15: Great Lakes).

Mississippi River

    The Mississippi River historically  has been
affected by both spring floods and drought. In 1988,
low flows  due  to  drought received national
attention.   Low flows disrupt navigation,  permit
saltwater  intrusion into  the  drinking  water of
southern Louisiana cities, reduce  the  dilution of
contaminants transported from upstream locations,
and reduce  the  inflow  of  water  to the  vast
Mississippi Delta wetlands (see Glantz, Volume J).


    Although  the  Northeast is humid,  cities and
powerplants demand  large  amounts  of water at

Chapter 9
localized points in a watershed, necessitating storage
and  interbasin transfers.   Because of the small
amount of storage in the Northeast, the region is
vulnerable to prolonged drought.  No new major
storage has been built in the Northeast during the
past 20 years, except the Bloomington Dam on the
Potomac  River.   Water  supply  in lower  New
England, New York, and Pennsylvania,  and power
production in the Northeast, remain vulnerable to
drought,  which  may  occur  more  frequently
(Schwartz, 1977; Kaplan  et  al.,  1981).   During
periodic droughts in the Northeast, such as those in
1962-65 and 1980-81, instream flow regulations
ration water and threaten shutdowns of electrical
powerplants (U.S. Army Corps of Engineers, 1977;
Schwartz,  1977; Kaplan et al., 1981).


    In the Southeast, the experience with drought
in recent years is increasing the use of groundwater
and  surface  water for irrigation and is prompting
farmers to consider shifting crops.  In Georgia, for
instance, the use of groundwater for irrigation has
grown quite  rapidly.  However, the GCMs disagree
on whether  the Southeast may become wetter or
drier (see Haines, Volume A;  Miller and Brock,
Volume A).  Most reservoirs in the  area have
sufficient  capacity to retain flood surges and to
maintain navigation, hydropower, water supply, and
instream uses (e.g.,  dilution,  wildlife) under both
wetter  and  drier  conditions  (see  Chapter  16:
Southeast;  Sheer  and  Randall,  Volume  A).
However,  drier conditions would pose  conflicts
between recreational uses (which would be hurt by
changes in reservoir levels) and all other instream
and  offstream uses.

    Should the Southeast become drier, a decline
in the inflow of freshwater could alter the estuarine
ecology of  the  gulf  coast, which may be  most
vulnerable to sea  level  rise  (see  Chapter  16:

    Decreases  in water availability and  quality,
increased  risk   of   flood  damages,  and  the
exacerbation of  conflicts  between water  users
competing for an increasingly scarce or difficult to
manage resource are the major potential impacts
of a global warming trend on the nation's water
resources.  How will we manage water resources
given the possibility of change and uncertainties
about its nature and timing?

    Policy approaches to water resources may be
grouped  under supply  (or  structural)  approaches
and demand (or nonstructural) approaches. Supply
approaches  mitigate  hydrologic  variability  and
climate   change;   demand   approaches   modify
behaviors that create vulnerability to such change.
For example, water shortages may be addressed
either by developing surface water storage capacity
and improving the quality of water from available
sources (supply approaches), or by decreasing water
use and consumption (a demand approach).

    Many of the policy approaches discussed below
have been recommended by water resource experts
for 20 years and are in use to address existing water
problems and vulnerabilities.  The  potential of
climate   change  provides  another  reason   for
expanded use of these approaches.

Supply and Structural Policy Approaches

    The supply-related  policy  approaches to water
resources include  design for  uncertainty,  surface
water development,  and optimization of water
resource systems.

Design for Uncertainty

    Most water resource decisions in the past have
been based on the assumption that the climate of a
region varies predictably around a stationary mean.
Water managers  develop  water resources plans
based   on   statistical   analyses  of  historical
climatological and hydrologic  data. However,  the
frequency of  extreme  events, which  has  been
assumed to be fixed or to be  modified only by the
urbanization  of  watersheds,  may be  changed
significantly by altered climatic conditions.

    In addition to being uncertain about hydrologic
conditions,  we  are   uncertain  about   future
demographic,  economic, and institutional factors
that affect offstream  water uses and  social and
economic values attached to instream uses.  As an
example, water withdrawals  in 1985 declined overall
from 1980, falling far  short  of projections made
starting in 1960 and as recently as 1978 (Solley et
al., 1988).

                                                                                      Water Resources
    Finally, we  are uncertain  about how  our
economic, regulatory, and institutional systems will
respond  to climate  change  in  the absence of
concerted  governmental action.   It  would be a
mistake to attempt to project the impacts of climate
change simply by superimposing projected future
hydrologic  conditions on today's social systems.

    The planners and designers of water resources
must  address such uncertainties.  Three  types of
response are often used to address conditions of
great  uncertainty:

    •    Avoid  inflexible,  large-scale, irreversible,
        and high-cost measures;  opt for shorter
        term,  less capital-intensive, smaller scale,
        and incremental measures.

    •    Conduct sensitivity analysis and  risk-cost
        exercises in the design of structural  and
        management  systems  to  address  the
        potential range of climate change impacts.
        Sensitivity analysis describes the sensitivity
        of projections to variables affecting their
        accuracy; risk-cost  analysis  identifies the
        costs, for various conditions other than
        those    projected,   associated  with
        underdesign or overdesign of a structure.
        The consideration of hydrologic extremes
        and the use of risk analysis in the design of
        specific projects  to  mitigate the adverse
        consequences of hydrologic variability may
        incidentally  mitigate many of the physical
        impacts of climate change (Hanchey et al.,

    •    Design structures and systems  for rare
        events.  Matalas and Fiering  (1977) found
        that many large systems have  substantial
        redundancy  (margins  of   safety)   and
        robustness  (ability to perform  under  a
        variety of conditions) that enable them to
        adapt technologically and institutionally to
        large stresses and uncertain future events.

    Although  the principle of  design for rare
extremes may provide robustness, it has a cost and
may conflict with the principle of maximizing the
economic return from a project.  Most public and
private water  developers  subject projects to  "net
present value" or "internal rate or return" analyses.
These analyses discount future benefits relative to
present benefits.  If a high discount rate is used in
decisionmaking, conditions beyond 10 or 20 years
may have little impact  on design and investment
decisions (see Chapter 19: Preparing for a Global
Warming; Hanchey et al., 1988).

Surface Water Development

    Surface water structures increase developed or
available water supply, provide for the regulation of
flows  for  instream uses,  prevent  flooding,  or
perform some combination of these functions. These
structures include  dams,  reservoirs, levees,  and
aqueducts. Because of high costs of construction,
adverse impacts on the environment, the  limited
number of sites available for  new structures,  and
opposition by citizen groups, the trend during the
past   decade  has  been  away  from  large
excess-capacity, capital-intensive projects. Only the
Central Utah Project  and the Central Arizona
Project have  gone forward in recent years. Only
one major project  in  the Northeast  has been
completed in past 20 years: the Bloomington Dam
on the Potomac River.  In 1982, California citizens
voted down funds for the proposed Peripheral Canal
that would have permitted increased diversion of
water from north to south in the state. In addition,
the  national  trend  toward increased local/state
financing and reduced federal financing for projects
has  reduced  funds available  for  large  projects
(National  Council on Public Works Improvement,

    These  current   trends in  water  resources
management may be reevaluated in light of possible
new demands for developed water caused by climate
changes. Pressure to build proposed projects such
as the Narrows Project in Colorado, the Garrison
Diversion in North Dakota, the Peripheral Canal in
California, and  structures to  divert water from
northern   New   England   to   southeastern
Massachusetts may be renewed if droughts reoccur
or demand increases.  The pace at which existing
projects are upgraded, modified, or expanded may
also accelerate.

Optimization of Water Resource Systems

    Water resources can be managed to maximize
the water availability from a given resource base
such as a dam, watershed, or aquifer. Adoption of
systemwide strategies for a large-scale water system
may allow for substantial operating flexibility related
to releases of stored water. This flexibility can have

Chapter 9
an enormous influence on the overall performance
and resilience (recovery abilities) of the system, and
may provide additional  yields  that mitigate the
impacts of climate change. For example, the U.S.
Department of the Interior's Bureau of Reclamation
(1987) is adopting operational, management,  or
physical changes to gain more output from the same
resources. Water management agencies nationwide
are implementing methods to protect groundwater
recharge areas  and to  use  ground  and surface
waters conjunctively (U-S. EPA, 1987b). Watershed
management practices also affect water supply; for
example, water yields can be significantly affected by
timber harvest practices.

    In the East, consolidation  of or coordination
among fragmented urban water supply authorities
can achieve economies of scale in water delivery,
decrease the risk of shortage  in any one subsystem
within a region, increase   yields,  and  provide
effective  drought management procedures. Sheer
(1985) estimated that coordinated water  authority
activities in the Potomac River basin eliminated the
need for new reservoirs, saving from $200 million to
$1 billion.

    River basin and aquifer boundaries in many
cases traverse or underlie portions of several states.
Regional  and interstate cooperation to  manage
water resources  has a long tradition in some U.S.
river basins. Although numerous opportunities exist
for additional coordination of water management
between states, within basins, or between basins, the
agreements required for regional  compacts and
operating procedures and sharing of water supplies
may require substantial and lengthy negotiations.

    Several  interstate   water   authorities   have
significant water allocation authority. For example,
the Delaware River Basin Commission allocates
water to users in the Delaware Basin and transfers
it to New York City under authority of a  1954
Supreme Court  ruling (347 U.S. 995) and federal
legislation, which established the Commission in
1961 and granted it regulatory, licensing, and project
construction powers. Similarly, water authorities in
the Washington,  D.C., metropolitan area operate
Potomac River water supply projects as integrated
systems  under  a 1982  agreement.    Both  the
Delaware and Potomac regional compacts include
provisions for drought allocations. (See Harkness et
al., 1985, for management  actions taken  by the
Delaware River Basin Commission during a 1984-85
Demand Management and Nonstructural
Policy Approaches

    Demand-related   adaptations   encourage   a
reduction in water demand and an increase in water
use efficiency through pricing, market exchange of
water  rights,  conservation,  protection of water
quality, education and extension service assistance,
technological innovation, and  drought management
planning. Policies that discourage activities in flood-
prone areas are the nonstructural counterparts for
reducing flood damage.

Water Pricing. Water Markets, and Water

    In the past, many people considered that water
was too essential a resource  or too insensitive to
price  to allow market forces to allocate its use,
especially during shortages. Policy took the form of
direct controls and appeals to conserve (Hrezo et
al., 1986).  In recent years, greater attention has
been   given  to   market-based   policies  and
mechanisms that allocate  limited water supplies
among   competing  uses  and  promote  water

    Water prices that reflect real or replacement
costs and the exchange of water rights by market
mechanisms can promote conservation and efficient
use. Since water use is sensitive to price (Gibbons,
1986) water users faced  with higher  prices will
conserve water and modify their technologies and
crop  selection to  use   less  without  substantial
reduction in output.  If there  is a market for water
rights, those willing to  pay  more  may  purchase
rights  from those  less,  willing to  pay.   As  a
consequence, water  will  be  transferred out  of
marginal uses and will be conserved.

    Three   related  pricing   and   conservation
approaches are irrigation conservation, municipal
and industrial water use,  and water markets and

Irrigation Conservation

    Relatively small reductions in irrigation demand
can make  large amounts  of water available for
urban and industrial uses. For instance, nearly 83%
of the withdrawals and 90% of the consumptive use
of western water is for irrigation. A  10% reduction
in irrigation use would  save 20 million  acre-feet
(maf) in water withdrawn and 10  maf  in water

                                                                                      Water Resources
consumed annually, effectively doubling the water
available for municipal and industrial uses in the
West (Frederick,  1986).   (For  comparison, the
average annual flow of the Upper Colorado River
Basin is 15 maf.)

    Inexpensive water was a key factor in the
settlement  of  the West and the  expansion of
agriculture  (Frederick, 1986).   The Bureau of
Reclamation was established early in this century to
promote the development of irrigation hi the West.
The Bureau provides irrigation for about 11 million
acres, more  than one-fifth of the total  irrigated
acreage.  Since the  Bureau  accounts for nearly
one-third of all surface water  deliveries and about
one-fifth of total water deliveries  hi the 17 western
states, actions by the Bureau to use this water more
efficiently have  an impact  throughout the West
(Frederick, 1986).

    In the past, demand for Bureau water was not
based on the real cost of the water, because more
than 90% of the Bureau's irrigation projects have
been subsidized, and payments on some projects no
longer even  pay for  operation and  maintenance
(Frederick and Hansen, 1982). Irrigators fortunate
enough to receive such inexpensive water may have
little or  no incentive  to  conserve.  However, the
Bureau's more recently  stated objectives include
revising  their water marketing policy, promoting
conservation, and  pricing water to reflect its  real
cost (U.S. Department of the Interior, 1987).

Municipal and Industrial Water Use

    Municipalities  throughout  the  country  are
finding it difficult  and expensive  to augment their
supplies  to meet the  demands of population and
economic growth and are finding that users would
rather use less  than  pay more  (Gibbons, 1986).
Traditional average-cost pricing provides adequate
service to customers and adequate returns  to water
companies, but is being reevaluated because it tends
to cause overinvestment in system capacity (U.S.
Congress, 1987).   Marginal-cost  pricing (charging
for the cost of the last-added and most expensive-
increment  of supply)  or  progressive-rate pricing
(charging more per unit to users of large amounts)
can   reduce  domestic   and  industrial  water
consumption because water use is sensitive to price
(Gibbons, 1986).
Water Markets and Transfers

    The "first  hi time, first hi right"  appropriation
doctrine, which favors the longest standing water
rights, governs much of the West's  surface water
and some groundwater. The appropriation doctrine
has the potential  to establish clear, transferable
property rights  to water  —  a precondition for
effective operation of water markets.  The potential
for water transfers to the highest value users has not
yet been  fully realized  because the nature and
transferability of the  rights are  obscured by legal
and  administrative  factors   (Trelease,  1977;
Frederick, 1986; Saliba et al.,  1987).  Following are
some examples:

    •    Rather  than grant  absolute  ownership,
        states with prior appropriation rules grant
        rights to use water for beneficial purposes.
        Water rights not put  to beneficial use may
        be forfeited.  This encourages a use-it-or-
        lose-it attitude.

    •    Federal and Native American water rights
        remain unquantified hi some areas such as
        the Colorado River Basin.

    •    The emergence hi law of the "public trust
        doctrine,"  which states  that all uses are
        subject to the public interest, has cast  a
        cloud over some water rights.   This has
        been true  in California, where  the public
        interest  has  driven  a  reexamination  of
        withdrawals  from Mono Lake, and where
        existing  permits have been modified  to
        protect the Sacramento-San Joaquin Delta
        from  saltwater  intrusion.   Montana is
        increasinglybasing water management plans
        on its instream flow requirements and is
        exploring ways to have these requirements
        for all future beneficial instream uses count
        as a bona fide use of the Missouri River to
        slow the growth rate of water diversion for
        offstream uses (Tarlock, 1987).

    •    In resolving  interstate water disputes,  a
        federal   common   law  of   "equitable
        apportionment" has developed under which
        an   informed   judgment,   based   on
        consideration of many factors, secures  a
        "just  and equitable" water allocation (see
       Strock,  1987).   The   Supreme Court

Chapter 9
        decided in Colorado v. New Mexico (456
        U.S.   176,   1982)   that   equitable
        apportionment may be used to  override
        prior appropriation priorities in  cases of
        major flow reductions. The Supreme Court
        specifically mentioned climatic conditions in
        ruling that prior  appropriation systems
        would otherwise protect arguably wasteful
        and inefficient uses of water at the expense
        of other uses (see Strock, 1987).

    •   Because   of  imperfect   competition,
        third-party   effects,  uncertainty   over
        administrative   rules,  and   equity
        considerations, water market prices may
        not  appropriately  measure water  values
        according to economic efficiency criteria
        (Gibbons, 1986; Saliba et al., 1987).

    •   It  is  possible  to  control  groundwater
        withdrawals, but for a number of reasons it
        is difficult to establish market mechanisms
        for groundwater  allocation.  Because all
        groundwater users essentially draw from a
        shared  pool, groundwater resources are
        treated as "common properly." As a result,
        property rights  are difficult  to define,
        third-party  impacts  of  transfers  of
        groundwater rights are  significant,  and
        interstate agreements concerning allocation
        of interstate aquifer water are difficult to
        attain (Emel, 1987).

    Despite the obstacles, transfer of water rights
among  users  —  especially  from  irrigators  to
municipalities and power companies seeking water
for urban expansion and electricity production — is
becoming common in many western  states  (Wahl
and Osterhoudt, 1986; Frederick, 1986). Methods
include negotiated purchases, short-term exchanges
during droughts, and  water  banks and  markets
(Wahl and Osterhoudt, 1985; Saliba et al., 1987;
WaM and Davis, 1986).

    Legislation  in  many  western   states  has
facilitated   water  transfers   (Frederick,  1986;
Frederick  and  Kneese,  1989).    For instance,
Arizona's new water law facilitates the purchase of
agricultural land for water rights, and the use of that
water  for  urban development.   Strict technical
standards imposing conservation on municipal and
industrial water uses, such as watering golf courses
with wastewater, are also part of Arizona's laws
(Saliba et al.,  1987).
    Frederick and Kneese (1989) caution that water
transfers occur gradually and are not likely to affect
more than a small percentage of agricultural water
rights for the foreseeable future.  However, legal
and institutional changes facilitating water markets
and demand  for water by high-value users may be
accelerated  under  the stress of  climate  change
(Trelease, 1977).

Drought Management Policies

    Integrating drought planning into water resource
management may assume greater priority if climate
change  aggravates water shortages.  The  Model
Water Use Act (Hrezo et al., 1986) advocates that
states or water supply authorities integrate drought
management  and  advance  planning into  their
policies by designating a governmental authority for
drought response and by adopting mechanisms for
automatically implementing and enforcing water-use
restrictions.    In  1986, only  seven  states had
comprehensive  management   plans  for   water
shortages (Hrezo et al., 1986). Most states rely on
water rights appropriations, emergency conservation
programs, and litigation to  allocate water during
shortages.     Improved  capabilities   in  surface
hydrology and  in  water system  modeling and
monitoring would be required to support broadened
drought contingency planning.

Water Quality

    Federal and state legislation and regulations for
control  of instream  water quality have  had  a
dramatic  effect  on  reducing  conventional  water
pollutants since  the enactment of the 1972  Clean
Water Act. The reduced riverflows and lake levels
that are possible under altered climate conditions
could necessitate more stringent controls on point
and nonpoint sources  to  meet  water  quality
standards.   Promotion of nonpolluting  products,
waste minimization, and agricultural practices that
reduce  the   application of  chemicals will  also
enhance water  quality,  making more water  of
suitable quality available for  use.

    Many states  have  adopted measures to protect
instream water uses. These include reserving flows
or granting rights for  particular instream uses and
directing agencies to review impacts before granting
new rights (U.S. Water Resources Council, 1980;
Frederick and Kneese, 1989). Regulations limiting

                                                                                     Water Resources
water use may have to be modified where climate
change  has  resulted  in  reduced  flows  during

Policies for Floodplains

    The National Flood Insurance  Program was
enacted in 1968, with major amendments in 1973.
The program provides subsidized  flood insurance
for existing structures in flood-prone areas, provided
that the community with jurisdiction regulates the
location  and construction  of new buildings  to
minimize future flood losses.  New structures that
comply  with the  restrictions  are  eligible  for
insurance at full actuarial rates.

    In 1979,  the program took in $140 million in
premiums  and  paid  $480  million  in claims.
Recently, the program was authorized to relocate
structures exposed to  repeated flood or erosion
damage rather than pay claims for such structures.

    Where  rainfall  and flooding increase,  the
100-year floodplain would expand, and rate maps
would need revision.  Premium payments and claims
would rise.

    Water is the  principal  medium  by which
changes in atmospheric conditions are transmitted
to the  environment,  the economy, and  society.
Hydrology is the key discipline that enables us to
understand and project these effects. Improvements
in both the GCMs and regional hydrologic models
are needed so that we may understand the impacts
of climate change  and devise appropriate water
resources management  strategies.   Specifically,
GCMs do not yet provide regional forecasts at the
level  of  certainty  and  temporal  and   spatial
resolution required for decisionmakers. To be more
helpful, the GCMs should provide forecasts specific
to individual  river basins or demand centers, and
should  describe hydrologic conditions  over  the
typical design-life of water resource structures.
    Research activities should include the following:

    •   Monitor   atmospheric,   oceanic,   and
        hydrologic conditions to detect evidence of
        water resources impacts of climate change.

    •   Continue to develop and refine regional
        hydrologic  models  that  are capable  of
        modeling  the  changes  in runoff,  water
        availability,  water  use,  and  evapotran-
        spiration   induced  by   changes   in
        temperature and  atmospheric conditions.
        This research should focus on vulnerable
        river basins where demand approaches or
        exceeds safe yield  or where hydrologic
        variability is high.

    •   Refine global climate change models and
        link them to regional hydrologic models so
        that  regional  water  resource  planners,
        engineers, and  managers can  use their
        projections more confidently.

    •   Study  the  sensitivity of existing  water
        systems to possible changes  in  climate

    At the same  time,  the  following research is
needed  to  identify opportunities   for adopting
measures to adjust and adapt to climate change.

    •   Quantify  federal  and  Native American
        water rights in the West.

    ซ   Examine  how  present  institutions  and
        markets can better  allocate  water among
        users  and provide incentives to conserve

    •   Assess the  extent  to which laws  and
        regulations may exacerbate the  effects  of
        climate change. (Examples include thermal
        controls for rivers and federal pricing and
        reallocation policies  for irrigation water.)

    •   Identify,   project,  and  quantify  the
        demographic and institutional adjustments
        that may occur in the  absence of public
        action  in response to  climate-induced
        impacts on water resources. This research
        will reduce uncertainty for  policymakers
        regarding where concerted public action
        may be or not be needed.

Chapter 9


Bajwa, S., M. Grosswhite, and J.E. Hostetler. 1987.
Agricultural   Irrigation   and   Water  Supply.
Agriculture  Information  Bulletin  532.    U.S.
Department of Agriculture. Washington, DC: U.S.
Government Printing Office.

Beran, M.   1986.  The water resource impact of
future climate change and variability.  In: Titus, J.,
ed. Effects of Changes in Stratospheric Ozone and
Global Climate. Vol. 1: Overview. Washington, DC:
U.S. Environmental Protection Agency, pp. 299-330.

Breitstein, L., and R.C.  Tucker.  1986.  Water
Reuse and  Recycle  in the U.S. Steam Electric
Generating Industry  - An Assessment of Current
Practice  and  Potential for Future Applications.
Report to U.S. Geological Survey, Grant No. 14-08-
0001-G1074.  Bethesda, MD: Dames and Moore.

Butcher, W.R.,  and  N.K. Whittlesey.   1986.
Competition between irrigation and hydropower in
the Pacific Northwest.   In:  Frederick, K.D., ed.
Scarce   Water   and   Institutional  Change.
Washington, DC: Resources for the Future

Dracup, JA. 1977. Impact on the Colorado River
Basin and  southwest  water  supply.   In:   The
National Research Council, ed. Climate, Climatic
Change,  and  Water Supply.   Washington, DC:
National Academy of Sciences, pp. 121-132.

Emel, J.L.   1987.  Groundwater rights:  definition
and transfer. Natural Resources Journal 27:653-674.

Farber, K.D.,  and  G.L. Rutledge. 1987.  Pollution
abatement and control expenditure, 1982-85. Survey
of Current Business  May:21-26.

Frederick, K.D., and J.C. Hansen. 1982. Water for
Western Agriculture. Washington, DC: Resources
for the Future.

Frederick, K.D.  1986.  In: Frederick, K.D., ed.
Scarce  Water   and    Institutional  Change.
Washington, DC:  Resources for the Future.

Frederick,   K.D.,  and  A.V.  Kneese.    1989.
Reallocation by markets and prices. In: Waggoner,
P.E.,   ed.    Climate Change  and  U.S.  Water
Resources.  New York: John Wiley and Sons.  In
Gibbons, D.C.   1986.   The Economic Value of
Water. Washington, DC: Resources for the Future.

Gleick,  P.H.     1987.     Regional  Hydrologic
Consequences of Increases in Atmospheric CO, and
Other Trace Gases. Climatic Change 10:137-161.

Gleick,  P.H.    1989.    The  sensitivities   and
vulnerabilities of water supply systems to climatic
changes.   In:   Waggoner, P.E.,  ed.    Climatic
Variability,  Climate  Change,  and U.S.  Water
Resources. New York:  John Wiley and Sons.  In

Hanchey,  J.R.,  K.E.  Schilling,  and  E.Z.  Stakhiv.
1988.  Water  resources planning under climate
uncertainty.  In:  Preparing for Climate  Change.
Proceedings  of   the   First   North   American
Conference  on Preparing  for  Climate  Change.
Rockville, MD: Government Institutes, Inc.   pp.

Hansen, J.; A. Lacis, D. Rind, G. Russell, I. Fung,
P. Ashcraft, S. Lebedeff, R. Ruedy,  and P. Stone.
1986. The greenhouse effect: projections of global
climate  change.   In: Titus, J.,  ed.  Effects of
Changes   in  Stratospheric Ozone  and  Global
Climate. Vol. 1: Overview.  Washington, DC: U.S.
Environmental Protection Agency, pp. 199-218.

Harkness, W.E., H.F. Lins, and W.M. Alley. 1985.
Drought in the Delaware River Basin. In: National
Water Summary.   1985:   Hydrologic Events  and
Surface-Water Resources.  Washington, DC:  U.S.
Geological  Survey,  U.S.   Government  Printing
Office, pp. 29-40.

Hobbs, B.F., and P.M. Meier. 1979. An analysis of
water resources constraints on power plant siting in
the Mid-Atlantic States.  Water Resources Bulletin

Hrezo, M.S., P.G. Bridgeman, and W.R. Walker.
1986.  Integrating drought planning  into  water
resources management. Natural Resources Journal

Jacoby, H.D. 1989. Water Quality. In: Waggoner,
P.E., ed.   Climate  Change  and U.S. Water
Resources.  New York:  John Wiley and Sons.  In

                                                                                     Water Resources
 Kaplan,  E., M. Rubino, and D.  Neuhaus.  1981.
 Water Resource Development and Energy Growth
 in the Northeast.  Brookhaven, NY: Brookhaven
 National Laboratory, Report BNL 51522.

 Klemes,  V., and J. Nemec.  1985.  Assessing the
 impact of climate change on the development of
 surface  water  resources.  In:   Klemes  V.,  ed.
 Sensitivity of Water Resource Systems to  Climate
 Variations. Geneva: World Meteorological Office.

 Kneese,  A.V., and G.  Bonem.  1986. Hypothetical
 shocks  to water  allocation  institutions  in  the
 Colorado Basin.  In:   Weatherford,  G., and F.L.
 Brown, eds. New Courses for the Colorado River:
 Major Issues for the Next Century. Albuquerque:
 University of New Mexico Press, pp. 87-108.

 Langbein, W.B. et al.  1949. Annual runoff in the
 United States.    Geological Survey Circular  5.
 Washington, DC:  U.S. Department of the Interior.

 Linsley,  R.K., and J.B. Franzini.   1979.  Water
 Resources Engineering. New York: McGraw-Hill,

 Matalas,   N.C.,   and   M.B.   Fiering.     1977.
 Water-resource systems planning. In:  The National
 Research Council, ed.  Climate, Climatic Change,
 and  Water Supply.  Washington,  DC:  National
 Academy of Sciences,  pp. 99-110.

 National Council on Public Works Improvement.
 1988. Fragile Foundations:  A Report on America's
 Public Works. Final Report to the President and to
 the Congress. Washington, DC: U.S. Government
 Printing Office.

 National   Research  Council.    1977.   Climate,
 Climatic  Change, and Water Supply.  Washington,
 DC:  National Academy Press.

 National  Research Council.  1988.    Estimating
 Probabilities of Extreme Floods. National Research
 Council,  Committee on Techniques for Estimating
 Probabilities of Extreme Floods, Water Science and
 Technology Board; and Commission  on Physical
 Sciences, Mathematics, and Resources. Washington,
 DC:  National Academy Press.

 Nemec, J., and J.C. Schaake. 1982. Sensitivity of
water  resource  systems   to  climate  variation.
Hydrological Sciences Journal 27:327-43.
 Revelle, R.R., and P.E. Waggoner. 1983. Effects of
 the carbon dioxide induced climatic change on water
 supplies  in  the  western  United  States.    In:
 Changing Climate:  Report of the Carbon Dioxide
 Assessment Committee. Washington, DC: National
 Academy Press.

 Rind, D., and S.Lebedeff. 1984. Potential Climatic
 Impacts of  Increasing  Atmosphere CO,  With
 Emphasis on Water Availability and Hydrology in
 the  United   States.     Washington,  DC:   U.S.
 Environmental Protection Agency.

 Rosenberg, NJ. 1988. Global climate change holds
 problems  and  uncertainties for agriculture.   In:
 Tutwiler,  MA., ed. U.S. Agriculture in a Global
 Setting.   Washington,  DC:   Resources for  the

 Saliba, B.C., D.B. Bush, W.E. Martinard, and  T.C.
 Brown.    1987.    Do  water  market  prices
 appropriately  measure  water  values?   Natural
 Resources Journal 27:617-652.

 Schwarz, H.E.  1977.   Climatic change and water
 supply:    how  sensitive  is  the  Northeast?    In:
 National Research Council,  ed. Climate, Climatic
 Change, and  Water  Supply.   Washington,  DC:
 National Academy of Sciences, pp. 111-120.

 Sheer, D.P.  1985.  Managing water supplies to
 increase water  availability.  In: National Water
 Summary.  Hydrologic Events and Surface-Water
 Resources.   Washington,  DC:  U.S.  Geological
 Survey,  U.S.  Government  Printing  Office,  pp.

 Solley, W.B., C.F. Merck, and RJR.  Pierce. 1988.
 Estimated use of water in the United States in 1985.
 U.S. Geological Survey. Circular 1004. Washington,
 DC: U.S. Government Printing Office.

 Stockton,  C.W.,   and  W.R.   Boggess.    1979.
 Geohydrological Implications of Climate Change on
Water Resource Development.  Fort Belvoir, VA:
U.S; Army Coastal Engineering Research Center.

Strock, J.M. 1987. Adjusting water allocation law
to meet water quality and availability concerns  in a
warming world. In: Preparing for Climate Change.
Proceedings  of   the   First  North  American
Conference  on  Preparing  for Climate  Change.
Rockville,  MD:  Government Institutes, Inc.,  pp.

Chapter 9
Tarlock,  A.D.   1987.   Damning the dams and
ditches:  a review of D. Worster, rivers of empire:
water aridity and  the American West.   Natural
Resources Journal 27:477-490.

Trelease, FJ.  1977.  Climatic  change and water
law.  In:  National Research Council, ed. Climate,
Climatic  Change, and Water Supply.  Washington,
DC:  National Academy of Sciences, pp. 70-84.

U.S. Army Corps of Engineers. 1977. Northeastern
United  States  Water  Supply  Study,  Summary
Report.  U.S. Army Corps of Engineers,  North
Atlantic Division.

U.S. Army Corps  of Engineers.  1988.   Lessons
Learned From the 1986 Drought. U.S. Army Corps
of Engineers, Institute for Water Resources. IWR
Policy Study 88 - PS - 1. Fort Belvoir, VA

ILS. Congress. 1987. U.S. Congress, Congressional
Budget Office. Financing Municipal Water Supply
Systems; May.  Washington,  DC:  Congressional
Budget Office.

U.S.  Congress.  1984. U.S. Congress,  Office  of
Technology Assessment.  Protecting the Nation's
Groundwater From Contamination.  Volume 1.
Washington, DC: TJ.S. Government Printing Office.

U.S.  Department  of the Interior.   1987.   U.S.
Department of the Interior, Bureau of Reclamation.
Assessment '87.. A New Direction for the Bureau of
Reclamation. Washington, DC: U.S. Department of
the Interior.

U.S. EPA.  1987a. U.S. Environmental Protection
Agency.  1986 Needs Survey Report to Congress:
Assessment of Needed Publicly Owned Wastewater
Treatment Facilities in the United States.  EPA-
430/9-87-001. Washington, DC: U.S. Environmental
Protection Agency.
U.S. EPA.  1987b.  U.S. Environmental Protection
Agency.  National Water Quality Inventory: 1986
Report   to   Congress.      EPA-440/4-87-008.
Washington, DC: U.S. Environmental Protection

U.S. Geological Survey.  1984.  National Water
Summary 1983  - Hydrological Events  and Issues.
Water Supply Paper 2250.  Washington, DC: U.S.
Government Printing Office.

U.S. Geological Survey.  1985.  National Water
Summary 1984.    Water  Supply  Paper  2275.
Washington, DC: U.S. Government Printing Office.

U.S. Geological Survey.  1986.  National Water
Summary 1984.    Water  Supply  Paper  2300.
Washington, DC: U.S. Government Printing Office.

U.S. Water Resources Council. 1980. State of the
States: Water Resources Planning and Management.
Washington, DC: U.S. Water Resources Council.

U.S. Water Resources Council. 1978. The Nation's
Water Resources 1975-2000: Second National Water
Assessment. Vol. 1:  Summary.  Washington, DC:
U.S. Government Printing Office.

ViessmanJr., W., and C. DeMoncada.  1980. State
and National Water Use Trends to the Year 2000.
Washington, DC: Congressional Research Service,
The  Library of Congress.

Wahl, R.W.,  and R.K. Davis.  1986.   Satisfying
Southern California's thirst  for water:   efficient
alternatives. In: Frederick, K.D., ed.  Scarce Water
and  Institutional  Change.    Washington,  DC:
Resources for the Future.

Wahl, R.W., and F.H. Osterhoudt; 1985. Voluntary
transfers of water in the West. In:  U.S. Geological
Survey.  National Water Summary 1985.  Hydrologic
Events and Surface-Water Resources. Washington,
DC: U.S. Government Printing Office, pp. 113-124.

                                       CHAPTER 10
                             ELECTRICITY DEMAND

 Global warming would increase electricity demand,
 generating capacity requirements, annual generation,
 and fuel costs nationally.  The impacts could be
 significant within a few decades and would increase
 substantially over time if global warming continues.

 •   The  new generating capacity requirements
    induced by climate change effects on electricity
    demand estimated for 2010 show an increase of
    25 to 55  gigawatts (GW), or 9 to  19% above
    estimated new capacity requirements assuming
    no change in climate. Between 2010 and 2055,
    climate change impacts on electricity demand
    could  accelerate,  increasing  new capacity
    requirements by 200 to 400 GW (14 to 23%)
    above what would be needed in the absence of
    climate change. These capacity increases would
    require investments of approximately $200 to
    $300 billion (in 1986 dollars). In the absence of
    climate  change,   population  and  economic
    growth   may   require  investments   of
    approximately $2.4 to 3.3 trillion through 2055.

 •   Estimated  increases  in  annual   electricity
    generation and  fuel use induced  by climate
    change represent  several thousand gigawatt-
    hours by 2055. The estimated increases are 1 to
    2% in 2010 and 4 to 6% in 2055. Annual fuel,
    operation, and  maintenance   cost to meet
    increased electricity demand would be several
    hundred million  dollars  in 2010 and several
    billion dollars in 2055. Without climate change,
    these annual costs would be $475 to 655 billion
    in 2055.

•   Estimated regional impacts differ substantially.
    The  largest  increases  could   occur  in  the
    Southeast and Southwest, where air-conditioning
    demands are large relative to heating. Northern
    border  states  may have a net reduction in
    electricity  generation relative   to  base  case
    requirements assuming no change  in climate.
    These  changes   could  be  exacerbated   by
    reductions in  hydropower production and
    increases in  demand for electricity to  run
    irrigation equipment.

 •   These results are sensitive to assumptions about
    the  rates of  economic growth, technological
    improvements, and the relationship between
    electricity use and climate. The potential savings
    in other energy sources (gas and oil) used for
    space heating and other end uses  sensitive to
    climate and the potentially significant impacts on
    hydroelectric   supplies  and  other   utility
    operations were not analyzed.

 Policy Implications

 •   Utility executives and planners should begin to
    consider climate change as a factor in planning
    new capacity  and future  operations.   The
    estimated impacts of climate change in some
    regions  are  similar  to the  range of  other
    uncertainties and issues utility planners need to
    consider  over  the  20- to  30-year  period.
    Additional climate and utility  analyses  are
    needed to develop  refined risk assessments and
    risk management strategies.

 •   The increased demand for electricity induced by
    climate change also  could  exacerbate  other
    environmental  problems,   such  as   the
    implementation of   "acid  rain"  strategies,
    adherence to  the international nitrogen  oxide
    treaty, state implementation  plans for ozone
    control, and thermal  pollution control permit
    requirements.   The Environmental Protection
    Agency should analyze the impacts of climate
    change  on  long-range policies and  should
    include climate change as an explicit criterion in
    making  risk  management   decisions  when

•   The increased demand for electricity could
    make policies to  stabilize  the atmosphere
    through energy conservation more  difficult to
    achieve.  The estimated increases in electricity
    generation induced by climate  change could
    increase annual CCL emissions, depending upon
    future   utility   technology  and  fuel  choice

Chapter 10
    decisions. Assuming no change in efficiency of
    energy production and demand,  reliance on
    coal-based technologies to meet the increased
    demands could increase CO2 emissions by 40 to
    65 million tons in 2010 and  by 250 to 500
    million tons in 2055. Use of other, lower CO2-
    emitting technologies  and fuels (e.g., efficient
    conversion  technologies  and nuclear  and
    renewable  resources) would  reduce  these
    incremental additions.  In addition, warmer
    winter temperatures could reduce the demand
    for oil and gas  in end uses such as residential
    furnaces  for  heating,  thereby lowering CO2
    emissions from  these sources.  Future analyses
    of national and  international strategies to limit
    greenhouse gases should include the changes in
    energy demand  created by global warming as a
    positive feedback.

    Climate change could affect a wide range of
energy sources and uses.  In the near  term, policies
aimed at reducing emissions of greenhouse gases
from  fossil fuel combustion could affect the level
and mix of fuel  consumption in various end-use
technologies and in the generation of electric power.
In  the longer  term, changes in  temperature,
precipitation,  and other climatic  conditions also
could affect energy resources.  For example, warmer
temperatures likely would reduce the demand for
fuels  used in the winter for space  heating and
increase the demand for  fuels used in the summer
for air-conditioning; and  reduced precipitation and
soil moisture in some regions could increase the use
of energy  to  pump  water for irrigation.  These
effects could be particularly significant for planning
in the  electric  utility industry based  upon  the
substantial amount of electric load accounted for by
weather-sensitive end uses, the variety of resources
used  to generate electric power, and the capital-
intensivity of the industry. One major consideration
is the potential impact of  climate change  on the
demand for electricity  and the  implications  of
changes  in  demand  on  utility  capacity and
generation requirements.

    Many  electrical  end uses vary  with weather
conditions. The principal weather-sensitive end uses
are space heating, cooling,  and irrigation pumping
and -- to a lesser degree  — water heating, cooking,
and refrigeration. These applications of electricity
may account for up to a third of total sales for some
utilities and may contribute an even larger portion
of seasonal and daily peak demands.

    Changes  in  weather-sensitive  demands  for
electricity can affect both the  amount and  the
characteristics of generating capacity that  a utility
must build and maintain to ensure reliable service.
These changes also can affect fuel requirements and
the  characteristics   of  efficient  utility  system
operations,   particularly   the   scheduling   and
dispatching of the utility's generating capacity. For
example, electric energy  used for air-conditioning
exceeds that used for space heating nationwide, and
the temperature sensitivity  associated with cooling
is  higher than that  associated with heating.  This
implies not only changes  in seasonal  electricity
demands but  also increases  hi  annual  electricity
demands as a result of higher temperatures.

    Similarly, utilities in  most regions experience
their peak demands in the summer.  A rise in air-
conditioning   and   other   temperature-sensitive
summer  loads  would significantly increase peak
loads  and,  as a result, would step  up utility
investments in new generating capacity  needed to
meet additional  demands and to maintain system

    Examples of other ways in which climate could
affect  electric  utilities  include  the  following.
Changes in precipitation, evaporation, and runoff
from mountain  snowpack  as  well as changes in
water management practices in response to climate
change could  affect the  annual  and seasonal
availability of streamflow to generate hydropower.
Reductions in hydropower would require utilities to
rely upon  other, possibly more costly and less
environmentally benign generation sources to meet
customer needs. Furthermore, reductions in water
resources would  adversely affect  the availability
and/or cost of water for powerplant cooling.

    Other  direct impacts  of  climate change on
electric utilities include the effects of temperatures
on powerplant operating efficiencies, the effects of
sea level rise  on the protection and siting of coastal
facilities, and the effects  of  changes in various
climate conditions  on the supply  of  renewable
energy resources such as solar  and wind power.
Also, legislation and regulations designed to limit
greenhouse gas emissions from utility sources could
significantly affect the supply and cost of electricity

    Although  some   of  these  impacts  could
significantly affect utility planning and operations
(particularly on a regional basis), they have not been
analyzed in detail and are not addressed in this
report. Further research and analysis are needed to
develop  a more complete  assessment  of utility

    A number of utilities conduct analyses relating
short-term variations in weather conditions with a
need to "weather-normalize" historical demand data
and to test the sensitivity of system reliability and
operations  to   these  short-term   variations.
Furthermore,  some  researchers  have  speculated
regarding the  potential  effects  of longer term
climate changes on electricity demand (e.g., Stokoe
et al., 1987).

    However, only one previous study has estimated
the potential implications  of longer term, global
warming-associated  temperature   changes   on
electricity demands and the effects of changes in
demand on utility investment and operating plans.
Linder et al. (1987) used general circulation model
(GCM) results to estimate the potential impacts of
temperature change on electricity demand (and on
the supply of hydropower) for selected case study
utility systems in two geographical areas: a utility
located in the  southeastern United States and the
major utilities in New York State, disaggregated
into upstate and downstate systems.
                             Electricity Demand

    Linder et al. found that temperature increase
 could  significantly heighten  annual  and  peak
 electricity demands by 2015, and that a temperature
 rise would require construction of new generating
 capacity and increases  in annual generation. The
 southeastern utility had higher estimated increases
 in electricity demand, generation, and  production
 costs than the New York utilities because of greater
 electricity  demands  for   air-conditioning.    In
 addition, streamflow used to generate hydropower
 in New York could be reduced, requiring increased
 use of fossil  fuel generation to meet customer
 demands for electricity.

Study Design

    Linder and Inglis (Volume  H) expanded the
case studies (Linder et al., 1987) of the sensitivity of-
electricity demand to climate change and conducted
a national analysis of electricity demand.  Relevant
regional results from the national studies of Linder
and Inglis are discussed in the regional chapters of
this report.

    The analytic approach developed by Linder et
al.  (1987)  formed the  basis for estimating  the
regional  and national impacts  described in this
report.  The principal steps in  the approach are
summarized in Figure 10-1 (see Volume H for more
details). Estimated impacts were developed for the
relatively near term  (from the present to  2010,

Sensitivity of
Electricity Demand

Impacts on
Utility Investments,
Operations, Costs
                    Figure 10-1. Analytic approach (Linder and Inglis, Volume H).

Chapter 10

within electric utility long-range resource planning
horizons of 20 to 30 years) and over the longer term
(to 2055), when  the  magnitude of  temperature
changes is expected to approach equilibrium levels
representative   of  a  doubling   of  atmospheric
concentrations  of CO,.  Linder  and Inglis  used
Goddard Institute for Space Studies (GISS) A and
B transient estimates of temperature change in 2010
and GISS A estimates for 2055 in their calculations.
The  scenario changes in annual temperatures for
the United States  range from about 1.0 to 1.4ฐ C in
2010   and  are  approximately  3.7ฐC  by  2055.
Regional  temperature  scenarios  show greater

    Linder and Inglis used actual utility demand and
temperature data  from the case study utilities, and
from five other large, geographically dispersed utility
systems, to  develop a  set  of  weather-sensitivity
parameters for utility areas.  On a weighted-average
basis (weighted by electricity sales), utility  peak
demands were estimated to  increase by about 3.1%
per change in degree Celsius (ranging from -1.35 to
5.40%  across  utility  areas), and annual energy
demands were estimated to  increase by about 1.0%
per change in degree Celsius (ranging from -0.54 to

    A number  of  uncertainties associated with the
data  and assumptions  used  to develop  these
weather-sensitivity relationships suggested that the
relationships may  understate customer response to
climate change, particularly at higher temperature
change levels occurring in the future. For example,
the approach did not explicitly account for probable
increases   in  the  market  saturation  of  air-
conditioning equipment  as  temperatures rise over
time. To address this possibility, an alternative case
was  designed  in  which the estimated weather-
sensitivity values were increased by 50%. This was
designated as the  "higher sensitivity case.

    Since this study is focused on estimating how
climate change may  affect key utility  planning
factors, Linder and Inglis used a planning scenario
assuming no change in  climate (a "base case") to
serve  as  a  basis for comparison  with  planning
scenarios under alternative  assumptions of climate
change for 2010 and 2055.  Thus, base case utility
plans  were  developed for  2010  and 2055, using
assumptions regarding future demands for electricity
in  the absence  of  climate  change  (reflecting
population   and  economic growth),   generating
technology option performance and costs, fuel costs,
and other utility characteristics.1  Linder and Inglis
assumed  that  future  capacity  and  generation
requirements will be met by investments either in
new coal-fired baseload  capacity  or in oil- and
natural gas-fired peaking capacity.  Other sources,
such as nuclear energy and renewables or innovative
fossil fuel-fired  technologies (e.g., fluidized bed
combustion),  were  not considered  (for further
details, see Linder et al., 1987).

    Demands  for electricity in  the absence  of
climate change can be related to the overall level of
economic activity as  represented by  the  gross
national product (GNP). Because economic growth
assumptions  are  critical to  estimates  of future
electricity demands,  alternative  GNP growth rates
were assumed in developing the base cases; these
ranged from  1.2 to  2.1%  per  year.2   These
alternative assumptions are  referred to as "lower
growth" and "higher growth," respectively.

    These assumptions served as inputs to a regional
planning model called the Coal and Electric Utilities
Model (CEUM).   CEUM outputs  include  the
amount  and  characteristics  of  new  generating
capacity  additions, electricity generation by fuel
type, and electricity production costs.


    The study extrapolated temperature-sensitivity
findings  for  some regions  and did  not include
specific analyses of temperature sensitivity for  all
utility  regions of the  United States.  It focused
narrowly on impact pathways, considering only the
potential effects of temperature  change on changes
in  electricity  demand.   Neither  the potentially
significant impacts of climate change on hydropower
availability nor   the  impacts  of  reduced  water
supplies for powerplant cooling  were  included.

    Furthermore,  the  study did not evaluate the
sensitivity of the results to different, doubled-CO2
 Note that the development and use of a base case reflecting
changes  in non-climate-related conditions  over time was
undertaken only for the electricity demand study, not for other
areas in this report. Changes in population and technology are
considered in Chapter 6: Agriculture.

HThese GNP growth rates are relatively conservative, but they
are comparable with GNP growth rates used by EPA in its
report to Congress on Policy Options for Stabilizing Global

                                                                                   Electricity Demand
GCM  climate  scenarios  (GFDL  and  OSU),
although the use of the GISS transient experiment
results  for  2010   and  2055  indicates  relative
sensitivities to small and large temperature changes.

    The  study  did not  consider  variations  in
temperature changes and the occurrence of extreme
events,  which  affect  powerplant  dispatch  and
determinations of peak demands, respectively, and
are important for utility planning.

    Many uncertainties exist regarding the concepts,
methods, and  assumptions involved in developing
and applying estimates of the temperature sensitivity
of demand. For example, a key assumption is that
the estimated sensitivities of demand to historical,
short-term variations in temperature are adequate
representations  of future  relationships  between
electricity demand and long-term changes in mean

    Uncertainties  also   exist  regarding  market,
regulatory, technological, and other conditions that
will face the utility industry in the future.  For
example, technological changes that improve  the
energy  efficiency   of  weather-sensitive  end-use
equipment or electricity-generating equipment will
continue to evolve. These changes would likely lead
to lower climate change impacts than estimated in
this report. On the other hand, regulatory changes
aimed at reducing the  emissions of greenhouse
gases from  electricity generation  could  limit a
utility's   future  fuel and technology  investment
options, leading to higher estimates of cost impacts
than reported here. Because of these limitations, it
is important to recall that the results presented in
the  next section  should not  be  considered  as
projections of actual powerplant  investments and
utility operations,  but  rather as  comparisons
providing estimates of the magnitude of sensitivities
to alternative climate; change assumptions.


    The potential national impacts for 2010 and
2055  are summarized  in Table 10-1.  The table
presents base case values (i.e., assuming no change
in  climate) for  each year and estimated impacts
represented by changes from the base case values.
The impacts for 2055 are presented for both the
lower growth  GNP and the higher growth GNP
cases. Also, where ranges of impacts are presented,
they  summarize the  estimates under alternative
climate change scenarios (GISS A and GISS B) and
assumptions of the weather sensitivity of demand
("estimated sensitivity' and "higher sensitivity").

    Estimated increases in peak demand over the
base case on a national basis range from 2 to 6% by
2010.   Changes  in  estimated annual   energy
requirements by 2010  are  more modest,  ranging
from 1 to 2%. In 2055, peak national demands are
estimated to increase by 13 to 20% above base case
values,  and  annual  energy   requirements  are
estimated to increase by 4 to 6%.

    By 2010, new climate change-induced generating
capacity requirements  increase by 6  to  19%, or
about  24  to 55   GW, representing  an  average
increase of up to 1 GW per state (approximately the
capacity of one to two large nuclear or coal-fired
baseload powerplants). The majority of the capacity
increase is for peaking capacity rather than baseload
capacity.  The investment associated with these
capacity increases is  several  billion dollars  (in
constant 1986 dollars).  By 2055, the change in new
capacity requirements increases in percentage terms
and represents several hundred GW.  Under high
GNP and higher weather-sensitivity assumptions, the
estimated increase attributable to climate change is
almost 400 GW, or 23%. To put these results into
perspective,   it should  be noted  that  current
generating capacity in the United States is about 700
GW.  The increase in new capacity requirements
under the  base case is 1,350 to X780 GW.

    Annual  generation increases for  the United
States are not as large in percentage terms as those
estimated for new generating capacity requirements,
but nonetheless, they account for several hundred
billion kWh by 2055.  In the  near term  (i.e., to
2010),  increased. levels and changing patterns of
climate change-induced electricity demand permit
utilities in some  areas having  excess generating
capacity to serve the growing needs of utilities in
other  areas  through substitution of lower  cost
baseload  generation  for  higher  cost   peaking
generation.  On net, peaking generation would be
lower  as a result  of climate change in 2010 (see
Under et al., 1987,  for further detail).  In 2055,
peaking generation is projected to increase along
with baseload generation, because all the excess
capacity that had existed in 2010 either would have
been fully used by growing demands to  2055 or
would  have been retired.  The estimated impacts of
climate change on national new generating capacity

 Chapter 10
             Table 10-1.  The Potential National Impacts of Climate Change on Electric Utilities

Lower GNP

Peak demand (GW)
New capacity requirements (GW)a
Annual sales (bkWh)
Annual generation15 (bkWh)
Cumulative capital costsฐ'd
Annual costsd
Higher GNP
^Includes reserve margin requirements; does not include "firm scheduled" capacity.
 Includes transmission and distribution losses.
ฐ"Base" values include regional capital expenditures for utility-related equipment hi addition to new generating
  capacity (e.g., new transmission facilities).
dln billions of 1986 dollars.
Abbreviations:  GW = gigawatts; bkWh = billion kilowatthours.
Source:  Linder and Inglis (Volume H).
requirements and annual generation are illustrated
in Figure 10-2.

    Table 10-1 also indicates that the increase in
annual  costs for capital, fuel, and operation and
maintenance associated with climate change-induced
modifications in utility investments and operations
are a few billion dollars in 2010 and are $33 to $73
billion by 2055, a 7 to 15% increase over base case
values of $475 to $655 billion for 2055.

    Figures 10-3 and 10-4 illustrate the diversity of
the estimated  results for generating  capacity on a
state-by-state  basis.    The  state  and  regional
differences  reflect  differences in current climate
conditions (e.g.,  seasonal temperature  patterns),
assumed future climate changes, and electricity end-
use and  utility system characteristics (e.g., market
saturation of  weather-sensitive  appliances  and

    Figure 10-3 shows that estimated reductions in
new capacity requirements induced by  climate
change are limited to  the whiter-peaking regions of
the extreme Northeast and Northwest. The Great
Lakes, northern Great Plains, and Mountain States

                                                                                           Electricity Demand
                                                                                  Additional Climate Change Impacts:
                                                                                  Higher Sensitivity

                                                                                  Climate Change Impacts:
                                                                                  Base Sensitivity

                                                                                  Base Case (No Climate Change)
             2010     2066     2055    2010     2056     2055
                      Lower    Higher            Lower    Higher
                      GNP     GNP              GNP      GNP
                    Assumption Assumption       Assumption Assumption
Figure 10-2.  Potential impacts of climate change on electric utilities, United States (Linder and Inglis, Volume
Figure 10-3.  Changes in electric utility capacity additions by state, induced by climate change in 2055 (derived
from Linder and Inglis, Volume H).  '

 Chapter 10
 are estimated to experience increased new capacity
 requirements by 2055 in the range of 0 to 10%.
 Increases greater than 20% are concentrated in the
 Southeast, southern Great Plains, and Southwest.

     Figure  10-4   shows  a  somewhat   similar
 geographic  pattern  of  impacts .for  electricity
 generation in 2055.  Reductions in generation are
 estimated  in the North, and the greatest increases
 are concentrated  in the  Southwest.    Despite
 substantial use of air-conditioning in the Southeast,
 the estimated increases in generation are only in the
 5 to 10% range. There is a relatively high market
 saturation of electric heat in the region,  and the
 increase in cooling is partly offset by, a decrease in
 heating as a result of warmer winters.

     Because regions  are affected differently,  the
 results indicate potential changes in the patterns of
 interregional bulk power exchanges and capacity
 sales over tune and as climate  changes.   For
 example,  under the assumption  of  increasing
 temperatures, some regions may require significant
 amounts of additional generating capacity to reliably
 meet increased demands  during  peak (cooling)
 seasons, but may experience lower demands in other
 (heating) seasons.  As a result, the region's needs
 may be  for powerplants that are  utilized heavily
 during only part of the year. Low annual utilization
 in the region would not justify construction of high-
. capital and low-fuel cost baseload powerplants that
 can produce electricity more cheaply (per kWh)
 than low-capital and high-fuel-cost peaking units.
 However, when considered across several regions,
 the least-cost  plan may be to construct baseload
 powerplants in certain regions, utilize them to an
 extent greater  than required by the region, and sell
 the "excess" electricity from these plants into other
 regions.    The  location  and amount  of these
 interregional sales would be subject to the transfer
 capabilities of transmission capacity in place.  An
 alternative to  increased interregional bulk power
 sales would be the development and application of
 efficient and effective energy storage technologies.

    Despite the limitations of the analysis and the
need  for more research to refine  the  data and
methods used,  the  results  are  judged to  be
Figure 10-4. Changes in electricity generation by state, induced by climate change in 2055 (derived from Linder
and IngUs, Volume H).

                                                                                     Electricity Demand
reasonable estimates of the sensitivity of electricity
demand   to   potential  climate  change.    Key
socioeconomic and environmental implications of
the: results  stem  from  the increases  irt electric
generating capacity and generation  requirements
associated with climate-induced changes in demand.
The implications include the following:

  •  •   Climate change could result in overall fuel
        mixes for electricity generation that differ
        from those expected  in the  absence of
        climate change.

    •   Climate change would  not  evenly affect
        regional demands for electricity. Greater
        impacts would occur  in regions  where
        weather-sensitive end uses (particularly air-
        conditioning)  are  important  sources  of
        electricity demand.  Substantially greater
        climate change impacts were estimated for
        the Southeast and Southwest than for other
        regions, especially the  northern tier of
        states.  Other impacts not addressed in this
:        study, such as the availability of water for
        hydropower  generation  and  powerplant
        cooling, also would be more important in
        some  regions  (e.g.,  the  West) than in

    •   Regional  differences  in  capacity   and
        generation  requirements  suggest  that
        important   new   opportunities   for
        interregional  bulk power exchanges  or
        capacity sales  may arise as a result of
        climate change.

    •   The impacts of uncertain climate conditions
        over the long term could pose significant
        planning and economic risks.  Because of
        long lead times required to plan and build
        economic baseload generating capacity, the
        ability  of utility  planners  to  correctly
        anticipate climate change could result in
        lower electricity production costs.  The
        magnitude of these risks in some regions
        (e.g., the Southeast and the southern Great
        Plains)  could  be   similar   to   other
        uncertainties  that utility  planners   and
        decisionmakers must face.

    •   If the result is  confirmed that the majority
        of new capacity requirements in response to
        climate change are for peaking capacity, a
        new technological and market focus would
        be directed toward this type of generating
        plant.  Related to this would be increased
        research and  development on  electricity
        storage technologies, which  would  allow
        lower cost, more efficient  powerplants to
        generate, at off-peak times, electricity for
        use during peak periods.

        Because increases in customer demands for
        electricity may be particularly concentrated
        in certain seasons and at peak periods,
        conservation   and   especially   load
        management programs  that  improve  the
        efficiency or  change  the  patterns   of
        customer uses of electricity could be more
        cost-effective  when  considered  in  the
        context of potential changes in climate.

        Increased electricity generation implies the
        potential   for   increased    adverse
        environmental  impacts depending  upon
        generating    technology   and   fuel-use
        assumptions.   Potential adverse impacts
        compared with the base case are associated
        with the following:

            air quality (e.g., emissions of sulfur
            dioxide, NOX, and other  pollutants);

            land use for new powerplant sites, fuel
            extraction, fuel storage, and solid waste

            water   quality  and  use  (e.g.,  for
            powerplant  cooling   and    fuel
            processing); and

            resource   depletion,   especially   of
            nonrenewable fuels such as natural gas.

        Of particular concern would  be additional
        water   withdrawal   and    consumption
        requirements in areas where water supplies
        may be reduced by climate change.3
 For example,  increased electricity  generation  induced by
climate  change in  northern  California  could  increase
requirements for water withdrawal by 600 to 1,200 million cubic
feet and for water consumption by 200 to 400 million cubic feet
in 2055. Comparable figures for the southern Great Plains in
2055 would be water withdrawal of 5,800 to 11,500 million cubic
feet and consumption of 1,800 to 3,500 million cubic feet.

 Chapter 10
        Increased electricity generation also implies
        increased emissions  of  COซ and  other
        greenhouse gases compared with base case
        emissions. For example, if the estimated
        increases  in   climate   change-induced
        generation reported in Table 10-1 were met
        by   conventional   technologies,   CO2
        emissions could increase by 40 to 65 million
        tons per year by 2010 and by 250 to 500
        million tons  per year by 2055.   Use of
        lower CO2-emitting technologies and fuels
        — such as efficient conversion technologies
        and nuclear or  renewable resources —
        would lower these estimated impacts.

    In general, the study results suggest that utility
planners and policymakers should begin now to
assess more fully and to consider climate change as
a  factor  affecting  their planning  analyses  and
decisions.  If more complete and more  detailed
analyses   support   the   socioeconomic   and
environmental implications of the climate change
effects described above, they should be  explicitly
addressed  in  planning analyses  and  decisions.
Specific policy implications related to the findings
include the following:

    •   In  formulating future  National Energy
        Plans, the Department of Energy may wish
        to consider the potential impacts of climate
        change on utility demands.

    •   The interactions of climate change and the
        current efforts  of the  Federal Energy
        Regulatory   Commission   (FERC)   to
        restructure the  electric utility industry are
        difficult  to  assess.   For example,  the
        industry's response to FERC policies could
        either  accelerate or reduce the rate of
        emissions of greenhouse gases, depending
        upon changes in the mix of generating fuels
        and effects on the efficiency of electricity
        production.     The  possible  alternative
        responses should be assessed, and FERC
 Note, however, that  these  increases  in emissions  from
electricity production could be offset, at least in part, by
reduced demand for space heating provided by natural gas and
oil furnaces or by other direct uses of fossil fuels.
        policies should be considered with respect
        to their potential implications  related to
        climate change issues.

        Increases in electricity demands induced by
        climate change will make achievement of
        energy conservation goals more  difficult.
        For  example, the  conference  statement
        from  "The   Changing  Atmosphere:
        Implications   for  Global   Strategy"
        (Environment Canada,  1988)  calls  for
        reductions in CO2 emissions to be achieved
        in part through increased efforts in energy
        efficiency and other conservation measures.
        An initial goal for wealthy, industrialized
        nations set by the conference is a reduction
        in CO, emissions through conservation of
        approximately 10% of 1988 emissions levels
        by 2005. The impacts of climate change to
        increase  electricity -demand should  be
        factored  into  the  policies  and  plans
        designed to achieve this conservation goal.

        Similarly,  climate  change impacts  may
        exacerbate   the  difficulties   or  costs
        associated  with  implementing  acid rain
        mitigation strategies being considered by
        the Congress.  However, these strategies
        center primarily on near-term solutions
        focusing  on  emissions reductions  from
        existing powerplants, and the impacts  of
        climate change may not  be large within that
        time frame.

        Although not addressed  directly in the
        analyses underlying  this report, state and
        federal agencies should  consider mitigation
        strategies that include energy conservation;
        increased  efficiency in  the  production,
        conversion, and  use of energy; and the
        development and reliance on  fuel sources
        with low CO2 emissions.

    Important areas for further climate change
research include improved methods for developing
and disseminating climate change scenarios, with
particular  emphasis on (1) improved estimates of
climate variables  (in  addition  to  temperature)
relevant  to   utility  impact   assessment  (e.g.,

                                                                                 Electricity Demand
hydrologic  factors, winds); (2) estimates  of the
possible impacts of global warming on variations in
weather conditions and the occurrence of extreme
events; (3)  continued attention to estimates of the
rate of climate change over time; and (4) estimates
of climate change at a more disaggregated regional
or local level.

    Follow-on research suggestions on the utility
side include  (1)  refinement  of  the  analytical
approach, in part  through lessons learned from
additional utility-specific analyses; (2) more detailed
and complete analyses of the weather sensitivity of
customer demand for electricity; (3) extension of the
approach to  consider other  pathways  (including
indirect and  secondary  effects)  through which
climate change could affect utility investments and
operations;  and (4) an assessment of the value of
improved climate  change  information to utility
planners and managers.
Environment Canada.   1988.   The Changing
Atmosphere:   Implications  for Global  Security
Conference  Statement.   Environment  Canada,
Government of Canada. Toronto; June 27-30.

Linder,  K.P., MJ. Gibbs, and M.R. Inglis.  ICF
Incorporated.  1987.  Potential Impacts of Climate
Change on Electric Utilities.  Report 88-2.  Albany,
New York: New York State Energy Research and
Development Authority.  (Note: This report was
also published  by the Electric  Power Research
Institute, Palo Alto,  California, in January 1989;
report no. EN-6249.)

Stokoe,  P.K.,  and M. LeBlanc.    P.  Lane and
Associates, Ltd., and Discovery Consultants, Ltd.
1987.  Socio-economic Assessment of the  Physical
and Ecological Impacts of Climate Change on the
Marine  Environment  of  the Atlantic Region  of
Canada, Phase  I.  Halifax, Nova Scotia,  Canada:
School for Resource and Environmental  Studies,
Dalhousie University.


                                        CHAPTER 11

                                      AIR QUALITY

•   Potential changes  in regional  temperatures,
    precipitation patterns, clouds, windspeed and
    direction, and atmospheric water vapor that will
    accompany global  climate change  will affect
    future air pollution levels and episodes in the
    United States.

•   While uncertainties remain, it is likely that an
    increase in global temperatures would have the
    following effects on air quality, if other variables
    remain  constant.   These  potential impacts
    should be interpreted as relative changes as
    compared with air quality levels without climate
    change. This chapter does not predict what will
    happen to  air  quality without climate change
    and does not consider changes in anthropogenic
    emissions or technology.

    —  Ozone  levels in many urban areas would
        increase because higher global temperatures
        would speed the reaction rates producing
        ozone in the atmosphere.

    —  Natural emissions of hydrocarbons would
        increase with a temperature rise. Natural
        emissions of sulfur would also change, but
        the   direction  is   uncertain.      The
        hydrocarbons   and   nitrogen  oxides
        participate in reactions that produce ozone.

    —  Manmade   emissions  of   hydrocarbons,
        nitrogen oxides, and sulfur oxides may rise
        if more fossil fuel is used to meet higher
        electricity needs (see Chapter 10: Electricity
        Demand)  and if  technology  does not

    —  The formation of acidic materials (such as
        sulfates)  would   increase  with warmer
        temperatures because sulfur and nitrogen
        oxides  would  oxidize more rapidly.  The
ultimate effect on acid deposition is difficult to
assess because of changes in clouds, winds, and

- Visibility  may  decrease  because  of the
   increase in hydrocarbon emissions and the
   rate at which sulfur dioxide is oxidized to

— The small increase in temperature will not
   significantly  affect   carbon  monoxide

Preliminary analyses of the effects of a scenario
of a 4ฐC temperature  increase  in the San
Francisco Bay area, with no change in emissions
or  other   climate   variables,   on   ozone
concentrations suggest that  maximum ozone
concentrations could increase by approximately
20%,  that  the area  in  which the National
Ambient Air Quality Standard (NAAQS) would
be exceeded would almost double, and that the
number of  people-hours of exposure would
triple.  The Midwest and Southeast also could
incur high concentrations and an increase in the
area of high ozone by a factor of three.

Increases in ambient ozone levels resulting from
climate change could increase the number of
nonattainment areas and make attainment more
expensive  in many  regions.    Preliminary
estimates suggest that an expenditure of several
million dollars per year may be necessary for
volatile organic compound (VOC)  controls
above those needed to meet standards without
climate change.  The total costs for additional
air pollution  controls that may be  needed
because of global warming cannot be estimated
at this time.

Because of the close relationship between air
pollution policies and global climate change, it
is appropriate for EPA to review the impact of

 Chapter 11
     global climate change on air policies and the
     impact of air pollution regulations on global
     climate change.

    The summer of 1988 provided direct evidence
 of the importance of weather to pollution episodes
 in the United States. Despite significant progress in
 reducing emissions of many pollutants over the last
 decade, the extended stagnation periods and high
 temperatures caused ozone levels in 76 cities across
 the country to exceed the national standard by at
 least  25%.   Whether this  recent summer is an
 appropriate  analog  for  the  future  cannot be
 determined with  certainty,  but  scientists  have
 recognized for some time  that air pollution does
 vary  with seasons  and is  directly  affected by
 ventilation,  circulation, and precipitation,  all of
 which could be  affected by future global climate


    Two major factors, referred to as "ventilation"
 when considered together,  control the dilution of
 pollutants by the atmosphere:  windspeed and the
 depth of the atmospheric mixing layer (frequently
 called the mixing depth).   If windspeed is high,
 more air is  available  to  dilute  pollutants,  thus
 lowering pollutant concentrations. The mixing layer
 (the distance  between the  ground and the first
 upper-layer inversion)  tends to trap  pollutants
 because the inversion above it acts as a barrier to
 vertical pollutant movement.   Thus,  pollutant
 concentrations decrease as mixing depth increases,
 providing greater dilution.

    The  ventilation  characteristics  of  an  area
 change, depending on whether a high- or  low-
 pressure system  is present.   Low-pressure systems
 usually  produce  good ventilation  because  they
 normally   have   greater  mixing   depths  and
windspeeds, and precipitation is often associated
with them.  High-pressure systems, on the other
hand, generally produce poor ventilation conditions
because they frequently have smaller mixing depths
 on their western sides and lower windspeeds. They
 also tend to move more slowly than lows, so more
 emissions can enter their circulation patterns.  In
 addition, they are frequently free of clouds, resulting
 in   maximum   sunlight  and  therefore   more
 photochemical  ozone production during the day.
 Also, during the evenings,  the  clear skies allow
 surface-based (see below) inversion layers to form,
 concentrating pollutants in a small volume  of air
 and often creating very high air pollution levels.

    Climatologically, certain places hi the country,
 such as the Great Plains and the Northeast (Figure
 11-1A), are frequently windy,  and others, such as
 the Southwest (Figure 11-1B), frequently have large
 mixing depths.  These areas will have cleaner-than-
 average  air  if  they  do not  contain too  many
 pollutant sources.  Areas, such as California, that
 are frequently affected by high-pressure  systems ~
 causing lower windspeeds and smaller mixing depths
 — will have more major air pollution episodes.


    Two semipermanent high-pressure systems are
 important to the  global circulation pattern and
 greatly influence U.S. air pollution climatology: the
 large Pacific high, which is often situated between
 the Hawaiian Islands and the west coast of North
 America, and the Bermuda high,  located over the
 western Atlantic Ocean.

    The Pacific high  often  results  in  extended
 periods of air stagnation over  the western United
 States from Oregon and  California to over the
 Rockies, and is responsible for many severe ozone
 episodes in southern California.   Air stagnation
 associated  with the westward extension of the
 Bermuda high occurs most often during the summer
 months and affects the eastern United States from
 southern Appalachia northward to New England.
Within  the Bermuda high,  pollutants are  slowly
 transported from the industrial areas of the Ohio
River  Valley  into  the populated  areas  of the
Northeast.  The Bermuda high is also responsible
for the general southwest-to-northeast airflow in the
summer, carrying pollutants along the metropolitan
corridor from Richmond to Boston and exacerbating
the ozone problem in the Northeast.


    Atmospheric pollutants in both particulate and
gaseous forms are  incorporated  into clouds and

                                                                                       Air, Quality
Figure 114, (A) Mean annual windspeed averaged through the afternoon mixing layer (speeds are in meters
per second); (B) mean annual afternoon mixing height, in hundreds of meters (adapted from Holzworth, 1972).
precipitation.    These  pollutants  can then  be
transported to the ground  through rainfall (wet
deposition).   Cloud-formation processes and the
consequent type of precipitation, together with the
intensity  and  duration  of  precipitation,  are
important  in determining  wet  deposition  of

    To protect the public health and welfare, the
U,S. EPA has promulgated National Ambient Air
Quality Standards (NAAQS). In 1986, more people
lived in counties with measured air quality levels
that violated the primary NAAQS for ozone (O3)
than for other pollutants (Figure 11-2).

    Although millions of people continue to breathe
air that is in violation  of the  primary NAAQS,
considerable progress is being made in reducing air
pollution levels. Nationally, long-term 10-year (1977-
86)   improvements  have   been  seen   for
a number of pollutants, including total suspended
particulates (TSP), O,,  carbon monoxide  (CO),
nitrogen  dioxide (NO^), lead, and sulfur dioxide
(SO2). This section  does not attempt to predict
future trends in emission levels.

Total Suspended Particulates

    Annual average TSP levels decreased by 23%
between  1977 and 1986, and particulate emissions
decreased by 25% for the same period. The more
recent TSP data (1982-86) show that concentrations
are leveling off, with a 3% decrease in ambient TSP
levels and a 4% decrease in estimated emissions
during that time.

    In the future, air quality may decrease  as the
benefits of current pollution control measures are
affected by increases  hi population and economic

Sulfur Dioxide

    Annual average SO2 levels decreased 37% from
1977 to  1986.   An  even greater  improvement

 Chapter 11
NO 2


) 10 20 30 40 50 60 70 80 90
mil lions of persons
Figure 11-2.  Number of persons living in counties
with air quality levels above the primary National
Ambient Air Quality Standards in 1986 (based on
1980 population data) (U.S. EPA, 1988).
was observed in the estimated number of violations
of  the  24-hour standard for  SO,  concentration,
which  decreased  by  98%.    These  decreases
correspond  to  a 21%  drop  in sulfur  dioxide
emissions during this 10-year period. However, most
of the violations and the improvements occurred at
source-oriented sites,  particularly a  few  smelter
sites.  Additional reductions may be more  difficult
to obtain. The higher concentrations were found in
the heavily populated Midwest and Northeast.


    A national standard for ambient levels of ozone
was established with the original Clean Air Act in
1972,  along with standards for five other pollutants.
While headway has been made in meeting all these
national air quality standards, progress in meeting
the ozone standard has been particularly slow and
frustrating  for   concerned    lawmakers   and
environmental officials  at all levels of government.
At the end of  1987, the date anticipated in the act
for final attainment of the ozone standard, more
than 60 areas had not met the standard.  In recent
years,  the number of nonattaimnent areas has
fluctuated with meteorology, often overwhelming the
progress being made through reduced emissions.
Thus   "bad"   weather   (summertime  conditions
favorable  to ozone  formation) in 1983 led to an
increased  number of  nonattainment  areas,  and
"good" conditions in 1986 led to a decreased number
of areas.
    Nationally, between 1979 and 1986, O3  levels
 decreased by 13%.  Emissions  of volatile organic
 compounds (VOCs), which are ozone precursors,
 decreased by  20% from  1979  to  1986.   The
 estimated  number  of  violations of  the ozone
 standard decreased by 38% between 1979 and 1986.
 The  highest  concentrations  were  in  southern
 California, but high levels  also  persisted hi  the
 Texas gulf coast, the northeast corridor, and other
 heavily populated regions.

 Acid Deposition

    Widespread concern exists  concerning  the
 effects of acid deposition on the environment.  With
 the present monitoring network density in eastern
 North America, it is  now possible  to quantify
 regional patterns of concentration and deposition of
 sulfate,  nitrate,   and   hydrogen  ions,  primary
 constituents of acid deposition.   In  Figures 11-3
 through 11-5, isopleth maps show the geographic
 pattern of acid deposition, as reflected by  the
 concentration and deposition of these three species
 (Seilkop and Finkelstein, 1987).

    For the relatively short  period from  1980 and
 1984, evidence indicates the total deposition and
 average  concentration  of  sulfate,  nitrate,   and
 hydrogen  ions in precipitation falling over eastern
 North America decreased by 15  to  20%.  The
 observed  decreases correspond   with  reported
 reductions in the U.S.  emissions  of sulfur oxides
 (SOX) and nitrogen oxides (NOX), and sulfate and
 nitrate precursors.   However, the  emission figures
 are subject to estimation error and should be used
 cautiously (Seilkop and Finkelstein, 1987).

    Some of the climate factors that could affect air
quality are listed in Table 11-1.  To explain these
relationships, two projects were undertaken for this
report to identify the potential impacts of climate
change on air quality:

    1.  Climate Change  and Its  Interactions with
      Air Chemistry: Perspectives and Research
       Needs - Penner, Connell, Wuebbles, and

                                                                                        Air Quality
Figure 11-3. Isopleth maps of average annual concentrations (mg/liter) and total annual deposition (g/m ) of
sulfates in 1980-84 (Seilkop and Finkelstein, 1987).
 Figure 11-4.  Isopleth maps of average annual concentration (mg/liter) and total annual deposition (g/m ) of
 nitrates in 1980-84 (Seilkop and Finkelstein, 1987).

  Chapter 11
                                               HYDROGEN IONS

                              CONCENTRATION                     DEPOSITION
 Figure 11-5.  Isopleth maps of average annual concentration (mg/liter) and total annual deposition (g/m^) of
 hydrogen ions in 1980-84 (Seilkop and Finkelstein, 1987).

                  Table 11-1.  Climate Change Factors Important for Regional Air Quality
 Changes in the following affect air quality:

 1.   the average maximum or minimum temperature
     and/or  changes in  their spatial distribution
     leading to a change  in reaction rates and the
     solubility of gases in  cloud water;

 2.   stratospheric O3 leading to a change in reaction

 3.   the frequency  and  pattern  of  cloud  cover
     leading to a change in reaction rates and rates
     of conversion of SO2 to acid deposition;

 4.   the frequency  and  intensity of stagnation
     episodes or a change in the mixing layer leading
     to  more  or less mixing of polluted air with
     background air;

 5.   background boundary layer concentrations of
     water  vapor,  hydrocarbons,  NO ,  and  O,,
     leading to more or less dilution ofpolluted air
     in the boundary layer  and altering the chemical
     transformation rates;
 6.  the   vegetative   and   soil   emissions   of
    hydrocarbons and NOX that  are sensitive  to
    temperature and light levels, leading to changes
    in their concentrations;

 7.  deposition rates to vegetative surfaces  whose
    absorption  of pollutants  is  a  function  of
    moisture, temperature, light intensity, and other
    factors, leading to changes in concentrations;

 8.  energy usage, leading to a change in energy-
    related emissions;

 9.  aerosol  formation,  leading  to  changes   in
    reaction rates  and  the  planetary  albedo
    (reflectivity); and

10.  circulation and precipitation patterns leading to
    a  change  in the  abundance of  pollutants
    deposited locally versus  these exported off the
Source: Adapted from Penner et al. (Volume F).

                                                                                          Air Quality
       Covey -  Lawrence Livermore National
       Laboratory (Volume F)

    2.  Examination of the Sensitivity of a Regional
       Oxidant  Model  to Climate  Variations -
       Morris,  Gery,  Liu,  Moore, Daly,  and
       Greenfield  -  Systems Applications,  Inc.
       (Volume F)

    The literature does  not contain studies on the
effects of climate change on air quality. Thus, these
studies should be considered as preliminary analyses
of the sensitivity of air quality to  climate change.

Climate Change and Its Interactions with
Air Chemistry

    Penner et al. conducted a literature review of
studies on the relationship of climate and air quality.
They also organized a workshop  on the issue.

Effect of Climate Change on Ozone Formation

    Changes in ventilation, circulation, precipitation,
and other aspects of climate affect the concentra-
tions of the  ozone  precursors (VOCs and NO ).
Climate changes can also increase or decrease the
rates at which these precursors react to form ozone.
The effects of change in global temperature and in
stratospheric ozone concentration on tropospheric
ozone precursor concentrations, reaction rates, and
tropospheric ozone  concentrations are discussed

Temperature Change

    Studies  of  the Effects  of  Temperature on
Ozone.  Smog chamber and modeling studies have
shown that ozone levels increase as temperature
increases.  Kamens et al. (1982) have shown in an
outdoor  smog chamber study that the maximum
ozone concentration increases as the daily maximum
temperature  increases   (holding light   intensity
constant).  Their data show that  there is no critical
"cut-off temperature that eliminates photochemical
ozone production.  Instead, a general gradient is
observed as a function of temperature.

     Samson (1988)  has recently studied  ambient
data for Muskegon, Michigan, and found  that the
number  of ozone excursions  above the standard
(0.12 ppm)  is almost  linearly  related to mean
maximum  temperature.    In  1988,  the  mean
maximum temperature was 77ฐF and there were 12
ozone  excursions.    In   1984,  with  a  mean
temperature  of  73.50ฐF,  there  was  only  one

    Temperature-dependent modeling studies were
conducted by Gery et al. (1987). For this modeling
effort,  Gery et al. used  the  OZIPM-3 trajectory
model, which is city specific. The scenarios for the
different cities used actual observed mixing heights,
solar  radiation and  zenith angle, and  pollutant
concentrations characteristic for the particular city
considered for June 24, 1980.  This base case was
chosen because it was a high-pollution  day, and
ambient  data  were  available.    The  increased
temperature   scenarios   applied   the  increase
throughout the day and were added to the base case
scenario. The light intensity increase was achieved
by  increasing the photolyses rates  for nitrogen
dioxide,  formaldehyde,  acetaldehyde,  hydrogen
peroxide, and ozone.  Results for New York in June
1980 are shown in Table 11-2.  In general, ozone
concentration increased withincreasing temperature.
The concentration of hydrogen peroxide (H2O2), a
strong oxidant that converts SO2 to sulfunc  acid,
was  also  observed to   increase  with  higher
temperatures. This is compatible with the increase
in ozone because  the entire photochemical reaction
process is accelerated when temperature rises. As
a result, cities currently violating the ozone NAAQS
will be in violation to a greater degree in the future,
and cities that are complying with the NAAQS now
could  be  forced out of  compliance just by  a
temperature  increase.   Figure  11-6  shows the
predicted  increase  in low-level  ozone for two
temperature increases in Los Angeles, New York,
Philadelphia, and Washington.

    Modeling studies by Penner et al. have shown
that the effect temperature has on ozone formation
also  depends  on the  ratio  of  volatile organic
compounds to nitrogen oxides, both of which are
ozone precursors.  Figure 11-7 shows that ozone
levels will generally go up, except in areas where the
ratio of VOCs to NOX is low.

    Temperature change has  a  direct  effect  on
ozone concentrations because it increases the rates
of  ozone-forming   reactions.      However,   a
temperature rise  can also affect ozone formation by
altering  four  other  aspects  of climate  or the
atmosphere:  cloud cover, frequency and intensity of

 Chapter 11

 Table 11-2. Maximum Hourly Concentrations and Percentage Changes for Ozone, H2O2, and PAN for the
            Future Sensitivity Tests Using an EKMA Model for the Simulation of June 24,1980, New York

  Change in Temp (ฐC)

  Stratospheric Ozone3

                 Concentration fppm)

                 0     +2    +5
                 0.125  0.130   0.138
                 0.150  0.157   0.167
                 0.165  0.170   0.178
                  Percent change
                    (from base)

                      +2    +5
 Hydrogen Peroxide (H2O2)

  Change in Temp (ฐC)

  Stratospheric Ozone3

                 Concentration fppb)
                0.05   0.06   0.08
                0.43   0.58   0.84
                3.08   3.31   3.60
                  Percent change
                    (from base)
                      20      60
              760.0  1060   1580
              6060.0  6520   7100
Peroxvacetvl Nitrate (PAN)

  Change in Temp (ฐC)

  Stratospheric Ozone3
                Concentration (ppb)
                   Percent change
                    (from base) .
-. 47
aBase refers to the present stratospheric ozone column.  The -16.6 and -33.3% refer to a depletion of the base
 value. Ultraviolet light will increase with the depletion (Gery et al., 1987).
stagnation  periods, mixing
reactant concentrations.
layer  thickness,  and
    Effect of Changes  in  Cloud  Cover.    The
reduction in light intensity caused by  increased
cloud cover can reduce ozone production. Penner et
al. (Volume F) calculate that a reduction hi light
intensity of 50% throughout the day will reduce the
ozone formation. However, the magnitude of ozone
reduction depends  on the time of day when the
cloud cover occurs. If clouds occur in the afternoon
or evening, little  effect  is observed in the  ozone
production, but if clouds occur during the morning
hours, photochemical reactions are slowed, and less
ozone is produced.   Jeffries et  al. (1989) suggest
that cloud cover can decrease ultraviolet radiation
by 7 to 14% in their outdoor smog chamber located
in North Carolina. Although a global temperature
change  would affect cloud  cover, the type  and
direction of the change are unknown.

                                                                                           Air Quality

         L.A.  N.Y. Phil. Wash.

  LA.  N.Y. Phil. Wash.

Figure 11-6. Percent increase in predicted O3 over
future base case  (0.12 ppm) for two temperature
increases in four cities (Gery, 1987).
15    20   25   30   35    40
                                                            BL = Boundary Layer
                          Figure 11-7. The effect of temperature on the peak
                          O,  concentrations predicted  in a  box  model
                          calculation of urban O3 formation. Calculations are
                          shown for three hydrocarbon to NOX ratios.  The
                          effect of increasing the boundary layer depth for the
                          case with a hydrocarbon to NQj  ratio of 7 is also
                          shown (Penner et al., Volume F).
    The Penner et  al. study assumes that cloud
cover causes an equal decrease in all wavelengths of
solar radiation.  However, clouds are not expected
to cause an equal decrease at all wavelengths. Solar
radiation is needed to form ozone. Since Penner et
al. may have underestimated  the intensity of some
wavelengths of light, they may have  overestimated
the decrease in ozone production.

    Effect of  Water  Vapor.   Water  vapor  is
involved in the formation of free radicals (reactive
compounds)  and  hydrogen  peroxide, which are
necessary for the formation of ozone.   Global
increases  in temperature  are expected  to  raise
tropospheric water vapor levels.

    If sources of water vapor are not perturbed by
vegetative changes, and if global circulation patterns
do not significantly affect precipitation events (an
unlikely assumption), then global water vapor levels
are   expected   to   increase  with   increasing
temperature. A temperature  increase of 2ฐC could
raise the water vapor  concentration by 10 to 30%
(Penner et al., Volume  F).   This  change should
affect  both oxidant  formation and sulfur dioxide
oxidation (acid deposition).
                             Smog chamber studies have shown that at high
                          pollutant  levels,  increases  in water vapor  can
                          significantly accelerate both the reaction rates of
                          VOCs and the rate of oxidant formation (Altshuller
                          and Bufalini, 1971).  Walcek (1988) has shown with
                          the use  of  a regional  acid  deposition model
                          (RADM)  that the ozone, hydrogen peroxide,  and
                          sulfate production rates in the boundary layer of the
                          troposphere all increase with increasing water vapor.

                             Effect  of Changes in Frequency and Intensity of
                          Stagnation Periods.   As noted previously, high-
                          pressure   systems  significantly  enhance  ozone
                          formation  potential.   During  a high-pressure
                          episode,   pollutants  are   exposed  to   high
                          temperatures and prolonged irradiation  (Research
                          Triangle Institute, 1975), resulting in high levels of
                          ozone.  If the intensity and  frequency  of high-
                          pressure  episodes increase  with  global  warming,
                          then ozone levels can be expected to be even higher.

                             Effect  of Changes in Mixing Layer Thickness.
                          As shown in Figure 11-7, increases in the mixing
                          layer height decrease ozone formation, presumably
                          because there are less ozone precursors per volume
                          of atmosphere.  An increase of global temperature

Chapter 11

would probably lead  to  an increase in average
mixing  depths  as  a result of greater convection,
which raises the mixing depth and increases mixing.

    Effect of Changes in Reactant Concentrations.
The concentrations of ozone precursor pollutants
(VOCs, NOX) play a large part in determining the
amount  of  ozone produced.    With  increasing
temperature, natural hydrocarbon emissions are
expected  to increase.   Also, unless  preventive
measures are  taken, manmade  emissions would
increase (vapor pressure of VOCs increases with
increasing temperature). If these  ozone precursors
increased in concentration, ozone production would

    Lamb et  al. (1985) have shown that natural
hydrocarbon (VOC)  emissions  from  deciduous
forests would increase by about a factor of three
with  a temperature  change  from  20  to  30ฐC.
However, as discussed in Chapter 5: Forests, the
abundance of some deciduous forests could decline
because of global warming. However, grasslands or
shrubs  that replace   forests  would  still  emit
hydrocarbons. The net effect is probably uncertain.
Emissions of NOX from powerplants would  grow
because of a greater demand for  electricity during
the summer months.  Soil microbial activity is also
expected to  increase with increasing temperature.
This  will increase  natural  emissions  of  Npx.
Evaporative emissions of VOCs from vehicles and
refueling  would also  be expected  to  rise  with
warmer temperatures.  However,  exact predictions
of the effects of all these factors on ozone formation
are  difficult to make because  the  relationship
between precursor emissions and ozone is extremely
complex and not  fully understood, and because
increases in emissions are difficult to quantify.

    An  example  of  this  complex  relationship
between ozone and its precursors is shown in Figure
11-8 (Dodge, 1977).  At high VOC levels and low
NOX, adding or reducing VOCs has very little effect
on  ozone  formation.   Likewise,  when  NOX
concentrations  are high and VOC concentrations
are low, increasing NOX reduces  ozone formation
while lowering NO  increases ozone  formation.
Thus, VOCs and NOX must be examined together
when considering any ozone  reduction strategy
based on controlling ozone-forming precursors.

Stratospheric Ozone Change

    Changes in stratospheric ozone concentration
can  also  affect tropospheric  ozone  formation
           •0.2 0.4 O.6 O.8  1.0 1.2 1.4 1.6  I..8
        O.O  0.2 0.4 0.6 0.8  1.0 1.2 1.4  1.6  1.8 2.0
                      VOC, PPMC
 Figure 11-8. Ozone isopleths as a function of NOX
 and volatile organic compounds (VOCs) (Dodge,
 because stratospheric ozone regulates the amount of
 ultraviolet (UV) radiation available for producing
 ozone  hi  the  troposphere.   Stratospheric ozone
 absorbs UV light from the sun and decreases the
 UV energy striking  the  Earth's  surface.   When
 stratospheric   ozone   is   depleted  by  the
 chlorofluorocarbons (CFCs) generated by human
 activity, more  UV radiation reaches  the  Earth's
 surface, which increases  the photolysis rates1  of
 compounds  that  absorb  solar  radiation  (NO2,
 formaldehyde, acetaldehyde, O3, andH2O2). Faster
 photolysis produces more free radicals (high-energy
 species) that increase the amount of smog. Thus,
 less stratospheric ozone will lead to enhanced ozone
 formation in the troposphere.

    Modeling results for New York from Gery et al.
 (1987) show that tropospheric ozone increased when
 stratospheric ozone  decreased  (see Table 11-2).
. They also  show that H2O2 and peroxyacetyl nitrate
 (PAN) yields increased. ELOj is  a strong oxidant
 that converts SO2 to sulfunc acid, and PAN is an
 air pollutant that damages plants and irritates eyes.
 The  16.6  and 33.3%  decreases  (Table 11-2)  in
 stratospheric   ozone   far  exceed the  expected
 decrease   resulting  from  the  buildup of CFC
 concentrations. This  is especially true  since the
  Photolysis is the breakdown of chemicals as a result of the
 absorption of solar radiation.

                                                                                           Air Quality
Montreal Protocol  agreement  will  limit  CFC
production.   These high values of stratospheric
ozone depletion are  used  only for  illustrative

Changes in Tropospheric Hydroxyl Radicals

    Hydroxyl radicals (reactive compounds) are the
most   important  free  radicals  found  in  the
atmosphere.    These  reactive  compounds  are
responsible   for   removing  many  atmospheric
pollutants (such as CH4, VOCs, methyl chloroform,
CO) from the atmosphere (Penner et al., Volume
F). Without these free radicals, pollutants would
not be removed from the atmosphere  and would
build  up to higher levels (global heating would be
greater).  Hydroxyl radicals in the free troposphere
are produced primarily by the decomposition of
ozone by sunlight and the subsequent reaction of
high-energy  oxygen with water.   In  the urban
atmosphere, hydroxyl radicals are produced through
a  complex  series of  reactions involving  VOCs,
nitrogen oxides, and sunlight. The solar photolysis
of hydrogen peroxide also gives rise to hydroxyl
radicals.  This occurs in both urban and rural areas.

    The  effect  of  global  climate  changes  on
hydroxyl radical abundance is  unclear.  In urban
areas  with  increases  in VOCs  and NO ,  a
temperature increase will  increase hydroxyl radical
concentration.  Also, if natural hydrocarbons and
NOX increase in rural areas, hydroxyl radicals are
expected to increase.  However, if methane, CO,
and  natural hydrocarbons increase without  an
additional increase in NOX, then hydroxyl radicals
will be  depleted.  A definitive prediction on the
effect  of  increasing  temperature  on  global
concentrations of hydroxyl radicals cannot be made
at this time.

Effect of Climate Change on Acid Deposition

    Rainwater  and surface waters are more acidic
than natural background levels because of industrial
and mobile emissions of SO2 and MX, which form
sulfuric and nitric acids in the atmosphere. In the
air, sulfuric acid (HgSO,) is produced primarily by
.the reaction  of SO2 with hydroxyl  radicals (high-
energy species); in clouds, the oxidation of SO2 to
H2SO4 is more complex, involving reactions with
hydrogen peroxide  and other  dissolved oxidants.
Nitric  acid  (HNO3) is  produced  in  air  by the
reaction of hydroxyl radicals with NOX.
   Organic acids, such as formic and acetic acids,
are also formed in the atmosphere. However, their
relative importance to the acid deposition problem
is unknown at present. Because they are weak acids
(compared to H^SO^ and HNO,), their contribution
to the problem is expected to be much less than
that  of the inorganic acids (Galloway et al., 1982;
Keene et al., 1983,1984; Norton, 1985).

   The acids produced  in the atmosphere can be
"dry  deposited" to the Earth's surface as gases or
aerosols, or they can be "wet deposited" as acid rain.
Changes in total acid levels depend on changes both
in   atmospheric   chemistry  and  changes   in
precipitation.  Wet deposition is  affected most by
the amount, duration, and location of precipitation.
Since  the  direction  of   regional precipitation
changes is unknown, it is  not known whether acid
fain  will  increase  or  decrease  in the  future.
However,  many of the same factors  that affect
ozone formation will also affect the total deposition
of acids.

Temperature Change

   Higher temperatures accelerate  the oxidation
rates of SO, and NOX to  sulfuric and nitric acids.
Gery et al. (1987) have  shown that a temperature
rise  would also  speed  the formation  of H2O2,
increasing the conversion of SO, to sulfuric acid
(see  Table 11-2).    Hales  (1988) studied  the
sensitivity to a 10ฐ C  temperature  rise using the
storm-cloud model PLUVIUS-2.  Considering only
the chemistry  occurring with a 10ฐC temperature
rise,  sulfate production increased 2.5 times.  No
modeling   was   performed   at  more  modest
temperature increases (e.g., ~4ฐC);  however, it is
likely that oxidation would also increase  with a
smaller increase in temperature. The limiting factor
in the oxidation of SO2 appears to be the availability
of H2O2.   The model  also suggested  that  a
temperature increase would cause more sulfuric acid
to form near the sources where SO2 is emitted.

   Effect of Global Circulation Pattern Changes.
Potential  changes in global circulation patterns
would  greatly affect local  acid deposition,  because
they  would alter  ventilation  and precipitation
patterns. Galloway et al. (1984) have calculated that
over 30%  of the sulfur emissions from the eastern
United States  are transported to the north and
farther east. Changes in circulation patterns would

Chapter 11
affect this transport, although the  direction  or
magnitude of the effect is unknown.  '

    Effects of Changes in Emissions. If electricity
demand rises with rising temperatures (see Chapter
10: Electricity Demand), if more fossil fuels are
burned, and if technology is not improved, SO2 and
NOX emissions will increase.  An. approximate 10%
growth in use of electricity in the summer could
increase  SCL emissions  during the summer  by
approximately 30% if present-day technology is used
in the future. This, in turn,  would increase acid

    Effects of Reduced Stratospheric Ozone.  A
decrease in stratospheric ozone due to  CFCs may
increase acid deposition because more UV radiation
would be available to drive the chemical reactions.
As discussed above, a modeling study by Gery et al.
(1987) showed an increase in the yield of H2O2
when stratospheric ozone  was reduced  by 16 and
33%. Because FJ^Ou is a strong oxidant,  SO2 would
probably also be oxidized more quickly into sulfate
aerosols  and acid rain, but  this  depends on the
availability of water  vapor  (e.g.,  clouds,  rain).
Implementation of the Montreal  Protocol should
help reduce CFC emissions.

    Reduced  Visibility.   The growth  in  natural
organic emissions and increases in sulfates resulting
from warmer temperatures should reduce visibility,
assuming  that the frequency  of rain events, wind
velocity, and dry deposition rates remain the same.
If rain  events increase,  washout/rainout should
increase   and visibility would be  better  than
predicted (see Chapter 3: Climate Variability).

Study Design

    Morris et al.  (Volume F) applied a regional
transport  model  RTM-HI to an area  covering
central  California  and  a  region  covering  the
midwestern  and the southeastern United States.
The model was run for the present-day conditions
and for a future climate.  For California, Morris et
al. used input data from August 5-10, 1981; for the
Midwest  and the Southeast, they used input data
from July 14-21,  1981.  These were periods with
high ozone  levels and may be  most sensitive to
changes in climate.  The scenario assumed that
temperatures would be 4ฐC warmer than in the base
case,  but  all other climate variables were  held
constant (relative humidity was held constant). The
scenario assumed no change in emission levels, no
change in boundary layer, and no change in  wind

   The RTM-III is a three-dimensional model that
represents  point  sources  embedded in a  grid
framework. The model has three prognostic vertical
layers and a diagnostic surface layer.   This means
that  the  surface layer is represented by actual
observations. The other three layers are predicted
by using the surface layer data.  The photochemical
reactions are based on the latest parameterized
chemical mechanism.


   Perhaps the most  important limitation is that
emission levels were held  constant. It is likely that
future emission levels will be different,  although this
study did  not estimate how.  The results of this
study are useful for indicating the  sensitivity  of
ozone formation to temperature, but should not be
considered as a prediction of future ozone levels.
The  model ignored future  increases  in emissions
that would occur with increased temperatures. The
estimates for ozone are only coarse approximations.
Morris et  al. used the National Acid  Precipitation
Assessment Program (NAPAP) emissions data of
1980.  These data appear  to underestimate actual
ratios of VOCs to NOX as  measured in urban areas.
Ching et al. (1986) state  that for  most cities, the
NAPAP data underestimate VOC emission values
by a factor of three or more. The model simplified
some  reactions of  the  hydrocarbons  (VOCs)
because the chemistry is not well known.

   This study  did  not estimate  climate-induced
alterations in most meteorological variables, except
temperature  and  water .vapor, which  is an
oversimplification. For example, this study assumed
that  the mixing heights remain unchanged for the
temperature increase  scenario; in reality,  mixing
heights could increase with  rising   temperature.
Holding  the mixing  heights  constant  probably
overemphasized the importance of temperature in
oxidant production, because an increased  mixing
layer depth might have had a dilution  effect.  Also,
as stated  earlier,  cloud  cover  will  affect ozone
production.  If cloud cover increases, then ozone is
expected to decrease. Frequency and intensity of

                                                                                         Air Quality
stagnation periods can also have profound effects on
ozone formation.  This modeling exercise did not
consider these factors.


Central California Study

    Table 11-3 summarizes the results from the
base case scenario and a climate sensitivity scenario
that used a 4ฐC temperature increase  and  an
attendant  increase in water vapor concentration.
All of the days studied show a larger area exposed
to high levels of ozone. An increase in temperature
may lengthen  the duration of  high ozone  levels,
although the maximum  levels may be the  same.
Figure 11-9  illustrates the August 6 base case and
climate sensitivity case.   The temperature change
increased   the   August   6   maximum   ozone
concentration  from 15 parts per hundred million
(pphm) to 18 pphm, a 20% increase in ozone. The
area in which the NAAQS was exceeded almost
doubled from 3,700 to 6,600 square kilometers.

    The temperature increases in the two main
cities in  the  San  Joaquin Valley (Fresno  and
Bakersfield) resulted in an approximate 0.5-pphm
increase (approximately  8%)  in maximum daily
ozone concentration. In regions farther away from
the emissions, such as the Sierra Nevada Mountains,
little change in ozone levels was observed with the
increased temperature.

Midwest and Southeast Study .        •

   The results  from  applying RTM-in  to  the
midwestern  and southeastern areas are  shown in
Table 11-4.  On one particular day (July 16), raising
the temperature caused maximum ozone to increase
from  12.5 pphm to 13.0  pphm  (Figure  11-10).
Although this is  only a slight increase  (0.5 pphm),
the predicted  area of exceedance  of the  ozone
NAAQS increased by almost a factor of three, from
9,800 to 27,000 square kilometers.  The differences
occurred mainly in the upper Midwest. In general,
the results range from a reduction of 2.4% to an
increase of  8.0% in  ozone levels.   Although a
temperature increase will generally increase ozone
formation, it is noted  in Table 11-4; that on two
days, July 14 and July 21, no ozone increases were
observed.  This occurs  when there are insufficient
precursors to sustain ozone formation. Under these
conditions, ozone is produced more quickly with
increasing  temperature but the  total amount
produced need not be  greater and could even be
less in some cases.

   Both  modeling   exercises   indicate   that
temperature change alone could increase  ozone
levels  over  what they would be  without  climate
Table 11-3.   Maximum Daily Ozone Concentrations Predicted by the RTM-III for Each Day of the Central
             California Modeling Episodes for the Base Case and the Case of Climate Sensitivity to Increased
             Temperature of 4ฐC
                                           Maximum daily ozone concentrations (pphm)
Date of Episode
August 5
August 6
August 7
August 8
August 9
August 10
4ฐC temperature
Source:  Morris et al. (1988)

 Chapter 11
                     Base Case
                   August 6, 1981
                                                       Exceeds Standard
                               <6    >6   >8
Climate Sensitivity Scenario No. 1
   4ฐC Temperature Increase
Figure 11-9.  Comparison of estimated maximum daily ozone concentrations (pphm) for the base case and
climate sensitivity scenario No. 1 (temperature and water increase) for August 6, 1981 (Morris et al., 1988).
Table  11-4.  Maximum Daily  Ozone  Concentrations  Predicted  by the  RTM-III  for Each Day of the
             Midwestern/Southeastern Episode for the Base Case and the Case of Increased Temperature of
             4ฐC                     .•••:'...                  .    .    .
Date of Episode
July 14
July 15
July 16
July 17
July 18
July 19
July 20
July 21

Maximum daily ozone concentrations
4ฐC temperature
, • • . • 5 •. ,.-...••,•
1 13.0 •'.•'•
' 11.2 .
, increase "".' , ,
. :. 'o'.o ,,.
•:.-.:"•.•; 3.5: '•••" 	 •
." :.' 2.6 •'; :- ;;. ..
: 8.0
• •"•"•. '72'
-2.4 V
Source: Morris et al. (1988).

                                                                                      Air Quality
                      Base Case
                                                          Climate Sensitivity Scenario
Figure 11-10. Comparison of predicted estimated maximum daily ozone concentrations (pphm) for the base case
and climate sensitivity scenario No. 1 (temperature and water increase) for July 16, 1980 (Morris et al., 1988).
Population Exposure

    As discussed  above, both the California and
Midwest/Southeast  studies  show  a  significant
increase in the area that is potentially exposed to
higher levels of ozone when the temperature is
increased as compared with base case conditions.
Data  taken from the 1980 census from  central
California and the midwestern and  southeastern
areas were used to determine the number of people
exposed to ozone for the base case and a 4ฐC
temperature rise scenario. Table 11-5 presents the
number of  people-hours of exposure to ozone
concentrations exceeding 8,12, and 16 pphm. These
estimates of human exposure were generated  by
multiplying the number of people in the grid cells by
the total number of hours that the estimated hourly
ozone concentration in those grid cells exceeded the
8-, 12-, or 16-pphm levels.  Actual exposure levels
may be less because indoor levels are generally
lower  than ambient air levels.


   An increase in ozone  levels  due to climate
change is important for several reasons:

   •  Ozone itself is a radiatively important gas
      and contributes to climate  change.  Ozone
      absorbs infrared energy much like carbon
      dioxide. It has been calculated that a 15%
      increase in tropospheric ozone could lead to
      a  0.1ฐC   rise  in  global   temperature
      (Ramanathan et al.,  1987).

   •  Ozone levels in many areas are just below
      the current standard. If emissions are not
      reduced, any  increase  in ozone  formation
      may push levels above the standard.

Chapter 11
Table 11-5.  Number of People-Hours of Exposure to Ozone Concentrations in Excess of 8,12, and 16 pphm
            for the Base Case and the Case of Climate Sensitivity to Increased Temperature
  Exposure to         Exposure to       Exposure to
  O3 > 8 pphm        O3 > 12 pphm     O3 > 16 pphm
            Base case

      Central California Modeling Episode

   70,509,216            660,876              0
            Base case

   Midwestern/Southeastern Modeling Episode

1,722,590,208          29,805,348              0
Source: Morris et al. (1988).
    •  Many inexpensive controls for ozone are
       already in place in nonattainment areas.
       Increases in ozone levels would require
       relatively expensive measures to sufficiently
       reduce ozone  precursors to  attain  the

    •  The standard itself is defined in terms of
       the highest levels of ozone experienced in
       an area, not average levels.  (As a yearly
       average, no  area  of the country would
       exceed the standard of 0.12 ppm.)  Thus, a
       factor such as temperature that may have a
       modest effect on average levels of ozone
       formation may have a much more significant
       effect on peak levels.

    A rough estimate of each of these factors can
illustrate the potential policy problems created by
a rising temperature scenario. The data in Figure
11-9 suggest that 4ฐC degree rise in temperature
may  lead   to  an   increase   in   peak  ozone
concentrations of around 10%.  A 10% increase in
                      peak ozone levels could affect a number of potential
                      ozone violations. In the 1983-85 period for example,
                      68 areas showed  measured  exceedances  of the
                      ozone air quality standards (for technical and legal
                      reasons,  not  all  these  areas  were  officially
                      designated nonattainment areas). A 10% increase
                      in ozone levels in that period doubled the number
                      of nonattainment areas to 136. This would  include
                      41 new metropolitan statistical areas (MSAs) added
                      to the  list  and  27 non-MSAs.    These new
                      nonattainment areas  would add most midsize and
                      some small cities in the Midwest, South, and East to
                      the list of nonattainment areas.

                         The  policy implications of this should be put
                      into context because the full effect of climate change
                      may  not be  felt until well  into the next century.
                      Over  the  next  several decades, various national
                      measures to reduce  ozone precursors, such as a
                      reduction in the volatility of gasoline, may  go into
                      effect. These would provide a cushion to marginal
                      areas and  could  offset  a  temperature   effect.
                      However,   other   factors  suggest  that   rising
                      temperatures could be a problem.

                                                                                           Air Quality
     Ozone levels and ozone precursors are closely
 related to  economic  expansion  and  population
 growth. Consumer solvents (e.g., paints, sprays, and
 even deodorants) are a major source of ozone
 precursors. These are very difficult to control and
 are likely to increase in the future in areas currently
 attaining the standards.  Growth in other  sources
 of  ozone  precursors  would bring  many areas
 relatively close to the limits of the ozone standard.
 Gradual increases in temperature would make
 remaining  in compliance with the standard more
 difficult.   Although  any sudden change  in the
 number  of nonattainment areas as  a result of a
 secular trend  toward increased  temperature  is
 unlikely, a number  of  small  to midsize cities
 eventually  may be forced to develop new  control

     The implications of  warmer temperatures for
 existing nonattainment areas can also be estimated.
 In  these  areas,  existing  and  planned  control
 measures  may not be  adequate to  reach  the
 standard, if additional ozone  forms.  In the past,
 EPA  has  attempted  to   project the  emission
 reductions  and costs associated  with the attempts
 of existing nonattainment areas to  reach the ozone
 standard. Using the same modeling approach, the
 effects of a temperature  increase were analyzed to
 estimate the additional tons and  costs associated
 with a projected temperature  rise.  Extrapolations
 of existing inventories to  the year 2000 suggest that
 higher temperatures could  require an additional
 reduction of 700,000 tons  of VOC from an inventory
 of about 6 million tons.  Given  that most  current
 nonattainment areas already will have implemented
 the most inexpensive measures, these additional
 reductions may cost as much as $5,000 per  ton per
year. Their aggregate cost could be as much as $3.5
billion each year.

    These   conclusions  should   be   viewed  as
 preliminary.  Nonetheless, they demonstrate that
the  potential  economic  consequences could be
significant  for  an already expensive  program  to
 combat ozone.

Acid Rain

    The global climate  change  is likely to affect
acidic  deposition in the near future  for  several
    First, emissions from  fossil  fuel powerplants
 both influence acid rain and contribute to global
 warming.   In the  future, global warming may
 increase energy demand and associated emissions.
 Because the growth in demand  for  electricity in
 northern  states   (see   Chapter  10:  Electricity
 Demand) may be  lower than in southern states,
 regional shifts in emissions  may occur  in the future.

    Second, global climate  change would influence
 atmospheric reaction rates and the deposition and
 form of acidic material.  It is  conceivable that
 regions of high deposition  may shift or that more
 acid  rain  may  be  transported  off the  North
 American continent. Strategies that seek to control
 powerplants in regions near sensitive areas may or
 may not be as effective, as global climate change

    Third,  global  climate  change may  alter the
 impacts of acid rain on ecological and other systems
 in as yet unpredictable ways.  For example:

    •  Changes in the amount of rainfall may dilute
       the effect of acid rain on many  sensitive

    •  Changes in clouds may alter the fertilization
       of high-elevation forests.

    •  Changes in humidity and frequency of rain
       may alter degradation rates for materials.

    •  Increased midcontinentaldryness would alter
       the amount  of calcium and magnesium in
       dust, neutralizing impacts on soils.

    •  Increased  numbers  of days without  frost
       would  decrease forest damage  associated
       with   frost   and   overfertilization  by
       atmospheric nitrogen.

    •   Changes in snowpack and the seasonality of
       rainfall would change acid  levels in streams
       and alter the timing and magnitude of spring
       shocks  on aquatic species.

    Finally,  solutions  to  both   problems  are
inextricably linked.  Some solutions, such as SO2
scrubbers and clean coal technologies, may abate
acid rain levels, but they may do little to improve air
quality or  may increase global warming. Other

Chapter 11
solutions, including increased energy efficiency and
switching fuels  to natural  gas  or to renewable
energy sources,  may provide positive solutions to
both problems.

    In  summary, an  examination  of  the  time
horizons of importance to both acid rain and global
climate change problems  suggests that these two
issues should not be viewed in isolation. Emissions,
atmospheric reaction rates, pollutant transport, and
environmental impacts  will likely be altered by
climate change.  This suggests that a more holistic
approach must be taken to air pollution problems
and that proposed solutions should be evaluated on
the basis of their contributions  to solving both

    The Environmental Protection Agency issues
air pollution regulations to improve air quality and
to protect public health and welfare.  In general,
current regulations to reduce oxidant levels will also
provide positive benefits toward a goal of limiting
the rate of growth in global warming.   Other
programs aimed at reducing carbon monoxide
levels, particularly from mobile sources, or CFCs to
protect the stratospheric ozone layer, also positively
affect  greenhouse gases and  the  rate  of  global
warming.  However, the regulatory activities of the
Agency have not been retrospectively reviewed to
determine their impacts on global warming.  In
some cases, there may be important benefits; for
example, current emission standards for automobiles
do not encourage more efficient use of gasoline. A
different  form   of  standard,  while  potentially
disruptive to  air  pollution efforts,  might produce
positive greenhouse gas benefits via reduced energy
consumption. These issues will have to be analyzed
in the future.

    Because  of  the  climate  change issue, the
following are some of the more important policy

    •  Air pollution control agencies such as EPA
       should  undertake  a  broad  review  to
       determine the impact of  global climate
       change  on air pollution  policies.   In
       particular, the  cost   of  added  controls
       resulting  from climate change should be
determined, perhaps as each significant regulation
is proposed or reevaluated.

    •   The impact of EPA regulations, particularly
       the impact on energy use; and greenhouse
       gases, should be a more important weight in
       future  regulatory  decisions.  Since  EPA
       regulations often serve as models for  other
       countries, the cost penalty for better energy
       usage, while sometimes small in the United
       States, may be important on a global basis.

    •   Future  reports to Congress  and  major
       assessments of ecological effects,  e.g., the
       1990 Acid Deposition Assessment document,
       should  include   sensitivity  analyses  of
       alternative  climates.   Risk  management
       decisions of the Agency could then be made
       with improved knowledge of climate impacts.

   Some  of the  key questions  that need to be
resolved regarding climate change and air quality
include the following:  How important will climate
change be  relative to  other  factors such as
population growth to future air pollution problems?
Is the impact of  climate change  likely  to be
significant enough to require totally different air
pollution strategies? What mix of control strategies
could be most cost effective in reducing acid rain,
global warming, tropospheric ozone,  and other
pollution problems? The research elements needed
to address  these issues  include basic research,
sensitivity analyses, full-scale atmospheric modeling,
and  cost-effectiveness  studies.   Examples  are
presented below:

Basic Research - There is  an important need to
understand how manmade and natural emissions of
hydrocarbons and other pollutants might change in
the future when  temperature,  CO2,  and  UV-B
radiation increase  and other climate  parameters

Sensitivity   Analyses   -   Analyses   of   ozone
concentrations are dependent on boundary layer
height, clouds, water  vapor,  windspeed,  UV-B
radiation,  and other parameters.  Sensitivity tests
using  single  models   could   improve   our
understanding of the relative importance  of these

                                                                                         Air Quality
 variables and could provide important information
 for general circulation modelers.   ;-,    ;

 Full-Scale  Modeling - Complete understanding of
 the interactions of climate change and air quality
 will ultimately  require  that  general .circulation
 models and mesoscale chemistry models be linked
 iri some direct or indirect manner. This will require
 the development of innovative approaches between
; the general circulation and air pollution modeling

 Cost-Effectiveness Studies - There are currently a
 number of congressional proposals to improve the
 Clean Air Act and to reduce global climate change.
 To assume that both air quality and global climate
 change-goals are  achieved, '.analyses, of-the cost-
 effectiveness of  alternating   strategies  will be
 necessary./,     ,   „ ••.'•••..'..    ., ;;  ,

 Altshuller,  A.P.,   and  J.J.   Bufalini.   1971.
 Photochemical  aspects  < of  air   pollution.
 Environmental Science and Technology 5:39-64.
          ' . " '       i l    . i   ,••'.-'.";   -   \ "
 Ching, J.K.S.,  J.H. Novak, K.L. Schere, and FA.
 Schiermeief.   1986.    Reconciliation   of  Urban
 Emissions   and  Corresponding   Ambient   Air
 Concentrations Using Mass'Flow.Rate Technique.
 Draft report prepared for EPA/ORD; April.'

 Dodge, M.C. 1977. .Effect, of Selected Parameters
 oh Predictions of a Photo-Chemical Model. EPA
 600/3-77/048.   Research Triangle Park, NC: U.S.
 Environmental Protection Agency.  June.      •

 Galloway, J.N., G.E. Likens, W.C. Keene, and M.M.
 Miller. 1982. The composition of precipitation in
 remote areas of the world. Journal of Geophysical
 Research 87:8771-8788.

 Galloway, J.N., D.M. Whelpdale, and G.T. Wolff.
 1984. The flux of sulfur and nitrogen eastward from
 North America.    Atmospheric    Environment
 18:2595-2607.   .   ;         -; •..,-.	

 Gery, M:W., R.D. Edmond, and G.Z. Whitten. 1987.
 Tropospheric Ultraviolet Radiation: Assessment of
 Existing  Data and Effect on Ozone Formation.
 EPA Report 600/3-87/047.-  Research Triangle
 Park, NC: U.S. Environmental Protection Agency.

 Hales, J. 1988. In: Sensitivity of Urban/Regional
 Chemistry to  Climate Change:  Report  of the
 Workshop, Chapel Hill, NC.  Wuebbles, D.J., and
 J.E.  Penner,  eds.   Livermore,  CA: Lawrence
 Livermore National Laboratory. Feb. 17-18.

 Holzworth,  G.C.     1972.    Mixing  Heights,
 Windspeeds and Potential for Urban Air Pollution
 Throughout the Contiguous United States. Research
 Triangle Park, NC: U.S. Environmental Protection
.Agency, Office of Air Programs.  Publication No.

 Jeffries, H.E., K.G. Sexton, J.R. Arnold, and T.L.
 Kole.     1989.    Validation   Testing  of  New
 Mechanisms with Outdoor Chamber Data, Volume
 4: Appendixes to Photochemical Reaction Photolysis
.Rates in the UNC Outdoor Chamber. EPA/600/3-
 89/010d.  .Research  Triangle Park,  NC:  U.S.
 Environmental Protection Agency.

 Kamens; R.M., H.E. Jeffries, K.G. Sexton, and AA.
 Gerhardt.  1982.  Smog Chamber Experiments to
.Test Oxidant-Related Control Strategy Issues. EPA
 600/3-82-014.  Research Triangle Park, NC: U.S.
.Environmental Protection Agency. August.

 JCeehe, :w.C., J.N. Galloway, and J.D. Holden Jr.
,1983.   Measurement  of weak  organic acidity in
 precipitation  from remote areas  of  the  world.
 Journal of Geophysical Research 88:5122-5130.

 Keene;  W.C., and J.N. Galloway.  1984. Organic
 acidity   in  precipitation   of  North  America.
Atmospheric Environment 18: 2491-2497.

Lamb B.K., H.H. Westberg, T. Quarles, and D.L.
Flyckt  1985.  Natural Hydrocarbon Emission Rate
Measurements   From   Selected   Forest   Sites.
EPA-600/3-84-001.  Research Triangle Park, NC:
U.S. Environmental Protection Agency.  October.

Norton, R.B.  1985. Measurements of Formate and
Acetate  in  Precipitation  at  Niwot Ridge  and
Boulder, Colorado.  Geophysical Research Letters

Ramanathan,  V., L. Callis, R. Cess, J. Hansen, I.
Isaksen, W. Kuhn, A. Lacis, F. Luther, J. Mahlman,

Chapter 11
R. Reck,  and M. Schlesinger. 1987.    Climate-
chemical  interactions  and  effects of  changing
atmospheric  trace  gases.    Review  Geophysics

Research Triangle Institute. 1975. Investigation of
Rural  Oxidant  Levels As  Related  to  Urban
Hydrocarbon Control Strategies.  EPA-450/3-75-
036.     Research  Triangle   Park,   NC:   U.S.
Environmental Protection Agency.  March.

Samson, PJ.  1988.   Linkages Between Global
Climate Warming and Ambient Air Quality. Paper
presented   at  the  Global   Climate  Linkages
Conference, Washington, DC;  Nov. 15-16.
Seilkop, S.K.,  and P.L. Finkelstein.  1987.  Acid
precipitation patterns and trends in eastern North
America, 1980-84. Journal of Climate and Applied
Meteorology 26(8):980-994.

U.S. EPA. 1988.  U.S. Environmental Protection
Agency,  Office  of Air  Quality  Planning  and
Standards.  National Air Quality  and Emissions
Trends Report, 1986. Research Triangle Park, NC:
U.S. Environmental Protection Agency. EPA report
No. 45014-88-001.  Research Triangle Park, NC:
U.S. Environmental Protection Agency.

Walcek C. 1988. In: Sensitivity of Urban/Regional
Chemistry to Climate Change: Report of Workshop,
Chapel Hill, NC. Wuebbles, DJ. and J.E. Penner,
eds.  Livermore, CA: Lawrence Livermore National
Laboratory. Feb. 17-18.

                                      CHAPTER 12
                                 HUMAN HEALTH

Global warming may lead to increases in human
illness (morbidity) and mortality during summer.
Populations at particular risk are the elderly and
very young (age  1  year  and below), particularly
those who are poor and/or homeless. These effects
may be more pronounced in some regions than in
others, with northern regions more vulnerable to
the effects of higher temperature episodes than
southern regions.   Milder winters may offset
increases in morbidity and mortality, although net
mortality may increase.  Mortality in southern cities
currently shows a lesser  effect  from heat waves,
presumably because populations  have acclimatized.
If  northern   populations   show   this   same
acclimatization, the  impact  of global warming on
summer mortality rates may be substantially lower
than estimated. The full scope  of the impacts of
climate change on human health remains uncertain
and is a subject for future research.

•   Although there may be an increase in weather-
    related summer deaths due to respiratory,
    cardiovascular,  and cerebrovascular  diseases,
    there may be a decrease in  weather-related
    winter deaths from the  same diseases.  In the
    United States, however, our studies suggest that
    an  increase  hi weather-related  deaths   hi
    summer would be greater than the decrease in
    weather-related deaths in winter. To draw firm
    conclusions, however, this area needs additional

•   Sudden changes in temperature are correlated
    with increases  hi deaths.    So  if  climate
    variability increases,  morbidity and mortality
    may also increase.  Conversely, a decrease  in
   , the frequency or intensity of climate extremes
    may be associated with a decrease in mortality
    and morbidity.

•   Seasonal variation hi perinatal mortality and
    preterm birth (higher hi the summers, lower
     hi the winters) have been observed hi several
     areas in the United States. The longer and
     hotter summers that may accompany climate
     change could increase infant mortality rates,
     although changes hi variability may be more
     important   than   average   changes   in

     Vector-borne diseases, such as those carried
     by ticks, fleas, and mosquitoes, could increase
     hi certain regions and decrease hi others.  In
     addition, climate change may alter habitats.
     For  example, some  forests  may  become
     grasslands, thereby modifying the incidence of
     vector-borne diseases.

     While  uncertainties  remain  about  the
     magnitude of other effects, climate change
     could have the following impacts:

     — If some farmland is abandoned or some
       forests become grasslands, a result could be
       an increased amount of weeds growing on
       cultivated land, and a potential increase hi
       the incidence of hay fever and asthma.

     — If humidity increases,  the  incidence and
       severity of skin infections and infestations
       such as ringworm, candidiasis, and scabies
       may also rise.

     — Increases hi the persistence and level of air
       pollution episodes associated with climate
       change may have  adverse health effects.

    Human illness and mortality are linked hi
many ways  to the environment (Figure 12-1).
Mortality rates, particularly for the aged and very ill,

Chapter 12
               Figure 12-1. Schematic showing how climate change can affect human health.
are influenced by the frequency  and severity  of
extreme temperatures. The life cycles of disease-
carrying insects, such as mosquitoes and ticks, are
affected by changes in temperature and rainfall,  as
well as by modifications in habitat that result from
climate change. Air pollution, frequently associated
with  climate  change, is  known to increase the
incidence or severity of respiratory diseases such as
emphysema  and  asthma   A variety of human
illnesses  show  sensitivity  to  the  changes  in
temperature (and/or humidity) that  accompany
changes in  season.   Stroke and  heart  attacks
increase with  very cold  or  very warm weather.
Allergic diseases  such as asthma and hay fever
increase in spring and summer when pollens are
released.  Diseases spread by insects such as St.
Louis encephalitis1  increase  in the warmth of
 StXouis encephalitis is an example of a vector-borne disease.
Such diseases are spread to humans or animals by arthropods
(e.g., mosquitoes or ticks).  The disease-causing organism, such
as a virus, is carried and transmitted by the vector, also known
as the agent. Some vectors, such as ticks, live on other animals,
such as deer and birds, which are called intermediate hosts.
For example, Lyme disease is caused by a bacteria (the agent),
which is carried by a certain type of tick (the vector), which
lives on deer and mice (the intermediate hosts).
summer when the mosquitoes that transmit it are
active. In addition, adverse effects on reproduction,
such as increased incidence of premature births,
show a summertime peak in some cities. Table 12-
1 lists the number of deaths  and the number of
physician  visits (used to estimate the incidence of
illness associated with a given effect) associated with
major causes of mortality and illness in the United

General Mortality and  Illness

     The   relationship  between  mortality and
weather has  been  studied for ; over  a century
(Kutschenreuter, 1959; Kalkstein,  Volume G), with
the relationship between mortality and temperature
receiving the most attention. Kutschenreuter (1959)
observed  "mortality is higher  during cold winters
and hot summers and lower during  warm winters
and cool summers."  The people  most sensitive to
temperature extremes are the elderly (White and
Hertz-Picciotto, 1985).   One explanation is the
increased susceptibility of the elderly is  that for
individuals  already  stressed  by the  circulatory
problems associated with vascular and heart disease,
heat waves (temperatures  above  100ฐF for   5

                                                                                        Human Health
              . Table 12-1.   Major Causes of Illness and Mortality in the United States (1984)a
Cause of illness or mortality
 number of
                                                                             Estimated mortality
Accidents and adverse effects
Cerebrovascular diseases'5
Chronic liver disease
  and cirrhosis
Chronic obstructive pulmonary
  diseases and allied conditions
Congenital anomalies
Diabetes mellitus
Heart diseases
Malignant neoplasms
Pneumonia and influenza
Suicides, homicides

  Total for potentially weather-
  sensitive diseases

   • Total all causes









^Causes are presented in alphabetical order and therefore are not ranked by severity.
 Conditions that can be influenced by changes hi weather and climate are indicated in bold type.
Source: CDC (1986).                                    ::         '      r
consecutive days) "overload"  the thermoregulatory
system,  which  is   struggling to  maintain  the
appropriate body temperature.  This results in heat
stress, heatstroke,  and  often mortality as well
(White and Hertz-Piccibtto, 1985).

     In addition to the elderly, people working in
hot  environments,  such as  steel  mills  and
construction sites, are  at  special risk from heat
waves  (Dukes-Dobos, 1981).   These workers face
even greater risk if  they have underlying medical
problems such as impaired circulation; higher than
normal body temperature due to disease; chronic
diseases such as alcoholism, diabetes, and obesity;
or other problems.

Cardiovascular, Cerebrovascular, and
Respiratory Diseases

    .Although  much of  the  earlier information
characterized the relationship between weather and
         total mortality from all causes, a growing body of
         literature evaluates the relationship of weather to
         specific Causes of death.  For example, changes in
         weather have been associated with mpacts on the
         cardiovascular,  Cerebrovascular, and  respiratory
         systems.   As previously shown in  Table 12-1,
         diseases of these three systems cause the majority of
         deaths  observed  on a yearly basis in the United
         States, as well as  significant illness;  Incidences of
         these diseases rise as climate extremes increase.

             The relationships  of weather  variables  to
         diseases  of  these'.' systems   are  diverse  and
         complicated.  Weather is 'riot the  main  causative
         factor in these diseases but,  rather, changes  in
         weather have an impact because they add stress to
         systems that have already been compromised for
         some other reason(s). For example, although it has
         been observed that  deaths  in individuals  with
         diseases of the cardiovascular system go up with
         heat waves, the .precise reason for this relationship
         is not known.     ' •> ', •.

Chapter 12

     To  understand  the  relationship  between
weather and these diseases, one must examine the
specific diseases that come under broad categories
such as "cardiovascular disease." For instance, heart
attack, coronary heart disease, and possibly coronary
arteriosclerosis and rheumatic heart disease  are
apparently sensitive  to  changes in temperature
(particularly cold and heat waves), whereas ischemic
heart disease is not (Vuori, 1987).

     That these different relationships  exist is not
unexpected given that different parts of the system
are   compromised  (e.g.,   the   arteries   in
arteriosclerosis and the heart muscle in rheumatic
heart disease), and that  different causes are also
likely (e.g., an infection-related process in rheumatic
heart  disease   and   diet   and   heredity   in
arteriosclerosis).   What this information  does
indicate,  however, is  that these relationships are
very complex and that unraveling them to predict
the  effects   of  global  warming  will  require
considerable analysis (Lopez and Salvaggio, 1983).

     The relationship between temperature changes
and illness (morbidity) from diseases such as heart
attack and  stroke is  not as well defined as the
relationship reported for mortality. Mortality has
national  reporting procedures, whereas morbidity
must be estimated from  such data  as  hospital
admission figures. A few studies have evaluated the
relationship of weather to hospital admissions from
cardiovascular or cerebrovascular  disease. These
have shown a relationship to weather changes, e.g.,
an increase in admissions for cardiovascular effects
with  heat waves,  similar to that  observed for
mortality (Sotaniemi et al., 1970; GUI et al., 1988).

     Morbidity  from   respiratory   diseases   is
somewhat  easier to estimate, principally because
two such diseases, asthma and hay fever,  affect as
much as  3  and  6%   of  the  U.S.  population,
respectively, causing significant losses of work time.
The most common seasonal pattern for the allergic
type of asthma and for  hay fever is an increased
springtime occurrence in response to grass pollens.
A nonseasonal form  of allergic asthma  may also
occur in response to allergens such as molds, which
are affected  by changes  in  precipitation  and

Vector-Borne Diseases

     Two  tick-borne diseases currently  posing a
public health problem in the United States, Rocky
Mountain spotted fever and Lyme disease, induce
similar initial symptoms: high fever, chills, headache,
backache, and profound fatigue.  Rocky Mountain
spotted fever can eventually result in hemorrhagic
areas that ulcerate, and Lyme disease may cause
permanent neurologic, cardiac, and rheumatologic
abnormalities (APHA, 1985). The ticks that spread
these  diseases,  and  therefore  the  geographic
distribution of the diseases themselves, are affected
both directly and  indirectly by climate variables.
Such  environmental  factors  as  temperature,
humidity,  and  vegetation   directly  affect  tick
populations and the  hosts of the tick populations,
e.g., deer, mice, and birds.

     Mosquito-borne diseases, such as malaria and
certain types of encephalitis  (inflammation of the
brain), are not a major  health  problem in the
United  States  today because  occurrences are
relatively rare.   However,  mosquitoes  are  also
weather-sensitive insects  favoring a warm, humid
climate.  The spread of mosquito populations and
the diseases they carry depends in part upon such
climate factors as temperature  and humidity, and
upon vegetation, which is also influenced by the

Human Reproduction

     Preterm delivery and perinatal mortality (i.e.,
death just before, during, or just after birth) are two
adverse reproductive outcomes  that are associated
with particular seasons and, thus, might be affected
by climate change. Statistically significant increases
in preterm births and in perinatal mortality in the
summer months have been documented (Keller and
Nugent,  1983; Copperstock  and Wolfe, 1986) (see
Figure 12-2).  The data  on total perinatal deaths
correspond closely with those on perinatal deaths
associated with  infection in the mother or infant,
suggesting that the observed seasonally in perinatal
death is  linked to a seasonality  of  reproductive
infections (Keller and Nugent, 1983).

     To assess the  effects  of  climate change  on
 human health, EPA sponsored three studies for this
 report  (Table 12-2).   Longstreth and  Wiseman
 (Volume G) reviewed the literature on the role of
 climate, season,  and  weather variables  in  the

                                                                                        Human Health
               F M  A  M
                        J  J  A S  O  N D
Figure 12-2. Probabilities of (A) perinatal death or
(B) preterm delivery (Keller and Nugent, 1983).
Table 12-2.  Studies Conducted for This Report
     The Impact of C(X. and Trace  Gas-Induced
     Climate Change  Upon Human  Mortality -
     Kalkstein, University of Delaware (Volume G)

     Computer Simulation of the Effects of Changes
     in Weather Pattern on Vector-Borne Disease
     Transmission  - Haile, UJS.  Department of
     Agriculture (Volume G)

     The Potential Impact of Climate Change on
     Patterns of Infectious Disease in the United
     States   -   Longstreth   and   Wiseman,
     ICF/Clement Associates, Inc. (Volume G)
incidence of, and mortality due to, vector-borne
diseases. In November 1987, they also conducted a
workshop of scientists  to evaluate  the potential
impacts of global climate change  on vector-borne
infectious diseases in the United States. Following
the  workshop,  Haile  (Volume   G)  conducted
modeling studies of the potential impact of climate
change on (1) the distribution of the American dog
tick, the vector of Rocky Mountain spotted fever;
and (2) the potential for malaria transmission in the
United States.   The third study, by  Kalkstein
(Volume G), as an extension of an earlier modeling
study that assessed the potential effects of global
climate change on the elderly and on total mortality
in New York (Kalkstein  et al., 1986).  Kalkstein
(Volume  G) expanded the New York analysis  to
include 14 other cities. A detailed  review of these
three studies, supplemented with other information
from the literature, is presented in this section.

General Mortality

     Preliminary analyses suggest that unless the
U.S.  population becomes fully  acclimatized2  to
higher  temperatures, climate change  will  be
associated with a sharply rising number of summer
deaths.  With full acclimatization  to the warmer
summers, heat-related mortality might increase less
dramatically or not at all.  In winter, the number of
weather-related   deaths   will  probably  decline
regardless of acclimatization.  It is not clear what
the net effect of these two offsetting trends may be.

     Only a few studies have evaluated the effects of
global  climate   change   on  human  mortality.
Kalkstein et  al. (1986)  developed a regression
equation involving nine weather elements, such  as
temperature, windspeed,  and humidity, to give the
best algorithm for describing the current impact  of
weather on mortality. The algorithm used mortality
data from New York City for 1964-66,1972-78, and

     The  analysis  revealed the  existence  of  a
summertime  "threshold   temperature"  ~  the
maximum  temperature   above  which  mortality
increases  — for  New York City of 92ฐF for total
deaths. This information was then used to assess
the potential impact of climate change under the
assumption   that  the  population would   not
 Estimations of the impact of warming on future mortality must
address the question of whether humans will acclimatize
(socially, psychologically, or physiologically adapt) to changing
weather. How quickly humans may become acclimatized  is a
topic of considerable controversy, so it is difficult to predict
whether the climate changes due to global warming will occur
slowly enough to permit acclimatization.

Chapter 12
acclimatize, as well as under the assumption that it
would  acclimatize.  Unacclimatized impacts were
estimated by combining the climate scenarios and
the historical weather algorithm described  above,
and  acclimatized impacts  were   estimated  by
analyzing analog cities that have values of weather
variables today that look like those New York is
estimated to have under climate change.

     Assuming full acclimatization  and a scenario
predicting that New York will be 3 to 4ฐC (5 to 7ฐF)
warmer than it is today, no additional deaths were
predicted. However, assuming no acclimatization,
the number  of summertime deaths attributable to
temperatures above the threshold (hereafter called
suprathreshold summer deaths) increased seven- to
tenfold.   Changes in winter weather, i.e.,  more
subthreshold temperatures, were not estimated to
affect mortality.
                             For  this  report,  Kalkstein  (Volume  G)
                        extended  the New  York  analysis to  cover  14
                        additional metropolitan  areas and to evaluate the
                        impact of two climate scenarios: the GISS doubled
                        CO2 scenario, and the GISS transient A scenario,
                        evaluated at  1994 to 2010 and  at  2024 to  2040.
                        Threshold temperatures were  calculated for each
                        city for summer and winter. Historical relationships
                        between mortality and temperature were derived
                        independently for each of these 15 cities for both
                        summer and winter.  Table 12-3 summarizes the
                        results for total mortality, by city and by season
                        (summer or winter),  for the doubled CCv, scenario
                        with and without  acclimatization.     The  cities
                        with   the   highest   estimated  number  of
                        suprathreshold summer  deaths  historically  were
                        New York City, Chicago, and Philadelphia; each
                        averaged over 100. All of the cities with the highest
                        average  number  of summer  deaths  are in the
     Table 12-3. Estimated Future Mortality Under Doubled CO2 Climate Conditions without
                and with Acclimatization
                                                 Number of deaths per season
Current    Without
Kansas City
Los Angeles
New Orleans
New York
Oklahoma City
St. Louis
San Francisco







Source: Kalkstein (Volume G).

                                                                                        Human Health
Midwest or Northeast,  and those with the  lowest
number are in the South.

     As would be expected, generally more  deaths
were predicted  for  populations  that  do  not
acclimatize.   However,  for  certain  cities,  e.g.,
Chicago, Kansas City, and Minneapolis, more deaths
were predicted with acclimatization than without.
Exactly why this occurred is uncertain. The  results
appear to  be very sensitive to  the choice  of the
analog city. For example, Chicago appears to have
more deaths if its population becomes acclimatized
than if it does not. It may be that the analog city
chosen to represent a particular acclimatized  city,
Chicago for instance, is more sensitive to weather
effects on mortality than Chicago currently is. More
research is planned  to  investigate this apparent
anomaly to refine the estimates  of what  global
warming will mean in terms of mortality.  Thus,
Kalkstein's results should not be used as predictions
of individual city behavior, but  as illustrations of

     In  the  absence   of any  acclimatization,
suprathreshold  summer mortality  in  the  United
Stated  under  conditions  of   doubled  CC^  is
estimated to rise from an estimated current total of
1,156 deaths to 7,402 deaths, with deaths  in the
elderly (aged 65 or over) subset  contributing about
60% of each figure (727 and 4,605, respectively).
Currently,  the  percentage of  elderly in the  U.S.
population  is  increasing.  Thus, the  mortality
estimated to result from climate  change may be
larger than that  found by Kalkstein because his
analysis is  predicated on today's age distribution.
Even with full  acclimatization, the  number of
weather-associated summer deaths almost doubles
to 2,198, possibly because  hot  weather increases
physiological stress.    Kalkstein's  analysis  also
estimates  a drop  in the  number  of  subthreshold
winter deaths. Historically, however, the number of
these deaths during the winter in the United States
is much smaller (243) than that observed for the
summer, and  subthreshold winter deaths were
estimated to fall to 52 without acclimatization and
to 159 with acclimatization. The net result for the
United  States is  an  increase in yearly mortality
associated with doubled CO2-

     This study is exploratory research in the field
of the  potential  impacts of  climate change on
human health.  Some aspects  of the analyses  that
led to these estimates need further investigation;
thus, the estimates should be accepted with caution.
The direction of predicted change, i.e., an increase,
is probably much more solid than the magnitude of
change. In addition, this research has  concentrated
on mortality occurring above a particular threshold
temperature for summer  or below  a  particular
threshold temperature for winter. Consideration of
a broader range of temperatures could conceivably
result in different conclusions being drawn.
Cardiovascular,   Cerebrovascular,
Respiratory Diseases
     Overall global warming  and climate change
may  exacerbate the  effects  of  cardiovascular,
cerebrovascular, and  respiratory diseases.   Data
from these studies show an  inverse relationship
between mortality and temperature (i.e., deaths go
down  as  temperature goes  up) for the  range
between  -5ฐC  and  about  +25ฐC, with  sharp
increases  at temperatures above  and below this
range, particularly  for the elderly and  for hot
weather  (White  and  Hertz-Picciotto, 1985); the
exact  range  appears to  depend  on  the  city.
Illustrations of this relationship for coronary heart
disease and stroke are shown in Figures 12-3 and
12-4, respectively (Rogot and Padgett, 1976). This
complex relationship precludes simple prediction of
the net effect of climate change. For example, it is
possible that hot weather-associated mortality from
these diseases may increase in some localities, but
this  trend may  be  offset, at  least  in part, by a
decrease in cold weather-associated mortality.

     Just  as  higher   summer  temperatures  are
associated  with increases  in  mortality   from
cardiovascular,  cerebrovascular, and  respiratory
diseases, they are also likely to be associated with
increases in morbidity from these diseases through
increases  in the number  or duration of hospital
admissions.  Particular stress  may be put on the
respiratory  system  because  climate change  can
potentially increase pollen, urban smog (discussed
below), and heat stress, all of which have an adverse
effect on the respiratory system.

     For example, if, as has been suggested in the
chapter on forests, climate change  encourages a
transition from forest  to grassland in some areas,
grass pollens  could increase.   This, hi turn, may
increase  cases of pollen-induced hay fever  and
allergic asthma.  (However,  the switch from forest
to grassland would  reduce the amount of  tree
pollens that also cause allergic responses in some

Chapter 12
            I I  I  I
                     I  I  I  I  III
100 F
38 C
              •18   -7   4   16

Figure 12-3.  Relationship of temperature to heart
disease mortality (adapted from Rogot and Padgett,
                                                             5. 20
                                                                     p LOS ANGELES
                                                                                        SAN FRANCISCO
                                                                  I I  I  I  I  I  I  I  I  I
-20  0   20   40   60   80  100 F
•29  -18   -7   4   16   27  38 C

                                    Figure 12-4.    Relationship  of  temperature  to
                                    mortality from  stroke (adapted  from Rogot and
                                    Padgett, 1976).
individuals.)  Rises in humidity also may affect the
incidence of mold-induced asthma and hay fever.

     As indicated in Chapter 11: Air Quality, global
warming may modify global and  regional  air
pollution because it may increase concentrations of
ozone and may also have impacts on acid deposition
and  general  oxidant  formation.   The increasing
occurrence of numerous respiratory diseases, such
as lung cancer, emphysema, bronchitis, and asthma,
has been attributed to the pollutants in urban smog
(Lopez  and Salvaggio, 1983).  Many of the trace
gases implicated  in global warming contribute to
these problems; other pollutants are created from
the interaction of ultraviolet light with these  and
other chemicals present in the atmosphere.

     The  component that  causes  the  greatest
concern in urban smog is ozone  (Grant,  1988). If
global warming causes an increase in tropospheric
                                     ozone, adverse consequences could result for adult
                                     asthmatics and people who suffer from acute or
                                     chronic bronchitis.

                                     Vector-Borne Diseases

                                         Potential changes in humidity and temperature
                                     could alter the geographic ranges and life cycles of
                                     plants, animals, insects, bacteria, and viruses. (For
                                     further discussion of forestry and agriculture, see
                                     Chapters 5 and 6, respectively.)  For example, the
                                     range of many plant pests may move northward by
                                     several hundred miles. Such changes could occur
                                     for insects that spread diseases to both humans and
                                     animals.  Vector-borne diseases that affect humans
                                     are  relatively  rare  in the United States.  The
                                     incidence of  most  of those found, however,  is
                                     increasing.  The incidence of some, such as Lyme
                                     disease, is increasing dramatically (CDC, 1986).

                                                                                       Human Health
Tick-Borne Diseases

     Both Rocky Mountain spotted fever and Lyme
disease are considered to be public health problems
in the United States. Although these two diseases
are  spread  by different  species  of ticks,  some
overlap   exists  in  their  geographic distribution
(Figure  12-5).  Because tick populations appear to
be limited by the size of their intermediate host
populations (such as white-tailed deer), the spread
of tick-borne diseases may be particularly sensitive
to any change that may affect the geographic range
of these hosts and, consequently,  the range of the
vector, or carrier.

     In  addition to the presence  of the host, tick
populations  also  depend upon the seasonally of
environmental  factors  such  as  temperature,
humidity, and vegetation.  Optimally, climate must
be warm enough to promote progression through
the life cycles, humid enough to prevent the drying
out of eggs, and cold enough in winter to initiate the
resting stage.

     As   for   many  tick-borne   diseases,  the
opportunity for a tick to acquire the infective agent
from an infected  animal is limited to the short
period when the level of the agent in the blood of
the host is high enough for the tick to  receive an
infective dose.  Higher temperatures may increase
the  amount of the agent (the  organism that is
transmitted by the carrier, such as a virus) and the
time it remains lodged on the host animal.  Both
these mechanisms would increase the  rate  of
infection of the carrier.  However, although higher
temperatures may favor the presence of the  agent,
there is some indication that they  could disrupt the
life  cycle of some tick species.   In these  cases,
warmer  temperatures  would  reduce  both tick
survival and the spread of diseases they carry.

     Tick populations  also vary  with the natural
vegetation of  an area.   The incidence of Rocky
Mountain spotted  fever,  in particular, has been
linked to natural vegetation and changes in climate.

     In examining the  potential impact  of climate
change  in the United  States on Rocky Mountain
spotted fever,  Haile (Volume G) used a weather-
based model, ATSIM, to evaluate  the impact of the
scenario climate changes on the distribution of the
American dog tick, the  primary carrier of  this
disease  (Haile, Volume G; Mount  and Haile, 1988).
The model uses data inputs from the three doubled
CO2 scenarios (GISS, GFDL, and OSU) to estimate
population dynamics, growth rate, and generation
time. Haile assumed that habitats and host density
did not change in response to  global wanning.
Sample results for six cities representing the most
southern,  the most northern, and the two middle
latitudes are presented in Figure 12-6. The results
indicate that  under all scenarios, tick populations
would shift from  south to  north and  would be
virtually   eliminated  from   the  most southern
locations (Jacksonville and San Antonio). However,
in the middle latitude cities, the results  are mixed
and depend on the scenario evaluated.  The model
does  not  estimate changes  in  incidence of  the

     In this analysis,  the  only  model inputs that
were changed to simulate climate change were the
weather inputs.  Other important parameters in the
model are the distribution of habitat between forests
and  meadows and the presence of suitable hosts.
Both parameters are likely to be changed relative to
current conditions under  climate change.   As
indicated  in  Chapter 5: Forests, a  change from
forests to  meadows may occur in certain areas of
the country; this would depress the tick population.
However,  the distribution of small mammals also
may  change.   If  small  mammal  populations
increased, tick populations would grow. In addition,
this  study did  not consider changes  in climate
variability, which may have a major  effect on the
outbreak of diseases.

     In a sensitivity analysis of their model, Mount
and Haile (1988) found that the model predictions
could vary sixteenfold, depending on the inputs used
for host density, whereas the variability conferred by
changes in the weather inputs is about fourfold.
Based on  the sensitivity analysis, host densities are
extremely important to these  predictions. Keeping
them constant,  as was done in this analysis, could
have underestimated or overestimated the impact of
climate change on the density of the American dog

Mosquito-Borne Diseases

     A second  category of  vector-borne diseases
that  can be affected by climate change consists of
diseases carried by mosquitoes.   Climate changes
resulting in more days between 16 and 35ฐC (61 to
95ฐF), with humidity between 25 and 60%, are likely
to favor   the growth of mosquitoes (White and
Hertz-Picciotto, 1985).  Mosquito populations are

Chapter 12
                                     LYME DISEASE
                          ROCKY MOUNTAIN SPOTTED FEVER
         States With Highest
         Incidence (Cases Per 100K)
         NC 3.6, SC 3.2, OK 2.3
Figure 12-5.  Geographic distribution of Lyme disease and Rocky Mountain spotted fever (Longstreth and
Wiseman, Volume G).

                                                                                         Human Health
              Richmond, VA
              Columbus, OH
            Jacksonville,  FL
            San Antonio, TX
               Hal fax, N.S.
              Missoula, MT
                                                 \\        I        I        I
                                                 15      20     25      30      35
                                      DENSITY (ADULT TICKS ON HOSTS/HECTARE)
Figure 12-6. Simulated tick densities for selected cities  under various scenarios  of  climate  change (Haile,
Volume G).
also sensitive to the presence of standing water. It
is not  clear whether standing water will generally
increase  or decrease  (see  Chapter  9:  Water

     Worldwide,  mosquito-borne  diseases   are
associated with significant illness and mortality.  In
the United States, however, vector control programs
and improved  hygiene have virtually  eliminated
endogenously transmitted cases of these  diseases,
with  the  exception  of  sporadic  outbreaks  of
arbovirus-encephalitis.  (Imported cases are  seen
occasionally.)   Numerous  mosquito species  are
present hi the  United States,  however.   Recent
restrictions on pesticide use, coupled with the influx
of visitors and immigrants who can serve as sources
of infectious agents, as well as the lack of available
vaccines for many of the potential  diseases, suggest
the potential for reintroduction and establishment of
these diseases in the United States — particularly if
global warming provides a more suitable climate for
their  growth and development (Longstreth and
Wiseman, Volume G).
     At a recent workshop, five of the numerous
mosquito-borne diseases were considered to pose a
potential risk to U.S. populations if the status quo
is disturbed by climate change (Longstreth and
Wiseman, Volume G).  Malaria, dengue fever, and
arbovirus-induced encephalitides were considered to
be significant risks, and yellow fever and Rift Valley
fever were considered to be possible risks.


     Malaria is an infectious disease transmitted by
mosquitoes and induced by parasites (Plasrnodia).
The symptoms are highly variable, depending on the
species of  the agent.  They include chills, sweats,
and headache, and in severe cases, may progress to
liver damage and even liver and renal failure.

     As a result of  effective vector control and
treatment programs, malaria is no longer indigenous
to the United  States.  However, imported cases
occur  regularly,   and  occasionally  indigenous
transmission has been  documented  (Longstreth

Chapter 12
and   Wiseman,  Volume  G).     Current  U.S.
demographic trends, including a  large number  of
legal  and illegal immigrants from locations where
malaria is endemic, could present a pool of infected
individuals that, in conjunction with climate changes,
may  create  sufficient  conditions  for  increased
disease incidence.

     Haile  used  the  weather-dependent   model
MALSIM  to  evaluate  the  potential impact  of
climate change on malaria in an infected population
living in  an area  where a competent  carrier  is
present.  The  model was originally developed  to
help predict malaria outbreaks in tropical countries
such as Kenya.  This is  the first application of the
model to the United States.  This analysis  did not
consider changes in climate variability, which may
be  important  for  the  spread  of malaria.  The
MALSIM  model  showed that several  cities in the
South (e.g., Miami, Key West, and Orlando), under
current climate conditions, are very favorable for
malaria transmission.   Using the climate change
scenarios in  MALSIM  did  little  to  affect the
estimated transmission potential of malaria in the
United States (Figure 12-7).  In a few cities, e.g.,
Richmond,  Nashville,  and  Atlanta, the  model
estimated large increases in one scenario relative to
those that would occur normally.  However, the
results varied with different climate scenarios, did
not occur at all locations, and should be considered
to be inconclusive.
"jfhe MALSIM estimates of malaria incidence by city under
current conditions were based on two assumptions: that there
were 100,000 female mosquitoes in the vicinity of each city and
that 100 infected people were added to the cities' populations.
Under those assumptions,  infection of virtually the  entire
population  of Miami was predicted to be possible  unless
protective measures were taken.

Miami, FL
Key West, FL
Orlando, FL
San Antonio, TX
Atlanta, CA
Nashville, TN
Tulsa, OK
Dallas, TX
Richmond, VA
Baltimore, MD
Indianapolis, IN
Boston, MA

1 l^g^BJaJ!'.. .. s
^ ioswt^'

1 1 1 1 1 1 1 1 1 1
1000 2000 3000 4000 SOOO 6000 7000 8000 9000 10000
Figure 12-7. Simulated incidence of malaria for selected cities under various scenarios of climate change (Haile,
Volume G).

                                                                                         Human Health
Dengue Fever

     Dengue fever is an arbovirus-induced  illness
characterized by fever, rash, and severe pain in the
joints. The dengue virus has four different types
(DEN 1  through DEN 4). Sequential infection by
different types is possible and has been suggested
to lead to an increased risk  of  developing a more
severe, hemorrhagic form of the disease that can be
fatal  in  the very  young and the elderly.   Like
malaria,  it is not currently endemic in the United
States, although potential carriers are present and
the disease is  imported here regularly by people
who have traveled abroad.

     The ability of the vector to transmit the agent
appears  to depend on temperature,  and current
conditions do not appear to be favorable for this
process.   Climate changes that raise temperatures,
however, may reduce the required incubation period
and increase the infectivity of the carrier, increasing
the potential transmission of the disease.

Arbovirus-Related Encephalitides

     Arbovirus-related encephalitides are  a group
of acute  inflammatory diseases that involve parts of
the brain, spinal cord, and meninges. In mild cases,
these  infections result in feverish headaches  or
aseptic meningitis; in more severe  cases, those
symptoms can  be accompanied by stupor, coma,
convulsions  (in infants), and occasionally spastic
paralysis (APHA, 1985).

     At   least  seven  types of  viruses  causing
encephalitis are present in the United States. These
include the three forms that also infect horses (the
western,   eastern,   and   Venezuelan   equine
encephalitis viruses) as well as four that are named
after the location of their discovery (the La Cross,
St. Louis, Powassan, and  California  encephalitis
viruses).  Cases range in severity depending on  the
type of virus, with yearly fatality rates between 0.3
and 60%. These infections are rare.  In 1984,  129
cases  were  reported  to  the Centers for Disease
Control,  which maintains  an  active  surveillance
program for them (CDC, 1986).
  An arbovirus is  a  virus transmitted by an  arthropod.
 Arthropods are a group of animals that includes insects and
 arachnids.  Examples of arthropods that transmit disease
 include mosquitoes and ticks.
     Outbreaks of encephalitis attributable to these
viruses are normally limited to specific geographic
locations and seasons for several reasons.   First,
warm temperatures are normally required for the
viruses to multiply and to be transmitted to a new
host. Higher  temperatures   may   quicken  the
transmission process and promote epidemic disease.
However, the extent of this effect depends largely
on the particular virus. Some viruses require cooler
weather  and  higher  moisture conditions.   Thus,
higher temperatures may reduce their prevalence.
Second,  environmental conditions  that favor the
presence of carriers and  hosts must prevail.  For
example,  relative  humidity may affect plant life
necessary for the feeding  of hosts.

Other Diseases

     The Incidence  of  a variety  of  other  U.S.
diseases  appears  to  be  sensitive  to  changes in
weather.   If humidity is higher,   an increased
incidence and severity of fungal skin diseases (such
as ringworm and athlete's foot) and yeast infections
(such as candidiasis) may be observed. Studies on
soldiers  stationed in  Vietnam  during the  war
indicated that outpatient visits for skin diseases (the
largest  single cause  of  outpatient  visits)  were
directly  correlated to  increases in  humidity but
showed  a  4-month  lag  with  relationship  to
temperature increases (Figure 12-8).  In addition,
excessively high temperatures  can lead to such skin
diseases as prickly heat and heat rash, which impair
the  ability of the skin to breathe and thus place
additional stress on people already  suffering from
overexposure to heat from other causes.

     Several diseases appear to be associated with
the  acquisition of winter infections.  If a reduction
in winter severity is also accompanied by a decrease
hi wintertime Infections, these diseases could be
reduced under global warming.

     For example, birth in cold winter months has
been associated with a higher risk of schizophrenia
in individuals whose  schizophrenia  is without an
apparent  genetic  component  (Kovelman  and
Scheibel, 1983). In addition, juvenile-onset diabetes,
which has been reported to be increasing over the
past several  decades,  has  been  shown   to  be
associated with a seasonal variation  in that the
month  of  first  admission peaks  in the   whiter
(Glatthaar et al., 1988; Patterson et al., 1988).  It is
a common clinical experience that  a minor viral
illness precedes the onset of symptoms.

 Chapter 12

1969 ,
                                                                   TEMPERATURE (ฐC) —'

                                                                     RELATIVE HUMIDITY
 Figure 12-8. Relationship of skin infections to humidity and temperature (Longstreth and Wiseman, Volume G),

     Demographic  and technological trends (the
 aging of the population, an influx of immigrants,
 advances in treatment techniques) make it difficult
 to analyze the potential impacts of climate change
 on human health. Although this  chapter attempts
 to identify those human health effects at risk from
 climate change, the analyses were not designed to
 consider adaptive responses  and should not be
 treated as  predictions of what will happen with
 climate change but as illustrations of sensitivities.
 Rather, the analyses presented  here  represent
 possible scenarios, in the absence of consideration
 of demographic trends or adaptive responses, that
 may either exacerbate or ameliorate  the impact of
 climate change on human health.  Societies possess
 considerable ability to adapt to change.   The
potential for climate to affect human health may be
considerably modified by adaptive responses, such
as immunizations, modification of the environmental
temperature  (e.g., use  of air conditioners), and
control of disease carriers.
                   Climate change may  affect  regional  and
              national health  care.  For instance, the treatment
              requirements for asthma may increase or decrease
              as locations  experience changes in the distribution
              and intensity of pollen concentrations. Increased
              resources maybe needed to treat premature infants
              if the number of preterm births increases.  If heart
              attacks, stroke, and respiratory problems increase,
              hospitalization costs and costs due to days lost from
              work may also increase. Higher health care costs
              might be  particularly obvious in  Medicaid  and
              Medicare  because those  below the poverty  line
              would be less able to take adaptive measures (e.g.,
              air-conditioning),   and  the   elderly  are  more
              susceptible to the ill effects of extreme  weather

                  The United States is already experiencing an
              infant mortality higher  than  that   of any  other
              industrialized nation (World Bank,  1987).  Some
              studies have found that perinatal mortality is higher
              hi  the  summer;  consequently,  the  increased
              temperatures expected with  global wanning may
              well exacerbate infant mortality (or at least neonatal

                                                                                      Human Health
     The need for irrigation may increase in many
regions of the  United  States (see  Chapter  6:
Agriculture).    Irrigation may result  in  greater
amounts of  standing  water  and can therefore
increase  mosquito   populations.     Arbovirus
encephalitis may become  a greater problem than at
present, and  other mosquito-borne diseases, such
as dengue  or yellow fever,  could  be  more easily
spread if introduced.

     One health impact of climate  change not
assessed in this report is the morbidity and mortality
associated with certain kinds of extreme events, e.g.,
tornadoes and hurricanes.  These currently cause
some mortality in the United States; however, it is
difficult to say whether there will be a change in the
mortality  induced by  these  events  with  global
warming.  As indicated in Chapter 3: Variability,
changes in the  frequency of such extreme events
cannot be predicted on the basis of an analysis  of
the  general  circulation  model  (GCM)  output,
although an increase in severity of some kinds  of
storms, e.g.,  hurricanes,  is  not inconsistent  with
current theories and more detailed models of storm

     The impact of global change on human health
will most likely be greater in the  lesser-developed
countries (LDCs) that do not have the resources to
take the adaptive or preventative measures available
to the United States.   Impacts on agriculture and
water resources in the LDCs could result in poor
nutrition  and water  shortages that  may make
populations more susceptible to disease.  Changes
in insect (arthropod) habitats may allow diseases to
flourish where they never have before.  Changes in
extreme events such as monsoons or  floods could
significantly affect mortality in the developing world.
Such external impacts on health might have  an
impact on the  United  States not  only  via the
potential  for introduction   of  diseases  already
discussed  but   also   via   our   participation  in
international aid and relief programs.

     The full impacts of climate change on human
health will require more research. Agencies such as
the  Department  of Health and Human  Services
should  consider  conducting studies on potential
     In the future, a cadre of trained professionals
may be needed to respond to outbreaks of diseases.
A shift in the distribution of carriers of human
disease  may  necessitate   regional  shifts   in
surveillance and eradication programs.  States that
do not have these programs may need to develop

     Although  information   evaluating   the
relationship of weather and season to various health
effects is plentiful, research into the significance of
these relationships in the context of global warming
is scarce.  A number of areas  requiring further
research are described below.

     A  number  of  studies   have   identified
relationships  between temperature  changes and
mortality from diseases of the heart,  respiratory
system, and cerebrovascular system. These studies
show slightly different relationships depending on
the  city that  provided the  data,  although  some
common elements exist. A statistical analysis of this
information might be warranted to determine if one
general relationship (across the United States,  or
perhaps related to latitude) could be developed for
each of these categories.  Such a relationship could
then be used to estimate the impact of global
warming by specific disease category.

     A companion study  to that proposed  above
should identify the  top  10 causes  of  deaths
associated with changes in weather in the Kalkstein
study.  The results could then be compared with the
information  derived  above  to  determine  other
causes of mortality that show great sensitivity to the

     The Kalkstein analysis did not look at deaths
occurring in the very young (aged  1 year and
below). Given the seasonality of perinatal mortality
and preterm  death observed in several studies,  an
investigation   of  the   relationship   between
temperature  and  mortality  in   the  very  young
probably would be  worthwhile.   More  baseline
information is needed for  the latter study.  Related
studies on perinatal mortality could examine the

Chapter 12
     •   Whether the South has a higher per capita
         incidence of perinatal mortality.

     •   Whether  infections,  which  have  been
         suggested  as  a potential  cause of the
         perinatal  mortality observed,  show  a
         seasonality  in   parallel   to  perinatal
         mortality,   and  whether   more  such
         infections occur in the South.

     •   The  principal  causes   (e.g.,  bacteria,
         viruses)  for  perinatal  infections,  and
         whether they  are the same as those for
         skin infections that increase with increases
         in humidity.

     •   Whether the incidence of preterm birth or
         perinatal mortality is related to  weather
         parameters such as temperature, humidity,
         or high-pressure systems.

     The following additional research areas are

     •   Synergism  between stratospheric ozone
         depletion  (due  to increases  in UV-B
         radiation) and global warming. Increased
         UV-B radiation and global warming might
         be  expected  to   exacerbate  infectious
         diseases.  UV-B  radiation  may  have an
         impact on the ability of an individual to
         respond to a disease, and global warming
         may  change  the  incidence  of  certain
         infectious   diseases.     For  example,
         leishmaniasis is an important  disease in
         many African countries.    In   animal
         models, UV-B irradiation adversely affects
         the   development   of  immunity   to
         Leishmania.   If climate change creates
         more favorable habitats for the  sand-fly
         vector of this disease, then a double insult
         to the  system could  occur:  a higher
         incidence, and a worse prognosis.

     •    The impacts on LDCs.  The Agency for
         International Development  is supporting
         the development  of a Famine  Early
         Warning System (FEWS) that will use a
         variety of inputs (many of them  weather
         related) to help predict  when conditions
         leading  to famine may be  occurring.
         Appropriate GCM outputs could be input
         into  this system  to evaluate how  the
 changes associated with global warming
 may   affect   famine   development.
 Similarly, the Department of Defense is
 using a number of models comparable to
 those used by Haile to attempt to predict
 where  infectious diseases are likely to
 pose problems for U.S. troops. It might
 be interesting to evaluate how the climate
 variables  from   the  GCM-generated
 scenarios would affect these  predictions,
 particularly in the LDCs where these
 diseases present a very real  problem to
 the health care systems.

 Introduction of infectious diseases into the
 United States via immigrants. Anecdotal
 information   indicates   that   many
 immigrants are not served by the health
 care  system;  consequently,  they could
 become a  population where diseases
 might  develop into full-blown epidemics
 before   initiation   of   treatment.
 Determining  whether  or   not  global
 warming will  affect this process, either
 directly  via  the  provision  of a more
 hospitable environment for the disease or
 indirectly via an increased  number of
 immigrants  and refugees, will require a
 better   characterization of the  current

 Intermediate hosts and their habitats. In
 the models used by Haile, two important
 input parameters that were held constant
 were the size of the intermediate host
 population and the distribution of habitat
 between forest and meadow.  It is likely
 that both  of these  parameters   will
 themselves be affected by climate change.
A better grasp of how climate  change will
 affect  these parameters  needs  to be
 developed   and  integrated   into   the
infectious disease models.

Irrigation and incidence of vector-borne
disease.   An increase hi irrigation is
possible, which could have a significant
impact  on mosquito  development and
therefore on  mosquito-borne diseases.
The importance of such water is time-
dependent, however (i.e., it must occur at
the right moment  hi the insect's  life-
cycle).    Thus an  analysis of how  the
growing season overlaps transmission of

                                                                                      Human Health
         diseases such as La Cross encephalitis
         might provide an indication of whether
         changes in irrigation practices should be a
         concern in terms of public health.

         Mortality from extreme events.  Another
         issue that might warrant  investigation is
         how  climate change  may affect  the
         mortality associated with extreme events,
         such as hurricanes and floods.

         Air pollution and respiratory disease. Air
         pollution is already a major contributing
         factor in the incidence and severity of
         respiratory disease in the United States.
         An analysis of  the  extent  that global
         warming  will exacerbate  air pollution is
         critical to an assessment of the potential
         health effects of climate change.

APHA. 1985. American Public Health Association.
In: Berenson, A.S., ed. Control of Communicable
Diseases in Man. Springfield, VA: John D. Lucas
Printing Company.

CDC. 1986. Centers for Disease Control.  Annual
Summary 1984. MMWR 33:54.

Cooperstock, M., and RA. Wolfe. 1986. Seasonally
of preterm birth in the collaborative perinatal
project: demographic factors. American Journal of
Epidemiology 124:234-41.

Dukes-Dobos, F., 1981. Hazards of heat exposure.
Scandinavian Journal of Work and Environmental
Health 73-83.

Gill, J.S., P. Davies, S.K. Gill, and D.G. Beevers.
1988.  Wind-chill and the  seasonal variation of
cerebrovascular disease. Journal of Clinical Epidem-
iology 41:225-230.

Glatthaar,  C., P. Whittall, T. Welborn, M. Giboon,
B. Brooks, M.M.  Ryan,  and  G. Byrne.   1988.
Diabetes   in   Western   Australian   children:
descriptive epidemiology.   Medical Journal of
Australia 148:117-123.

Glenzen, W.P. 1982. Serious illness and mortality
associated  with   influenza   epidemics.
Epidemiological Reviews 4:25-44.
Grant, L.D. 1988. Health effects issues associated
with regional and global air pollution problems.
Prepared for World Conference on the Changing
Atmosphere, Toronto.  Draft document.

Harris, R.E., F.E. Harrell, K.D. Patil, and R. Al-
Rashid. 1987. The seasonal risk of pediatric/juvenile
acute lymphocytic leukemia in  the Unites States.
Journal of Chronic Diseases 40:915-923.

Kalkstein, L.S., R.E. Davis, JA. Skindlov, and K.M.
Valimont.  1986. The impact  of human-induced
climate warming upon human mortality:  a New
York case study.  Proceedings of the International
Conference on Health and Environmental Effects of
Ozone   Modification   and   Climate   Change,
Washington, DC; June.

Kalkstein, L.S.,  and K.M. Valimont. 1987.  Effect
on human health.  In: Tirpak, D., ed. Potential
Effects of Future Climate Changes on Forest and
Vegetation, Agriculture, Water Resources,  and
Human Health, Vol. V, pp. 122-152.  Washington,
DC: U.S. Environmental Protection Agency.  EPA

Keller, CA., and R.P. Nugent.   1983. Seasonal
patterns in perinatal mortality and preterm delivery.
American Journal of Epidemiology 118:689-98.

Kovelman, J., and A. Scheibel.  1986. Biological
substrates  of  schizophrenia.    Acta  Neurologica
Scandinavica 73:1-32.

Kutschenreuter, P.H. 1959. A study of the effect of
weather  on mortality.  New  York Academy of
Sciences 22:126-138.

Lopez, M., and  J.E. Salvaggio. 1983.  Climate-
weather-air pollution. In: Middleton, E., and  C.E.
Reed, eds. Allergy, Chapter 54. St.  Louis, MO:
C.V. Mosby Company.

Mount, GA., and D.G. Haile.  1988. Computer
simulation of population dynamics of the American
dog tick, Dermacentor variabilis (Acari: ixodidae).
Journal of Medical Entomology. In press.

Patterson, C., P. Smith, J. Webb, M. Heasman, and
J.  Mann.  1988. Geographical  variation in  the
incidence of diabetes mellitus  in Scottish children
during the period 1977-1983.   Diabetic Medicine

Chapter 12

Rogot, E., and S J. Padgett.  1976.  Associations of
coronary and stroke mortality with temperature and
snowfall in selected areas of the United States 1962-
1966. American Journal of Epidemiology 103:565-

Sontaniemi, E., U. Vuopala, E.  Huhta, and  J.
Takkunem. 1970. Effect of temperature on hospital
admissions for myocardial infarction in a subarctic
area. British Medical Journal 4:150-1.

Vuori, I. 1987.  The heart and the cold.  Annals of
Clinical Research 19:156-162.
White, M.R., and I. Hertz-Picciotto.  1985. Human
health: analysis  of climate related  to health.  In:
White, M.R., ed. Characterization of Information
Requirements  for Studies of CO2 Effects: Water
Resources,  Agriculture,  Fisheries,  Forests,  and
Human Health.  Washington, DC: Department of
Energy. DOE/ER/0236.

World Bank.  1987.  World Development Report
1987. New York: Oxford University Press.

                                      CHAPTER 13
                         URBAN INFRASTRUCTURE

Global climate change could require U.S. cities to
make major changes in capital investments  and
operating budgets. Areas most likely to be affected
include water supplies, roads, and bridges; storm
sewers and flood control levees; and energy demand
in municipal buildings and schools.

•    Most urban infrastructure in the United States
     will turn over in the next 35 to 50 years. If
     potential  changes in climate are considered,
     this turnover will allow cities to prepare for
     climate change at lower costs. In some cases,
     the  risk  of  climate  change  should  be
     incorporated into decisions beginning today,
     such as coastal drainage systems that are likely
     to last for 50 to 100 years.

Northern and Southern Cities

•    Northern cities, such as Cleveland, may incur
     a change in the mix of their expenditures. In
     such locations, increased electricity costs for
     air-conditioning could be offset by reductions
     in expenditures for heating fuel, snow and ice
     control, and  road  maintenance.  Southern
     cities could see increases in operating budgets
     due  to  the   demand  for   additional  air-

Coastal Cities

•    Coastal cities, including 12 of the 20 largest
     metropolitan areas, may face somewhat larger
     impacts, such as the following:

     — Sea level rise or more frequent droughts
        would increase  the  salinity  of  shallow
        coastal aquifers  and tidal surface waters.
        Cities that rely on water from these sources
        should examine water supply options. Such
        areas  as Dade County, Florida, or  New
        York City would probably be vulnerable.
    ~  As  sea level rises, some coastal cities
        would require levees to hold back the sea
        or fill to raise the land surface area.  In the
        case of Miami, the cost of these activities
        might exceed $500 million over the next
        50 to 75 years, necessitating an average
        increase  of  1  to  2% in annual capital
        spending in Greater Miami.

Water Supply and Demand

•   Climate change will influence the supply and
    demand for water in many cities. A lengthened
    summer season and higher temperatures would
    increase the use  of water for air conditioners,
    lawns, and gardens.   Changes in rainfall
    patterns, runoff, and flood control measures
    may alter water supplies. In the Hudson River
    Basin, summer water demand could increase by
    5% over the demand for water without climate
    change, while supplies might fall.  Such a
    change  would require new institutional and
    management approaches for both the Delaware
    and Hudson Rivers.

Policy Implications

•   Climate  change has  implications for many
    national  programs  and policies, including the

    —  The National Flood Insurance Program may
        react to  climate  change  by redrawing
        floodplain maps and  adjusting insurance
        rates to account for  sea level rise  and
        changes hi riverflows.  This program might
        consider  discouraging development that
        would be vulnerable to sea level rise.

    -  Because of the key role federal programs
        play  in  the  development  of  cities, the
        Department  of   Housing  and  Urban
        Development   should   examine  the
        implications of climate change on long-term
        policies. A minimum response might be to

 Chapter 13
        provide guidance  on the  certainties and
        uncertainties of climate change to groups
        such as the National League of Cities, the
        U.S.  Conference  of  Mayors, and  the
        American Planning Association.

        Because water supply infrastructure may last
        for several centuries, improved planning is
        important.  The TJ.S. Geological Survey
        should study the probable impacts of global
        climate change and sea level rise on the
        water supplies of major cities.  The  U.S.
        Army  Corps of Engineers should factor
        climate change into  the design of major

        Given the assumption that modest changes
        in the design  and  location  of  many
        transportation systems may facilitate an
        accommodation  to climate  change,  the
        Department of Transportation should factor
        climate change into  the  design of  roads,
        bridges, and mass transit facilities.

        Voluntary standards organizations, such as
        the American Society of Civil Engineers, the
        Building Officials and Code Administrators
        International, and the American Society of
        Heating   and  Refrigerating  and   Air
        Conditioning Engineers should examine the
        need for changes  in existing energy  and
        safety factors to account for the possibility
        of climate change.

     Three-quarters  of  the U.S.  population is
concentrated in urban areas (Statistical Abstract,
1988). The majority of the nation's investment in
water supply, wastewater transport and treatment
facilities, drainage, roadways, airports, mass transit,
electric power, solid waste disposal sites, and public
buildings  serves these urban  areas.  The  current
value  of  selected  infrastructure   nationwide,
displayed  in Table 13-1, provides insight into the
aggregate investment at  stake if  climate changes.
Most of these items could be considered  part of
urban  infrastructure; their locations  and  designs
have   been  based   on  historic   meteorologic
information. Annually, governments add an average
 of $45 billion to the capital stock (National Council
 on Public Works Improvement, 1988).

     Of the 20 most populated U.S. urban areas, 18
 have access to oceans, major lakes, or rivers and
 have invested in infrastructure for port facilities and
 flood  control.     The  expenditure  required to
 construct coastal defense structures — which prevent
 inundation  by the sea, slow oceanfront erosion,
 control storm  surges, slow saltwater  advance up
 rivers, and reduce saltwater intrusion into aquifers
 — is now minimal.
 Table  13-1. Value   of  the  Nation's  Stock  of
             Selected Infrastructure (billions of
             1984 dollars)
Water supply
Urban drainage
Public airports
Mass transit
Electric power
(private only)b
470 .
     Public buildings
 Based on a useful life of 35 to 50 years for most
 assets, and 10 to 20 years for transit vehicles.
bAbout 77% of electric power is privately produced.
Source: Statistical Abstract (1988); National Council
on Public Works Improvement (1988).
 Of the 20 most populated urban areas in the United States, 12
are tidal waterfront cities (Baltimore, Boston, Houston, Los
Angeles, Miami,  New York,  Philadelphia/Wilmington, San
Francisco/Oakland, San Diego, Seattle, Tampa/St. Petersburg,
and Washington, DC), 3 are located  on the Great Lakes
(Chicago, Cleveland, and Detroit), 3 are on navigable  rivers
(Minneapolis, Pittsburgh, and St. Louis), and 2 are not on a
navigable waterway (Atlanta and Dallas).

                                                                               Urban Infrastructure
    Although actual practice varies, the nominal
replacement cycle for most infrastructure is 35 to 50
years   (National  Council  on   Public  Works
Improvement,  1988).     Some  water  supply
investments have 100-year cycles between planned
replacement;  however, sea level rise, temperature
change, and changes in precipitation patterns could
alter the balance between water supply and demand.
The nature and pattern of precipitation could affect
drainage requirements as well as highway design
and maintenance.

    The heat wave of 1988 illustrated some of the
potential  impacts.    Hundred-degree  weather
distorted railroad tracks, forcing Amtrak to  cut
speeds from 200 to 128 kilometers per hour between
Washington and Philadelphia (Bruske,  1988) and
possibly contributing to a train wreck that injured
160 people  on a  Chicago-Seattle  run  (The
Washington Post, 1988).  A U.S. Army Corps of
Engineers contractor worked around the clock for
2 weeks to build a 170-meter-wide, 9-meter-high silt
wall across the bottom 40% of the Mississippi River
channel,  48   kilometers  below New Orleans
(Sossaman, 1988a,b). This $2 million wall, designed
to wash away when spring snowmelt demands the
full capacity of the  channel, slowed an advancing
wedge of saltwater that threatened the water supply
in New Orleans and nearby parishes. In Manhattan,
heat exacerbated the effects of longstanding leaks in
256 kilometers of steam pipes, causing the asphalt
to  soften.  As  vehicles kneaded the soft asphalt,
thousands  of  bumps  formed  on city  streets,
requiring extensive repairs (Hirsch, 1988).  In the
suburbs of Washington, DC, steel expansion joints
bubbled along a 21-kilometer stretch of Interstate
66 (Lewis, 1988).

    The following  sections of  this chapter  will
examine such issues  as  the  portions  of  the
infrastructure that will be  significantly affected, and
anticipated costs and who will bear them.

     Available literature on the potential effects of
 global climate change on urban infrastructure is
 sparse. Rhoads et al. (1987) examined the potential
 impacts of sea level rise on water supply and flood
protection in Dade and Broward Counties, Florida,
and concluded that the effects might be substantial.
Linder et  al.  (1987) estimated that CO2 doubling
might require raising electric capacity by 21% in a
southeastern utility and by 10 to 19% in New York
State. Hull and Titus (1986) analyzed the potential
impact of sea level rise on water supply  in the
Philadelphia-Wilmington-Trenton area and found
that a rise of 0.3 meters could require adding 140
million cubic meters of reservoir capacity, about a
12% increase, to prevent saltwater from advancing
past  water  intakes on  the  Delaware  River.
Additional investment would be required to prevent
or respond to saltwater infiltration into underground
aquifers.   Cohen  (1987)  estimated  that  large
municipalities along the Great Lakes might increase
water withdrawals by 5.2 to 5.6% during May to
September because of increased lawn watering.

    Two recent studies illustrate the importance of
considering  sea  level  rise  in  urban coastal
infrastructure planning and the uncertain nature of
the decisions  involved. Wilcoxen (1986) examined
the impact of sea level  rise on a portion of San
Francisco's sewage transport system buried near the
shoreline.  The study estimated that if sea level rose
0.6 meters by  the year 2100,  an expenditure of
roughly $70 million on beach nourishment might be
required to prevent damage to a structure that cost
$100 million to build in the late 1970s. The author
suggested  that consideration (at no additional cost)
of sea level rise in siting the structure could have
prevented these  expenses.   Titus  et  al.  (1987)
examined  the impact of sea level rise on a proposed
coastal  drainage system  in  Charleston,   South
Carolina,  and estimated  that a 0.3-meter sea level
rise by 2025 would  require almost $2.5 million in
additional investments to maintain the target level
of flood protection.  The present value of these
investments is $730,000.  In contrast, only about
$260,000,  one-third  of the  cost of responding in
2025, would  be  required  to  add  this level of
protection at  initial  construction.   Thus, the
investment would be worthwhile if the probability of
sea level rising this rapidly exceeds 35%.

     Several  studies undertaken  for  this report
examined some of  the  implications  of climate
change in relationship to urban infrastructure. One

 Chapter 13
 study  comprehensively examined the impacts  on
 infrastructure in several cities:

     •  Impact of Global Climate Change on Urban
        Infrastructure - Walker, Miller, Kingsley,
        and Hyman, The  Urban Institute (Volume

 The following studies, referenced in this chapter,
 covered issues relating to urban infrastructure:

     •  The Potential Impacts  of Climate Change
        on Electric Utilities: Regional and National
        Estimates - Linder  and Inglis,  ICF Inc.
        (Volume H)

     •  Impacts  of Extremes  in Lake Michigan
        Levels Along  the Illinois Shoreline: Low
        Levels -  Changnon, Leffler, and  Shealy,
        University of Illinois  (Volume H)

     •  Methods  for  Evaluating  the  Potential
        Impacts of Global Climate Change: Case
        Studies of the Water Supply Systems of the
        State of California and Atlanta. Georgia^-
        Sheer and  Randall,  Water Resources
        Management Inc.  (Volume A)

     •   National Assessment of Beach Nourishment
        Requirements  Associated with Sea Level
        Rise - Leatherman, University of Maryland
        (Volume B)

     •   The  Costs   of   Defending  Developed
        Shorelines Along  Sheltered Waters of the
        United States  from a Two-Meter Rise in
       Mean   Sea  Level
 -   Weggel,  Brown,
and  Doheny, Drexel
       Escajadillo,  Breen,
       University (Volume B)

       Effect of Climate Change on  Shipping
       Within Lake Superior and Lake  Erie  -
       Keith, DeAvila, and  Willis,  Engineering
       Computer Optecnomics (Volume H)

Impacts on Miami, Cleveland,  and New
York City

    Walker et al. examined three cities distinctly
affected by climate change to determine a range of
impacts on urban infrastructure.
 Study Design

     The study was based on a critical review of
 existing infrastructure studies in the three  cities,
 discussions of likely impacts with local infrastructure
 experts, analyses undertaken by these experts, and
 preliminary calculations  of probable  impacts.
 Experts were presented with GCM scenarios for
 CO2 doubling, and scenarios were used to calculate
 effects on  energy demand,  roadways, and  other
 systems.  The study also derived conclusions  based
 on experiences in  other  cities  where  current
 temperatures   are  analogous  to  temperatures
 projected  for  the cities under study,  using the
 analogs identified by Kalkstein (Volume G).


     The principal limitation  of the overall study is
 the limited use of hydrologic and other modeling.
 In  addition,   experts   were  asked  to  derive
 conclusions  regarding   conditions  beyond   their
 experience. Since only three cities are included, the
 full range of effects on urban infrastructure was not
 covered.  The authors did not perform engineering
 analyses of cost-effective responses, and they did not
 assess the potential for  reducing impacts through
 technological change.  Thus, these results should be
 considered as approximations  of the costs of impacts
 and  as illustrative  of  the  sensitivity  of urban
 infrastructure to climate  change.

 Results and Implications

Miami's Infrastructure

     Greater Miami is  bounded by water on all
 sides during the rainy season.  An extensive network
 of canals and levees has been  built to control ocean
 and freshwater flooding and to recharge the aquifer
beneath the area.  Miami has one of the world's
most porous aquifers, which lies less than 1.5 meters
below the surface in one-third  of the developed
area.  Federal  law requires  that roughly 15%  of
Miami's freshwater be released into the Everglades
National Park.

    The Miami case study examined the probable
impacts of  climate change and  sea level  rise on
Dade County's water supply, water control and
drainage  systems, building  foundations, roads,
bridges, airports,  solid  waste disposal sites, and
sewage transport and treatment systems, assuming
that  a gradual  sea level rise would be managed
through strategies such as  raising the land in

                                                                                  Urban Infrastructure
low-lying areas, upgrading levees  and dikes with
pumped outflows, retreating selectively from some
areas, and increasing the freshwater head roughly in
proportion to sea level rise to  prevent saltwater
infiltration into the aquifer.

    As Table  13-2 shows, global  climate change
could require more than $500  million in capital
investment in Greater Miami over the next century.
Because needed investments in many systems could
not  be   estimated  and  because  a  complete
engineering  analysis  was  not  performed,  these
results  should  be  considered  only  as  rough
estimates. They imply an increase  of 1% to 2% in
Greater Miami's capital spending for the next 100
years, no more than $20 per household per year at
1985 population levels (Metropolitan Bade County
Planning Department, 1988).

    Because  the  south Florida aquifer  extends
under the ocean, the typical urban response to a
rising sea ~ diking the water at the surface and
pumping out the seepage from ditches behind the
dikes ~ appears to be unworkable.  Unless the dike
extended  downward more  than 45 meters, rising
seawater pressure would cause the  sea to rush into
the aquifer below the surface and push freshwater
upward, almost to the surface.

    The  one-time  capital costs   for  upgrading
existing canals and levees in response to a 1-meter
sea level rise could be about $60 million. However,
almost  $50  million  in  new control structures,
including  extensive  pumping  capacity,  might be
required  for  the canals used  to maintain the
freshwater head. Large-scale pumping along canals
also could involve substantial  operating costs, but
these have not been estimated.  Storm sewers and
drainage would need upgrading, requiring invest-
ment of  several hundred million dollars  above
normal replacement costs.

     Building foundations generally should remain
stable if the freshwater head rises 1 meter because
houses are built on concrete slabs, most buildings in
newer areas already are built on raised lots to meet
Dade County's flood  control ordinance,  and the
foundations of many larger buildings are designed to
extend into  the water table.

     Conversely, the water table  could infiltrate the
base of about a third of Dade County streets, which
Table  13-2. Probable  Infrastructure Needs and
         ,   Investment in Miami in Response to a
            Doubling of CO2 (millions of  1987
  Infrastructure need
Raising canals/levees
Canal control structures
Raising streets
Raising yards and houses
Pumped sewer connections
Raising lots at reconstruction
Raising bridges
Sewer pipe corrosion
Water supply
Electric generation
not estimated
250 added to
not estimated
not estimated
not estimated
not estimated
not estimated
20-30% capacity
aCosts are partially based on Weggel et al.,
 Volume B.
Source:  Walker et  al. (Volume H); Linder and
Inglis (Volume H).
would have to be raised or risk collapse. If sea level
rose gradually, thereby permitting raising of streets
and related sewer mains during scheduled recon-
struction,  the   added   public   cost  might  be
approximately $250 million. Building owners would
incur substantial  costs to improve drainage,  raise
yards, raise lots at reconstruction, and pump sewage
to mains. Miami's airport also would need better
drainage, requiring an approximately $30 million

     A  1-meter rise  in  sea level would require
raising most bridges to ensure adequate clearances
and reduce vulnerability to  storm surges during

     It is unclear  what effect climate change will
have on hurricanes.   Without increased  hurricane
activity, climate change probably would exacerbate
water shortages that  are expected to result  from
population growth in Greater Miami. Thus, climate

Chapter 13
change could accelerate Miami's long-range plan
for large-scale production of desalinated water at
three times  current water prices.  If hurricanes
increase, Miami's added expense for water supply
might be roughly $100 million to move some wells
farther inland.   Conversely, increased  hurricane
frequency  and  intensity  could cause billions  of
dollars in property damage.

     Analysis of Miami's coastal defense and water
supply options provides insight into the impacts of
sea level rise on cities built on coral reefs, but not
into the response of most mainland cities on the
U.S. coastline.  Dade County  is unusual because
readily extracted  fill is extensively available  on
public lands  having easy access to a canal system
that can be  navigated by flat-bottomed barges.
Nevertheless, this case  study suggests that global
climate change could cause large coastal cities  to
invest billions of dollars  over the next 50 to 75 years
to add and upgrade infrastructure.

Cleveland's Infrastructure

     The Cleveland case study examined impacts of
climate change on snow and ice control costs, road
construction  and maintenance, heating and cooling
costs and equipment needs, water supply, and storm
and wastewater transport.  The study also included
a preliminary analysis of the effects of a drop in the
level of Lake Erie  as  estimated by Croley (see
Chapter 15: Great Lakes). The impact on the snow
and ice control budget was estimated by analogy to
the budget in Nashville, Tennessee.

     Results  are displayed in  Table 13-3, which
shows  that the net impact of  climate change on
Cleveland's annual infrastructure costs  could be
negligible,  although  expenditures  probably would
shift between categories.  In addition to the costs
shown in Table 13-3, a one-time capital expenditure
of $68 to $80 million could be required to add air
conditioners in public buildings. Also, many private
residences probably would install air conditioners.

     Walker  et al. estimated that global climate
change could cause annual snowfall in Cleveland to
drop from 1.25 to roughly 0.2 meters (4.1 to 0.7
feet), reducing annual snow and ice control costs by
about  $4.5 million.   Decreased frost damage  to
roads  and bridges  could  yield  further  savings
estimated at  $700,000 per year.  A drop of $2.3
million per year in heating costs for public buildings
also was estimated. Conversely, annual public air-
Table 13-3.   Estimated Impacts of a CO2 Doubling
             on Cleveland's Annual Infrastructure
             Costs (millions of 1987 dollars)
Infrastructure category
operating costs
Snow and ice control
Frost damage to roads
Road maintenance
Road reconstruction
Mass transit
River dredging
Water supply
Storm water system
+ 6.6-9.3
   summer increase
    offsets winter
   less than 0.5
   -1.6 to +$1.1
Source:  Walker et  al. (Volume H); Keith et al.
(Volume H).
conditioning costs seemed likely to rise by $6.6 to
$9.3 million.  The impacts on the transit operating
budget seemed likely to mirror the impacts on the
general budget, with reduced mishaps  and traffic
delays in ice and snow offsetting increased fuel costs
for vehicle cooling.

     The  study  suggested Cleveland might spend
about $65 to $80 million to add air-conditioning to
older schools and to large nonoffice spaces such as
gyms and repair garages. Much of this expenditure
would  occur   as  buildings  were  replaced  or
refurbished and might have  occurred even without
climate change.

     The rise in winter temperatures associated with
a doubling of CO2 might allow Cleveland to use
thinner pavement, resulting in possible savings of
about  3% in  road resurfacing  costs and 1% in
reconstruction costs. The net savings could average
about $200,000 per year or 1.3% of the city's current
capital budget. Engineering standards (AASHTO,
1987)  suggested  that   the  rate  of  pavement
deterioration probably also should decline as whiter
temperatures rise, saving roughly $500,000 per year.

                                                                                 Urban Infrastructure
    A climate-induced drop in the level of Lake
Erie probably would not adversely affect Cleveland,
although some dredging might be required in the
Cuyahoga River and port area (Keith et al., Volume
H).  Upgrading  of the city's combined storm and
wastewater  collection system  appeared   to  be
unnecessary,  although this  would  depend upon
rainfall variability.

    If temperature  rises several  degrees, most
northern cities probably could anticipate savings in
snow  and ice  control,   heating,   and  roadway
construction and maintenance costs similar to those
described for Cleveland.   These  savings might
approximately offset the increase in air-conditioning
costs.   More southern  cities could  experience
modest budget increases.

    Cleveland could become a more attractive
location  for  water-intensive industry  if water
supplies  in  other  areas become  less reliable.
Resulting in-migration could bring further  growth-
related infrastructure costs..  Lower Great Lakes
levels could require dredging, modification to ports,
and relocation  of some  water intakes.  (For a
further discussion of these issues, see Chapter 15:
Great Lakes.)

New York City's Water Supply

    New York City's infrastructure maybe affected
in  many  ways  by global  climate  change.
Temperature change could affect the same capital
expense  categories  in both New York City and
Cleveland.    In  addition, the city  may have  to
gradually  raise  its  dikes   and  better  protect
underground  infrastructure   from   seawater
infiltration.   Interpolating  from Weggel  et al.
(Volume B),  approximately  $120 million might be
invested to protect shorelines from a sea level rise
of 1 meter. The most pressing, and perhaps largest,
problem facing the city may be the effects of global
climate change on the adequacy of the city's water
supply.  The  New York City study focused on that
issue. Table  13-4 provides estimates drawn from a
number  of studies about possible  infrastructure
impacts on New  York City.

    The New York metropolitan area draws water
from  the adjoining Hudson and Delaware River
Basins and  from  underground aquifers  serving
coastal New Jersey and Long Island.  Figure 13-1
shows the region and its water supply sources.
Table 13-4.  Probable Impacts of a CO, Doubling
            on Selected Infrastructure in the New
            York Metropolitan Area (millions of
            1987 dollars)
Upgrading levees


Sewer outflows

Water supply

Snow and ice control

increased flooding in low-
  lying areas, minimal
  sewer system changes

more frequent inspection


reduced substantially
Road maintenance and  winter savings, offset by
  reconstruction           melting asphalt in
Mass transit
summer increase offsets
  winter savings
Electricity production   65-150
NOTE: Impacts  on underground  infrastructure,
        airports, and ports have not been probed,
        but a discussion of these impacts among
        Port Authority representatives and  other
        experts at  the Second North American
        Conference  on Preparing  for   Climate
        Change,  Washington, DC, December 7,
        1988, suggested they might be small.
Source: Walker et al. (Volume H); Weggel et al.
(Volume B); Linder et al. (1987); Schwarz  and
Dillard (1989).
     The water supply network is in deficit.  The
Mayor's  Task  Force  (1987)  has  recommended
remedying New York City's portion of the deficit
through better management of water demand and
detailed study of the possibility of reactivation of a

Chapter 13
                                                              ATLANTIC OCEAN
Figure 13-1. The sources of New York City's water supply (New York City Municipal Water Finance Authority,

                                                                                Urban Infrastructure
water intake at  Chelsea, a $223 to $391  million
investment that would yield 375 to 750 million liters
of water daily.

    Walker  et  al. estimated changes  in water
demand using design  standards for commercial
cooling-tower  demand,  changes  in  electricity
demand estimated by  Linder et al. (1987), and
historic residential summer water use. The impact
of sea level rise on water supply was estimated by
analogy using Hull and Titus (1986), which analyzes
possible saltwater advance up  the Delaware River.
The impact on reservoir supply also was estimated
by  analogy,  using a Great Lakes  water balance
model (Linder et al., 1987).  Walker et al. assumed
that baseline demand  would  not increase above
projected   demand   in    2030,   potentially
underestimating the increased  demand for water.

    Walker  et   al.  estimated  that  a  rise  in
temperatures consistent with the GISS and GFDL
scenarios would  mean about  a 20% increase in
cooling  degree days.   In response,  average daily
demand for  water used in cooling large buildings
could increase by 190 million liters during  the
summer, and increased lawn watering could raise
demand by  110  million liters  per day,  thereby
generating a 5% rise in annual demand.

    Higher   temperatures    could    increase
evaporation and evapotranspiration, decreasing the
ability  to  store  water  efficiently  in   surface
impoundments. The water balance model indicated
the supply loss could range from 10 to 24%.

    Saltwater infiltration due to rising sea level
would further reduce supply.  The study suggested
that a  1-meter  sea level  rise could  place  the
proposed $300 million Chelsea intake below the salt
line during the peak summer demand period in mild
drought years,   reducing  supply another 13%.
Larger sea level  rise or greater droughts might
prevent  use  of the existing Poughkeepsie intake
during severe droughts,  further reducing supply.  In
addition, subsurface infiltration could reduce  the
supply available from the Long Island aquifer.

    In summary, a doubled CO2 atmosphere could
produce a shortfall equal to 28 to 42% of planned
supply in the Hudson River Basin.
Implications  Arising  from  Other  EPA
Studies in This Report

     Linder and Inglis (Volume H; Chapter 10:
Electricity  Demand) suggest that  increased  air-
conditioning use could raise peak electricity demand
by 10 to 30% in the southern half of the United
States.     Nationally,   utilities  supplying   the
northernmost  cities  could experience  decreased
demand, while those  supplying  cities   in  the
remainder  of  the  country  could  experience
electricity needs higher than they have anticipated.
Sheer's study of California (see Chapter 14) water
supply   suggests   that   new  surface   water
impoundments maybe needed to meet urban water
needs and  other demands.  The coastal defense
strategies suggested hi  Chapter  7: Sea Level Rise
would apply to most urban coastal areas, especially
those along the Atlantic and Gulf coasts.

     Changnon et al. (Volume H) conclude that a
falling lake level might  prompt investment of $200
to  $400   million   to   adapt  recreational   and
commercial harbors  and beach facilities, and an
investment  of $20  million to adjust water supply
intakes  and   sewer  outfalls along  the  Illinois
shoreline of Lake Michigan, with similar costs likely
on the other  Great Lakes.  The Keith study (see
Chapter  15:   Great  Lakes)  suggests that  each
commercial harbor  on Great  Lakes  Erie  and
Superior could spend $5 to $30 million on dredging
to maintain harbor access.

Metropolitan Water Supply

    Schwarz  and   Dillard  (1989)   conducted
telephone  interviews  with  local  infrastructure
managers to identify the probable impacts of global
climate  change on water supply and drainage in
several  metropolitan areas.  Results from some
cities are discussed here.

Washington. DC

    Longer hot spells could warm the Potomac
River and  cause trihalomethane formed  during
chlorination  to   rise  above   allowable  limits.
Remedying this could require a  capital investment
of roughly  $50 to $70 million and  could increase

 Chapter 13
treatment  costs.   Also, lawn  watering  probably
would increase during  long  spells of  hot,  dry
weather. Although a substantial decrease in runoff
could reduce supply in  parts of the system,  the
availability of  additional storage  capacity would
make a shortage unlikely.

New Orleans

     Sea level  rise could necessitate moving  the
water intakes considerably farther up the Mississippi
and replacing  cast iron water mains that would
corrode if exposed to saltwater. Reduced riverflow
also  could  increase   settling  and   treatment
requirements.   Rising sea level  could increase
saltwater infiltration into the water system and could
require increased pumping capacity.

New York City

     This study raised many of the same concerns
regarding water supply and demand as the study by
Walker et al. (Volume H) and indicated that even
a 0.25-meter sea level rise would mean the proposed
Chelsea  intake was too far  downstream.   The
sanitary  and storm  sewage system capacity and
design  probably   would   not  need   revision.
Nevertheless, in a few low-lying areas, higher  sea
level could increase  sewer backups, ponding, and
basement flooding when high tides coincided with
high runoffs.


     Tucson is depleting its aquifer despite substan-
tial  conservation efforts  and lawn  watering with
treated  wastewater.  Higher temperatures would
increase  demand  and  tighten supply,  possibly
jeopardizing the city's ability to draw on water from
the Central Arizona Project on the already strained
Colorado River.  While  modest savings might be
achieved through stricter conservation measures and
more wastewater use, purchase of water in  the
regional  market most likely would be  the only
practical response to climate-related shortfalls.

    The implications of climate change for urban
America vary spatially. Some  localities, especially
those along the  Great Lakes, might  experience
roughly  offsetting  gains  and  losses.    Others
especially those along the coastlines and in water-
short areas,  could  bear  increased infrastructure
costs. The costs would be especially high if changes
came through abrupt "sawtooth" shifts or increases
in extreme  events, making it difficult to adapt
infrastructure primarily during normal repair and
replacement.  The  likely impacts of an  effective
doubling of atmospheric CO,  could  affect a wide
range of infrastructure. Additional climate change
effects beyond doubled CO2 or sea level rise above
1 meter could result in even greater costs.


     Hotter   temperatures  could  cause  faster
evaporation  of groundwater and raise the demand
for water to support commercial air-conditioning
systems and lawn watering.  Earlier snowmelt in the
West could force a lowering of dam levels to ensure
availability of enough capacity  to  control flood
waters. At the same time, sea level rise could cause
saltwater to advance up rivers and to infiltrate into
coastal aquifers. In droughts, many existing water
intakes might deliver brackish water.

     The solution to these problems could involve
strong conservation measures, such as miles of
aqueducts from new water intakes at higher river
elevations, new reservoirs, sewage effluent recycling
systems to support commercial cooling  or  lawn
watering, and perhaps  desalinization  efforts along
the  coasts.   The  solution for  the  New York-
Philadelphia corridor alone is likely to cost $3 to $7
billion. Communities in the Delaware River Basin,
northern New Jersey, the lower Hudson, and Long
Island  might well form a multistate  water supply
and management district of unprecedented size and
complexity  to   handle  financing   and   capital

Drainage and Wastewater Systems

     Increased storm size and intensity could tax
many storm sewer systems. Sea level rise also could
reduce coastal flood protection levels in low-lying
areas.  The  resulting  increases in flooding  and
releases of untreated waste into watercourses from
combined storm and wastewater systems probably
would motivate new sewer investments.  In Bade
County alone, costs  to maintain flood  protection at
existing levels could be $200 to $300 million if sea
level rose 1 meter.

                                                                                  Urban Infrastructure
     Temperature rise  could increase hydrogen
sulfide formation in sewer pipes, leading to internal
corrosion and eventual failure. In coastal areas with
increased ocean flooding, storm sewers would carry
corrosive saltwater with increased frequency.  Sea
level rise also could cause more pipes in coastal
areas to face the external risk of corrosive seawater.
More frequent inspection and earlier replacement of
much existing pipe, as well as a gradual shift to
more corrosion-resistant pipe with plastic lining,
might be required.

Coastal Defenses

     Protection from  a rising sea  could require
periodic   investment   in   many  major  coastal
communities.  In urban areas, a common approach
might  be   the   New  Orleans  solution,  where
extensively developed coastal areas are protected by
dikes, and covered drainage ditches behind the dikes
are pumped to keep out the saltwater.


     Rising temperatures could reduce the costs of
road construction and maintenance. Snow and ice
control costs might drop dramatically. A decrease
in deep freezes and freeze-thaw  cycles also would
mean fewer potholes.  Warmer temperatures and
the  improved  drainage  resulting  from  higher
evaporation rates could  permit  use  of thinner
pavements  in  many areas,  but  could  require
enhanced expansion capabilities.


     Sea level rise  and increased storm  intensity
could require upgrading of many bridges either
through costly retrofit or as part of normal recon-
struction. The range of temperature accommodated
by expansion joints also might  need to be increased.
The costs  might be  modest if bridge  planners
upgraded in anticipation of climate change.

Mass Transit

     In   the  North,  buses  and  railcars could
experience fewer  snow-related delays.  Conversely,
slight increases  in  fuel costs could result  from
increased use of air conditioners.
Electricity and Air-Conditioning

     Hotter temperatures could increase  air-con-
ditioning use. Consequently, peak load capacity to
generate electric power might have to increase in
response to global climate change. Fortunately, air-
conditioning equipment is replaced frequently, so
increased loads on  existing equipment could be
accommodated incrementally.   Some houses and
public buildings in northern climates might need to
add air-conditioning, but such retrofitting has been
performed since the first window air conditioners
were introduced.

     The  possibility  of  global  climate  change
increases  the  risks of  infrastructure investment.
Application of design standards and extrapolation
from historical data still may not provide reasonable
assurance  that  water  and power  supply,  dam
strength and capacity, bridge clearances, or storm
sewerage capacity will  be adequate  for  the 35-,
50-,  and 100-year design cycles of these facilities.
For   example,  the   National  Flood Insurance
Program's maps identifying the historical 100-year
floodplain and 500-year floodway may no longer
provide a reliable basis for local building and zoning
ordinances designed to minimize flood losses to life
and property.

Investment  Analysis Methods

     Especially in coastal  areas, the possibility of
global climate change may soon  require careful
decisions regarding how and when  to adapt the
infrastructure.    A  strong emphasis  on  lifecycle
costing  and upgrading  during reconstruction  in
anticipation of future changes  could yield large,
long-term cost savings.  To accomplish this goal,
such institutions as the Department of Housing and
Urban Development might work with the American
Public Works Association, the National League of
Cities,  the  U.S. Conference   of  Mayors,   the
American Planning Association, and similar groups
to  educate their  constituencies  regarding   the
uncertainties and ways to incorporate them into the
decisionmaking process.

Chapter 13
Water Supply

     Water supply is of particular concern because
decades are required to plan and complete projects,
which then might last 100 years. Dams, reservoirs,
and water intakes currently being planned and built
could become obsolete or inadequate as a result of
global  climate  change.   Elsewhere, communities
might be allowing development of land needed for
reservoirs to  meet the water shortages that would
result from climate change.

     Such federal agencies as the U.S. Geological
Survey, U.S. Army Corps of Engineers,  and EPA
may wish to work with states and municipalities to
study the possible impacts of climate change  on the
water supply of major metropolitan areas.

     Water supply  investments frequently affect
multistate  areas,  creating a need  for  federal
coordination. The Supreme Court has  been forced
to settle previous water rights disputes concerning
many major rivers, and global climate change might
well generate new disputes. Cost-effective response
to climate change also might require new multistate
water projects. For example, a major project on the
Hudson  River  that  allowed  New York City to
reduce its use of Delaware River water might be the
least costly  way to  increase  water  supply in
Philadelphia.   The upcoming state debates over
water supply  financing should be informed by the
lesson  of past infrastructure crises: water  piping
and  pumping costs resulting  from global climate
change should be fully recovered from the water
users to avoid stimulating artificial demand for
bargain water.

Infrastructure Standards

     Voluntary standards organizations, such as the
American Society of Civil Engineers, the Building
Officials  and Code Administrators International,
and the American Association of State Highway and
Transportation Officials, may wish to educate their
committees on global climate change.  Growing
uncertainty   concerning   future   temperature,
precipitation,   and  sea   levels  might dictate  a
reassessment of existing standards and safety factors
for ventilation,  drainage, flood protection, facility
siting,  thermal tolerances, resistance to corrosion,
and  so forth.  Conversely, prompt detection of
lasting   changes  could  allow   adjustment  of
geographically based standards — for example, on
roadbed depth and home insulation levels --  and
provide significant savings.  Thus, the standard-
making organizations might beneficially establish
policies concerning how and when their committees
should account for global climate change or educate
their committees about the prospects.

     The following are recommended for  further

1.   More  case studies  of  urban impacts,  with
     priority on a west coast city and an inland city.
     Issues of particular interest include the effects
     on subsidence  problems in cities similar to
     Phoenix, the implications for sewage treatment
     capacity in areas where more frequent and
     intense periods of low riverflow could reduce
     acceptable effluent discharge rates, the impact
     on bridge replacement costs, and the potential
     for and probable consequences of saltwater
     infiltration  into  pipes  in  older   coastal

2.   The probable impacts of global climate change
     on domestic and international migration flows
     and the infrastructure demands these flows
     produce.  Heat  and high water prices might
     drive jobs and people away from some regions,
     while  others might  flourish.  Infrastructure
     investment in new water supply, for example,
     might be unnecessary in areas that would lose
     population, but extra capacity might be needed
     in  areas   where  population  would grow.
     Similarly,  as climate change shifts the  best
     growing areas for specific crops, new farm-to-
     market transportation networks might need to
     be developed. Rights-of-way for these systems
     might best be set aside now, before land prices
     rise in response to climate change.

AASHTO. 1987.  American Association of State
Highway and Transportation Officials.'  Manual for
the Design of Permanent Structures, Appendix A.
Treatment of Roadbed Swelling and/or Frost Heave

                                                                                Urban Infrastructure
 in Design. Washington, DC: American Association
 of State Highway and Transportation Officials.

 Bruske, E. 1988.  104 (phew!) degrees hottest in 52
 years.  The Washington Post 111(225):A1, A6. July

 Cohen, S.J.  1987. Projected increases in municipal
 water use in the Great Lakes due to CCyinduced
 climatic  change.    Water  Resources  Bulletin
Hirsch, J.   1988.    As  streets  melt,  cars  are
flummoxed by hummocks.  The New York Times
137(47599):B1, B5.  August 16.

Hull, C.H.J., and J.G. Titus, ed. 1986. Greenhouse
Effect, Sea Level Rise, and Salinity in the Delaware
Estuary.   Washington,  DC:  U.S.  Environmental
Protection Agency.  Publication No. 230-05-86-010.

Lewis, N.  1988.  Two more heat records fall as
summer of 1988 boils on. The Washington Post 111
(257):A1, A10, All. August 18.

Linder, K.P., MJ. Gibbs, and M.R. Inglis.  ICF
Incorporated. 1987.  Potential Impacts of Global
Climate Change on Electric Utilities. Albany, NY:
New York State Energy Research and Development
Authority. Publication No. 824-CON-AEP-86.

Mayor's Intergovernmental Task  Force on New
York City Water Supply Needs.  1987.  Managing
for the Present, Planning for the Future; December.

Metropolitan Dade  County Planning Department.
1988. Comprehensive Development Master Plan for
Metropolitan Dade County, Florida. July 1979, July
1985, June 1987, and April 1988.

National Council on Public Works Improvement.
1988.  Fragile Foundations - Final Report to the
President and Congress, Washington, DC; February.
 New York City Municipal Water Finance Authority.
 1986.  Water and Sewer System Revenue Bonds,
 Fiscal 1986, Series A. Prospectus. New York.

 Rhoads, P.B., G.C. Shih, and R.L. Hamrick.  1987.
 Water resource planning concerns and changing
 climate:  a Florida perspective. In:  Proceedings of
 the Symposium on Climate Change in the Southern
 United States:  Future Impacts and Present Policy
 Issues. Norman, OK: University of Oklahoma, pp.

 Schwarz, H.E., and L. Dillard. 1989. Urban water.
 Chapter III-D. In: Waggoner, P.E.,  ed. Climatic
 Variability,  Climate  Change,  and  U.S.   Water
 Resources.  New York:  John Wiley and Sons.  In

 Sossaman, BA.  1988a. News release. U.S. Army
 Corps of Engineers, New Orleans District. June 28.

 Sossaman, BA.  1988b. News release. U.S. Army
 Corps of Engineers, New Orleans District. July 15.

 Statistical Abstract of the  United States.   1988.
 Washington, DC: U.S. Government Printing Office.

 Titus, J.G., C.Y. Kuo, MJ. Gibbs,  T.B. LaRoche,
 M.K. Webb, and J.O. Waddell. 1987.  Greenhouse
 effect, sea level rise, and coastal drainage systems.
 Journal   of  Water   Resources   Planning  and
 Management 113(2):216-227.

The Washington Post. 1988.  Warped rails checked
 in Amtrak wreck.  111(246)A5.  August 7.

Wilcoxen, PJ. 1986.  Coastal erosion and sea level
rise:   implications for  Ocean Beach and San
Francisco's Westside Transport Project.  Coastal
Zone Management Journal 14(3):173-191.


                                       CHAPTER 14


Global warming could cause higher winter runoff
and lower spring runoff in California and increase
the difficulty of meeting water supply needs.  It
could also increase salinity in the San Francisco Bay
and the Sacramento-San Joaquin Delta and increase
the relative abundance of marine species in the bay;
degrade  water quality  in  subalpine lakes; raise
ambient  ozone levels; increase electricity demand;
and raise the demand for water for irrigation.

Water Resources

•   Higher temperatures  would lead to  higher
    winter runoff from the mountains surrounding
    the  Central Valley, because less precipitation
    would fall as  snow, and the snowpack would
    melt earlier.  Runoff  in the late spring and
    summer consequently would be reduced.

•   As a result, the amount and reliability of the
    water supply  from reservoirs  in the Central
    Valley Basin would decrease.  Annual water
    deliveries from the State Water Project (SWF)
    could be reduced by 200,000 to 400,000 acre-
    feet or 7 to 16%. In comparison, the statewide
    increase for water from the SWP, due to non-
    climate factors such as population growth, may
    total 1.4 million acre-feet by 2010.  Even if
    operating   rules  were  changed,  current
    reservoirs would not have the capacity to store
    the heavier winter run