POTENTIAL IMPACTS OF CLIMATE CHANGE ON ELECTRIC UTILITIES New York State Energy Research and Development Authorlty% From the digital collections of the New York State Library. The New York State Energy Research and Development Authority (NYSERDA) is a public benefit corporation chartered by the New York State Legislature. It is governed by a 13-member Board of Directors ap- pointed by the Governor with the consent of the Senate. State Energy Commissioner William D. Cotter ls Chairman of the Board and the Chief Exemtive Officer. A President manages the Authority's programs, staff, and facilities. . As expressed in its enabling legislation, the underlying rationale for establishing the Authority is: . . . that accelerated development and use within the State of new energy technologies to supplement energy derived from existing sources will promote the State?s economic growth, protect its environmental values and be in the best interests of the health and welfare of the State?s population . . . The legislation further outlines the Authority?s mission as: . . . the development and utilization of safe, dependable, renewable and economic energy sources and the conservation of energy and energy resources. The Authority?s policy and program stress well-designed research, devel0pment and demonstration projects, based on technol- ogies with potential for near-term commercialization and application in New York State. The Authority seeks to accelerate the introduction of alternative energy sources and energy-efficient technologies and to im- prove environmental acceptability of existing fuels and energy pro- cesses. The Authority also seeks to ensure that Federal research pro grams reflect the needs of the State. The use of New York contractors and an awareness of energy-related growth opportunities are part of the Authority?s effort to support in- dustry in New York. Concentrating on these objectives ensures that NYSERDA's programs will produce maximum benefits for the citizens and businesses of New York, while attracting the participation of both the private sector and the Federal Government. NYSERDA derives its research and development revenues from an assessment upon the intrastate sales of the State?s investor-owned gas and electric utilities. The Authority also derives income from the invest- ment of retained earnings and leased property, as well as from bond financings of pollution control facilities and special energy projects. Further information about NYSERDA's programs may be obtained by writing or calling the Department of Communications, New York State Energy Research and Development Authority, Two Rockefeller Plaza, Albany, NY. 12223; (518) 465-6251. Mario M. Cuomo William D. Cotter Governor Chairman State of New York New York State Energy Research and Development Authority From the digital collections of the New York State Library. POTENTIAL IMPACTS OF CLIMATE CHANGE ON ELECTRIC UTILITIES Final Report Prepared for NEW YORK STATE ENERGY RESEARCH AND DEVELOPMENT AUTHORITY Project Manager Dr. Fred V. Strnisa and EDISON ELECTRIC INSTITUTE c. Project Manager Richard Bozek, Jr. and ELECTRIC POWER RESEARCH INSTITUTE Project Manager Dr. Richard Richels and u.s. ENVIRONMENTAL PROTECTION AGENCY Project Manager Dr. Dennis A. Tirpak Prepared by reF Incorporated 1850 K Street, Northwest Washington, D.C. 20006 Project Managers Kenneth P. Linder Michael J. Gibbs Mark R. Inglis 824-CON-AEP-86 Energy Authority Report 88'-2 December 1987 From the digital collections of the New York State Library. NOTICE This report was prepared by ICF Incorporated in the course of performing work contracted for and sponsored by the New York State Energy Research and Development Authority, Edison Electric Institute, Electric Power Research Institute and the u.s. Environmental Protection Agency (hereafter the "Sponsors"). The opinions expressed in this report do not necessarily reflect those of the Sponsors or the State of New York and reference to any specific product, service, process or method does not necessarily constitute an implied or expressed recommendation or endorsement of same. Further, the Sponsors and the State of New York make no warranties or representations, expressed or implied, as to the fitness for particular purpose, merchantability of any product, apparatus or service or the usefulness, completeness or accuracy of any processes, methods or other information contained, described, disclosed or referred to in this report. The Sponsors and the State of New York and the contractor make no representation that the use of any product, apparatus, process, method or other information will not infringe privately owned rights and will assume no liability for any loss, injury, or damage resulting from, or occurring in connection with, the use of information contained, described, disclosed, or referred to in this report. From the digital collections of the New York State Library. TABLE OF CORTERTS Section .flu S-l SUMMARY ~ 1. INTRODUCTION 2. SCENARIOS OF POTENTIAL FUTURE CLIMATE CHANGE e e a a a a a a. •• 000 ••••• ••••• ••••••••••••••••• Approach Temperature Change Scenarios Stream Flow Scenarios 3. ••••••• •••••••••• UTILITY IMPACT METHODOLOGY 3-1 SOUTHEASTERN UTILITY CASE STUDy 3-1 3-1 3-6 3-10 3-12 3-12 3-14 4-1 ~ 6. eee 00. "...... CONCLUSIONS 4·12 4-15 4-17 5-1 e 0 •••••••••••••••• 00 ••••••• 0 • • • 0 ••••••••••••• • • • • • • • • • • • • • • • • • • • • •• 0 ••••••••••• •••••••••••••••••••••••••••••• 0 0 • • • • • • • •• ' . . . . . . . . . . . . . . . . . . . . . . . . . . . . .. Comparison of Case Study Results Implications of Findings Suggestions for Further Research APPENDIX A: UTILITY PlANNING FACTORS 4-1 4-1 4·5 NEW YORK CASE STUDY Introduction and Overview Base Case Utility Plan Weather-Sensitivity of Demand Weather-Sensitivity of Supply Alternate Planning Scenarios Utility Impacts 2-8 2-15 Introduction and Overview Base Case Utility Plan Weather-Sensitivity of Demand Weather-Sensitivity of Supply Alternative Planning Scenarios Utility Impacts 5. 2-1 2- 2 0 Overview Task 1: Develop Base Case Utility Plan Task 2: Develop Weather Sensitivity of Demand Task 3: Develop Weather Sensitivity of Supply Task 4: Develop Alternative Scenario Outputs Task 5: Compare Base Case/Alternative Case Outputs Sensitivity Analysis: Specification of Alternative Utility Planning Scenarios ~ 4. 1-1 5-1 5-3 5-6 5-14 5-20 5-21 6-1 6-1 6-4 6-9 '. . . . . . . . . . . . . . . . . . . . .. From the digital collections of the New York State Library. A-I LIST OF FIGURES Figure S-1 Relative Contribution of Key Assumptions to "High Impact" Results 8-5 1-1 Analytic Approach 1-3 2-1 Geographic Resolution of the GCMs Over the United States 2-9 3-1 Steps in Task 1: 3-4 3-2 Example Load Duration Curve 3-5 3-3 Steps in Task 2: of Demand 3-7 Development of Base Case Scenario Development of Weather Sensitivity 3-4 Structural Approach to Estimating Utility Demands 3-9 3-5 Steps in Task 3: of Supply Development of Weather Sensitivity 3-11 Steps in Task 4: Scenarios Development of Case Study 3-6 4-1 4-2 4-3 4-4 4-5 4-6 5-1 5-2 5-3 3-13 Southeastern Utility Percent Change in Annual Energy Requirements Alternate Climate Change Scenarios 4-14 Southeastern Utility Percent Change in Annual Peak Demand: Alternate Climate Change Scenarios 4-14 Southeastern Utility Cumulative Capacity Requirements by 2015 4-18 Southeastern Utility Cumulative Capacity Requirements Induced by Climate Change 4-19 Southeastern Utility Change in Annual Production Costs for 2015 Induced by Climate Change 4-22 Southeastern Utility Cost Impact from Utility Planning Perspective -- 2015 4-25- New York State Impact of Temperature Change on Energy Demand by Time of Day -- Weekday, August 2015 5-10 New York State Impact of Temperature Change on Energy Demand by Month in 2015 5-12 New York State Impact of Temperature Change on System Energy Demand in 2015 5-13 From the digital collections of the New York State Library. Figure 5-4 5-5 5-6 5-7 5-8 6-1 New York State Impact of Temperature Change on System Peak Demand in 2015 5-15 Stream Flow and Electricity Generation, Niagara Hydro Project 5-16 Stream Flow and Electricity Generation, St. Lawrence Hydro Project 5-17 New York State Potential Impact of Changes in Stream Flow on Hydro Generation in 2015 5-19 New York State Cost Impact from Utility Planning Perspective -- 2015 5-31 New York State Relative Contribution of Key Assumptions to "High Impact" Results 6-6 From the digital collections of the New York State Library. LIST OF TABLES Comparison of New York and Southeastern Utility Case Studies: Demand Sensitivity S-2 Comparison of New York and Southeastern Utility Case Studies: Generating Capacity-Requirements 8-3 Comparison of New York and Southeastern Utility Case Studies: Impact on Total Electricity Production Costs in 2015 8-3 1-1 Project Sponsors and Advisors 1-5 2-1 Results of 1-Dimensional Model Showing Global Warming Estimates Under Alternative Assumptions 2-6 Estimates of Equilibrium Temperature Change Due to Doubled CO 2-10 Temperature Change Estimates by Season Based on the GCMs, the Transient Run, and the 1-D Model 2-12 High Temperature Change Scenario Used to Evaluate Utility Impacts 2-14 Temperature Change Scenarios Used to Contrast with High Scenario 2-16 Change in Net Basin Supply for Five Scenarios of GeM Results 2-18 3-1 Summary of Analytical Tasks 3-2 4-1 Southeastern Utility Distribution of 1985 Sales by Type of Customer 4-2 Southeastern Utility Distribution of 1985 Capacity and Generation by Fuel Type 4-2 Southeastern Utility Average Annual Growth Rates in Demand for Electric Energy 4-4 Southeastern Utility Base Case Capacity Expansion Post-1995 4-6 Southeastern Utility Average Annual Growth Rates in Real Fuel Prices 4-6 Southeastern Utility Example of Change in Energy Requirements (Percent) by Day Type and Time of Day 4-10 8-1 8-2 8-3 2-2 2 2-3 2-4 2-5 2-6 4-2 4-3 4-4 4-5 4-6 From the digital collections of the New York State Library. Southeastern Utility Estimated Changes in Temperature and Energy Requirem~nts by Month, 2015 4-11 Southeastern Utility Estimated Impact of Climate Change on Annual Energy Requirements, Selected Years 4-13 Southeastern Utility Estimated Impact of Climate Change on Annual Peak Demand, Selected Years 4-13 Southeastern Utility Poten~ial Impacts of Temperature Change on Heat Rates and Effective Capacity for a Typical 400 MW Thermal Plant 4-16 Southeastern Utility Potential Impacts of Climate Change on Fuel Utilization, 2015 4-21 Southeastern Utility Impact of Alternative Planning Assumptions on Total Electricity Production Costs in 2015 4-24 New York State Distribution of 1985 Sales by Type of Customer 5-2 New York State Distribution of 1985 Capacity and Generation by Plant Type 5-2 New York State Composition of Upstate and Downstate Utility Regions 5-4 New York State Average Annual Growth Rates in Electrical Demand 5-4 5-5 New York State Base Case Capacity Expansion Plan 5-7 5-6 New York State Average Annual Growth Rates in Real Fuel Prices 5-7 5-7 New York State Summary of Impacts in 2015 5-22 5-8 New York State Impact on Capacity Requirements in 2015 5-23 5-9 New York State Impact on Non-Hydro Generation Requirements in 2015 5-25 5-10 New York State Impact on Generation Mix in 2015 5-26 5-11 New York State Impact on Total Electricity Production Costs in 2015 5-28 New York State Impact Alternative Planning Assumptions on Total Electricity Production Costs in 2015 5-29 4-7 4-8 4-9 4-10 4-11 4-12 5-1 5-2 5-3 5-4 5-12 From the digital collections of the New York State Library. 6-1 6-2 6-3 Comparison of New York and Southeastern Utility Case Studies: Demand Sensitivity 6-2 Comparison of New York and Southeastern Utility Case Studies: Generating Capacity Requirements 6-3 Comparison of New York and Southeastern Utility Case Studies: Impact on Total Costs in 2015 6-5 From the digital collections of the New York State Library. S1lMMAR.Y This report summarizes the analytic approach and preliminary findings of a study jointly sponsored by the Edison Electric Institute, the Electric Power Research Institute, the New York State Energy Research and Development Authori ty, and the U. S. Envirorunental Pro cec t Lon Agency. In addi tion to these organizations, an advisory board of global c l Lmaee scientists and utility executives provided comment and direction during the design phase of the study. The proj ect examines the potential impacts of greenhouse-gas-induced climate change on the demand for electricity and electric utility planning and operations in two case studies. The study focuses on the next 30 years, a period of direct relevance to utility planning, but a period of modest temperature changes compared with predictions for the latter part of the 21st Century. SUMKARY OF CASE STUDY RESULTS Table S-1 presents estimates of the changes in peak demand and total energy estimated for ~ the Southeastern utility and for Upstate and Downstate New York utilities under "high" climate change assumptions. These changes in demand are above changes over time caused by increases in population, GNP, and other factors. The high climate scenario represents the largest temperature increases estimated using General Circulation Model (GeM) results made available for this project. Although this scenario is termed "high," ut:lc~rtainties in the GeMs and other climate modeling factors do not preclude the possibility that the actual climate response will exceed this high scenario. 1 The table indicates that the estimated percent change in peak demand and electric energy requirements is greater for the Southeastern utility than for the New York utilities. This result is a product of: (1) higher estimated summer temperature changes in the Southeastern utility case study (l.87 oF versus l.46 oF in New York); and (2) higher es tima ted weather- sensitivity coefficients (1.19 to 2.27 for New York versus 3.76 for the Southeastern utility). The estimates for Upstate and Downstate New York show the range of resul ts obtained from two different modeling approaches (statistical and structural) used to estimate the weather-sensitivity of demand in the New York case studies. The estimated sensitivity of peak demand to changes in temperature in Downstate New York shown in Table 5-1 is nearly as large as the peak-sensitivity estimated for the Southeastern utility. Given the same change in summer temperature, there would be similar estimates of percent change in peak demand in the two regions. 1 Factors that affect the rate of climate change investigated in this study are: rate of trace gas concentration growth; climate sensitivity to changing atmospheric concentrations; and rate of heat transfer to the oceans. Other processes, such as cloud formation and land use changes could have significant impacts on the rate of climate change. S - 1 From the digital collections of the New York State Library. S_ TABLE 5-1 allPAlUSOR OF REV YORIt ARD UTILITY CASB STUDIES: DEMAND SENSITIVITY (HIGH TlKPIRATOU SCElWlIO) Change in Summer Temp. Nev York (OF) Peak SensitiVity ('1°F) Change in Peak Demand (') Change in Total Energy (') -0.27 to -0.21 0.49 to 1.04 Upstate Downstate 1.46 1.46 0.66 to 1.47 1.51 to 2.77 0.96 to 2.14 2.20 to 4.04 System 1.46 1.19 to 2.27 1.74 to 3.32 Southeastern Utility 1.87 3.76 7.04 0.13 to 0.45 3.40 The substantially larger percentage change in total energy consumption for the Southeastern utility (3.4 percent versus 0.13 to 0.45 percent for New York) is related. in large degree to the importance of air conditioning lo'ads in all seasons for that utility, offsetting reduced winter heating loads. In New York, air conditioning is almost exclusively a summertime use of electricity. In fact, in Upstate New York, total energy requirements fall in response to temperature increases (due to reduced winter heating loads). Table 5-2 indicates the estimated increase in generating capacity required by the year 2015, to maintain system reliability for the case study utilities. The "base requirements" shown in the table, assume that no climate' change occurs. Increases in capacity requirements due to climate change are similar in the two regions. The range of additional capacity induced by climate change in New York is 746-1429 MY and is 1417 MY for the Southeastern utility, both under the high temperature scenario assumptions. These capacities are equivalent to 1-2 large, central station power plants in each region. The percentage increases compared with base case additions during the period, range from 10-19% in New York and is estimated as 21' for the Southeastern utility. The potential impact of climate change on annual electricity production costs (annualized capital costs and annual fuel costs) for the case studies are summarized in Table S-3. Costs in New York range from $48 million to $241 million in 2015 (1985 dollars) , depending upon the approach used to estimate weather-sensitivity of demand and upon the assumed impact of stream flow changes on the availability of hydroelectric generation. S - 2 From the digital collections of the New York State Library. Stream flow changes were estimated based on the temperature change scenario and the results of a water balance model. Investigation of changes in hydrological variables (such as stream flow) due to climate change are in their early stages. Nevertheless, the scenario examined here indicates that the cost implications of climate change impacts on hydro generation as reported in Table 5-3 are significant in New York.' The high case costs in 2015 for the Southeastern utility (where hydro generation is not a factor) are similar in magnitude, ($267 million in 1985 dollars). TABLE 8-2 COKPARISON OF REV YOU. ARD SOUTIIBASTElUf UTILITY CASE STUDIES: GENERATING CAPACITY REQUIUKENTS (HIGH TlllPIllATORI SCEMlUO) Additional Requirements Induced by Climate Change Base Requirements Nev York (MW) Upstate Downstate 160 7331 155 - System 7491' 746 Southeastern Utility 6749 (MW, ') 349 591 - 1080 1429 (10t-19t) 1417 (21%) TABLE S-3 COKPARISON OF REV YOU. ARD SOlJ'J:lllASTBU UTILITY CASE STUDIES: IKPAct OR TOTAL ILEC1'llICITY PRODUctION COSTS IN 2015* (HIGB TlllPIllATORI SCERAlUO) Fuel Cost Capital Cost Total Cost Upstate Downstate -22 to +79 +44 to +112 +5 to +12 +21 to +38 -17 to +91 +65 to +150 System +22 to +191 +26 to +50 +48 to +241 +50 +267 Nev York +217 Southeastern Utility * Millions of 1985 $ S - 3 From the digital collections of the New York State Library. IKPLlCATIONS OF FINDINGS There are many uncertainties associated with developing estimates of potential climate change impacts. This project addressed uncertainties in: climate modeling; weather-sensitivity modeling; and other economic, technological, and behavioral conditions. Because these uncertainties make it difficult to predict the future with precision, the results are driven by assumptions about these factors. The relative contributions of key assumptions to the results are illustrated in Figure S -1 for the New York State "High Impact" case, which is based on the high temperature and stream flow change scenarios. Regarding the estimates of additional capacity requirements (left-hand bar), use of the data and assumptions in the statistical approach to modeling the weather-sensitivity of demand, results in an estimate of 746 MW by 2015. Alternatively, use of the structural approach and assumptions of a constant saturation of air conditioning equipment results in an estimate of 1158 MW. The additional assumption that air conditioning saturation increases over time, pushes the estimate to 1429 MW. The right-hand bar illustrates the impact of these factors on total annual electricity production costs. This bar emphasizes the importance of the estimated effects of stream flow reduction on hydro generation, and the assumptions regarding the utilities' response to these changes. The substitution of oil generation and off-system electricity purchases to offset the reduction in hydropower availability accounts for over half (52%) of the estimated total cost impact of $241 ·million (in 1985$). Smaller increments are attributed to the statistical approach to demand response ($39 million), and the assumption of increased air conditioning saturation ($28 million). Although the results are sensitive to the assumptions about these factors, this situation is little different than forecasting demand, technological change, and customer response to utility conservation or marketing programs. These types of analyses, commonly conducted by utility planners, also involve substantial uncertainties and require many assumptions. Although not precise, the estimated impacts are judged to be reasonable. The findings indicate that the potentiai impacts of climate change on electric utilities are not insignificant and that these impacts may start to occur within the typical time frame of current utility planning studies and decisions. Although the case studies have been conducted on different types of utility systems in different regions of the U.S., it is difficult to generalize to regions or the nation as a whole based solely on these results (e.g., consider the different results obtained for Upstate and Downstate New York). However, the analyses suggest the following general conclusions: S - 4 From the digital collections of the New York State Library. FIGURE 8-1 ULATIVE COBTRIBOTIOR OF KEY ASSUllPTIORS TO -HIGH DlPAcr- RESULTS (2015) Additional Capacity Requirements 100 Total Costs 1429 241 .-- b-AlC Saturation Demand Response: Structural Approach % of "High Impact" 50 746 /' Demand Response: Statistical Approach --. ~ --. 174 126 Hydro Reduction o Additional Capacity Requirements in Megawatts Total Costs in Millions of 1985$ 5 - 5 From the digital collections of the New York State Library. Climate Chanse The temperature change scenarios developed for the two case studies indicate that current general circulation model (GCM) estimates of potential regional climate change due to a doubling of atmospheric concentrations of C02 are quite diverse. Climate model outputs indicate that the rate of climate change may be uneven over time, and may vary substantially from one location to another. These results are consistent with expectations regarding how the climate system would respond to increased greenhouse gas concentrations. In light of the diversity of the GeM results, and the relative inexperience of using GeMs to perform transient analyses, the climate change scenarios must not be considered as forecasts. Although the scenarios reflect the diversity of current estimates, future climate change outside the range of estimates presented here cannot be ruled out. In terms of suggestions for further research, important areas would include improved methods for developing and disseminating climate change scenarios, with particular focus on: 1) estimates of variables (in addd t Lon to temperature) relevant for impact assessment (e. g., hydrologic variables, winds); 2) estimates of climate change over time; and 3) estimates of climate change at the regional or l~cal level. Utility Impacts It appears that climate change will have greater direct impacts on the demand for electricity than on characteristics of the supply of electricity for most utility systems: The impacts resulting from demand response to climate change are more likely to be significant for utilities with large, summer, weather-sensitive (air conditioning) loads. This is especially true for regions in the southern U. s. where air conditioning saturation and utilization is high, and for urban areas in northern climate zones where the potential for increased air conditioning saturation is high. Because of the nature and patterns of these weather-sensitive loads, response to climate change is likely to have greater impacts on peak demand (capacity requirements) than on energy consumption (generation requirements). The order of magnitude of temperature changes examined here is unlikely to have significant impacts on the effective capacity or operating efficiency of thermal generating units. However, there can be significant implications for utilities where hydro is an important source of generation. As indicated by the New York case studies, hydro generation is critical for some utilities, and the potential planning uncertainties associated with possible climate change-induced stream flow changes are large. S - 6 From the digital collections of the New York State Library. We have found that the utility capacity and cost implications of climate change potentially are significant. In the two case study analyses: (1) generating capacity additions induced by climate change are on the order of 10-20% of base case (i.e., no climate change) additions through 2015 under high temperature change assumptions; and (2) annualized capital cost and annual fuel and O&M costs induced by climate change exceed $200 million (1985 dollars) in 2015. Because of long lead-times and the capital intensity of the most efficient electric generating units, there are economic benefits associated with being able to anticipate climate change correctly. The magnitude of the potential cost savings depends on the base case planning assumptions, and in these two case studies may be as high as $50 to $70 million per year (1985 dollars) by 2015. Utility planners should start now to consider climate change as a factor affecting their planning analyses and decisions. Large impacts are not imminent, but the importance of climate change impacts for utility planning are likely to increase over time. Climate change will likely increase the uncertainties faced by utility planners and will interact with other issues they must address, including: the level and patterns of future electric demands; the availability and mix of future generating resources; and investment and financial planning. Follow-on research suggestions include: development of both broad, regional estimates of climate change impacts and additional utility-specific analyses; more detailed and complete analyses of the sensitivity of customer demand for electricity; weather- consideration of the impacts of climate change on renewable energy sources such as solar and wind; increased attention to potential secondary and indirect effects of climate change on utility investments and operations; and a more complete assessment of the value of improved climate change information to utility planners and managers. S - 7 From the digital collections of the New York State Library. From the digital collections of the New York State Library. SKerlOK 1 INTRODUCTION This project examines the potential impacts of greenhouse-gas-induced climate change on the demand for electricity and electric utility planning and operations in two case studies. The ~~ational Academy of Sciences reports that a change in the radiative properties of the Earth's atmosphere associated with a . doubling of the concentration of atmospheric carbon dioxide (C02) will raise the earth's temperature by 1.SoC to 4.SoC. 1 Increases in the atmospheric concentrations of C02 and other greenhouse gases (such as methane, chlorofluorocarbons, and nitrous oxide) have been measured,2 and one recent estimate of the potential rate of warming that may resul t from these increased concentrations is a lOC (1. 8oF) increase in global temperature by 2000. 3 Although the potential magnitude and rate of climate change remain uncertain, a recent meeting of scientists and policy makers in Villach, Austria, recommended that analyses ~ of the potential implications of alternative climate change possibilities be undertaken in order to begin to assess the importance of climate change for man's activities. 4 This study of the potential implications of climate change for electric utility planning and operations is one such study. Analysis of electric utility planning and operations is relevant for two reasons: The demand for and syp;ly of electricity is sensitive to local weather conditions. Utility studies of customer demands have shown that daily and seasonal peak electric loads are determined in large part by demands for services provided by weather-sensitive appliances and equipment, principally heating and air conditioning equipment. Further, a substantial portion of seasonal and annual electric sales for many utilities also are. determined by the use of such equipment. 1 J. Charney, Chairman, Climate Research Board, Carbon Dioxide and Climate: A Scientific Assessment, Washington, D.C., National Academy of Sciences Press, 1979. 2 Atmospheric Ozone, WHO Global Ozone Research and Monitoring Project, Report No. 16, Geneva, Switzerland, 1986. 3 J. Hansen, et a1., "The Greenhouse Effect: Projections of Global Climate Change," in Effects of Chanies in Stratospheric Ozone and Global Climate, J.G. Titus ed., U.S. EPA and UNEP, Washington, D.C., August 1986. 4 Report of the International Conference on the Assessment of the Role of Carbon Dioxide and of Other Greenhouse Gases in Climate VariatioDs and Associated Impacts, WHO-No. 661, Vi1lach, Austria, 9-15 October 1985. 1 - 1 From the digital collections of the New York State Library. On the supply side, the operating efficiency of electric generation, transmission, and distribution equipment is affected directly by temperature and other weather variables. Also, stream flow (driven by precipitation and runoff into streams, lakes, and reservoirs and evaporation from streams, lakes, and reservoirs) affects the availability of hydropower which is an important source of electricity in certain regions of the u.s. The industry is vet:y capital- intensive and has a lonK plannin& horizon, so that uncertainty in future demand and supply associated with potential changes in climate may pose substantial economic risks. These two characteristics of electric utilities indicate that changes in climate may have an important influence on supply and demand for electricity within the time horizon considered for curent investment decisions. Case studies were used to assess the potential impacts of climate change on utility planning and operations. By evaluating two utility systems in detail, the relative importance of various climate change scenarios and planning assumptions was assessed. The experience gained from these detailed case studies is instructive for analyzing groups of geographically dispersed utilities. The study focuses on the period 1986 to 2015. This 30-year period was chosen as representative of the time horizon of current utility planning decisions. The approach used to perform the case studies is illustrated in Figure 1-1. The initial steps in the analysis are to: (1) develop climate change scenarios (i.e., changes in average seasonal temperatures for a particular area of the u.s. over time); and (2) estimate the sensitivity of electricity demand and supply to changes in weather conditions. This information is used to evaluate the potential implications of climate change for the future demand for and supply of electricity. This assessment is used with a set of utility planning assumptions as inputs to a utility planning model. The model is used to evaluate the implications of climate change for generating capacity requirements, fuel utilization, electricity production costs, and other utility planning factors. Climate change impacts are evaluated by comparing these planning model outputs with base case outputs (i.e., outputs based on current climate conditions). The utility planning assumptions. were also varied to evaluate the economic risks associated with alternative planning decisions. Given that the extent and timing of future changes in climate are uncertain, utilities must make decisions today with imperfect climate-related information. The standard planning assumption implicitly employed today is that the future climate will be the same as the past climate. If the climate does change significantly within the time horizon of decisions that are based on this assumption, costly responses to changing conditions may be required in the 1 - 2 From the digital collections of the New York State Library. nGUIB 1-1 AlW.ftIC APPROAaI a J-male Chaaae • SceaanGI Weadaer-Se~liYilJ of Elecanc1'Y... DemaadlSupp., IIIIC Utility BV_tl, OD Ope_ou, COItI U&iJj&y l'Ia!miDI AllumPbODl 1 • 3 From the digital collections of the New York State Library. future By varying the planning assumptions, the potential costs and benefits of planning ahead for various amounts of climate change were evaluated. e The study benefited from direction from the study sponsors, advice from a group of reviewers, and the cooperation and assitance from representatives of the utilities being examined. The study sponsors include representatives of the electric utility industry (Edison Electric Institute and Electric Power Research Institute), the federal government (U.S. Environmental Protection Agency), and a state government agency (New York State Energy Research and Development Authority). Scientists from the climate research and modeling communities provided valuable insights during the study design and data that enabled the climate change scenarios to be developed. Table 1-1 lists the sponsors and the advisors. The remainder of this report is organized as follows: Section 2 presents the method used to develop the climate change scenarios and the scenarios used in the case studies; Section 3 presents the utility impact methods; Section 4 presents the results for the Southeastern Utility Case Study; Section 5 presents the results for the New York Case Study; and Section 6 compares the presents conclusions. results of the two case 1 - 4 From the digital collections of the New York State Library. studies and TABLE 1-1 PROJECT SPONSORS AND ADVISORS Sponsors: Edison Electric Institute, Environment and Fossil Fuels Department Electric Power Research Institute, Environment Division Energy Analysis and New York State Energy Research and Development Authority u.s. Environmental Protection Agency Advisors: W. Washington - National Center for Atmospheric Research (NCAR) J. Hansen - NASA/Columbia University S. Manabe - Geophysical Fluid Dynamics Laboratory (GFDL) J .M. Mitchell National Administration (NOAA) Oceanic and D. Rind - NASA/Columbia University T. Wigley - University of East Anglia 1 - 5 From the digital collections of the New York State Library. Atmospheric From the digital collections of the New York State Library. SEcrIOR 2 SCENARIOS OF POn;NTIAL FUTURE CLIMATE CHARGE This section describes scenarios of potential future climate changes. Based on these scenarios, a "high case" was developed for purposes of assessing the potential impacts of climate change on t-W'o electric utility systems in the United States. Two lower cases were also developed as a contrast to the high case. The th·ree cases are referred to below as scenarios A, B, and C. There is little doubt that increasing concentrations of carbon dioxide (C02) and other trace gases have the potential to modify the Earth's climate. 1 Virtually all studies indicate that increasing concentrations of these gases will significantly increase the global average temperature, although regional and seasonal impacts vary across the studies. 2 For purposes of this study, a range of scenarios of possible changes in temperature and stream flow during the period 1986 to 2015 was examined. These two climate variables were chosen as the primary climate-related factors that affect the demand for electricity and electric utility operations. 3 The term "scenario" is used deliberately in this context because although substantial progress has been made in the past decade in analyzing and modeling the potential for climate change, substantial questions and uncertainties remain. In particular, models are limited in their abilities to evaluate potential future regional changes in climate. The state of the science in evaluating climate change does not permit precise forecasts to be made at this time. 4 The scenarios developed below reflect the diversity of the estimates of climate change that have been made. However, it should be noted that the potential for climate change to be larser or smaller than these scenarios cannot be ruled out. 1 Projectins the Climate Effects of Increasine Carbon Dioxide, MacCracken, M.C. and F.M. Luther editors, U.S. Department of Energy, DOE/ER-0237, Washington, D.C., December 1985,.p. ix. This report is one of a series published by the U.S. Department of Energy on the topic of climate change and climate change impacts. 2 Ibid. 3 Additional climate-related factors that influence electricity demand and utility operations not examined here include: humidity; winds; storm frequency and severity; and soil moisture (which may affect the demand for electricity for irrigation). Increasing atmospheric concentrations of greenhouse gases will likely influence these factors. 4 The U.S. Department of Energy has published a series of reports on the state of the science in assessing the potential for climate change and climate change impacts. See for example, Projectinc the Climate Effects of lnereasini Carbon Dioxide, Maceracken, M.C. and F.M. Luther editors, U.S. Department of Energy, DOE/ER-0237, Washington, D.C., December 1985. 2 - 1 From the digital collections of the New York State Library. Tnis secc~on describes ~ne overall approach used to develop a range of climate change scenarios. Descriptions of the temperature and stream flow scenarios used in the evaluation of utility impacts are also presented. APPROACH Scenarios of future climate change were developed based on the results of computer models that describe the manner in which changing concentrations of greenhouse gases may affect the Earth's climate. The computer models use equations to represent the physical laws that govern the behavior of the Earth's climate, including the conservation of momentum, heat, and moisture. These physical laws and the physics of the various interactive processes that link them together are relatively well known for the climate system. 5 The ability of the models to simulate the current climate of the Earth, as well as the current climate of planets with very different atmospheres (Venus and Mars), provides some confidence in them. In principle, the models may represent a rational physical basis for studying the Earth's climate system, and possible perturbati~ns to it. 6 Like all models, however, the climate models have important limitations. Significant uncertainties remain regarding the proper means of representing some important phenomena, principally clouds, convection, and ocean circulation. 7 These phenomena are represented crudely in' current models. The potential for these phenomena to produce feedbacks within the climate system that magnify or reduce the impacts of greenhouse gases beyond what is currently expec ted is the subj ect of ongoing research. Should research indicate that the current representations of these phenomena are biased, future estimates of climate change due to increasing concentrations of greenhouse gases may be larger or smaller than currently expected. 8 5 Gates, W. Lawrence, "Modeling as a Means of Studying the Climate System," in Proj ecting the Climate Effects of Increasing Carbon Dioxide, MacCracken, M.C. and F.M. Luther editors, U.S. Department of Energy, DOE/ER-0237, Washington, D.C., December 1985, p. 66. 6 Ibid., p. 65. 7 Ibid., p. 77. 8 Based on a review of climate model estimates, the National Academy of Sciences reported that a doubling of atmospheric carbon dioxide would likely lead to an increase in the average global temperature of 3. ovc (S.4 0F), plus or minus 1.S oC (Z.7oF). This estimate includes evaluations of the potential for feedbacks (positive and negative) among the climate processes, and is commonly referred to as the climate sensitivity to a radiative forcing equal to a doubling of atmospheric COZ. See: Chan&in& Climate, Report of the Carbon Dioxide Assessment COmmittee, W.A. Nierenberg, Chairman, National Academy Press, Washington, D.C., 1983. 2 - 2 From the digital collections of the New York State Library. Due to the limitations of existing climate models, researchers have also relied on historical climate analogies for purposes of assessing the potential implications of climate changes. For example. meteorologipal observations from unusually warm periods have been evaluated as examples of how the full set of climate variables may change in response to global warming. The advantage of this approach is that actual climate observations are used, so that the precipitation, . humidity, and other values are consistent with the observed wa~er temperatures. The drawback of the historical analogy approach is that the factors that caused the warm period to occur may differ from the driving factors of the greenhouse warming. Consequently, the observations during an historical warm anomoly may not be representative of the changes that may occur as the climate system is driven by increasing concentrations of greenhouse gases. Nevertheless, the historical analogy approach is a tool that can be used in climate change impact analyses. in particular as a contrast to scenarios based on model estimates. However, developing scenarios based on historical data was beyond the scope of this study. To assess the potential for greenhouse-gas-induced climate change, the results of climate change experiments from the three major three-dimensional (3-0) general circulation models (GCMs) of the atmosphere were examined: Goddard Institute for Space Studies, Principal investigator Dr. James Hansen;9 ~. NASA, National Center For Atmospheric Research, NOAA, Principal Investigator Dr. Warren Washington;lO and ~. Geophysical Fluid Dynamics Laboratory, Principal Investigator Dr. Syukuro Manabe. l l ~. NOAA, These models represent the climate of the entire globe, and (as described below) perform their analysis at various levels of geographic resolution. The results of these three models were used to identify the ma&nitude of the potential change in temperature that may result from a doubling or quadrupling of the atmospheric concentration of carbon dioxide. The results of similar experiments from each of the three models were made available by the Principal Investigators. The experiments were generally of the type: 9 Hansen, J., et al., "Climate Sensitivity: Analysis of Feedback Mechanisms," in Climate Processes and Climate Sensitivity, J. Hansen, ed., American Geophysical Union, Washington, D.C., 1984, pp. 130-163. 10 Washington, W' .M., and D.L. Williamson, itA Description of the NCAR Global Circulation Models in Methods of Computational Physics," General Circulation Models of the Atmosphere, J. Change, ed., Academic Press: New York, 1977. 11 Kanabe, S. and R.J. Wetherald, "The Effects of Doubling the C02 Concentration on the Climate of a General Circulation Kodel," Journal of Atmospheric Sciences, 32:3-15, 1975. 2 - 3 From the digital collections of the New York State Library. 1. Run the model with the current composition of the atmosphere and estimate climate statistics such as average annual temperature. This is the baseline XYD. 2. Run the model with a lDodified composition of the atmosphere (e.g., double the concentration of carbon dioxide) and estimate climate statistics once the climate system has reached an equilibrium state. This is the perturbation run or the doubled C02 ,nm. 3. The differences in the climate statistics from the baseline run and the doubled C02 run are the estimates of the magnitude of the potential changes in climate that may result from increases in greenhouse gases. For doubled C02, the GISS and NCAR model results indicate a potential increase in annual global temperature of about 4. 2°C (7. 6oF) and 3. SoC (6.8 0F) respectively. For a quadrupling of C02, the GFDL results indicated an increase of about 4°C (7.2 oF). (GFDL results for a doubling of C02 were not available for this study.) Although the estimates of global annual temp~rature change are similar across the three models, the seasonal and regional changes are very different. Consequently, there is considerable diversity among the different model results. The GCM model results are limited for purposes of developing climate change scenarios for the period 1986 to 2015 because they represent what the climate may be like when the composition of the atmosphere has chanced significantly, and when the climate· system has come into equilibrium with the new atmospheric composition. First, the atmospheric changes examined with the models (e.g., a doubling of C02) are not expected to be reached by 2015. Second, the climate system will not be in equilibrium as the composition of the atmosphere changes slowly over time. The amount of climate change that will be observed (i.e., that will be experienced over time) will behind the amount of change that would be expected under equilibrium conditions. 12 Therefore, the climate changes indicated by the GeM experiments are larler than the changes that may be expected by 2015 for two reasons: (1) the atmosphere is not expected to change as significantly by 2015 as the assumptions used in the GCM experiments; and (2) the observed climate change is expected to be less than the change indicated by the equilibrium results. .w To use the GCM estimates, a separate analysis of the potential future ~ of clt.ate change was performed using a parameterized representation of a one-dime~lonal (1-0) radiative-convective model of the atmosphere and oceans developed by Hansen (in 19S1) , and subsequently modified by Lacis (in 1981), Hoffman, Keyes, and Titus (in 1983) and Hoffman, Wells, and Titus (in 12 The observed climate change will lag behind the expected equilibrium change because the oceans are slow to absorb heat. See for example, Hansen, J. et al., "Climate Response Times: Dependence on Climate Sensitivity and Ocean Mixing," Science, .30 August 1985, pp. 857-859. 2 - 4 From the digital collections of the New York State Library. 1986) .13 This 1-D model evaluates global warming over time by estimating the equilibrium warming expected due to the changing radiative characteristics of the atmosphere, and computing the rate at which heat may be transferred into the oceans. The model therefor. provides an indication of the potential rate of warming that may b. expected over time in relation to the potential equilibrium warming that.1s estimated. 14 To evaluate the potential rate of climate change using the l·D model, assumptions were required regarding the rate of change of the composition of the atmosphere and the rate at which heat lIay be transferred' into the oceans. To describe the changing composition of the atmosphere, two scenarios of potential future trace ga. concentrations (low and high) were taken from Hoffman, Wells, and Titus, which summarize. work by Ramanathan and Quinn. 15 The trace gases included in these scenarios are: C02; CH4; N20; CFC-1l; CFC-12; 502; CF4; C2F6; CH2C12; CeLF3; CHC13; C2H2, CFC-22; CC14; CH3CCL3; and Halon 1301. The range of assumptions about the•• trace gases result ina changed composition of the atmosphere that is equivalent to a climate forcing of apprOXimately a doubled concentration of C02 by 2025 in the high case and by 2055 in the low case. The rate at which heat may be transferred to the oceans can be described using an average eddy diffusion coefficient. Using tracer., Broecker and Peng estimated an average value for the oceaa. of 1.7 cm2/sec. 16 As a range of assumptions, one-half this value (0.85 cm2/sec) and twice this value (3.4 cm2/sec) are used as low and high assumptions. Using the 1-D model, the two trace gas scenarios and the three diffusion coefficient assumptions were used to develop six estimates of the rate of global w~rming for each of the three GeM climate sensitivities (4.2 oC, 3.8 o C , and 2°C). The simulated global warming for each of the scenarios is presented in Table 2-1. The following factors are seen in the table: 13 Lacis, A., et al., "Greenhouse Effect of Trace Gases," GeophYsical Research Letters, 8:1035-1038, 1981. Hoffman, J.S., D. Keyes, ana J.G. Titus, Projectins Fyture5ea Level Rise: Methodology, Estimates to the Year 2100, and Research Needs, U.5. EPA, Washington, D.C., 1983. Hoffman, J .S., J .B. Wells, and J .G. Titus, "Future Global Warming and Sea Level Rise," U.S. EPA and The Bruce Company, 1986. 14 Because the 1-D model provides average results for the entire globe, it does not indicate how the climate may change in particular parts of the globe. 15 Hoffman, Wells, and Titus, 2R. ~ Work by Ramanathan reported in Ramanathan, V. et a1., "Trace Gas Trends and Their Potential Role in Climate Change," Journal of Geopbys iea1 Research (in press). The work by Quinn is reported in: Quinn, T.H., et al., Potential Use, Emissions and Banks of Potential Ozone Depleting Substances, The RAND C.orporation, Washington, D.C. 1984. 16 Broecker, W.S., and T. Peng, Tracers in the Sea, Palisades, N.Y.: Lamont-Doherty Geophysical Observatory, Columbia University, 1982. 2 - 5 From the digital collections of the New York State Library. TABLE 2-1 RESULTS OF I-DIKENSIONAL KODEL SHOVING GLOBAL VARKIltG ESTIMATES UNDER ALTERNATIVE ASSllKPTIONS (S~1at~d Climate Sensitivity AI Global Wa~ing from 1986 to 2015 in °C) Diffusion Coefficient hi Trace Case Scenarios £/____. Low High 4.2 0C 0.85 cm2/sec 0.63 0C 0.94 0C 4.2 0C 1.70 cm2/sec 0.53 0C 0.76 0C 4.2 0C 3.40 cm2/sec . 0.48 0C 0.66°C 3.8 0 C 0.85 cm2;sec 0.60 0 C 0.86 0C 3.8 0C 1.70 cm2/sec 0.55 0 C 0.74 0C 3.8 0C 3.40 cm2/sec 0.47 0 C 0.62 0C 2.0 0C 0.85 cm2;sec 0.42 oC O.60 o C 2.0 0C 1.70 cm2/sec 0.38 0 C 0.55 0C 2.0 0C 3.40 cm2;sec 0.35 0C 0.49 0C sf Climate sensitivity is defined as the equilibrium global warming anticipated in response to a doubling of atmospheric C02 concentrations. Range based on GeM results. hi The diffusion coefficient indicates how rapidly heat may be transferred to the oceans. A range of O. 5 to 2.0 times an average value of 1. 7 cm2 / s ec is assumed. £/ The range of trace gas scenarios results in an effective radiative forcing of doubled C02 by 2025 in, the high case and by 2055 in the low case. 2 - 6 From the digital collections of the New York State Library. Greater climate sensitivity results in larger estimates of warming. However, doubling the sensitivity does not result in a doubling of the estimate of warming for this time period. Smaller diffusion coefficients result in larger estimates of warming. The trace gas scenarios have a large influence on the estimates of global wa~ing in this period. For a given climate sensitivity, this assumption has a larger impact than does the diffusion coefficient assumption. Overall, for a given climate sensitivity, the lowest value estimated (diffusion coefficient - 3.4 cm 2/sec, low trace gas scenario) is about 55 percent of the highest value estimated (diffusion coefficient - 0.85 cm 2/sec, high trace scenario). To develop climate change scenarios, the rates of climate change in Table 2-1 can be used to scale the equilibrium estimates from the three GeMs to estimate potential changes in temp~rature over time. Using these rates in this manner implicitly assumes: (1) that global annual average temperatures will change smoothly over time in response to gradual changes in atmospheric composition; and (2) that seasonal and regional temperatures will change over time at the same rate as the global annual average temperature. In fact, it has been hypothesized that the rate of warming may not be smooth, and may not be uniform across the globe. l 7 This potential phenomenon would not be reflected in scenarios developed by scaling the GCK results with the 1-D model results in Table 2-1. Only one GCM has been used to evaluate how the climate may change over time as the atmosphere changes. 1S The results of this experiment (referred to as a "transient run") with the GISS model were made available for this study and are compared below to the results obtained by scaling the equilibrium results of the three GCMs using the results of the 1-D model. The transient run results show the potential for the rate of climate change to vary by season and location. Additional transient runs with GCMs are expected in the future. In the case study of the New York State utilities, the potential impact of climate change on stream. flow for hydro-electric power generation was examined. The results of a water balance model of the entire Great Lakes watershed developed by Cohen were used. 19 This model evaluates potential 17 See for example Broecker, W.S., D.M. Peteet and D. Rind, "Does the Ocean-Atmosphere System Have More Than One Stable Mode of Operation?" Nature, Vol. 315, May 1985, pp. 21-25. 18 Hansen, J., et al., "The Greenhouse Effect: Projections of Global Climate Change, It in Effects of Chances in Stratospheric Ozone and Global Climate, J .G. Titus ed. ,U.S. EPA and UNEP, Washington, D.C., August 1986, p. 205. 19 Cohen, Stewart J., "Impacts of C02-induced Climatic Change on Water Resources in the Great Lakes Basin," Climatic Chanle, 1985. 2 - 7 From the digital collections of the New York State Library. changes in Net Basin Supply (NBS), the amount of water flowing out of the Great Lakes Basin. NBS includes the flow of water in the St. Lawrence River, the site of one of New York's hydro-electric generating stations. Historical data indicate that the flow in the Niagara River (the site of New York's largest hydro-electric generating station) is well correlated with the flow in the St. Lawrence River. Therefore, the change in NBS estimated from the water balance model can also be applied to the Niagara River. Like the GeMs, this water balance model produces equilibrium results. The temperature change scenarios developed based on the GeMs and the 1-0 model were used to scale the equilibrium water balance results to create scenarios of stream flow over time. The following two sections present the temperature and stream flow scenarios, and the data upon which they are based. TmlPERATURE CHARGE SCERARIOS Scenarios of potential changes in average seasonal temperatures were developed for the New York State and Southeastern case studies using the 1-0 model, the GCM equilibrium results, and the GISS transient run. Equilibrium temperature change estimates for each of the 12 months of the year from the GISS and GFDL models were averaged to create seasonal estimates of temperature change. The NCAR results were supplied to us as seasonal values. The geographic resolution of the models is displayed in Figure 2-1. The GISS model results were supplied for an SO latitude x 100 longitude grid. The values for New York State are the averages of the values for grid numbers 5 and 6. The NCAR and GFDL results were supplied for a 4° latitude x 5° longitude grid. Grid number 17 was used for the New York State scenarios. Single grid squares cover the Southeastern utility area frOID each model. Table 2-2 lists the equilibrium seasonal temperature change results for doubled C02 for the grid squares used to develop the scenarios (the 4xC02 values from the GFDL model have been divided by 2). As shown in the Table 2- 2 , the potential change in temperature varies across the three models' results. 20 The equilibrium values in Table 2-2 can be scaled to estimate changes The scaling is accomplished by multiplying the GCM equilibrium value by the ratio of the warming estimate for 2015, to the assumed climate sensitivity. For example, to get the highest winter GISS value for New York State, the 5.l oC (Ta~le 2-2) is multiplied by 0.94 oC (from the high trace gas scenario and a diffusion coefficient of 0.85 cm2/sec) and divided by 4.2 oC (the climate sensitivity for GISS). The result is 5.1 x .94/4.2 - 1.1 oC. This procedure will yield 18 (6 times 3) scenarios of changes in temperature for each season. Using the rates from the leD model with the equilibrium results for each season, implicitly assumes that the seasonal changes will occur uniformly at the same rate as the global annual average changes. by 2015, using the leD model results presented in Table 2-1. 20 It should be noted that subsequent modifications to the GFDL GCM have yielded higher estimates of changes associated with doubled C02 than are presented here. 2 - 8 From the digital collections of the New York State Library. fiGUU 2-1 GIOGRAPIIIC USOLDTIOlI OF mB GQfa OVER TIIB URITID STAra GISS NCAR/GFDL 2 - 9 From the digital collections of the New York State Library. TABLE 2-2 ESTIMATES OF EQUILIBRIUK TEMPERATURE CHANGE DUE TO DOUBLED C02 (oC) New York State Season ~/ Southeastern Utility g/ GISS NCAR GFDL GISS NCAR GFDL Winter 5.1 3.6 4.5 2.5 2.2 2.0 Spring 4.2 2.6 2.6 3.5 3.7 1.9 Sununer 3.6 2.0 2.4 3.3 3.3 1.7 Fall 3.7 3.6 3.8 3.2 3.0 1.6 §../ Values reported here are the 4xC02 values divided by 2. 2 - 10 From the digital collections of the New York State Library. rn addition to the scenarios that can be developed from the leD model and the equilibrium GeM results, a scenario was developed from the GISS transient run. The results provided from the transient run experiment were estimates of seasonal temperatures each year for the years 1958 to 2019. The data displayed considerable interannual variability, so II-year running averages were computed. 21 The II-year period prior to and including 1985 was adopted as the baseline period, and changes (in the ll-year running average) from that period were estimated for each season. These changes in seasonal temperatures form a scenario based on the transient run. Table 2- 3 displays seven scenarios of temperature change for each season for 2015 (note that estimates can be performed for each year). For each of the equilibriUm GeM experiments, the lowest and highest estimates based on the 1-0 scaling are presented. For example, for winter in New York State the GISS-based estimate ranges from .58 oC to 1.10C. Similar ranges are shown for the other models, for each of the seasons. The GISS Transient Run is the seventh scenario shown in Table 2-3. In the New York State estimates, the GISS Transient Run values are highest for two of the four seasons. In the Southeastern Utility estimates, the GISS Transient Run values are highest for three of the four seasons. Despite the diversity of scaling assumptions applied to the equilibrium GeM results, ·the GISS Transient Run generally shows larger potential climate changes by 2015. The larger estimates from the transient run occur even though its rate of trace gas growth is s lower than the high trace gas scenario used for the scaling (the transient run has an effective doubling of C02 by about 2030, as opposed to about 2025 in the high trace gas scenario). In assessing the potential impacts of climate change on electric utilities, various sets of scenarios could be constructed based on the estimates presented in Tables 2-1 and 2-2. Ranges of scenarios from each of the GeM equilibrium experiments can be developed, the bounds of which are presented in Table 2-3. For purposes of assessing whether climate change will have an impact on electric utilities, a "highest" scenario has been developed. This scenario, based on the highest estimate for each season presented in Table 2-3, is summarized in Table 2-4. By taking the largest estimate for each season (most of which come from the Transient Run estimates), the largest temperature increases currently indicated are assessed together. The drawback of this approach is that the resulting scenario may not be internally consistent because it mixes model results and underlying assumptions. Although this scenario is constituted as a "high" case, limitations in the models do not preclude the possibility that climate changes will in fact be larger than the values used here. 21 The choice of the averaging period has a small influence on the estimates of changes in temperature. The period of 11 years was chosen as the longest period that the available data would support, while providing an estimate out to 2015. 2 - 11 From the digital collections of the New York State Library. TABLE 2-3 TEKPDATUU CHANGE ESTIllATES BY SBASOR BASED OIfmE GQIa, mE TllABSIBRT R.UB, AlII) mE I-D IIODIL (OC by the Year 2015) Hey YOJ:k State GISS Low III High l2./ Low High GFDL Low H1gh NCAR GISS Transient Run £I Winter 0.58 1.10 0.44 0.81 0.79 1.30 1.30 Spring 0.48 0.94 0.32 0.59 0.46 0.78 0.81 Summer 0.40 0.78 0.25 0.45 0.42 0.72 0.67 Fall 0.42 0.83 0.45 0.81 0.67 1.10 1.10 !I Rate of climate change based on the following assumptions used in a 1-0 model: trace gas growth - low; diffusion coefficient - 3.4 cm 2/sec. hi Rate of climate change based on the following assumption used in a 1-0 model: sf trace gas growth - high; diffusion coefficient - 0.85 cm 2/sec. Computed using 11-year running average. 2 - 12 From the digital collections of the New York State Library. 'lABLB 2-3 (CoDt:1nuecl) TFJIPERATOIlE CHARGB ESTIlIATES BYSIASOR BASED OR TIIB Gals. TIIB TBABSIaT R1JB. AlII) TIIB 1-D IIODBL (OC by the Year 2015) Southeastern Utility GISS Low AI High W NCAR Low High GFDL Low High GISS Transient Run s/ Winter 0.28 0.56 0.27 0.50 0.35 0.60 1.10 Spring 0.40 0.78 0.46 0.84 0.33 0.57 0.70 Summer 0.38 0.74 0.41 0.75 0.30 0.51 1.00 Fall 0.37 0.72 0.37 0.68 0.28 0.48 0.79 AI Rate of climate change based on the following assumptions used in a 1-D model: trace gas growth - low; diffusion coefficient - 3.4 cm2/sec. W Rate of climate change based on the following assumption used in a 1-D model: trace gas growth - high; diffusion coefficient - 0.85 cm2/sec. £I Computed using 11-year running average. 2 - 13 From the digital collections of the New York State Library. TABLE 2-4 HIGH ~ERATURE CHANGE SCENARIO USED TO EVALUATE UTILITY IMPACTS Hew York Case Study Season Temperature Change by 2015 Basis (oC) Winter 1.3 Transient Run and GFDL-Based High Estimate Spring 0.94 GISS-Based High Estimate Summer 0.81 GISS-Based High Estimate Fall 1.1 Transient Estimate Run and GFDL-Based High Southeastern Utility Season Temperature Change by 2015 Basis (oC) Winter 1.1 Transient Run Spring 0.84 NCAR-Based High Estimate Summer 1.0 Transient Run Fall 0.79 Transient Run 2 . . 14 From the digital collections of the New York State Library. As a contrast to this high scenario, two lower scenarios were developed. The GCM results were scaled using a subset of the climate sensitivity, trace gas, and diffusion coefficient assumptions described above. 22 The two scenarios shown in Table 2-5 represent the lowest and the median estimates produced by these assumptions. These three scenarios from Tables 2-4 and 2-5, were used to evaluate potential impacts on electric utilities. Because of the uncertainties in the models employed to create the scenarios, it is possible that actual climate changes may fall outside the range presented here. Therefore, the scenarios are not referred to as "low, middle, and high." Instead, the labels "A, B, and CIf are employed. In using these climate change scenarios to evaluate impacts in the case studies, it is implicitly assumed that the distributions of daily, monthly, and seasonal temperatures all change over time by the amounts shown in Tables 2-4 and 2-5. Additional research is required to assess whether the shapes of the distributions may also change, for example, indicating greater (or less) variability.. To the extent that impacts are driven by extreme events, the omission of an assessment of the potential change in variability is an important limitation of the scenarios. ~ STlUWI FIDV SCERAlUOS Changes in climate due to increased atmospheric concentrations of greenhouse gases are expected to have significant impacts on water resources around the world. Not only are temperatures expected to increase, but changes in all aspects of the hydrologic cycle are anticipated, including changes in precipitation, evaporation, runoff, soil moisture, and evapotranspiration. These changes in the hydrologic cycle may influence the availab.ility and quality of water for virtually all of its uses. Several studies have examined the potential impacts of changing climate on water resources. 23 These studies) based on GCM estimates of climate change, all indicate that a potential reduction in overall water availability is likely in the Great Lakes Region of the United States and Canada, despite projected increases in precipitation. The increases in evaporation driven by higher temperatures appear to dominate the hydrologic response, although one 22 The high diffusion coefficient was used with .the low trace gas scenarios and the low climate sensitivity. These assumptions maximize the differences between the leD-simulated global warming estimates. 23 See for example: Cohen, ~ ~ Rind, D. and S. Lebedeff, "Potential Climate Impacts of Increasing Atmospheric C02 with Emphasis on Water Availability and Hydrology in the United States," NASA Goodard Space Flight Center Institute for Space Studies, prepared for the U.s. EPA, Washington, D.C., April 1984. Rind, D., "The Doubled C02 Climate and Future Water Availability in the u. S. ," NASA Goddard Space Flight Center Institute for Space .Studies, New York, New York, in press. 2 - 15 From the digital collections of the New York State Library. TABLE 2-5 TEMPERATURE CHANGE SCENARIOS USED to cotrrRAST VIm HIGH SCENARIO (TEMPERATURE CHANGE BY 2015. °e) Rev York· Case Study Season Scenarig A Scenario B Winter Spring Swmner 0.64 0.46 0.36 0.65 0.92 0.59 0.52 0.76 Season Scenario A Scenario B Winter Spring Summer 0.35 0.35 0.31 0.30 0.45 0.69 0.64 0.60 Fall Southeastern Utility Fall 2 - 16 From the digital collections of the New York State Library. study24 found that projected changes in average wind conditions could be equally important. Evaluations of the potential for climate change to influence water resources remain very crude. The processes that drive water quality and availability operate on small scales (such as drainage basins and watersheds) and are not well-modeled in the GeMs. The current· assessments of the impacts of climate change on water resources must be viewed as very preliminary, as considerable research remains to be performed. 25 To evaluate the potential implications of climate change on hydro-electric generation in the New York case study, a scenario of changes in stream flow was developed based on the results of a water balance model for the Great Lakes Region. 26 The water balance model is empirical, and provides estimates of potential and actual evapotranspiration, water surplus (or deficit), and runoff based on temperature, precipitation, and soil moisture storage capacity . inputs. 27 This water balance approach to modeling has been found to perform well in climates like the one found in the Great Lakes Region. 28 Table 2-6 displays the results of the water balance model for five scenarios of doubled-C02 GeM results from GISS and GFDL. All scenarios assume no change in the consumptive use of water as th. result of climate change,29 and despite increases in precipitation in sOlDe areas in SOll8 cases, all five scenarios examined show significant reductions in NBS relative to the long-term flow average (based on 1959 to 1982). These estimates are based on the equilibriUII doubled-C02 results from the GCKs and do not represent the changes that may be observed over the next 30 years. Similar to the temperature change scenarios discussed above, the equilibrium stream flow scenarios must be scaled to the next 30 years. 24 Coh en, 2R. .£&&. ~.~ 25 Research is required not only on how climate variables such as precipitation and evaporation may change at relevant scales, but also on the potential feedbacks between climate change and the demand for water for human activities (such as irrigation water and cooling water) and the demand for water by natural vegetation (which will change in response to climate change). See: Characterization of Information Requirements for Studies of C02 Effects: Water Resources. Airiculture. Fisheries. Forests. and Human Health, M.R. White ed., U.S. Department of Energy, DOE/ER-0236, Washington, D.C., December 1985. 26 Cohen, 2R. ill. 27 .llWI. 28 llli. 29 The consumptive use of water in the region is expected to grow from 170 ems in 1985 to 720 cms by 2035, for growth of about 2.9 percent per year. This growth in use is included in the analysis of these scenarios. Chanaes in this growth as a consequ~nce of climate change are not considered. 2 - 17 From the digital collections of the New York State Library. TABLE 2-6 CHANGE IN NET BASIN SUPPLY FOR FIVE SCERAlUOS OF Gal RESULTS Change in ~ Net Basin Supply Scenario 1. GISS 2xC02 -28.9' 2. GISS 2xC02 with Wind Speed at 80' of current values QI -11.8' 3. GFDL 2xC02 -26.4' 4. GFDL 2xC02 with Wind Speed at 80% of current values Qj -11.7' 5. GFDL 2xC02 with GFDL-estimated wind conditions s/ sf Percent change from long-term average flow conditions (1959-1982 period). Q/ Reduced wind speeds examined based on the hypothesis that greater warming at the poles than the equator could lead to reduced temperature differentials, and consequently reduced winds. s/ Uses wind speeds computed by the GFDL model. Source: Cohen, Stewart J., "Impacts of C02-Induced Climate Change on Water Resources in the Great Lakes Basin," Climate Chan&e, 1985. Note: Values for GFDL were computed by Cohen by dividing the 4xC02 estimates by 2. . 2 - 18 From the digital collections of the New York State Library. Because the stream flow changes are expected to be driven primarily by changes in temperature, the temperature change scenarios developed above are used to scale the equilibrium NBS results. The following steps were performed for each year for the five NBS scenarios: 1. Compute a scaling factor by dividing the annual temperature change value for the scenario by the equilibrium annual temperature change value from the GeM. 2. Compute the change in evaporation, precipitation, and runoff that is realized this year by multiplying the scaling factor from step 1 by the equilibrium values for evaporation, precipitation and runoff. 3. Compute the change in consumptive uses of water by interpolating between the current use and the projected future use in 2035. 4. Sum the changes in from steps 2 and 3 to compute the change in NBS in this year. The results from the five NBS scenarios were then averaged to create a representative stream flow scenario. By 2015, this method implies a 7 percent reduction in stream flow based on the New York State temperature scenario presented in Table 2-4 (i.e., scenario C). Temperature scenarios A and B imply smaller reductions in stream flow, 5.1 percent and 5.8 percent respectively. As with the temperature scenarios, the actual stream flow response to climate change may fall outside the range identified here. Therefore, the chree scenarios are labelled X, Y, and Z. The largest reduction, scenario Z, is used in the New York Case Study to assess the potential complications of stream flow reductions. The approach used to develop the stream flow scenarios implicitly assumes that all the key climate variables will change over time at the same rate as temperature. In fact, precipitation patterns may change faster or slower than temperature, and may not change smoothly over time. Similarly, changes in the demand for water due to climate change that are not considered here could influence the stream flow. The stream flow scenarios developed here should be considered as illustrative of the magnitude of estimates of the changes that have been developed to date for the Great Lakes Region. This scenario is useful for assessing the potential implications of these changes for hydro-electric power generation as described in Section 5, but should not be considered as forecasts. 2 - 19 From the digital collections of the New York State Library. From the digital collections of the New York State Library. SECTIOR 3 UTILITY IMPACT METHODOLOGY OVERVIEW Estimates of future values for several characteristics which describe the demand for and supply of electricity are important considerations for utilities in planning future capacity and operations. These characteristics --referred to as utility planning factors--include: electricity sales; annual and seasonal peak demand; generating capacity requirements; fuel utilization by type of fuel; and capital and operating costs. l Our approach is to develop and compare estimates of future values of these utility planning factors for each case study utility, under alternative climate change scenarios. The changes in these planning factors are indications of the potential impacts of climate change on the case study utility. Five tasks were performed to evaluate the potential impacts of climate change on these planning factors given the climate change scenarios described in Section 2. These five tasks are illustrated in Table 3-1. The tasks are to: 1. 2. 3. 4. 5. develop develop develop develop compare base case utility planning scenarios; weather-sensitivity of demand; weather-sensitivity of supply; alternative planning scenario outputs; and base case and alternative case scenario outputs. The specific steps to be undertaken to complete these tasks are described in this section. The data inputs and applications of the methodology to the case study utilities are described in Sections 4 and 5. TASlC 1: DEVELOP BASE CASE UTILITY PIAR Given (1) characteristics of current generating capacity and operating costs as well as (2) performance and cost characteristics of future investment options and operating strategies, a principal objective of utility planning is to meet future generation requirements (i.e., demand for 1 A more complete description of these utility planning factors are presented in Appendix A. 3 - 1 From the digital collections of the New York State Library. Table 3-1 SUMMARY Of ANALYTICAL TASKS Inputs Methods Develop 885e Case Utility '.anning SCenarios o Utility de.and forecasts o Utility supply characteri st iC6 o Utility planning assu.ptions o Seasona I load durat ion curves o Icr Planning Model o Sales o Capacity o Costs Develop Weather-sensitivity 0 Utility forecasts/load research studios o De.and and weather dala o [cono.ic. deltOgraphic. appliance saturation data o Statistical analysis: De.and = f(weather, other para.eters) o Structural analysi6: De_and = Appliance stock x Utilization per appl ianee o Weather ela5ticjties of dCliand o AdjustEtd de. . nds or Supply Develop Weather-Sensitivity o Utility studies o Engineering studies o Hydro generation o Statistical analysis: hydro generation = rCwater supply) Engineering relationships: Power plant efficiency = f( tellps ra ture) o Weather' e last Ic it ies of supply o Adjusted supply characteri st iC6 Develop AILernative Planning SCenario Outputs o Alt. cli.ate scenarios o Weather elasticities o seasonal load duration curves o Icr Planning Model Task 2 3 or De..nd \.#oJ N _ o AIL. utility behavioral assu.ption5 Outouts o o o Sales Capacity Costs Peak de.and Fuel 0 0 0 0 Peak de.allld fue' (anticipation of cli.ale change) o Alt. utility de. .nd forecasts ~ eo.pare Base Casel A' t.ernat i ve Case Scena rio OUtputs o Outputs of Task 1 o Outputs of Task 4 From the digital collections of the New York State Library. o ASales o ~apaclty o ~osts o APeak o 6Fuet .. nd oe~ electricity) at minimum cost. Generally, utilities prepare plans to meet this objective assuming future climatic conditions are the same as or similar to those that have occurred in the recent past. That is, utilities plan assuming continuation of "typical weather conditions," and do not explicitly consider the possibility of climate change. 2 Because our focus is on estimating how climate change may affect key utility planning factors, plans developed by the case study utilities (assuming no change in climate) can serve as a basis for comparison in Task 5, with plans developed by reF under alternative assumptions of climate change. In making these comparisons, absolute values of estimated future temperatures and other climate variables are of secondary importance; rather, the assumed direction and magnitude of changes in these climate variables from "typical" values are of primary interest. To implement this approach, a first task is to characterize the case study utilities' investment and operating plans. These plans are referred to as the "base case" planning scenarios. The six steps to accomplish Task 1 are illustrated in Figure 3 -1. Base case scenarios are developed from data and studies (demand forecasts, supply planning analyses) prepared by the case study utilities. Utility forecasts of energy sales (kwh) and peak demand (kw) are direct inputs to the analysis (Step 1.1). Typically, utilities prepare demand forecasts which cover a 10-20 year planning horizon. Because our horizon is 30 years (1986-2015), the utility demand forecasts must be extended for the out-years. This is accomplished through application of straightforward extrapolation methods (e.g., apply the average annual growth rate in demand calculated from the last five years of the utility forecast). Utility loss multipliers are applied to the forecasts of energy sales and peak demand to account for transmission and distribution (T&D) losses (Step 1.2). These multipliers are available from utility filings with regulatory agencies and directly from utility planning studies. Together electricity demands and T&D losses define total generation requirements (Step 1.3). The level and pattern of these requirements are important for determining (1) the amount of capacity and (2) the types of capacity which constitute the utility's optimal (i.e., least cost) investment plan. A device for summarizing the characteristics of total generation requirements is a load duration curve (LDC). An LDC is an ordering of hourly generation values from highest demand to lowest demand over a period of time (month, season, or year). An example of an LDC is presented in Figure 3-2. LDC's often are used to represent customer demands and total generation requirements in models which optimize utility investments in capacity additions and optimize dispatch (utilization) of capacity. In this study, t 2 Exceptions are utilities for which hydropower is an important source of electricity. These utilities often conduct extensive analyses of the potential effects of changes in rainfall or run off on the availability of hydropower. However, these analyses typically consider variability around the existing means of these weather variables. 3 - 3 From the digital collections of the New York State Library. FIGURE 3-1 STEPS IN TASK 1: DIVELOPIIIIT OF BASE CASI SCIllARIO 1.1. Base Sales and Peak Demand 1. 3. 1.2. Base Transmission and Distribution Losses 1.4. Total Generation Total Sources Requiremen~s (LDC) 1.6. Base Case Results: - Sales - Peak - Fuel - Capacity - Costs 3 - 4 From the digital collections of the New York State Library. FIGUU 3·2 Demaad (MW) .. .. .. .........•... ••...•.••. . :::~ .. ":':':':.:.:':' ...••..•........... Hours 3 5 From the digital collections of the New York State Library. LOCs for four seasons per year are used to summarize generation requirements under base case (Task 1) and alternative case (Task 4) conditions. Base case LDCs are developed directly from case study utility forecasts. Similarly, case study utility data are used to characterize the investment options available to the utility (Step 1.4) and its base case supply plan (i.e., its choices from the .investment options available, Step 1.5). Tne investment options include technologies which are widely utilized currently (e.g., large thermal plants), as well as technologies which are not currently widely utilized, but in the utilities' estimation have potential for significant future utilization (e. g. , integrated gasification/combined cycle units, fuel cells). The utility planning data base includes, by type of generating unit: capacity (kw), potential generation (kwh), operating and maintenance costs, fuel costs ($/BTU), heat rates (BTU/kwh), and outage rates (the expected amount of time when the plant is unavailable for generation). This data base . is summarized for each forecast year by season. The information changes over time as older units are retired from service and new units are constructed, and as costs of fuel change. For hydro units, the information varies within a year by season because of different seasonal water flows. Operating parameters for other types of units vary by season due to such differences as weather and expected load. Step 1. 6 provides a summary of the characteristics of the base case planning scenario. The characteristics include selected seasonal and annual values over the forecast horizon of utility planning factors estimated using the case study utility data. Principal planning factors are: electricity sales; peak demand; generating capacity additions; fuel use by type; and electricity production costs (capital, fuel, and O&M). TASK 2: DEVElDP VEAl'HER. SENSITIVITY OF DIHARD The purposes of this task are (1) to develop quantitative relationships between changes in weather conditions (i.e., temperature) and changes in utility sales and peak demand and (2) to apply these relationships using temperature data from the climate change scenarios to develop adjusted sales and peak demand estimates and adjusted total electricity requirements. The steps in this task are illustrated in Figure 3-3. The first step (Step 2.1) is to develop the weather sensitivity relationships. Two basic approaches can be applied in this step depending upon data availability. The first approach develops the weather sensitivity relationships using temperature data and aggregate electricity demand data. Statistical relationships are developed with regression analysis techniques applied to historical temperature and utility system load data. 3 - 6 From the digital collections of the New York State Library. FIGURE 3-3 STEPS IN TASK 2: DEVELOPMENT Of WEATHER SENSITIVITY OF DEMAND 2.1 WEATHER SENSITIVITY RELATIONSHIPS CHANGES IN WEATHER (FROM CLIMATE CHANGE SCENARIOS) BASE SALES AND PEAK DEMAND (FROM TASK 1) 2.2 ADJUSTED SALES AND PEAK DEMAND T&D LOSSES (FROM TASK 1) 2.3 ADJUSTED TOTAL REQUIREMENTS (LDe) 3 - 7 From the digital collections of the New York State Library. For each month, type of day (weekday or weekend day), and tri-hourly time block, regressions (ordinary least squares) are run using the foliowing model specification: % deviation of load from mean (deviation of (deviation of &1 heating degree + &2 cooling degree hours) hours) and data for a recent period of 5-7 years. This estimation procedure (1) takes advantage of tri-hourly weather observations available from NOAA and (2) is designed to minimize the influence of changing economic, electricity price, and other factors affecting the historical data on the estimated coefficients. The estimated coefficients, Al and &2, represent the weather sensitivity of demand. This first approach is designated the "statistical approach" and is applied -- with some modifications described in Sections 4 and 5 -- to the Southeastern and the New York State case studies. An important point is that because historical data are used to estimate the statistical relationships, the coefficients represent historical equipment saturations, energy efficiency, and utilization of weather-sensitive appliances and equipment. In addition, because the utility system load data ~nlcude T&D losses as well as electricity delivered to customers, the coefficients represent the weather-sensitivity of utility total requirements, and no explicit adjustments need to be made to account for T&D losses. The second approach -- designated the "structural approach" -- operates on disaggregated, utility end-use data. The structural approach is illustrated in Figure 3-4. Here, the saturation and utilization of individual types of appliances or other customer end-uses are analyzed and the end-use loads are aggregated to the customer class or utility system load levels. Adjustments for T&D losses are made explicitly. Weather sensitivity relationships are estimated using statistical techniques and engineering simulation models for end-uses such as air conditioning and electric heat. The statistical estimation procedures in this case are similar to those described above, but are applied to end-use demand data. The principal advantage of the structural approach is that explicit assumptions can be made regarding the market saturation of weather-sensitive appliances over time, and the implications of these assumptions for utility energy requirements can be estimated. Because of limited data availability for the Southeastern utility, the structural approach is applied only in the New York case study.3 Once the weather sensitivity relationships are developed using -either the statistical or the structural approach, Steps 2.2 and 2.3 are straightforward. The changes in average seasonal temperature developed in the climate change 3 A structural load shape analysis and forecasting model -- the Hourly Electric Load Model (HELM) - - developed by ICF for the Electric Power Research Institute -- is used to implement the structural approach in the New York case study analyses. 3 - 8 From the digital collections of the New York State Library. nGQRI 3-4 - .....,.... APPIOAQI to ISTIIIATIS UTILITY DDIABDS QassUJads Systtm Load M'•••la,••• MW c .. ........... Houn 01 tb. Da, 3 - 9 From the digital collections of the New York State Library. scenarios are assQmed to apply to all tri-hourly time periods and day types within that season. (No information is available to characterize the pattern of temperature changes in a finer level of detail.) The changes in temperature are entered into the weather sensitivity relationships and the estimated changes in demand are calculated. In the case of the structural approach, adjustments are made for T&D losses through use of a loss multiplier prpvided by the case study utilities to estimate adjusted total requirements. 4 TASK 3: DEVELOP VEAl1IER. SENSITIVITY OF SUPPLY The purpose of this task is to develop a set of quantitative relationships between utility operational characteristics and changes in weather conditions. Figure 3-5 illustrates the steps to accomplish this task. The first step (Step 3.1) is to develop a relationship between hydro generation availability and changes in climate. Hydro generation availability is a function of streamflow, or runoff water which will pass through the generators. Runoff water, in turn, is a function of precipitation and evaporation. Relatively small percentage changes in precipitation can lead to much larger percentage changes in runoff and hydro generation availability. Hydro is an important source of power in New York, but not for the Southeastern case study utility. Historical stream flow data for the Niagara River and the St. Lawrence River at the points of New York's two principal hydro sites and annual hydro generation data are related using regression analysis. The changes in Great Lakes net basin supply and stream flow developed from the climate change scenarios in Section 2 are used with this statistical model to estimate the potential impacts on future hydropower generation availability. The details of this analysis are reported in Section 5. The second step in this task is to develop the relationship between temperature changes and the availability of electrical generation from non-hydro sources. Climate conditions can alter the effective capacity and operating efficiency of gas turbines (used primarily for generating power during periods of peak requirements) as well as fossil-fired and nuclear steam generators (used to serve baseload and intermediate load requirements). The largest relative impact is the relationship between gas turbine efficiency (i.e., effective unit capacity and heat rate)5 and ambient air temperature: turbine efficiency falls with higher ambient air temperatures. 4 That is: + AT&D losses Ademand Aenergy requirements t t t (1 + loss multiplier) Ademand t 5 "Effective capacity" measures the maximum potential output (in megawatts (MW» that can be obtained from a generating unit based upon values for various operating conditions. "Heat rate" measures the amount of energy provided by fuel (in British thermal units (Btu's) that is required to gen~rate one kilowatt hour (kwh) of electricity. 3 - 10 From the digital collections of the New York State Library. FIGURE 3-5 STEPS IN TASK 3: DEVELOPMENT OF WEATHER SENSITIVITY OF SUPPLY Chanae. in Weather (from climate cqanl. scenario) Base Total Sourc•• (fro. Task 1) 3.1. 3.2. Adjust Hydro Production 3.3 Adjusted Non-Hydro Production Acljusted Total Sourc •• 3 - 11 From the digital collections of the New York State Library. Similarly, the efficiency of steam generators is sensitive to air or water temperatures used for cooling during the condensing stage of the steam cycle. An inverse relationship also is found here: the higher the air or water temperature used for cooling, the lower the generator efficiency.6 In order to schedule plant maintenance and plan system dispatch, the Southeastern utility has used generating unit test data to calculate effective unit capacity as air and cooling water temperatures change on a seasonal basis. IeF requested that the Southeastern utility exercise these relationships using the changes in average seasonal temperatures estimated for that region as reported in Section 2. The results of these analyses and the implications of climate change for electricity supply from gas turbines and steam generating units are discussed in Section 4. Taken together (Step 3.3), these two steps define the relationship between changes in utility production operations and changes in temperature. TASK 4: DEVEIDP ALTERRATIVE SCERARIO 0UTPU1'S The purpose of this task is to use the information developed in Tasks 2 and 3 to determine the impact of climate change scenarios on key utility parameters listed in Step 1.6. Figure 3-6 illustrates the steps to accomplish this task. The adjusted LOes calculated in Task 2 and adjusted supply characteristics calculated in Task 3 are used to calculate optimal utility investment and operating strategies under these alternative conditions. For LOCs, climate-related changes include changes in both the amplitude and shape of the LDCs as compared with the base case curves (Step 4.1) . Changes in supply characteristics include new levels of hydro generation availability and effective capacity of other types of generating units (Step 4.2). The impacts of climate change are estimated using the ICF Integrated Planning Model (IPM) (Step 4.3). The IPM is formulated as a linear programming model which combines capacity planning and powerp1ant dispatching decisions to provide the lowest net present value of generation costs over the planning horizon. It can be used to determine the optimum capacity expansion plan given a set of investment options, demand growth forecasts, and reliability criteria. Also, given a capacity expansion schedule (determined by the model or input to the model), it provides the optimum utilization (dispatch) of power plants given their operating characteristics, fuel prices, and known operational constraints. The outputs of the IPM (Step 4.4) include capacity expansion schedule, fuel use by type, and capital and operating costs. T.ASl( 5: allPARE BASE CASEIALTERRATIVE CASE 00TP0'l'S The purpose of and approach to this task are straightforward. The results of the base case (Task 1) are compared with the results of the alternative cases (Task 4), and differences in estimated values for utility planning factors are calculated. These differences in sales, peak demand, 6 Air temperature also can affect the loss of cooling water from evaporation in generating units using closed cycle cooling technologies (e.g., natural draft or mechanical draft cooling towers). 3 - 12 From the digital collections of the New York State Library. nauu 3·6 STIrS III TASK 4: DIVElDPMRn: OF AL'lIIIIATIU CASB STUDY SCIRAIlIOS 4.1 4.2 Total Requir..antI (fro. Task 2) Adjua~.d Adjusted Total SouZ'c•• (frOll Tat 3) 4.3 Utility P1Aftftini Hoelel (IPM) 4.4 Adjuated C... a.sult.: • Sale. • Peak ~ Fuel • Capac:ity • Costs 3 - 13 From the digital collections of the New York State Library. capacity, operations (e.g., fuel use by type) and costs represent the potential impacts of climate change on the case study utilities. Sensitivity analysis is performed to estimate ranges of impacts under alternative assumptions and conditions. Design of some of the more important sensitivity analyses is described below. SENSITIVITY ANALYSIS: SPECIFICATION OF ALTERRATIVE UTILITY PLANNING SCERARIOS Because there are substantial lead times associated with installing utility capacity and implementing operating strategies, utility planning studies must consider the occurrence of future events. Future economic, technological, climatic, and other conditions and events can affect both the need for power and the options available to the utility to meet those needs. These future conditions and events are uncertain. As a result, utility planning decisions are risky. A utility planner may optimize utility investments and operating strategies based upon a particular forecast of future conditions. However, if a different set of conditions occurs, the utility plans may not be optimal. For example, an unanticipated increase in peak demand can have serious operating consequences for a utility, which may have to: operate high fuel-cost or inefficient equipment; buy emergency power from neighboring utilities; cut voltage levels to its customers; interrupt service to some customers; or build expensive, short lead-time capacity. Unless the utility has planned for possible increased seasonal sales or peak demands associated with increased temperatures, it may not be able to provide low cost, reliable service. On the other hand, if demand is lower than anticipated, the utility may have substantial idle generating capacity. The costs associated with this idle capacity will have to be borne by ratepayers or utility stockholders. To the extent that these changes in demand can be anticipated, capacity and operating decisions can be optimized, and the utility's objectives met. The consequences of inaccurate forecasts of future conditions and events--i.e., forecasts which are too high or too low compared with actual conditions and events--can be substantial. Utilities often conduct sensitivity analyses to estimate the economic risks associated with "overbuilding" or "underbuilding" new capacity; that is, to estimate the economic consequences of inaccurate forecasts. The results of these sensitivity analyses, for example, may indicate that the costs associated with underbuilding capacity (1) are substantial and (2) greatly exceed the costs associated with overbuilding capacity. In such a case, a prudent utility planner would recommend that the utility build more capacity than suggested by the most likely or mean value of the distribution of uncertain forecasts. Forecasts of the level and pattern of climate change are very uncertain. In this study we estimate the order of magnitude of ecoucnd,c 3 - 14 From the digital collections of the New York State Library. risks associated with ina.ccurately anticipating climate change. In other words, we are interested in studying the potential costs associated with making planning decisions that are not proven to be economically optimal. Are these costs likely to be significant or not? If they are significant, should utilities spend more effort to incorporate climate change into their planning studies, or is climate change of little consequence? Our approach for studying this issue is as follows. For selected climate change scenarios, values for utility planning factors -- including the present value of capital and operating costs for a future year (an output of the IPM) - - are estimated for four cases corresponding to the cells of the following matrix: Climate Change Occurs? No Yes Utility plans incorporate impacts of climate change? No 1 2 Yes 3 4 Cell #1 represents the base case utility planning assumptions and implicitly the climate. change outcome expected by utility planners. Cell #4 represents the case in which planners correctly anticipate climate change over time and optimize their long-term investment and operating plans accounting for these changes. Comparison of the results of Cell #1 with Cell #4 provides estimates of the impacts of climate change on utility planning factors derived from implementation of the process described previously in this section. In both of these cells, the utility is assumed to anticipate climate change correctly and optimize its capacity investments and operations. Cells #2 and #3 represent cases in which the utility incorrectly anticipates climate change. In cell #3, the utility builds capacity in anticipation of climate changes which do not occur. In this case, costs associated with idle capacity will be hi~her than in cell #1, and will bean estimate of the costs of "overbuilding". In cell #2, the utility will not have included sufficient capacity in its long-term plan to meet changes in demand and operating efficiency caused by changes in climate. Short term response in the form of coping strategies will have to be employed to meet customer demands. Short-term responses include building short lead-time generating capacity, operating high fuel cost or inefficient generating equipment for more hours than planned, and purchasing high cost emergency 7 Implicit in this discussion are the assumptions that climate changes associated with global warming will (1) increase the demand for electricity and (2) decrease the operating efficiency of electricity generation. These assumptions are likely to be valid for most, but not all, utilities in the u.s. 3 - 15 From the digital collections of the New York State Library. power from neighboring utility systems. The costs associated with these short-term responses provide an estimate of the costs of "underbuilding".8 A final set of sensitivity analyses addresses the potential imp~cts of climate change under alternative assumptions of growth rates in demand caused by events and conditions other than climate change. In particular the process described by Tasks 1-5, is implemented using demand growth assumptions higher than those provided by the case study utilities. The rationale for these alternative calculations is that some utility analysts anticipate greater economic growth and higher utilization of electricity than the "expected value" growth rates provided by the case study utilities. Climate changes which increase electricity demand would exacerbate these trends, thereby increasing the amount of production capacity that would be needed to meet customer requirements. The results of the sensitivity analyses indicate the range of climate change impacts under the "base growth" case and "high growth" case assumptions" Sections 4 and 5, which document the Southeastern and New York case studies, respectively, are organized parallel to this section. An initial subsection describes the utility and its base case planning assumptions. Subsequent subsections address the weather-sensitivity of demand, the weather-sesitivity of supply, development of alternative cases and sensitivities, and the estimated impacts of climate change on the case study utilities' planning factors. A discussion of the overall conclusions and recommendations developed from the case study utilities is presented in Section 6. 8 Note that the cases described by cells #1 and #4 assume that the utility planners foresight regarding climate change is perfect. In cells #2 and #3 his foresight is totally imperfect. In reality, climate change will occur gradually over time and planners will have opportunities to adjust their plans as the uncertainties of climate change are resolved. In this sense, the estimated cost impacts developed by comparing the alternative cases should be considered as "bounding" cases. 3 - 16 From the digital collections of the New York State Library. SacrIQR 4 _ IRTRODUcrIOR AlQ) urILIft CASE S'ftJD'f OVERVIEV1 The Southeastern case study utility is located in a relatively high growth area of the Sun Belt. Table 4-1 indicates that the large majority of its sales go to residential and commercial customers. Among these customers, air conditioning is an important use of electricity. Air conditioning systems are utilized heavily in the summer, but there are also periods of air conditioning use in the other seasons, including the winter. There is very little heating load in the region, but electric systems capture a substantial portion of the heating market because of low capital and installation costs compared with natural gas or oil systems. A problem the utility faces is that electric heating systems - - particularly in the residential sector -- exhibit very low load factors. That is, these systems are turned on for only short periods of time, typically on cool winter mornings. The utility must maintain suff~cient capacity to meet these "spiked" demands, but the peaking capacity is not utilized highly at other times. In sum, a substantial portion of the utility customers' demands consists of weather-sensitive load. Climate changes which increase average temperatures are likely to increase the utility's high air conditioning loads and decrease the smaller, but spiked, heating loads. Historically, and because of its location, a substantial portion of the Southeastern utility's generating capacity is gas-fired or oil-fired (see Table 4-2). Changing market conditions during the 1970's have resulted in this suboptimal capacity mix (i.e., not least-cost). In the near term, the utility is purchasing lower cost electricity from neighboring, coal-based utilities and is planning to build coal-fired, gasification combined-cycle units as new capacity is needed over time. Hydropower is not an important source of generation for the Southeastern utility. BASE CASE UTILITY PLAR DeII8Dd Grcnrth Forecasts of future growth in annual demand for electri-c energy were provided by the case study utility. Using 1985 as a base year, the 1 We wish to acknowledge the invaluable assistance of the Southeastern utility's staff in providing data and comments on interim results. The utility also analyzed the potential impact of climate change on generating unit effective capacity and operating efficiency at our request using a model developed by the utility. 4 - 1 From the digital collections of the New York State Library. TABLE 4-1 SOUTHEASTERN UTILITY DISTRIBUTION OF 1985 SALES BY TYPE OF CUSTOMER (percent) Residential Commercial Industrial Other Source: 50.5 39.7 7.9 1.9 100.0% Southeastern case study utility. TABLE 4-2 SOUTHEASTERN UTILITY DISTRIBUTION OF 1985 CAPACITY AND GENERATION BY FUEL TYPE (percent) Annual Capacity Nuclear Coal Steam-Gas Steam·Oil Combined Cycle Combustion Turbine Purchases Cogeneration .18.9 0.0 33.2 18.1 3.5 14.8 10.9 0.5 100.0% Source: Generation 38.6 0.0 16.3 13.2 3.3 0.0 27.9 0.6 100.0% Southeastern case study utility. 4 - 2 From the digital collections of the New York State Library. forecasts covered the period 1986-2000. 2 rCF extended the forecasts to 2015 assuming a continuation of the constant growth rate provided by the utility for 1995-2000. Table 4-3 summarizes the electric energy forecasts. The forecast growth rates average 2.65% per year through 1990, 2.06% per year between 1990 and 1995, and 1.35% per year thereafter. by Many assumptions underlie the case study utility's forecast, including their success in implementing conservation and load management programs to reduce growth rates in future customer demands. Although forecasts of economic growth suggest that the case study utility's service area is likely to be economically healthier than most other regions of the country, the forecast electric energy growth rate of 1.35% in the out years is quite low. As described in the previous section, future demand growth is uncertain (particularly beyond a ten-year horizon) and Ls an issue to be addressed through sensitivity analysis. As an alternative to the case study utility's forecast, a growth rate of 2.3% was assumed for all years after 1995. This rate approximates the average of several forecasts of electricity growth for the U.S. as a whole, and was suggested by the technical representatives of the project sponsors. Not only is the forecast of total, annual electric energy demand (in kwh) important for utility planners, but so is the forecast pattern of demand. Particularly important are forecasts of monthly or seasonal sales and peak demands for these periods. Based upon conversations with case study utility staff, we assumed that annual peak demand (in MY) would grow at the same rate as electric energy. The pattern of demand across seasons, months and hours (i.e., normalized utility system load shape) was assumed to remain constant over time. Load shapes describing this pattern were taken from 1984. This year was described by the case study utility staff as being reasonably typical in terms of factors affecting the pattern of demand, principally economic activity and seasonal weather conditions. Resource Plan The case study utility's capacity expansion plan was assumed through 1995. Staff also provided us with detailed information on firm capacity and economy energy purchases from neighboring utilities and on the potential development of cogeneration resources in the utility's service area. Beyond 1995, it was assumed that new generating capacity would be built to meet new load growth and maintain reliability levels measured by a 15% reserve margin (i. e., available, effective capacity in excess of forecast annual peak demand) . The Southeastern c ase study utility is planning to meet forecast load growth primarily with combined-cycle generating units. 3 Because of the 2 Base year electric energy demands were approximately 53,363 GWH. 3 These units are based on gas turbine technology. However, they are much more efficient than conventional gas turbine because they make effective use of waste heat from "first stage" generation to power a "second stage" turbine. Combined cycle units also have several advantages as compared with conventional baseload generating units. These include their ability to be built in smaller increments and shorter lead times. Their operation also tends to be more environmentally benign. 4 - 3 From the digital collections of the New York State Library. TABLE 4-3 SOUTHEASTERN UTILITY AVERAGE ANNUAL GROV11I RATES IN DEMAND FOR ELECTRIC ENERGY (percent) Period Base Case Annual Growth Rate High Case Annual Growth Rate 1986-90 1991-95 1996-00 2001-05 2006-10 2011-15 2.65 2.06 1.35 1.35 1.35 1.35 2.65 2.06 2.30 2.30 2.30 2.30 Source: 1986-2000: 2001-2015: Southeastern case study utility. reF. 4 - 4 From the digital collections of the New York State Library. state of technological development and forecast fuel market conditions, these uni ts are assumed to be fired by natural gas prior to 2000, at an investment cost of $700/kw (in 1985$). Between 2000 and 2015, new additions are asswned to include coal gasifiers so that coal can be used as the principal generating fuel. These units are assumed to cost $1500/kw (in 1985$) . If demand grows faster than' forecast (e. g., as a result of unanticipated changes in climate), the utility's short-term response is assumed to be construction of short lead-time gas turbines ($350/kw in 1985$) to maintain its planning reserve margin of 15'. Table 4-4 summarizes the base case capacity expansion plan for the Southeastern utility assuming "base load growth" conditions (i.e., 1.35% per year after 1995) and "high load growth" conditions (i.e., 2.30% per year after 1995). Fuel Prices Forecasts of future fuel prices are an important determinant of utility capacity planning and operating decisions. Relative fuel prices determine the most economic type of capacity to build and the most economic utilization, or dispatch, or available capacity. Assumed fuel price forecas ts are presented in Table 4- 5 . These represent forecas ts from the IeF Energr Service which were current at the time the analyses were conducted. The forecasts indicate the potential, long-term economic advantage to the utility of coal-fired generating units. Section 3 described two approaches for estimating the sensitivity of demands for electricity to changes in weather conditions. The statistical approach uses regression analysis to relate data on historical weather conditions to changes in aggregate, utility system demands. The structural approach operates on more disaggregated, end-use data and then "adds up" the end-use impacts to estimate the utility system impacts. Disaggregated, end-use data were not available for the Southeastern case study utility, so the structural approach could not be applied. Although data were not available for weather-sensitive end-use loads, the Southeastern utility sponsored a detailed weather-normalization study, and the resul ts were made available to us . The purpose of weather-normalization studies is to develop adjustment factors to "correct" actual utility load data for the effects of abnormal or atypical weather conditions, usually on a monthly or seasonal basis. Utilities use the adjusted data for short-term and long.-term operational planning. These adjustment factors, in essence, estimate the sensitivity of utility loads to changing weather conditions. The adjustment factors are developed using statistical techniques similar to the description of the statistical approach in Section 3. The 4 The reF Energy Service is a semi-annual subscription service which reports current energy market events and short- and long-term forecasts. 4 - 5 From the digital collections of the New York State Library. TABLE 4-4 SOUTHEASTERN UTILITY BASE CASE CAPACITY EXPANSION POST-1995 (megawatts) Gas-Fired Combined Cycle o 939 1995-2000 2000-2015 Source: Base Load Growth Coal-Fired Combined Turbine/ Cycle Other o o o 5810 High Load Growth Gas-Fired Coal-Fired Combined Combined Turbine/ Cycle Cycle Other o 1889 o o o 9221 ICF Incorporated estimates. TABLE 4-5 SOUTHEASTERN UTILITY AVERAGE ANNUAL GROWTH RATES IN REAL FUEL PRICES (percent) Period Oi1-1% Resid 1986-90 1991-95 1996-00 2001-05 2006-10 2011-15 -2.3 4.3 3.0 3.0 3.0 0.0 Source: reF Oil-Distillate -4.9 2.4 2.6 2.6 2.6 0.0 Natural Gas -1.7 4.2 3.0 4~1 1.9 0.0 Energy Service, Fall 1985. 4 - 6 From the digital collections of the New York State Library. Coal -1.0 -0.1 1.5 0.9 1.0 1.1 Nuclear 0.0 0.0 0.0 0.0 0.0 0.0 relationship and estimation techniques in the weather-normalization study, however, are somewhat different and are more complex. Specifically, in the weather-normalization study (1) economic activity, electricity prices, and other variables were included explicitly in the relationships to account for the effects of changes in these variables in addition to weather on electricity demands and (2) the parameters of the relationships were estimated using monthly weather and sales data as opposed to tri-hourly data. S Our review of the weather-normalization study indicated that a sound analysis had been conducted of the complexities of weather-utility demand relationships. The results of the analysis were used to estimate the impacts of climate change on monthly electric demands. However, an important consideration for utility planning is how the pattern of demands might change within monthly periods; specifically, are the changes likely to be concentrated on weekdays or weekend days, at night or during the day, or during periods of high electricity production costs vs. low cost periods? To address these~ considerations, a set of statistical relationships was developed using tri-hourly demand and weather data provided by the Southeastern utility to allocate the monthly demand changes to various periods during the month. To implement this approach, data for January 1980 through August 1985 were used. The following illustrates the steps involved in developing the relationships: Step 1. For all July weekdays in the data base (i.e., for July 1980, 1981, . .. , 1985), calculate load observations as the average hourly load between (e.g.) 3:00 pm and 6:00 pm. Step 2. Group observations by year, observations for each year. and calculate the mean of the 5 The general form of the relationship is as follows: monthly jurisdictional electricity use per customer is a function of monthly cooling and heating degree days, electricity price, per capita personal income, and the saturation of air conditioning and electric space heating. For e~ample, in summer months: In (EU) = 7.12 E-6 * CDD6S * ACC - .351 * In (PRICE) + .472 * where: ED = monthly electricity use per customer, CDD65 = monthly cooling degree days calculated on a 65 0 base, Ace = saturation of air conditioning, PRICE = real price of electricity, INCOME = per capital personal income. 4 - 7 From the digital collections of the New York State Library. In (INCOME) Step 3. For each observation calculate the percent deviation from the mean for that ~ear. Step 4 . . Repeat steps 1-3 with weather using 72 0 base) observations. 6 (cooling degree days calculated Step 5. Regress the resul ts of step 3 on the resul ts of step 4; i . e . , estimate the percent deviation of the load observations as a function of the percent deviation of the weather observations. Step 6. Repeat steps 1-5 (a) for all months, (b) for weekdays and weekend days, and (c) for all daily, tri-hourly periods. An example of the estimated relationships (for July weekdays between 3:00 pm and 6:00 pm) is as follows: ~Load(t) - .3311 * ~Weather(t) where: ~Load(t) ~ ~Weather(t) - Percent deviation of weather (cooling degree days calculated using 72 0 base) from the mean weather for the given year, month, and time period. Percent deviation of load on day t from the mean load for the given year, month, and time period, The statistical fits generally were good, and aggregating results to months provided results similar to the findings of the weather normalization study. An important assumption implicit in the estimation of the relationships is that the weather-sensitive portion of the Southeastern utility's system loads will stay constant over time at the level exhibited by the 1980-85 historical data. While this assumption is a source of uncertainty in estimating future climate change impacts, it is reasonable for the Southeastern utility; the utility's service area already exhibits high market saturations of electric air conditioning and space heating equipment. Warmer temperatures and economic changes over the period of analysis are unlikely to affect these market saturations by a large ·amount. The final step to estimate the impact of climate change on electricity demand during the tri-hourly periods is to input the temperature change scenarios into the estimated relationships. An example of these calculations for July 2015 on weekdays between 3:00 p.m. and 6:00 p.m. is: 6 This procedure was followed for the Southeastern utility case study. However, for the New York case study when using weather data in step 3, the actual deviations rather than the percent deviations were calculated. This resulted in a more stable model and more reasonable results. 4 - 8 From the digital collections of the New York State Library. Description Source Value I (1) Mean Cooling Degree Days 11.85 1.87 (2) Change in Cooling Degree Days Assumed Base Weather Conditions Climate Change Scenario C Due to Climate Change (3) Percent Change in Cooling Degree Days 15.80 (2) / (1) (4) Coefficient of Regression 0.3311 Regression Analysis 5.23 (5) Percent Change in Load Due to Climate Change (3) * (4) Table 4-6 illustrates the results of these calculations by providing estimated percent changes in tri-hourly energy requirements for an average weekday and weekend day in July 2015. The illustration was developed assuming the case C (or "high") temperature change scenario (i.e., a change in average summer temperature of approximately 1.9 0F). As would be expected, from typical utility load patterns, the use of air conditioning equipment results in loads being most sensitive to temperature changes in the afternoon and evening hours, and least sensitive in the early morning hours. Also as expected, the percent change in demand is lower on the weekend day as compared with the weekday, but the patterns among tri-hourly values are similar. As will be discussed below, changes in demand on the order of 3-4% are significant in terms of utility planning and operations. Table 4-7 illustrates the estimated changes in monthly energy requirements calculated by aggregating across .tri-hourly periods and day types. As described above, these aggregations are similar to resul ts obtained using the weather normalization study. An important point is that the Southeastern case study utility must satisfy customer demands for cooling in all twelve months of the year. The winter months of December through March are cold enough on certain days to require the use of space heating equipment. For these months the weather sensitivity relationships were estimated separately using 1) heating degree day temperature data and 2) cooling degree day temperature data. These data series relate actual temperatures to nominal, "base" temperatures or comfort zones for which the use of space conditioning equipment is not required. Note that in che winter months of December, January, and Feburary, the assumed temperature increase leads to a net reduction in system electricity demands. In March, however, the increased demand for cooling exceeds the reduced demand for heating. Table 4-7 indicates that the estimated net change in annual demand, or energy requirements, under high climate change and base load growth condi tions is 3.40%. Seasonal and annual peak demand also are important planning factors for utility capacity investment decisions and operating decisions such as maintenance planning. The Southeastern utility typically achieves its summer -- and annual -- peak on an August weekday in the late afternoon. The estimated peak demand sensitivity from the weather 4 - 9 From the digital collections of the New York State Library. TABLE 4.,.6 SOUTIlEASTERN UTILITY EXAMPLE OF CHANGE IN ENERGY REQUIREMENTS (PERCENT) BY DAY TYPE AND TIME OF DAY JULY. 2015* Time Period 12:00 3:00 6:00 9:00 12:00 3:00 6:00 9:00 AM - Whole Day * 3:00 AM AM 6:00 AM AM - 9:00 AM AM - 12:00 PM PM - 3:00 PM PM - 6:00 PM PM - 9:00 PM PM - 12:00 AM Weekday Change In Requirements (%) Weekend Change In Requirements (%) 612.4 553.9 661.4 873.6 965.7 988.6 937.3 811.3 4.09 3.68 2.88 3.26 4.25 5.23 5.39 5.07 3.64 3.27 2.03 2.51 3.72 4.91 5.17 4.93 6,404.2 4.33 3.87 Weekday Base Dsase (GWH) Assumes temperature change scenario C and base load growth. 4 - 10 From the digital collections of the New York State Library. TABLE 4-7 SOUTHEASTERN UTIUTY ESTIMATED CHANGES IN TFJlPERATURE AND ENERGY REQUIREMENTS BY MONTH, 2015* Month Temperature Change (F°1- Demand Change (%) Heating Cooling Net January February 2.02 2.02 -2.06 -1.42 0.52 1.01 -1.54 -0.41 March April May 1.53 1.53 1.53 -0.50 0.0 0.0 1.44 2.81 4.70 0.94 2.81 4.70 June July August 1.87 1.87 1.87 0.0 0.0 0.0 6.37 5.65 4.74 6.37 5.65 4.74 September October Novmeber 1.42 1.42 1.42 0.0 0.0 -0.21 5.19 4.80 1.72 5.19 4.80 1.51 December 2.02 -0.72 l.47 0.75 Annual 1.71 -0.32 3.72 3.40 * Assumes temperature change scenario C and base load growth. 4 - 11 From the digital collections of the New York State Library. normalizaton study was similar to the estimated change in demand for the tri-hourly period 3:00 pm-6:00 pm for weekdays in August and is taken as our estimate of the potential effects of climate change on annual peak demand.7 The estimated change in 2015 under high climate change and base load growth assumptions is 7.04%. Because the percent increase in peak is greater than the percent increase in energy requirements for 2015, the estimated load factor for the Southeastern utility would be lower than in the absence of climate change. 8 This implies relatively greater requirements for generation of high cost, peak load electricity than for generation of lower cost, base10ad electricity. Our estimates of the potential impacts of climate change on energy requirements and annual peak demand for the Southeastern utility are summarized in Tables 4-8 and 4-9 and in Figures 4-1 and 4-2. These tables and figures show the percent impacts for selected future years under low (A), medium (B), and high (C) temperature change scenarios relative to base case (no climate change) assumptions. On an annual basis, the impacts show increases in energy requirements and peak demands 1) over time and 2) with higher assumed temperature changes. By 2015, the range of estimated impacts on energy requirements is 1.2% to 3.4% and on peak demand is 2.1% to 7.0%.9 The reader is cautioned that uncertainties in the estimates of future temperature changes and uncertainties in the estimation of weather sensitivity relationships could result in actual impacts being higher or lower than the bounds of these ranges. However, our judgment is that the ranges are reasonable given the current state of information. VFATIIER.-SENSITIVITY OF SUPPLY As described in Section 3, climate change can affect the operation of utility resources for supplying electricity. Several factors can combine to 7 Periods of utility peak demand usually are associated with abnormal or extreme weather conditions. As described in Section 2, we are unable to estimate the potential effects of global climate change on daily weather patterns and extreme events. Climate change could either exacerbate or smooth the patterns. In the absence of additional information, the average seasonal temperature changes reported in Section 2 are used to estimate potential climate change impacts on utility peak demands. 8 Appendix A describes the calculation of load factor. 9 Because our estimation methods assume that weather sensitive load is a constant proportion of utility system demands over time, the percentage impacts of climate change are independent of the underlying load growth rates assumed. Thus, the percent changes in energy requirements and peak demand are identical under "base If load growth and "high If load growth assumptions, although the potential impacts measured in GWH and MW would differ. The GWH and MW impacts under high load growth assumptions are used in sensitivity analyses described later in this section. 4 - 12 From the digital collections of the New York State Library. TABLE 4-8 SOUTHEASTERR UTILITY ESTDfATED IMPACl OF CLDIATE CHARGE ON ANNUAL ENERGY REQUIREMENTS, SELECTED YEARS* (gigawatt-hours) Base Case Increase Due to Climate Chanie (%) Scenario A Scenario C 2000 72,714 362 (0.50%) 1,954 (2.69%) 2015 88,911 1,045 (1.18%) 3,019 (3.40%) * Assumes Base Utility Load Growth Rates. TABLE 4-9 SOUTHEASTERN UTILITY ESTIllATED IMPACT OF CLDIATE CHARGE ON ANNUAL PEAK DEKAND, SELECTED YEARS* (aegavatts) Base Case Increase Due to Climate Change (,) Scenario A Scenario C 2000 14,316 126 (0.88%) 678 (4.74%) 2015 17,505 367 (2.10%) 1232 (7.04%) * Assumes Base Utility Load Growth Rates. 4 - 13 From the digital collections of the New York State Library. FIGURE 4-1 SOUTHEASTERN UTI LITY PERCENT CHANGE IN ANNUAL ENERGY REQUIREMENTS ALTERNATE CLIMATE CHANGE SCENAR lOS 4....-.----~--------------------, 3 1 O-WI,~~~ ----,.----~---,.....----r-----"" 1990 1985 2010 200S 2000 1995 Year FIGURE 4-2 SOUTHEASTERN UTILITY PERCENT CHANGE IN ANNUAL PEAK DEMAND: ALTERNATE CLIMATE CHANGE SCENARIOS 8..,.------------------------.. . . 7 6 s 3 .... 2 _ 1 _...... i«-·'" . ~ .... tC."··"· ........... o~, 1985 ....... . . . . '" • • tc ~ ~ # _ • ,.-.,. Scenario A ."'.-_.--- -......,.----.,....-----...,...-----,...---......,...------1 1995 2010 2015 1'90 1000 2005 Year 4 - 14 From the digital collections of the New York State Library. 2015 affect the availability of stream flow to generate electricity with hydropower plants. The efficiency of fossil-fueled and nuclear plants can be affected by the temperature of air used in gas turbines or of air and water used for cooling in steam plants. Hydropower is not a significant source of electricity for the Southeastern case study utility, and was ignored in this analysis. The potential impacts of climate change on thermal plant efficiencies, however, has been analyzed. This issue is important to the Southeastern utility on a seasonal basis primarily because of the impacts of seasonal temperature changes on the effective capacity of generating units (measured in MW) and the operating efficiency, or heat rate, of these units (measured in BTU's of fuel per KWH of electricity generated). Based upon test data for its generating units, the Southeastern utility developed relationships between generating unit heat rates and effective capacity on the one hand, and air and cooling water temperatures on the other. The Southeastern utility exercised these relationships using the temperature changes contained in the climate change scenarios. Our expectations were that the estimated impacts on heat rate and effective capacity would be small, because the increases in average temperatures over time from the climate change scenarios were small compared with changes in seasonal temperatures within a given year. These expectations are borne out by the results presented in Table 4-10. The figures indicate, for example, that effective capacity varies by up to 3 MW across seasons assuming no climate change, while the maximum impacts of high climate change assumptions by 2015 are 0.3 to 0.8 MW. This represents a potential reduction of one- to two-tenths of a percent of the unit's effective capacity. Similar results are estimated for changes in heat rates. Because these impacts are small absolutely and in comparison with the potential impacts on changes in demands for electricity, the estimated supply impacts were not included in our utility modeling and estimation of capacity requirements, fuel utilization, and costs. ALTERNATE PLANNING SCENARIOS The next subsection reports the potential impacts of climate change on utility planning factors for a number of alternate scenarios. The first set of results assumes that future climate changes are known with certainty and that utility planners are able to optimize their investment and operating decisions by minimizing the costs of electricity production. New capacity is built to maintain a 15% reserve margin (gas-fired or coal gasification combined cycle units). Values for utility planning factors are calculated (1) for three alternative climate change scenarios: no climate change (base case assumptions), low (Scenario A) temperature change, and high (Scenario C) temperature change and (2) for two alterntive sets of demand growth rates: base and high. In the second set of results, we vary the assumption that utility planners are able to anticipate climate change correctly. These alternative analyses were described in Section 3. To the extent that the utility is 4 - 15 From the digital collections of the New York State Library. TABLE 4-10 SOUTHEASTERN UTILITY POTENTIAL IMPACTS OF TEMPERATURE CHANGE ON HEAT RATES AND EFFECTIVE CAPACITY FOR A TYPICAL 400 KW THERMAL PlANT Heat Rate (BTU/KWH) Scenario C No Climate Climate Change Change Impact. 2015 Winter Spring Summer Fall 9256.0 9291.4 9391.4 9291.4 + 7.5 +13.4 +20.5 +11.3 (0.08%) (0.14%) (0.22%) (0.12%) Effective Capacity (M'W) Scenario C Climate Change No Climate Impact. 2015 Change 370 369 367 369 4 - 16 From the digital collections of the New York State Library. -0.3 (-0.08%) -0.3 (-0.14%) -0.8 (-0.22%) -0.4 (-0.11%) unable to anticipate climate changes correctly in the long term, the utility is not able to optimize its choices, and it is faced with additional costs associated with "overbuilding" or "underbui1ding" its system. The magnitude of these additional costs for the base load growth and high climate change assumptions are reported below. UTILITx IKPACTS This subsection summarizes the estimates of potential climate change impacts on capacity requirements, fuel utilization, and costs for the Southeastern case study utility using reF's Integrated Planning Model (IPM). Figure 4-3 indicates a range of capacity additions which may be required by the Southeastern utility through 2015 in order to meet customer demands and maintain a 15% planning reserve margin. Results for different scenarios are presented, repres.enting alternative climate change and load growth assumptions. Assuming no climate change, 6,749 and 11,110 MW of capacity would need to be added in the base and high load growth cases, respectively. In addition, 422 to 1,417 MY would be required under base load growth assumptions to meet increased demands in response to climate change. This represents an increase in new capacity requirements of 6 to 21%. Comparable figures under high load growth assumptions are an addition of 514 to 1,724 MW, or 5 to 15%, in new capacity requirements in response to climate change. Differences in the underlying growth rates in demand can lead to substantial differences in capacity requirements. Utility planners generally recognize these uncertainties in their planning analyses. However, the occurrence of and customer response to climate change also is uncertain and can exacerbate these differences. The results above indicate the possibility of additional capacity requirements on the order of 10% or more in response to climate change. Figure 4-4 presents the estimated capacity requirements induced by climate change for 2000 and 2015. The amount "induced by climate change" is calculated by subtracting the capacity requirements for the no climate change scenarios from the capacity requirements for the low (A) or high (C) temperature change scenarios. What is important about this figure is that about one-third to one-half of the impacts discussed in the previous paragraphs could occur by 2000, a period well within utility intermediateto long-term planning horizons. Table 4-11 describes fuel utilization by type in 2015 under the base case (no climate change) conditions and low (A) and high (C) temperature change scenarios. In all cases, base load growth is assumed. Two alternative assumptions regarding utility response to climate charge are reported in the table. In one case (designated "long-term plan"), the capacity requirements induced by climate change as presented in Figures 4-3 and 4-4 are anticipated and are included in the utility's long-term plan. Here, gas-fired combined-cycle units are added through 2000, and coal gasification/combined cycle units are added after 2000 to meet the utility's capacity requirements. 4 - 17 From the digital collections of the New York State Library. FIGURE 4-3 SOUTHEASTERN UTIUTY CUKULATlVE CAPACITY REQUlREKENTS BY 2015 15000 6 = Climate Change Assumption 12,834 13000 11 110 11000 MW 9000 7000 5000 3000 8166 n 6749 1000 No6 '::; Lowf:). NoA Hilh6 ~ Base Case Load Growth J " Low~ Hlah6 -......,.a High Load Growth 4 - 18 From the digital collections of the New York State Library. # FIGURE 4-4 SOUTHEASTERN UTILITY CUMULATIVE CAPACITY REQUIREMENTS INDUCED BY CLIMATE CHANGE 2000 1800 1724 1600 1400 1200 MW 1000 800 600 422 ~OO 200 o~-- 2000 2015 Scenario A .. , 2000 2015 . ~~en~~o C.,., 2000 2015 Scenario A '= ..... " 2000 High Load Growth Base Case Load Growth 4 - 19 From the digital collections of the New York State Library. 2015 Scenario C# In the other case, climate change is not anticipated by the utility planners, but it does occur. Additional capacity requirements can be met through purchases of firm capacity from neighboring utility systems or construction of short lead time gas turbines. l O The figures in Table 4-11 assume that turbines are built. The table indicates that anticipating climate change results in increased construction and utilization of coal units. The long-term plan cases result in utilization of 17 to 40 billion Btu's more coal in 2015. The high temperature change case results in an increase in coal utilization of almost 10 percent. If the increased generation requirements are met as a result of short-term response to climate change, the utility would tend to rely more heavily on existing oil-fired steam plants and economy purchases when available. The short- term plan case assuming the high temperatures change scenario causes an increase in oil utilization of almost 13 percent and a small increase in purchases. Finally, the last row of Table 4-11 indicates the difference in fuel costs across the various cases. The results emphasize the high costs of oil-fired generation under the short-term planning assumptions. Figure 4-5 indicates the total costs -- capital plus fuel -- in 2015 associated with anticipating climate change correctly. Capital costs are calculated as a 1eve1ized cost per kwh. As can be seen from the figure, climate change could result in annual costs on the order of $61 to $212 million (1985$) in 2015 under base case load growth assumptions and $76 to $257 million under high load growth assumptions. The maj ori ty of these costs are for new capacity. Recall from the discussion of demand sensitivity in Section 4 that the estimated changes in average temperatures resulted in greater percentage changes in peak demand than in energy requirements. This means that low fuel cost, coal gasification/combined cycle units built to meet the additional capacity requirements associated with climate change after 2000 are available 1) to serve the additional, climate-induced energy requirements and 2) to "back out" generation from existing, high fuel cost, oil-fired units. In other words, the coal-fired units are available to operate more hours than required to serve the climate-induced energy requirements, and savings can be realized by substituting coal-fired generation for oil-fired generation. This can be seen more clearly in Table 4-12. Here we fill out the cost matrix described in Section 3 which illustrates the economic risks associated with incorrectly anticipating climate change. Box #1 represents the case in which the utility does not plan for climate change and no 10 There is a distinction between firm capacity purchases and economy purchases. Firm purchases must be satisfied on demand by the neighboring utility, and typically are subject to "take or pay" contracts. Economy purchases result from the periodic availability of baseload capacity in excess of customer demands, and mayor may not be available at a particular point in time. Economy purchases can be used to satisfy generation requirements when available, but cannot be counted toward meeting capacity (reliability) requireme~ts. 4 - 20 From the digital collections of the New York State Library. TABLE 4-11 SOUTHEASTERN UTILITY POTENTIAL IMPACTS OF CLIMATE CHANGE ON FUEL UTILIZATION, 2015 (Btu x 1012 ) Increase Due to Climate Change Long-Term Plan Short-Term Plan Fuel Type Base Case Nuclear 58.8 0 0 0 247.6 -2.7 -1.7 +31.6 Oil Natural Gas Scenario 76.7 A Scenario C Scenario C 0 0 0 Coal 417.7 +17.4 +40.0 0 Purchases/Other .az,z --::U ~ +1.0 Total 888.0 +9.3 +29.8 +32.6 +9 +62 Total Fuel Costs* * 3267 Millions of 1985 dollars. 4 - 21 From the digital collections of the New York State Library. +217 FIGURE 4-5 SOUTHEASTERN UTILITY CHANGE IN ANNUAL PRODUCTION COSTS FOR 2015 INDUCED BY CLIMATE CHANGE 320 _ c=J Capital Costs Fuel Costs 112 280 240 200 Millions of 19855 160 120 80 61 40· 0 • Scenario A '=: I Scenario C " "" Base Case Load Growth Scenario A "'tn Scenario C . y: High Load Growth 4 - 22 From the digital collections of the New York State Library. climate change occurs. This is the base case and indicates no additional capital and fuel costs due to climate change. Box #4 represents correct anticipation of and long-term planning for the high climate change outcome. The total, annual costs of $212 million (1985$) are the same as presented in Figure 4-5. Box #2 is the case in which the utility does not plan for climate change, but the high climate change scenario occurs. To meet its reliability constraints the utility builds 1,417 MW of short lead-time capacity,ll and satisfies the increased energy requirements through economy purchases and increased utilization of existing oil- and gas - fired units. Total costs in 2015 in this case are $267 million. Finally, Box #3 is the case in which the utility plans for high climate change conditions, but no climate change occurs. The capital costs are the same as in Box #4, but significant fuel savings are fossible. Thus, the total costs are lower than in either Box #2 or Box #4. 2 This situation is typical of some but not all utilities. 1 3 It suggests that, ceteris paribus, total costs would be less for the Sputheastern utility if its long-term plans included the potential for "overbuilding" in anticipation of climate change rather than "underbuilding. n The costs presented in Table 4-12 represent the total costs from a societal perspective. That is, in Box #4 the utility planner has anticipated climate change correctly, yet societal costs are estimated to be $212 million higher in 2015 than they would have been in the absence of climate change. Figure 4-6 presents these costs from a somewhat different perspective. From the utility planner's perspective, if he plans for climate change over the long run and it does occur, he has planned correctly, even if societal costs have increased by $212 million annually. What is important to the planner is the risk associated with his relative ability to anticipate future conditions. Thus, both the top branch and the bottom branch in Figure 4-6 have a potential zero cost outcome from the utility planner's perspective. If the utility plans for climate change and it does not occur, the additional cost is $11 million (representing the difference between Boxes #1 and #3 in Table 4-12). The utility will have built more capacity than is necessary to meet its demands for electricity. The remaining branch in Figure 4- 6 represents the case in which the utility does not plan for climate change and it does occur. In this case, the additional costs are $55 million in 2015 (representing the difference between Boxes #2 and #4 in Table 4-12). 11 If it did not build any new capacity, its reserve margin would fall to 7.4%, resulting in a much less reliable system to satisfy its customers demands. 12 The utility's reserve margin is 23.1% in this case. 13 The results for the New York case study reported in Section 5 will also indicate the potential for significant fuel cost savings in that state. 4 - 23 From the digital collections of the New York State Library. TABLE 4-12 SOUTIlEASTERN UTILITY IMPACT OF ALTERNATIVE PLANNING ASSUMPTIONS ON TOTAL ELECTRICITY PRODUCTION COSTS IN 2015 (millions of 1985$) Build Combustion Turbines in Response No Climate Change Temperature Scenario C 1. 2. 0 Fuel Cost Capacity Cost Total Build Combined Cycle Plants in Anticipation _0_ +217 + 50 0 +267 4. 3. Fuel Cost Capacity Cost -139 +150 + 62 +150 Total + 11 +212 4 - 24 From the digital collections of the New York State Library. FIGURE 4-6 COST DlPAGr ftl(I( SOUTHEASTERN UTILITY UTILITY PLADDIG PlRSPEcrIVI - - 2015 (1985$) Plan for Climate Change? Does Climate Change Occur? Cost of Not Anticipating Climate Chana_ Correctly o Yes Yes ~o $11 million Yes $55 million ~o o No 4 - 25 From the digital collections of the New York State Library. From the digital collections of the New York State Library. SECTION 5 NEW' YORK CASE STUDY INTRODUCTION AND OVERVIEW! New York State represents a wide diversity of economic activity and uses of electricity. Like much of the Northeast, the New York economy was hit particularly hard by the energy crisis and subsequent economic recessions of the 1970' s and early 1980' s . The State is recovering and economic growth is returning. However, the character of the State's economy continues to shift more towards services and away from heavy industry. A stable population and stable economy suggest low to moderate growth rates in future electricity consumption. In contrast to the Southeastern case study utility, the current distribution of electricity use is much more even among residential, commercial, and industrial customers. ~ The distribution of 1984 electricity sales is presented in Table 5-1. A critical component of this distribution, for purposes of our analysis, is the market saturation of weather-sensitive electrical end-uses. First, industrial uses of electricity tend not to be weather- sensi tive. The high portion of total sales associated with these customers limits the overall sensitivity of electrical use to changes in weather conditions. Second, the saturation and use of air conditioning equipment certainly is lower in New York than it is in the Southeastern U. S., but air conditioning does constitute a significant electrical end-use among residential and commercial customers in the Southeastern (or "Downstate") section of New York. Summer temperatures and humidity levels are lower in other sections of the State ("Upstate"), and air conditioning utilization is less. As discussed in a later section, higher summer temperatures associated with future changes in climate could increase the saturation of air conditioning in New York. Third, high electricity prices have resulted in limited use of electricity for space heating in New York. However, existing heating loads tend to be relatively higher in the colder, Upstate region. Therefore, the response to warmer winter temperatures in the future would be somewhat more concentrated in this region. Utilities in New York tend to have a wide variety of generating capacity. The distribution of utility capacity and electrical generation by plant type for 1985 is presented in Table 5-2. The increase in oil prices in the 1970s 1 Much of the utility data and many insights on the utility planning process in New York were provided to us by members of the staff of the New York Power Pool Planning Committee. We wish to thank them for their assistance. 5 - 1 From the digital collections of the New York State Library. TABLE 5-1 NEW' YORK STATE DISTRIBUTIOR OF 1985 SALES BY TYPE OF CUSTOMER (percent) Residential Commercial Industrial Other Source: 30.0 35.3 26.1 9.6 100.0% "Statistical Yearbook of the Electric Utility Industry Industry, 1985," Edison Electric Institute. TABLE 5-2 NEW" YORK STATE DISTRIBUTION OF 1985 CAPACITY AND GENERATION BY PLANT TYPE (percent) Nuclear Coal Steam-Gas* Steam-Oil* Combustion Turbine Hydro Purchases Cogeneration Pumped Storage Off-System Sales Total * Generation 19.1 19.1 12.9 20.4 0.0 22.3 14.2 0.8 -0.8 -8.0 100.0% Capacity 11.6 12.6 1.3 44.4 11.2 12.6 3.8 0.4 3.1 -0.8 100.0% These units can burn either gas or oil. The distinction in this table is based upon primary fuel consumed in 1985. Source: "New York Power Pool Long Range Plan: Electric Supply and Demand, 1986-2002," Planning Committee, April 1986. 5 - 2 From the digital collections of the New York State Library. and early 1980s made much of the oil-fired capacity uneconomic. 2 Combined with negative or low, positive growth in demand, some utilities in the State have high reserve margins, and oil generation is at the margin. As indicated in Table 5-2, hydro generation accounted for almost one-fourth of total electricity generation in the State in 1985. As suggested by the stream flow analysis presented in Section- 2, the potential impact of climate change on the availability of stream flow to generate electricity could have significant implications for electricity supply in New York State. An important characteristic of the energy supply situation in New York is its uneveness. In particular, much of the lower cost coal, nuclear, and hydro generation is located Upstate, and much of the higher cost oil-fired generation is located Downstate. Further, growth rates in demand have been lower Upstate than Downstate such that the Upstate utilities have higher reserve margins and capacity available to supply off-system sales. Available generating capacity is becoming a constraint on system reliability for some utilities Downstate. There are transmission lines which link the two regions of the State, and significant transfers of electric energy do occur from Upstate to Downstate. Inter-utility transfers are facilitated through economic dispatch of generating units statewide by the New York Power Pool (NYPP). However, the limited capacity of the transmission lines constrains the amount of transfer and reduces the opportunities for reducing the costs associated with the imbalance in location of 1) lower cost electricity supply, and 2) higher growth in electricity demand. Because of the contrasting characteristics of electricity demand and supply in New York, our analysis of the potential impacts of climate change on electric utilities in the State considers the Upstate and Downstate regions separately. We account for interactions between the regions in our utility modeling (i.e., central. dispatch of generating units), and report results for Upstate utilities, Downstate _ utilities, and the State as a whole. Table 5-3 indicates the division of the State's utilities into the two regions. BASE CASE UTILITY PLAN Demand Growth Forecast growth rates in electric energy demand by utility for the period through 2002 were made available by NYPP. These forecasts were extrapolated to 2015 by rCF. The forecasts are summarized in Table 5-4. Comparing these forecasts with those for the Southeastern case study utility indicates lower forecast growth rates in New York through 1995, and somewhat higher growth rates in the out-years. Within New York, the growth rates are higher Downstate, particularly before 1995. 2 Falling oil prices in recent months has resulted in oil displacing coal generation from older, less efficient units for some of the State's utilities. 5 - 3 From the digital collections of the New York State Library. TABLE 5-3 NEW YOPJ{ STATE COMPOSITION OF UPSTATE AND DOWNSTATE UTILITY REGIONS Upstate Utilities New York Power Authority - Upper New York New York State Electric and Gas Niagara·Mohawk Rochester Gas and Electric Downstate Utilities Central Hudson Gas and Electric Consolidated Edison Long Island Lighting New York Power Authority - Southeastern New York Orange and Rockland TABLE 5-4 NEW' YORK STATE AVERAGE ANNUAL GROVTH RATES IN ELECTRICAL DEMAND (percent) 1984 - 1995 1995 - 2015 Upstate Downstate 1.09 1.89 1.51 1.59 State System Peak State System Energy 1.58 1.49 1.56 1.55 Source: 1984-2002: 2003-2015: New York Power Pool IeF Incorporated estimate. 5 - 4 From the digital collections of the New York State Library. As discussed in Sections 3 and 4, utility system load shapes are important for planning purposes to characterize the seasonal and daily patterns of customer demands. Load shape data for 1984 by utility were made available by NYPP, and these were aggregated into Upstate and Downstate regional load shapes. These load shapes (on a normalized basis) were assumed to remain constant over time in the absence of climate change in ~ establishing the base case utility plan. J • Differences in the sizes and the load shapes of individual utilities lead to differences in the growth rates of energy consumption and peak demand on a statewide basis. The bottom two rows in Table 5-4 indicate that the State system peak demand is forecast to grow slightly faster than State system energy consumption. - Annual system peak demand is an important characteristic determining the capacity requirements to maintain a reliable electrical system. An important feature in New York for our analysis is that the Upstate region typically achieves its system peak in the winter season while the Downstate region is summer peaking. This means that a trend toward warmer temperatures across seasons would tend to reduce capacity requirements Upstate and increase capacity requirem~nts Downstate in comparison with no change in temperatures. These potential impacts are estimated below. As described in Section 4, utility load forecasts are uncertain. Actual loads in the future could be higher or lower than forecast. To test the sensitivity of our results to the assumed underlying growth rates in electricity demand, a modified base case (i.e., assuming no climate change) and climate change case were developed assuming growth in electricity demand averaging 2.3% per year after 1995. This rate approximates the average of several forecasts of electricity growth for the U.S. as a whole. Resource Plan The New York electricity supply system is not optimized to existing energy market conditions. Conceptually, there are opportunities over the long- term to make economic investments in new generating and transmission capacity to reduce the costs of electricity production in the State. In particular, from an economic point of view it might be desirable to replace old, inefficient coal units and oil- fired capacity with new, coal-based technologies. However, financial and environmental constraints as well as uncertainties inherent in load forecasts limit the ability of utility planners and managers to make "wholesale changes 11 in electricity supply systems. Rather, movements toward a more optimal supply can be made only over time. Everything else equal, the lower the growth in demand, the longer it would take for a utility system to move from a suboptimal to an optimal position. 3 Use of a load shape model in New York to account for the effects of changing customer mix and end-use mix is discussed later in this Section. 5 - 5 From the digital collections of the New York State Library. For our New York case study, utility resource plans through 2002 compiled by NYPP were used in establishing the base case. 4 After 2002, units are added only to meet the Power Pool's criterion for maintaining a reliable level of electricity generating capacity, i.e., a statewide reserve margin of 22% above the combined utility summer peak demand. The long-term plan is to add coal- fired units at a cost of $1500/kw in 1985 dollars. Short- term response to unanticipated growth in peak demand (e. g., as a result of climate change) is assumed to be met with construction of combustion turbines at $350/kw in 1985 dollars. 5 The base case assumes that generating capacity built Upstate can serve increased Downstate loads, but subject to transmission constraints. There will be some expansion of these transmission links prior to 2000, but no new transmission capacity between the two regions is assumed to be constructed after 2000. Up to 4,500 MW of Upstate capacity is assumed to be available to serve Downstate loads and satisfy reserve margin requirements after 2000. Table 5-5 summarizes the base case generating capacity expansion plan under the base load growth and high load growth assumptions. Fuel Prices The fuel price assumptions are summarized in Table 5-6. Prices through 1990 were taken from NYPP base price forecasts. Real escalation rates are assumed to be the same as those in the Southeastern utility case study, and were taken from the Fall 1985 reF Energy Service. These prices are important determinants of the most economic types of capacity to build as well as the most economic dispatch of available capacity. VEATHER-SENSITIVITY OF DEHAND Section 3 described two alternate approaches for estimating the sensitivity of utility system demand to changes in temperature. One is referred to as the statistical approach and the other as the structural approach. Section 4 detailed the application of the statistical approach in the Southeastern utility case study. Sufficient, detailed data were not available to implement the structural approach for the Southeastern utility. However, based upon work reF independently conducted for NYPP, sufficient data were available to implement both the statistical and structural approaches in the New York case study. Among our findings for New York: 4 Source: "New York Power Pool Long Range Plan: Demand, 1985-2002," Planning Committee, April 1986. Electric Supply and 5 Unanticipated increases in generation (as opposed to capacity) requirements also can be met through power purchases from utility systems outside the State, principally from Canada. This is discussed further in this section. 5 - 6 From the digital collections of the New York State Library. TABLE 5-5 Nml YORK STATE BASE CASE CAPACITY EXPANSION PIAR (aegawatts) Base Load Growth Turbine/ Coal Other Hiih Load Growth Turbine/ Coal Other 1995-2000 0 0 0 0 2000-2015 7,491 0 14,230 0 Source: 1995-2002: 2003-2015: New York Power Pool Long Range Plan, April 1986, rCF Incorporated estimate. ~cit, TABLE 5-6 NEW YORK STATE AVERAGE ANNUAL GROWTH RATES IN REAL FUEL PRICES 011-0.3% Resid* 1990 Price (1985$/MMBtu) Oil-Distillate 308 Coal (1i)* Nuclear 430 229 50 Growth Rates (% per year) 1990-1995 4.3 2.4 1.0 0.0 1995-2015 3.0 3.0 1.0 0.0 * Prices of all grades of oil-resid and coal are assumed to grow at the same rates as 0.3% resid and 1% coal, respectively Source: 1990 prices: New York Power Pool Growth rates: reF Energy Service, Fall 1985 5 - 7 From the digital collections of the New York State Library. Both approaches can produce estimates of changes in load due to climate change by region (Upstate and Downstate), month, type of day (weekday, weekendfholiday), and time of day (tri-hourly time block); Both approaches show similar patterns of sensitivity between regions, months, and time of day; However, the structural approach shows an overall higher sensitivity of demand to temperature changes. These findings are described in more detail below. The steps followed to implement the statistical approach for the Southeastern utility case study were listed and described in Section 4. These steps also were followed for the New York case study.6 In particular for each region, month, type of day, and time block, historical system load and weather data were used to estimate coefficients for the following relationship: I % deviation of load from mean I Ideviation ofl I heating + f3 fJ I degree 2 1 I days I from mean Ideviation ofl cooling I degree I days I from mean I I Note one difference in the functional form of the estimated relationships. The independent variables for the Southeastern case study were percent deviations in weather variables, and absolute deviations were used in New York. The difference was necessitated by the lower winter temperatures in New York: small absolute temperature changes would result in large percentage temperature changes and instability in the parameter estimates using the Southeastern case study approach. Coefficients were estimated using ordinary least-squares regression applied to utility load data and NOAA tri-hourly weather observations from 1981 to 1985. The load and weather data were provided to us by NYPP. As with our findings for the Southeastern case study utility, the estimated coefficients exhibited patterns which are consistent in sign and magnitude with expected behavior. The data and techniques used in these analyses imply that the estimated coefficients reflect historical appliance and equipment saturations, energy efficiency, and patterns of weather-sensitive appliance and equipment utilization. 6 A weather-normalization study was available for the Southeastern utility and was used to estimate the weather-sensitivity of monthly energy demands. The reF statistical approach was used to develop relationships to distribute these demands to day-types and tri-hourly periods within the months. A similar weather-normalization study was not available for the New York utilities. In this case the tri-hourly and daily impacts estimated using the IeF statistical approach were aggregated to develop monthly impacts. 5 - 8 From the digital collections of the New York State Library. The structural approach was implemented through data development and application of the Hourly Electric Load Model, or HELM (see Section 3), to each New York utility. The essence of HELM is to develop hourly loads associated with individual electric end-uses or customer classes, and to aggregate these end-uses and classes to the utility system level. HELM includes model routines which permit the development of weather response functions (using techniques similar to our statistical approach for estimating weather sensitivity) for individual end-uses and classes. In our application of HELM for the New York utilities, we modeled the following loads as weather sensitive: Residential space heating; Residential room air conditioning; Residential central air conditioning; and the Commercial class as a whole. Weather response functions were developed for each which permitted us to estimate the impacts of changes in temperature on the levels and patterns of these loads. An important attribute of HELM is that it permits the analyst to make explicit assumptions regarding future changes in customer mix or in the saturation or utilization of electric end-uses, thereby developing estimates of the impacts of these changes on utility system load shapes. For our analysis of climate change, we estimated the potential impacts of the temperature change scenarios from Section 2 under two assumptions of future residential air conditioning saturation. Under one assumption, air conditioning saturation was proj ected to increase at a rate which would maintain its current share (about 2%) of total annual electricity consumption across the· total of all end-uses and classes. The second assumption was to increase residential air conditioning saturation at a substantially higher rate such that its share of total consumption would be almost 50% higher (a share of 2.9%) by 2015. The rationale for this latter assumption is that increasingly warmer temperatures due to global warmin~ could induce an increase in the saturation of weather-sensitive appliances. The results of our analysis of the weather sensitivity of demand in New York and the potential impacts of climate change on demand are illustrated in Figures 5-1 through 5-4. Figure 5-1 shows the estimated impacts of the high climate change scenario C (i.e., the "high" temperature scenario), for an August weekday in 2015. These estimates were developed using the statistical approach for estimating weather sensitivity. As would be expected, the results indicate a much greater sensitivity to higher summer temperatures Downstate than Upstate. The percentage impacts average about 3% across the tri-hour1y periods Downstate as compared with impacts of less than 1% Upstate. This reflects the much higher saturation and utilization of air conditioning equipment Downstate. 7 Recall from our discussion of the Southeastern utility that air conditioning saturation already is in excess of 90% of all customers. 5 - 9 From the digital collections of the New York State Library. FIGURE 5-1 REV YORK STATE IHPACT OF TEKPERATOU CHANGE ON ENERGY DIIWlID BY TDIB OF DAY - - VDKD&Y, AUGUST 2015 5 - 10 From the digital collections of the New York State Library. The result that the percentage impacts are greater during night and early morning hours may appear curious and counter-intuitive. However, there are two factors that explain these patterns. First, the figures are in percentage terms. If the scale were "NW" rather than "''', the figure would show somewhat higher megawatt impacts during the daytime hours. This explanation is consistent with our finding for the Southeastern utility as reported in Table 4-6. The second factor is that a very high percentage of air conditioners in New York are room air conditioners as opposed to central air conditioners. The patterns of utilization of room units is different from central units. Room units are used much more frequently and heavily during "at-home" and sleeping hours in New York, while central units tend to be operated more evenly across hours of the day. Figure 5-2 presents the estimated changes in total monthly energy demand in 2015, again using the climate change scenario C and the statistical weather-sensitivity approach. The figure indicates the expected monthly pattern of load changes associated with increasing temperatures in New York: increasing energy demands in the summer months and decreasing energy demands in the winter months. The swings are much greater Downstate, but this primarily reflects the higher summer air conditioning loads; the winter impacts on a percentage basis are comparable between the two regions. On an annual basis, Figure 5-2 indicates that the high climate change scenario would reduce energy requirements Upstate and increase them Downstate. These two effects tend to offset each other on an annual basis. The net annual figure is a useful measure for determining the impacts on some operational requirements such as total fuel utilization. But, from the point of view of utility planning and systems operations,the impacts of climate change on the patterns of demand within the year are more important. In this regard, the periods of increasing demand Upstate and Downstate tend to coincide. Thus, the patterns tend to exacerbate rather than mitigate each other. J Figure 5-3 reports the impacts on annual energy demand for the State as a whole in 2015 across alternate temperature scenarios and weather sensitivity estimating techniques. Using the statistical approach, the change in annual energy demand across the three temperature change scenarios developed in Section 2 ranges from 82 GWH to 255 GWH. The upper end of the range represents only a 0.1% increase compared with the base case. Using a similar set of assumptions and temperature scenario C, the structural approach results in an estimated impact of 621 GWH, almost 2.5 times higher than the statistical approach. No obvious inconsistencies or discrepancies lead to this result. Rather, they follow from the different modeling techniques and different data sets we used. More diagnostic work would be required to reconcile the differences. However, our judgment is that the structural approach and its more detailed treatment of weather-sensitive end-uses potentially offers opportunities for more meaningful analysis of climate change impacts. 5 - 11 From the digital collections of the New York State Library. FIGURE 5-2 NEW YORK STATE IMPACT OF TEHPERATORE CHANGE ON ENERGY DIIfAlU) BY IIOlIl1I IR 2015 3 Upstate 2 1 'fO Chance 0 -1 -2 Annual Month J F M A M J J. A S 0 N D -3 3 Downstate 2 1 ~ Chance 0 . I I -1 Annual Month -1 J F M A M J J A S 0 N D -3 5 - 12 From the digital collections of the New York State Library. fiGORB 5-3 NEW YOH STATE DlPAcr OF TlllPIIATUU CBARGI OR SYSTEK ENIRGY DBKAND IN 2015 900 100 100 Chanl' In EnerlY Demand (GWII) '21 600 500 300 lSS 200 130 100 Temperature Scenarios: 82 04.-.....--...................---c B A C Ale Satantloa C AlC Saturation Constant Increasinl "t y Statistical Approach Structural Approach 5 - 13 From the digital collections of the New York State Library. ? Figure 5-3 also indicates the potential impacts of increasing air conditioning saturation on system energy demand. The assumed increase in residential air conditioning saturation using the structural model increased the total climate change impact from 621 GWH to 886 GWH, an increase of approximately 45%. A similar set of bar charts is included in Figure 5-4, which shows the imp ac ts 0 f the a 1 te rna te temperature change scenarios and weather-sensitivity modeling approaches on system peak demand. Using the statistical approach, the alternate temperature scenarios result in peak demand increases of 272 to 612 MW. The structural approach leads to an estimated increase in peak demand of 949 MW (about 50% higher than the statistical approach) assuming a constant relative saturation of air conditioning, and an 1,171 MW increase assuming an increase in air conditioning saturation. This latter figure is about a 3.3% increase in peak demand compared with the base case forecast. The impacts of temperature change on system peak demand were estimated alternately assuming high load growth rates (2.3% per year). These estimates are approximately 15% higher than under the base load growth rates. Using the structural approach and assuming temperature scenario C and increasing air conditioning saturation, results in an estimated increase in peak demand of 1,352 MW. WEATHER-SENSITIVITY OF SUPPLY As described in Section 3, climate change potentially can affect electricity supply in two ways: 1) the effective capacity and/or operating efficiency of thermal generating units and 2) the availability of stream flow to provide hydro generation. The analysis reported in Section 4 based upon modeling of thermal unit efficiency by the Southeastern utility, indicated that the potential impact of climate change on thermal units is negligible. We accept these resul ts as they might affect thermal uni t generation in New York given that the order of magnitude of temperature changes in the New York scenarios is similar to that in the Southeast. However, climate change potentially can affect hydro generation in New York as a result of the stream flow analysis reported in Section 2. That section reported future temperature changes could reduce water availability and stream flow in the Great Lakes Basin. Hydro plants supplied about 22% of New York's electricity requirements in 1984. The New York Power Authority's Niagara and St. Lawrence hydro proj ects accounted for 85% of this supply. These proj ects are tied directly to stream flow through the Great Lakes Basin. The relationships between stream flow and hydro generation were quantified to estimate the potential impacts of climate change on this important source of electricity supply in New York. Figures 5-5 and 5-6 present plots of historical stream flow data for the Niagara and 5 - 14 From the digital collections of the New York State Library. FIGDRI 5-4 NEW YORK STATE DlPACr OF TIIIPBIIATUI.B aIABGB OR SYSTFJI PEAK DEIfAND IN 2015 1111 1200 lOGO H' Chanl_ 800 In Peak Demand (MW) '-12 600 - 393 272 %00 0 Temperature Scenarios: A B "t ¥ C AlC Satun dOD Constant C t Ale C SaturatloD IDcreasinf" "en ¥ Statistical Approach Structural Approach 5 - 15 From the digital collections of the New York State Library. FIGURE 5-5 STREAK FLOW AND ELECTRICITY GENERATION NIAGARA HYDRO PROJECT 20----------------------------, 10 oL-------#-~~4_-------------_, Deviation from Mean -10 (o/c) -20 -30 -40L-----------------------:-:::--.. 1977 1982 1972 1967 1962 5 - 16 From the digital collections of the New York State Library. FIGURE 5-6 STREAK FLOW AND ELECTRICITY GENERATION ST. lAWRENCE HYDRO PROJECT 10 O-r----------::iI~-.l~____:IJ~---------IIt..------------.I Deviation from Mean -10 (%) -20 -30 -40.,....-------------...,..------------------J 1972 1977 1967 1982 1962 5 - 17 From the digital collections of the New York State Library. St. Lawrence rivers and electricity generated from the hydro projects on these rivers. 8 Visually, one can see a very close relationship between the two series in each figure. The changes in stream flow match changes in generation closely. These annual time-series data were used to estimate coefficients for the following relationship: In I I I Istreaml I flow I hydro I I generation I I a + P ln I I I I The coefficient p of this model indicates the estimated sensitivity of hydro generation to changes in stream flow. Ordinary least-squares regression coefficients. The results are as follows: Niagara St. Lawrence was used to estimate Standard Error Adjusted R2 0.06 0.03 0.96 0.95 1.43 0.74 the These results indicate a high statistical significance for the estimated coefficients and that the model specification explain 95-96% of the variation in the data series. The greater sensitivity of generation at Niagara (i.e., the higher coefficient) than at St. Lawrence is evident from comparing the turning points in the data series in Figures 5-5 and 5-6. For purposes of our analysis, the estimated coefficients for the two hydro projects were weighted based upon mean annual generation to produce a system sensitivity coefficient of 1.22. The interpretation of this coefficient is that a 1% change in stream flow is expected to result in a 1.22% change in hydro generation. This coefficient was applied to the alternate stream flow scenarios described in Section 2. The results are illustrated in Figure 5-7. In 2015, the reduction in stream flow could reduce hydro generation in New York by 6.2 to 8.5%, or 1,500 to 2,066 GWH. Given the importance of hydro generation in the State and the fact that it is New York's least expensive source of power, this is a very significant result. The reduction in hydro availability implies that electric energy requirements would have to be supplied from other, more expensive resources. The estimated increase in costs is presented below. 8 Data provided by W.P. Palazzo memorandum to reF, September 10, 1986. of New York Power 5 - 18 From the digital collections of the New York State Library. Authority in FIGURE 5-7 NEW YORK STATE POTENTIAL DlPAcr OF CHARGES IB STIUWI FI.l7J OR HYDRO GENERATION IN 2015 x Stream Flow Scenario y 0.,...----....,.-.,.------..-.....----- z Chanle In Hydro Generation 1000 (GWII) -1500 -7.1" -1705 %000 -1"' 5 - 19 From the digital collections of the New York State Library. ALTERNATE PLANNING SCENARIOS The previous sections have discussed uncertainty and variation in a number of factors which could influence electric utility planning in New York for the future. These include alternate: temperature change scenarios; stream flow scenarios; weather sensitivity modeling approaches; assumptions of future saturation of weather- sensitive appliances and equipment; and growth rates in electricity demand. In addition, there are issues associated with the costs and risks associated with anticipating climate change correctly and the utility's investment response to these changes. There are a large number of utility planning scenarios that could be developed and analyzed across all these dimensions. In order to limit the scenarios for which we report results, our main discussion in the next subsection centers around: base case load growth; short- term utility supply response to climate change; and four alternate case assumptions as illustrated in the following table. Weather Sensitivity Model Statistical Approach Structural ARproach Climate Change Assumptions Temperature Scenario Stream Flow Scenario A/e Saturation C Base N/A C C Z N/A Base Constant C Z Increasing These four cases indicate that we assume the "high" temperature change scenario in all cases, but vary the stream flow asswnption (base or "high change" scenario Z), the residential air conditioning saturation assumption J and the estimated weather sensitivity of demand resulting from our application of both the statistical and structural approaches. The fourth column describes a scenario in which the largest changes in temperature (C) and stream flow (Z) are assumed to occur, and demand response is estimated at its highest level. This will be referred to as the "high impact" case. In addition to these cases, we report the results of sensitivity analysis in which we assume (1) the high load growth rates previously discussed, and (2) that utilities anticipate climate change and make capacity expansion decision over the long term. Our view is that recent historical and current utility supply and financial conditions in New York would make it difficult to justify construction of long lead-time, capital intensive generating plants in anticipation of climate change which is very uncertain. In other words, despite possible economic advantages to a long term response to climate change, the perception among utility planners and managers is that there are severe risks and penalties associated with "overbuilding. n Our assumptions of short term response to climate change 5 - 20 From the digital collections of the New York State Library. (i. e . , construction of combustion turbines) is considered as the likely response, and long-term anticipation of climate change is treated in sensitivity analyses. Results for both of these situations are reported and compared in the next subsection. UTILITY IMPACTS Our estimated impacts of climate change in New York State in 2015 are summarized in Table 5-7. We indicate the estimated change in peak demand, capacity requirements, non-hydro generation, and electricity production costs for four cases assuming base load growth rates and for one sensitivity case assuming high load growth rates. Climate change scenario C ("high temperature change) results in an increase in summer peak demand for the State of 611-1,171 MW under base load growth conditions and up to 1,352 MY under high load growth conditions. Adding a reserve margin of 22% to satisfy electrical system reliability requirements, translates to an increase in generating capacity requirements of 746-1,429 MW under base load growth conditions and 1,649 MW under high load growth conditions. As discussed in more detail below, . these are significant changes in requirements, particularly when compared with estimated capacity additions over the period in the base case (i.e., no climate change). II The impact of climate change on stream flow and, therefore, the contribution of hydro generation can be seen in the next row of the table. Under base load growth and base stream flow assumptions, generation requirements from thermal and other non-hydro sources increase by 225 GWH on a statewide basis. Additional tables presented below indicate how t-hese requirements vary Upstate vs. Downstate. Under assumed stream flow scenario Z (i.e., high climate change impact), however, non-hydro generation requirements increase by 2,313 GWH, an. addition of over 2,000 GWH. This increases further to 2,944 GWH under base load growth, "high impact" assumptions and 3,073 GWH in our high load growth sensitivity case. The importance of these potential reductions in hydro generation is emphasized in our estimated cost impacts. Changes in the cost of electricity production in 2015 induced by climate change are estimated to range from $48-241 million under base load growth conditions and up to $248 million assuming high load growth. (These costs are in constant, 1985 dollars.) Importantly, however, changes in fuel costs account for about half the total change in costs assuming base stream flow conditions, but 80% or more of the cost impact assuming stream flow scenario Z. In essence, the estimated reduction in hydro generation must be replaced by higher cost generation from marginal, thermal units or by purchases from other utility systems outside the state. Table 5-8 provides more detail on the estimated impact of climate change on generating capacity requirements in 2015. The first column in the table indicates that under base case planning assumptions, load growth in the State would lead to requirements for an additional 7,491 MW of 5 - 21 From the digital collections of the New York State Library. TABLE 5-7 NEW YORK STATE SUMMARY OF IMPACTS IN 2015 Temperature Scenario: Stream Flow Scenario: Statistical Approach C C Base Z C Base Structural Approach High Load "High Growth Impact" Peak Demand (MY) Capacity Requirements (MW) +611 +746 +611 +746 +949 +1158 +1171 +1429 +1352 +1649 Non-Hydro Generation (GWH) +255 +2313 +621 +2944 +3073 +22 +26 +48 +148 + 26 +174 +46 +41 +87 +191 +50 +241 +190 + 58 +248 Annual Cost of Climate Change (Millions of 1985$) Fuel Cost Capacity Cost Total Cost 5 - 22 From the digital collections of the New York State Library. TABLE 5-8 NEW YORK STATE DlPACl OR CAPACITY REQUIREKENTS IN 2015 Temperature Scenario: Stream Flow Scenario: Base Base Additional Requirements Induced by Climate Chanie (MW) Structural Approach Statistical Approach High Load C C C "High Growth Base Impact" Base Z 160 155 155 267 349 403 Downstate* 7331 591 591 891 1080 1246 System 7491 746 746 1158 1429 1649 10.0% 10.0% 15.5% 19.1% 11.6%** Upstate* % Change * Based on location of load, not location of power plants. ** Based on additions of 14,230 MY in the high load growth base case. 5 - 23 From the digital collections of the New York State Library. generating capacity. These requirements are concentrated Downstate because of 1) higher load growth rates Downstate and 2) very high current reserve margins Upstate. The second factor dominates: Upstate utilities are projected to be able to serve large increases in load growth from existing capacity. Additional capacity reqtlirements induced by climate change also are concentrated Downstate. This results from the higher saturation and use of air conditioning equipment in that region, and the subsequent greater weather sensitivity of demand. The concentration of additional generating capacity requirements Downstate is problematic because of greater difficulty in siting new generating units for environmental and safety reasons and because of potentially higher construction and operating costs. In other words, climate change would tend to exacerbate the situation in the State in which there is an imbalance between the location of new demands for electricity and the location of existing or new economic generating capacity. Among the possible implications is that climate change could contribute to arguments for construction of increased transmission capacity between Upstate and Downstate. The bottom row of Table 5-8 indicates that additional generating capacity requirements induced by climate change over the next 30 years could be as much as 10 - 20% higher than under base case condi tions . This is significant and worthy of consideration by utility planners. However, possible changes in climate represent only one set of uncertainties planners must face. The results of the high load growth sensitivity case indicate the potentially large impacts of uncertainty in demand growth tends to enhance rather than mitigate the impacts of these uncertainties. Table 5-9 details the potential changes in generation requirements in 2015 induced by climate change. Assuming there is no change in stream flow and hydro generation (i.e., base stream flow scenario), increasing temperatures decrease the annual generation requirements Upstate where a significant portion of weather-sensitive demands occur in the Winter. These reductions help to offset increasing requirements Downstate. As a result, temperature scenario C in the absence of any changes in hydro availability leads to an increase in generation requirements statewide of less than 0.5%. Under stream flow scenario Z, however, generation requirements from non-hydro units increase both Upstate and Downstate. The total impact in New York of these latter cases could be an increase on the order of 1.3 to 1.7%. These percentage increases are measured relative to total generation in the State estimated for 2015. The percentages of total additional requirements over the next 30 years (which may be a more relevant consideration for utility planners) are much larger: 3.0 to 3.9%. Table 5-10 indicates that the increased generation requirements induced by climate change are met by increased utilization of existing, high cost oil-fired units and increased purchases of electricity from systems out of the State. Both of these sources are utilized even more heavily assuming reductions in stream flow and hydropower availability in New York. 5 - 24 From the digital collections of the New York State Library. TABLE 5-9 NEW' YORK STATE IMPACT ON NON-HYDRO GENERATION REQUIREMENTS IN 2015* Temperature Scenario: Stream Flow Scenario: Base Base Change Induced by Climate Change (GWH) Statistical Approach Structural Approach C High Load C C "High Base Base Impact" Growth Z Upstate 77364 -253 +1162 -172 +1219 +1185 Downstate 96597 +508 +1151 +793 +1725 +1888 173961 +255 +2313 +621 +2944 +3073 0.1% 1.3% System % Change * Assumes that 37% of ** Based on non-hydro 0.4% 1.7% 1.5%** NYPA's hydro generation is consumed Downstate. generation in 2015 of 205,394 GWH in the high load growth base case. 5 - 25 From the digital collections of the New York State Library. TABLE 5-10 NEW YORK STATE IMPACT ON GENERATION MIX IN 2015 Temperature Scenario: Stream Flow Scenario: Base Base Change Induced by Climate Change (GWH)* Structural Approach Statistical Approach C C High Load C "High Base Base Growth Impact" Z Purchases 16783 +137 +1441 +312 +1402 +1806 Cogeneration 10528 0 0 0 0 0 Oil 33461 +122 +877 +315 +1544 +1263 Coal 81141 0 0 0 0 0 Nuclear 32227 0 0 0 0 0 Hydro 24186 0 -2058 0 -2058 -2058 -Subtotal Transmission Losses Total * 198325 +259 +260 +627 +880 +1011 -177 --::..!:± _ _-_5 __-_6 +6 ---±A 198148 +255 +255 +621 +886 1015 Assumes no change in availability or price of purchases in alternative cases. 5 - 26 From the digital collections of the New York State Library. A note of caution is offered here. Purchases by New York utilities are poss-ible because of the existence of excess generating capacity at other utilities or different patterns of seasonal demands exhibited by these utilities' customers. An important source of purchases for New York is hydro power imports from Canada. Logically, if climate change is affecting the availability of excess generating capacity at other utilities in ways similar to those in New York, the increased generation requirements in New York may not be able to be satisfied - - even in part - - by purchases. Further, many of these purchases enter the State from the North, limiting the amount of energy that can be transmitted to the higher load growth areas Downstate. Under these conditions an even greater amount of the State's generation would come from high cost oil-fired units. The cost.impacts of climate change, therefore, would be even higher than indicated in the following table. Table 5-11 presents our estimates of the potential impacts of climate change on electricity production costs in 2015. The costs are presented in constant 1985 dollars. For the four cases, assuming base load growth assumptions, the range of estimated annual cost impacts ranges from $48 million to $241 million. The annual cost in the high load growth sensitivity case is $248 million. Reflecting the discussion of capacity and generation requirements in the paragraphs above, the estimates in Table 5-11 indicate 1) the higher costs associated with climate-induced changes in demand Downstate, and 2) the importance to the total cost impacts of changes in generation requirements and the availability of hydro power. The largest increases in costs are associated with the assumed occurrence of stream flow scenario Z. We have assumed that increased capacity requirements induced by climate change would be satisfied by construction of short lead-time combustion turbines which are gas- or oil-fired. What if, however, these increased requirements could be anticipated and met with longer lead-time, higher capital cost, but lower fuel cost coal-fired units? Table 5-12 reports the results of a sensitivity analysis conducted by varying our planning assumption and allowing New York utilities to build coal-fired combined cycle units in anticipation of climate change. The statistical approach to estimating weather-sensitivity of demand was used in this sensitivity analysis. Similar to our analysis for the Southeastern utility, we compare the costs associated with building plants in anticipation of climate change which mayor may not occur. The upper portion of Table 5-12 repeats some of the results from Table 5-11. Here, it is assumed that short lead-time combustion turbines are built. In the case of no climate change occurring, no new capacity is built and no additional costs are incurred. The lower portion of the table presents the results assuming construction of coal-fired combined cycle units. Here, capacity costs are significantly higher, but fuel cost savings are even more substantial. Building a coal-fired plant in anticipation of climate change which does not occur result in cost savings of $39 million. Th1s seemingly curious result 5 - 27 From the digital collections of the New York State Library. TABLE 5-11 NEW YORK STATE IMPACT ON TOTAL ELECTRICITY PRODUCTION COSTS IN 2015 (millions of 1985$) Temperature Scenario: Stream Flow Scenario: Change Induced by Climate Change Structural Approach Statistical Approach C C C "High High Load Base Base Impact" Growth Z Upstate Fuel Cost Capacity Cost Total Cost -22 + 5 -17 +74 + 5 +79 -13 +10 - 3 +79 +12 +91 +73 +14 +87 Downstate Fuel Cost Capacity Cost Total Cost +44 +21 +65 +74 +21 +95 +59 +31 +90 +112 + 38 +150 +117 + 44 +161 System-Total Cost +48 +174 +87 ;-L41. 5 - 28 From the digital collections of the New York State Library. • I) J. 1 +248 TABLE 5-12 NEW YORK STATE IMPACT ALTERNATIVE PLANNING ASSUMPTIONS ON TOTAL ELECTRICITY PRODUCTION COSTS IN 2015* (millions of 1985$) Build Combustion Turbines in Response No Climate Change Temperature Scenario C 1. 2. Fuel Cost Capacity Cost Total Build Coal Plant in Anticipation Fuel Cost Capacity Cost Total * Temperature Scenario C Stream Flow Scenario Z +22 +26 +48 +148 +26 +174 -152 +113 -133 +113 -39 -20 -11 +113 +102 0 _0_ 0 4. 3. Estimated using statistical approach to weather-sensitivity of demand. 5 - 29 From the digital collections of the New York State Library. is the direct consequence of the suboptimal economic position of the electricity supply system in New York. Given all of our assumptions about future costs and conditions in the State, our analysis indicates opportunities for economic displacement of existing oil-fired capacity with new coal-fired capacity independent of any potential impacts of climate change. If climate change does occur, however, the differences in total costs between the short-term response (combustion turbines) and the long-term response (combined cycle units) is even greater. The implications of these results are that on an economic basis, the long lead-time, higher capital cost plants should be built. However, environmental and financial constraints and risks are likely to preclude this option. Figure 5-8 is analogous to Figure 4-6 in Section 4. It shows the costs from a utility planner's perspective in light of the relative ability to anticipate and plan for future climate change correctly. Respectively, the top and bottom branches of Figure 5-8 are cases for which the utility builds additional capacity in anticipation of climate change (top) and in response to climate change (bottom). The previous paragraph explained that a suboptimal economic situation of electricity supply in New York could lead to a reduction in costs as a result of building and operating this additional capacity. The savings would be on the order of $39 million in 2015. Correctly anticipating climate change (yes or no) produces a zero outcome from the utility planner perspective. The remaining branch of the figure indicates the costs associated with the utility not planning for climate change that does occur. Two cases are presented: one considering temperature change scenario C but no change in stream flow, and one considering temperature scenario C and stream flow scenario Z. The range of additional costs for these two cases is $68 to $72 million. The range is narrow in comparison. with the estimates of total cost impacts from a societal perspective presented in Table 5-12. This is because climate change impacts on stream. flow are assumed to affect the short-run variable costs of electricity production (fuel and O&M) but not the long-run fixed costs of generating capacity. It is these long-run fixed costs that are the focus of the utility planning decisions described in this report. A broader planning focus including long-term contracting for fuel purchases would expand the range. 5 - 30 From the digital collections of the New York State Library. nGUU 5-8 NEW YOIUC STATE COST IKPACT mOIl UTILITY PLARRDIG PERSPECTIVE - - 2015* (1985$) Plan for Climate Change? Does Clima'te Chang- Occur? Cost of Not Anticipatina Climate Chanae Correctly Temperature Temperatura Scenario C Scenario C Stream Flow Scenario Z o o Yes Yes No -$39 million -$39 million Yes $68 million $72 million No o No * o Estimated using statistical approach to weather-sensitivity of demand. 5 - 31 From the digital collections of the New York State Library. From the digital collections of the New York State Library. SECI'ION 6 CONCIJJSIONS This section contains three subsections: the first presents swnmary tables that compare the results of the Southeastern utility and New York utility case studies; the second subsection discusses several implications of our findings; and the final subsection offers recommendations for additional research. COMPARISON OF CASE STUDY RESULTS Table 6-1 presents estimates of the changes in peak demand and total energy estimated for the Southeastern utility and for Upstate and Downstate New York Utilities under temperature scenario C (high change) assumptions. The table indicates that the estimated percent change in peak demand and electric energy requirements is greater for the Southeastern utility than for New York. This result is a product of: (1) higher estimated summer temperature changes in the Southeastern utility case study (1.87 oF versus 1.46 oF in New York); and (2) higher estimated weather-sensitivity coefficients (1.19 to 2.27 for New York versus 3.76 for the Southeastern utility). The estimates for Upstate and Downstate New York show the range of results obtained from the two modeling approaches (statistical and structural) used in the New York case studies. Recall that the two approaches differ in the data and level of detail used to estimate weather-sensitivity of demand. The estimated sensitivity of peak demand to changes in temperature in Downstate New York is nearly as large as the peak-sensitivity estimated for the Southeastern utility. Given the same change in summer temperature, there would be similar estimates of percent change in peak demand in the two regions. The substantially larger percentage change in total energy consumption for the Southeastern utility (3.4 percent versus 0.13 to 0.45 percent for New York) is related in large degree to the importance of air conditioning loads in all seasons for that utility. In New York, air conditioning is almost exclusively a summertime use of electricity. In fact, total energy requirements in Upstate New York fall in response to the temperature increase (due to reduced winter heating loads). Table 6-2 indicates the estimated increase in generating capacity required by 2015 to maintain system reserve margins in the case studies. The "base requirements" shown in the table, assume base case load growth assumptions and no climate change. Although the total existing generating capacity in New York exceeds that for the Southeastern utility, the estimated additions in the base case are similar for the two regions because reserve margins among utilities in New York are currently very high. Increases in capacity requirements due to climate change are similar in the two regions. The range of additional capacity induced by climate change in New York is 746-1,429 MW and is 1,417 MW for the Southeastern utility, both under temperature scenario C assumptions. The percentage increases compared with base case additions during the period range from 10-19% in New York and is estimated as 21% for the Southeastern utility. 6 - 1 From the digital collections of the New York State Library. TABLE 6-1 COMPARISON OF REV YORK AND SOUTIIFASTERR UTILITY CASE STUDIES: DEMAND SENSITIVITY (Temperature Scenario C) Change in Summer Temp. New York (OF) Peak Sensitivity (%/oF) Change in Peak Demand (%) Change in Total Energy (%) Upstate Downstate 1.46 1.46 0.66 to 1.47 1.51 to 2.77 0.96 to 2.14 2.20 to 4.04 -0.27 to -0.21 0.49 to 1.04 System 1.46 1.19 to 2.27 1.74 to 3.32 0.13 to 0.45 Southeastern Utility 1.87 3.76 7.04 3.40 6 - 2 From the digital collections of the New York State Library. TABLE 6-2 COIIPARISOR OF NEW YORK AND S01J11lEASTERB UTILITY CASE STUDIES: GENERATING CAPACITY REQUIREMENTS (Temperature Scenario C) Additional Requirements Induced by Climate Change Base Requirements New York (MW) Upstate Downstate 160 7331 155 - 349 591 - 1080 System 7491 746 - 1429 (10%-19%) Southeastern Utility 6749 (MW,%) 1417 (21%) 6 - 3 From the digital collections of the New York State Library. The potential impact of climate change on annual electricity production costs (annualized capital costs and annual fuel costs) for the case studies are summarized in Table 6-3. Costs in New York range from $48 million to $241 million in 2015 (1985 dollars). As discussed in Section 5, the cost implications of climate change impacts on hydro generation are significant in New York. The high case costs in 2015 for the Southeastern utility are similar in magnitude, $267 million (in 1985$). These costs represent the potential impacts of climate change from a societal perspective. From a utility planning perspective, the costs associated with uncertainty concerning the utility planner's ability to anticipate future climate conditions ranges from $11 to $55 million for the Southeastern utility and from $39 to $72 million for the New York utilities. IKPLICATIONS OF FINDINGS The discussions of our methods, analyses, and results in earlier sections indicate that there are many uncertainties associated with developing estimates of potential climate change impacts. We have addressed uncertainties in: climate modeling; weather-sensitivity modeling; and other economic, technological, and behavioral conditions. Because these uncertainties make it difficult to predict the future with precision, the results are driven by assumptions about these factors. The relative contributions of key assumptions to the results are illustrated in Figure 6-1 for the New York State "High Impact" case. Recall that this case assumes temperature scenario C and stream flow scenario Z. Regarding the estimates of additional capacity requirements (left-hand bar), use of the data and assumptions in the statistical approach to modeling the weather-sensitivity of demand results in an estimate of 746 MW by 2015. Alternatively, use of the structural approach and assumptions of a constant saturation of air conditioning equipment results in an estimate of 1,158 MW. The additional assumption that air conditioning saturation increases over time pushes the estimate to 1,429 MW. The right-hand bar illustrates the impact of these factors on total annual electricity production costs. This bar emphasizes the importance of the estimated effects of stream flow reduction on hydro generation, and the assumptions regarding the utilities' response to these changes. The substitution of oil generation and off-system electricity purchases for the reduction in hydro power availability accounts for over half (52%) of the estimated total cost impact of $241 million (in 1985$). Smaller increments are attributed to the statistical approach to demand response ($48 million), the structural approach to demand response ($39 million), and the assumption of increased air conditioning saturation ($28 million). Although the results are sensitive to the assumptions about these factors, this situation is little different than forecasting demand, 6 - 4 From the digital collections of the New York State Library. TABLE 6-3 COMPARISON OF NEW YORK AND SOUTHEASTERN UTILITY CASE STUDIES: IMPACl OR TOTAL COSTS IN 2015~1 (millions of 1985$) New York Temp. Scenario C Stream Flow Base Upstate Downstate -17 +65 +79 +95 +91 +150 System +48 +174 +241 Southeastern Utility ~/ Q/ £/ hi Temp. Scenario C Stream Flow Z hi "High Impact" +267 Assumes no anticipation of climate change. Statistical approach used to estimate weather-sensitivity of demand. Structural approach used to estimate weather-sensitivity of demand. 6 - 5 From the digital collections of the New York State Library. £1 FIGURE 6-1 NEW YORK STATE IlKIATIVE CORTlUBOTIOR OF ICEY ASSUllPTIORS -HIGH IKPAcr- REST_TLTS ro (2015) 100 Additional Capacity Requirements 1429 Total Costs 241 .--- b. Ale Saturation .--- Demand Response: Structural Approach 1158 % of -+ .> 213 174 "High Impact" SO Demand Response: Statistical Approach -: -+ Hydro Reduction Additional Capacity Requirements in Megawatts 6 - 6 From the digital collections of the New York State Library. 126 technological change, and customer response to utility conservation or marketing programs. These types of analyses, commonly conducted by utility planners, also involve substantial uncertainties and require many assumptions. Although not precise, the estimated impacts are judged to be reasonable. The findings indicate that the potential impacts of climate change on electric utilities are not insignificant and that these impacts'may start to occur within the typical time frame of current utility planning studies and decisions. Although the case studies have been conducted on different types of utility systems in different regions of the U. S., it is difficult to generalize to regions or the nation as a whole based solely on these results (e.g., consider the different results obtained for Upstate and Downstate New York). However, the analyses suggest the following general conclusions: Climate Change The temperature change scenarios developed for the two case studies indicate that current GeM estimates of potential regional climate change due to a doubling of atmospheric concentration of C02 are quite diverse. The GISS transient run indicates that the rate of climate change may be uneven over time, and may vary subtantially from one location to another. The results of the transient run with the GISS GCM display considerably more diversity than the results of transient analyses performed with the l-D model. This result is consistent with expectations regarding how the climate system would respond to increased greenhouse gas concentrations. In light of the diversity of the GCM results, and the relative inexperience of using GeMs to perform transient analyses, the climate change scenarios must not be considered to be forecasts. Although the scenarios reflect the diversity of current estimates, future climate change outside the range of estimates presented here cannot be ruled out. Utility Impacts It appears that climate change will have greater direct impac ts on the demand for e lec trici ty than on characteristics of the supply of electricity for most utility systems: The impacts resulting from demand response to climate change are more likely to be significant for utilities with large, summer, weather-sensitive (air conditioning) loads. This is true for regions in the southern u. s. where air conditioning saturation and utilization is high, and for urban areas in northern climate zones where the potential for increased air 6 - 7 From the digital collections of the New York State Library. conditioning saturation is high. Because of the nature and patterns of these weather-sensitive loads, response to climate change is likely to have greater impacts on peak demand (capacity requirements) than on energy consumption (generation requirements). The order of magnitude of temperature changes examined here is unlikely to have significant impacts on the effective capacity or operating efficiency of thermal generating units. However, there can be significant implications for utilities where hydro is an important source of generation. As indicated by the New York case studies, hydro generation is critical for some utilities, and the planning uncertainties associated wi th possible climate-change- induced stream flow changes are large. We have found that the utility capacity and cost implications of climate change potentially are significant. In the two case study analyses generating capacity additions induced by climate change are on the order of 10-20% of base case (i.e., no climate change) additions through 2015 under scenario C temperature change assumptions (the highest case examined here), and annualized capital cost and annual fuel and O&M costs induced by climate change exceed $200 million (1985 dollars) in 2015. Because of long lead-times and the capital intensity of the most efficient electric generating units, there are economic benefits associated with being able to anticipate climate change correctly. The magnitude of the potential cost savings depends on the base case planning assumptions, and in these two case studies may be $50 to $70 million per year (1985 dollars) by 2015. Utility planners should start now to consider climate change as a factor affecting their planning analyses and decisions. Large impacts are not imminent, but the importance of climate change impacts for utility planning is likely to increase over time. Climate change is likely to increase the uncertainties utility planners must face and to interact with other issues they must address, including: the level and patterns of future electric demands, the availability resources, and and mix of future generating investment and financial planning. 6 - 8 From the digital collections of the New York State Library. SUGGESTIONS FOR FUR1.'HER. RESEARCH Climate Change Several recent studies have assessed the research needs for developing assessments of potential future climate change, and its impacts.! In performing this study, more precise estimates of potential future changes in a wide variety of climate variables would have been valuable. It is recognized that significant efforts are currently underway to improve GCMs and produce better estimates, as is recommended by these recent reports. To help make the ongoing efforts to improve the GCMs usefulness for performing impact assessments such as this one, several areas of effort should be considered: Improve methods for change scenarios. developing and disseminating climate Develop methods of linking GeM estimates to historical estimates of variables relevant for impact assessment (e.g., hydrologic variables, variability, winds). Improve methods for evaluating potential climate changes over time, particularly at the regional level. Making the current data available, and improving methods for evaluating impacts over time, would enable additional assessments to be performed. These assessments would provide additional feedback to the climate community regarding research priorities. For some locations dependent upon water for hydropower and cooling, water resources are an important consideration. The New York case study demonstrated the need to develop improved methods to estimate hydrologic variables including precipitation, evaporation, soil moisture, and runoff. Modeling Utility Impacts The case study analyses described in this report represent a first attempt at analyzing the potential impacts of climate change on individual electric utilities. Therefore, there are many opportunities for improving the data, 1 Projecting the Climatic Effects of Increasing Carbon Dioxide, M.C. MacCracken and F .M. Luther, ed., U. S. Department of Energy, DOE/ER-0237, Washington, D.C., December 1985. Characterization of Information Requirements for Studies of C02 Effects: Water Resources. Agricu1 ture , Fisheries. Fares ts and Human Heal th, U. S . Department of Energy, DOE/ER-0237, Washington, D.C., December 1985. 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, WMO-No. 661, Villach, Austria, 9-15 October 1985. 6 - 9 From the digital collections of the New York State Library. assumptions, analytic approaches, and estimating methods that have been used. Principal recommendations for further research include the following: Based upon the knowledge and experience gained in conducting the Southeastern utility and New York State case studies, broad, regional and national estimates of potential climate change impacts should be developed. These estimates would provide a broader perspective on the likely importance of climate change to the industry as well as identify opportunities and needs for policy development at the regional or national level. An adjunct to this would be a more complete assessment of the value of improved climate change information to utility planners and managers. As an adjunct to the regional/national analyses, additional detailed studies of individual utility systems should be conducted. Regional studies could "average out" or otherwise mask important impacts of climate change on individual systems. The results in Section 5 indicated some significant differences between Upstate and Downstate New York. Given different demand and supply characteristics, there also could be significant differences in impacts for neighboring utility systems such as Consolidated Edison and Long Island Lighting. A recent paper prepared by IeF and the Edison Electric Institute, found that despite the large number of utility holding companies and regional power pools, the principal focus for utility planning studies and decision-making in the u.S. is the individual utility system. 2 The case study conclusions indicate that the most significant impacts are likely to result in changes in the level and pattern of electrici ty demand of utility customers. The case studies also reflect significant uncertainties in modeling the weather-sensitivity of demand. Further development of data and methods to estimate the sensitivity of demand to climate change over time would be valuable. This would include analysis of weather variables (e.g., extreme temperatures) and secondary effects and interactions (e.g., the impacts of climate change on the rate of improvement in air conditioning equipment efficiency) not addressed in these case studies. The more detailed analyses imply an emphasis on development of disaggregated customer and end-use data as well as application of detailed analytic methods such as 2 "A Survey of Electric Utility Resource Options and Planning Practices," by Kenneth P. Linder (ICF Incorporated) and Bruce G. Humphrey (Edison Electric Institute), June 1986. 6 - 10 From the digital collections of the New York State Library. embodied in the structural approach used in the New York case study to model the weather-sensitivity of demand. Renewable resources such as solar energy and wind are likely to become more important sources of electric power over time. With the exception of hydro power in New York, the case studies have not addressed these resources. Clearly, climate change could affect the availability of these resources and, therefore, energy policies and plans for the future. Analysis of these potential impacts is imperative. The case study analyses focused on the possible direct consequences of climate change on electricity demand and supply because of the attention these factors receive in long-term utility planning. However, climate change potentially could have significant impacts on other factors including the implications of: changes in the seasonal patterns of electricity demand for maintenance scheduling (maintenance schedules were assumed to be fixed in the case studies); the indirect impacts of climate change for utility operations and planning (e.g., the potential impacts of sea level use on the siting and operation of generating units in coastal areas); and changes in the occurrence of extreme events on utility system design and operations (e.g., the effects of more frequent or more severe storms on transmission and distribution system design and maintenance). 6 - 11 From the digital collections of the New York State Library. From the digital collections of the New York State Library. APPENDIX A UTILITY PLANNING FACTORS In assessing the impacts of a l ce rnatdve economic and technological conditions, utility analysts often measure their impacts as changes in a set of utility planning factors. These planning factors are measures of the services, or products, provided by the utility as well as the economic and financial well-being of the utility and its customers. Estimated values for a set of commonly used planning factors are used to measure the potential impacts of climate change. The following subsections describe these planning factors, indicate why each factor is an important consideration for utility planning, describe the relationship of the factor to climatic conditions, and present hypotheses of how climate changes are likely to affect these factors. 1. SALES TO CUSTOHERS (SEASONAL AND ANNUAL) Sales to customers are measured in kilowatt-hours (kWh). (1) determine the total amount of fuel utilities must consume energy needs of its customers and (2) are directly related operation and maintenance costs. Fuel and O&M costs are a portion of total utility costs and are principal inputs to prices a utility must charge in order to cover its costs financially viable. Total sales to meet the to variable substantial calculating and remain Sales to customers are related to climate changes, principally through the demand for weather - sens i tive energy services. These services could account for up to a quarter or more of a utility's total energy requirements depending upon the mix of electricity sales among types of customers (residential, commercial, and industrial), customer equipment stocks, electricity prices, and climate. Principal weather-sensitive energy services include space heating, air conditioning, water heating, and -- for some utilities -- irrigation pumping. Some other energy services, such as refrigeration, also are somewhat affected by weather conditions, although these relationships are secondary. Climate changes associated with global warming could either increase or decrease electricity sales. Warmer temperatures are likely to increase the demand for air conditioning services, but are likely to decrease the demand for space heating services. The net impact will depend upon the relative market penetration and utilization of these alternative technologies. Air conditioning accounts for a greater percentage of utility sales than space heating for most utilities. A principal reason is that non-electric space heating technologies (e.g., gas and oil) tend to be more cost-effective in many regions; exceptions include the Pacific Northwest where electric energy is relatively inexpensive; and the southeast where total space heating requirements are small. In some regions, where space heating and air conditioning requirements are fairly equal, electric heat pumps are displacing gas heating in new housing markets. Increasing temperatures may increase this trend for some utilities (e.g., in the mid-Atlantic and Northeast) . A -1 From the digital collections of the New York State Library. Water heating is a maj or residential and commercial end-use for some utilities. Hot water usage is determined principally by life style (e.g., number and ages of household members) and market penetration of hot water using appliances (e.g., dishwashers). Secondar Ll.y , the amount of energy needed to heat water to a given temperature is dependent upon water inlet temperatures which in turn depends upon ground' water temperature.. Ground water temperature varies by region and season (affecting seasonal electric sales), and could be affected by warming trends. For some utilities with a substantial agricultural load (e.g., in the plains states), irrigation pumping can be a significant load. Typically, sales for irrigation pumping vary from year to year depending principally on rainfall and temperatures. Increases in average temperatures over time and reductions in rainfall could greatly increase demands for irrigation pumping. This particularly would be the case if prolonged periods of low rainfall would lower water tables and increase the depth of wells required to access water. 2. PEAK DEKARD (SEASONAL AND ANROAL) Peak demand measures the maximum level of electricity consumption at a point of time. TypicallYt it is measured over a 15 minute or hour interval and is expressed in kilowatts (KW) or megawatts (MY). Because electricity cannot be stored and must be produced at the same time it is consumed, the utility must plan for enough capacity to meet this maximum hourly demand. Thus, peak demand is a very important factor in determining its capacity requirements. The relative levels of monthly or seasonal peaks also are important for maintenance planning. Electric generating units often need to be out of service for a month or more each year for maintenance. These periods are planned for "o ffi-peak" seasons usually the spring and fall when weather-sensitive demands typically are at their lowest levels . Annual and seasonal peak demands typically are correlated with extreme temperatures and other weather conditions. That iS t the demand for weathersensitive services is a principal determinant of the timing and level of peak demand. Maximum summer temperatures (along with humidity) tend to drive summer peaks and minimum winter temperatures are associated with winter peak demands for utility systems on which electric space heating is an important energy use. A general warming will tend to increase swnmer peaks and decrease winter peaks. These changes are correlated with changes in seasonal sales. Since most utility systems experience their annual peaks during the summer, a relative increase in these peaks could have important implications for utility capacity requirements and the characteristics of that capacity. 3• ELECTRICITY GENERATION Electricity generation is the total amount of energy produced over a period of time. It is measured in kilowatt-hours (kWh). Generation differs from sales principally by the amount of losses experienced in transmitting electricity to areas where it is being demanded, and then distributing it to A -2 From the digital collections of the New York State Library. individual customers. 1 That is, generation is measured at the point of production and sales are measured at the point of consumption. The difference between generation and sales is transmission and distribution (T&D) losses. On an annual basis T&D losses typically account for 7-10 percent of generation. Because generation is a direct function of sales, climate change effects on the latter will affect the former similarly. However, there is another factor to consider. That is, the amount of T&D losses are related (1) to levels of demand and (2) temperature. Thus, climate change can affect the level of losses indirectly through increased load levels and directly through increased temperatures. These effects are related to the physical characteristics of transmission lines and the physics of transmitting electricity. The level of load is more important than the direct effects of temperature in determining losses. For example, in a system experiencing an average annual loss level of 10 percent, losses may be 12 percent during high demand hours and seven percent during low demand hours. 4. LOAD FACTOR Load factor is a relationship between total electricity generation and peak generation. 2 It is a ratio and it is dimensionless. It is a measure of the diversity of daily, seasonal or annual patterns of electricity generation. It can be related directly to the utilization of generation capacity and T&D plant and, therefore, is associated with capacity requirements and operating efficiency. The ratio (often referred to as a percent) typically ranges from 0.55 to 0.65 on an annual basis. A higher load factor is associated with greater overall utilization of utility capacity and (with some exceptions) greater operating efficiency of the utility. Therefore, a comparison of load factor across climate change scenarios can serve as one indication of impacts on operating efficiency. t 5. FUEL USE Fuel use is measured in British thermal units (btu) or kWh. It is a measure of the total fuel needed to meet generation requirements. Of particular importance for utility planners and managers is the projected fuel use by type (e.g., gas, oil, coal). These projections are needed to formulate fuel procurement and inventory strategies. Further, the mix of fuel by type is a principal determinant of total operating costs and prices to consumers. Because of higher costs and for national security reasons, utilities have been implementing plans for reductions in use of critical fuels principally oil for electricity generation. Impacts of alternative conditions affecting utility operations on total fuel use and costs and on the use of critical fuels is an important consideration for utility planners. 1 The word "principally" is used because the utility uses small amounts of electricity itself at generation points. Thi's auxiliary use also is part of the difference between generation and sales. 2 Load Factor = Total Generation (kWh) Peak Generation (KW) x Hours in the Period A -3 From the digital collections of the New York State Library. The impacts of climate change in total fuel use and the mix of fuels will operate principally through changes in customer demands. These impacts on fuel use relate to the overall level of demand (i.e., sales) and the patterns of demand (i.e., conswnption during "peak" and "off-peak" periods). Utilities maintain different types of generating capacity to produce electricity economically. High capital· (fixed) cost, low fuel (variable) cost plans are used to meet "bas e Lo ad" requirements; that is they operate for many hours a year. Low capital cost, high fuel cost plants are used to meet daily or seasonal high load hours and peaks. They are not operated many hours. Therefore, the relative influence of climate change on customer sales vs. peak demand will influence the mix of fuels. Further, to the extent that planners do not anticipate increased demands associated with climate changes, existing high fuel cost or inefficient generating units may be utilized more extensively than planned, thereby affecting total fuel use and mix. 6. CAPACITY ADDITIONS Capacity additions refer to investments in generation, transmission, and distribution capacity needed to satisfy customer demands over the planning horizon. Generation capacity is measured in kw by plant or fuel type. Typically, sufficient capacity is built and maintained to meet firm customer demands (i.e., the utility is committed to provide service upon demand) under all but extreme conditions. 1 This ability to meet firm demands is referred to as system reliability., There are several measures of reliability including reserve margin (i.e., building sufficient capacity to meet demands as much as 20-30 percent in excess of forecast peak demand) and loss of load probability (i.e., a statistical measure indicating the likelihood that firm customer demands will not be able to be met, say, more often than 1 or 2 days in ten years). As described with respect to fuel use, capacity requirements will be affected by climate change primarily through the level and pattern of customer demands. A secondary influence will be through the direct impacts on operating characteristics (e. g., the effects of temperature on steam plant efficiency and effective capacity ratings). These impacts are measured through changes in the amount and type of capacity required to meet customer demands reliably and at minimum cost. 7 .. CAPACITY AND OPERATING COSTS Total utility capacity and operating costs are important for determining the revenues which need to be collected from customers to meet operating expenses, investments in new capacity, and taxes and to provide a return on utility stockholder investments. Dividing total costs by total sales provides a proxy of average electricity rates (expressed as ¢/kWh) charged to the utility's customers. This cost per unit (expressed in 1 Some types of customers have non-firm contracts with utilities and can be denied power if sufficient capacity is not available. A -4 From the digital collections of the New York State Library. levelized, present value form) is an accepted proxy for electricity rates in utility planning analyses, and changes in levelized cost per unit provide indications of the direction and magnitude of the impacts of climate change on electricity prices to consumers. These impacts reflect both the direct and indirect effects of climate change on utility investments and operations. It is anticipated that increasing temperatures will increase levelized costs per unit for most utilities accounting for increased capital and operating costs to meet increased demands and reduced operating efficiencies. The magni tude of these impacts, however, are likely to be small over the next several years. A -5 From the digital collections of the New York State Library. From the digital collections of the New York State Library. A A44 . For further information on NYSERDA reports or publications contact: Department of Communications NYS Energy Research and Development Authority Two Rockefeller Plaza Albany, NY. 12223 (518) 465-6251 From the digital collections of the New York State Library. Energy Authority Report 88-2 State of New York Mario M. Cuomo, Governor New York State Energy Research and Development Authority William D. Cotter, Chairman Irvin L. White, President From the digital collections of the New York State Library.