Integrated Resource Plan

Transcription

1 June 2015 Integrated Resource Plan APPENDIX A Sales and Load Forecast 2015

2 SAFE HARBOR STATEMENT This document may contain forward-looking statements, and it is important to note that the future results could differ materially from those discussed. A full discussion of the factors that could cause future results to differ materially can be found in Idaho Power s filings with the Securities and Exchange Commission.

3 June 2015 Integrated Resource Plan 2015 ACKNOWLEDGMENT APPENDIX A Sales and Load Forecast Resource planning is an ongoing process at Idaho Power. Idaho Power prepares, files, and publishes an Integrated Resource Plan every two years. Idaho Power expects that the experience gained over the next few years will likely modify the 20-year resource plan presented in this document. Idaho Power invited outside participation to help develop the 2015 Integrated Resource Plan. Idaho Power values the knowledgeable input, comments, and discussion provided by the Integrated Resource Plan Advisory Council and other concerned citizens and customers. It takes approximately one year for a dedicated team of individuals at Idaho Power to prepare the Integrated Resource Plan. The Idaho Power team is comprised of individuals that represent many different departments within the company. The Integrated Resource Plan team members are responsible for preparing forecasts, working with the advisory council and the public, and performing all the analyses necessary to prepare the resource plan. Idaho Power looks forward to continuing the resource planning process with customers, public interest groups, regulatory agencies, and other interested parties. You can learn more about the Idaho Power resource planning process at Printed on recycled paper

4

5 Idaho Power Company Appendix A Sales and Load Forecast TABLE OF CONTENTS Table of Contents... i List of Tables... ii List of Figures... ii List of Appendices... iii Introduction IRP Sales and Load Forecast...3 Average Load...3 Peak-Hour Demands...4 Overview of the Forecast...7 Fuel Prices...8 Electric Vehicles...10 Forecast Probabilities...11 Load Forecasts Based on Weather Variability...11 Load Forecasts Based on Economic Uncertainty...12 Residential...15 Commercial...17 Irrigation...21 Industrial...23 Additional Firm Load...25 Micron Technology...26 Simplot Fertilizer...26 Idaho National Laboratory...26 Company System Peak...27 Company System Load...29 Contract Off-System Load...31 Energy Efficiency and Demand Response...33 Energy Efficiency...33 Demand Response Integrated Resource Plan Page i

6 Appendix A Sales and Load Forecast Idaho Power Company Table 1. LIST OF TABLES Residential fuel-price escalation ( ) (average annual percent change)...8 Table 2. Average load and peak-demand forecast scenarios...12 Table 3. Forecast probabilities...13 Table 4. System load growth (amw)...14 Table 5. Residential load growth (amw)...15 Table 6. Commercial load growth (amw)...17 Table 7. Irrigation load growth (amw)...21 Table 8. Industrial load growth (amw)...23 Table 9. Additional firm load growth (amw)...25 Table 10. System summer peak load growth (MW)...27 Table 11. System winter peak load growth (MW)...28 Table 12. System load growth (amw)...29 LIST OF FIGURES Figure 1. Forecast residential electricity prices (cents per kwh)...9 Figure 2. Forecast residential natural gas prices (dollars per therm)...10 Figure 3. Forecast system load (amw)...14 Figure 4. Forecast residential load (amw)...15 Figure 5. Forecast residential use per customer (weather-adjusted kwh)...16 Figure 6. Forecast commercial load (amw)...17 Figure 7. Commercial building share...18 Figure 8. Forecast commercial use per customer (weather-adjusted kwh)...19 Figure 9. Commercial categories UPC, 2014 relative to Figure 10. Forecast irrigation load (amw)...21 Figure 11. Forecast industrial load (amw)...23 Figure 12. Industrial electricity consumption by industry group (based on 2014 sales)...24 Figure 13. Forecast additional firm load (amw)...25 Figure 14. Forecast system summer peak (MW)...27 Figure 15. Forecast system winter peak (MW)...28 Page ii 2015 Integrated Resource Plan

7 Idaho Power Company Appendix A Sales and Load Forecast Figure 16. Forecast system load (amw)...29 Figure 17. Composition of system company electricity sales (thousands of MWh)...30 LIST OF APPENDICES Appendix A1. Historical and Projected Sales and Load...35 Residential Load...35 Historical Residential Sales and Load, (weather adjusted)...35 Projected Residential Sales and Load, Commercial Load...37 Historical Commercial Sales and Load, (weather adjusted)...37 Projected Commercial Sales and Load, Irrigation Load...39 Historical Irrigation Sales and Load, (weather adjusted)...39 Projected Irrigation Sales and Load, Industrial Load...41 Historical Industrial Sales and Load, (not weather adjusted)...41 Projected Industrial Sales and Load, Additional Firm Sales and Load*...43 Historical Additional Firm Sales and Load, Projected Additional Firm Sales and Load, Company System Load (excluding Astaris)...45 Historical Company System Sales and Load, (weather adjusted)...45 Projected Company System Sales and Load, Integrated Resource Plan Page iii

8 Appendix A Sales and Load Forecast Idaho Power Company This page left blank intentionally. Page iv 2015 Integrated Resource Plan

9 Idaho Power Company Appendix A Sales and Load Forecast INTRODUCTION Idaho Power has prepared Appendix A Sales and Load Forecast as part of the 2015 Integrated Resource Plan (IRP). The sales and load forecast is Idaho Power s best estimate of the future demand for electricity within the company s service area. The forecast covers the 20-year period from 2015 through The expected-case monthly average load forecast represents Idaho Power s estimate of the most probable outcome for load growth during the planning period and is based on the most recent economic forecast for Idaho Power s service area. To account for inherent uncertainty and variability, four additional load forecasts were prepared two that provide a range of possible load growths due to economic uncertainty and two that address the load variability associated with abnormal weather. The high and low economic growth scenarios provide a range of possible load growths over the planning period due to variable economic, demographic, and other non-weather-related influences. The high-growth and low-growth scenarios were prepared based on statistical analyses to empirically reflect uncertainty inherent in the load forecast. The 70 th -percentile and 90 th -percentile load forecast scenarios were developed to assist Idaho Power in reviewing the resource requirements that would result from higher loads due to more adverse weather conditions. While the expected-case load forecast assumes median historical values (50 th percentile) for temperatures and median rainfall, the weather scenarios are developed with a 70 th -percentile and a 90 th -percentile weather probability assumption. The 70 th -percentile load forecast assumes monthly loads that can be exceeded in 3 out of 10 years (30% of the time). The 90 th -percentile load forecast assumes monthly loads that can be exceeded in 1 out of 10 years (10% of the time). Idaho Power uses the 70 th -percentile load forecast in IRP resource planning to account for the risk associated with weather impacts on load. In the expected-case scenario, Idaho Power s system load is forecast to increase to 2,240 average megawatts (amw) by 2034 from 1,786 amw in 2015, representing an average yearly growth rate of 1.2 percent over the 20-year planning period ( ). In the more critical 70 th -percentile load forecast used for resource planning, the system load is forecast to reach 2,292 amw by 2034 (1.2% average annual growth). The Idaho Power system peak load (95 th percentile) is forecast to grow to 4,773 megawatts (MW) in 2034 from the actual system summer peak of 3,407 MW that occurred on Tuesday, July 2, 2013, at 4:00 p.m. In the expected-case scenario, the Idaho Power system peak increases at an average growth rate of 1.5 percent per year over the 20-year planning period ( ). This year s economic forecast was based on a forecast of national and regional economic activity developed by Moody s Analytics, Inc. The national, state, metropolitan statistical area (MSA) and county econometric projections are tailored to Idaho Power s service area using an in-house economic forecast model and database. Specific demographic projections are also developed for the service area from national and local census data. National economic drivers from Moody s Analytics were also used in the development of Appendix A Sales and Load Forecast. The number of Idaho Power active retail customers is expected to increase from the December 2014 level of 514,700 customers to nearly 710,000 customers by year-end Integrated Resource Plan Page 1

10 Appendix A Sales and Load Forecast Idaho Power Company Economic growth assumptions influence several classes of service growth rates. The number of households in Idaho is projected to grow at an annual rate of 1.9 percent during the forecast period. The growth in the number of households within individual counties in Idaho Power s service area is projected to grow faster than the remainder of the state over the planning period. In addition, the number of households, incomes, employment, economic output, real retail electricity prices, and customer consumption patterns are used to develop load projections. In addition to the economic assumptions used to drive the expected-case forecast scenario, several assumptions were incorporated into the forecasts of the residential, commercial, industrial, and irrigation sectors. Further discussions of these assumptions are presented below. Conservation influences on the load forecast, including Idaho Power energy efficiency demand-side management (DSM) programs, statutory programs, and non-programmatic trends in conservation, are included in the load forecasts of each sector. Idaho Power DSM programs are described in detail in Idaho Power s Demand-Side Management 2014 Annual Report, which is incorporated into this IRP document as Appendix B. During the 20-year forecast horizon, major changes in the electric utility industry (e.g., carbon regulations and subsequent higher electricity prices impacting future electricity demand) could influence the load forecast. In addition, the price and volatility of substitute fuels, such as natural gas, may also impact future demand for electricity. The high degree of uncertainty associated with such changes is reflected in the economic high and low load growth scenarios described previously. The impact of carbon legislation on the load forecast is reflected in the retail electricity price variable for each forecasted customer sector. The alternative sales and load scenarios in Appendix A Sales and Load Forecast were prepared under the assumption that Idaho Power s geographic service area remains unchanged during the planning period. Data describing the historical and projected figures for the sales and load forecast are presented in Appendix A1 of this report. Page Integrated Resource Plan

11 Idaho Power Company Appendix A Sales and Load Forecast Average Load 2015 IRP SALES AND LOAD FORECAST The 2015 IRP average annual system load forecast reflects the continued improvement in the service area economy. While economic conditions during the development of the 2013 IRP were positive, they were less optimistic than the actual performance experienced in the interim period leading up to the 2015 IRP. The improved economic and demographic variables driving the 2015 forecast are reflected by a more positive sales outlook throughout the planning period. The stalled recovery in the national and, to a lesser extent, service-area economy caused load growth to stall through However, in 2012, the recovery was evident, with strength exhibited in most all economic drivers to date. Retail electricity price projections for the 2015 IRP are lower relative to the 2013 IRP, serving to increase the forecast of average loads, especially in the second 10 years of the forecast period. Significant factors and considerations that influenced the outcome of the 2015 IRP load forecast include the following: The load forecast used for the 2015 IRP reflects a near-term recovery in the service-area economy following a severe recession in 2008 and 2009 that kept sales from growing through The collapse in the housing sector beginning in 2008 held new construction and customer growth to a near standstill until However, beginning in 2012, acceleration of in-migration and business investment resulted in renewed growth in the residential and commercial connections along with increased industrial activity. By 2017, customer additions are forecast to approach sustainable growth rates experienced prior to the housing bubble ( ). The electricity price forecast used to prepare the sales and load forecast in the 2015 IRP reflects the impact of additional plant investment and associated variable costs of integrating new resources identified in the 2013 IRP preferred portfolio, including the expected costs of carbon emissions. As discussed previously, when compared to the electricity price forecast used to prepare the 2013 IRP sales and load forecast, the 2015 IRP price forecast yields lower future prices. The retail prices are most evident in the second 10 years of the planning period and impact the sales forecast positively, a consequence of the inverse relationship between electricity prices and electricity demand. There continues to be significant uncertainty associated with the industrial and special-contract sales forecasts due to the number of parties that contact Idaho Power expressing interest in locating operations within Idaho Power s service area, typically with an unknown magnitude of the energy and peak-demand requirements. Nonetheless, the expected load forecast reflects only those industrial customers that have made a sufficient and significant binding investment indicating a commitment of the highest probability of locating in the service area. Therefore, the large numbers of prospective businesses that have indicated an interest in locating in Idaho Power s service area but have not made sufficient commitments are not included in the current sales and load forecast Integrated Resource Plan Page 3

12 Appendix A Sales and Load Forecast Idaho Power Company Conservation impacts, including DSM energy efficiency programs and codes and standards, are considered and integrated into the sales forecast. Impacts of demand response programs (on peak) are accounted for in the load and resource balance analysis within supply-side planning (i.e., are treated as a supply-side peaking resource). The amount of committed and implemented DSM programs for each month of the planning period is shown in the load and resource balance in Appendix C Technical Appendix. The 2015 irrigation sales forecast is higher than the 2013 IRP forecast throughout the entire forecast period due to the significant trend toward more water-intensive crops, primarily alfalfa and corn, due to growth in the dairy industry. Also, farmers have taken advantage of higher market prices over the past few years and have put high-lift acreage back into production. Additionally, load increases have come from the conversion of flood/furrow irrigation to sprinkler irrigation, primarily related to farmers trying to reduce labor costs. Peak-Hour Demands Peak-day temperatures and the growth in average loads drive the peak forecasting model regressions. The peak forecast results and comparisons with previous forecasts differ for a number of reasons that include the following: The 2015 IRP peak-demand forecast considers the impact of committed and implemented energy efficiency DSM programs on peak demand. The 2015 IRP peak-demand forecast model explicitly excludes the impact of demand response programs to establish peak impacts. The exclusion allows for planning for demand response programs and supply-side resources in meeting peak demand. Demand response program impacts are accounted for in the IRP load and resource balance and are reflected as a reduction in peak demand. The peak model develops peak-scenario impacts based on historical probabilities of peak-day temperatures at the 50 th, 90 th, and 95 th percentiles of occurrence for each month of the year. The all-time system summer peak demand was 3,407 MW (recorded on Tuesday, July 2, 2013, at 4:00 p.m.) and serves as a benchmark for the forecasting model. The previous summer peak demand was 3,245 MW, occurring on Thursday, July 12, 2012, at 4:00 p.m. Historical peak-demand data serve as the basis in the peak-model regressions. Historical new peak loads were reached in July 2007, June 2008, July 2012, and July The summer system peak load growth accelerated from 1998 to 2008 as a record number of residential and commercial customers were added to the system and air conditioning (A/C) became standard in nearly all new residential homes and new commercial buildings. Page Integrated Resource Plan

13 Idaho Power Company Appendix A Sales and Load Forecast Idaho Power uses a median peak-day temperature driver in lieu of an average peak-day temperature driver in the 50/50 peak-demand forecast scenario. The median peak-day temperature has a 50-percent probability of being exceeded. Peak-day temperatures are not normally distributed and can be skewed by one or more extreme observations as referred to in the previous bulleted item; therefore, the median temperature better reflects expected temperatures within the context of probabilistic percentiles. The weighted average peak-day temperature drivers are calculated over the 1984 to 2013 time period (the most recent 30 years) Integrated Resource Plan Page 5

14 Appendix A Sales and Load Forecast Idaho Power Company This page left blank intentionally. Page Integrated Resource Plan

15 Idaho Power Company Appendix A Sales and Load Forecast OVERVIEW OF THE FORECAST The sales and load forecast is constructed by developing a separate forecast for each of the major customer classes: residential, commercial, irrigation, and industrial. Individual energy and peak-demand forecasts are developed for special-contract customers, including Micron Technology, Inc.; Simplot Fertilizer Company (Simplot Fertilizer); and the Idaho National Laboratory (INL). These three special-contract customers are reported as a single forecast category labeled additional firm load. Currently, Idaho Power has no long-term contracts to provide off-system customers with firm energy and demand. The assumptions for each of the individual categories are described in greater detail in the respective sections. Since the residential, commercial, irrigation, and industrial sales forecasts provide a forecast of sales as billed, it is necessary to adjust these billed sales to the proper time frame to reflect the required generation needed in each calendar month. To determine calendar-month sales from billed sales, the billed sales must first be converted from billed periods to calendar months to synchronize them with the time period in which load is generated. The calendar-month sales are then converted to calendar-month average load by adding losses and dividing by the number of hours in each month. Loss factors are determined by Idaho Power s Transmission Planning department. The annual average energy loss coefficients are multiplied by the calendar-month load, yielding the system load, including losses. A system loss study of 2012 was completed in May The results of the study concluded that on average, the loss coefficients are lower than those applied to the 2013 IRP generation forecast. This resulted in a permanent reduction of nearly 20 amw to the load forecast annually. The peak-load forecast was prepared in conjunction with the 2015 sales forecast. Idaho Power has two peak periods: 1) a winter peak, resulting primarily from space-heating demand that normally occurs in December, January, or February and 2) a larger summer peak that normally occurs in late June or July. The summer peak generally occurs when extensive A/C use coincides with significant irrigation demand. Peak loads are forecast using 12 regression equations and are a function of average peak-day temperatures, the historical monthly average load, and precipitation (summer only). The peak forecast uses statistically derived peak-day temperatures based on the most recent 30 years of climate data for each month. Peak loads for the INL, Micron Technology, and Simplot Fertilizer are forecast based on a historical analysis, customer-provided input, and any contractual considerations. The primary external factors influencing the forecast are economic and demographic in nature. Moody s Analytics serves as the primary provider for this data. The national, state, MSA, and county economic and demographic projections are tailored to Idaho Power s service area using an in-house economic database. Specific demographic projections are also developed for the service area from national and local census data. Additional data sources used to substantiate Moody s data include the Idaho Department of Labor, Construction Monitor, and Federal Reserve Economic Databases Integrated Resource Plan Page 7

16 Appendix A Sales and Load Forecast Idaho Power Company Fuel Prices Fuel prices, in combination with service-area demographic and economic drivers, impact long-term trends in electricity sales. Changes in relative fuel prices can also have significant impacts on the future demand for electricity. The sales and load forecast is also influenced by the estimated impact of proposed carbon legislation on retail electricity prices. In addition to supply-side influences, carbon-reduction legislation creates an upward trend in retail electricity prices throughout the forecast period, resulting in reduced future electricity sales. Class-level and economic-sector-level regression models were used to identify the relationships between real historical electricity prices and their impact on historical electricity sales. The estimated coefficients from these models were used as drivers in the individual sales forecast models. Short-term and long-term nominal electricity price increases are generated internally from Idaho Power financial models. The United States (US) Energy Information Administration (EIA) provides the forecasts of long-term changes in nominal natural gas prices. The nominal price estimates are adjusted for projected inflation by applying the appropriate economic deflators to arrive at real fuel prices. The projected average annual growth rates of fuel prices in nominal and real terms (adjusted for inflation) are presented in Table 1. The growth rates shown are for residential fuel prices and can be used as a proxy for fuel-price growth rates in the commercial, industrial, and irrigation sectors. Table 1. Residential fuel-price escalation ( ) (average annual percent change) Nominal Electricity 2015 IRP % 0.0% Electricity 2013 IRP % 1.0% Natural Gas % 1.3% * Adjusted for inflation Figure 1 illustrates the average electricity price paid by Idaho Power s residential customers over the historical period 1979 to 2014 and over the forecast period 2015 to Both nominal and real prices are shown. In the 2015 IRP, nominal electricity prices are expected to climb to about 14 cents per kilowatt-hour (kwh) by the end of the forecast period in Real electricity prices (inflation adjusted) are expected to remain flat over the forecast period at an average rate of 0.0 percent annually. In the 2013 IRP, nominal electricity prices were assumed to climb to about 18 cents per kwh by 2034, and real electricity prices (inflation adjusted) were expected to slowly increase over the forecast period at an average rate of 1.0 percent annually. The impact of the lower real electricity price forecast on the 2015 IRP load forecast serves to positively influence the growth in electricity sales, especially in the last 10 years of the forecast period. The electricity price forecast used to prepare the sales and load forecast in the 2015 IRP reflected the additional plant investment and variable costs of integrating the resources identified in the 2013 IRP preferred portfolio, including the expected costs of carbon emissions. When compared to the electricity price forecast used to prepare the 2013 IRP sales and load forecast, the 2015 IRP price forecast yielded lower future prices. The retail prices are more evidently lower in the Real* Page Integrated Resource Plan

17 Idaho Power Company Appendix A Sales and Load Forecast second 10 years of the planning period and impact the sales forecast positively, a consequence of the inverse relationship between electricity prices and electricity demand Hi Nominal Real Nominal 2013 IRP Nominal 2015 IRP Real 2013 IRP Real 2015 IRP Figure 1. Forecast residential electricity prices (cents per kwh) Electricity prices for Idaho Power customers increased significantly in 2001 and 2002 because of the power cost adjustment (PCA) impact on rates, a direct result of the western US energy crisis of 2000 and Prior to 2001, Idaho Power s electricity prices were historically quite stable. From 1990 to 2000, electricity prices rose only 8 percent overall, an annual average compound growth rate of 0.8 percent annually. Figure 2 illustrates the average natural gas price paid by Intermountain Gas Company s residential customers over the historical period 1979 to 2013 and forecast prices from 2014 to Natural gas prices remained stable and flat throughout the 1990s before moving sharply higher in Since spiking in 2001, natural gas prices moved downward for a couple of years before moving sharply upward in 2004 through Since 2006, natural gas prices have experienced a steady decline, matching prices from over a decade ago. Nominal natural gas prices are expected to remain flat through 2017, then rise at a steady pace throughout the remainder of the forecast period until nearly doubling by 2034, growing at an average rate of 3.2 percent per year. Real natural gas prices (adjusted for inflation) are expected to increase over the same period at an average rate of 1.3 percent annually Integrated Resource Plan Page 9

18 Appendix A Sales and Load Forecast Idaho Power Company $1.80 $1.60 $1.40 $1.20 $1.00 $0.80 $0.60 $0.40 $0.20 $ Nominal Actual Nominal Forecast Real Actual Real Forecast Figure 2. Forecast residential natural gas prices (dollars per therm) If future natural gas price increases outpace electricity price increases, the operating costs of space heating and water heating with electricity would become more advantageous when compared to that of natural gas. However, in the 2015 IRP price forecast, the long-term growth rates of electricity and natural gas prices are nearly identical. Electric Vehicles The load forecast includes an update of the impact of plug-in electric vehicles (PEV) on system load to reflect the future impact of this relatively new and evolving source of energy use. While PEV consumer adoption rates in Idaho Power s service area remain relatively low, with continued technological advancement, such limiting attributes of vehicle range and re-fueling time continue to improve the competiveness of these vehicles to non-electric models. Since the first introduction of the Chevy Volt and Nissan Leaf, the number of PEVs offered in the marketplace has proliferated to over 50 models since Early in this period, PEVs were sold with unique model names (e.g., VOLT); however, as the market grows, the plug-in technology is increasingly offered as an option to existing models (e.g., Ford Focus). Initially, the Idaho Power forecast for PEV impact relied on third-party forecasts from the Electric Power Research Institute (EPRI) and Oak Ridge National Laboratory due to a lack of service-area vehicle registration data; however, beginning with the 2011 IRP, sufficient service-area data became available via vehicle registration data provided by the Idaho Transportation Department (ITD). This data provides a basis from which to develop service-area adoption rates and support the collection of charging behavior. The methodology continues to integrate the fuel and technology share forecasts of the Department of Energy s (DOE) National Energy Model (NEM). The Idaho Power vehicle share forecast uses these models as well as a Bass consumer adoption model as informed by registration data. Load impacts from the share model output are derived from assumptions of battery-only and hybrid plug-in shares evident from Idaho Power observations and informed by the DOE. Page Integrated Resource Plan

19 Idaho Power Company Appendix A Sales and Load Forecast Currently, the registration data collection methodology is being revised to capture vehicles sold with PEV technology as an option (e.g., Ford Focus). The methodology will require the unique string of characters within the vehicle identification number (VIN) to be identified and serve as a key value in the ITD data extraction. The PEV forecast in this IRP did include registration data for the Toyota Prius PEV but did not capture all models for which PEV technology is sold as an option; however, to capture the impact of these models on future adoption, the forecast used the forecast national share assumptions from the DOE. The net effect was to rely less on the registration data than the 2013 IRP model and more on third-party assumptions, as was the case in earlier forecasts. Forecast Probabilities Load Forecasts Based on Weather Variability The future demand for electricity by customers in Idaho Power s service area is represented by three load forecasts reflecting a range of load uncertainty due to weather. The expected-case load forecast represents the most probable projection of system load growth during the planning period and is based on the most recent national, state, MSA, and county economic forecasts from Moody s Analytics and the resulting derived economic forecast for Idaho Power s service area. The expected-case load forecast assumes median temperatures and median precipitation (i.e., there is a 50-percent chance loads will be higher or lower than the expected-case loads due to colder-than-median or hotter-than-median temperatures or wetter-than-median or drier-than-median precipitation). Since actual loads can vary significantly depending on weather conditions, two alternative scenarios were considered that address load variability due to weather. Maximum load occurs when the highest recorded levels of heating degree days (HDD) are assumed in winter and the highest recorded levels of cooling and growing degree days (CDD and GDD) combined with the lowest recorded level of precipitation are assumed in summer. Conversely, the minimum load occurs when the lowest recorded levels of HDD are assumed in winter and the lowest recorded levels of CDD and GDD, combined with the highest level of precipitation, are assumed in summer. For example, at the Boise Weather Service office, the median HDD in December from 1984 to 2013 (the most recent 30 years) was 1,039. The 70 th -percentile HDD is 1,074 and would be exceeded in 3 out of 10 years. The 90 th -percentile HDD is 1,268 and would be exceeded in 1 out of 10 years. The 100 th -percentile HDD (the coldest December over the 30 years) is 1,619 and occurred in December This same concept was applied in each month throughout the year in only the weather-sensitive customer classes: residential, commercial, and irrigation. In the 70 th -percentile residential and commercial load forecasts, temperatures in each month were assumed to be at the 70 th percentile of HDD in wintertime and at the 70 th percentile of CDD in summertime. In the 70 th -percentile irrigation load forecast, GDD were assumed to be at the 70 th percentile and precipitation at the 30 th percentile, reflecting drier-than-median weather. The 90 th -percentile load forecast was similarly constructed Integrated Resource Plan Page 11

20 Appendix A Sales and Load Forecast Idaho Power Company Idaho Power loads are highly dependent on weather, and these two scenarios allow the careful examination of load variability and how it may impact future resource requirements. It is important to understand that the probabilities associated with these forecasts apply to any given month. To assume temperatures and precipitation would maintain a 70 th -percentile or 90 th -percentile level continuously, month after month throughout an entire year, would be much less probable. Monthly forecast numbers are evaluated for resource planning, and caution should be used in interpreting the meaning of the annual average load figures being reported and graphed for the 70 th -percentile or 90 th -percentile forecasts. Table 2 summarizes the load scenarios prepared for the 2015 IRP. Three average load scenarios were prepared based on a statistical analysis of the historical monthly weather variables listed. The probability associated with each average load scenario is also indicated in the table. In addition, three peak-demand scenarios were prepared based on a statistical analysis of historical peak-day average temperatures, and the probability associated with each peak-demand scenario is also indicated in Table 2. Table 2. Average load and peak-demand forecast scenarios Scenario Weather Probability Probability of Exceeding Weather Driver Forecasts of Average Load 90 th Percentile 90% 1 in 10 years HDD, CDD, GDD, precipitation 70 th Percentile 70% 3 in 10 years HDD, CDD, GDD, precipitation Expected Case 50% 1 in 2 years HDD, CDD, GDD, precipitation Forecasts of Peak Demand 95 th Percentile 95% 1 in 20 years Peak-day temperatures 90 th Percentile 90% 1 in 10 years Peak-day temperatures 50 th Percentile 50% 1 in 2 years Peak-day temperatures The analysis of resource requirements is based on the 70 th -percentile average load forecast coupled with the 95 th -percentile peak-demand forecast to provide a more adverse representation of the average load and peak demand to be considered. In other Idaho Power planning, such as the preparation of the financial forecast or the operating plan, the expected-case (50 th percentile) average-load forecast and the 90 th -percentile peak-demand forecast are typically used. Load Forecasts Based on Economic Uncertainty The expected-case load forecast is based on the most recent economic forecast for Idaho Power s service area and represents Idaho Power s most probable outcome for load growth during the planning period. To provide risk assessment to economic uncertainty, two additional load forecasts for Idaho Power s service area were prepared. The forecasts provide a range of possible load growths for the 2015 to 2034 planning period due to high and low economic and demographic conditions. The high and low economic-growth scenarios were prepared based on a statistical analysis to empirically reflect the uncertainty inherent in the load forecast. The average growth rates for the high and low growth scenarios were derived from the historical distribution of one-year growth rates over the past 25 years ( ). Page Integrated Resource Plan

21 Idaho Power Company Appendix A Sales and Load Forecast The estimated probabilities for the three load scenarios are reported in Table 2. The standard deviation observed during the historical time period is used to estimate the dispersion around the expected-case scenario. The probability estimates assume the expected forecast is the median growth path (i.e., there is a 50-percent probability the actual growth rate will be less than the expected-case growth rate and a 50-percent chance the actual growth rate will be greater than the expected-case growth rate). In addition, the probability estimates assume the variation in growth rates will be equivalent to the variation in growth rates observed over the past 25 years ( ). Two views of probable outcomes from the forecast scenarios the probability of exceeding and the probability of occurrenceare are reported in Table 3. The probability of exceeding shows the likelihood the actual load growth will be greater than the projected growth rate in the specified scenario. For example, over the next 20 years, there is a 10-percent probability the actual growth rate will exceed the growth rate projected in the high scenario; conversely, there is a 10-percent chance the actual growth rate will fall below that of the low scenario. In other words, over a 20-year period, there is an 80-percent probability the actual growth rate of system load will fall between the growth rates projected in the high and low scenarios. The second probability estimate, the probability of occurrence, indicates the likelihood the actual growth will be closer to the growth rate specified in that scenario than to the growth rate specified in any other scenario. For example, there is a 26-percent probability the actual growth rate will be closer to the high scenario than to any other forecast scenario for the entire 20-year planning horizon. Probabilities for shorter, 1-year, 5-year, and 10-year time periods are also shown in Table 3. Table 3. Forecast probabilities Probability of Exceeding Scenario 1-year 5-year 10-year 20-year Low Growth... 90% 90% 90% 90% Expected Case... 50% 50% 50% 50% High Growth... 10% 10% 10% 10% Probability of Occurrence Scenario 1-year 5-year 10-year 20-year Low Growth... 26% 26% 26% 26% Expected Case... 48% 48% 48% 48% High Growth... 26% 26% 26% 26% The system load is the sum of the individual loads of residential, commercial, industrial, and irrigation customers, as well as special contracts (including past sales to Astaris, Inc.) and on-system contracts (including past sales to Raft River Coop and the City of Weiser). Idaho Power system load projections are reported in Table 4 and shown in Figure 3. The expected-case system load-forecast growth rate averages 1.1 percent per year over the 20-year planning period. The low scenario projects that the system load will increase at an average rate of 0.7 percent per year throughout the forecast period. The high scenario projects a load growth of 1.6 percent per year. Idaho Power has experienced both the high- and low-growth rates in the past. These forecasts provide a range of projected growth rates that cover approximately 80 percent of the probable outcomes as measured by Idaho Power s historical experience Integrated Resource Plan Page 13

22 Appendix A Sales and Load Forecast Idaho Power Company Table 4. System load growth (amw) Growth Annual Growth Rate Low... 1,776 1,802 1,871 2, % Expected... 1,786 1,900 2,012 2, % High... 1,864 2,009 2,181 2, % 2,800 2,600 2,400 2,200 2,000 1,800 1,600 1,400 1,200 1, Weather Adjusted (excluding Astaris) Expected 70th Percentile High Low Figure 3. Forecast system load (amw) Page Integrated Resource Plan

23 Idaho Power Company Appendix A Sales and Load Forecast RESIDENTIAL The expected-case residential load is forecast to increase from 588 amw in 2015 to 755 amw in 2034, an average annual compound growth rate of 1.3 percent. In the 70 th -percentile scenario, the residential load is forecast to increase from 608 amw in 2015 to 780 amw in 2034, matching the expected-case residential growth rate. The residential load forecasts are reported in Table 5 and shown in Figure 4. Table 5. Residential load growth (amw) Growth Annual Growth Rate th Percentile % 70 th Percentile % Expected Case % 1, Weather Adjusted Expected Case 70th Percentile 90th Percentile Figure 4. Forecast residential load (amw) Sales to residential customers made up 32 percent of Idaho Power s system sales in 1984 and 36 percent of system sales in The residential customer proportion of system sales is forecast to be approximately 37 percent in The number of residential customers is projected to increase to approximately 591,000 by December The average sales per residential customer increased to over 14,800 kwh in 1979 before declining to 13,200 kwh in In 2002 and 2003, residential use per customer dropped dramatically over 500 kwh per customer from 2001 the result of two years of significantly higher electricity prices combined with a weak national and service-area economy. The reduction in electricity prices in June 2003 and a recovery in the service-area economy caused residential 2015 Integrated Resource Plan Page 15

24 Appendix A Sales and Load Forecast Idaho Power Company use per customer to stabilize and rise through However, the recession in 2008 and 2009, combined with conservation programs designed to reduce electricity use, slowed the growth in residential use per customer. The average sales per residential customer are expected to slowly decline to approximately 11,200 kwh per year in Average annual sales per residential customer are shown in Figure 5. 20,000 18,000 16,000 14,000 Hi 12,000 10,000 8,000 6,000 4,000 2, Actual Forecast Figure 5. Forecast residential use per customer (weather-adjusted kwh) The residential-use-per-customer forecast is based on a forecast of the number of residential customers and an econometric analysis of residential-sector sales. The number of residential customers being added each year is a direct function of the number of new service-area households as derived from Moody s Analytics July 2014 forecast of county housing stock and demographic data. The residential-customer forecast for 2015 to 2034 shows an average annual growth rate of 1.6 percent. The residential sales forecast equation considers several factors affecting electricity sales to the residential sector. Residential sales are a function of HDD (wintertime); CDD (summertime); the number of service-area households as derived from Moody s Analytics forecasts of county housing stock; the real price of electricity; and the real price of natural gas. The forecast of residential use per customer is arrived at by dividing the residential sales forecast, which considers the impact of forecast DSM, by the residential-customer forecast. Page Integrated Resource Plan

25 Idaho Power Company Appendix A Sales and Load Forecast COMMERCIAL The commercial category is primarily made up of Idaho Power s small general-service and large general-service customers. Other customers associated with this category include unmetered general-service, street-lighting service, traffic-control signal lighting service, and dusk-to-dawn customer lighting. In total, within the expected-case scenario, the commercial load is projected to increase from 466 amw in 2015 to 559 amw in The average annual compound-growth rate of the commercial load is 1.0 percent during the forecast period. As summarized in Table 6, the commercial load in the 70 th -percentile scenario is projected to increase from 472 amw in 2015 to 568 amw in The commercial load forecasts are illustrated in Figure 6. Table 6. Commercial load growth (amw) Growth Annual Growth Rate th Percentile % 70 th Percentile % Expected Case % Weather Adjusted Expected Case 70th Percentile 90th Percentile Figure 6. Forecast commercial load (amw) With a customer base of over 67,000, the commercial class represents the diversity of the service area economy, ranging from residential subdivision pressurized irrigation to manufacturing. Due to this diversity, the category is further segmented into categories associated with common elements of energy-use influences, such as economic variables (e.g., employment), industry (e.g., manufacturing), and building structure characteristics (e.g., offices). Figure 7 shows the breakdown of the categories and their relative sizes based on 2014 billed energy sales Integrated Resource Plan Page 17

26 Appendix A Sales and Load Forecast Idaho Power Company The commercial-customer forecast for 2015 to 2034 shows an average annual growth rate of 1.7 percent. 3% 35% 3% 14% 14% 5% 3% 7% 3% 5% 3% 5% Agriculture 14% Assembly 5% Communication 3% Education 7% Health Care 4% Lodging 5% Mfg/Food 3% Mfg/Ind. 5% Office 14% Other 3% Retail Goods/Svcs. 35% Warehouse 3% Figure 7. Commercial building share As indicated in Figure 7, retail goods and service providers represent the majority of customers, with 35% of the total in The number of commercial customers is expected to increase at an average annual rate of 1.7 percent, reaching 94,900 customers by December Much of the future commercial customer growth is expected to come from retail goods and services. Historically, this category growth is a function of the growth in residential customers. Recent trends indicate continued growth in communications and general manufacturing and small industrial categories. In 1984, customers in the commercial category consumed approximately 17 percent of Idaho Power system sales, growing to 28 percent by This share is forecast to remain at the upper end of this range throughout the planning period. Figure 8 shows historical and forecast average use per customer (UPC) for the entire category. The commercial-use-per-customer metric in Figure 8 represents an aggregated metric for a highly diverse group of customers with significant differences in total energy use per customer, but it is instructive in aggregate for comparative purposes. The UPC peaked in 2001 at 67,400 kwh and has declined at approximately 0.31 percent compounded annually to The UPC is forecast to decrease at an annual rate of 0.46 percent over the planning period. For the category as a whole, common elements that drive use down include increases in electricity prices, business-cycle recessions, and the adoption of energy efficiency technology. Within the sub-categories, the UPC varies widely from manufacturing/industrial at 159,500 kwh per customer to communications at 44,250 kwh (2014 basis). Page Integrated Resource Plan

27 Idaho Power Company Appendix A Sales and Load Forecast 100,000 90,000 80,000 70,000 Hi 60,000 50,000 40,000 30,000 20,000 10, Actual Forecast Figure 8. Forecast commercial use per customer (weather-adjusted kwh) Figure 9 shows the diversity in the commercial segments UPC as well as the trend for these sectors. The figure shows the 2014 UPC for each segment relative to the 2010 UPC. A value of 1.0 indicates the UPC has not changed over this period. The figure supports the general decline of the aggregated trend of Figure 7 but highlights differences in energy and economic dynamics within the commercial category not evident in the residential category Figure 9. Commercial categories UPC, 2014 relative to 2010 Energy efficiency implementation is a large determinant in UPC decline, particularly in high-growth categories, such as retail goods and services, communication, and office, where many structures are new and subject to efficient building code requirements. Increases in the UPC, such as in the water/agriculture (Water/Ag.) category are indicative of an increasing density of pumps and water treatment consolidation. Other influences include a difference in 2015 Integrated Resource Plan Page 19

28 Appendix A Sales and Load Forecast Idaho Power Company price sensitivity, sensitivity to business cycles and weather changes, and degree and trends in automation. In addition, aggregate commercial UPC can vary when a customer s use increases to the point where it must, by tariff rules, migrate to an industrial (Rate 19) category. The commercial-sales forecast equations consider several varying factors, as informed by the regression models, and vary depending on the sub-category. Typical variables include weather: HDD (wintertime); CDD (summertime); specific industry growth characteristics and outlook; service-area demographics and their derivatives, such as households, employment, and small business conditions; the real price of electricity; and conservation adoption. Page Integrated Resource Plan

29 Idaho Power Company Appendix A Sales and Load Forecast IRRIGATION The irrigation category is comprised of agricultural irrigation service customers. Service under this schedule is applicable to power and energy supplied to agricultural-use customers at one point-of-delivery for operating water pumping or water-delivery systems to irrigate agricultural crops or pasturage. The expected-case irrigation load is forecast to increase slowly from 213 amw in 2015 to 235 amw in 2034, an average annual compound growth rate of 0.5 percent. The expected-case, 70 th -percentile, and 90 th -percentile scenarios forecast slow growth in irrigation load from 2015 to In the 70 th -percentile scenario, irrigation load is projected to be 226 amw in 2015 and 248 amw in The individual irrigation load forecasts (Table 7 and Figure 10) illustrate the poorer economic conditions and dramatic reduction in land put into production in the mid-1980s. Table 7. Irrigation load growth (amw) Growth Annual Growth Rate th Percentile % 70 th Percentile % Expected Case % Hi Weather Adjusted Expected Case 70th Percentile 90th Percentile Figure 10. Forecast irrigation load (amw) The annual average loads in Table 7 and Figure 10 are calculated using the 8,760 hours in a typical year. In the highly seasonal irrigation sector, over 97 percent of the annual energy is billed during the six months from May through October, and nearly half of the annual energy is billed in just two months, July and August. During the summer, hourly irrigation loads can reach nearly 900 MW. In a normal July, irrigation pumping accounts for roughly 25 percent of the energy consumed during the hour of the annual system peak and nearly 30 percent of the energy 2015 Integrated Resource Plan Page 21