demand forecasts

Electric Load Forecast 2005/06 to 2025/26 Market Forecasting Power Planning and Portfolio Management Distribution Line of Business BC Hydro December 2005 Forecast PAGE i PAGE ii Preamble BC Hydro's Load Forecast is typically published annually. The forecast is a 21year projection of customers’ annual future electricity demands. The forecast also projects the highest, or peak hourly load that the BC Hydro system will experience in a given year. There are several major types of forecast uncertainties that could cause forecast error. BC Hydro undertakes several strategies to attempt to mitigate these sources of forecast error. These uncertainties are: 1. Model Uncertainty – The forecasting model which captures the relationship between the drivers and load cannot exactly match the detailed behaviour of BC Hydro’s customers or the operation of its system. BC Hydro attempts to develop stable models that capture the relationship between drivers and sales correctly; however this relationship may change during the forecast period. 2. Data Uncertainty – In any forecast there is uncertainty in the model drivers or predictors used to develop the forecast. In order to mitigate this uncertainty, BC Hydro uses reliable and credible sources of drivers of the forecast. 3. Outcome Uncertainty – external events, weather and world events such as 9/11, Severe Acute Respiratory Syndrome (SARS), trade disputes and changes in commodity markets. Additionally, unexpected structural changes in the market place, such as the introduction of new products due to changes in consumer behaviour or tastes, can impact the variance and the forecast. These three factors lead to forecast variances from actual demand. BC Hydro continuously attempts to improve the accuracy of its forecasting process through a process of internal and external validation. This testing includes backcasting, monitoring trends on forecasting approaches, and tracking developments that may affect the forecast. As a means to assess forecast uncertainties that exist in the annual long-term forecasts of electricity demand, ranges for the forecasts are developed. High and low forecast uncertainty bands are used to represent ranges around the annual long-term forecasts of electricity demand. The Load Forecast is used to provide decision-making support for several aspects of BC Hydro’s business including financial projections, revenue requirements and rate design, integrated resource planning, and BC Hydro's service plan. Annual reference forecasts along with ranges in the forecasts are developed and published. Annual forecast uncertainty bands are used to represent that various ranges around the annual reference forecasts since there is uncertainty in the variables that predict sales and the forecasting models, which establish the relationships between sales and the predictors of future sales. To support the various decision making requirements, the energy forecast is compromised of a number of components that include sales and gross requirements as discussed below. The energy sales forecast is derived by making projections of the sales to customer segments – the summation of which is the Total BC Hydro Service Area Sales. An estimate of the firm sales to other PAGE iii utilities are added to the Total BC Hydro Service Area Sales to get the Total Firm Sales. Finally, an estimate of the energy lost (commonly known as “losses”) in delivering the electricity to customers is added to the Total Firm Sales to get the Total Gross Requirements. This total does not include forecasts of future transactions of power sales by BC Hydro's marketing arm Powerex. Last year, BC Hydro’s Total Gross Requirements were 55,747 GWh The peak load forecast is derived from the actual peak load from the previous year, adjusted to reflect average weather conditions. Peak forecasts include an allowance for growth in the demand requirements based on a number of predictors. BC Hydro's total peak demand for the last full winter (2004/05) was 10,110 MW, after adjustments to reflect average cold winter weather conditions. The Load forecast for both energy and peak is prepared using the most recent information on economic projections from the public sector (i.e. the B.C. Ministry of Finance) and augmented by information from private sector consultants. The main variables or predictors of future load growth are: • • • • • Housing starts, Gross Domestic Product (GDP) Growth, DSM (Demand Side Management) Weather; and Electricity Rates. These predictors are applied to make forecasts of sales for the customer segments (residential, commercial and industrial) as follows: • The residential customer segment represents approximately 30 per cent of BC Hydro's of total energy sales volume to all customer segments. The sales forecast reflects future projections of residential accounts, driven by housing starts, and the expected average usage per account. Variability in the weather also impacts residential sales and peak demand requirements. To account for this, BC Hydro prepares its residential and peak forecast with adjustments reflecting average historical weather conditions. The commercial segment represents approximately 30 per cent of BC Hydro’s total energy sales volume. This segment closely tracks economic growth or GDP. Forecasts of future growth in GDP are used to develop a forecast for commercial sales. The industrial segment represents approximately 40 per cent of BC Hydro’s total energy sales volume. Industrial customer sales are highly correlated to GDP; therefore forecasts of GDP growth are used to project industrial sales. Sales in all segments are impacted in varying amounts by electricity rates. This Load Forecast is based on currently approved electricity rates, adjusted in future years to reflect inflation. The Load Forecast is normally updated following any changes in BC Hydro’s approved revenue requirements and electricity rates. In addition, BC Hydro uses scenario analysis to assess how changes in electricity rates could impact demand. Typically, if electricity rates increase, the initial response is a small reduction in demand. The opposite is also typically true; if prices fall, demand rises. If natural gas prices were to remain very high on a sustained basis, there may be more substitution leading to higher electricity consumption. The level of demand response is different for • • • PAGE iv each customer segment. There are limitations on the accuracy of demand response analysis, given experience to date. • Sales in all segments are also impacted by the effects of initiatives to conserve electricity (Demand Side Management or DSM). Forecasts are prepared both with and without the impact of planned DSM initiatives. In some segments the impacts of DSM can be accurately estimated because they involve changes in equipment that can be metered or verified. In other segments, they involve changes that are verified through surveys of a sample of participants and non-participants. For this reason, there is some uncertainty regarding some DSM results, which is reflected in the overall forecast uncertainty band. The uncertainty around DSM results is minimized through comprehensive monitoring, verification and evaluation activities. • Expected total gross requirements for fiscal 2005/06 are 57,537 GWh with DSM savings. This is above last year's forecast for fiscal 2005/06 by 1,110 GWh or 2.0 per cent. Five years into the forecast, the difference between this year's forecast and the previously published forecast is 2,488 GWh or 4.3 per cent, after 11 years, the difference is 2,297 GWh or 3.6 per cent and after 20 years the difference is 2,312 GWh or 3.1 per cent. The increase in the forecast reflects: 1. An increase in historical sales and GDP. For example, industrial electricity sales grew by 4.9 per cent last year and GDP grew by 3.9 per cent. These latest increases and previous growth trends in industrial sales and its predictor (i.e., GDP) form the basis for establishing the current forecast. The growth in sales is attributed to higher commodity prices and a recovery in B.C's mining sector. The growth in GDP has been a result of strong demand for BC exports, higher consumer confidence and consumption, and favorable world economic conditions. 2. An increase in the variables that drive the forecast (e.g. GDP, housing starts). In particular, the previous forecast assumed GDP would grow 3.1 per cent in 2005 and the current forecast used by BC Hydro is for GDP growth in 2005 of 3.4 per cent. A higher GDP forecast has resulted in a higher load forecast. Since load projections are based on estimates of future conditions, there is inherent uncertainty in the annual point forecasts referred to above. To assess the uncertainty around the current forecast, BC Hydro undertakes a sensitivity analysis on the drivers discussed above, including effects of weather variability, to produce high and low electricity forecasts. The high and low forecasts provide an estimated range of values under the assumption that actual sales will be within the range 80 per cent of the time. The estimated ranges for total gross requirements forecast (GWh) with DSM for the current year, the next 11 years and the next 21 years are: PAGE v High Fiscal 2005/06 Fiscal 2015/16 Fiscal 2025/26 58,330 68,373 82,127 Most Likely (Reference) 57,537 65,775 77,024 Low 56,715 63,313 72,835 The long-term forecast information is used for long-term planning, as contained in the IEP, and it also is used to support the need for future infrastructure, as documented in CPCNs and capital plans. Over the long-term, there is more uncertainty around the accuracy of the drivers and this is reflected in the larger width of the uncertainty band. The actual drivers of the forecast may differ from their forecast due to several factors including: fluctuations in the economy (i.e. business cycles), changes in external demand for BC’s raw materials, and structural changes such as an increase in demand for electronics which impact use rate patterns and projections. These changes create longer-term risks such as over or under investing in infrastructure or potential contracting new generation for too much or too little power. In the short-term, the forecast is used primarily for financial projections of revenues and as an input into BC Hydro’s Marginal Cost model for resource optimization. In this case, the forecast uncertainty band tends to be smaller recognizing that near-term projections of drivers such as GDP, industry trends, account growth and DSM should be more accurate. Changes in the key drivers of the load and unexpected events contribute to short-term variances in the forecast and can also create significant financial risks, such as revenues falling below projected revenues or increased energy purchase required to meet unexpected load. Historical variance information is recorded and traced back to the sources as accurately as possible. In the short-term, variance analysis of the sales and the drivers is used to make adjustments to both the short-term and the long-term forecast projections, if necessary. Table P1.1 below shows the variance of actual sales to the planned billed sales projections in the previously filed 2005 and 2006 revenue requirements. Table P1.1 Total Firm Sales Variance F2004 Forecast with DSM as per 2005 and 2006 RRA Actual Total Variance DSM Variance Model and Driver Variance Notes: 1. Actual sales for F2006 are projected based on 11 months of actual sales and 1 month of forecast. 2. DSM variance for F2006 is projected based on 6 months actuals and 6 months forecast. F2005 49,323 51,088 1,765 54 1819 F2006 49,637 52,1731 2,5362 165 2,701 49,180 50,273 1,093 -2 1,091 PAGE vi The current projected growth rate of total firm sales, with DSM, over the next five years is 1.6%. This compares closely with the previous five-year historical growth rate of 1.8%. Even though the near term projected sales growth is close to historical sales growth, changes in export markets, domestic labor conditions, international trade policy and domestic policy on regulation of BC’s key industries as well as changes in DSM plans for large customers, can lead to significant volatility in sales. For example, industrial sales declined by 4.5% between F2001 and F2002 and more recently increased by 4.9% between F2004 and F2005. PAGE vii Table of Contents Preamble. .............................................................................................................iii Highlights............................................................................................................xiii Background and Context.....................................................................................xv Sectors and Methodology...................................................................................xvi Residential Forecast...........................................................................................xvi Commercial Forecast ........................................................................................xvii Industrial Forecast.............................................................................................xvii Gross Requirements ........................................................................................ xviii Peak Demand................................................................................................... xviii Energy and Peak Forecast Before DSM (Demand Side Management) .............xix Sensitivity Analysis and Risks ............................................................................xxi 1 Introduction........................................................................................................... 1 1.1. Background and Context ...................................................................... 1 1.2. Role of Forecasting at BC Hydro .......................................................... 1 Regulatory Background........................................................................................ 3 Forecast Drivers, Data Sources and Assumptions............................................... 4 3.1. Forecast Drivers ................................................................................... 4 3.2. Growth Assumptions............................................................................. 5 Forecast Process and Methodologies .................................................................. 7 4.1. Residential Forecast Methodology ....................................................... 7 4.2. Commercial Forecast Methodology ...................................................... 7 4.3. Industrial Forecast Methodology........................................................... 8 4.4. Peak Forecast Methodology ................................................................. 8 Reference Forecast............................................................................................ 11 5.1. Reference Forecast Before DSM........................................................ 11 2 3 4 5 Table 5.1. Reference Forecast Before DSM .............................................................. 12 5.2. Reference Forecast With DSM ........................................................... 13 6 Comparison Between 2004 and 2005 Forecasts ............................................... 16 6.1. Total Gross Requirements.................................................................. 16 6.2. Integrated Gross Requirements.......................................................... 17 Table 6.2. Comparison of Integrated Gross System Requirements with DSM ......... 17 6.3. Total Integrated Peak Sales ............................................................... 18 Table 6.3. Comparison of Reference Peak Forecasts With DSM (Integrated System) ........................................................................................................................... 18 7 Sensitivity Analysis............................................................................................. 19 7.1. Monte Carlo Analysis.......................................................................... 21 7.2. Uncertainty Assumptions .................................................................... 21 Residential Forecast........................................................................................... 27 8.1. Summary ............................................................................................ 27 8.2. Forecast Overview.............................................................................. 28 8.3. Forecast Methodology ........................................................................ 29 8.4. Residential Forecast Comparison....................................................... 29 Commercial Forecast ......................................................................................... 33 9.1. Summary ............................................................................................ 33 9.2. Major Trends....................................................................................... 34 9.3. Methodology ....................................................................................... 37 9.4. Commercial Forecast Comparison ..................................................... 37 8. 9 PAGE viii 10 Industrial Forecast.............................................................................................. 38 10.1. Summary ............................................................................................ 38 10.2.1. Medium-Term Forestry Outlook.......................................................... 39 10.2.2. Medium-Term Pulp and Paper Outlook .............................................. 39 10.2.3. Medium-Term Mining Outlook ............................................................ 40 10.3. Methodology ...................................................................................... 41 10.4. Industrial Forecast Comparison.......................................................... 41 10.5. Risks and Uncertainties ...................................................................... 44 Peak Forecast .................................................................................................... 45 11.1 Introduction ......................................................................................... 45 11.2 Methodology and Procedure............................................................... 46 11.3 Peak Forecast Comparison ................................................................ 47 11 12 Demand Side Management .................................................................................. 49 12.1. Demand Side Management Background............................................ 49 12.2. Forecast.............................................................................................. 49 13 14 Glossary ............................................................................................................. 50 References ......................................................................................................... 53 Appendix 1. Weather Normalization........................................................................... 54 Appendix 2. Ordinary Least Squares-Based Forecasts ............................................. 56 Appendix 3. Maximum Likelihood-Based Forecasts .................................................. 58 Appendix 4. Commercial Sector Regressions............................................................ 61 Appendix 5. Industrial Sector Regressions ................................................................ 63 Appendix 6. Monte Carlo Methods ............................................................................. 68 Appendix 7. Forecast Tables ..................................................................................... 72 Table A7.1. Regional Non-Coincident and Coincident Distribution Peaks Before DSM ........................................................................................................................... 73 Table A7.2. Regional Non-Coincident and Coincident Transmission Peaks Forecast Before DSM........................................................................................................ 74 Table A7.3. Regional Non-Coincident and Coincident Distribution Peaks Forecast With DSM ........................................................................................................... 75 Table A7.4. Regional Non-Coincident and Coincident Transmission Peaks Forecast With DSM ........................................................................................................... 76 Table A7.9. 2005 BC Hydro, Reference Load Forecast Before DSM ........................ 81 Table A7.10. 2005 BC Hydro, Reference Load Forecast With DSM.......................... 82 Table A7.11. 2005 BC Hydro, High Load Forecast Before DSM ............................... 83 Table A7.12. 2005 BC Hydro, Low Load Forecast Before DSM ................................ 84 Table A7.12. 2005 BC Hydro, Low Load Forecast Before DSM ................................ 84 Table A7.13. 2005 BC Hydro, High Load Forecast With DSM................................... 85 Table A7.14. 2005 BC Hydro, Low Load Forecast With DSM ................................... 86 PAGE ix Tables Table P1.1 Total Firm Sales Variance .......................................................................vi Table 1. Energy and Peak Forecast Before DSM for Selected Years.........................xx Table 2. Energy and Peak Forecast With DSM for Selected Years ............................xx Table 2.1. B.C. Utilities Commission Comments and Actions...................................... 3 Table 3.1. Key Forecast Drivers................................................................................... 4 Table 3.2. Growth Assumptions (Annual rate of growth).............................................. 5 Table 3.3. Data Sources and Uses for Growth Assumptions ....................................... 6 Table 5.1. Reference Forecast Before DSM .............................................................. 12 Table 5.2. Reference Forecast With DSM ................................................................. 14 Table 6.2. Comparison of Integrated Gross System Requirements with DSM ......... 17 Table 6.3. Comparison of Reference Peak Forecasts With DSM (Integrated System) ............................................................................................................................ 18 Table 7.1 Monte Carlo Analysis – Energy and Peak Before DSM ............................. 23 Table 7.2 Monte Carlo Analysis – Energy and Peak With DSM................................. 25 Table 8.1. Residential Sales Before DSM .................................................................. 31 Table 8.2. Residential Sales With DSM ..................................................................... 32 Table 9.1. Commercial Sales Before DSM (GWh) ..................................................... 35 Table 9.2. Commercial Sales With DSM (GWh) ........................................................ 36 Table 10.1. Industrial Sales by Sector Before DSM (GWh) ....................................... 42 Table 10.2. Industrial Sales by Sector With DSM (GWh)........................................... 43 Table 11.1 Integrated System Peak Before and With DSM ...................................... 48 Table 12.1 Forecast Demand Side Management Savings......................................... 49 Table A1.1. Actual and Weather-Normalized Sales for BC Hydro Service Territory.. 55 Table A4.1. Econometric Model of Commercial Sales ............................................... 61 Table A4.2 Forecast Total Commercial Sales Before DSM (GWh)........................... 62 Table A5.1. Econometric Model of Industrial Transmission Sales (Model 1) ............. 63 Table A5.2. Econometric Model of Industrial Transmission Sales (Model 2) ............. 64 Table A5.3. Econometric Model of Industrial Distribution Sales (Model 1) ................ 64 Table A5.4. Econometric Model of Industrial Distribution Sales (Model 2) ................ 65 Table A5.5. Forecast Industrial Sales Before DSM (GWh) Model 1 .......................... 66 Table A5.6. Forecast Industrial Sales Before DSM (GWh) Model 2 .......................... 67 Table A7.1. Regional Non-Coincident and Coincident Distribution Peaks Before DSM ............................................................................................................................ 73 Table A7.2. Regional Non-Coincident and Coincident Transmission Peaks Forecast Before DSM ........................................................................................................ 74 Table A7.3. Regional Non-Coincident and Coincident Distribution Peaks Forecast With DSM............................................................................................................ 75 PAGE x Table A7.4. Regional Non-Coincident and Coincident Transmission Peaks Forecast With DSM............................................................................................................ 76 Table A7.5. Domestic System and Regional Peak Forecast Before DSM ................. 77 Table A7.6. Domestic System and Regional Peak Forecast with DSM ..................... 78 Table A7.7. Weather-Adjusted Peak Forecast Before DSM ...................................... 78 Table A7.8. Weather-Adjusted Peak Forecast With DSM.......................................... 80 Table A7.9. 2005 BC Hydro, Reference Load Forecast Before DSM ........................ 81 Table A7.10. 2005 BC Hydro, Reference Load Forecast With DSM.......................... 82 Table A7.11. 2005 BC Hydro, High Load Forecast Before DSM ............................... 83 Table A7.12. 2005 BC Hydro, Low Load Forecast Before DSM ................................ 84 Table A7.12. 2005 BC Hydro, Low Load Forecast Before DSM ................................ 84 Table A7.13. 2005 BC Hydro, High Load Forecast With DSM................................... 85 Table A7.14. 2005 BC Hydro, Low Load Forecast With DSM ................................... 86 PAGE xi Figures Figure 7.1. Sources of Forecast Uncertainty.............................................................. 20 Figure 7.2. Total Gross Requirements with DSM....................................................... 20 Figure 7.2. Monte Carlo Analysis – Energy before DSM............................................ 24 Figure 7.3. Monte Carlo Analysis – Energy with DSM ............................................... 26 PAGE xii Executive Summary Highlights • Economic Outlook. Real economic growth in British Columbia is forecast to be about 3.4% in 2005 and to average about 3.2% per year for the five years from 2005 to 2009. This compares with real economic growth of 3.9% in 2004 and an average of 3.0% per year for the five years from 2000 to 2004. Over the first ten years of the forecast period, average forecast Gross Domestic Product (GDP) growth for British Columbia is 2.8% per year. The markets for pulp and paper are reasonably stable, although prices are currently below their long-term trends. The markets for the other main exports including wood and wood products, copper and coal all continue to be strong. This strong outlook for the external and domestic economic environment has strong implications for increased electricity demand and sales in British Columbia. Sales Forecast before DSM (Demand Side Management) 1. Billed domestic sales, which include sales to the residential, commercial and industrial sectors, as well as sales to New Westminster and Fortis BC, were approximately 50,800 GWh in 2004/05. Billed domestic sales are forecast to grow from about 50,800 GWh in 2004/05 to 57,000 GWh in 2009/10, 62,000 GWh in 2015/16, and 71,900 GWh in 2025/26. These increases represent growth rates of 2.3% over the next five years (2004/05 to 2009/10), 1.8% over the next 11 years (2004/05 to 2015/16), and 1.7% over the next 21 years of the forecast (2004/05 to 2025/26). Sales Forecast with DSM (Demand Side Management). Billed domestic sales, with DSM reductions, are forecast to grow from about 50,800 GWh in 2004/05 to 54,900 GWh in 2009/10, 59,500 GWh in 2015/16, and 69,700 GWh in 2025/26. These increases represent growth rates of 1.6% over the next five years (2004/05 to 2009/10), 1.5% over the next 11 years (2004/05 to 2015/16), and 1.5% over the next 21 years of the forecast (2004/05 to 2025/26). Peak Demand before DSM (Demand Side Management). Peak demand is composed of the demand for electricity at the distribution level, transmission level plus inter-utility sales and transmission and distribution losses. Total integrated system peak was 9,762 MW (10,110 MW weather-adjusted 2) in 2004/05. Forecast peak before DSM, is expected to grow from 9,762 MW (10,110 MW weatheradjusted) in 2004/05 to 11,151 MW in 2009/10, 11,875 MW in 2015/16, and 13,756 MW in 2025/26. These increases represent growth rates, on a weatheradjusted basis, of 2.0% over the next five years (2004/05 to 2009/10), 1.5% over the next 11 years (2004/05 to 2015/16), and 1.5% over the next 21 years of the forecast (2004/05 to 2025/26). Peak Demand with DSM (Demand Side Management). Forecast total integrated system peak with DSM, is expected to grow from 9,762 MW (10,110 MW weather-adjusted) in 2004/05 to 10,821 MW in 2009/10, 11,487 MW in 2015/16, and 13,413 MW in 2025/26. These increases represent growth rates, on a weather-adjusted basis, of 1.4% over the next five years (2004/05 to 2009/10), • • • • 1 2 Demand Side Management is defined in the Glossary. Weather has the greatest impact on the residential sector sales and peak demand. To account for the variability of weather on sales and demand, residential sales and peak demand are adjusted to reflect normal weather conditions. Normal weather is an average of previous weather conditions. PAGE xiii 1.2% over the next 11 years (2004/05 to 2015/16), and 1.4% over the next 21 years of the forecast (2004/05 to 2025/26). • Rate Impacts and Elasticities. The Reference Forecast assumes that electricity rates will be constant in real terms for the entire of the forecast period. Further analysis on the rate change impact will be published. BC Hydro uses scenario analysis to assess how changes in electricity rates could impact demand. Typically, if electricity rates increase, the initial response is a small reduction in demand. The opposite is also typically true; if prices fall, demand rises. The level of demand response is different for each customer segment. Gross Requirements before DSM (Demand Side Management). Gross requirements include sales to BC Hydro’s residential, commercial and industrial customers, sales to other utilities, non-integrated areas, and transmission and distribution losses. Gross requirements were approximately 55,700 GWh in 2004/05. Gross requirements, before DSM impacts, are forecast to grow from about 55,700 GWh in 2004/05 to 62,900 GWh in 2009/10, 68,500 GWh in 2015/16, and 79,400 GWh in 2025/26. These increases represent growth rates of 2.4% over the next five years (2004/05 to 2009/10), 1.9% over the next 11 years (2004/05 to 2015/16), and 1.7% over the next 21 years of the forecast (2004/05 to 2025/26). Gross Requirements with DSM (Demand Side Management). Gross requirements, with DSM impacts, are forecast to grow from about 55,700 GWh in 2004/05 to 60,700 GWh in 2009/10, 65,800 GWh in 2015/16, and 77,000 GWh in 2025/26. These increases represent growth rates of 1.7% over the next five years (2004/05 to 2009/10), 1.5% over the next 11 years (2004/05 to 2015/16), and 1.6% over the next 21 years of the forecast (2004/05 to 2025/26). Forecast Uncertainty Ranges. BC Hydro’s load forecast is sensitive to a number of variables including weather, economic conditions, and electricity prices. Since load projections are based on estimates of future conditions, there is inherent uncertainty in the forecasts of sales and total gross requirements. To assess the uncertainty around the current forecast, BC Hydro undertakes a sensitivity analysis on drivers of the forecast, including effects of weather variability, to estimate high and low electricity forecast ranges. The high and low forecasts provide an estimated range of values under the assumption that actual sales will be within the range 80 per cent of the time. The high forecast for Gross requirements, before DSM impacts, is forecast to grow from about 55,700 GWh in 2004/05 to 64,400 in 2009/10, 71,000 in 2015/16, and 84,500 in 2025/26. The low forecast for Gross requirements, before DSM impacts, is forecast to grow from about 55,700 GWh in 2004/05 to 61,500 in 2009/10, 66,000 in 2015/16, and 75,200 in 2025/26. The high forecast for Gross requirements, with DSM impacts, is forecast to grow from about 55,700 GWh in 2004/05 to 62,100 in 2009/10, 68,400 in 2015/16, and 82,100 in 2025/26. The low forecast for Gross requirements, with DSM impacts, is forecast to grow from about 55,700 GWh in 2004/05 to 59,200 in 2009/10, 63,300 in 2015/16, and 72,800 in 2025/26. • Factors Leading to Lower than Forecast Sales. The main factors that could lead to lower than forecast sales include the following: • Increase in the value of the Canadian dollar could reduce growing commodity exports from BC, with the pulp and paper sector particularly vulnerable if the Canadian dollar rises above US $0.90 on a sustained basis; • • • PAGE xiv • Rising interest rates in the United States and Canada could reduce North American housing demand and the demand for BC lumber, although in the near term this would be offset by the reconstruction efforts following Hurricane Katrina; and Reduction in growth in China could lead to a slowing of commodity demand and commodity prices for base metals. • • Factors Leading to Higher than Forecast Sales. The main non-economic risks that could lead to higher than forecast sales to the forecast include assumptions about load reductions due to elasticity response to price increases applied to the forecast that do not materialize. The main economic risks that could lead to higher sales are as follows: • Strengthening world demand for market pulp and energy-intensive paper grades could increase electricity demand in the pulp and paper sector, although the BC industry appears to be operating close to full capacity; Resolution of the softwood lumber dispute at a time when there is an ample supply of fibre in the interior of BC could increase electricity demand in the sawmilling sector; Strengthening business confidence could increase investment in BC leading to increased industrial and commercial demand for electricity; and Economic spillover of the Winter Olympics could increase investment and economic activity in British Columbia with consequent impact on electricity sales. • • • • • Comparison With 2004 Forecast. The current domestic sales forecast, before DSM, is higher than the December 2004 Forecast, before DSM, for all years of the forecast period. The main reasons for the increase in the forecast are a significant uplift in commercial and industrial sales forecasts compared to the previous forecast. Sales to these sectors follow closely with economic conditions. Strong demand for BC commodities, a robust housing market and an increase in consumer confidence and spending has lead to growth in GDP. Projections of economic growth, measured by GDP, are used to estimate future growth in electricity sales. Recent projections of GDP have increased relative to previous estimates and this has led to a higher sales projection. The residential forecast has also increased relative to the previous forecast, which reflects a slight increase in the projected residential use rate over the long run. Overall, the current forecast domestic sales, before DSM, for 2005/06 are 52,400 GWh compared to the forecast of 52,000 GWh for 2005/06 in the 2004 forecast. Background and Context BC Hydro is the third largest utility in Canada and generates nearly 80% of the electricity produced in British Columbia. The company’s generating capacity is over 11,000 MW and gross requirements including losses were over 55,700 GWh in 2004/05. Sales to BC Hydro’s residential, commercial, industrial and all other utilities were over 51,000 GWh in 2004/05. Load forecasting is central to BC Hydro’s long-term planning, medium-term investment and short-term and operational and forecasting activities. The BC Hydro PAGE xv Electric Load Forecast is produced annually and published in the fall. The forecast is based on several comprehensive end-use and econometric models that use billed data up to March 31 of the relevant year as historical information, combined with a wide variety of economic forecasts and inputs from internal, governmental and third party sources. The forecast outputs are validated through additional tests and information including time series econometric models. The primary purpose of the electric load forecast is to provide decision-making support on the questions of “where, when and how much” electricity is expected to be required on the BC Hydro system. BC Hydro’s load forecasting activities centre on the production of a number of termspecific and location-specific forecasts of energy sales and peak demand requirements to meet user needs for decision-support information. A variety of related products including quarterly forecast updates, monthly variance reports, inputs for the revenue forecast, load shape analysis and small area forecasts are produced to supplement the base forecasts presented in this report. Sectors and Methodology BC Hydro’s load forecast is assembled from the following components: the residential forecast; the commercial forecast (distribution voltage and transmission voltage), the industrial forecast (distribution voltage and transmission voltage) and the peak forecast. A variety of end-use and econometric models are used to produce the forecast. These specific forecast methods are based on their predictive value and their ability to most appropriately meet the needs of users. Residential Forecast Of the three customer classes, residential, commercial, and industrial, the residential sector is the most stable. Over the longer term, growth in the number of residential accounts is currently about 1.5%. After many years of growth in use per account, growth in use per account is forecast to moderate. Key features of the residential forecast include the following: • Electricity Use - BC Hydro’s residential sector currently consumes about 31% of BC Hydro’s total annual billed sales. This electricity is used to provide a range of services including space heating, water heating, refrigeration, and miscellaneous plug-in load which includes computer equipment and home entertainment systems. Drivers - The drivers of the residential forecast are number of accounts and average annual use per account. Number of accounts is driven by estimates of housing starts. With growth expected in both the number of residential accounts and use per account, annual growth in sales with DSM is forecast to average 1.8 % over the entire forecast period. Trends – Sales to the residential sector, before weather adjustment, fell by about 1.8% in 2004/05, but on a weather adjusted basis, sales grew by 0.3% in 2004/05. Compared to last year’s forecast, there is a slight reduction in residential sales. This reduction is due to a decline in the number of residential accounts. Use rate is forecast to grow at a rate of less than 1% over the 21-year term. Forecast - Billed residential sales were approximately 15,600 GWh in 2004/05. Forecast sales to residential customers, before DSM, are expected to grow from about 15,600 GWh in 2004/05 to 17,900 GWh in 2009/10, 19,900 GWh in 2015/16, and 23,200 GWh in 2025/26. These increases represent growth rates of 2.8% over the next five years (2004/05 to 2009/10), 2.2% over the next 11 years • • • PAGE xvi (2004/05 to 2015/16), and 1.9% over the next 21 years of the forecast (2004/05 to 2025/26). Please refer to Table 2 for the forecast summary with DSM. Commercial Forecast BC Hydro’s commercial sector encompasses a wide a variety of commercial and publicly provided services. It includes a diverse set of BC Hydro customers who operate a wide range of facilities such as office buildings, retail stores, institutions (i.e., hospitals and schools) and transportation infrastructure. The remainder of the sector is comprised of facilities (non-buildings) such as transportation infrastructure and public utilities. Key features of the commercial forecast include the following: • Electricity Use– BC Hydro’s commercial sector currently consumes 28% of BC Hydro’s total annual billed sales. This electricity is used to provide a range of services (often called end-uses) such as lighting, ventilation, heating, cooling, refrigeration, hot water, etc. These needs vary considerably between the different types of buildings. Drivers – Commercial sales are related to growth in total floor area by building type and end uses. Consumption in the commercial sector is also tied closely with economic activity in the province; there is a strong relationship between consumption and economic indicators such as the provincial GDP and employment rates. As a result, future economic trends are a good indicator of future electricity consumption in the commercial sector. Trends – Electricity consumption of the commercial sector can vary considerably from year to year reflecting the level of activity in BC’s service sector. Sales to BC Hydro’s commercial sector grew by 1.5% in 2004/05, reflecting the strong performance of the BC economy. Growth consistent with a strong economy is expected to continue. Forecast – Billed commercial sales were approximately 14,400 GWh in 2004/05. Commercial sales, before DSM impacts, are forecast to grow from about 14,400 GWh in 2004/05 to 16,500 GWh in 2009/10, 18,600 GWh in 2015/16, and 22,000 GWh in 2025/26. These increases represent growth rates of 2.8% over the next five years (2004/05 to 2009/10), 2.4% over the next 11 years (2004/05 to 2015/16), and 2.1% over the next 21 years of the forecast (2004/05 to 2025/26). • • • Please refer to Table 2 for the forecast summary with DSM. Industrial Forecast BC Hydro’s industrial sector is concentrated in a limited number of industries, the most important of which are pulp and paper, wood products, chemicals, metal mining and coal mining. The remaining industrial load is made up of a large number of small and medium sized manufacturing establishments. Key features of the industrial forecast include the following: • Electricity Use - BC Hydro’s industrial sector currently consumes 38% of BC Hydro’s total annual billed sales. This electricity is used to provide a range of services including fans, pumps, compression, conveyance, processes such as cutting, grinding, stamping and welding and electrolysis. Drivers – As in the case of the commercial sector, industrial electricity consumption is tied closely with the level of economic activity in the province. In other words, there is a strong relationship between industrial electricity • PAGE xvii consumption and provincial Gross Domestic Product. Future economic trends are a good indicator of future electricity consumption in the industrial sector. • Trends - Electricity consumption in the industrial sector is quite volatile, driven substantially by economic conditions in the United States, China and Japan that affect commodity markets. Export sales to these three countries are a key determinant of domestic industrial output and of the industrial demand for electricity. Trends for sales to these markets are positive. Sales to BC Hydro’s industrial sector grew by 4.9% in 2004/05, reflecting the strong performance of the BC economy. Forecast - Billed industrial sales were approximately 19,600 GWh in 2004/05. Forecast sales to industrial customers, before DSM, are expected to grow from about 19,600 GWh in 2004/05 to 21,300 GWh in 2009/10, 22,100 GWh in 2015/16, and 25,100 GWh in 2025/26. These increases represent growth rates of 1.6% over the next five years (2004/05 to 2009/10), 1.1% over the next 11 years (2004/05 to 2015/16), and 1.2% over the next 21 years of the forecast (2004/05 to 2025/26). • Please refer to Table 2 for the forecast summary with DSM. Gross Requirements Gross requirements were approximately 55,700 GWh in 2004/05. Gross requirements, before DSM impacts, are forecast to grow from about 55,700 GWh in 2004/05 to 62,900 GWh in 2009/10, 68,500 GWh in 2015/16, and 79,400 GWh in 2025/26. These increases represent growth rates of 2.4% over the next five years (2004/05 to 2009/10), 1.9% over the next 11 years (2004/05 to 2015/16), and 1.7% over the next 21 years of the forecast (2004/05 to 2025/26). Forecast transmission losses have been set at 7% of billed sales and forecast distribution losses have been set at 4% of billed sales. These loss factors combine to provide an estimate of total losses that is consistent with historical trends. The non-integrated area (NIA) forecast is the difference between the total gross requirements forecast and the total integrated requirements forecast as stated in the tables A7.9 and A7.10 in Appendix 7. The non-integrated consist of communities that are not connected to BC Hydro’s transmission grid. The areas included in the nonintegrated sales forecast are Masset, Sandspit, Atlin, Dease Lake, Eddontenajon, Telegraph Creek, Anahim Lake, Bella-Bella, Bella Coola, and Fort Nelson. The nonintegrated sales, including Fort Nelson, are about 0.5% of the Total Gross Requirements. The non-integrated sales, before DSM impacts, are forecast to grow from 294 GWh in 2004/05, 318 GWh in 2009/10, 352 GWh in 2015/16 and 403 GWh in 2025/26.These increases represent growth rates of 1.6% over the next five years (2004/05 to 2009/10), 1.7% over the next 11 years (2004/05 to 2015/16), and 1.5% over the next 21 years of the forecast (2004/05 to 2025/26). Please refer to Table 2 for the forecast summary with DSM. Peak Demand Peak demand is composed of the demand for electricity at the distribution level, transmission level plus inter-utility sales and transmission losses on the integrated system (excluding NIA). Key features of the peak forecast include the following: • Electricity Use – Peak demand is forecast as the maximum expected one-hour demand during the year. For BC Hydro’s load, this peak demand occurs in winter with the peak driven particularly by space heating load. BC Hydro peak forecast is PAGE xviii based on the demand attributed to the average coldest day of the year, over the most recent 30 years; • Drivers – Key drivers of electricity peak include the level of economic activity, number of accounts, use rate per account, and the historical winter day cold temperatures. Trends – Since peak demand is defined as the extreme event of the maximum expected one hour load, it is more volatile than electricity sales. This volatility can mask trends in peak demand growth, as the peak can move up or down substantially from year to year in response to a few days of extreme winter weather. Nevertheless, the long-term trend in peak is strongly upward. Forecast – Total integrated system peak was 9,762 MW (10,110 MW weatheradjusted) in 2004/05. Forecast peak before DSM, is expected to grow from 9,762 MW (10,110 MW weather-adjusted) in 2004/05 to 11,151 MW in 2009/10, 11,875 MW in 2015/16, and 13,756 MW in 2025/26. These increases represent growth rates of 2.0% over the next five years (2004/05 to 2009/10), 1.5% over the next 11 years (2004/05 to 2015/16), and 1.5% over the next 21 years of the forecast (2004/05 to 2025/26). • • Please refer to Table 2 for the forecast summary with DSM. Energy and Peak Forecast Before DSM (Demand Side Management) Table 1 and Table 2 provides a summary of historical and forecast sales and peak for selected years, both before and with DSM. PAGE xix Table 1. Energy and Peak Forecast Before DSM for Selected Years Residential (GWh) Commercial (GWh) Industrial (GWh) Total Domestic Sales (GWh) 46,376 50,787 52,426 56,970 61,978 71,894 Total Gross Requirements (GWh) 51,534 55,747 58,109 62,917 68,456 79,375 Total Integrated System Peak * (MW) 8,646 9,762 (10,110) 10,298 11,151 11,875 13,756 1999/00 2004/05 2005/06 2009/10 2015/16 2025/26 Growth Rates 3 5 years: 99/00 to 04/05 5 years: 04/05 to 09/10 11 years: 04/05 to 15/16 21 years: 04/05 to 25/26 14,572 15,620 16,245 17,916 19,935 23,182 13,176 14,362 14,859 16,493 18,550 22,008 17,890 19,635 20,237 21,307 22,120 25,148 1.4% 2.8% 2.2% 1.9% 1.7% 2.8% 2.4% 2.1% 1.9% 1.6% 1.1% 1.2% 1.8% 2.3% 1.8% 1.7% 1.6% 2.4% 1.9% 1.7% 3.2% 2.0% 1.5% 1.5% * Values shown in brackets are weather normalized actuals. Unless otherwise noted, the forecast of the growth rates for the peak is on a weather-adjusted basis. Table 2. Energy and Peak Forecast With DSM for Selected Years Residential (GWh) Commercial (GWh) Industrial (GWh) Total Domestic Sales (GWh) 46,376 50,787 51,898 54,913 59,520 69,749 Total Gross Requirements (GWh) 51,534 55,747 57,537 60,676 65,775 77,024 Total Integrated System Peak* (MW) 8,646 9,762 (10,110) 10,205 10,821 11,487 13,413 1999/00 2004/05 2005/06 2009/10 2015/16 2025/26 Growth Rates 4 5 years: 99/00 to 04/05 5 years: 04/05 to 09/10 11 years: 04/05 to 15/16 21 years: 04/05 to 25/26 14,572 15,620 16,116 17,613 19,469 22,536 13,176 14,362 14,786 16,143 18,171 21,629 17,890 19,635 19,911 19,903 20,507 24,028 1.4% 2.4% 2.0% 1.8% 1.7% 2.4% 2.2% 2.0% 1.9% 0.3% 0.4% 1.0% 1.8% 1.6% 1.5% 1.5% 1.6% 1.7% 1.5% 1.6% 3.2% 1.4% 1.2% 1.4% * Values shown in brackets are weather normalized actuals. Unless otherwise noted, the forecast of the growth rates for the peak is on a weather-adjusted basis. 3 4 Unless otherwise noted, growth rates are calculated as annual compound growth rates. Unless otherwise noted, growth rates are calculated as annual compound growth rates. PAGE xx Sensitivity Analysis and Risks BC Hydro’s load forecast is sensitive to a number of variables, for example weather, economic conditions, and price. A composite sensitivity analysis (Monte Carlo study) has been completed to look at the sensitivity of the load to a combination of six causal factors that affect the forecast. This composite sensitivity analysis is used to derive an uncertainty band around the reference forecast. Six major causal factors were analyzed to determine the range of forecasts that produce an 80 per cent confidence level encompassing the reference forecast. For more detail on Monte Carlo methods see Appendix 6. PAGE xxi 1 Introduction 1.1. Background and Context BC Hydro is the third largest electric utility in Canada and generates nearly 80 per cent of the electricity produced in British Columbia. B.C. Hydro’s generating capacity is over 11,000 MW, with about 90 per cent of this capacity consisting of hydroelectric generation and the balance thermal generation. The remainder of the provincial electric generation capacity includes Alcan’s Kemano facility, Fortis BC’s plants, industry self-generation, particularly in the forest products sector, independent power producers, and small, off-grid installations, particularly in the northern part of British Columbia. BC Hydro’s Gross electricity requirements, including losses were 55,747 GWh in 2004/05, while firm sales were 51,088 GWh in 2004/05. The BC Hydro Electric Load Forecast is produced annually and normally published in the late fall. The forecast is based on comprehensive model runs that use billed data up to March 31 of the relevant year as historical information, combined with a wide variety of forecasts and inputs from internal, governmental and third party sources. In addition to reflect short term trends, this year’s forecast has been adjusted to reflect actual sales and variance information over the current fiscal year. The primary purpose of the electric load forecast is to provide decision-making support on the questions of “where, when and how much” electricity is expected to be required on the BC Hydro system. The forecast includes only domestic load and firm sales to other utilities. The forecast does not take into account the possibility of additional sales to other utilities in the event of an excess water year generating surplus supply, nor does it reflect transactions by Powerex. 1.2. Role of Forecasting at BC Hydro Load forecasting is central to BC Hydro’s long-term planning, medium-term investment and operational and reporting activities. As such, BC Hydro’s load forecasting activities centre on the production of a number of term-specific and location-specific forecasts of energy sales and peak demand requirements to meet user needs for decision support information. A variety of related products including quarterly forecast updates, monthly variance reports, inputs for the revenue forecast, load shape analysis and small area forecasts are produced to supplement the base forecasts. Additionally, analytical, statistical and modelling support for a number of special or one time projects, including the Vancouver Island Call for Tender, the Network Integrated Transmission Service (NITS) Application, Cost of New Energy Supply, Rate Design and the Integrated Electricity Plan, are also provided. Forecast requirements for electric utilities are changing in response to a number of changes in the industry. These changes include: • • Increased risks as the system operates closer to capacity, increasing need for more frequent, shorter-term forecasting; Future uncertainty, resulting in more need for stress testing, and a focus on risk/sensitivity analysis; and PAGE 1 • Increased interest on the part of the regulator and stakeholders, reinforcing a need to ensure methods are transparent, consistent and defensible in regulatory context. The main users and uses of forecast products include the following: • • • • • • Generation: operations system planning; Rates: rate design and rate structure; Distribution: revenue forecasting, energy planning, distribution planning and investment; Transmission: transmission planning and investment; and Corporate: financial forecasts, service plan and budget reports and regulatory filings; BC Utilities Commission: as part of the regulatory reviews. The key focus for the current year is to ensure that the forecast function is evolving appropriately in response to these trends and to implement the Forecast Renewal Project. The Forecast Renewal Project involves building new forecasting models that are more accurate, transparent and better supported than the existing end-use models. PAGE 2 2 Regulatory Background In November 2002, the Government of British Columbia released its new energy plan, Energy for our Future: A Plan for B.C. Following the release, the Utilities Commission Act was amended, in part, to provide a mandate consistent with the new energy plan. In particular, the amendments clarified the regulatory role of the British Columbia Utilities Commission (BCUC) with respect to the planning requirements of utilities. In July 2003, the BCUC issued draft resource planning guidelines. Section III (2) of the draft guidelines considers the development of gross demand forecasts (before considering the effect of demand-side management programs) and states: “In making a demand forecast, it is necessary to distinguish between demographic, social, economic and technological factors unaffected by utility actions, and those actions that the utility can take to influence demand, (e.g. rates, DSM programs). The latter actions should not be reflected in the utility’s gross demand forecasts. More than one forecast would generally be required in order to reflect uncertainty about the future: probabilities or qualitative statements may be used to indicate that one forecast is considered to be more likely than others…” In its decision of October 29, 2004, on the Regulatory Requirements Application, the BCUC found that the “Electric Load Forecast 2003/04 to 2023/24 documenting BC Hydro’s forecasting approach was informative and responsive to the comments made in the September 2003 BCUC Decision on the VIGP CPCN application”. In addition in its decision, the BCUC provided a number of comments pertaining to the Load Forecast. Actions taken in response to these comments are summarized in Table 2.1. Table 2.1. B.C. Utilities Commission Comments and Actions Comments from 2004/05 and 2005/06 Revenue Requirements Decision 1. Streamline the load forecasting function. Action 1. New econometric based forecasting models have been built for the, commercial and industrial sectors 2. New short-term models by rate class have been developed 3. Rate impacts were included in the December 2004 Reference Load Forecast. In the current Reference Load forecast, rates are assumed to be constant in real terms for the forecast period pending further information 2. Prepare more rigorous short-term forecasting 3. Include impacts of the rate increase in the Reference Load Forecast The improvements in the Load forecast, subsequent to the BCUC comments in the F2005 and F2006 RRA decision, are detailed in section 4. PAGE 3 3 Forecast Drivers, Data Sources and Assumptions 3.1. Forecast Drivers Table 3.1 provides an overview of the key drivers for the reference forecast. Each forecast component is described as activity variables, use rate variables and data requirements for these variables. With respect to the commercial and industrial forecast, use rates are derived from the regression models. Within the forecast, the activity variable drives the scale of forecast electricity use. The use rate is a variable that represents the intensity of electricity use. Table 3.1. Key Forecast Drivers Forecast Component 1. Residential Forecast Activity • Number of residential accounts by housing type, heating type and region • GDP Use Rate • Consumption per account based on Residential EndUse Energy Planning System (REEPS) • Consumption per dollar of GDP (Based on regression model) • Consumption per dollar of GDP (Based on regression model) • Consumption per dollar of GDP (Based on regression model) • Consumption per account (based on REEPS) Data • Current number of accounts as base • Housing starts • Appliance saturation rates from Residential End Use Survey (REUS) • Billing data • GDP 2. Commercial Forecast 3. Industrial Distribution Forecast 4. Industrial Transmission Forecast 5. Non-Integrated Area Forecast • GDP • Billing data • GDP • GDP • Billing data • GDP • Number of accounts 6. Peak Forecast • Number of accounts by housing and heating type • Sales to general sector • Industrial activity and trend • Residential – kW/Account • General – kW/kWh • Transmission – peak demand (kVA) from hourly data • Current number of accounts as base • Local conditions in the short term • Population forecast for longer term • Appliance saturation rates from REUS • Employment forecast and housing starts • Weather data and load research data for use rate • Transmission customer hourly data • Economic and demographic forecasts PAGE 4 3.2. Growth Assumptions Growth assumptions for key drivers used in the reference forecast are displayed in Table 3.2 below. Table 3.2. Growth Assumptions (Annual rate of growth) Fiscal Year Actual 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Residential Accounts (%) 1.5 1.7 1.7 1.8 1.8 1.7 1.7 1.6 1.5 1.5 1.4 1.4 1.4 1.3 1.3 1.3 1.3 1.3 1.2 1.2 1.2 1.2 Calendar Year Employment (%) Real GDP (%) 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019 2020 2021 2022 2023 2024 2025 2.3 2.6 2.0 1.9 1.6 1.6 1.5 1.1 1.2 1.3 1.2 1.0 1.0 1.0 0.8 0.8 0.8 0.5 0.5 0.3 0.3 0.2 3.9 3.4 3.2 3.1 3.1 3.1 2.7 2.5 2.2 2.2 2.2 2.2 2.2 1.9 2.0 2.1 2.0 2.0 2.0 1.8 1.8 1.9 Information on the sources and the uses of the growth assumptions is shown in Table 3.3. PAGE 5 3.3. Variable Population GDP (Gov) 5 Data Sources Application • Used as a check on account growth • Commercial and industrial energy forecast • Commercial and industrial energy forecast • Residential accounts forecast • General rate class sales growth for peak Forecast Period • 2005-2024 Source • B.C. Statistics, B.C. Population Forecast, December 2004 • B.C. Ministry of Finance, September Budget Update – 2005/06 to 2007/08, September 2005 • Growth rates based on Conference Board of Canada, Provincial Data, June 2005 • Conference Board of Canada, Provincial Data, June 2005 • R.A. Malatest, July 2004 • Conference Board of Canada, Provincial Data, June 2005 Table 3.3. Data Sources and Uses for Growth Assumptions • 2005-2009 GDP (third party) Housing Starts Employment • 2010-2025 • 2005-2025 • 2005-2025 The GDP forecast is available from the BC Provincial government for only the first five years of the forecast period thus requiring use of an additional source. 5 PAGE 6 4 Forecast Process and Methodologies There are a number of key components to the load and sales forecast: the residential forecast; the commercial forecast, the industrial forecast (distribution voltage and transmission voltage), and the peak forecast (distribution voltage and transmission voltage). This section covers the methodology used for each of these forecast components. 4.1. Residential Forecast Methodology The residential energy forecast is determined by forecasting the number of accounts times rate of use based on the following expression: (4.1) where: • • • • RES is residential consumption; R is the number of residential accounts; RUR is the residential use rate; i indexes 20 appliances (space heating, space cooling, water heater, refrigerator, freezer, clothes washer, clothes dryer, dishwasher, range, lighting and so on); j indexes four housing types (single/duplex, row, apartment and other); and k indexes four regions (Lower Mainland, Northern Region, South Interior and Vancouver Island). RES = Σk Σj Σi Rijk*RURijk, • • The residential energy forecast is determined by forecasting the number of accounts multiplied by the rate of use. The forecast in the growth of the number of residential accounts is based on a forecast of housing starts provided by a third party. The number of residential accounts is then the current number of residential accounts plus the forecast of additional accounts to be added each year. Use rates are forecast from appliance saturation rates and unit energy consumption per end use (as well as their trends) to determine the average use rate by dwelling type, by region and changes in these rates over time. Appliance saturation rates and unit energy consumption come from the Residential End-Use Energy Planning System model (REEPS) as updated using the Residential End Use Survey (REUS) and the Conservation Potential Review (CPR). 4.2. Commercial Forecast Methodology The commercial energy forecast is based on the following expression: (4.2) COM = α + β*GDP PAGE 7 where: • • COM is commercial energy sales. α and β are the regression coefficients from a time series regression of commercial electricity sales on provincial GDP. Commercial forecast results are presented in Appendix 4. 4.3. Industrial Forecast Methodology The industrial distribution energy forecast is based on the following expression: (4.3) where: • • INDC is industrial distribution sales α and β are the regression coefficients from a time series regression of industrial distribution sales on provincial GDP. INDC = α + β*GDP The industrial transmission voltage energy forecast is based on the following expression: (4.4) where: • • INDD is industrial energy sales for transmission voltage customers α and β are the regression coefficients from a time series regression of industrial transmission sales on provincial GDP. INDD = α + β*GDP A modified version is also estimated that incorporates the impacts of industrial strikes: (4.5) where: • Binary variable is defined in the glossary Industrial forecast results are presented in Appendix 5. 4.4. Peak Forecast Methodology The peak demand forecast is built up in four stages, each with several steps. First, the substation peak in MVA non-coincident 6; second, the area peak in MVA non-coincident; third, the region peak in MW on a region coincident basis; and, fourth, the system peak in MW on a system coincident basis. The substation peak forecast is first built up in several steps: (a) actual and normalized peak loads by substation/area; (b) area substation peak forecast 6 INDDt = α + β*GDPt + χ*binary variable for strike years. Non-coincident is defined in the glossary. PAGE 8 guidelines are developed from an econometric model; (c) an 11-year substation forecast for each substation; and (d) the substation and guideline peak forecast are averaged together. The first step is analysis of last year’s substation peak using the following: (4.6) where: • • • KVA is the read peak load; and min is the minimum mean temperature for the coldest day during the reading period. α and β are the regression coefficients of transmission from a time series regression of peak substations readings on temperatures. KVA = α + β*min Using the estimated regression coefficients, the weather-normalized peak is then calculated based on the design day temperature for that substation 7: (4.7) where: • • NKVA is weather-normalized peak; and designmin is the design mean temperature for the substation. NKVA = α + β*designmin Historical growth rates of normalized substations peak, load transfers between substations and large substation load additions are used to develop short-term substation peak forecast. The second step is the 11-year substation guideline (econometric model): (4.8) SKit = [α1SFDHTG + α2SFDNON + α3MULTHTG + α4MULNON + α5U35E + α6O35E] where: • • • • • • SKit is the total substation peak for the ith planning area; SFDHTG is the number of single-family electrically heated homes; SFDNON is the number of single-family non-electrically heated homes; MULTHTG is the number of multi-family electrically heated homes; MULTNON is the number of multi-family non-electrically heated homes; U35E is annual energy consumption General under 35 kW; Alternatives to a linear regression approach to weather adjusting individual substation peak data are under review. The alternatives include expanding the regression model to a nonlinear approach to estimate the relationship between peak and temperature. 7 PAGE 9 • • O35E is annual energy consumption General over 35kW; the coefficients α1, α2, α3, and α4 are kW contribution to the distribution peak per dwelling in area i, for the four dwelling types under normal temperature conditions; and the coefficients, α5 and α6 represent the increase in peak demand due to a one-kWh increase in the General rate class Under 35 and Over 35 kW energy consumption. • As a third step, a longer-term (11-year) substation forecast is prepared for each substation based on the guidelines, trends in substation growth, load transfers between substations and large substation load additions. The fourth step is calculation of the blend/average of the long-term substation peak forecast, the peak guideline forecast, and the bottom up and top down weather normalization values. The blending of the historical and forecast information reduces the substation peak forecast from 15 planning areas to 4 regional distribution peaks. The regional peak forecast, for any of the four regions, is computed using: (4.9) where: • • • • • • • PK is the distribution peak; DCF is the distribution peak coincidence factor; PF is the power factor ; TP is the transmission peak; TCF is the transmission coincident factor; OP is the other utility peak sales; and OCF is the other utility coincident factor. RPKjt = Σj [PKit*DCFj*PFj + TPj*TCFj*PFj + OPj*OCFj] Finally, system peak is the sum of coincidence-adjusted regional peaks and includes transmission losses: (4.10) where: • • TL is the transmission loss factor; and SCF are the system coincidence factors for each of the regions’ distribution and transmission peaks. SPK = (1 + TL)*Σj RPKjt *SCFj PAGE 10 5 Reference Forecast 5.1. Reference Forecast Before DSM Table 5.1 provides a summary of sales by customer class, gross requirements, peak demand and the reference forecast before DSM, that is, before considering the effects of BC Hydro’s demand-side management programs (See Sections 8, 9 and 10 for the individual sector forecasts). BC Hydro’s total domestic sales before DSM include residential, commercial and industrial sales for the BC Hydro service area as well as sales to New Westminster and Fortis BC. BC Hydro’s total domestic sales before DSM are expected to grow from 50,787 GWh in 2004/05 to 71,894 GWh in 2025/26. BC Hydro’s total gross requirements are for both the integrated area and the NIA, and include total domestic sales, other firm utility sales and losses. BC Hydro’s total gross requirements before DSM are expected to grow from 55,747 GWh in 2004/05 to 79,375 GWh in 2025/26. BC Hydro’s total integrated system peak (system coincident basis excluding Powerex and related losses) before DSM is expected to grow from 9,762 MW (10,110 MW weather adjusted) in 2004/05 to 13,756 MW in 2025/26. Growth rates of sales vary significantly by sector but within a given sector are fairly consistent over time. For the residential sector, the growth rates of sales before DSM are 2.8 per cent for the five years from 2004/05 to 2009/10; 2.2 per cent for the 11 years from 2004/05 to 2015/16; and 1.9 per cent in the 21 years of the forecast 2004/05 to 2025/26. For the commercial sector, the growth rates of sales before DSM are 2.8 per cent for the five years from 2004/05 to 2009/10; 2.4 per cent for the 11 years from 2004/05 to 2015/16; and 2.1 per cent in the 21 years of the forecast, 2004/05 to 2025/26. For the industrial sector, the growth rates of sales before DSM are 1.6 per cent for the five years from 2004/05 to 2009/10; 1.1 per cent for the 11 years from 2004/05 to 2015/16; and 1.2 per cent in the 21 years of the forecast, 2004/05 to 2025/26. For total gross requirements, the growth rates of sales before DSM are 2.4 per cent for the five years from 2004/05 to 2009/10; 1.9 per cent for the 11 years from 2004/05 to 2015/16; and 1.7 per cent in the 21 years of the forecast, 2004/05 to 2025/26. For the peak, the growth rates before DSM on a weather-adjusted basis are 2.0% for the five years from 2004/05 to 2009/10; 1.5 per cent for the 11 years from 2004/05 to 2015/16; and 1.5 per cent in the 21 years of the forecast 2004/05 to 2025/26. Table A7.9 in Appendix 7 provides additional details of the current forecast before DSM. PAGE 11 Table 5.1. Reference Forecast Before DSM Residential (GWh) Commercial (GWh) Industrial (GWh) Total Domestic Sales (GWh) Total Gross Requirements (GWh) Total Integrated System Peak (MW) Actual (not weather-normalized) 1999/00 14,572 13,176 17,890 46,376 51,534 2000/01 14,573 13,654 18,579 47,891 52,978 2001/02 15,090 13,583 17,739 47,473 52,567 2002/03 15,287 13,729 18,596 48,685 53,298 2003/04 15,899 14,151 18,725 49,960 55,051 2004/05 15,620 (15,976) 14,362 19,635 50,787 (51,143) 55,747 (56,142) Forecast (Residential Energy and System Peak forecasts assume “normal weather”) 2005/06 16,245 14,859 20,237 52,426 58,109 2006/07 16,854 15,310 20,513 53,875 59,515 2007/08 17,202 15,717 20,781 54,926 60,670 2008/09 17,559 16,092 21,036 55,923 61,766 2009/10 17,916 16,493 21,307 56,970 62,917 2010/11 18,265 16,862 21,557 57,956 64,006 2011/12 18,625 17,215 21,863 58,994 65,147 2012/13 18,959 17,542 22,149 59,961 66,209 2013/14 19,289 17,873 22,437 60,930 67,273 2014/15 19,613 18,211 21,829 61,004 67,382 2015/16 19,935 18,550 22,120 61,978 68,456 2016/17 20,260 18,906 22,425 62,985 69,560 2017/18 20,586 19,224 22,703 63,925 70,596 2018/19 20,911 19,460 22,997 64,798 71,559 2019/20 21,235 19,926 23,312 65,921 72,794 2020/21 21,557 20,278 23,617 66,919 73,890 2021/22 21,881 20,633 23,929 67,928 75,002 2022/23 22,206 20,980 24,234 68,924 76,099 2023/24 22,533 21,315 24,534 69,905 77,184 2024/25 22,858 21,656 24,834 70,888 78,266 2025/26 23,182 22,008 25,148 71,894 79,375 Growth Rates 5 years: 99/00 to 04/05 1.4% 1.7% 1.9% 1.8% 1.6% 5 years: 04/05 to 09/10 2.8% (2.3%) 2.8% 1.6% 2.3% (2.2%) 2.4% (2.3%) 11 years: 04/05 to 15/16 2.2% (2.0%) 2.4% 1.1% 1.8% (1.8%) 1.9% (1.8%) 21 years: 04/05 to 25/26 1.9% (1.8%) 2.1% 1.2% 1.7% (1.6%) 1.7% (1.7%) 8,646 9,320 9,003 8,824 9,911 9,762 (10,110) 8 10,298 10,511 10,772 10,978 11,151 11,274 11,425 11,565 11,720 11,756 11,875 12,050 12,229 12,409 12,592 12,779 12,968 13,161 13,356 13,554 13,756 2.5% 2.7% (2.0%) 1.8% (1.5%) 1.6% (1.5%) 8 Values shown in brackets are based on weather normalized actual values. PAGE 12 5.2. Reference Forecast With DSM Table 5.2 provides a summary of historical sales, requirements, peaks and the reference forecast with DSM, that is, including the effects of BC Hydro’s demand-side management program. BC Hydro’s total domestic sales for the integrated area and the NIA, with DSM include residential, commercial and industrial sales for the BC Hydro service area as well as sales to New Westminster and Fortis BC. BC Hydro’s total domestic sales with DSM are expected to grow from 50,787 GWh in 2004/05 to 69,749 GWh in 2025/26. BC Hydro’s total gross requirements include total domestic sales, firm exports and losses. BC Hydro’s total gross requirements with DSM are expected to grow from 55,747 GWh in 2004/05 to 77,024 GWh in 2025/26. BC Hydro’s total integrated system peak (system coincident basis excluding Powerex and related losses) with DSM is expected to grow from 9,762 MW (10,110 MW weather adjusted) in 2004/05 to 13,413 MW in 2025/26. The DSM electricity savings included in the load forecast do not include any additional savings beyond those reflected in BC Hydro’s current DSM plans. Again, growth rates of sales vary significantly by sector but within a given sector are fairly consistent over time. For the residential sector, the growth rates of sales with DSM are 2.4 per cent for the five years from 2004/05 to 2009/10; 2.0 per cent for the 11 years from 2004/05 to 2014/15; and 1.8 per cent for the 21 years from 2004/05 to 2025/26. For the commercial sector, the growth rates of sales with DSM are 2.4 per cent for the five years from 2004/05 to 2009/10; 2.2 per cent for the 11 years from 2004/05 to 2015/16; and 2.0 per cent for the 21 years from 2004/05 to 2025/26. For the industrial sector, the growth rates of sales with DSM are 0.3 per cent for the five years from 2004/05 to 2009/10; 0.4 per cent for the 11 years from 2004/05 to 2015/16; and 1.0 per cent for the 21 years from 2004/05 to 2025/26. For total gross requirements, the growth rates of sales with DSM are 1.7 per cent for the five years from 2004/05 to 2009/10; 1.5 per cent for the 11 years from 2004/05 to 2015/16; and 1.6 per cent for the 21 years from 2004/05 to 2025/26. For peak, the growth rates of sales with DSM on a weather-adjusted basis, are 1.4 per cent for the five years from 2004/05 to 2009/10; 1.2 per cent for the 11 years from 2004/05 to 2015/16; and 1.4 per cent for the 21 years from 2004/05 to 2025/26. The following points are worth noting with respect to the peak forecast: • The distribution peak forecast reflects the change in the design temperature to a rolling 30 year historical period used to determine the average coldest day. The transmission peak forecast has been updated to reflect recent trends among B.C.’s large industrial customers. The forecast also includes the closing of Highland Valley Copper beginning in 2013/14 in the South Interior. • PAGE 13 Table 5.2. Reference Forecast With DSM Residential (GWh) Commercial (GWh) Industrial (GWh) Total Domestic Sales (GWh) Total Gross Requirements (GWh) Total Integrated System Peak (MW) 8,646 9,320 9,003 8,824 9,911 9,762 (10,110) 9 10,205 10,340 10,526 10,680 10,821 10,912 11,037 11,166 11,321 11,363 11,487 11,655 11,827 12,006 12,188 12,395 12,625 12,817 13,013 13,211 13,413 Actual (not weather-normalized) 1999/00 14,572 13,176 17,890 46,376 2000/01 14,573 13,654 18,579 47,891 2001/02 15,090 13,583 17,739 47,473 2002/03 15,287 13,729 18,596 48,685 2003/04 15,899 14,151 18,725 49,960 2004/05 15,620 (15,976) 14,362 19,635 50,787 (51,143) Forecast (Residential Energy and System Peak forecasts assume “normal weather”) 2005/06 16,116 14,786 19,911 51,898 2006/07 16,651 15,197 19,816 52,862 2007/08 16,976 15,497 19,713 53,412 2008/09 17,295 15,796 19,752 54,079 2009/10 17,613 16,143 19,903 54,913 2010/11 17,925 16,482 20,021 55,700 2011/12 18,251 16,827 20,179 56,548 2012/13 18,565 17,167 20,383 57,426 2013/14 18,882 17,498 20,670 58,381 2014/15 19,184 17,835 20,132 58,502 2015/16 19,469 18,171 20,507 59,520 2016/17 19,746 18,527 20,814 60,481 2017/18 20,031 18,846 21,097 61,386 2018/19 20,332 19,081 21,392 62,235 2019/20 20,635 19,547 21,706 63,336 2020/21 20,928 19,900 22,187 64,482 2021/22 21,235 20,255 22,809 65,784 2022/23 21,560 20,602 23,114 66,780 2023/24 21,887 20,936 23,414 67,760 2024/25 22,212 21,278 23,714 68,744 2025/26 22,537 21,629 24,028 69,749 Growth Rates 5 years: 99/00 to 04/05 1.4% 1.7% 1.9% 1.8% 5 years: 04/05 to 09/10 2.4% (2.0%) 2.4% 0.3% 1.6% (1.4%) 11 years: 04/05 to 15/16 2.0% (1.8%) 2.2% 0.4% 1.5% (1.4%) 21 years: 04/05 to 25/26 1.8% (1.7%) 2.0% 1.0% 1.5% (1.5%) 51,534 52,978 52,567 53,298 55,051 55,747 (56,142) 57,537 58,416 59,022 59,757 60,676 61,547 62,481 63,447 64,496 64,655 65,775 66,828 67,824 68,761 69,971 71,227 72,652 73,750 74,833 75,917 77,024 1.6% 1.7% (1.6%) 1.5% (1.5%) 1.6% (1.5%) 2.5% 2.1% (1.4%) 1.5% (1.2%) 1.5% (1.4%) Comparing the growth rates in Tables 5.1 and 5.2 for energy and for peak both before and with DSM, it can be noted: 9 Values shown in brackets are based on weather normalized actuals PAGE 14 • That the energy forecast before DSM tracks the key economic driver real GDP fairly closely. This largely reflects the importance of real GDP as a forecast driver and its historical relationship to energy consumption. Applying the DSM estimates reduces energy growth substantially for the first 11 years, but has less impact over latter part of the forecast period. This reflects the absence of new planned/approved DSM activity beyond the projected savings under the current DSM plan; and Peak demand, after weather adjustments, is growing more slowly than energy. This can be explained by: 1. First, weather adjustment of sales and peak demand is done differently. Sales are weather adjusted to account for the different in actual degree days compared to normal degree days, while peak is done by adjusting the actual peak for the difference in actual daily average temperature and design temperature. 2. Recovery and gains in the mining sector have resulted in a higher increase in energy demand relative to the increase in peak demand. 3. Transmission customers may respond in the near to term to favourable economic and prices conditions by increase their energy consumption by adding additional shifts or using idle capacity. These changes may not directly lead to an increase in their non-coincident peak demands. • • PAGE 15 6 Comparison Between 2004 and 2005 Forecasts 6.1. Total Gross Requirements Table 6.1 compares total gross requirements for the current 2005 reference forecast including the effects of DSM with the December 2004 forecast including DSM. For all forecast years, the current forecast is higher than the 2004 forecast. For 2005/06, the 2005 forecast is above the 2004 forecast by 1,110 GWh. Much of the increase in 2005/06 is a result of a higher industrial sales forecast compared to previous forecast. As well the new forecast for 2005/06 reflects sales and variance information from the current year. For 2015/16, the 2005 forecast is 2,297 GWh above the 2004 forecast, and is 2,312 GWh above the 2004 forecast in 2024/25. Table 6.1. Comparison of Reference Gross System Requirements With DSM December 2005 Forecast (GWh) 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 NB. * = actuals 51,534* 52,978* 52,567* 53,298* 55,051* 55,747* 57,537 58,416 59,022 59,757 60,676 61,547 62,481 63,447 64,496 64,655 65,775 66,828 67,824 68,761 69,971 71,227 72,652 73,750 74,833 75,917 77,024 December 2004 Forecast (GWh) 10 51,354* 52,978* 52,567* 53,298* 55,051* 55,935 56,427 56,847 57,183 57,844 58,188 59,184 59,492 60,403 61,341 62,416 63,478 64,539 65,571 66,653 67,742 68,992 70,224 71,337 72,455 73,605 December 2005 minus December 2004 (GWh) -187 1,110 1,569 1,839 1,913 2,488 2,363 2,989 3,044 3,155 2,239 2,297 2,289 2,253 2,108 2,229 2,235 2,428 2,413 2,378 2,312 Change (%) -0.3 2.0 2.8 3.2 3.3 4.3 4.0 5.0 5.0 5.1 3.6 3.6 3.5 3.4 3.2 3.3 3.2 3.5 3.4 3.3 3.1 - 10 The December 2004 gross requirements forecast in Table 6.1 excludes the effects of the Weyerhaeuser load displacement project. All comparisons to the December 2004 forecast, unless otherwise stated, are exclusive of the impact of the project. With the impact of Weyerhaeuser, the adjustment to the December 2004 forecast is 157 GWh annually and 21 MW annually for peak. PAGE 16 6.2. Integrated Gross Requirements Compared to December 2004 forecast, the current peak forecast is increased to reflect a revised transmission peak forecast, forecast drivers and DSM estimates. On a system total basis, the 2005 forecast including DSM is 1,111 GWh higher for 2005/06; 2,296 GWh higher for 2015/16; and 2,137 GWh higher for 2024/25. Table 6.2. Comparison of Integrated Gross System Requirements with DSM December 2005 Forecast (GWh) 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 NB. * = actuals December 2004 Forecast (GWh) 11 51,279* 52,718* 52,292* 53,010* 54,756* 55,636 56,127 56,544 56,877 57,532 57,870 58,860 59,162 60,068 61,001 62,071 63,127 64,182 65,208 66,283 67,366 68,610 69,836 70,944 72,056 73,201 - 51,279* 52,718* 52,292* 53,010* 54,756* 55,453* 57,238 58,111 58,713 59,445 60,358 61,223 62,151 63,111 64,155 64,308 65,423 66,470 67,461 68,393 69,597 70,848 72,268 73,361 74,439 75,518 76,621 December 2005 minus December 2004 (GWh) -183 1,111 1,567 1,836 1,913 2,488 2,363 2,989 3,043 3,154 2,237 2,296 2,288 2,253 2,110 2,231 2,238 2,432 2,417 2,383 2,317 - Change (%) -0.3 2.0 2.8 3.2 3.3 4.3 4.0 5.1 5.1 5.2 3.6 3.6 3.6 3.5 3.2 3.3 3.3 3.5 3.4 3.3 3.2 - 11 The December 2004 gross requirements forecast in Table 6.1 excludes the effects of the Weyerhaeuser load displacement project. All comparisons to the December 2004 forecast, unless otherwise stated, are exclusive of the impact of the project. With the impact of Weyerhaeuser, the adjustment to the December 2004 forecast is 157 GWh annually and 21 MW annually for peak. PAGE 17 6.3. Total Integrated Peak Sales Compared to December 2004 forecast, the current peak forecast is increased to reflect a revised transmission peak forecast, forecast drivers and DSM estimates. On a system total basis, the 2005 forecast including DSM is 136 MW higher for 2005/06; 169 MW higher for 2015/16; and 292 MW higher for 2024/25. Table 6.3. Comparison of Reference Peak Forecasts With DSM (Integrated System) December 2005 Forecast (MW) 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 NB. * = actuals ** = On a weather-adjusted basis, the peak is 10,110 MW and the difference is 77 MW or 0.8%. 12 December 2004 Forecast (MW) 12 8,646* 9,320* 9,003* 8,824* 9,911* 10,033 10,069 10,161 10,232 10,372 10,473 10,633 10,713 10,862 11,009 11,153 11,318 11,481 11,651 11,821 11,995 12,175 12,361 12,543 12,731 12,919 - 8,646* 9,320* 9,003* 8,824* 9,911* 9,762** 10,205 10,340 10,526 10,680 10,821 10,912 10,037 11,166 11,321 11,363 11,487 11,655 11,826 12,006 12,188 12,395 12,625 12,817 13,013 13,211 13,413 December 2005 minus December 2004 (MW) -271 136 179 294 308 348 279 324 304 312 210 169 174 175 185 193 220 264 274 282 292 - Change (%) -2.7 1.4 1.8 2.9 3.0 3.3 2.6 3.0 2.8 2.8 1.9 1.5 1.5 1.5 1.6 1.6 1.8 2.1 2.2 2.2 2.3 - The December 2004 gross requirements forecast in Table 6.1 excludes the effects of the Weyerhaeuser load displacement project. All comparisons to the December 2004 forecast, unless otherwise stated, are exclusive of the impact of the project. With the impact of Weyerhaeuser, the adjustment to the December 2004 forecast is 157 GWh annually and 21 MW annually for peak. PAGE 18 7 Sensitivity Analysis Each year the load forecasts are developed based on assumptions and forecast of independent variables or drivers that are incorporated into BC Hydro’s forecasting models for the residential, commercial and industrial sectors. There are several major types of forecast uncertainties that could cause forecast error. BC Hydro undertakes several strategies to attempt to mitigate these sources of forecast error. These uncertainties are: 1. Model Uncertainty – The forecasting model which captures the relationship between the drivers and load cannot exactly match the detailed behaviour of BC Hydro’s customers or the operation of its system. BC Hydro attempts to develop stable models that capture the relationship between drivers and sales correctly. 2. Data Uncertainty – In any forecast there is uncertainty in the model drivers or predictors used to develop the forecast. In order to mitigate this uncertainty, BC Hydro uses reliable and credible sources of drivers of the forecast. 3. Outcome Uncertainty – external events, weather and world events such as 9/11, Severe Acute Respiratory Syndrome (SARS), trade disputes and shocks in commodity markets. These factors lead to forecast variances from actual demand. BC Hydro continuously attempts to improve the accuracy of its forecasting process through a process of internal and external validation. This testing includes backcasting, monitoring industry trends on forecasting approaches, and tracking of industry developments that may affect the forecast. Figure 7.1 summarizes the major types of forecast uncertainties that could cause forecast error, over what timeframe those uncertainties could affect the accuracy of the forecast, and strategies by which BC Hydro attempts to mitigate these sources of forecast error. PAGE 19 Figure 7.1. Sources of Forecast Uncertainty Major Types Duration Mitigation Strategies •Data Uncertainty - reliable and credible sources of drivers of the forecast •Model Uncertainty - are the models stable, capture the relationship between drivers and sales correctly - are the reports on industry outlooks plausible and supported by other sources •Outcome Uncertainty - External Shocks, Weather, & World Events. Longer-term Impacts •multiple sources for key drivers •analysis of Regional Economic reports •sharing forecast driver information with other utilities •internal validation- goodness of fit, error specification, model predicts well within sample •external validation - other jurisdictions, trends •other checks - load factor, previous forecasts •forecasts prepared to normal weather and actuals adjusted for normal weather •acknowledgment that shocks can impact short-term trends Shorter-term Impacts To the extent of which BC Hydro’s forecasts have tracked actual load is shown in the next figure. Figure 7.2 shows a history of gross requirements as predicted by BC Hydro forecasts compared to actual gross requirements. Figure 7.2. Total Gross Requirements with DSM Total Gross Requirements with DSM 75,000 Forecast Vintage 70,000 1995 1996 1997 1998 65,000 1999 2000 2001 2002 GWh 60,000 2003 2004 2005 Actual 55,000 50,000 45,000 1990/91 1991/92 1992/93 1993/94 1994/95 1995/96 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 PAGE 20 7.1. Monte Carlo Analysis BC Hydro uses a Monte Carlo model to assess the sensitivity of BC Hydro’s load forecast to a number of variables including weather, economic conditions, and electricity prices. The analysis examines the sensitivity of the load to a combination of key drivers to produce a forecast uncertainty band. An analysis, involving Monte Carlo simulation methods, was completed for the reference forecast to reflect a range of uncertainties implicit in the load forecast that includes factors beyond GDP and price. Monte Carlo simulation method is a technique for estimating probabilities of an outcome as defined by a model. Random sampling techniques are used to generate a range of outcomes. Probabilities are estimated from an analysis of this range of outcomes Six major causal factors were used to analyze the sensitivity of the forecast. These include: economic growth rate (reflected by GDP), the electricity rate, the effective energy reduction achieved by demand-side management (DSM) programs, the response to electricity price changes (price elasticity), electricity intensity and heating degree-days. Probability distributions were assigned to each of these factors. The parameters of these distributions were established to reflect possible future levels of each of these factors. A Monte Carlo model, that includes simulation methods, was used to quantify and combine the probability distributions, reflecting the relationships between the six causal factors and electricity consumption. A probability distribution for the load forecast was thus obtained which showed the likelihood of various load levels resulting from the combined effect of the six factors. This distribution is banded by: • • The low scenario: There is a 10 per cent chance the outcome will be below this value. The high scenario: There is a 10 per cent chance that the outcome will exceed this value. Table 7.1 summarizes the results of Monte Carlo model and simulations analysis on the reference energy and peak forecast before DSM. Table 7.2 summarizes the results with DSM. 7.2. Uncertainty Assumptions For each of the six major causal factors, a probability distribution is defined as follows: (a) Long-Term Economic Growth (GDP) GDP growth rate is assigned a normal distribution with mean equal the base case annual percent change and with standard deviation equal to 1.70% (b) Electricity Rates The base case assumes that electricity rates will increase at the rate of inflation (i.e. no increase in real terms). The probability distribution for electricity price is then defined to be a triangular distribution with three parameters; a most likely value equal to the base case annual percentage growth rate, a minimum that is 2.5% lower than the base case and a maximum that is 2.5% higher than the base case. PAGE 21 ( c ) With respect to the forecasts with DSM, energy reduction due to DSM programs are assigned a triangular distribution with a most probable value equal to the base case, a minimum that is 25% lower and a maximum that is 25% higher. (d) Response to Electricity Price Changes (Elasticity) The elasticity of electricity demand with respect to BC Hydro rates is the ratio of the percentage change in consumption due to a price change to the percentage changes in the price itself. Estimates of elasticity are subject to considerable uncertainty. This uncertainty is modelled by assuming that price elasticity is a triangular distribution with its most probable value equal to the base case elasticity. Specific triangular distributions are assumed for residential, commercial and industrial, and for long and short run. Details are given in Appendix 6. (e) Electricity Intensity Electricity intensity is assumed to have a distribution whose mean is the base case electricity intensity. The distribution is generated by a growth rate that has a triangular distribution with a mean of zero, a 10th percentile of –0.2% and a 90th percentile of +0.2%. (f) Weather The impact of weather is modeled using annual heating degree days. Annual heating degree days are modeled by a Beta distribution which provided the best fit to 50 years of temperature data. See Appendix 6. Monte Carlo Methods for a more detailed description of this distribution. Uncertainty of Peak Demand The peak forecasts and uncertainty bands described in Table 7.1 and Table 7.2 are based on ratios of the low and the high energy uncertainty bands to the reference energy forecast. PAGE 22 Table 7.1 Monte Carlo Analysis – Energy and Peak Before DSM Lower Confidence Band (10%) Total Gross Requirements Total System Peak Reference Forecast Total Gross Requirements Total System Peak Upper Confidence Band (90%) Total Gross Requirements Total System Peak 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 (GWh) 55,747 57,285 58,540 59,550 60,481 61,467 62,405 63,387 64,269 65,179 65,116 65,999 66,941 67,843 68,592 69,661 70,613 71,506 72,457 73,360 74,241 75,191 (MW) 9,825 10,208 10,396 10,631 10,808 10,954 11,053 11,178 11,289 11,419 11,426 11,515 11,663 11,819 11,964 12,121 12,284 12,436 12,605 12,769 12,933 13,108 (GWh) 55,747 58,109 59,515 60,670 61,766 62,917 64,006 65,147 66,209 67,273 67,382 68,456 69,560 70,596 71,559 72,794 73,890 75,002 76,099 77,184 78,266 79,375 (MW) 9,825 10,355 10,569 10,831 11,038 11,212 11,336 11,489 11,630 11,786 11,823 11,944 12,120 12,299 12,481 12,666 12,854 13,044 13,238 13,435 13,634 13,837 (GWh) 55,747 58,901 60,468 61,763 63,032 64,359 65,621 66,967 68,228 69,491 69,793 71,036 72,410 73,643 74,851 76,296 77,662 79,050 80,419 81,733 83,103 84,469 (MW) 9,825 10,496 10,738 11,026 11,264 11,469 11,622 11,810 11,984 12,175 12,246 12,394 12,616 12,830 13,055 13,275 13,510 13,748 13,990 14,226 14,476 14,725 PAGE 23 Figure 7.2. Monte Carlo Analysis – Energy before DSM 90,000 85,000 80,000 75,000 GWh 70,000 65,000 60,000 55,000 50,000 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Upper Uncertainty Band (90%) Probable Forecast Lower Uncertainty Band (10%) Note: Probable forecast is the Reference forecast PAGE 24 Table 7.2 Monte Carlo Analysis – Energy and Peak With DSM Lower Confidence Band (10%) Total Gross Requirements Total System Peak Reference Forecast Total Gross Requirements Total System Peak Upper Confidence Band (90%) Total Gross Requirements Total System Peak 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 (GWh) 55,747 56,715 57,439 57,892 58,455 59,218 59,922 60,730 61,492 62,379 62,379 63,313 64,240 65,045 65,793 66,842 67,956 69,164 70,109 71,014 71,915 72,835 (MW) 9,825 10,115 10,224 10,382 10,506 10,621 10,685 10,790 10,885 11,013 11,028 11,123 11,271 11,410 11,557 11,713 11,897 12,091 12,258 12,424 12,590 12,760 (GWh) 55,747 57,537 58,416 59,022 59,757 60,676 61,547 62,481 63,447 64,496 64,655 65,775 66,828 67,824 68,761 69,971 71,227 72,652 73,750 74,833 75,917 77,024 (MW) 9,825 10,262 10,398 10,585 10,740 10,882 10,974 11,101 11,231 11,387 11,430 11,556 11,725 11,897 12,078 12,262 12,470 12,701 12,894 13,092 13,291 13,494 (GWh) 55,747 58,330 59,385 60,120 61,033 62,132 63,157 64,322 65,476 66,724 67,081 68,373 69,689 70,859 72,062 73,470 75,038 76,706 78,071 79,374 80,769 82,127 (MW) 9,825 10,403 10,570 10,782 10,969 11,143 11,261 11,428 11,590 11,780 11,859 12,012 12,227 12,429 12,658 12,875 13,137 13,410 13,650 13,886 14,140 14,388 PAGE 25 Figure 7.3. Monte Carlo Analysis – Energy with DSM 85,000 80,000 75,000 70,000 GWh 65,000 Upper Uncertainty Band (90%) Probable Forecast 55,000 Lower Uncertainty Band (10%) 2004/05 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 60,000 50,000 Note: Probable forecast is the Reference forecast PAGE 26 8. Residential Forecast 8.1. Summary Characteristics - Of the three customer classes, residential, commercial, and industrial, the residential sector is the most stable. Growth in the number of residential accounts is currently about 1.1% per year. After many years of growth in use per account, growth in use per account is forecast to be moderate growing at less than 1% over the entire forecast period. Of the 1.48 million residential accounts served at the end of fiscal year 2004/05, 59% were single/duplex, 8% were row houses, 25% were apartments, and 8% were mobile and miscellaneous. Geographically, 58% of the residential accounts are in the Lower Mainland, 10% are in the North Region, 11% are in the South Interior, and 21% are on Vancouver Island. Vancouver Island has the highest percentage of electrically heated accounts because of the relatively recent availability of natural gas. On a sales basis, 53% of residential sales were made in the Lower Mainland, 9% in the North Region, 12% in the South Interior, and 26% in Vancouver Island. Drivers – The drivers of the residential forecast are number of accounts and average annual use per account. The number of accounts is driven by housing starts. Since household size is gradually decreasing, account growth is expected to be somewhat higher than population growth. Account growth can vary considerably from year to year in response to BC’s economy. In the mid 1990's, about 38,000 accounts were added annually, but by the early 2000's, that number fell to about 13,000. Now that the economy has pulled out of the slump of the early 2000’s, account growth is forecast to be about 26,000 per year. This almost doubles the number of new accounts of a few years ago, but only about two thirds of the account rate growth 10 years ago. With growth expected in both the number of accounts and use per account, growth in sales with DSM is forecast to be 1.8% over the entire forecast period. Trends – Sales to the residential sector fell by about 1.8% in 2004/05 mainly due to warmer weather in that year. Based on the current forecast for housing starts, the new forecast for number of accounts at the end of 2009/10 is 1,619,453 which is 15,513 or 0.5% below the previous forecast. This leads to a slight reduction in the sales forecast for the short term compared to last year’s forecast. However, the lower forecast for number of accounts is offset by a higher use per account forecast, which is expected to grow by 0.29% per year, up from 0.23% last year. Electricity Use – BC Hydro’s residential sector currently consumes about 31% of BC Hydro’s total annual billed firm sales. This electricity is used to provide a range of services (often called end-uses). The largest end-uses are space heating, water heating, refrigeration, and miscellaneous plug-in load which includes computer equipment and home entertainment systems such as plasma TVs. Because space and water heating loads are dependent on the outside temperature, monthly residential sales can be strongly affected by the weather, but sales variations due to the weather tend to cancel out over the long term. Forecast - Billed residential sales were 15,620 GWh in 2004/05 but on a weather normalized basis were higher at 15,976 GWh. Residential sales, before DSM impacts, are forecast to grow from 15,620 GWh in 2004/05 to 17,916 GWh in 2009/10, 19,935 GWh in 2015/16, and to 23,182 GWh in 2025/26. These increases represent growth rates of 2.8 per cent over the next five years (2004/05 to 2009/10), 2.2 per cent over the next 11 years (2004/05 to PAGE 27 2015/16), and 1.9 per cent over the next 21 years of the forecast (2004/05 to 2025/26). 8.2. Forecast Overview Table 8.1 shows the forecast for residential sales before DSM, including sales by region. Table 8.2 shows the forecast for residential sales after DSM, including sales by region. The use rate forecast is based on projections of factors such as housing mix (single family, row house, apartment, etc.), heating fuel choices (electric versus non-electric), appliance penetration rates, appliance life span and changes in electricity demands. Currently, 19 per cent of BC Hydro’s residential accounts are heated electrically, and on average they require about 14,600 kWh per year. The average usage may not change much for reasons stated below. Ten years ago the average residential weather-normalized use rate was 10,590 kWh per year, and was increasing by about 85 kWh per year. However, growth in use rate has declined since then, with use rate currently at 10,850 kWh per year, and growing by only an average of 25 kWh per year for the last ten years. Improvements in building insulation and appliance efficiency are the main reasons for the moderation of growth in the annual residential use rate. Over the longer term, the use rate is not expected to change significantly because of the offsetting effects of the following residential trends. First, increased electric space heating market share is expected to be offset by smaller housing units. Due to limited availability of land for residential development, the trend in major metropolitan centres is expected to be towards denser housing. Since row houses and apartments are more likely to be built with electric heat than single family homes, the market share for electrically heated housing is expected to increase. Although new row houses and apartments tend to be larger than existing similar dwellings, they generally have a smaller floor area than detached single family homes, and therefore have lower space heating load requirements. The increase in market share of electric space heating is also offset to some extent by improvements in building standards, and by the construction of the Vancouver Island gas pipeline, which has made gas space heat available to Island residents. However, natural gas prices are projected to be higher for Vancouver Island compared to the Mainland over the entire forecast period. As a result, the growth in the penetration rate of gas heating is anticipated to be slower for Vancouver Island than it was for the Mainland. Second, there is the issue of more efficient appliances versus higher penetration. Manufacturers throughout Canada and the United States are expected to continue to improve the energy efficiency of major electrical appliances. As older models wear out and are replaced by newer ones, electricity consumption for major appliances such as refrigerators, freezers, ovens and ranges is forecast to decrease. However, new models of these major appliances tend to be larger than models currently in use. Some of the reduction in electricity use resulting from improvements in electricity efficiency will be offset by an increase in appliance size. Third, a projected decrease in the number of people per household would tend to reduce electricity use per account. However, this reduction is expected to be PAGE 28 offset by an increase in the penetration level of small appliances. An increase in electricity use is also projected from lifestyle changes and technological improvements. The latter are expected to cause an increase in demand for electronic, entertainment and telecommunication devices in the home. A trend towards home offices is also expected to produce a long-term increase in residential electricity consumption. In the long term, the expected overall impact of these various trends is that the factors working to increase the use rate will be offset by the factors working to decrease it, leading to the use rate levelling. 8.3. Forecast Methodology The forecast for residential sales is calculated as forecast number of accounts times forecast use per account. These two main drivers are discussed in detail below. For the first year of the forecast, growth in number of accounts is forecast based on recent trends. For all subsequent years, percentage growth in number of accounts is set equal to percentage growth in forecast housing starts, which are provided by a third party. To develop the residential sales forecast for the entire BC Hydro service area, the total service area is divided into four customer service regions, and a forecast was prepared for each region. These regions are Lower Mainland, Northern Region, South Interior and Vancouver Island. For each region, a third party housing stock forecast was prepared based on the housing starts forecast in the region, and on other regional factors such as trends in housing mix and gas availability. A use rate forecast is also developed for each region based on projections of penetration rates and individual consumption levels by end use (space heating, water heating, major appliances and small lifestyle appliances). The residential sales forecast for a region is the sum of the requirements for each end use. The requirements for each end use are the product of the number of accounts having that end use and the energy used by an average account having that end use. 8.4. Residential Forecast Comparison The December 2005 residential forecast is higher than the December 2004 forecast for most of the forecast period. The differences in the residential sales forecast compared to the previous forecast, with DSM, is -29 GWh (-0.2%) for 2005/06, +283 GWh (+1.6%) for 2009/10, +140 GWh (+0.7%) for 2015/16 and 250 GWh (-1.1%) for 2024/25. The main reasons for the difference between the previous forecast and the current forecast are: (a) higher anchor point; (b) lower number of accounts forecast; and (c) higher forecast for use rate. (a) Anchor Point: In 2004, the forecast called for 2004/05 billed sales with DSM impacts to be 15,698 GWh. Actual weather normalized billed sales for 2004/05 were 15,976 GWh, 278 GWh or 1.7% higher than forecast. Higher sales for fiscal 2004/05 put upward pressure on the current forecast. PAGE 29 (b) Number of Accounts: The ending number of accounts for 2004/05 was 1,484,339. This was 6,661 accounts or 0.45% below the forecast value of 1,491,000. Since forecast sales are calculated by multiplying forecast use rate by forecast accounts, the forecast for the number of accounts can have a significant impact on the sales forecast. Fiscal year 2004/05 ended with about 0.5% fewer accounts than last year’s forecast. The new forecast for number of accounts, therefore starts with a lower prediction for 2005/06, and also has a slightly lower growth rate, down by about 2.3% in 2025/26. This put downward pressure on the current forecast and explains why the current sales forecast is not above the previous forecast for all years of the forecast period. (c) Use Rate: The main reason that weather normalized 2004/05 sales were above the previous forecast is because the annual use per account was higher than expected. Weather normalized use rate in 2004/05 was 227 KWh or 2.14% higher than forecast. A higher than expected use rate put upward pressure on the current forecast. The previous forecast called for use rate to show a small decline in 2004/05 in response to the 4.85% increase in electricity rates applied in 2004, followed by very slow growth (less than 0.25% annually) over the rest of the forecast period. The current forecast assumes that the impact from the rate increase is fully reflected in the historical sales. The current forecast calls for use rate to increase over the entire forecast period. The current forecast for the growth in the use rate is 0.7% over the next 5 years, 0.4% over the next 10 years, and 0.3% over the next 20 years. PAGE 30 Table 8.1. Residential Sales Before DSM Lower Vancouver South Northern Mainland Island Interior Region Sales Sales Sales Sales (GWh) (GWh) (GWh) (GWh) Actual (not weather-adjusted) 1999/00 7,670 3,909 1,583 1,409 2000/01 7,695 3,863 1,617 1,397 2001/02 7,975 4,001 1,656 1,458 2002/03 8,120 3,981 1,729 1,457 2003/04 8,447 4,123 1,803 1,526 2004/05 8,316 4,021 1,805 1,478 Forecast (Residential sales forecast based on “normal weather”) 2005/06 8,670 4,213 1,839 1,522 2006/07 8,993 4,377 1,903 1,580 2007/08 9,189 4,475 1,938 1,600 2008/09 9,387 4,576 1,974 1,622 2009/10 9,584 4,675 2,012 1,645 2010/11 9,778 4,770 2,048 1,668 2011/12 9,980 4,866 2,086 1,693 2012/13 10,168 4,957 2,120 1,714 2013/14 10,355 5,044 2,154 1,735 2014/15 10,541 5,129 2,188 1,755 2015/16 10,726 5,213 2,221 1,775 2016/17 10,913 5,297 2,255 1,795 2017/18 11,101 5,381 2,289 1,815 2018/19 11,287 5,466 2,323 1,835 2019/20 11,473 5,549 2,357 1,855 2020/21 11,659 5,631 2,392 1,875 2021/22 11,845 5,713 2,426 1,896 2022/23 12,033 5,796 2,461 1,916 2023/24 12,222 5,879 2,496 1,936 2024/25 12,409 5,962 2,531 1,956 2025/26 12,596 6,045 2,566 1,976 Growth Rates 5 years: 99/00 to 04/05 1.6% 0.6% 2.7% 1.0% 5 years: 04/05 to 09/10 2.9% 3.1% 2.2% 2.2% 11 years: 09/10 to 15/16 2.3% 2.4% 1.9% 1.7% 21 years: 15/16 to 25/26 2.0% 2.0% 1.7% 1.4% BC Hydro Total Sales (GWh) 14,572 14,573 15,090 15,287 15,899 15,620 16,245 16,854 17,202 17,559 17,916 18,265 18,625 18,959 19,289 19,613 19,935 20,260 20,586 20,911 21,235 21,557 21,881 22,206 22,533 22,858 23,182 1.4% 2.8% 2.2% 1.9% PAGE 31 Table 8.2. Residential Sales With DSM Lower Vancouver South Northern Mainland Island Interior Region Sales Sales Sales Sales (GWh) (GWh) (GWh) (GWh) Actual (not weather-adjusted) 1999/00 7,670 3,909 1,583 1,409 2000/01 7,695 3,863 1,617 1,397 2001/02 7,975 4,001 1,656 1,458 2002/03 8,120 3,981 1,729 1,457 2003/04 8,447 4,123 1,803 1,526 2004/05 8,316 4,021 1,805 1,478 Forecast (Residential sales forecast based on “normal weather”) 2005/06 8,604 4,176 1,824 1,510 2006/07 8,895 4,310 1,882 1,563 2007/08 9,066 4,418 1,912 1,579 2008/09 9,243 4,509 1,944 1,597 2009/10 9,420 4,599 1,978 1,617 2010/11 9,594 4,684 2,010 1,637 2011/12 9,777 4,772 2,044 1,658 2012/13 9,954 4,858 2,076 1,678 2013/14 10,135 4,941 2,108 1,697 2014/15 10,308 5,021 2,139 1,715 2015/16 10,473 5,095 2,169 1,732 2016/17 10,634 5,167 2,197 1,748 2017/18 10,800 5,241 2,227 1,764 2018/19 10,973 5,320 2,258 1,782 2019/20 11,147 5,397 2,290 1,800 2020/21 11,318 5,472 2,321 1,817 2021/22 11,495 5,550 2,354 1,836 2022/23 11,683 5,633 2,389 1,856 2023/24 11,871 5,716 2,424 1,876 2024/25 12,059 5,799 2,459 1,896 2025/26 12,245 5,882 2,493 1,916 Growth Rates 5 years: 99/00 to 04/05 1.6% 0.6% 2.7% 1.0% 5 years: 04/05 to 09/10 2.5% 2.7% 1.8% 1.8% 11 years: 09/10 to 15/16 2.1% 2.2% 1.7% 1.4% 21 years: 15/16 to 25/26 1.9% 1.8% 1.6% 1.2% BC Hydro Total Sales (GWh) 14,572 14,573 15,090 15,287 15,899 15,620 16,116 16,651 16,976 17,295 17,613 17,925 18,251 18,565 18,882 19,184 19,469 19,746 20,031 20,332 20,635 20,928 21,235 21,560 21,887 22,212 22,536 1.4% 2.4% 2.0% 1.8% PAGE 32 9 Commercial Forecast 9.1. Summary Characteristics - BC Hydro’s commercial sector provides electricity to British Columbia’s service sector. This is a diverse set of BC Hydro customers who operate a wide range of facilities such as office buildings, retail stores, institutions and transportation infrastructure. The largest portions of these facilities are buildings, with the remaining “non-buildings” including facilities and infrastructure such as transportation systems and public utilities. Electricity Use – BC Hydro’s commercial sector currently consumes about 28% of BC Hydro’s total annual billed sales. This electricity is used to provide a range of services such as lighting, ventilation, heating, cooling, refrigeration, and domestic hot water. These vary considerably between the different types of buildings. Unlike the residential sector, the diversity of use in the commercial sector means that consumption in the sector is not strongly correlated to weather. Drivers – At an aggregate level, consumption in the commercial sector is tied closely with economic activity in the province - the stronger the economy the more services needed, the greater the electricity consumption of the commercial sector. As a result future economic trends, such as expected provincial Gross Domestic Product, are good indicators of future electricity consumption in the sector. At a more detailed level, the consumption in the commercial sector is related to the growth in the number of buildings and facilities and the amount of energy required to meet various energy needs. Trends – Electricity consumption of the commercial sector can vary considerably from year to year reflecting the level of activity in BC’s service sector. During periods where the economy is strong, electricity sales tend to be high. Sales to BC Hydro’s commercial sector in 2004/05 grew by 1.5 per cent over the previous year. The BC Economy is anticipated to continue on a positive upswing, with implications for strong commercial sector electricity sales growth. Forecast – Billed commercial sales were 14,362 GWh in 2004/05. Commercial sales, before DSM impacts, are forecast to grow from 14,362 GWh in 2004/05 to 16,493 GWh in 2009/10, 18,550 GWh in 2015/16, and 22,008 GWh in 2025/26. These increases represent growth rates of 2.8 per cent over the next five years (2004/05 to 2009/10), 2.4 per cent over the next 11 years (2004/05 to 2015/16), and 2.1 per cent over the next 21 years of the forecast (2004/05 to 2025/26). PAGE 33 9.2. Major Trends The BC economy continues to be strongly influenced by primary resource industries and their associated international markets. BC’s service sector has however been growing significantly in recent years and currently employs 80 per cent of the total provincial population and is responsible for approximately 76 per cent of the province’s GDP. This compares to 20 per cent and 24 per cent for employment and GDP for the goods-producing sector. As a result, the service sector has been the primary employment growth engine for the province and this trend is expected to continue. In addition, the BC service sector is also less susceptible to fluctuations in international markets than the goods-producing sector, which contributes to more stable growth. In the short term, variations in the economy on a year-to-year basis will affect electricity sales in terms of occupancy rates and/or performance of the province’s commercial buildings and facilities. In the long term, the size and structure of the economy as well as the size and age of the population will dictate what types of commercial buildings and facilities are constructed to meet the needs of the province’s service sector. Over the long term, commercial sales growth is likely to be influenced by factors including: • • • • • • • • • • A growing population, which increases the demand for most general services; A gradual shift in the structure of British Columbia from a goods-based to a more service-based economy; An aging population, which will require increased heath care services; Increases in electric intensity, a result of greater use of electronic and information end-use technologies; Continued growth in the tourism sector; New electricity-using technologies becoming more common in commercial establishments; Continued growth of Vancouver as an international finance centre; Growth in pipeline infrastructure used to transport oil and gas; BC’s continued role as Canada’s link with Pacific markets; and The potential for further development of a high tech sector within the province. Table 9.1 provides a summary of historical and forecast by distribution and transmission customers. Table 9.2 summarizes the same information as Table 9.1 with DSM. PAGE 34 Table 9.1. Commercial Sales Before DSM (GWh) Distribution Sales Actuals 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates 5 years: 99/00 to 04/05 5 years: 04/05 to 09/10 11 years: 04/05 to 15/16 21 years 04/05 to 25/25 12,541 12,991 12,924 13,119 13,509 13,696 14,216 14,664 14,939 15,285 15,620 15,988 16,241 16,545 16,814 17,133 17,448 17,740 18,058 18,353 18,692 19,037 19,406 19,765 20,145 20,518 20,887 Transmission Sales 635 663 659 610 642 666 643 646 778 807 873 874 974 997 1,059 1,078 1,102 1,166 1,166 1,107 1,234 1,241 1,227 1,215 1,170 1,138 1,121 Total Sales 13,176 13,654 13,583 13,729 14,151 14,362 14,859 15,310 15,717 16,092 16,493 16,862 17,215 17,542 17,873 18,211 18,550 18,906 19,224 19,460 19,926 20,278 20,633 20,980 21,315 21,656 22,008 1.8% 2.7% 2.2% 2.0% 1.0% 5.6% 4.7% 2.5% 1.7% 2.8% 2.4% 2.1% PAGE 35 Table 9.2. Commercial Sales With DSM (GWh) Distribution Sales Actuals 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates 5 years: 99/00 to 04/05 5 years: 04/05 to 09/10 11 years: 04/05 to 15/16 21 years 04/05 to 25/25 12,541 12,991 12,924 13,119 13,509 13,696 14,147 14,573 14,754 15,032 15,320 15,662 15,918 16,239 16,511 16,859 17,166 17,460 17,778 18,068 18,413 18,747 19,105 19,463 19,839 20,211 20,578 Transmission Sales 635 663 659 610 642 666 639 624 743 764 823 820 909 928 987 976 1,005 1,067 1,068 1,013 1,134 1,153 1,150 1,139 1,097 1,067 1,051 Total Sales 13,176 13,654 13,583 13,729 14,151 14,362 14,786 15,197 15,497 15,796 16,143 16,482 16,827 17,167 17,498 17,835 18,171 18,527 18,846 19,081 19,547 19,900 20,255 20,602 20,936 21,278 21,629 1.8% 2.3% 2.1% 2.0% 1.0% 4.3% 3.8% 2.2% 1.7% 2.4% 2.2% 2.0% PAGE 36 9.3. Methodology The main determinant of the commercial electricity sales forecast is the level of future economic activity in the province. The methodology for the commercial electricity sales forecast is described in Section 4.2. 9.4. Commercial Forecast Comparison The December 2005 commercial forecast is higher than the December 2004 forecast for all years in the forecast period. The increase in the forecast compared to the previous forecast reflects: (i) the growth in Commercial sales and GDP between 2003/04 and 2004/05; and (ii) the GDP forecast as the main driver of this year’s forecast. Commercial sales grew by 211 GWh (1.5%) and GDP grew by 3.9% between 2003/04 and 2004/05. As a result of this growth, this puts upward pressure on this year’s forecast as the forecast incorporates the historical growth in sales and GDP. The increase in the commercial sales forecast compared to the previous forecast, with DSM, is 29 GWh (0.2%) for 2005/06, 460 GWh (2.9%) for 2009/10, 786 GWh (4.5%) for 2015/16 and 802 GWh (3.9%) for 2024/25. PAGE 37 10 Industrial Forecast 10.1. Summary Characteristics - BC Hydro’s industrial sector is concentrated in a limited number of industries, the most important of which are pulp and paper, wood products, chemicals, metal mining and coal mining. The remaining industrial load is made up of a large number of small and medium sized manufacturing establishments. Electricity Use - BC Hydro’s industrial sector currently accounts for 38% of BC Hydro’s total annual billed sales. This electricity is used to provide a range of services including fans, pumps, compression, conveyance, processes such as cutting, grinding, stamping and welding and electrolysis. In comparison to the commercial sector, space conditioning, lighting, refrigeration and freezing loads are relatively less important. They can, however, be significant in small and medium sized facilities. Unlike the residential sector, the diversity of use in the industrial sector is not strongly correlated to weather. Trends - Electricity consumption in the industrial sector is quite volatile, driven substantially by economic conditions in the United States, China, Japan, and other Asian economies. Volatility in industrial electricity consumption accounts for approximately 45% of year over year changes in total billed sales. Export sales are a key determinant of domestic industrial output and of the industrial demand for electricity. Other key determinants of the industrial load include increasing levels of self-generation and co-generation. Drivers – As in the case of the commercial sector, industrial electricity consumption is tied closely with the level of economic activity in the province. There is a strong relationship between industrial electricity consumption and provincial Gross Domestic Product. Forecast - Billed industrial sales were 19,635 GWh in 2004/05. Forecast sales to industrial customers before DSM, are expected to grow from 19,635 GWh in 2004/05 to 21,307 GWh in 2009/10, 22,120 GWh in 2015/16, and 25,148 GWh in 2025/26. These increases represent growth rates of 1.6% over the next five years (2004/05 to 2009/10), 1.1% over the next 11 years (2004/05 to 2015/16), and 1.2% over the next 21 years of the forecast (2004/05 to 2025/26). Forecast industrial sales before DSM are shown in Table 10.1 and the forecast with DSM are shown in Table 10.2. 10.2. Sector Outlooks Resource extraction and processing form the basis of BC’s industrial economy. Key industrial activities include metal mining, coal mining, ore processing and smelting, wood extraction, saw milling, pulp and paper and chemical production. Approximately 80 per cent of BC Hydro’s sales to the industrial sector are made PAGE 38 to large-scale customers involved in the extraction and processing of natural resources and the remaining are smaller manufacturing companies. Given the importance of the forestry and saw-milling, pulp and paper and mining sectors for BC Hydro’s electricity load, this section summarises the current situation and prospects in each of these sectors. It should be noted that all three of these sectors have experienced large swings in the level of economic activity over the past decade. 10.2.1. Medium-Term Forestry Outlook Shipments of lumber from British Columbia, which have recently been averaging between 12 billion and 13 billion board feet per year, now go almost entirely to the United States, Japan and other parts of Canada. Shipments to European and other markets, which peaked at about 1.5 billion board feet in the late 1980s, have fallen to about 500 million board feet. A majority of interior timber is sold into the US market, while a majority of coastal timber is shipped to Japan. The health of the B.C. lumber market depends critically on the strength of the American and Japanese economies as well as the degree of market access for B.C. lumber. Performance of the American economy has been stable, while Japanese performance is accelerating. Key issues for B.C. lumber sales in the medium term include: • Changes in domestic timber supply. Move to smaller, poorer quality, second growth timber on Vancouver Island and beetle destruction of large volumes of wood are expected to eventually raise fiber costs. In the shorter term there has been a significant increase in the allowable cut to take advantage of dead and beetle-damaged timber. • Changes in lumber demand. On-going oversupply in the North American market is likely to continue to provide downward pressure on prices over the medium term and lead to more rationalisation of mills and mill closures. Increases in interest rates have created a more balanced housing market, which should moderate lumber demand. • Changes in productivity. Many BC mills have substantially improved productivity in recent years by improving product flow, replacing older equipment with newer and more efficient equipment, and improving use of labour on the mill floor. Two catalysts for these changes are duties imposed by the United States in the softwood lumber dispute and the high Canadian dollar. 10.2.2. Medium-Term Pulp and Paper Outlook BC pulp capacity is about 8.3 million metric tons per year with about 55% bleached softwood kraft pulp, 33% thermal mechanical pulp (TMP) or chemical thermal mechanical pulp (CTMP), 7% unbleached kraft and about 5% in other grades. Consolidation in the pulp and paper industry, coupled with four years of relatively PAGE 39 poor markets in both North America and overseas through 2005, have led to strong pressures to rationalise production and reduce costs. A number of BC pulp and paper mills are reasonably high cost producers compared to competitors in Latin America and Asia, rising production costs combined with a strong Canadian dollar have furthered weakened profitability. For the near term, however, the pulp and paper outlook is neutral, because Asian demand remains strong and interior fiber costs have been low due to beetle killed wood. Key issues for BC pulp and paper sales in the medium term include: • Ongoing decline in the North American newsprint market, which is expected to foreshadow a slow decline in the world newsprint market. A number of Canadian producers, who operate mills in BC, have announced closures at operations in Newfoundland, Quebec, Ontario, and Saskatchewan in an attempt to stabilise North American prices. Advertising revenue will continue to move away from newspapers into internet media as print subscriptions decline. Slowing demand growth for most printing and writing paper grades, with some limited bright spots such as directory paper (example: Yellow Pages). Several BC companies appear to be making a successful transition away from kraft pulp and newsprint to higher value products. Growing demand for paper products on the part of China (a positive factor for B.C.) combined with expected increased supply in China from new and very large mills (a negative factor for B.C.) as well as continued expansion of production of newer, low cost mills in developing countries. Rationalisation of the pulp and paper industry, leading to decisions on investment and upgrading made on a global basis, with negative implications for high cost paper machines and pulp and paper mills, including some B.C. facilities. • • • 10.2.3. Medium-Term Mining Outlook The mining sector had gross sales of some $3.5 billion in 2004. Some 20 mines currently purchase power from BC Hydro, with more metal mines and coal mines currently in development. Most production is for export and there is relatively little domestic processing or manufacturing based on mineral production. Coal is sold primarily to Japan and China with the demand for metallurgical coal decreasing in Japan but increasing in China. Key issues for B.C. mining sales in the medium term include: • There is renewed interest in the BC mining industry with increasing exploration spending and numerous companies submitting mining projects to the Environmental Assessment Office. Few high quality ore deposits have been found in BC in recent years, while major finds have taken place in Latin America, Africa and • PAGE 40 Indonesia. However, a number of large, new mines have been proposed in Northwest BC. • • Spot prices for gold, copper, and coal are at record highs, but are expected to return to more normal levels within 3-5 years. An increase in the Canadian dollar has tended to reduce profitability because most costs are in Canadian dollars. At the same time, copper and gold prices have been strong so that the overall impact on cash flow has been positive. At least two BC mines are nearing the end of reserve life. If metal prices remain high, these mines would be expected to continue operating for at least the next 3-4 years. • 10.3. Methodology The main determinant of industrial electricity sales is the level of future economic activity in the industrial sector. Since long-term forecasts for industrial GDP are not available, the forecast uses total provincial GDP as a proxy driver. At present, two sources of information on GDP are used for the load forecast. These sources are the B.C. Ministry of Finance, September 2005 Budget Update and the Conference Board of Canada. For the period 2005 through 2009, the forecast is based on provincial real GDP estimates of 3.4% (2005), 3.2% (2006), and 3.1% (2007-2009). For the period 2010 to 2025, the forecast is based on long term Conference Board of Canada GDP estimates of growth. Different sources are used for the following reasons: • The B.C. Ministry of Finance forecast is not available for the 20 years needed for the load forecast so the B.C. Ministry of Finance forecast needs to be supplemented by other forecasts; and The Conference Board is one of only a few organisations that provides long term GDP forecasts. • An adjustment to the regression results for the industrial transmission forecast is made to reflect Highland Valley Copper planned shut down in 2013/14. Other adjustments to the transmission regression results reflect sales and variance information from the current years forecast. 10.4. Industrial Forecast Comparison The December 2005 industrial forecast is higher than the December 2004 forecast for all years in the forecast period. The main reason is an increase in the GDP growth assumptions, which is a key driver of the forecast. The increase in the industrial sales forecast compared to the previous forecast, with DSM, is 896 GWh (4.7%) for 2005/06, 1,607 GWh (8.8%) for 2009/10, 1,254 GWh (6.5%) for 2015/16 and 1,633 GWh (7.4%) for 2024/25. PAGE 41 Table 10.1. Industrial Sales by Sector Before DSM (GWh) Transmission Voltage Customers Metal Mines Actual 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates 5 years: 99/00 to 04/05 5 years: 04/05 to 09/10 11 years: 04/05 to 15/16 21 years: 04/05 to 25/26 1,619 1,996 1,952 1,873 1,906 1,990 2,299 2,554 2,583 2,653 2,997 3,122 3,509 3,560 3,609 2,683 2,718 2,756 2,789 2,824 2,862 2,899 2,936 2,973 3,008 3,043 3,079 4.2% 8.5% 3.0% 2.1% Coal Mines 558 547 554 516 467 507 514 535 554 583 642 647 639 646 653 664 676 690 702 715 729 743 756 770 784 797 812 -1.9% 4.8% 2.7% 2.3% Wood Paper Chemical Other Distribution All Sectors 3,828 3,627 3,884 4,046 3,893 4,208 4,179 4,217 4,254 4,287 4,323 4,385 4,434 4,482 4,531 4,585 4,640 4,690 4,745 4,800 4,859 4,912 4,974 5,029 5,090 5,154 5,218 1.9% 0.5% 0.9% 1.0% Total Sales 17,890 18,579 17,739 18,596 18,725 19,635 20,237 20,513 20,781 21,036 21,307 21,557 21,863 22,149 22,437 21,829 22,120 22,425 22,703 22,997 23,312 23,617 23,929 24,234 24,534 24,834 25,148 1.9% 1.6% 1.1% 1.2% 826 892 885 928 937 1,003 1,059 1,092 1,108 1,121 1,110 1,122 1,114 1,130 1,145 1,166 1,183 1,201 1,217 1,234 1,252 1,270 1,288 1,306 1,323 1,340 1,358 4.0% 2.0% 1.5% 1.5% 8,685 8,937 7,957 8,534 8,785 9,178 9,349 9,413 9,527 9,580 9,446 9,454 9,363 9,473 9,599 9,772 9,912 10,062 10,194 10,336 10,487 10,637 10,785 10,934 11,076 11,216 11,365 1.1% 0.6% 0.6% 1.0% 1,710 1,724 1,626 1,798 1,787 1,812 1,893 1,741 1,775 1,815 1,798 1,825 1,813 1,855 1,885 1,926 1,944 1,963 1,979 1,996 2,015 2,033 2,051 2,068 2,084 2,100 2,117 1.2% -0.2% 0.6% 0.7% 663 856 880 902 950 938 944 961 980 997 991 1,002 991 1,003 1,015 1,033 1,047 1,063 1,077 1,092 1,108 1,123 1,139 1,154 1,169 1,184 1,199 7.2% 1.1% 1.0% 1.2% PAGE 42 Table 10.2. Industrial Sales by Sector With DSM (GWh) Transmission Voltage Customers Metal Mines Actual 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates 5 years: 99/00 to 04/05 5 years: 04/05 to 09/10 11 years: 04/05 to 15/16 21 years: 04/05 to 25/26 1,619 1,996 1,952 1,873 1,906 1,990 2,260 2,463 2,414 2,451 2,737 2,835 3,175 3,215 3,265 2,397 2,441 2,480 2,514 2,548 2,588 2,643 2,818 2,855 2,889 2,924 2,960 4.2% 6.6% 1.9% 1.9% Coal Mines 558 547 554 516 467 507 514 532 534 557 607 609 598 603 610 622 637 651 663 676 689 708 726 740 754 767 782 -1.9% 3.7% 2.1% 2.1% Wood Paper Chemical Other Distribution All Sectors 3,828 3,627 3,884 4,046 3,893 4,208 4,152 4,159 4,005 3,977 3,977 4,002 4,002 4,024 4,070 4,113 4,195 4,242 4,299 4,359 4,412 4,523 4,621 4,677 4,741 4,807 4,872 1.9% -1.1% 0.0% 0.7% Total Sales 17,890 18,579 17,739 18,596 18,725 19,635 19,911 19,816 19,713 19,752 19,903 20,021 20,179 20,383 20,670 20,132 20,507 20,814 21,097 21,392 21,706 22,187 22,809 23,114 23,414 23,714 24,028 1.9% 0.3% 0.4% 1.0% 826 892 885 928 937 1,003 1,014 1,015 1,033 1,037 1,031 1,039 1,031 1,044 1,059 1,078 1,099 1,117 1,133 1,150 1,168 1,194 1,237 1,255 1,271 1,288 1,306 4.0% 0.5% 0.8% 1.3% 8685 8,937 7,957 8,534 8,785 9,178 9,197 9,096 9,073 9,044 8,900 8,862 8,736 8,816 8,943 9,137 9,309 9,461 9,594 9,734 9,887 10,104 10,346 10,494 10,635 10,774 10,922 1.1% -0.6% 0.1% 0.8% 1,710 1,724 1,626 1,798 1,787 1,812 1,840 1,626 1,707 1,730 1,705 1,722 1,701 1,736 1,766 1,812 1,836 1,856 1,873 1,890 1,910 1,941 1,968 1,985 2,001 2,017 2,034 1.2% -1.2% 0.1% 0.6% 663 856 880 902 950 938 934 925 947 956 946 952 936 945 957 973 990 1,007 1,021 1,035 1,052 1,074 1,093 1,108 1,123 1,137 1,152 7.2% 0.2% 0.5% 1.0% PAGE 43 10.5. Risks and Uncertainties Beyond the general risk of higher or lower than economic growth materialising, a significant risk to the industrial forecast pertains to discrete one-time changes in sales to the base metal mining and pulp and paper sectors. Historically, these two sectors have accounted for 70-80% of variation in industrial load on a year/year basis. Risks not represented in the forecast relate to large one-time additions or contractions of load as a result of new investment, strikes or closure of major facilities. Because it is difficult to assess the likelihood of these events, the approach has been to relate the industrial load to GDP rather than attempt to forecast these individual discrete events. The one exception is Highland Valley Copper as described in section 10.3. PAGE 44 11 Peak Forecast 11.1 Introduction Characteristics: BC Hydro’s forecast peak demand is defined as the expected maximum amount of electricity consumed in a single hour under an average coldest day assumption established as the design temperature. BC Hydro is a winter peaking utility because the system has a greater share of winter heating load than summer air conditioning load. The distribution peak is the most sensitive to temperature. The transmission peak is considered to be responsive to external market conditions and changes in demands for BC’s key industrial commodities such as wood, pulp and paper and metals. Trends: The peak demand is quite variable from year to year and reacts strongly to temperature. In the winter of 2004/05, BC Hydro’s recorded Domestic Peak was 9,437 MW, which occurred at a daily average temperature of -1.9OC. The weather adjusted Domestic peak for fiscal 2004/05 is 9,787 MW. Including sales to other utilities, the actual integrated system peak is 9,762 MW and the weather-adjusted peak is 10,110 MW. Design Temperature: Since temperature is the greatest driver in the variability of the peak demand around its long term trend, the peak forecast is prepared based on a design temperature. BC Hydro’s design temperature is based on an average of the coldest daily average temperatures using the most recent 30 years of weather data. Using the most recent 30 years of temperatures is consistent with British Columbia Utilities Commission direction in the VIGP decision. The design temperature for the System is -5.3 OC. 13 Drivers: BC Hydro’s peak forecast is based on the peak demands from its distribution, transmission and wholesale customers including BC Hydro sales to other utilities such as Fortis BC. The growth in the distribution peak forecast reflects the forecast of accounts, forecast of energy sales to the General Service rate class and the forecasts of BC Hydro’s substations. Housing starts and employment forecasts are the main drivers of the accounts forecast and general sales forecast. Information from BC Hydro’s Key Account Managers, industry trends and previous trends in the peaks from individual transmission accounts are the main drivers of the transmission peak forecasts. Forecast: The 2005 integrated system peak forecast for 2005/06 before DSM is 10,298 MW. The 2005 integrated system peak forecast with DSM is 10,205 MW. The Vancouver Island peak forecast with transmission losses for 2005/06 before DSM is 2,331 MW and with DSM is 2,318 MW. The peak forecast is sum of four regional distribution peaks, four regional transmission peaks, the peak requirements for the other utilities and system losses. Each of the four regional distribution peaks have unique design temperatures of: -5.3OC for Lower Mainland, -3.6 OC for Vancouver Island, -16.4 OC O for Southern Interior, and -28.5 C for Northern Region. 13 PAGE 45 11.2 Methodology and Procedure 11.2.1 Distribution Peak The annual distribution peak forecast includes a distribution peak forecast guideline for the 15 planning areas. The forecast acts as a guideline for the expected total level of growth for each of substations located in the various planning areas. The guidelines 14 are provided to BC Hydro distribution planners in order to produce a forecast for each of BC Hydro’s 220 substations. The growth forecast from the guidelines is combined with growth rates from the individual substations to produce a total non-coincident regional distribution growth forecast for each of BC Hydro’s four main service regions which include the Lower Mainland, Vancouver Island, Northern Region and South Interior. The region’s total growth forecast along with region’s total weather adjusted peak produces the total non-coincident distribution peak forecast. During the preparations of the total regional distribution peak forecasts, the results of a top down weather adjusted procedure are compared to the results of bottom up weather adjusted procedure. The process of reconciling the weather adjusted peaks from the two approaches involves a blending or weighting of both approaches to weather normalization. Nevertheless, both approaches reduce the risk of relying upon any single approach. The top-down approach to weather normalization includes a daily peak model, which is cubic in temperature data, and a weather-normalization procedure where the model is simulated using the most recent 30 years of weather. The bottom-up approach to weather normalization involves determining a weather normalized peak for each substation using regression methods that involve substation peak data and temperatures recorded from weather stations located closest to the substation. 11.2.2 Transmission Peak BC Hydro prepares a non-coincident transmission peak forecast for each of its transmission accounts. A total non-coincident transmission account forecast for each of the four main service regions is prepared as the sum of the individual accounts forecast in that region. 11.2.3 Regional Total and System Total Peak After the regional non-coincident distribution and transmission peak forecasts are completed, BC Hydro calculates peak forecast in each region that is equal to the coincident sum of the regional distribution and transmission peaks and applicable other utilities’ peak sales. In any region, the total regional peak is: (11.1) Regional Peakt = Regional Transmission Peak t + Regional Distribution Peakt + Other Utilities Peakt The regional peak forecast and system peak forecast are developed using BC Hydro’s regional and system peak forecast model. The model also incorporates The guideline forecast provides a guide for the total non-coincident growth for all of the substations located in a planning area. 14 PAGE 46 DSM savings and estimated rate impacts for distribution and transmission peaks as appropriate. The integrated system is the sum across the four regions, i =1, 2, 3, 4 and t = 1, 2, 3…21 years: (11.2) System Peakt = Σ Coincidence Factori * Distribution Peakti + Σ Coincidence Factori * Transmission Peakti + Σ Coincidence Factori * Other Utilities Peakti + Lossest 11.3 Peak Forecast Comparison Table 11.1 provides the integrated system peak forecast before and with DSM. The forecast of the integrated system peak before DSM for the winter of 2005/06 is 10,298 MW, and 13,756 MW for the winter of 2025/26. Most of the increase in the 2005 forecast is attributed to the increase in the industrial activity of last year. The increase in demand for commodities, the increase in mining activity and the increase of prices for copper and pulp and paper all have contributed to a strong growth in the transmission peak. Growth in the transmission sector peak between 2003/04 and 2004/05 was 220 MW or 12.5% on a coincident basis. The previous year’s strong growth has raised the transmission contribution to this year’s forecast. Growth in the distribution sector peak between 2003/04 and 2004/05 was 105 MW or 1.5% on a coincident basis. A lower rate of growth in the accounts and employment forecast has led to a slower increase in the distribution peak forecast compared to last year. PAGE 47 Table 11.1 Integrated System Peak Before and With DSM BEFORE WITH DSM DSM Total Total (MW) (MW) Actual 2004/05 9,762 Weather-adjusted 2004/05 10,110 Forecast 2005/06 10,298 2006/07 10,511 2007/08 10,772 2008/09 10,978 2009/10 11,151 2010/11 11,274 2011/12 11,425 2012/13 11,565 2013/14 11,720 2014/15 11,756 2015/16 11,875 2016/17 12,050 2017/18 12,229 2018/19 12,409 2019/20 12,592 2020/21 12,779 2021/22 12,968 2022/23 13,161 2023/24 13,356 2024/25 13,554 2025/26 13,756 Growth Rates 2.0% 5 years 04/05 to 09/10 1.5% 11 years 04/05 to 15/16 1.5% 21years 04/05 to 25/26 9,762 10,110 10,205 10,340 10,526 10,680 10,821 10,912 11,037 11,166 11,321 11,363 11,487 11,655 11,826 12,006 12,188 12,395 12,625 12,817 13,013 13,211 13,413 1.4% 1.2% 1.4% PAGE 48 12 Demand Side Management 12.1. Demand Side Management Background BC Hydro launched its Power Smart demand side management (DSM) initiative in 1989, achieving electricity savings of 2,500 GWh per year by the end of the 1990’s. In 2001/02, BC Hydro launched a new round of DSM programs to save electricity through energy efficiency and load displacement. In BC Hydro’s resource planning process, this second round of DSM is referred to as DSM 2 or Power Smart 2. 12.2. Forecast Table 12.1 is a forecast of incremental DSM energy and peak savings from 2005/06 through 2025/26 from current and planned DSM activities. These energy savings are incremental to the DSM energy and peak savings due to historical DSM activities that are reflected in actual electricity consumption during the load forecast’s reference year. Table 12.1 Forecast Demand Side Management Savings 15 ENERGY Energy DSM Efficiency (GWh) (GWh) 84 282 535 1,177 528 1,013 1,515 1,844 2,057 2,259 2,446 2,535 2,549 2,504 2,457 2,503 2,541 2,564 2,585 2,438 2,144 2,144 2,144 2,144 2,144 84 282 535 1,022 320 516 848 1,102 1,309 1,511 1,697 1,787 1,801 1,825 1,863 1,911 1,952 1,977 1,998 2,026 2,043 2,043 2,043 2,043 2,043 Load Displacement (GWh) 155 209 497 667 743 748 748 748 748 748 678 593 592 589 587 587 411 101 101 101 101 101 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 PEAK Energy DSM Efficiency (MW) (MW) 13 42 80 173 84 155 222 268 298 325 350 360 360 354 349 356 361 362 364 345 308 308 308 308 308 13 42 80 153 57 90 135 172 201 228 252 262 263 266 271 279 285 286 287 292 294 294 294 294 294 Load Displacement (MW) 20 27 65 87 97 97 97 97 97 97 88 77 77 77 76 76 53 13 13 13 13 13 Fiscal Year Actual 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Fiscal Year The energy and peak DSM savings in Table 12.1 are before distribution and transmission losses. 15 PAGE 49 13 Glossary Binary Variable is a variable whose value is either zero or one. Binary variables are often used as independent variables in regression models in order to account for events that either occur or do not occur. In this latter context, binary variables are often referred to as “dummy variables” in regression. Coincidence Factor A ratio reflecting the relative magnitude of a region’s (or customer’s or group of customers’) demand at the time of the system’s maximum peak demand to the region’s (or customer’s or group of customers’) maximum peak demand. Consumer Price Index (CPI) An inflation index calculated by comparing the price of a typical bundle of goods in the year in question to the price of the same goods in a set reference year. Demand-Side Management (DSM) Refers to activities that occur on the demand side of the revenue meter and are influenced by the utility. DSM activities results in a change in electricity sales and revenues. Demand Side Management savings include load displacement and energy efficiency. Demand Side Management savings include incremental load displacement and energy efficiency savings. Note that B.C. Hydro reports historical sales information as including the impact of DSM savings realized up to that year. Distribution voltage customer A BC Hydro customer who receives electricity via distribution lines that operate at relatively low voltage (35 kV and less). Diversity That quality or characteristic by which individual maximum demands occur at different times. Diversity may be examined on an hourly, daily, monthly or yearly basis. Domestic System Peak includes the peak requirements for BC Hydro’s distribution and transmission customers in its service territory; sales to the City of New Westminster and system transmission and distribution losses. Econometric modelling The use of statistical techniques, typically regression analysis of time-series and/or cross-sectional data, to detect statistically verifiable relationships, coherent with economic theory, between an explained variable (e.g. electricity consumption) and explanatory variables (e.g. industry output, prices of alternative energy inputs and GDP). Elasticity The proportionate change in a dependent variable, (e.g. electricity consumption, divided by the proportionate change in a specified independent variable; electricity price). A dependent variable is highly elastic with respect to a given independent variable if the calculated elasticity is much greater than one. The dependent variable is inelastic if the elasticity is less than one. End-use model A model used to analyze and forecast energy demand, which focuses on the end uses or services provided by energy. Typical end uses are lighting, process heat and motor drive. For a given industry, the model estimates the influence of prices and technological change on the evolution of the secondary energy inputs required to satisfy the industry's end uses over time. Energy The amount of electricity delivered or consumed over a certain time period, measured in multiples of watt-hours. A 100-watt bulb consumes 200 watt-hours in two hours. A typical BC Hydro residential account consumes about 10,000 kWh (10 million watt-hours) annually. PAGE 50 Energy Efficiency Includes both being more efficient with electricity used (i.e. producing the same, equivalent or greater output while using less electricity per unit of output) and using less electricity (i.e. conservation). Intensity A unitized measure of energy consumption, typically in kilowatt-hours per unit of stock. For example, kWh per account in the residential sector or kWh per unit of production in the industrial sector. Gross Domestic Product (GDP) A measure of the total flow of goods and services produced by the economy over a specified time period, normally a year or quarter. It is obtained by valuing outputs of goods and services at market prices (alternatively at factor cost), and then aggregating the total of all goods and services. Gigawatt-hour (GWh) A measure of electrical energy, equivalent to one million kilowatt-hours. (See Units of Measure.) Integrated System That portion of the BC Hydro system which is connected as one whole. Non-integrated facilities refer to generating facilities that are not connected to the system, located in remote areas of the province. Integrated System Peak includes the peak requirements for BC Hydro’s distribution and transmission customers in its service territory; sales to Other Utilities, which includes Seattle City Light, New Westminster, Fortis BC, ATCO Electric and Tongass Power and Light Co. Inc.; and system transmission and distribution losses Kilowatt-hour (kWh) A measure of electrical energy, equivalent to the energy consumed by a 100-watt bulb in 10 hours. (See Units of Measure.) Load The total amount of electrical power demanded by the utility's customers at any given time, typically measured in megawatts (MW). Load Displacement Projects that involve the installation of new self-generation facilities at customer sites, with the electricity generated being use don-site by the customer, with a resultant decrease in the purchase of electricity from BC Hydro. Megawatt (MW) A unit used to measure the capacity or potential to generate or consume electricity. One MW equals one million watts. (See Units of Measure.) Monte Carlo method A technique for estimating probabilities involving the construction of a model and the simulation of the outcome of an activity a large number of times. Random sampling techniques are used to generate a range of outcomes. Probabilities are estimated from an analysis of this range of outcomes. MVA Megavolt-Amps – a unit of apparent power. Apparent power is real power in MW divided by power factor. Natural conservation The increase in energy efficiency that would occur in the absence of any utility-induced demand-side management program, all other things being equal. Non-coincident refers to use of a coincidence factor that is a ratio reflecting the relative magnitude of a region’s (or customer’s or group of customers’) demand at the time of the system maximum peak demand to the region’s (or customer’s or group of customers’) maximum peak demand. Normalization The correction of actual customer sales and peak demand for factors such as unusually warm or cold weather. PAGE 51 Price elasticity of demand The percentage change in quantity demanded, divided by the percentage change in price that caused the change in quantity demanded. Real price increases that have been adjusted for changes in prices of all goods. The nominal price of an item may rise by 10 per cent over a year, but inflation (and assumed wages) may have risen by seven per cent over the same time period. Therefore the effective price increase faced by the consumer is three per cent. It is necessary to deflate current prices by an appropriate inflation index (the CPI in Canada) to convert money values to constant prices or real terms. Region A geographical sub-division of the BC Hydro service area. Four regions exist: Lower Mainland, Vancouver Island, South Interior and the Northern Region. Stock A quantity representing a number of energy consuming units. For example, in the residential sector, stock is the number of accounts or housing units; in the commercial sector, stock is represented by the floor area of commercial building space. System peak demand The greatest combined demand of all BC Hydro customers faced by the generation system during a given fiscal year. Total system peak is the sum of total integrated system peak and the total peak demand from the non-integrated areas. Transmission voltage customer A BC Hydro customer that is supplied its electricity via high-voltage transmission lines (60 kV or above). Units of measure The large amounts of electricity generated and consumed on a system-wide basis are discussed in multiples of the basic units of watt and watthours. Kilowatts and megawatts are used to measure power, and kilowatthours, megawatt-hours, and gigawatt-hours are used to measure energy. The equivalence are: 1 kilowatt (kW) 1 megawatt (MW) 1 kilowatt-hour (kWh) 1 megawatt-hour (MWh) 1 gigawatt-hour (GWh) = = = = = 1000 watts 1000 kilowatts or 1 million watts 1000 watt-hours 1000 kilowatt-hours or 1 million watt-hours 1000 megawatt-hours or 1 billion watt-hours PAGE 52 14 References BC Hydro, Conservation Potential Review, 1994, 2003. B.C. Ministry of Energy and Mines, Energy for our Future: A Plan for B.C., November 2002. B.C. Ministry of Finance, September Budget Update – 2005/06 to 2007/08, September 2005 B.C. Statistics, B.C. Population Forecast, June 2004. Conference Board of Canada, Economic Indicators, June 2005. Gellings, C.W. ed., Demand Forecasting in the Electric Utility Industry, Second Edition, 1996. Malatest, R.A., British Columbia Regional Economic Outlook 2004 to 2023, July 2004. PAGE 53 Appendix 1. Weather Normalization Weather-normalized sales are an estimate of the sales that would have been made if normal weather had been experienced (ie a 10 year rolling average of degree days). Sales are adjusted using heating degree-days (a standard approach used by the utility industry). A degree-day is measure of coldness, defined by the number of degrees below 18 degrees Celsius in (A1.1), for the average daily temperature. For example, if the average temperature on day t is 12 degrees Celsius then that day has 18-12 = 6 heating degree-days. The heating degree-days for a month are the sum of the heating degree-days for the days in that month. Formally, for day t heating degree-days is defined in (A1.1) where max is the maximum function. (A1.1) heating degree-dayt = max (18°C – average daily temperature, zero) Note that degree-days are never negative because space heating systems are not required to produce heat at temperatures above 18°C. We assume that the monthly residential use rate for a given class of residential accounts can be modelled using the following cubic polynomial (A1.2.). (A1.2) use ratet = α + β*HDDt + χ*HDDt2 + δ*HDDt3 + εt The most recent 36 months of data available is used to estimate each regression, which is modelled using ordinary least squares. To calculate the weather-adjusted use rate for a particular period, the heating degree-days for the period are substituted into the estimated regression equation (A1.2). It is important to note the following points: • First, weather normalization is undertaken for the residential sector only since only limited evidence exists of weather response for the commercial and industrial sectors. This means that when weather-normalized totals are reported, only the residential part of the total is actually weatheradjusted. Although this is not viewed as a major source of error, research is being conducted to determine if and how the commercial and industrial loads should be weather normalized. Second, the model actually normalizes the use per account or the use rate rather than sales per se. Normalized sales are then calculated as normalized use rate multiplied by the average number of accounts for the class. Eight classes are used in these calculations, namely a heating and non-heating class in each of the four regions. Third, because this forecast uses billed sales rather than the unknown actual consumption by class, monthly heating degree-days are allocated using a 35/50/15 per cent adjustment to match the assumed pattern of meter reading. For example, to weather normalize the month of November, November sales would be regressed against the summation of 1) 35% of November degree days 2) 50% of October degree days and 3) 15% of September degree days. • • PAGE 54 Table A1.1 compares the actual and weather-normalized sales for BC Hydro’s service territory for the fiscal years 1993/94 to 2004/05. Table A1.1. Actual and Weather-Normalized Sales for BC Hydro Service Territory 16 Year Actual (GWh) 40,979 41,616 42,851 43,598 42,607 44,863 45,638 46,806 46,412 47,612 48,774 49,618 Weather Normalized (GWh) 41,367 41,992 43,055 43,095 43,115 45,418 45,542 46,628 46,252 47,789 48,776 49,973 1993/94 1994/95 1995/96 1996/97 1997/98 1998/99 1999/00 2000/01 2001/02 2002/03 2003/04 2004/05 BC Hydro Service Territory sales is the total sales to the residential, commercial and industrial customers. 16 PAGE 55 Appendix 2. Ordinary Least Squares-Based Forecasts Most economic analysis deals with situations where the outcome variables can be assumed to be continuous and normally distributed. These include decisions about how much of a product to purchase, how much of a product to produce and what price to charge for a product. In each of these cases, explaining the determinants of the variable typically involves modelling the outcome variable as a continuous function of a set of k explanatory or independent variables, a set of k associated parameters, plus an error term ε assumed to be normally distributed with mean zero, constant variance σ2 and covariance’s equal to zero (there is no correlation between errors for different observations). The basic idea of least squares regression is to choose the parameters to minimize the sum of squares of the errors. The assumption of normally distributed errors is not necessary to apply ordinary least squares regression, but some assumption on the distribution of errors is needed to generate test statistics for the parameters. The rationale for using minimum least squared error as the criterion for choosing parameter values makes intuitive sense since large errors are more important than are small errors. Equally important is the fact that ordinary least squares estimators have desirable properties in the classical regression context. In particular, ordinary least squares estimates are unbiased and have minimum variance in the class of linear unbiased estimators. In other words, they are the best estimator for this class of regression problem. In the typical set-up, then, the regression model is given by: (A2.1) Here: • • • • yt is the dependent variable at observation; t is a k×1 vector of independent variables at observation t; β is a k×1 vector of parameters assumed constant for all observations; and T is the number of observations. yt = x′tβ + εt , where εt ∼ N(0, σ2) and t = 1,2, … T In other words, (A2.1) is a set of T equations where the value of yt at time t is a linear function of k variables, x1t, x2t, … , xkt. It is convenient for what follows to write equation (A2.1) in matrix form as follows (A2.2) where: • • • • y is a T×1 vector; X is a k×T matrix; β is a T×1 vector; and ε is a T×1 vector. y = X′β + ε, PAGE 56 Assuming that X is a non-stochastic matrix of full rank k ≤ T that satisfies the regularity condition limT→∞ (X′X/T) = Q, where Q is a finite and non-singular matrix, and using E for the expectation operator, note that the E(ε) is zero since the expectation of each of its components is zero. Defining the sum of the squared errors as S, note that the variance of the errors is the expectation of S: (A2.3) S ≡ ε′ε = (y – X′β)′ (y – X′β) and Eεε’ = σ2I The ordinary least squares estimators of the vector of parameters β* and the variance of the errors σ2* are found by minimizing the sum of the squared errors (A2.4) ∂S/∂β = -2X′y + 2X′Xβ* = 0 Solving (A4.4) for β* the estimated value of β yields the following expression (A2.5) β* = (X′ X)-1 X′ y The ordinary least squares estimate of the variance of the errors σ2 is given by the following expression: (A2.6) σ2* = ε*′ε* / (T – k) where ε* = y - Xβ* The estimate of σ2 is used to estimate confidence intervals and to conduct hypotheses tests for the parameters. Further, the least squares estimates of the parameters can be shown to be unbiased and consistent estimates of the population parameters. It may be useful to find the effect of a change in an independent variable on the outcome variable. This partial effect is found by calculating the relevant partial derivative and in the linear model is just the value of the regression coefficient for that variable. (A2.7) ∂y/∂x = ∂(Xβ + ε)∂x = β, since the derivative of ε is zero. Finally, the measure of goodness of fit for an ordinary least squares regression is R-squared adjusted for degrees of freedom, which is calculated as the explained sum of square divided by the total sum of squares times T – k divided by T. PAGE 57 Appendix 3. Maximum Likelihood-Based Forecasts The main alternative to least squares estimation is maximum likelihood estimation. It is normally used in circumstances where the underlying assumptions of the standard linear model are not met, but it is convenient to first review maximum likelihood estimation of the standard linear model (described in Appendix 2) before considering the more complicated case of auto-correlated residuals. The basic idea of maximum likelihood estimation is to choose estimates for the parameter values that maximize the probability that the distribution represented by the estimated parameters generated the observed sample. Formally, consider the normal linear regression model considered in Appendix 2, the joint likelihood for the T observations is the product of T normal densities as follows: (A3.1) L = f(y1, y2, …, yT) = (2πσ2)-T/2 exp{-(2σ2)-1(y - Xβ)′(y - Xβ)} Taking the log of this expression yields: (A3.2) ln L = -T/2 ln(2π) – T/2 ln(σ2) – (2σ2)-1(y - Xβ)′ (y - Xβ) Maximizing the log likelihood function with respect to the parameters yields the first order conditions given by expressions (A5.3) and (A5.4): (A3.3) (A3.4) ∂L/∂β = -σ-2 (-X′y + X′Xβ) = 0 ∂L/∂σ2 = -T (2σ2)-1 + (2σ4)-1 (y - Xβ)′(y - Xβ) = 0 Solving these equations for the unknown parameters yields the estimators (A3.5) and (A3.6): (A3.5) (A3.6) β** = (X′X)-1X′y σ2** = ε**′ε**/T = (T – k)/T σ2* The maximum likelihood estimate of β is the same as the ordinary least squares estimate for this model and is unbiased and consistent. The maximum likelihood estimate of σ2 is different from the ordinary least squares estimate by the factor T/(t – k), and is therefore a biased estimator, but it is a consistent estimate since as T→∞ the bias goes to zero. In fact, the strength of maximum likelihood estimators is that under fairly general conditions they are consistent, asymptotically normal and asymptotically efficient. These features account for their widespread use in econometrics in situations where least squares estimates are inappropriate because the requirements of the classical linear regression model are not met. PAGE 58 Up to now we have assumed that covariance’s of the errors are zero or that there is no auto-correlation. However in many cases, errors are correlated over time, often due to persistent shocks reflecting the inertia of economic processes or due to omitted variables that are hopefully uncorrelated to variables in the model. Consider the linear model: (A3.7) where: • • • • y is a T×1 vector; X is a k×T matrix; β is a T×1 vector; and ε is a T×1 vector; y = X′β + ε, but where: (A3.8) εt = ρεt-1 + ut, t = 1,2, …,T Assuming that the absolute value of the parameter ρ is less than one, the ut are independently and identically distributed with variance σu2, and εt are generated by a stationary stochastic process beginning in the infinite past. Roughly speaking, a stochastic process is stationary if the mean, variance and covariance’s for given lags are constant over time. The form of the errors is awkward to work with and the calculations can be simplified by expanding the previous expression by making successive substitutions for εt to yield: (A3.9) εt = Σρiut-1 , where the sum runs over i = 0,1,…,∞ Using the assumptions on ut and the formula for the sum of a converging series gives the variance of εt as follows: (A3.10) E(εt2) = ρ0E(ut2) + ρ2(ut2) + ρ4E(uu4) + … = σu2/(1 - ρ2) = σε2 Finally, the covariance of εt with εt-i is needed, which is: (A3.11) E(εtεt-i) = E([ut + ρut-1 + ρ2ut-2 + …]*[ut + ρut-1 + ρ2ut-2 +…]) = ρiσε2 This gives all the variances and covariance’s in the variance-covariance matrix for εt. Noting that every term contains σu2, this common term can be extracted and the variance-covariance matrix can be written as follows: (A3.12) Eεε’ = σu2Ω PAGE 59 If the value of ρ were known, the value of β could be found that minimizes this sum of squares as with for the ordinary least squares estimator to yield the generalized least squares estimator: (A3.13) β* = (X’Ω-1X)-1X’Ω-1y But since the value of ρ is not known, a maximum likelihood estimator can be used, which gives us consistent and asymptotically efficient estimates of the parameters. Starting by formulating the likelihood function in the usual way and taking its log that yields: (A3.14) ln L(y, X, β, σu2, ρ) = -T/2 ln(2π) – 1/2 ln|σu2Ω| – (2σu2)-1(y - Xβ)′ Ω-1(y - Xβ) This expression can be simplified by partially maximizing with respect to β(ρ) and σu2(ρ) which, noting that these expressions are functions of ρ, yields the simpler concentrated likelihood function: (A3.15) ln L*(ρ, y, X) = -T/2{ln(2π) + 1} -T/2 ln{[σu2(ρ)][(1 - ρ2)-1/T]} Maximizing this function with respect to ρ is then a relatively straightforward numerical estimation problem. The method of Beach-MacKinnon (1978) can be used to maximize this function. PAGE 60 Appendix 4. Commercial Sector Regressions Table A4.1 summarizes econometric estimates of the determinants of electricity consumption for the commercial sector. The table shows results from ordinary least squares (OLS) regression and maximum likelihood regression of commercial sales on real GDP. Ordinary least squares regression is a method of choosing parameters to minimise the sum of squares of errors produced as a function of a set of variables (See Appendix 2). Maximum likelihood estimation is an alternative to ordinary least squares regression (See Appendix 3). Table A4.1. Econometric Model of Commercial Sales Variable Constant GDP Adjusted R-sq Log likelihood Durbin – Watson Estimated Auto OLS 1810 (770) 17 0.093 (0.0064) 0.96 0.6151 0.6925 Maximum Likelihood 2522 (1110) 0.087 (0.0091) -69.43 1.3067 0.3467 The OLS commercial sales equation for Model 1 has a very good fit with an adjusted R-squared values of 0.96. The Durbin Watson statistic of 0.62 indicates the presence of autocorrelation in the OLS model. (The Durbin-Watson is a measure of auto-correlation, which means that the errors are correlated over time rather than being independent as assumed in the ordinary least squares model.) If the errors are auto-correlated, then use of a maximum likelihood estimation procedure with correction for autocorrelation can lead to statistically superior estimates. Table A4.2 shows the forecast results of the OLS regression and maximum likelihood regression. 17 Numbers in parenthesis are standard errors. PAGE 61 Table A4.2 Forecast Total Commercial Sales Before DSM (GWh) Model 1 Commercial Commercial OLS ML 11,611 11,611 11,928 11,928 12,226 12,226 12,465 12,465 12,814 12,814 13,176 13,176 13,654 13,654 13,583 13,583 13,729 13,729 14,151 14,151 14,362 14,362 15,210 14,818 15,606 15,182 16,040 15,580 16,455 15,961 16,894 16,363 17,301 16,737 17,689 17,093 18,051 17,425 18,414 17,759 18,785 18,098 19,156 18,439 19,548 18,799 19,898 19,120 20,270 19,462 20,671 19,829 21,058 20,184 21,450 20,544 21,832 20,895 22,204 21,236 22,579 21,580 22,969 21,938 Year 1994/95* 1995/96* 1996/97* 1997/98* 1998.99* 1999/00* 2000/01* 2001/02* 2002/03* 2003/04* 2004/05* 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Note: * = Actual PAGE 62 Appendix 5. Industrial Sector Regressions Tables A5.1 through A5.4 summarise alternative econometric estimates of the determinants of electricity consumption for the industrial sector while Tables A5.5 and A5.6 provide the results. The regressions are as follows: OLS (Transmission). Ordinary least squares regression of transmission industrial sales on GDP and with/without a binary variable for 1997 and 2001 because of work stoppages (Model 2 and Model 1 respectively). Ordinary least squares regression is a method of choosing parameters to minimise the sum of squares of errors produced as a function of a set of variables (See Appendix 2 for additional details regarding OLS); ML (Transmission). Maximum likelihood regression of industrial sales on GDP and with/without a binary variable for 1997 and 2001 (Model 2 and Model 1 respectively). Maximum likelihood regression is a method to choose estimates for parameter values that maximise the probability that estimated parameters will represent an observed sample (See Appendix 3 for additional details regarding ML); OLS (Distribution). Ordinary least squares regression of distribution industrial sales on GDP and with/without a binary variable for 2000 because of low economic activity with the recession (Model 2 and Model 1 respectively). ; ML (Distribution). Maximum likelihood regression of distribution industrial sales on GDP and with/without a binary variable for 2000 (Model 2 and Model 1 respectively). The OLS industrial transmission sales equation for Model 1 has an adequate fit with an adjusted R-squared values of 0.53, although the Durbin-Watson statistic suggests the possible presence of auto-correlation. (The Durbin-Watson is a measure of auto-correlation, which means that the errors are correlated over time rather than being independent as assumed in the ordinary least squares model. If the errors are auto-correlated, then use of a maximum likelihood estimation procedure may lead to statistically superior estimates.) The ML industrial transmission sales equation looks reasonable with coefficients having the anticipated signs. The Durbin-Watson statistic is better (closer to the desired value of 2.0) suggesting that auto-correlation has been reduced. Table A5.1. Econometric Model of Industrial Transmission Sales (Model 1) Variable Constant GDP binarytran Adjusted R-sq Log likelihood Durbin-Watson Estimated Auto OLS 6437 (2219) 18 0.0636 (0.0183) 0.53 2.50 -0.25 ML 6545 (1834) 0.0627 (0.0151) -83.8 2.22 -0.11 18 Numbers in parenthesis are standard errors. PAGE 63 The OLS industrial transmission sales equation for Model 2 has a good fit with an adjusted R-squared values of 0.82, and the Durbin-Watson statistic suggests there is no auto-correlation. The ML industrial transmission sales equation looks reasonable with coefficients having the anticipated signs. The Durbin-Watson statistic is better (very close to the desired value of 2.0). According to this model, a one billion dollar increase in provincial GDP increases the industrial transmission demand for electricity by 62 MWh. Table A5.2. Econometric Model of Industrial Transmission Sales (Model 2) Variable Constant GDP binarytran Adjusted R-sq Log likelihood Durbin-Watson Estimated Auto OLS 6875 (1379) 19 0.0616 (0.011) -1069 (272) 0.82 2.07 -0.03 ML 6887 (1338) 0.0615 (0.0110) -1064 (274) -78.2 1.94 0.04 The OLS industrial distribution sales equation for Model 1 has an adequate fit with an adjusted R-squared values of 0.46, and the Durbin-Watson statistic is close to 2.0. The maximum likelihood equation for distribution sales again looks reasonable. The maximum likelihood equation has a Durbin-Watson statistic that is better (very close to the desired value of 2.0) suggesting that autocorrelation has been reduced. According to this model, a one billion dollar increase in provincial GDP increases the industrial distribution voltage transmission demand for electricity by 12 MWh. Table A5.3. Econometric Model of Industrial Distribution Sales (Model 1) Variable Constant GDP binarydist Adjusted R-sq Log likelihood Durbin-Watson Estimated Auto Durbin-Watson OLS 2398 (474) 20 0.0120 (0.0039) 0.46 2.28 -0.14 ML 2423 (379) 0.0118 (0.0031) -66.7 2.10 -0.05 The OLS industrial distribution sales equation for Model 2 has a good fit with an adjusted R-squared values of 0.76, although the Durbin-Watson statistic suggests the presence of auto-correlation. The maximum likelihood equation for distribution sales again looks reasonable. The maximum likelihood equation 19 20 Numbers in parenthesis are standard errors. Numbers in parenthesis are standard errors. PAGE 64 has a Durbin-Watson statistic that is much better (close to the desired value of 2.0) suggesting that auto-correlation has been reduced. According to this model, a one billion dollar increase in provincial GDP increases the industrial distribution voltage transmission demand for electricity by 13 MWh. Table A5.4. Econometric Model of Industrial Distribution Sales (Model 2) Variable Constant GDP binarydist Adjusted R-sq Log likelihood Durbin-Watson Estimated Auto Durbin-Watson OLS 2290 (318) 21 0.0131 (0.0026) -295 (85) 0.76 3.14 -0.57 ML 2398 (91) 0.0122 (0.00076) -311 (31) -54.7 2.60 -0.30 21 Numbers in parenthesis are standard errors. PAGE 65 Table A5.5. Forecast Industrial Sales Before DSM (GWh) Model 1 Year Transmission OLS 13,347 13,948 13,613 12,553 14,257 14,062 15,052 13,855 14,550 14,819 15,412 15,602 15,873 16,170 16,454 16,754 17,032 17,297 17,545 17,794 18,047 18,301 18,569 18,808 19,063 19,337 19,602 19,870 20,132 20,385 20,642 20,909 Transmission ML 13,347 13,948 13,613 12,553 14,257 14,062 15,052 13,855 14,550 14,819 15,412 15,575 15,842 16,135 16,414 16,710 16,985 17,246 17,489 17,735 17,984 18,235 18,499 18,734 18,986 19,256 19,516 19,781 20,038 20,228 20,541 20,804 Distribution OLS 3,740 3,682 3,834 3,786 3,820 3,828 3,627 3,884 4,046 3,893 4,208 4,128 4,179 4,235 4,289 4,346 4,398 4,448 4,495 4,542 4,590 4,638 4,689 4,734 4,783 4,834 4,885 4,935 4,985 5,033 5,081 5,132 Distribution ML 3,740 3,682 3,834 3,786 3,820 3,828 3,627 3,884 4,046 3,893 4,208 4,116 4,167 4,222 4,274 4,330 4,381 4,430 4,476 4,522 4,569 4,616 4,665 4,709 4,756 4,807 4,856 4,906 4,954 5,001 5,048 5,098 Industrial OLS 17,087 17,630 17,447 16,339 18,077 17,890 18,679 17,739 18,596 18,712 19,620 19,730 20,052 20,405 20,743 21,100 21,430 21,745 22,040 22,336 22,637 22,939 23,258 23,542 23,846 24,171 24,487 24,805 25,117 25,418 25,723 26,041 Industrial ML 17,087 17,630 17,447 16,339 18,077 17,890 18,679 17,739 18,596 18,712 19,620 19,691 20,009 20,357 20,688 21,040 21,366 21,676 21,965 22,257 22,553 22,851 23,164 23,443 23,742 24,063 24,372 24,687 24,992 25,229 25,589 25,902 1994/95* 1995/96* 1996/97* 1997/98* 1998.99* 1999/00* 2000/01* 2001/02* 2002/03* 2003/04* 2004/05* 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 * = actuals PAGE 66 Table A5.6. Forecast Industrial Sales Before DSM (GWh) Model 2 Year Transmission OLS 13,347 13,948 13,613 12,553 14,257 14,062 15,052 13,855 14,550 14,819 15,412 15,750 16,012 16,300 16,575 16,865 17,135 17,392 17,632 17,872 18,118 18,364 18,624 18,885 19,102 19,367 19,623 19,883 20,136 20,382 20,631 20,889 Transmission ML 13,347 13,948 13,613 12,553 14,257 14,062 15,052 13,855 14,550 14,819 15,412 15,747 16,009 16,296 16,570 16,860 17,130 17,386 17,625 17,866 18,110 18,356 18,615 18,847 19,093 19,358 19,613 19,873 20,126 20,371 20,619 20,877 Distribution OLS 3,740 3,682 3,834 3,786 3,820 3,828 3,627 3,884 4,046 3,893 4,208 4,179 4,235 4,296 4,355 4,417 4,474 4,529 4,580 4,631 4,683 4,736 4,791 4,840 4,893 4,949 5,004 5,059 5,113 5,165 5,218 5,273 Distribution ML 3,740 3,682 3,834 3,786 3,820 3,828 3,627 3,884 4,046 3,893 4,208 4,152 4,204 4,261 4,315 4,372 4,425 4,476 4,523 4,571 4,619 4,668 4,719 4,765 4,814 4,866 4,917 4,968 5,018 5,067 5,116 5,167 Industrial OLS 17,087 17,630 17,447 16,339 18,077 17,890 18,679 17,725 18,596 18,712 19,620 19,929 20,247 20,596 20,930 21,282 21,609 21,921 22,212 22,503 22,801 23,100 23,415 23,725 23,995 24,316 24,627 24,942 25,301 25,547 25,849 26,162 Industrial ML 17,087 17,630 17,447 16,339 18,077 17,890 18,579 17,739 18,596 18,712 19,620 19,899 20,213 20,557 20,885 21,232 21,555 21,862 22,148 22,437 22,729 23,024 23,334 23,612 23,907 24,224 24,530 24,841 25,144 25,438 25,735 26,044 1994/95* 1995/96* 1996/97* 1997/98* 1998.99* 1999/00* 2000/01* 2001/02* 2002/03* 2003/04* 2004/05* 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 * = actuals PAGE 67 Appendix 6. Monte Carlo Methods This Appendix describes the Monte Carlo model that is used to assess the uncertainty associated with BC Hydro's Load Forecast. This description includes a discussion of the methodology used, model results and assumptions on the parameters included in the model. Load forecasting involves considerable uncertainty. The demand for electricity depends on a large number of factors which fluctuate widely with time and which are difficult to measure. Some of these factors include population, gross domestic product, weather, technology, conservation, alternate energy source options, business climate experienced by major customers and the changing tastes and behavior of end use customers. The challenge of assessing the uncertainty of the load forecast is to quantify the way in which uncertainty in the major causal factors flows through to impact the resultant load. To quantify load forecast uncertainty, BC Hydro uses a Monte Carlo model and Monte Carlo simulation techniques. The model and simulation analysis proceeds as follows: • First, several major input variables or causal factors are identified. These are: Economic Growth (measured by GDP); Price of Electricity (electricity rates); DSM; Weather (measured by heating degree days); Electric Intensity (the impact of technology, consumer tastes and other residual variables) and Elasticity of Load (with respect to GDP and Electricity Price). Second, probability distributions are assigned to each input variable and a model is specified that defines the mathematical relationship between the input variable and the output variables. Third, a large number of random samples are taken from the input probability distributions. The model is used, with each sample as input, to calculate a large number of simulations of the output variables. These simulations are used to construct probability distributions for the output variables. • • The Monte Carlo model calculates the impact of the major causal factors that drive load. The model perturbs the reference case forecast by calculating the impacts for each of the factors. The impact factors are random variables. Each of the sectors Residential, Commercial and Industrial is perturbed separately, and has separate impact factors, but essentially the same methodology is used for all of them. The model is implemented in Microsoft EXCEL augmented with @RISK software. Energy demand for each sector is computed by the following equation. (A6.1) Et =0 Et ItP ItG ItW ItU ItD Here 0Et is base case energy demand, Et perturbed energy demand, and the impact factors are identified by their superscripts; P for electricity price (rates), G for GDP, W for weather, U for electric intensity and D for DSM. PAGE 68 Equation (A6.1) is used to calculate the random variable for energy demand before DSM. A random variable for DSM savings is then calculated and subtracted to give energy after DSM. Impact of GDP Uncertainty: In order to assess the impact of uncertainty in future GDP, the base case GDP is perturbed. The base case GDP is denoted by 0Gt and the perturbed GDP is denoted by Gt . The perturbed GDP starts off being equal to the base case GDP in the first year. It then grows at a growth rate equal to the base case GDP growth rate ( 0gt ) plus a random perturbation growth rate ( gt ). This random perturbation is a normally distributed random variable with zero mean and a standard deviation of 1.70%. That is: (A6.2) (A6.3.) (A6.4) gt ∼ N(0,1.70%) Gt = Gt-1 [ 1 + 0gt + gt ] . IGt = exp( α0 Gt / 0Gt ) The perturbed GDP is calculated by: The impact factor for GDP is then given by the following equation: where α0 is the elasticity of load with respect to GDP. Impact of Price Uncertainty: Calculation of the impact factor for price changes ( IPt ) is similar. A random variable, the perturbed price Pt , is calculated starting from the base case price 0Pt . The perturbed price starts out being equal to the base case price in the initial year. It then grows at a rate equal to the base case growth rate plus a random perturbation. In the model, the random perturbation has a triangular distribution with parameters (-2.5%, 0, +2.5%). However, unlike the case of GDP, the impact of price change is assumed to take place with time lags. This assumption is made because it may take customers some time to adjust their consumption to price changes. It takes time to process information and to arrive at decisions. Moreover, capital investments may have to be made and may take time to complete. The time lagged effect of price changes is modeled by introducing different elasticities for price changes that occurred at different time periods in the past and by making these elasticities decline geometrically as they refer to times more remote from the present. Let εk be the Elasticity of Load at time t with respect to a price change at time tk , k=0,1,2,..., and let λ be a parameter such that 0 < λ < 1. Assume that: (A6.5) εk = ε0 λk It follows that as k increases, that is as one goes back in time from time t, the elasticity defined here goes to zero, because the lag parameter is less than one. Impact of Elasticity Uncertainty: The elasticity parameters used in the Monte Carlo model represent the best judgement of BC Hydro. Elasticity research has produced estimates with considerable variation. PAGE 69 Table 6.1 gives the elasticity parameters used in the current Monte Carlo model. Table 6.1. Elasticity Parameter for Monte Carlo Model Mean Short-term Price Elasticities Residential Commercial Industrial Long-term Price Elasticities Residential Commercial Industrial Lag Parameter Residential Commercial Industrial GDP Elasticity Residential Commercial Industrial -0.200 -0.100 -0.200 -0.270 -0.350 -0.280 0.259 0.714 0.286 0.670 0.780 0.500 Probability Distribution Triangular (-0.30, -0.20, -0.10) Triangular (-0.20, -0.10, 0.00) Triangular (-0.30, -0.20, -0.10) Triangular (-0.37, -0.27, -0.17) Triangular (-0.45, -0.35, -0.25) Triangular (-0.38, -0.28, -0.18) Not Applicable Not Applicable Not Applicable Triangular (0.47, 0.67, 0.87) Triangular (0.58, 0.78, 0.98) Triangular (0.30, 0.50, 0.70) In Table 6.1, Triang( a , b , c ) refers to a probability distribution known as a triangular distribution because its graph is a triangle. This distribution is zero for values of its random variable less than a or greater than c. It has a maximum (most probable) value at b. Impact of Electric Intensity Uncertainty: Electric intensity is a composite variable that tries to capture the residual effect on load of other factors such as changes in use rate, technology, consumer taste, household structure, business type, and inter-regional differences. The electric intensity factor starts out at 1.00 in the base year and grows at a rate that is, in each year, a random variable with the triangular distribution. The impact factor is defined by the following equations: (A6.6) IUt = IUt-1 ( 1 + gUt) IU0 = 1 where gUt denotes a random variable with a triangular distribution. As usual, the @RISK software allows the specification of probability distributions in Excel. Impact of Demand Side Management Uncertainty. The impact of uncertainty in energy savings due to DSM is treated separately from the other impacts. The base case DSM energy savings, 0St , is converted to a random variable DSM savings St by multiplying by an impact factor IDt . This impact factor for DSM is assumed to have a triangular distribution. (A6.7) IDt ∼ Triang(50% , 100% , 150%) PAGE 70 A random variable for energy after DSM is then calculated by adding this DSM savings random variable to the previously calculated random variable for energy demand before DSM. (A6.8) Etafter = Etbefore - St Impact of Weather: A preliminary analysis has been undertaken in order to add weather uncertainty to the model. The analysis is restricted to the Residential sector because it is here that the impact of weather is assumed to be most important. The weather analysis is based on 50 years of daily temperature data at Vancouver International Airport. For every day, the number of heating degree days is calculated by the formula: HDD=max(0, Daily Temperature -18). HDD is a measure of how much energy is needed to heat housing up to a typical comfortable temperature of 18 degrees C. Then, the annual sum of HDD was calculated for each year. A standard probability distribution of the Beta type was found to provide the best fit to this data. The Beta distribution has 4 parameters, and is written Beta(a1,a2,min,max). Min and max are the maximum and minimum, while a1 and a2 determine the shape of the distribution. The weather impact factor is calculated by: (A6.9) IWt = exp{ εW log( HDDt / 2,978) } where εW is the elasticity of Residential load with respect to HDD. εW is estimated judgmentally to be 0.374. The number 2,978 is the mean value of HDD in the Lower Mainland. IWt is a random variable as are the other impact factors. However it differs from the other impact factors in that its properties are the same for all years. This is because weather in each year is independent of weather in all other years. Therefore the width of the 80% confidence region for IWt does not increase with time. Peak Load: The Monte Carlo Model produces a stochastic peak load forecast from the energy forecast using load factors. This is achieved as follows: (A6.10) (A6.11) LPt ≡ Peak load at time t. L P t f ≡ Load factor = Et ( 1/8.76) ( 1/f ) PAGE 71 Appendix 7. Forecast Tables Tables A7.1 to A7.4 show the regional non-coincident (MVA) and coincident peak (MW) forecast for both distribution and transmission before and with DSM. Tables A7.5 and A7.6 show the domestic and regional peak forecast before and with DSM. Tables A7.7 and 7.8 show the weather-adjusted peak forecast before and with DSM. Tables A7.9 to A7.10 summarize BC Hydro’s reference load forecast including the effect before and with DSM. Table A7.11 to A7.14 summarize BC Hydro’s scenarios resulting from the Monte Carlo uncertainty analysis including the effects before and with DSM. PAGE 72 Table A7.1. Regional Non-Coincident and Coincident Distribution Peaks Before DSM Lower Mainland NonCoinc. Coinc. Peak Peak (MVA) (MW) Actual 2004/05 4,215 3,645 Weather-Adjusted Actual 2004/05 4,356 3,914 Forecast (Weather-Adjusted) 2005/06 4,512 4,054 2006/07 4,638 4,166 2007/08 4,766 4,282 2008/09 4,879 4,383 2009/10 4,960 4,456 2010/11 5.016 4,506 2011/12 5.081 4,564 2012/13 5.143 4,620 2013/14 5.204 4,675 2014/15 5.266 4,730 2015/16 5.327 4,786 2016/17 5.417 4,867 2017/18 5.509 4,949 2018/19 5.602 5,033 2019/20 5.697 5,118 2020/21 5.794 5,205 2021/22 5.892 5,293 2022/23 5.992 5,383 2023/24 6.093 5,474 2024/25 6.196 5,567 2025/26 6.301 5,661 Vancouver Island NonCoinc. Coinc. Peak Peak (MVA) (MW) 1757 1827 1867 1908 1942 1978 2011 2043 2071 2097 2123 2148 2172 2207 2242 2279 2315 2353 2391 2430 2469 2509 2549 1,701 1,682 1,719 1,757 1,788 1,821 1,851 1,881 1,906 1,931 1,955 1,978 1,999 2,032 2,065 2,098 2,132 2,166 2,201 2,237 2,273 2,310 2,347 South Interior NonCoinc. Coinc. Peak Peak (MVA) (MW) 918 904 932 950 966 982 999 1,014 1,025 1,037 1,048 1,059 1,071 1,088 1,105 1,122 1,140 1,158 1,177 1,195 1,214 1,234 1,253 811 815 841 857 872 886 901 915 925 935 945 956 966 981 997 1,013 1,029 1,045 1,061 1,078 1,095 1,113 1,130 Northern Region NonCoinc. Coinc. Peak Peak (MVA) (MW) 740 730 756 780 792 800 809 816 824 832 839 846 854 863 872 882 891 901 911 921 931 941 951 647 652 675 696 707 714 722 729 736 743 749 756 762 771 779 787 796 804 813 822 831 840 849 5 years 04/05 to 09/10 2.6% 2.6% 1.9% 1.9% 2.0% 2.0% 2.1% 2.1% 11 years 04/05 to 15/16 1.8% 1.8% 1.6% 1.6% 1.6% 1.6% 1.4% 1.4% 21years 04/05 to 25/26 1.8% 1.8% 1.6% 1.6% 1.6% 1.6% 1.3% 1.3% Notes: 1. Distribution peak forecast based on average of Substation forecast and Distribution Peak Guideline Forecast. 2. Growth rates based on weather adjusted peaks. PAGE 73 Table A7.2. Regional Non-Coincident and Coincident Transmission Peaks Forecast Before DSM Lower Mainland NonCoinc. Coinc. Peak Peak (MVA) (MW) Actual 2004/05 633 476 Weather-Adjusted Actual 2004/05 633 476 Forecast (Weather-Adjusted) 2005/06 633 443 2006/07 637 446 2007/08 659 462 2008/09 670 469 2009/10 673 471 2010/11 679 475 2011/12 688 481 2012/13 697 488 2013/14 713 499 2014/15 720 504 2015/16 727 509 2016/17 734 514 2017/18 741 519 2018/19 749 524 2019/20 756 529 2020/21 764 535 2021/22 771 540 2022/23 779 545 2023/24 786 550 2024/25 794 556 2025/26 802 561 5 years 04/05 to 09/10 11 years 04/05 to 15/16 21years 04/05 to 25/26 Vancouver Island NonCoinc. Coinc. Peak Peak (MVA) (MW) 646 646 621 624 637 646 649 655 649 655 665 670 676 684 693 702 710 719 728 738 747 756 766 501 501 497 500 509 517 519 524 519 524 532 536 541 547 554 561 568 576 583 590 598 605 613 South Interior NonCoinc. Coinc. Peak Peak (MVA) (MW) 331 331 334 339 377 380 379 382 385 393 405 298 298 302 306 310 314 318 322 326 330 334 338 234 234 255 259 289 291 290 292 294 301 310 228 228 231 234 237 240 243 246 249 252 256 259 Northern Region NonCoinc. Coinc. Peak Peak (MVA) (MW) 1070 1070 1061 1062 1083 1109 1149 1147 1193 1207 1237 1247 1257 1273 1288 1304 1320 1336 1352 1369 1385 1402 1419 828 828 769 769 785 804 833 831 864 874 897 904 911 922 934 945 956 968 980 992 1,004 1,016 1,029 1.2% 1.3% 1.1% -0.2% 0.6% 0.8% 0.1% 0.4% 0.8% 0.7% 0.7% 1.0% 2.8% -0.9% 0.1% 4.3% -0.2% 0.5% 1.4% 1.5% 1.4% 0.1% 0.9% 1.0% Notes: 1. Distribution peak forecast based on average of Substation forecast and Distribution Peak Guideline Forecast. 2. Growth rates based on weather adjusted peaks. PAGE 74 Table A7.3. Regional Non-Coincident and Coincident Distribution Peaks Forecast With DSM Lower Mainland NonCoinc. Coinc. Peak Peak (MVA) (MW) Actual 2004/05 4,215 3,645 Weather-Adjusted Actual 2004/05 4,356 3,914 Forecast (Weather-Adjusted) 2005/06 4,484 4,028 2006/07 4,596 4,129 2007/08 4,695 4,218 2008/09 4,791 4,304 2009/10 4,860 4,366 2010/11 4,907 4,408 2011/12 4,966 4,461 2012/13 5,027 4,516 2013/14 5,087 4,570 2014/15 5,148 4,625 2015/16 5,208 4,679 2016/17 5,294 4,756 2017/18 5,382 4,835 2018/19 5,474 4,918 2019/20 5,569 5,003 2020/21 5,666 5,090 2021/22 5,764 5,179 2022/23 5,864 5,268 2023/24 5,966 5,360 2024/25 6,069 5,452 2025/26 6,174 5,547 Vancouver Island NonCoinc. Coinc. Peak Peak (MVA) (MW) 1757 1827 1855 1887 1917 1947 1977 2005 2031 2057 2083 2108 2130 2163 2197 2233 2269 2306 2344 2382 2422 2461 2502 1701 1682 1708 1737 1765 1793 1820 1846 1870 1894 1918 1941 1961 1991 2023 2056 2089 2123 2158 2193 2230 2266 2304 South Interior NonCoinc. Coinc. Peak Peak (MVA) (MW) 918 904 926 942 951 963 977 990 1,000 1,011 1,022 1,033 1,044 1,060 1,077 1,094 1,112 1,130 1,149 1,167 1,186 1,206 1,225 811 815 836 850 857 869 881 893 902 912 922 932 942 957 971 987 1,003 1,020 1,036 1,053 1,070 1,088 1,105 Northern Region NonCoinc. Coinc. Peak Peak (MVA) (MW) 740 730 750 771 774 778 784 789 794 802 809 817 824 833 841 850 860 871 881 891 901 911 921 647 652 669 689 691 695 700 704 709 716 722 729 736 743 751 759 768 777 787 795 804 813 822 5 years 04/05 to 09/10 2.2% 2.2% 1.6% 1.6% 1.6% 1.6% 1.4% 1.4% 11 years 04/05 to 15/16 1.6% 1.6% 1.4% 1.4% 1.3% 1.3% 1.1% 1.1% 21years 04/05 to 25/26 1.7% 1.7% 1.5% 1.5% 1.5% 1.5% 1.1% 1.1% Notes: 1. Distribution peak forecast based on average of Substation forecast and Distribution Peak Guideline Forecast. 2. Growth rates based on weather adjusted peaks. PAGE 75 Table A7.4. Regional Non-Coincident and Coincident Transmission Peaks Forecast With DSM Lower Mainland NonCoinc. Coinc. Peak Peak (MVA) (MW) Actual 2004/05 633 476 Weather-Adjusted Actual 2004/05 633 476 Forecast (Weather-Adjusted) 2005/06 628 439 2006/07 605 424 2007/08 631 441 2008/09 634 444 2009/10 633 443 2010/11 634 444 2011/12 638 447 2012/13 645 452 2013/14 660 462 2014/15 670 469 2015/16 680 476 2016/17 687 481 2017/18 694 486 2018/19 702 491 2019/20 709 496 2020/21 722 505 2021/22 733 513 2022/23 740 518 2023/24 748 524 2024/25 756 529 2025/26 764 535 5 years 04/05 to 09/10 11 years 04/05 to 15/16 21years 04/05 to 25/26 Notes: 1. Distribution peak forecast based on average of Substation forecast and Distribution Peak Guideline Forecast. 2. Growth rates based on weather adjusted peaks. Vancouver Island NonCoinc. Coinc. Peak Peak (MVA) (MW) 646 646 619 620 610 612 611 612 602 606 615 623 631 640 648 657 666 681 693 702 711 721 730 501 501 495 496 488 489 489 490 481 485 492 498 505 512 519 526 533 544 554 562 569 577 584 South Interior NonCoinc. Coinc. Peak Peak (MVA) (MW) 331 331 332 335 363 363 359 360 361 368 379 274 275 279 283 287 291 298 303 307 312 316 320 234 234 254 256 278 278 275 276 276 281 290 210 211 214 217 220 223 228 232 235 238 242 245 Northern Region NonCoinc. Coinc. Peak Peak (MVA) (MW) 1,070 1,070 1,014 980 1,000 1,015 1,048 1,039 1,077 1,087 1,118 1,131 1,146 1,161 1,177 1,192 1,208 1,233 1,294 1,311 1,327 1,344 1,361 828 828 735 710 724 735 759 753 781 788 810 820 830 841 853 864 876 894 938 950 962 974 987 0.0% 0.6% 0.9% -1.4% 0.0% 0.6% -1.1% -0.2% 0.6% -0.5% 0.1% 0.7% 1.7% -1.6% -0.2% 3.2% -1.0% 0.2% -0.4% 0.6% 1.2% -1.7% 0.0% 0.8% PAGE 76 Table A7.5. Domestic System and Regional Peak Forecast Before DSM Lower Mainland (MVA) Vancouver Southern Island Interior (MW) (MVA) 2,202 2,183 2,216 2,256 2,298 2,337 2,371 2,405 2,425 2,455 2,487 2,514 2,540 2,579 2,619 2,659 2,700 2,742 2,784 2,827 2,871 2,915 2,960 1,285 1,231 1,277 1,298 1,342 1,358 1,376 1,396 1,412 1,432 1,451 1,380 1,390 1,408 1,427 1,446 1,465 1,484 1,504 1,524 1,544 1,564 1,585 Northern Region (MW) 1,476 1,480 1,444 1,467 1,493 1,519 1,555 1,561 1,601 1,618 1,647 1,661 1,674 1,694 1,713 1,733 1,753 1,774 1,794 1,815 1,836 1,857 1,878 Domestic Vancouver Island System with Trans. Losses (MW) (MW) 9,418 9,787 9,976 10,188 10,449 10,655 10,825 10,943 11,090 11,227 11,382 11,418 11,536 11,712 11,890 12,071 12,254 12,441 12,630 12,822 13,018 13,216 13,417 2,313 2,294 2,331 2,373 2,417 2,455 2,490 2,525 2,547 2,578 2,611 2,640 2,667 2,708 2,749 2,792 2,834 2,878 2,922 2,967 3,013 3,059 3,106 Actual 2004/05 4,287 Weather-Adjusted Actual 2004/05 4,591 Forecast (Weather-Adjusted) 2005/06 4,697 2006/07 4,813 2007/08 4,945 2008/09 5,054 2009/10 5,129 2010/11 5,184 2011/12 5,250 2012/13 5,313 2013/14 5,380 2014/15 5,441 2015/16 5,502 2016/17 5,589 2017/18 5,678 2018/19 5,768 2019/20 5,859 2020/21 5,952 2021/22 6,046 2022/23 6,142 2023/24 6,240 2024/25 6,339 2025/26 6,440 Growth Rates 5 years 04/05 to 09/10 2.2% 11 years 04/05 to 15/16 1.7% 21years 04/05 to 25/26 1.6% 1.7% 1.4% 1.5% 2.3% 1.1% 1.2% 1.0% 1.1% 1.1% 2.0% 1.5% 1.5% 1.7% 1.4% 1.5% Notes: 1. Regional peaks include distribution losses but not transmission losses, unless otherwise stated. 2. The Domestic Peak is defined in the Glossary. 3. Lower Mainland peak includes peak sales to City of New Westminster and sales to Seattle City Lights. 4. Southern Interior peak includes sales to Fortis B.C. 5. Peak forecasts and growth rates are on a weather-adjusted basis. 6. The recorded domestic peak was 9,437 MW on January 13, 2005. The actual domestic peak value as stated above has been reduced to account for losses associated with peak sales to other utilities including Fortis B.C. and Seattle City Light. 7. Vancouver Island Peak with transmission losses is the Vancouver Island Regional peak before DSM adjusted for estimated transmission losses. PAGE 77 Table A7.6. Domestic System and Regional Peak Forecast with DSM Lower Mainland (MVA) Vancouver Southern Northern Island Interior Region (MW) (MVA) (MW) 2,202 2,183 2,203 2,233 2,253 2,282 2,309 2,336 2,351 2,379 2,410 2,439 2,466 2,503 2,541 2,581 2,622 2,668 2,712 2,755 2,798 2,843 2,888 1,285 1,231 1,271 1,288 1,317 1,327 1,341 1,357 1,370 1,389 1,408 1,338 1,349 1,366 1,384 1,403 1,421 1,443 1,464 1,484 1,505 1,525 1,546 1,476 1,480 1,405 1,399 1,416 1,431 1,460 1,458 1,491 1,504 1,533 1,550 1,567 1,586 1,605 1,625 1,644 1,672 1,726 1,746 1,767 1,789 1,810 Domestic Vancouver Island System with Trans. Losses (MW) (MW) 9,418 9,787 9,882 10,018 10,203 10,358 10,494 10,582 10,702 10,828 10,982 11,024 11,149 11,316 11,488 11,668 11,850 12,057 12,287 12,479 12,675 12,873 13,074 2,313 2,294 2,318 2,349 2,370 2,397 2,425 2,454 2,470 2,498 2,531 2,561 2,589 2,628 2,668 2,710 2,753 2,800 2,847 2,892 2,937 2,984 3,031 Actual 2004/05 4,287 Weather-Adjusted Actual 2004/05 4,591 Forecast (Weather-Adjusted) 2005/06 4,668 2006/07 4,753 2007/08 4,860 2008/09 4,950 2009/10 5,012 2010/11 5,055 2011/12 5,111 2012/13 5,172 2013/14 5,238 2014/15 5,301 2015/16 5,362 2016/17 5,445 2017/18 5,530 2018/19 5,620 2019/20 5,711 2020/21 5,808 2021/22 5,905 2022/23 6,001 2023/24 6,099 2024/25 6,198 2025/26 6,299 Growth Rates 5 years 04/05 to 09/10 1.8% 11 years 04/05 to 15/16 1.4% 21years 04/05 to 25/26 1.5% 1.1% 1.1% 1.3% 1.7% 0.8% 1.1% -0.3% 0.5% 1.0% 1.4% 1.2% 1.4% 1.1% 1.1% 1.3% Notes: 1. Regional peaks include distribution losses but not transmission losses, unless otherwise stated. 2. The Domestic Peak is defined in the Glossary. 3. Lower Mainland peak includes peak sales to City of New Westminster and sales to Seattle City Lights. 4. Southern Interior peak includes sales to Fortis B.C. 5. Peak forecasts and growth rates are on a weather-adjusted basis. 6. The recorded domestic peak was 9,437 MW on January 13, 2005. The actual domestic peak value as stated above has been reduced to account for losses associated with peak sales to other utilities including Fortis B.C. and Seattle City Light. 7. Vancouver Island Peak with transmission losses is the Vancouver Island Regional peak before DSM adjusted for estimated transmission losses. Table A7.7. Weather-Adjusted Peak Forecast Before DSM Distribution WeatherAdjusted Peak MW Historical 2003/04 6,838 Transmission Domestic System Peak WeatherAdjusted Peak MW 9,431 Peak MW 1,760 PAGE 78 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates 5 years 04/05 to 09/10 11 years 04/05 to 15/16 21years 04/05 to 25/26 6,944 7,166 7,352 7,522 7,675 7,799 7,898 7,997 8,093 8,187 8,281 8,373 8,508 8,645 8,784 8,926 9,070 9,216 9,365 9,516 9,669 9,825 1,980 1,933 1,943 2,012 2,048 2,080 2,089 2,124 2,152 2,201 2,138 2,155 2,180 2,206 2,232 2,258 2,285 2,312 2,339 2,367 2,395 2,423 9,787 9,976 10,188 10,449 10,655 10,825 10,943 11,090 11,227 11,382 11,418 11,536 11,712 11,890 12,071 12,254 12,441 12,630 12,822 13,018 13,216 13,417 2.3% 1.7% 1.7% 1.0% 0.8% 1.0% 2.0% 1.5% 1.5% Notes: 1. Distribution peak forecast includes distribution losses. 2. Domestic Peak forecast includes distribution peak, transmission peak, peak sales to the City of New Westminster and losses. PAGE 79 Table A7.8. Weather-Adjusted Peak Forecast With DSM Distribution WeatherAdjusted Peak MW Historical 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates 5 years 04/05 to 09/10 11 years 04/05 to 15/16 21years 04/05 to 25/26 6,838 6,944 7,119 7,281 7,406 7,534 7,639 7,722 7,811 7,905 7,998 8,092 8,181 8,308 8,439 8,577 8,718 8,863 9,010 9,159 9,310 9,463 9,619 Transmission Domestic System Peak WeatherAdjusted Peak MW 9,431 9,787 9,882 10,018 10,203 10,358 10,494 10,582 10,702 10,828 10,982 11,024 11,149 11,316 11,488 11,668 11,850 12,057 12,287 12,479 12,675 12,873 13,074 Peak MW 1,760 1,980 1,894 1,857 1,901 1,916 1,935 1,932 1,953 1,973 2,022 1,966 1,990 2,016 2,042 2,068 2,095 2,138 2,202 2,229 2,257 2,285 2,313 1.9% 1.5% 1.6% -0.5% 0.0% 0.7% 1.4% 1.2% 1.4% Notes: 1. Distribution peak forecast includes distribution losses. 2. Domestic Peak forecast includes distribution peak, transmission peak, peak sales to the City of New Westminster and losses. PAGE 80 Table A7.9. 2005 BC Hydro, Reference Load Forecast Before DSM BC Hydro Service Area Sales Residential Commercial Industrial Total BCH Nwest Fortis BC (GW.h) 1,085 1,062 1,072 1,185 1,169 (GW.h) Actual 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates: 5 yrs 04/0509/10 11 yrs 04/0515/16 21 yrs 04/0524/25 14,573 15,090 15,287 15,899 15,620 (GW.h) 13,654 13,583 13,729 14,151 14,362 (GW.h) 18,579 17,739 18,596 18,725 19,635 (GW.h) 46,805 46,412 47,612 48,775 49,618 Total Domestic Sales (GW.h) 47,891 47,473 48,685 49,960 50,787 Firm Export (GW.h) 314 314 314 313 301 Total Firm Sales (GW.h) 48,204 47,787 48,999 50,273 51,088 Losses (GW.h) 4,774 4,780 4,299 4,778 4,659 Total Gross Requirement (GW.h) 52,978 52,567 53,298 55,051 55,747 Total System Peak (MW) 9,370 9,054 8,876 9,970 9,825 Integrated System Total Peak Gross Requirement (GW.h) (MW) 52,718 52,292 53,010 54,756 55,453 9,320 9,003 8,824 9,911 9,762 16,245 16,854 17,202 17,559 17,916 18,265 18,625 18,959 19,289 19,613 19,935 20,260 20,586 20,911 21,235 21,557 21,881 22,206 22,533 22,858 23,182 2.8% 2.2% 1.9% 14,859 15,310 15,717 16,092 16,493 16,862 17,215 17,542 17,873 18,211 18,550 18,906 19,224 19,460 19,926 20,278 20,633 20,980 21,315 21,656 22,008 2.8% 2.4% 2.1% 20,237 20,513 20,781 21,036 21,307 21,557 21,863 22,149 22,437 21,829 22,120 22,425 22,703 22,997 23,312 23,617 23,929 24,234 24,534 24,834 25,148 1.6% 1.1% 1.2% 51,341 52,677 53,700 54,687 55,716 56,684 57,703 58,650 59,599 59,653 60,605 61,591 62,513 63,368 64,473 65,452 66,443 67,420 68,382 69,348 70,338 2.3% 1.8% 1.7% 1,085 1,198 1,226 1,236 1,254 1,272 1,291 1,311 1,331 1,351 1,373 1,394 1,412 1,430 1,448 1,467 1,485 1,504 1,523 1,540 1,556 1.4% 1.5% 1.4% 52,426 53,875 54,926 55,923 56,970 57,956 58,994 59,961 60,930 61,004 61,978 62,985 63,925 64,798 65,921 66,919 67,928 68,924 69,905 70,888 71,894 2.3% 1.8% 1.7% 315 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 0.6% 0.3% 0.2% 52,741 54,186 55,239 56,234 57,281 58,267 59,307 60,272 61,241 61,315 62,291 63,296 64,236 65,109 66,234 67,230 68,239 69,235 70,218 71,199 72,205 2.3% 1.8% 1.7% 5,368 5,329 5,431 5,532 5,636 5,739 5,840 5,937 6,032 6,067 6,165 6,264 6,360 6,450 6,560 6,660 6,763 6,864 6,966 7,067 7,170 3.9% 2.6% 2.1% 58,109 59,515 60,670 61,766 62,917 64,006 65,147 66,209 67,273 67,382 68,456 69,560 70,596 71,559 72,794 73,890 75,002 76,099 77,184 78,266 79,375 2.4% 1.9% 1.7% 10,355 10,569 10,831 11,038 11,212 11,336 11,489 11,630 11,786 11,823 11,944 12,120 12,299 12,481 12,666 12,854 13,044 13,238 13,435 13,634 13,837 2.7% 1.8% 1.6% 57,810 59,210 60,362 61,454 62,599 63,682 64,816 65,873 66,932 67,035 68,103 69,203 70,233 71,190 72,420 73,511 74,618 75,710 76,790 77,867 78,972 2.5% 1.9% 1.7% 10,298 10,511 10,772 10,978 11,151 11,274 11,425 11,565 11,720 11,756 11,875 12,050 12,228 12,409 12,592 12,779 12,968 13,161 13,356 13,554 13,756 2.7% 1.8% 1.6% PAGE 81 Table A7.10. 2005 BC Hydro, Reference Load Forecast With DSM BC Hydro Service Area Sales Residential Commercial Industrial Total BCH Integrated System Total Peak Gross Requirement (GW.h) (MW) 52,718 52,292 53,010 54,756 55,453 9,320 9,003 8,824 9,911 9,762 Nwest Fortis BC (GW.h) 1,085 1,062 1,072 1,185 1,169 (GW.h) Actual 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates: 5 yrs 04/0509/10 11 yrs 04/0515/16 21 yrs 04/0524/25 14,573 15,090 15,287 15,899 15,620 (GW.h) 13,654 13,583 13,729 14,151 14,362 (GW.h) 18,579 17,739 18,596 18,725 19,635 (GW.h) 46,805 46,412 47,612 48,775 49,618 Total Domestic Sales (GW.h) 47,891 47,473 48,685 49,960 50,787 Firm Export (GW.h) 314 314 314 313 301 Total Firm Sales (GW.h) 48,204 47,787 48,999 50,273 51,088 Losses (GW.h) 4,774 4,780 4,299 4,778 4,659 Total Gross Requirement (GW.h) 52,978 52,567 53,298 55,051 55,747 Total System Peak (MW) 9,370 9,054 8,876 9,970 9,825 16,116 16,651 16,976 17,295 17,613 17,925 18,251 18,565 18,882 19,184 19,469 19,746 20,031 20,332 20,635 20,928 21,235 21,560 21,887 22,212 22,536 2.4% 2.0% 1.8% 14,786 15,197 15,497 15,796 16,143 16,482 16,827 17,167 17,498 17,835 18,171 18,527 18,846 19,081 19,547 19,900 20,255 20,602 20,936 21,278 21,629 2.4% 2.2% 2.0% 19,911 19,816 19,713 19,752 19,903 20,021 20,179 20,383 20,670 20,132 20,507 20,814 21,097 21,392 21,706 22,187 22,809 23,114 23,414 23,714 24,028 0.3% 0.4% 1.0% 50,813 51,664 52,186 52,843 53,659 54,428 55,257 56,115 57,050 57,151 58,147 59,087 59,974 60,805 61,888 63,015 64,299 65,276 66,237 67,204 68,193 1.6% 1.5% 1.5% 1,085 1,198 1,226 1,236 1,254 1,272 1,291 1,311 1,331 1,351 1,373 1,394 1,412 1,430 1,448 1,467 1,485 1,504 1,523 1,540 1,556 1.4% 1.5% 1.4% 51,898 52,862 53,412 54,079 54,913 55,700 56,548 57,426 58,381 58,502 59,520 60,481 61,386 62,235 63,336 64,482 65,784 66,780 67,760 68,744 69,749 1.6% 1.5% 1.5% 315 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 0.6% 0.3% 0.2% 52,213 53,173 53,725 54,390 55,224 56,011 56,861 57,737 58,692 58,813 59,833 60,792 61,697 62,546 63,649 64,793 66,095 67,091 68,073 69,055 70,060 1.6% 1.4% 1.5% 5,324 5,243 5,297 5,367 5,452 5,536 5,620 5,710 5,804 5,842 5,942 6,036 6,127 6,215 6,322 6,434 6,557 6,659 6,760 6,862 6,964 3.2% 2.2% 1.9% 57,537 58,416 59,022 59,757 60,676 61,547 62,481 63,447 64,496 64,655 65,775 66,828 67,824 68,761 69,971 71,227 72,652 73,750 74,833 75,917 77,024 1.7% 1.5% 1.6% 10,262 10,398 10,585 10,740 10,882 10,974 11,101 11,231 11,387 11,430 11,556 11,725 11,897 12,078 12,262 12,470 12,701 12,894 13,092 13,291 13,494 2.1% 1.5% 1.5% 57,238 58,111 58,713 59,445 60,358 61,223 62,151 63,111 64,155 64,308 65,423 66,470 67,461 68,393 69,597 70,848 72,268 73,361 74,439 75,518 76,621 1.7% 1.5% 1.6% 10,205 10,340 10,526 10,680 10,821 10,912 11,037 11,166 11,321 11,363 11,487 11,655 11,826 12,006 12,188 12,395 12,625 12,817 13,013 13,211 13,413 2.1% 1.5% 1.5% PAGE 82 Table A7.11. 2005 BC Hydro, High Load Forecast Before DSM BC Hydro Service Area Sales Residential Commercial Industrial Total BCH Integrated System Total Peak Gross Requirement (GW.h) (MW) 52,718 52,292 53,010 54,756 55,453 9,320 9,003 8,824 9,911 9,762 Nwest Fortis BC (GW.h) 1,085 1,062 1,072 1,185 1,169 (GW.h) Actual 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates: 5 yrs 04/0509/10 11 yrs 04/0515/16 21 yrs 04/0524/25 14,573 15,090 15,287 15,899 15,620 (GW.h) 13,654 13,583 13,729 14,151 14,362 (GW.h) 18,579 17,739 18,596 18,725 19,635 (GW.h) 46,805 46,412 47,612 48,775 49,618 Total Domestic Sales (GW.h) 47,891 47,473 48,685 49,960 50,787 Firm Export (GW.h) 314 314 314 313 301 Total Firm Sales (GW.h) 48,204 47,787 48,999 50,273 51,088 Losses (GW.h) 4,774 4,780 4,299 4,778 4,659 Total Gross Requirement (GW.h) 52,978 52,567 53,298 55,051 55,747 Total System Peak (MW) 9,370 9,054 8,876 9,970 9,825 16,813 17,402 17,770 18,142 18,533 18,896 19,300 19,674 20,038 20,407 20,735 21,108 21,456 21,819 22,153 22,534 22,895 23,256 23,625 24,022 24,340 3.5% 2.6% 2.1% 15,002 15,475 15,932 16,366 16,826 17,251 17,659 18,037 18,425 18,801 19,194 19,609 19,973 20,272 20,795 21,209 21,640 22,049 22,432 22,822 23,242 3.2% 2.7% 2.3% 20,411 20,661 20,987 21,326 21,664 22,003 22,397 22,774 23,152 22,637 23,023 23,468 23,859 24,276 24,717 25,147 25,599 26,055 26,474 26,916 27,405 2.0% 1.5% 1.6% 52,226 53,538 54,689 55,834 57,023 58,150 59,356 60,485 61,615 61,845 62,952 64,185 65,288 66,367 67,665 68,890 70,134 71,360 72,531 73,760 74,987 2.8% 2.2% 2.0% 1,085 1,198 1,226 1,236 1,254 1,272 1,291 1,311 1,331 1,351 1,373 1,394 1,412 1,430 1,448 1,467 1,485 1,504 1,523 1,540 1,556 1.4% 1.5% 1.4% 53,311 54,736 55,915 57,070 58,277 59,422 60,647 61,796 62,946 63,196 64,325 65,579 66,700 67,797 69,113 70,357 71,619 72,864 74,054 75,300 76,543 2.8% 2.2% 2.0% 315 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 0.6% 0.3% 0.2% 53,626 55,047 56,228 57,381 58,588 59,733 60,960 62,107 63,257 63,507 64,638 65,890 67,011 68,108 69,426 70,668 71,930 73,175 74,367 75,611 76,854 2.8% 2.2% 2.0% 5,275 5,421 5,535 5,651 5,771 5,888 6,007 6,121 6,234 6,286 6,398 6,520 6,632 6,743 6,870 6,994 7,120 7,244 7,366 7,492 7,615 4.4% 2.9% 2.4% 58,901 60,468 61,763 63,032 64,359 65,621 66,967 68,228 69,491 69,793 71,036 72,410 73,643 74,851 76,296 77,662 79,050 80,419 81,733 83,103 84,469 2.9% 2.2% 2.0% 10,496 10,738 11,026 11,264 11,469 11,622 11,810 11,984 12,175 12,246 12,394 12,616 12,830 13,055 13,275 13,510 13,748 13,990 14,226 14,476 14,725 3.1% 2.1% 1.9% 58,595 60,154 61,443 62,706 64,026 65,282 66,622 67,877 69,134 69,431 70,668 72,036 73,264 74,466 75,905 77,266 78,649 80,012 81,321 82,686 84,048 2.9% 2.2% 2.0% 10,438 10,678 10,965 11,202 11,405 11,557 11,743 11,917 12,106 12,176 12,322 12,543 12,756 12,980 13,198 13,432 13,668 13,909 14,144 14,393 14,640 3.2% 2.1% 1.9% PAGE 83 Table A7.12. 2005 BC Hydro, Low Load Forecast Before DSM BC Hydro Service Area Sales Residential Commercial Industrial Total BCH Nwest Total Fortis BC Domestic Sales (GW.h) (GW.h) 1,085 1,062 1,072 1,185 1,169 47,891 47,473 48,685 49,960 50,787 Firm Export (GW.h) 314 314 314 313 301 Total Firm Sales (GW.h) 48,204 47,787 48,999 50,273 51,088 Losses Total Gross Requirement (GW.h) 52,978 52,567 53,298 55,051 55,747 Total System Peak (MW) 9,370 9,054 8,876 9,970 9,825 Integrated System Total Peak Gross Requirement (GW.h) (MW) 52,718 52,292 53,010 54,756 55,453 9,320 9,003 8,824 9,911 9,762 (GW.h) Actual 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates: 5 yrs 04/0509/10 11 yrs 04/0515/16 21 yrs 04/0524/25 14,573 15,090 15,287 15,899 15,620 (GW.h) 13,654 13,583 13,729 14,151 14,362 (GW.h) 18,579 17,739 18,596 18,725 19,635 (GW.h) 46,805 46,412 47,612 48,775 49,618 (GW.h) 4,774 4,780 4,299 4,778 4,659 15,767 16,297 16,623 16,957 17,283 17,606 17,950 18,236 18,538 18,836 19,104 19,410 19,708 19,995 20,305 20,607 20,878 21,176 21,480 21,761 22,053 2.0% 1.8% 1.7% 14,809 15,147 15,501 15,817 16,159 16,476 16,779 17,061 17,354 17,637 17,933 18,231 18,510 18,689 19,112 19,428 19,713 20,032 20,301 20,591 20,908 2.4% 2.0% 1.8% 20,191 20,352 20,563 20,749 20,959 21,150 21,377 21,592 21,804 21,121 21,334 21,569 21,792 21,985 22,205 22,434 22,670 22,896 23,119 23,331 23,567 1.3% 0.8% 0.9% 50,767 51,796 52,687 53,523 54,401 55,232 56,106 56,889 57,696 57,594 58,371 59,210 60,010 60,669 61,622 62,469 63,261 64,104 64,900 65,683 66,528 1.9% 1.5% 1.4% 1,085 1,198 1,226 1,236 1,254 1,272 1,291 1,311 1,331 1,351 1,373 1,394 1,412 1,430 1,448 1,467 1,485 1,504 1,523 1,540 1,556 1.4% 1.5% 1.4% 51,852 52,994 53,913 54,759 55,655 56,504 57,397 58,200 59,027 58,945 59,744 60,604 61,422 62,099 63,070 63,936 64,746 65,608 66,423 67,223 68,084 1.8% 1.5% 1.4% 315 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 0.6% 0.3% 0.2% 52,167 53,305 54,226 55,070 55,966 56,815 57,710 58,511 59,338 59,256 60,057 60,915 61,733 62,410 63,383 64,247 65,057 65,919 66,736 67,534 68,395 1.8% 1.5% 1.4% 5,118 5,235 5,324 5,411 5,501 5,590 5,677 5,758 5,841 5,860 5,942 6,026 6,110 6,182 6,278 6,366 6,449 6,538 6,624 6,707 6,796 3.4% 2.2% 1.8% 57,285 58,540 59,550 60,481 61,467 62,405 63,387 64,269 65,179 65,116 65,999 66,941 67,843 68,592 69,661 70,613 71,506 72,457 73,360 74,241 75,191 2.0% 1.5% 1.4% 10,208 10,396 10,631 10,808 10,954 11,053 11,178 11,289 11,419 11,426 11,515 11,663 11,819 11,964 12,121 12,284 12,436 12,605 12,769 12,933 13,108 2.2% 1.5% 1.4% 56,998 58,249 59,257 60,187 61,167 62,099 63,076 63,953 64,857 64,789 65,667 66,604 67,501 68,245 69,309 70,256 71,144 72,091 72,989 73,866 74,811 2.0% 1.5% 1.4% 10,153 10,340 10,575 10,752 10,896 10,994 11,118 11,228 11,357 11,362 11,450 11,598 11,752 11,896 12,051 12,213 12,364 12,532 12,695 12,857 13,031 2.2% 1.5% 1.4% PAGE 84 Table A7.13. 2005 BC Hydro, High Load Forecast With DSM BC Hydro Service Area Sales Residential Commercial Industrial Total BCH (GW.h) Actual 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates: 5 yrs 04/0509/10 11 yrs 04/0515/16 21 yrs 04/0524/25 14,573 15,090 15,287 15,899 15,620 (GW.h) 13,654 13,583 13,729 14,151 14,362 (GW.h) 18,579 17,739 18,596 18,725 19,635 (GW.h) 46,805 46,412 47,612 48,775 49,618 Nwest Total Fortis BC Domestic Sales (GW.h) (GW.h) 1,085 1,062 1,072 1,185 1,169 47,891 47,473 48,685 49,960 50,787 Firm Export (GW.h) 314 314 314 313 301 Total Firm Sales (GW.h) 48,204 47,787 48,999 50,273 51,088 Losses (GW.h) 4,774 4,780 4,299 4,778 4,659 Total Gross Requirement (GW.h) 52,978 52,567 53,298 55,051 55,747 Total System Peak (MW) 9,370 9,054 8,876 9,970 9,825 Integrated System Total Peak Gross Requirement (GW.h) (MW) 52,718 52,292 53,010 54,756 55,453 9,320 9,003 8,824 9,911 9,762 16,685 17,199 17,542 17,871 18,225 18,548 18,923 19,273 19,617 19,973 20,263 20,592 20,887 21,228 21,546 21,907 22,243 22,610 22,975 23,380 23,704 3.1% 2.4% 2.0% 14,930 15,365 15,706 16,067 16,473 16,859 17,270 17,654 18,040 18,410 18,801 19,224 19,575 19,882 20,401 20,831 21,259 21,665 22,044 22,440 22,851 2.8% 2.5% 2.2% 20,088 19,980 19,932 20,060 20,281 20,481 20,734 21,030 21,417 20,974 21,446 21,875 22,275 22,702 23,129 23,751 24,491 24,940 25,358 25,808 26,292 0.6% 0.8% 1.4% 51,703 52,544 53,180 53,998 54,979 55,888 56,927 57,957 59,074 59,357 60,510 61,691 62,737 63,812 65,076 66,489 67,993 69,215 70,377 71,628 72,847 2.1% 1.8% 1.8% 1,085 1,198 1,226 1,236 1,254 1,272 1,291 1,311 1,331 1,351 1,373 1,394 1,412 1,430 1,448 1,467 1,485 1,504 1,523 1,540 1,556 1.4% 1.5% 1.4% 52,788 53,742 54,406 55,234 56,233 57,160 58,218 59,268 60,405 60,708 61,883 63,085 64,149 65,242 66,524 67,956 69,478 70,719 71,900 73,168 74,403 2.1% 1.8% 1.8% 315 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 0.6% 0.3% 0.2% 53,103 54,053 54,719 55,545 56,544 57,471 58,531 59,579 60,716 61,019 62,196 63,396 64,460 65,553 66,837 68,267 69,789 71,030 72,213 73,479 74,714 2.1% 1.8% 1.8% 5,227 5,332 5,401 5,488 5,588 5,686 5,791 5,897 6,008 6,062 6,177 6,293 6,399 6,509 6,633 6,771 6,917 7,041 7,161 7,290 7,413 3.7% 2.6% 2.2% 58,330 59,385 60,120 61,033 62,132 63,157 64,322 65,476 66,724 67,081 68,373 69,689 70,859 72,062 73,470 75,038 76,706 78,071 79,374 80,769 82,127 2.2% 1.9% 1.9% 10,403 10,570 10,782 10,969 11,143 11,261 11,428 11,590 11,780 11,859 12,012 12,227 12,429 12,658 12,875 13,137 13,410 13,650 13,886 14,140 14,388 2.6% 1.8% 1.8% 58,024 59,071 59,800 60,707 61,799 62,818 63,977 65,125 66,367 66,719 68,005 69,315 70,480 71,677 73,079 74,642 76,305 77,664 78,962 80,352 81,706 2.2% 1.9% 1.9% 10,345 10,511 10,721 10,907 11,079 11,196 11,361 11,522 11,711 11,789 11,940 12,154 12,355 12,583 12,798 13,059 13,330 13,569 13,804 14,057 14,303 2.6% 1.8% 1.8% PAGE 85 Table A7.14. 2005 BC Hydro, Low Load Forecast With DSM BC Hydro Service Area Sales Residential Commercial Industrial Total BCH Nwest Fortis BC (GW.h) 1,085 1,062 1,072 1,185 1,169 Total Domestic Sales (GW.h) 47,891 47,473 48,685 49,960 50,787 Firm Export (GW.h) 314 314 314 313 301 Total Firm Sales (GW.h) 48,204 47,787 48,999 50,273 51,088 Losses Total Gross Requirement (GW.h) 52,978 52,567 53,298 55,051 55,747 Total System Peak (MW) 9,370 9,054 8,876 9,970 9,825 Integrated System Total Peak Gross Requirement (GW.h) (MW) 52,718 52,292 53,010 54,756 55,453 9,320 9,003 8,824 9,911 9,762 (GW.h) Actual 2000/01 2001/02 2002/03 2003/04 2004/05 Forecast 2005/06 2006/07 2007/08 2008/09 2009/10 2010/11 2011/12 2012/13 2013/14 2014/15 2015/16 2016/17 2017/18 2018/19 2019/20 2020/21 2021/22 2022/23 2023/24 2024/25 2025/26 Growth Rates: 5 yrs 04/0509/10 11 yrs 04/0515/16 21 yrs 04/0524/25 14,573 15,090 15,287 15,899 15,620 (GW.h) 13,654 13,583 13,729 14,151 14,362 (GW.h) 18,579 17,739 18,596 18,725 19,635 (GW.h) 46,805 46,412 47,612 48,775 49,618 (GW.h) 4,774 4,780 4,299 4,778 4,659 15,638 16,095 16,403 16,694 16,989 17,269 17,590 17,854 18,134 18,414 18,646 18,922 19,155 19,425 19,716 19,986 20,230 20,533 20,840 21,131 21,411 1.7% 1.6% 1.5% 14,740 15,039 15,284 15,527 15,813 16,094 16,400 16,694 16,983 17,269 17,566 17,875 18,141 18,327 18,748 19,069 19,358 19,672 19,944 20,233 20,546 1.9% 1.8% 1.7% 19,865 19,647 19,477 19,443 19,536 19,592 19,678 19,793 20,011 19,402 19,699 19,939 20,153 20,356 20,579 20,985 21,537 21,758 21,977 22,197 22,422 -0.1% 0.0% 0.6% 50,243 50,781 51,164 51,664 52,338 52,955 53,668 54,341 55,128 55,085 55,911 56,736 57,449 58,108 59,043 60,040 61,125 61,963 62,761 63,561 64,379 1.1% 1.1% 1.2% 1,085 1,198 1,226 1,236 1,254 1,272 1,291 1,311 1,331 1,351 1,373 1,394 1,412 1,430 1,448 1,467 1,485 1,504 1,523 1,540 1,556 1.4% 1.5% 1.4% 51,328 51,979 52,390 52,900 53,592 54,227 54,959 55,652 56,459 56,436 57,284 58,130 58,861 59,538 60,491 61,507 62,610 63,467 64,284 65,101 65,935 1.1% 1.1% 1.3% 315 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 311 313 311 311 0.6% 0.3% 0.2% 51,643 52,290 52,703 53,211 53,903 54,538 55,272 55,963 56,770 56,747 57,597 58,441 59,172 59,849 60,804 61,818 62,921 63,778 64,597 65,412 66,246 1.1% 1.1% 1.2% 5,072 5,149 5,189 5,244 5,315 5,384 5,458 5,529 5,609 5,632 5,716 5,799 5,873 5,944 6,038 6,138 6,243 6,331 6,417 6,503 6,589 2.7% 1.9% 1.7% 56,715 57,439 57,892 58,455 59,218 59,922 60,730 61,492 62,379 62,379 63,313 64,240 65,045 65,793 66,842 67,956 69,164 70,109 71,014 71,915 72,835 1.2% 1.2% 1.3% 10,115 10,224 10,382 10,506 10,621 10,685 10,790 10,885 11,013 11,028 11,123 11,271 11,410 11,557 11,713 11,897 12,091 12,258 12,424 12,590 12,760 1.6% 1.1% 1.3% 56,428 57,148 57,599 58,161 58,918 59,616 60,419 61,176 62,057 62,052 62,981 63,903 64,703 65,446 66,490 67,599 68,802 69,743 70,643 71,540 72,455 1.2% 1.2% 1.3% 10,061 10,169 10,326 10,449 10,563 10,626 10,729 10,824 10,951 10,964 11,058 11,205 11,343 11,489 11,644 11,827 12,020 12,185 12,349 12,515 12,684 1.6% 1.1% 1.3% PAGE 86