The URL can be used to link to this page

Home My WebLink AboutNome Electrical Load Forecast, October 1990October 26, 1990 Nome Electrical Load Forecast Prepared for: Arctic Slope Consulting Group P.O. Box 650 Barrow, AK 99723 Prepared by: Alan Mitchell Analysis North 911 West 8th Avenue, Suite 204 Anchorage, AK 99501 and Steve Colt EMI Consulting 1408 P Street #A Anchorage, AK 99501 3.2 Sa3 3.4 3.5 3.6 3.7 3.8 3.9 Nome Electrical Load Forecast CONTENTS Geer cero lbrerslor Vg ot co Go Go oO oO OG Glo oO oo oo Oo Go 8) Load Forecasting Methods . ...........4++ +. 3 3.2.1 Residential Model . . <2. 2. ss se es 552s 3 3.2.2 Commercial/Other Model ..........-.-.-3 3.2.3 Gold Mining Model. .....+4++¢e¢4e¢+2 ec ee o FD 3.2.4 Forecasting Uncertainty ... oe ee ew ew ew e 8 3.2.5 Losses and Peak Generation Requirements ec «see 3 Employment and Population Projections ........3 3.3.1 Background . « * e ° =—s— +e 3.3.2 Historical Employment and Population Data eon 3.3.3 Economic Projection Methodology and Assumptions 3 - 3.3.4 Economic Projections Results ........ 3 Residential Load Forecast ......++-++-e+-+e- 3 3.4.1 People per Customer Trend ........+.. 3 3.4.2 Use Per Customer Trend ......+.. +++. 3 3.4.3 Summary of Residential Forecast ....... 3 oo ercial/Other Forecast ....:.++-+-+e4e-. 3.5.1 Use per Employee ..... +++ +++ eee Non-Mining Summary . .-....--.-. +++ +2. .- 3 Mining Forecast .....-. + 2+ 2. + 2+ 2 e+ 2 2+ ee - 3 3.7.1 General Background ....... ScmDmonne! 3.7.2 Recent Mine Operators and Potential Future Operators ...... ee - - 3 3.7.3 Low, Mid, and High Mining Assumptions, and Forecast Summary . . S . : . 73 3.7.4 Likelihood of NJUS supplying Mining Loads ~—«_- Combined Peak Generation Requirement for Non-Mining and Mining Loads... .- +--+ + 2 2 2 2 2 © ee ee ee 3 Forecast Summary .. .. + + © «© «© © © © © © © © ee 3 3.10 References .....- + 2 © © © © © © © © © © we © es 3 OPRWNPP PB oMmnunM r= 31 32 so 34 40 42 48 49 59 LIST OF TABLES Table 3.1 - Historical Employment and Population Data: City OLSNOMEC marropers erento ore mete ns romeo Mote re rele Mit oUtt otitis mmrodtt olf ott eae, Table 3.2 - Statewide Economic Assumptions by Case.. ...3-9 Table 3.3 - State Spending Assumptions.. . att ae ee Table 3.4 Projected Employment and Population. ee ae Table 3.5 Adjustments to NJUS Historical Data oie Sea 4 Table 3.6 End Use Forecast Results for Fairbanks . . 3-24 Table 3.7 Residential Forecast Assumptions by Case .. 3 - 25 Table 3.8 Assumptions for Commercial/Other Forecast. . 3 - 31 Table 3.9 Assumptions for the Mining Forecast by Case . 3 - 42 Table 3.10 - Relative Monthly Peak Demands ....... 3 - 49 Table 3.11 - Growth Rates of Total Net Generation Requirements and Peak Generation Requirements. ws = oe Table 3.12 - Mid Case Forecast Results, Non-Mining Loads 3 - 53 Table 3.13 - Mid Case Forecast Results, Mining Loads and Totals .... Semon oueiioteemireltetoiiite Holi ttelilt elineme ellie mia) Table 3.14 - Low case Forecast Results, Non-Mining Loads 3 - 55 Table 3.15 - Low Case Forecast Results, Mining Loads and Totals .. ° =e °° atten tage et aieapsynnnnner< meso -) Table 3.16 - High Case Forecast Results, Non-Mining Loads 3-— 57 Table 3.17 - High Case Forecast Results, Mining Loads and MOEA Sires tome otto omer re iioiee call oto lile lieth ollie elit ice ate tS =a LIST OF FIGURES Figure 3.1 - Indoor Employment Projections ....... 3-11 Figure 3.2 - Nome Population Forecast .......... 3-12 Figure 3.3 - People per Residential Customer . . et HS ed Figure 3.4 Use per Residential Customer Mio iret at eM eer Figure 3.5 Income per Nome Residential Customer .... 3 - 16 Figure 3.6 Nome Residential Electric Rates ...... 3-17 Figure 3.7 Residential Energy Sales Forecast ..... 3 - 26 Figure 3.8 Use per Employee ie om cuiemiciiei-sieiieiion tes ae t= — rai, Figure 3.9 Commercial/Other kWh Sales Forecast. .... 3 - 31 Figure 3.10 - Non-Mining Net Generation Requirements. . . 3 - 32 Figure 3.11 - Historical Load Factor for Nome Non-Mining LLORAS a scmromeomeclinel nemsommollreiiieiliiameclire reli iien Nels elie oir or oHitS ean SS Figure 3.12 - Historical Gold Production in the Nome District ° Riel iret oHiTomeHi otiy sii! simntotalis Hil nS aman 3 4 Figure 3.13 - Us “wholesale Gold Seine ee eos > Figure 3.14 - Gold Operators in the Nome District, 1989 sn 0) Figure 3.15 - Gold Mining Production Forecast ...... 3 - 43 Figure 3.16 - Mining Net Generation Forecast ...... 3 - 44 Figure 3.17 - Total Net Generation Requirements, Non-Mining and Mining Loads ei oitene eo ttelertottone . aaa oe Figure 3.18 - Total Peak Generation Requirements, Non-Mining and Mining. .....- .- ee Hore ioMite ite Mite elltitemeatoiairroitt tiiies mts ks Figure 3.19 - Total Net Generation Requirements, Mid Case 3 - 52 3.0 NOME ELECTRICAL LOAD FORECAST 3.1 Introduction This report presents a 25-year forecast (1990-2014) of electric loads in the vicinity of Nome, Alaska. Both the total annual electrical energy use (kilowatt-hours) and the annual peak demand (megawatts) were forecast over the analysis period. Because of the uncertainty present in any forecasting effort, a Low, Mid, and High forecast are provided. These forecasts were selected so that there is approximately an 80% probability that the actual load in future years will fall between the Low and High forecasts. 3.2 Load Forecasting Methods We utilized relatively simple models to forecast the electrical load in Nome. We believe that the level of sophistication present in the models matches the type and quality of historical data available and is appropriate given the level of unavoidable uncertainty present in a long-term forecasting effort. Electrical use in the Nome area was forecast separately for the following 3 customer classes: + Residential + Commercial, Community Facilities, and Street Lighting + Mining 3.2.21 Residential Model The model used for residential electrical use was: Residential Use = (# of customers) * (use per customer) and (# of customers) was modeled as: # of Customers = Population / (people per customer) Using this model, three variables need to be forecast: population, people per customer, and use per customer. We utilized the Rural Alaska Model (RAM), a regional economic and demographic model developed at the University of Alaska, Institute of Social and Economic Research (ISER), to forecast population for the Nome area (Knapp, 1990). "People per customer" is essentially a measure of household size. We forecast this household size variable by examining statewide forecasts developed by ISER (Goldsmith, 1990). Finally, we forecast use per customer by examining historical trends in this variable for Nome, estimating the impacts from probable changes in Nome electric use patterns, and examining other statewide and national forecasts that have relevance to the Nome situation. 3.2.2 Commercial/Other Model All non-residential use except for mining electric consumption was grouped together in one category. Commercial use, community facility use, and street lighting are the main uses in this category. The model used for this type of electric use was: Commercial/Other Use = (Indoor Employment) * (Use/Employee) Indoor employment is a measure of the number of employees working in-buildings and was calculated as total employment minus mining employment and minus a portion of construction employment. The indoor employment forecast is provided by the RAM model referred to above. Use per employee is forecast by examining historical Nome data and investigating other forecasts concerning the energy intensity of commercial activities. Some electricity uses in the commercial/other category, such as street lighting, are not closely related to the number of indoor employees. Street lighting use is more closely related to the total population of Nome. However, the total use associated with street lighting is not large. We did not complicate the forecasting effort by separating out such uses. In addition, indoor employment is closely correlated with population, so forecasting street lighting use from indoor employment will not result in significant errors. 3.2.3 Gold Mining Model The final category of electric use is gold mining use. The electrical use of gold mining operations is a large fraction of the total electrical requirements in the Nome area. Further, total mine electrical use is dependent on factors different from factors determining the level of other electrical uses. Thus, a separate forecast of mining electrical use was prepared. Gold mining was divided into three different types of mining operations: onshore mining that occurs during the summer only, onshore mining that occurs year-round, and offshore mining. This division was made because the seasonal distribution of power requirements and the total power requirements vary substantially among these different types of operations. For offshore mining, we only considered the wintertime (off- season) power requirements. These are the power requirements that occur while the dredges are docked at the Nome causeway between mining seasons. For the onshore mining operations, we only included operations and potential operations that would have economical access to the Nome Joint Utilities System (NJUS) transmission and distribution grid. The electrical use of each different type of mining was forecast using the following model structure: Mining Use = (Ounces of Gold Mined) * (Electric Use per Ounce) For each type of use, we forecast the amount of gold mined by consulting with state mining experts and interviewing mine operators and potential operators. To forecast electric use per ounce of gold mined for the different types of operations, we gathered such data from existing operations and utilized mine operator estimates for potential future operations. 3.2.4 Forecasting Uncertainty For each type of use, three scenarios were projected--a mid case, a low case, and a high case. The mid case was chosen such that there is a 50% chance that the actual load will be higher than the mid case and a 50% chance that it will be lower. The low case represents a combination of forecast assumptions that result in a low growth rate in electric use, and the high case represents assumptions that result in rate of growth higher than the mid case. We attempted to chose the assumptions for the low and high cases such that the likelihood of the actual electrical load being somewhere between those cases is roughly 80%, i.e. there is a 10% chance that the actual electrical load will exceed our high scenario projection and a 10% chance that the actual load will be less than our low scenario projection. However, these probability figures themselves are substantially uncertain. The range of assumptions chosen for an individual variable may seem small. For example, there is probably more than a 10% probability that gold production will exceed our high case estimate. However, the High gold assumption was combined with the High population growth assumption and the High use per customer assumption to determine the High load growth projection. Because of the unlikelihood of all of these variables assuming high values simultaneously, we restrained See the high estimates for the individual variables. Likewise, the low estimates were restrained. 3.2.5 Losses and Peak Generation Requirements Since the objective of the overall study is to evaluate alternative generation options, the consumption figures were converted to a generation requirement by adding the transmission and distribution losses associated with serving the loads with off-site generation. In section 3.7.4 we discuss the likelihood of various loads being served from the NJUS generation systen. As well as forecasting annual generation requirements in kilowatt-hours, we also forecast the peak generation requirements in megawatts based on typical load factors for each load. Since mining loads do not necessarily peak at the same time as non-mining loads, the combined peak generation requirement of the loads is less than the simple addition of their individual peaks. We account for this non-coincidence in calculating a Nome area peak electrical demand. Employment and Population Projections 3.3.1 Background The city of Nome is the regional hub of about 15 Inupiaq villages surrounding Norton Sound. Nome was founded in the 1898 gold rush and briefly enjoyed the status of Alaska’s largest community at the turn of the century (Waring 1988, p. i). Since that time, its cash economy has been driven by military spending, continued mining, and, especially during the 1980s, public spending for government services, public works construction, education and health care. 3.3.2 Historical Employment and Population Data Table 3.1 shows historical average annual employment by sector for the City of Nome during the 1980s. These data are from the Alaska Department of Labor’s ES-202 reports, with two substantial adjustments* made to better represent employment actually occurring within the city. Total "Adjusted DOL Employment" grew at an average annual rate of 3.3 percent between 1980 and 1989. As Knapp (1990) points out, ES-202 data undercounts some actual employment because it does not include proprietors or workers working in Nome for firms headquartered in other cities. On the other hand, the data erroneously includes workers working outside of Nome for firms headquartered in Nome. Even with these limitations, the ES- 202 data are excellent for reviewing trends in employment. Because they are based on monthly reporting, they accurately reflect the highly seasonal nature of many rural Alaskan jobs. An alternative employment data set, generated by a 1986 survey of employers, exists for the period 1980-1986 (Impact Assessment 1987). These data are somewhat more detailed and are theoretically free of the defects noted above. However, they are biased downward for earlier years because it is likely that there are several firms which were in business during the early part of the decade but no longer in business when the survey was conducted. In order to combine the best attributes of each data source, we further adjusted the entire DOL data series so that total "benchmarked" DOL non-mining employment was equal to Impact Assessment measured non-mining employment in the year 1986. ‘These are: (1) Removing estimated Bering Straits School District employment since the district is actually located in Unalakleet; (2) Removing one-time CETA program summer employment from 1980 and 1981 services employment. 3-6 Table 3.1 - Historical Employment and Population Data: City of Nome. 1985 1986 1987 EMPLOYMENT (AK Dept of Labor) Mining (1) 57 65—s«143 Construction 41 31 19 Trans/Comm/Util 4 77 105129 Trade 202 «221 = 208 F.LR.E. 61 40 31 0 Services (2) 471 455) 464 449 Fed Government 98 93 89 90 State Government wo 26 «62 211 Local Government 604 8598 86580 628 Total DOL Employment 1,858 1,847 1,853 2,045 Less: Bering Sts School Dist 429 372400 400 Less: CETA workers Adjusted DOL Employment 1,453 1,645 Less: Mining 143-269 Less: 1/2 Construction 9 14 Less: Benchmarking to LAI (3) 97 Benchmarked Indoor Employment ; Ff 1,204 POPULATION City of Nome AK Dept. of Labor (4) Sources: Minerals Mgmt Service 1990 (employment): Waring 1989 (population); Personal Communication Judy Hallanger, ADOL, 10/15/90 (population 1987-89). Notes: (1) Mining is partially suppressed in years 1980-86. Estimated during those years as a residual, using total and sum of all other industries. (2) When CETA service workers (1980-1981) are removed, average growth rate is 2.6%. (3) Data collected in a survey by Impact Assessment, Inc. (1987) are believed by Knapp (1990) to be best single year snapshot of employment and are used in the RAM forecasting model. Benchmarking adjustment forces DOL data series to be equal to LAI series in 1986. (4) Figure for 1980 is US Census, believed to be a significant undercount. Average growth rate for 1981-89 is 1.8%. Further removal of one half of the construction employment yields a data series labeled Benchmarked Indoor Employment which is most useful for looking at trends in electricity use per employee. A single definitive data series for population is not available. There are two population series available, from the City of Nome and from the State Demographer through the Department of Labor. These two series are also shown in ees Table 3.1. The two series are in agreement for the years 1981 and 1982. For the period 1983-1986, they diverge substantially, with the city estimates growing more rapidly. Waring (1989) discusses the probable accuracy of these two series at length and concludes: We conclude that the City’s post-1982 population estimates overstate the City’s true population. The City’s 1982-1985 population estimates were based on an annual count of housing units multiplied by the vacancy rates and average household size that prevailed at the time of the City’s 1981 population count. We believe this method is prone to yield increasingly inflated population estimates under the housing market conditions that prevailed at Nome [in 1981]. Specifically, in the following four years, a residential construction boom enlarged the housing stock by 331 units (34 percent). Under these changing market conditions, we believe it is unrealistic to hold vacancy rates and average household sizes fixed at 1981 levels, as the city did in its estimation technology... All things considered, we are persuaded to accept the Alaska Department oof Labor’s 1983-1986 population estimates over the City of Nome’s official estimates as more consistent with other available population indicators. (page 55) Based on this logic, we agree with Waring that the DOL population series is more plausible for the period 1981-89. We have therefore projected population using a 1989 starting point consistent with DOL data. It is important to remember that either series could be used as a basis for projection, since it is the relative growth in population that is relevant to the electric load forecast. 3.3.3 Economic Projection Methodology and Assumptions We used the Rural Alaska Model (RAM) developed by Knapp (1990) and adjusted to reflect 1988 and 1989 data in order to project ce GC City of Nome employment and population under various sets of driving assumptions. The RAM model relies on assumptions about the following driving variables: + future basic sector employment + future levels of state spending (operating and capital) * future growth in "exogenous" support employment (serving tourist and out-of-region demand) Future basic sector employment consists almost entirely of offshore and onshore mining. Scenarios for Low, Mid, and High case mining employment levels are discussed in section 3.7.3 of this report. Other basic employment (about three full time equivalent fishermen) is assumed to remain constant. Table 3.2 - Statewide Economic Assumptions by Case. Source: Goldsmith (1990). [_ assumption _| Oil Prices in 1989 $ 1990 $18 $18 $18 2000 $14 $19 $26 | $14 $21 $35 Tourism 3% per year 3% per year 5% per year | Growth West Sak, | West Sak, West Sak, 1990’s Production after 2000 1990’s ANWR, after | 2000 | | | Petroleum | See fe ee oe Three scenarios for future levels of state spending have been taken from recent econometric projections prepared by the Institute of Social and Economic Research using the MAP model (Goldsmith 1990). Table 3.2 shows the key assumptions used by Goldsmith to develop these projections. More detailed assumptions are contained in Goldsmith (1990). sa=e9 Table 3.3 - State Spending Assumptions. Results of MAP Model Projections. Source: Goldsmith (1990). Average Annual Growth Rate, 1990-2010 oe 1989 a erakias Expenditures Capital | Expenditures Table 3.3 shows the projected annual growth rates in state operating and capital spending which result from the MAP model runs. These are used as inputs to the RAM model to place realistic limits on the number of future state-supported jobs in Nome. Future growth in exogenous support employment is assumed to be 2.0 percent annually. This is Knapp’s best guess for the Nome region and is consistent with the 3.0 percent growth rate in tourist visits assumed by Goldsmith in the MAP statewide model projections. Because it is highly unlikely that low mining employment, low state spending, and low tourism growth will all occur at the same time, we have chosen not to vary this assumption across cases. 3.3.4 Economic Projections Results Table 3.4 summarizes the results of the RAM model runs. Population grows at between 1.3 and 2.3 percent annually, while indoor employment grows at between 0.7 and 1.6 percent.’ 2These growth rates do not exactly correspond to the growth rates in the final forecast summary tables. This is because the electric forecast extends to the year 2014 while RAM only projects 3 LO Table 3.4 -— Projected Employment and Population. Source: RAM Model. Employees 500 0 1981 1986 1991 1996 Year 2001 2006 2011 —® Historical —— Mid — Low —= High Figure 3.1 - Indoor Employment Projections Population grows faster than employment due to the effects of natural increase. Natural increase in population is not dependent on employment growth. Figure 3.1 and Figure 3.2 to the year 2010. We extended the RAM results to 2014 for use in the electric forecast by using the 2000-2010 growth rate. Sear graphically depict the indoor employment and population projections. Nome Population 7,000 6,000 5,000 4,000 People 3,000 2,000 1,000 g 981 1986 1991 1996 2001 2006 2011 Year —®— Historical —— Mid —— Low —=- High Figure 3.2 - Nome Population Forecast 3.4 Residential Load Forecast As stated in section 3.2.1 our model requires that three variables be forecast to project residential sales: population, people per residential customer, and use per residential customer. The population forecast is an output of the economic/demographic forecast discussed in previous section. The forecast of the remaining two variables will be addressed in this section. 3.4.1 People per Customer Trend Figure 3.3 shows the history of the people per customer variable, a measure of household size, for the 1981 through 3 = 12 1989 period, based on Alaska Department of Labor population data, and NJUS residential customer counts adjusted for misclassified customers. Prior to 1988, a number of apartment units were classified by NJUS as commercial customers. An estimate of these misclassified customers was made, and pre- 1988 customer counts were adjusted. Table 3.5 shows these adjustments, the associated kWh sales adjustments, and the separation of mining electrical use from the NJUS sales data (relevant in subsequent sections of the report). People per Residential Customer Nome 3.5 3 2.5 2 -3.1% / year 1.5 People/Customer 0.5 1981 1982 1983 1984 1985 1986 1987 1988 1989 Year Figure 3.3 — People per Residential Customer. Population data from Alaska Department of Labor. Residential customer data from NJUS, adjusted for misclassified residential customers. The graph shows that the people per customer variable drops sharply between 1981 and 1983, and then remains nearly constant for the remaining years of the historical period. A housing boom occurred in Nome during the early 80s. The housing stock increased faster than population, resulting in 3 - 13 Table 3.5 - Adjustments to NJUS Historical Data. Residential customers erroneously classified as commercial are added back to residential customer counts and sales. Also, mining sales are separated from commercial sales. ss sss ts 198s 1585 sos 1067 1088 1580] a) Total Sales, MWh 14,629 16,148 17,534 18,515 19,024 20,379 20,785 21,915 24,740 b) Residential Class Sales, MWh c) Residential Customers d) Use/Resid. Cust., kWh, b * 1000/c 4,530 5,327 6,004 6,354 6,257 6,492 6,995 8,084 7,934 829-953 -:1,119 1,164 1,160 1,151 1,197 1,344 1,365 5,464 5,590 5,366 5,459 5,394 5,640 5,844 6,015 5,812 e) Resid. Cust. classified as Commercial] 93 107 125 130 130 129 134 0 0 (f) Sales to Misclass. Cust., MWh, d xe 507.596 672711 700 727_—S— 783. 0 g) Adj. Residential Sales, MWh,b +f | 5,037 5,923 6,676 7,065 6,957 7,219 7,778 922 1,060 1,244 1,294 1,290 1,280 1,331 1,344 1,365 5,464 5,590 5,366 5,459 5,394 5,640 5,844 5,812 h) Adj. Residential Customers, c + € Ki) Adj. Use/Resid. Cust., kWh, g * 1000 j) Mining Company Sales, MWh 228 228 228 228 215 192 304 1,593 3,091 k) Net Commercial, Commun. Facil., 9,364 9,997 10,630 11,222 11,851 12,969 12,703 12,238 13,715 & Street Lighting Sales, MWh, a - g -j a reduction in the average household size during this period. The stability shown in the 1983 through 1989 period is consistent with more general evidence concerning long-term household size trends. The variable does change over time but in a slow manner. For our forecast of this variable, we rely on the forecast of Alaskan household size used in the ISER MAP forecast (Goldsmith, 1990). In this forecast, household size was projected to decline 0.25%/year through the year 2010. Because of its minor effect on load relative to other variables, the household size trend is assumed to be the same in all three residential forecasts--mid, low, and high. 3.4.2 Use Per Customer Trend Figure 3.4 shows the history of residential use per customer for Nome, Alaska, and the US. Use per customer in Nome has trended upward at a rate of 0.8% per year for the period 1981 through 1989. Sar Use per Residential Customer 10,000 9,000 8,000 7,000 6,000 5,000 4,000 3,000 1.1% / year -2.4% / year bt LL eae Te 0.8% / year kWh/year or) | toe fee ee . pee) ee T 1981 1982 1983 1984 1985 1986 1987 1988 1989 Year —=— Nome —— US —— Alaska Figure 3.4 - Use per Residential Customer. Sources: Nome - adjusted NJUS data (adjustments shown in Table 3.5); Alaska - Penny Haldane, Alaska Energy Authority; US - US Statistical Abstract, 1989. Use per customer can be thought of as the amount of electric services used by a customer multiplied by the energy intensity of providing those services. An example of an "electrical service" is food refrigeration or residential lighting. The “energy intensity" of providing a service indicates the amount of electricity required per unit of service. A change in use per customer can be explained by changes in the amount or energy intensity of electrical services used by customers. The amount of electrical services used by residential customers is affected by a number of factors. The household income affects the amount of services that can be afforded by the customer. Figure 3.5 shows how per customer income varied throughout the 1981-89 period. Per capita income rises by 3— 5 Income per Nome Residential Customer Inflation-Adjusted 70 60 = 50 Ss -2.2% / year 5 5 ‘s. oO 3 & ® 20 10 — ee T <9 ep eager 5 x 1981 1982 1983 1984 1985 1986 1987 1988 1989 Year Figure 3.5 - Income per Nome Residential Customer. Sources: Income per capita - Bureau of Econ. Analysis, adjusted for differences in Census Area and Nome City incomes; Population - AK Dept. of Labor; Customer count - NJUS. about 0.9% per year throughout the period; however, the figure shows that household income declines initially because of decreasing household size. The average change in per customer income over the entire period is a decline of 2.2% per year. Such a reduction in household income would exert a downward influence on the amount of electrical services used over the period. Another factor affecting the amount of electrical services used is the price of electricity. The more expensive electricity is, the less electrical services will be used. Figure 3.6 shows the history of residential electrical rates in Nome. The rates shown are adjusted for inflation. Two data series are given; one represents the "full" residential 3 -=—16 Nome Residential Electric Rates Inflation-Adjusted 30 25 6.7% / year # a S 20 £15 = 2 40 -8.5% / year S Oo 5 —— T aay T rape 2 1981 1982 1983 1984 1985 1986 1987 1988 1989 Year —#= Total Rate —— Subsidized Rate Figure 3.6 - Nome Residential Electric Rates. Source: NJUS. US Consumer Price Index used to adjust for inflation. rate and the lesser series shows the net rate after state subsidies have been deducted. The Power Cost Assistance program began subsidizing rural electric rates in 1981. This program was supplanted by the Power Cost Equalization program in 1985, which is still in existence today. For customers using less than 750 kWh per month, as do most residential customers in Nome, the rate paid by the customer for additional kilowatt-hours is the subsidized rate. Their decisions to use more or less electricity should, in theory, be influenced by this rate. The figure shows that the subsidized rate has substantially declined since 1981, both because of the higher level of assistance received under the PCE program relative to the PCA program and because of a reduction in the total electric rate. Seely, The total electric rate has declined in inflation-adjusted terms because of a number of factors. Electric rates in the early 80s were set above electric production costs, and the surplus was used to fund other city services. This cross- subsidy has diminished in the latter part of the 80s. Also, the cost-efficiency of the utility has improved. Fuel prices have decreased and the fuel efficiency of generation has improved. Finally, sales have increased, which has caused the per kilowatt-hour impact of fixed costs to decline. The average decline of the subsidized residential electric rate over the 1981-89 period is 8.5% per year, although the rate has flattened in the latter half of the period. Sucha decline would tend to have caused additional use of electrical services, since electricity became cheaper relative to other goods and services. Another factor influencing the level of electrical services used is the development of new services that use electricity. VCR‘s and home computers are examples of relatively new appliances that use electricity. Even without household income increases, these new services can be attractive enough to draw away dollars from other uses of household income. However, new uses for electricity can sometimes reduce electrical usage for older electrical services. More time spent with a home computer may mean less time in front of the television, Finally, the availability and price of substitute energy sources affects the amount of electrical services used by consumers. The drop in statewide Alaskan residential use per customer in the mid 1980s is probably due to substitution of natural gas space and water heating for electrical heating. 3=—=18 The above discussion addresses changes in the amount of electrical services used by customers. Also important are changes in the energy intensity of those electrical services. As well as affecting the amount of electric services used, the price of electricity affects the energy intensity of how those services are provided. Consumers buy more energy-efficient lights and appliances when electricity is more expensive. The extra cost of the more efficient appliances is more quickly paid back through energy savings when electric rates are higher. Other factors affect the energy intensity of electrical services other than the electric price paid by consumers. The range of appliance energy efficiency available to consumers is often determined by the electric prices seen by United States consumers as a whole. Frequently, the spectrum of appliances does not include units optimized for the high electric rates paid in rural Alaska. Appliance energy intensities are also affected by government-mandated energy-efficiency standards. This general discussion assists in interpreting Figure 3.4 and in projecting the change in future use per residential customer. In Figure 3.4, the Nome use per customer is substantially less than the Alaska and US use per customer. This is predominantly explained by the fact that there is little use of electricity for space heating in Nome relative to Alaska and the rest of the US. The severity of the climate combined with the substantial price advantage of fuel oil give electric space heat a very small market share. A recent forecast of electrical load for the Alaskan Railbelt (Colt, 1989) concluded that about 1,800 kWh/year of the average use per Railbelt customer is due to electric heat. Although space heating requirements in Southeast Alaska are less, a higher market share for electric space heat probably compensates. 3a—e Lo Another factor differentiating the Alaska and Nome use per customer curves from the US curve is lack of the air conditioning end use in Alaska. About 1,400 kWh/customer of the US figure is due to air-conditioning. Most relevant to this load forecasting effort are the changes in Nome residential use per customer over time. The growth rate over the 1981-89 period was 0.8% per year. It is useful to see if this rate corresponds to that derived from a simple econometric analysis of historical price and income data. Price during the period declined at the rate of 8.5% per year. From a cross-sectional analysis of 93 rural Alaskan communities, Yang (1989) determined a long-term residential electric price elasticity of -0.15 (t-statistic = 2.5), i.e. a 1% increase in price causes a 0.15% decline in usage.* Applying this elasticity to the 8.5%/year decline in inflation-adjusted electric price for Nome implies a 1.3%/year increase in use due to the price effect. However, an 8 year period may not be long enough to justify applying a long-term elasticity measure. Residential electricity-using appliances often have lives of 10-20 years. A large portion of today’s consumption is determined by appliance choices and prices in effect 10 years ago. °"This price elasticity is very low relative to elasticities determined on total US data. Griffin and Steele (1980, p. 232) report a residential/commercial price elasticity of 0.88. One explanation is that price signals in rural Alaska do not determine the range of appliance efficiencies offered. Rural Alaskan consumers respond to higher prices by choosing a more efficient appliance from a range of appliance efficiencies determined by US electric prices. Si —20 Yang also determined a residential income elasticity for electricity usage of 0.31 (t-statistic = 5.7).‘ Applying this to the 2.2%/year decline in income per customer over the 1981- 89 period gives an income effect of -0.7%/year. The sum of the price effect and the income effect is an increase in usage of 0.6%/year, relatively close to the actual rate of increase of 0.8%/year. This close correspondence may be somewhat coincidental. Yang found that price and income effects do not explain a large fraction of the variation in use per customer. The r-squared for his equation was 0.32. Nonetheless, it is useful to note that the historical use per customer trend is not radically inconsistent with historical prices, income, and associated elasticities. To develop a mid case projection of use per customer, we make the following assumptions. First, we assume that there will not be any substantial changes in electric price or income over the forecast period. The inflation-adjusted, subsidized electric price in Nome has stabilized since 1985. Because of the structure of the PCE formula, future variations in the unsubsidized price will be greatly dampened by the PCE program. Assuming residential rates track electric costs, a 1 cent/kWh increase in electric rate is dampened by a 0.95 cent/kWh increase in the PCE subsidy.* In the mid case, we ‘This value corresponds well with other estimates of income elasticity for electricity use. See Ross and Williams (1981), page 28. SBecause the 8.5 cent/kWh base rate in the PCE formula is not adjusted for inflation, the program actually results in a nearly constant nominal (not inflation-adjusted) subsidized price for electricity. A constant nominal price for electricity corresponds to a real (inflation-adjusted) price for electricity that declines at the rate of inflation. The eventual effect of this will be that residential consumers in some rural communities, such as Nome, will Spe also assume that income effects will be negligible. In the base case of his MAP forecasts, Goldsmith (1990) projects a nearly constant per capita income for Alaska through 2010. Declining permanent fund dividends and a structural shift in the economy towards lower-paying jobs are the assumptions underlying this projection. With a 0.25%/year decrease in household size, this per capita income projection corresponds to a decline in household income of 0.25%/year. With an income elasticity of 0.31, the income effect amounts to a decline in usage of less than 0.1%/year, a trivial effect. One factor we do consider in our mid case forecast is the recently enacted National Appliance Energy Conservation Act of 1987 (NAECA). The NAECA set energy-efficiency standards for a number of different residential appliances, including refrigerators, freezers, and water heaters. A subsequent amendment set standards for fluorescent lamp ballasts, the devices that start and control power flow to fluorescent lamps. The effective dates for the requirements vary by appliance, but most requirements will be in force by 1991. As an example of the effect of this standard, Colt (1989) projected an 1.2% annual decline in the refrigerator electric use per customer in Fairbanks. The forecast period was 1988 - 2010, and the decline was largely due to the standard. The model used was a detailed end use model that explicitly accounts for the additions and replacements of appliances and associated effects on electricity use. be paying less than urban communities such as Fairbanks and Juneau for electricity. We doubt that the Alaska legislature will accept this situation for long, and expect that the 8.5 cent/kWh base rate will be adjusted upwards because of inflationary trends. An ongoing inflation adjustment to the base rate will result in nearly constant inflation-adjusted price for electricity. 3 - 22 Colt also assumed an increase over time in miscellaneous residential uses of electricity. Future development of new electricity-using appliances and equipment (that do not substitute for other electric uses) would be consistent with this projection. We use Colt’s load forecast results for Fairbanks as the basis for our Nome projections. The price and income trends for Fairbanks used in the forecast were similar to our mid case assumptions concerning Nome.* Also, the electrical end use structure for Fairbanks is somewhat similar to Nome’s. Thus, use per customer trends should be comparable. Table 3.6 shows the results of the end use load forecast for Fairbanks. The average use per customer is shown for each of the end uses. The bottom row in the table applies two adjustments to the Fairbanks end use results. First, the electric space heating end use is removed because we believe electric space heating is negligible in Nome. Next, 31% of the water heating use is removed to account for a lower market share of electric water heating in Nome.” The adjusted use per customer declines at a rate of 0.2% per year, and this is taken to be our mid case projection for Nome. The results show that the effect of the efficiency standard is almost canceled by a growth in miscellaneous uses of electricity. To develop a low case estimate for the use per customer trend, we make one adjustment to the mid case estimate. We assume ‘The real price of electricity in Fairbanks was projected to grow at a rate of 0.4% per year. With a 0.15 price elasticity, the effect on consumption would be a decline of only 0.06% per year. 7The initial electric water heater market share in Fairbanks was estimated to be 36%. Phil Kaluza of Arctic Energy Systems in Nome estimates that approximately 25% of Nome households have electric water heaters. 2 “ae Table 3.6 - End Use Forecast Results for Fairbanks. Source: Colt (1989). Average Use per Fairbanks Customer kWh per year Space Heating Water Heating Refrigerators Freezers Cooking Clothes Drying Lighting Miscellaneous i ’ . |rotar_ | 8,699 | 8,315 | -0.218 Total - Heat - 7,426 Tana. -0.19% 31% Water that the Power Cost Equalization Program is phased out by the end of the forecast period, 2014. If NJUS acquires the Alaska Gold Company load, as we believe they will, loss of the PCE subsidy will cause residential rates to rise about 40%. With a price elasticity of 0.15, this price increase will reduce consumption by the end of the period by 6%. This lowers the use per customer trend from -0.2% per year in the mid case to -0.45% per year in the low case. To develop a high case estimate for the use per customer trend, we make one adjustment to the mid case estimate. The steep drop in inflation-adjusted Nome residential electrical rates during the 1980s may not have yet realized its full impact. Electric water heaters have substantially less initial cost than oil-fired water heating systems, especially for homes with forced-air space heating systems. With a oo ae greatly diminished price difference between fuel oil and electricity, electric water heaters in new construction and as replacements for existing water heaters may become more prevalent. Electric water heater consumption averages about 4,800 kWh/year. If the average market share increases by 10 percentage points, average use per customer will be 480 kWh/year higher than otherwise. This increases our mid case trend from -0.2% per year to our high case trend of +0.16% per year. 3.4.3 Summary of Residential Forecast Table 3.7 - Residential Forecast Assumptions by Case Assumption Population People per -0.25% per -0.25% per -0.25% per Customer year year year Use per -0.45% per -0.2% per 0.16% per Customer year year year Table 3.7 summarizes the assumptions that were used to generate the low, mid, and high residential forecasts. Figure 3.7 shows the results of the forecast in terms of sales of kilowatt-hours. The numeric values are presented at the end of the load forecast report in section 3.9. 3.5 Commercial/Other Forecast All non-residential electric use, except for mining, is forecast as one category. The category primarily includes commercial buildings, community facilities, and street lighting. our forecasting method involves forecasting two variables: number of indoor employees (total employment - mining employees - 1/2 of Site) Nome Residential Sales million kWh/year 24 | © oa ann AES UL UR UE RR RL ee en enn en ee 1981 1986 1991 1996 2001 2006 2011 Year . —=— Historical —— Mid ow: ———— gh Figure 3.7 - Residential Energy Sales Forecast construction employees) and electric use per employee. The indoor employment forecast was described in section 3.3. The following section discusses the use per employee forecast. 3.5.1 Use per Employee Figure 3.8 shows the historical pattern of use per employee for the Nome, US, and Alaskan commercial sector. All data series show growth during the 80s of from 1.6% to 3.1% per year. The Nome and the Alaska series exhibit more growth than the US series, especially during the 1984 through 1987 period. One possible explanation for this is the addition of a large number of public buildings with relatively high electric use per employee. The state spending boom of the early 80s funded 3 - 26 Use per Commercial Employee 12 104 = L g 87 —>_—— oo oO 2 => o 6-5 S36 Growth Rates: x~¢E Nome, 1981-89, 2.7% / yr —<—" US, 1982-88, 1.6% / yr a Alaska, 1983-89, 3.1% / yr 1981 1982 1983 1984 1985 1986 1987 1988 1989 Year —™- Nome —— US — Alaska Figure 3.8 - Use per Employee. Sources: Nome - NJUS data, US - US Statistical Abstract, 1989, Alaska - Penny Haldane, Alaska Energy Authority. the construction of a number of buildings that began consuming power during this 84 through 87 period. Because of a large amount of building floorspace per employee (e.g. recreation centers, schools), the addition of these buildings drive up the electric use per employee measure. Another important factor during this period was the economic recession beginning in late 1985. On a statewide basis, this reduced the number of employees but it did not reduce the amount of commercial floorspace.* Although some of this *The Nome employment figures do not show a decrease in employment during the recession. However, the growth in employees was reduced. Because of the time required to construct buildings, the growth in commercial floorspace was probably not reduced to the degree that employment growth was reduced. (Buildings under 3 =_27 floorspace was left vacant by the recession, vacant floorspace still consumes electricity. The ratio of electric use to number of employees was therefore increased. In general, use per employee is determined by the type of work that is being performed, the amount of work performed per employee, and the electric intensity of the process used to perform the work. The type of work being performed is important because different activities require differing amount of electricity. As discussed before, operating and managing large public buildings requires few employees relative to the amount of electricity consumed. Structural changes in the economy that change the overall composition of the type of work being formed will have effects on average electricity use per employee. Changes in the amount of work performed per employee can have effects on electricity use. If implementing a bar code pricing and inventory system in a grocery store reduces the number of employees required to operate a 10,000 square foot store, electricity use per employee will increase. Other productivity-enhancing changes have a much less dramatic impact on electricity use per employee. Enhanced communication systems and information access systems are not very electricity intensive but can result in substantial productivity improvements. Finally, changes in the electricity intensity of the process used to perform the work will affect electricity use per employee. Approximately half of commercial electricity use is for lighting. Technological progress has produced substantial improvements in the energy-efficiency of lighting technologies, and their implementation reduces the electricity construction when the recession hit were still completed). 3 - 28 intensity of work processes. Other changes can cause more electricity use. Providing additional and higher quality lighting in retail spaces has been found to be an effective marketing tool in some situations. To forecast use per employee, we once again rely on the more detailed work performed by Colt (1989) for the Railbelt of Alaska. Colt utilized an end use model that projected usage levels for specific types of electric services such as lighting and ventilation. For the Fairbanks forecast, the electric price trajectory was relatively flat. Even though inflation-adjusted prices in Nome have been dropping dramatically during the previous decade, we do not expect significant further price decreases.*® As explained in section 3.4.2, much of the price decrease was due to reduction of the profit present in the electric rate, which was used to fund other non-electric utility services. Much of this profit has already been eliminated. The end use modeling for Fairbanks projects significant declines in the amount of lighting electricity use per square foot, a 23% reduction between 1988 and 2010. It also projected increases in miscellaneous electric use to account for additional uses of equipment such as personal computers. The overall change in use per square foot for Fairbanks between 1988 and 2010 was a reduction of 0.5% per year. We utilize this as our mid case estimate for Nome. However, our Nome forecast is based on use per employee. A trend in commercial electric use per square foot is not equivalent to a trend in electric use per employee if the amount of floorspace per employee is also changing. The National °In the Nome commercial sector, the price relevant for determining consumption levels is the unsubsidized price, since most commercial customers consume more than the 750 kWh/month limit present in the PCE program. 3 - 29 3.6 Academy of Sciences (1979) project an increase in non- residential floorspace per employee of 0.6% per year. This projected increase in floorspace per employee combined with Colt’s projected decrease in use per square foot produce our Mid case estimate of a 0.1% per year growth in use per employee. As a Low estimate for the use per employee trend, we assume that floorspace per employee grows at only 0.3% per year. Electric use per employee therefore declines at 0.2% per year in the Low case. For the High case use per employee trend, we utilize a forecast from the Edison Electric Institute (EEI), a trade association for the electric utility industry (forecast discussed in Oatman and Talbert, 1989, p.2). Some analysts claim EEI’s forecasts are consistently high (Ross and Williams, p.23). EEI projects use per square foot for commercial buildings to grow at 0.8% per year through 2000 and then remain constant from then on. The composite growth rate is 0.4% per year for the 1989 through 2014 period. Adding the 0.6% per year growth in floorspace per employee gives a use per employee growth of 1.0% per year for the High case. Table 38 summarizes the assumptions used in the commercial/other forecast. Figure 3.9 shows the resultant forecast of commercial/other kilowatt-hour sales, not including losses. Non-Mining Summary Net generation requirements ("net" means not including usage within the power plant) must also include transmission and distribution losses. The forecast summary tables in section 3.9 show the historical losses for 1981-89. We project future losses at 6.5%. SPO Table 3.8 - Assumptions for Commercial/Other Forecast. Assumption Indoor Low Mid High Employment Use per -0.5% per year -0.5% per 0.4% per Square Foot year year Floorspace 0.3% per 0.6% per 0.6% per per Employee year year year Use per -0.2% per 0.1% per 1.0% per Employee year year year Nome Commercial/Other Sales 307 255 S 204 = £ = 15; Ee ond S ; 105 55 Se ee te tt ae Tt to O98) 1986 1991 1996 2001 2006 2011 Year —#®— Historical —— Mid — Low — High Figure 3.9 - Commercial/Other kWh Sales Forecast. Figure 3.10 shows the combined net generation requirements for residential and commercial/other loads, including losses. It is also necessary to forecast the peak generation requirement as well as the total requirement. Historical measurements of NJUS a Non-Mining Net Generation million kWh/year 54 [© ho San Sn Sn Sn SL SS SU Sn BS Sn Sk Sk SS Se Se en oe ee een ee en een ee ee 1981 1986 1991 1996 2001 2006 2011 Year —#—- Historical —— Mid ——-—Low —=— High Figure 3.10 - Non-Mining Net Generation Requirements. peak requirements have been for the system load as a whole. No detailed analysis has been done divide the peak requirement by customer class. Because we forecasted non-mining (i.e. residential and commercial/other) and mining peak requirements separately, we needed to separate historical system peaks into non-mining and mining components. This separation was only required for the 1989 peak, when the WestGold Bima dredge contributed 1.3 MW to the system peak. Figure 3.11 shows the non-mining load factor for Nome from 1981-89. We project the future non-mining load factor to be 60%. The shape of the non-mining peak requirements forecast is similar to the net generation requirements forecast shown in the previous graph. The numeric detail of the forecast is provided in the forecast summary tables found in section 3.9. St oe Non-Mining Load Factor Nome 70%7— aad aN 50%7 40% 30% 5 Load Factor 20% 4 10% o———————— T —— 1981 1982 1983 1984 1985 1986 1987 1988 1989 Year Figure 3.11 - Historical Load Factor for Nome Non-Mining Loads. Source: NJUS data adjusted for mining sales. 3.7 Mining Forecast 3.7.1 General Background A close examination of the Nome mining sector is important to the electrical load forecast because gold mining is very electricity intensive. Also, mining supports a significant fraction of the total Nome population and thus affects non- mining electricity use. Gold production in Nome over the past century has amounted to 4.6 million ounces, worth about $1.8 billion dollars at today’s prices. Figure 3.12 shows how the production has varied over the time period. Very high production levels were realized during the gold rush of the early 1900s. The other 3 = 33 Nome District Gold Production 300 2507 200+ 1505 100+ Thousands of Ounces / Year 507 0 1897 1907 1917 1927 1937 1947 1957 1967 1977 1987 Year Figure 3.12 - Historical Gold Production in the Nome District. Source: T.K. Bundtzen, Alaska Division of Geological and Geophysical Survey. significant feature of the graph is the tremendous variability of production levels. Mining in Nome was reactivated in the mid 70s after a period of very slow activity. Since then, gold prices and production have risen substantially, as shown by Figure 3.13. 3.7.2 Recent Mine Operators and Potential Future Operators Figure 3.14 shows who the major gold mining operators were during 1989. Alaska Gold Company has had the most operating history in the Nome area. They reopened their operation in 1975. They operate two onshore dredges that mine low-grade placer deposits, gold found in gravel along stream beds and beaches. Although their production in 1989 was 21,000 ounces, it has been as high as 27,000 ounces. Alaska Gold Company has 3 => 34 US Wholesale Gold Price Inflation-Adjusted $800 $700- $600- $5005 $4007 $300- Price, 1989 $/troy oz. 8.5% / year $200- $100- $0 T T sa T T T T T T T T T ee ——— T erry 1971 1973 1975 1977 1979 1981 1983 1985 1987 Year Figure 3.13 - US Wholesale Gold Prices. Source: Metal Stat, Handy & Harman. Inflation-adjusted by the US Producer Price Index. very large reserves in the Nome area, which should last through the period of this forecast. Production costs are in the $300 - $350 per ounce range (not counting fixed costs), whereas current gold prices are $390 per ounce. While their financial statements do not look encouraging because of large debts to their parent corporation, variable production costs are low enough that continued operation is likely if gold prices remain above $350 per ounce. Electrical power is required for their two dredges and associated water pumps. Accurate power requirements data was available for 1989, indicating consumption of about 450 kWh per ounce of gold produced. Because they only mine from mid- June through mid-November (a 160 day season is typical), their annual load factor is low, approximately 29% for a typical Sates © Nome District Gold Production 1989 Windfall Mining 5,450 oz. Alaska Gold Company 21,000 oz. WestGold, Bima 30,660 oz. Figure 3.14 - Gold Operators in the Nome District, 1989. Source: Bundtzen, et al. (1990). season. However, while they are mining consumption remains relatively constant for 24 hours per day, except for equipment down-time periods. The Alaska Gold Company has historically produced their own power, except for minor wintertime needs. They have a power plant in the city of Nome and deliver power to their dredges outside of town via their own distribution system. However, as we discuss in section 3.7.4, we believe it is likely that NJUS will supply the majority of the Gold Company’s needs starting in 1991. Western Gold Exploration and Mining began operating their very large Bima offshore dredge in 1986. After a major mechanical failure in September of 1990, they announced that they would 3 - 36 cease operations in Nome and sell the Bima. They have stated that operation of the Bima has not been profitable. The Bima was expected to produce 50,000 ounces per year of gold from its Nome operations. However, mechanical difficulties and poor weather kept its maximum take to 36,700 ounces in 1987. Power requirements are substantial. Very approximate estimates indicate that power use while mining offshore during the summer amount to 240 kWh per ounce of gold mined. In addition, the dredge docked during the winters at the Nome causeway. For protection, snow was made and bermed around the dredge. Snow-making power use and on-board wintertime requirements amounted to about 80 kWh per ounce mined. Had levels of gold production closer to design levels been achieved, the fixed wintertime use would have amounted to about 60 kWh per ounce mined. The annual load factor of this wintertime use is about 22%. For the 88-89 winter and the 89-90 winter, the Bima purchased its wintertime power requirements while docked at the causeway from NJUS. While offshore, power was generated by on-board diesel generators with efficiencies of approximately 12 kWh/gallon. Offshore tin dredges in Indonesia are fed by submarine cables from onshore power plants, and there is speculation that a similar arrangement could be used in Nome if more offshore mining occurs. Windfall Mining operated in the Nome area for about 5 years ending in 1989. In 1989, they produced only 55% of what they projected for the year, and the operation shut down. Although the 1989 take was 5,450 ounces, past production peaked at about 13,000 ounces in 1987. Windfall has onsite generation, but began purchasing a portion of their needs from NJUS in 1988. No estimates were available Sao for their total power production needs. However, they process the same kind of material as the Alaska Gold Company, and we expect that their energy requirements are similar per ounce of gold produced. Windfall only mined during the summer season. Of the other 2,390 ounces of gold produced in 1989 in the Nome District, a significant portion was from Anvil Mining Company. The mine is still operative at the time of this report. Although they have onsite generation, they began purchasing all of their power requirements from NJUS in 1988. An approximate estimate of their power requirements is 200 kWh per ounce produced. There are a number of potential future mining operations in the Nome area. Cyprus Mining currently has employees in Nome that are investigating a strip mining operation that would utilize an electric dragline. Production estimates are 20 - 40,000 ounces of gold per year. Company officials give the operation a 50-50 chance of opening during the 1991 season.*° Power requirements estimates were very uncertain but calculated out to about 350 kWh per ounce mined. One design is based on a year-round operation that would involve stripping during the winter to avoid problems with mud, and the design would involve processing material during the summer. An alternative design utilizes a summer-only operation. The company has indicated that they would probably buy power from NJUS, as they thought the prices were reasonable. The cost of interconnection would be very small, about $5,000, because the operation would be located near an existing transmission line to the Nome Beltz school. personal communication with Jay at Cyprus Mining in Nome, October 10, 1990. 3 =~ 38 Aspen Exploration in partnership with Tenneco Minerals and other entities is conducting a lode gold exploration program in the Rock Creek and Anvil Creek area near Nome. Reserves have been identified amounting to about 500,000 ounces of gold. If the project comes to fruition, it would be located about 8 miles north of Nome and could start as early as 1994.** The open pit mine would probably produce 50,000 ounces per year for 10 years.** T. K. Bundtzen, Division of Geological & Geophysical Surveys, puts the probability of this mine operating at more than 10% but less than 50%. Estimates for power requirements are about 360 kWh per ounce of gold produced with a very high load factor of 80% because of year-round production. The companies have indicated a willingness to buy power from the city, citing a reasonable price and elimination of a large fuel storage system as reasons. Serving the mine from the NJUS system would require the construction of 5 miles of transmission line at a cost of $50,000 per mile.*® The levelized cost (constant 1989 $) for transmission of the power calculates to about 0.8 cents/kWh.** However, the transmission of electrical power eliminates the need to transport fuel from Nome to the mine site for power generation. Fuel transport costs for onsite generation would be at least 15 cents/gallon, amounting to 1 Personal communication with William Newlin, Tenneco Minerals, October 9, 1990. 22personal communication with T. K. Bundtzen, Division of Geological and Geophysical Survey, October 11, 1990. 22personal communication with Joe Murphy, Nome Joint Utilities System, September 27, 1990. 24§250,000 capital cost, 10 year life, 4.5% real interest rate, operation and maintenance of line at 1.5% of capital cost per year, 6.5% losses on generated power costs of 8.5 cents/kWh, and annual transmission of 18 million kWh. 3 = 39 cent/kWh. There would be no substantial uses for the waste heat from onsite generation. Other potential mining operations in the Nome area include Coastal Hills and Coastal Plains where Aspen Exploration is currently doing exploration work. The project is more speculative than the Rock Creek project and we did not obtain production estimates. The Big Hurrah lode gold mine is another potential project about 40 miles from Nome. fT. K. Bundtzen puts the probability of production at a higher level than the Rock Creek project. Production would be 30-40,000 ounces per year, and power requirements would be similar to the Rock Creek project, 360 kWh/ounce. Because of the distance, transmission costs would be higher, about 3 cents/kWh. 3.7.3 Low, Mid, and High Mining Assumptions, and Forecast Summary For the Mid mining case, we assume that onshore gold production will consist primarily of Alaska Gold Company and some small other summertime producers, totalling 25,000 ounces per year. We assume that power requirements will be approximately 450 kWh/ounce with a 29% annual load factor. In addition, we assume that a smaller offshore dredge will begin mining in 1995 producing 17,000 ounces per year. Tee Ke Bundtzen indicates that the experience with Bima indicated that the appropriate size for an offshore dredge is a dredge much smaller than the Bima. He thinks it is likely that one will begin operation in the not too distant future. We only include the wintertime (off-season) power requirements in the load forecast, and we estimate these to be 60 kWh per ounce of gold mined with an annual load factor of 22%. Thus, the effect on the load forecast is small relative to the onshore operations. 3 - 40 For the Low forecast, we assume that production from summertime onshore operations declines to 12,000 ounces per year by 1992. This case is consistent with the shutdown of one of the Alaska Gold Company dredges, as occurred during 1985 and 1986. We assumed no offshore production in this scenario. For the High forecast, we assume that onshore summertime production increases from 25,000 ounces per year to 32,000 ounces per year by 1992. Also, we assume that the Cyprus strip mining operation begins in 1992 at a production level of 30,000 ounces per year. We assume that it is a year-round operation that uses 380 kWh per ounce produced, and the load factor is 80%.*° We assume that 2 small offshore dredges are in service by 1993 producing 34,000 ounces of gold per year, with per ounce power requirements as in the mid case.** Table 3.9 summarizes the assumptions used in the mining forecast. Figure 3.15 shows the mining production forecast for the Low, Mid, and High cases. Figure 3.16 shows the mining net generation forecast results. 6.5% losses were assumed to calculate net generation requirements. To convert these mining production forecasts into mining employment forecasts for use in the population/employment modeling, we assumed that each 1,000 ounces per year of onshore gold production requires 2.6 annual average employees, and each 1,000 ounces per year of offshore gold production requires 1.4 annual average employees. These employment *sBecause of the uncertainty of the Cyprus power requirements estimate, 350 kWh/ounce, we use an estimate closer to the Alaska Gold Company electricity intensity, because the AK Gold Company figure is substantially more certain. *~T, K. Bundtzen states that 100,000 ounces per year of offshore gold is conceivable. 3 - 41 Table 3.9 - Assumptions for the Mining Forecast by Case ee Rise from 25,000 Decline to 12,000 25,000 ounces/yr, | ounces/yr to 32,000 Summertime | ounces/yr by 1992. 450 kWh/ounce, ounces/yr by 1992, Gold Mining 450 kWh/ounce, 31% load factor. 450 kWh/ounce, 31% load factor. 31% load factor. Cyprus starts in 1992 at 30,000 ounces/yr, 350 kWh/ounce, 80% load factor. Onshore, Year-Round Gold Mining 17,000 ounce/yr dredge begins operation in 1995. 60 kWh/ounce onshore use, 22% load factor. Two dredges by 1993, 34,000 ounces/yr, 60 kWh/ounce, 22% load factor. Offshore Gold Mining ratios were derived from data for the Alaska Gold Company and the WestGold Bima operations. 3.7.4 Likelihood of NJUS Supplying Mining Loads This load forecasting effort addresses loads that have economical access to the present NJUS transmission and distribution system. To determine whether these loads will actually be served by the NJUS generation system requires an examination of the relative costs of on-site generation versus NJUS generation. In this section we address the relative cost of onsite generation versus NJUS diesel-fired generation. We do not address the relative cost of on-site generation with alternative forms of NJUS generation, such as coal. However, use of any alternative form of NJUS generation that is more cost-effective than diesel generation will increase the 3. 42 Nome Gold Forecast 100 ——- ee ae = —s 907 805 707 605 504 405 305 2075, 105 Thousands of Ounces / Year g 981 1986 1991 1996 2001 2006 2011 Year —#— Historical —— Mid = Low —— High Figure 3.15 - Gold Mining Production Forecast likelihood of loads being served from the NJUS systen. For small commercial and residential users, the economies of scale inherent in NJUS generation and the presence of the Power Cost Equalization subsidy clearly tilt the economics in favor of purchase from NJUS instead of on-site generation. There are no significant instances of onsite generation for small users that have access to the NJUS grid. For larger users, the economics of on-site generation improve. Historically, the mining operations in the Nome area have produced their own power. The Alaska Gold Company, the oldest mining operation in the Nome area, has always supplied their own power, except office and shop needs during the winter when mining is not occurring. Mine operators indicate that NJUS has never had the generation capacity to supply the mine’s 3)— 43 Mining Net Generation 40 355 307 20- 155 *y 5 ge TT te st es ee 9 981 1986 1991 1996 2001 2006 2011 Year million kKWh/year —®- Historical —— Mid =, LOW ===. High Figure 3.16 - Mining Net Generation Forecast needs. For example, Alaska Gold Company power demands peak at 4.5 MW, and the city of Nome’s needs peaked at 2.7 MW in 1977. This situation has recently changed. During the 1980s the non-mining loads in Nome grew because of the boom in the overall Alaskan economy. Non-mining loads now peak at about 4.2 MW, comparable to the mining loads in the area. Further, the new NJUS general manager, Mr. Joe Murphy, has aggressively pursued the acquisition of the mining loads since he began his position in 1988. NJUS has offered interruptible power purchase contracts to the mining operations at a rate less than average cost but still in excess of the NJUS incremental generation cost. The NJUS board has accepted this pricing policy because of its benefits for the other customers on the systen. 3 - 44 Windfall Mining and Anvil Mining began purchasing NJUS power in 1988. The WestGold offshore dredge, the Bima, started buying NJUS power while docked at the Nome causeway during the 1988-89 winter season. The Bima peak demand was 1.3 MW and the dredge used about 2.7 million kWh during the winter of 1989-90. Power was sold to the Bima for 12.5 cents/kWh for the first 100,000 kWh/month, 11.9 cents/kWh for the next 100,000 kWh/month, and 10.9 cents/kWh for any additional use. The standard three-phase commercial rate in effect at the time was 12.5 cents/kWh plus a $10/kW/month demand charge. These actual power purchase decisions provide evidence that NJUS diesel generation is less expensive than on-site generation for a variety of situations. The Anvil and Windfall mining operations characterize small to medium size summertime mining operations. The WestGold Bima decision to buy power while docked at the Nome causeway is indicative of the wintertime power purchase decision of a large offshore gold mining dredge. The remaining existing load of substantial import that still self-generates is the Alaska Gold Company, a load with a 4.5 MW peak demand and an annual electricity use of about 10 million kWh. NJUS is currently negotiating with Alaska Gold to supply their total annual load (NJUS currently supplies their minimal winter load). NJUS believes a contract will be signed before the 1991 mining season. Mr. Joe Fisher, general manager of Alaska Gold, and Mr. Gary Butcher, power plant operator for Alaska Gold, also believe that a contract is likely. After examining the relative costs of generation, we concur that a contract is likely. The factors relevant to our conclusion are: + The fuel efficiency of NJUS generation is better than Alaska Gold generation. The Alaska Gold Company diesel generation has a fuel efficiency of about 13.1 kWh generated (net of station use) per gallon of fuel used. 1989 average generation efficiency for NJUS was 14.3 3 - 45 kWh/gallon. NJUS is currently in the process of installing a new 3.7 MW generator with a fuel efficiency at 75% load of 15.4 kWh/gallon. This unit will be base- loaded and much of its annual energy generation will be used by non-mining loads. However, the extra generation required to served mining loads will come from generators with efficiencies near 14.5 kWh/gallon. Differences in the fuel efficiency of NJUS generation and on-site generation may persist over the long-term because NJUS generation operates at higher annual capacity factors than mining generation. The extra capital cost of more efficient generation is more readily cost-justified for generation units with high utilization. + NJUS has recently purchased and stored fuel at less cost than Alaska Gold. During 1989, the Alaska Gold Company paid 25 cents/gallon more for fuel than NJUS.*” NJUS benefits from cooperative fuel purchasing with other communities in western Alaska. The cooperative purchases approximately 6 million gallons of fuel annually (NJUS uses about 1.8 million gallons without the Alaska Gold load). The combined effect of better fuel efficiency and lower fuel prices currently gives NJUS approximately a 2.4 cent/kWh advantage over onsite generation. + The seasonal load shapes of non-mining loads and mining loads are complementary. The onshore placer mining loads are present during the summer when non-mining loads are at a minimum. The NJUS generation that was needed to meet winter peaks is available to supply a portion of the summer mining loads. Serving this portion of the mining loads does not require an additional capital investment in generating capacity. In addition, the reserves necessary to back up the required generation are less on a large system than they are on an independent, on-site systen. + Alaska Gold has had environmental problems with their fuel storage system that have caused examination by the EPA. ._ Buying power from NJUS will greatly reduce their fuel storage needs and alleviate some of these problems. - One typical advantage of self-generation is the reduction or elimination of transmission and distribution costs and losses. In the case of Alaska Gold, the dredges and water pumps that consume electricity are located outside of town, but the company’s power plant is located within town. Thus, self-generation still requires’ the 7Ppersonal Communication with Mr. Joe Murphy, NJUS, September 27, 1990. 3 - 46 transmission of power a significant distance from the power plant. If NJUS acquires the Alaska Gold load, they will simply connect to the existing Alaska Gold transmission facility, which passes directly by the NJUS Snake River power plant. Interconnection costs are minimal, and the transmission and distribution losses from the NJUS plant are similar to those from the Alaska Gold plant. + Another typical advantage of self-generation is the ability to inexpensively utilize the waste heat from generation on-site. For a self-generation situation in Nome where all of the water jacket waste heat can be utilized, the value of that heat is approximately 3 cents per kilowatt-hour produced. In the case of Alaska Gold, there are few uses for the waste heat during the summer when the bulk of the electrical consumption occurs. Thus, this advantage does not exist. + There appear to be economies of scale in the operation and maintenance of diesel generation equipment. The Alaska Gold Company employs approximately 8 power plant operators during the summer mining season. NJUS claims that they will only need to add 1 year-round employee if they acquire the Alaska Gold Company load. The net savings is about 1 cent/kWh. . + NJUS has access to tax-free financing that private mining companies do not have access to. MThis lowers capital costs per kWh for NJUS generation relative to onsite generation. One of the largest factors acting against a power sales contract is the effect that the sale will have on Power Cost Equalization receipts. In Nome’s case, PCE will be calculated on the average cost of electricity for the utility. Because the incremental cost of generating the Alaska Gold Company load is less than the existing average cost, addition of the load will cause the utility’s average cost to drop. This drop will result in a drop in the PCE rate. We calculated the drop to be approximately 2 cents/kWh, and it will apply to the 9.8 million kWh currently subsidized by the PCE program. To compensate for this loss of revenue, margins earned on sales to the Alaska Gold Company need to be at least 1.9 cents/kWh. 3 - 47 It appears as though contract prices currently under discussion will provide at least this level of margin. 3.8 Combined Peak Generation Requirement for Non-Mining and Mining Loads Non-mining, onshore summertime mining, onshore year-round mining, and the winter power requirements of offshore mining have seasonal and hourly variations that are much different from each other. Thus, the total peak generation requirements for the Nome area cannot be calculated as the sum of the separate peak requirements. To calculate the Nome-area coincident peak generation requirement, we used the following technique. First, Table 3.10 was developed, which gives monthly peak generation requirements for each load component. Using the non- mining load component as an example, the 100% indicates that the non-mining loads typically peak during January. The 71% for July indicates that the July peak for the non-mining loads is 71% of the annual peak. Given an annual peak requirements estimate for non- mining, multiplying by the appropriate figure in the table gives an estimate of the peak requirement for a particular month. The table was derived from 1988-89 monthly data from NJUS for non-mining loads, from knowledge of the mining season for onshore mining loads, and from WestGold billing history for the shore-component of the offshore mining loads. We next assumed total coincidence among the monthly peaks for each load component. For example, if the non-mining peak requirement in July is 5.0 MW, the onshore mining peak is 4.0 in July, and the offshore load is 0 MW in July, we assumed that the total coincident peak in July is 5.0 + 4.0 = 9.0 MW. This assumption implies that the peak load for each individual component occurs at the same time during the month. While this obviously is not exactly correct, it suffices given the overall accuracy of this forecasting effort. 3 - 48 Table 3.10 - Relative Monthly Peak Demands Off-shore Mining | Mining 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% 100.0% The hourly load variation for the non-mining loads is very constant during the daytime. Also, the mining loads tend to be quite constant when the dredging or snow-making equipment is running. The more constant the loads are, the higher the probability that peak loads will be coincident. Once total Nome-area peak requirements are calculated for each month, the annual peak requirement is simply the maximum of the monthly requirements. This technique was used for each of the years in the analysis. 3.9 Forecast Summary The tables and figures in this section summarize the results of the forecast. Figure 3.17 shows the total net generation 3 = 49 requirements for the non-mining and mining loads combined. The associated peak generation requirements are shown in Figure 3.18. The area chart, Figure 3.19, shows how the net generation requirements in the Mid case divide between residential, commercial/other, and mining uses. Annual average growth rates of total net generation requirements and peak generation requirements for the period 1980 through 2014 are shown in Table 3.11. Finally, Table 3.12 through Table 3.17 give the numeric detail of the forecast. There are two tables for each case--Mid, Low, and High. The first table gives the results for the residential and commercial/other forecasts for that case. The second table gives the results for the mining load forecast and gives the non-mining and mining totals. The bottom row of the table gives the annual average growth rate for the period 1989 through 2014 for each variable in the forecast. SEO Total Net Generation 807—— 705 60-5 a S 505 io = 404 § = 30+ [= 20+ 107 So 981 1986 1991 1996 2001 2006 2011 Year —#— Historical —— Mid — Low — High Figure 3.17 - Total Net Generation Requirements, Non-Mining and Mining Loads. Total Peak Generation Megawatts 2 oO 1981 1986 1991 1996 2001 2006 2011 Year —#- Historical —— Mid — Low —= High Figure 3.18 - Total Peak Generation Requirements, Non-Mining and Mining. Saeed Total Net Generation Mid Case 0 1981 1986 1991 1996 2001 2006 2011 ¥i Residential HE Commercial/Other KK Mining gur. 1 ati qu da ‘abl wth Rat f tal qu da ak c equir ic es - € Non-Mining Forecast Case: Mid ‘Commerical, Commun. Facil., and Street Lighting Use per Total Indoor Employce Sales Employees __kWh MWh Residential Non-Mining Totals T&D T&D Net Load Peak Losses Losses Generat. Factor Generat, % MWh MWh % Use per Total People/ Custome Cust kWh ; 5,037 3,071 1,060 2.90 5,588 5,923 3,102 1,244 2.49 5.366 = 6.676 3,146 1,294 2.43 5.459 7,065 3,236 1,290 2.51 5,394 6,957 3,208 1,280 2.51 5.640 7.219 3,306 1,331 2.48 5,844 7,778 6.015 8,084 5812 7,934 x 5 5,801 8.041 ‘ 3,530 1,384 2.55 5,789 8.011 1,310 10,346 =—-13,556 3,515 1,382 2.54 5,778 7,983 1,286 10,357 13,323 3,535 1,393 2.54 5,766 8,031 1,291 10,367 13.386 3,578 1,414 2.53 5.755 8,134 1,306 10,377 13,549 3,696 1,464 2.53 5,743 8,406 1,340 10,388 13,924 3811 1,513 2.52 5,732 8.671 1,378 10,398 = 14,323 3,910 1,556 251 5,720 8.901 1,407 10,409 14,641 3,999 1,596 2.51 5,709 9,109 1,431 10,419 14,911 4,081 1,632 2.50 5.697 = 9,301 1,453 10,429 15,150 4,164 1,670 2.49 5.686 = 9,495 1475 10,440 15,394 4,210 1,693 2.49 5,674 9,604 1,465 10,450 15,305 4,262 1,718 2.48 5,663 9,727 1,472 10,461 15,396 4,320 1,745 2.48 5,652 9,865 1,483 10,471 15,527 4,392 1,779 2.47 5,641 10,035 1,504 10,482 15,765 4,479 1,819 2.46 5,629 10,239 1,530 10.492 16.053 4,593 1,870 2.46 5,618 10,504 1,564 10,503 16,429 4,716 1,925 2.45 5,607 10,790 1,601 10,513 16,835 4846 1,983 2.44 5,596 11,094 1,641 10,524 17,270 4,935 2.024 2.44 5,584 11,303 1,659 10,534 17.472 5,009 2,060 2.43 5,573 11,478 1,662 10,545 17,521 5,102 2,103 2.43 5,562 11,698 1,682 10,555 17.750 5.197 2,148 2.42 5SS1 11,922 1,702 10,566 17,981 2,193 241 5,540 12,150 1,722 10,576 18,215 240 241 5,529 __ 12,383 1,743 10,587 18.453 1989 - 2014 Annual Average Growth Rates: 1.7% 2.0% 0.25% _-0.20% 1.8% 1.1% 0.1% 1.2% 0.1% 1.3% 14% 0.2% 6.5% 1,499 23,070 = 60.0% 4.39 6.5% 1,498 23,066 60.0% 4.39 6.5% 1,480 22,78 0.0% 4.33 6.5% 1,488 22,05 = 0.0% 4.35 6.5% 1,507 23,190 60.0% 441 6.5% 1,552 23,882 60.0% 4.54 6.5% 1,598 24,592 60.0% 4.68 6.5% 1,636 25,178 = 60.0% 4.79 6.5% 1,669 25,689 60.0% 4.88 6.5% 1,699 26,150 60.0% 497 6.5% 1,729 26,619 60.0% 5.06 6.5% 1,731 26,640 60.0% 5.07 6.5% 1,746 26,868 60.0% 5.11 6.5% 1,764 27,155 0.0% 5.16 6.5% 1,793 27,593 60.0% 5.25 6.5% 1,827 28,119 = 60.0% 5.35 6.5% 1,871 28.803 60.0% 5.48 6.5% 1,919 =. 29,545 60.0% 5.62 6.5% 1,971 034 60.0% 5.77 6.5% 1,999 37715 = 0.0% 5.85 6.5% 2,015 31,014 60.0% 5.90 6.5% 2,046 31,494 = 60.0% 5.99 6.5% 2,078 31,980 = 60.0% 6.08 6.5% 2,110 = 32,475 60.0% 6.17 6.5% 2,142 32,978 __ 60.0% 6.27 SpeoT HUTUTW-UON ‘SR[NSeYy 4SedeIO™, eseD PIW - ZI°E STqGeL 0S Mining Forecast and Totals Case: Mid Onshore Gold, Summer Only Onshore Gold, Year-Round Offshore Gold, Shore Elect. Req't Mining Total ining + Non-Mining Total Total Net Peak Gold Total Net Peak Gold Onshore Net Peak Net Peak Net Peak Use Generat. Generat.| Mined Use Generat. Generat.| Mined Use Generat. Generat.| Generat. Generat.]} Generat. Generat. MWh MWh ___MW_| ,00007__MWh MWh __MW_| .00007__MWh MWh __MW MW MW 7,470 7,989 3.14 0.0 0.00 0.0 0 0 0.00 8,325 8,904, 3.50 0.0 0.00 0.0 0 0 0.00 9,225 9,866 3.88 0.0 0.00 0.0 0 0 0.00 9,675 10,348 4.07 0.0 0.00 0.0 0 0 0.00 10,080 = 10,781 4.24 0.0 0.00 0.0 0 0 0.00 11,565 12,369 4.87 0.0 0.00 3.0 0 0 0.00 17,595 18,818 7.40 0.0 13,725 14,679 5.77 0.0 12,960 __ 13,861 5.45 0.0 . 11,250 12,032 4.73 0.0 1991} 25.0 11,250 12,032 4.73 0.0 0.00 67 2,350 2,513 1.41 0.00 35.5 2,350 2,513 1.41 0.00 30.7 2,337 2,499 1.41 0.00 24.1 2,373 2,538 1.41 14,570 4.73 37,640 8.76 0.00 0.0 0 0 0.00 12,032 4.73} 35,098 8.76 1992} 25.0 11,250 12,032 4.73 0.0 0.00 0.0 0 0 0.00 12,032 4.73 34,818 8.71 1993} 25.0 11,250 12,032 4.73 0.0 0.00 0.0 0 0 0.00 12,032 4.73 | 34,937 8.73 1994 25.0 11,250 12,032 4.73 0.0 0.00 0.0 0 0 0.00 12,032 4.73 | 35,222 8.78 1995} 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 37,005 8.90 1996} 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 37,715 9.03 1997} 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 38,301 9.13 1998} 25.0 11,250 12,032 4.73 0.0 1999} 25.0 11,250 12,032 4.73 0.0 2000} 25.0 11,250 12,032 4.73 0.0 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 38,812 9.22 0.00 17.0 1,020 1,091 0.57 13,123 4.73} 39,273 9.30 0.00 17.0 1,020 1,091 0.57 13,123 4.73 39,742 9.38 0.00 17.0 1,020 1,091 0.57 13,123 4.73 39,763 9.38 OmHAANAM-ORV 2002} 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 39,991 9.42 2003} 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 || 40,278 9.47 2004} 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 40,716 9.55 2005 | + 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 |] 41,242 9.64 2006 | 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73} 41,926 9.76 2007 | =25.0 11,250 12,032 4.73 0.0 2008 | 25.0 11,250 12,032 4.73 0.0 2009} 25.0 11,250 12,032 4.73 0.0 2010} 25.0 11,250 12,032 4.73 0.0 2011] 25.0 11,250 12,032 4.73 0.0 2012 25.0 11,250 12,032 4.73 0.0 2013} 25.0 11,250 12,032 4.73 0.0 0.00 170 1,020 1,091 0.57 13,123 4.73 42,668 9.89 0.00 17.0 1,020 1,091 0.57 13,123 4.73 43,457 10.03 0.00 17.0 1,020 1,091 0.57 13,123 4.73 }} 43,898 10.10 0.00 17.0 1,020 1,091 0.57 13,123 4.73 44,137 10.15 0.00 17.0 1,020 1,091 0.57 13,123 4.73 44,617 10.23 0.00 17.0 1,020 1,091 0.57 13,123 4.73 45,103 10.31 0.00 17.0 1.020 1,091 0.57 13,123 4.73 45,598 10.40 ecoooeooeoosoosoSooSoSeooeo ooo oOSooojeoSsoSoSoSoSoOSoSoS cecooococoosooosoooeoeoosoooeoSo ooo ScjeoosoooSooSoSo 2014 25.0 11,250 12,032 4.73 0.0 0.00 17.0 1,020 1,091 0.57 13,123 4.73 46,101 10.49 1989 - 2014 Annual Average Growth Rates: 0.6% 06% 0.6% 0.6% 2.3% -3.3%_ _—-3.3% __—-3.6% 0.9% 0.6% 0.6% 0.5% sTejO] pue speoyT HbuTutW ‘Sz TNsey 4SedetTog esed PTW - ET°E FTQeL ss 922 Residential People/ Custome 3.30 Use per kWh 5,463 Non-Mining Forecast Case: Low Commerical, Commun. Facil., and Street Lighting Use per Total Employee Sales kWh MWh Non-Mining Totals T&D T&D Net Load Peak Losses Losses Generat. Factor Generat, % MWh Indoor I < S 1982} 3,071 1,060 2.90 5,588 5,923 1,190 8,398 9,997 T 1983] 3,102 1,244 2.49 5,366 6,676 1,252 8,493 10,630 O 1984] 3,146 1,294 2.43 5,459 7,065 1,258 8,922 11,222 R 19859 3,236 1,290 2.51 5,394 6,957 1,254 9.454 11.851 1 19869 3,208 1,280 2.51 5.640 7219 1,297 10,002 12.969 C 1987] 3,306 1,331 2.48 5,844 7,778 1,204 10,554 12,703 A 1988] 3,403 1,344 2.53 6.015 8,084 1,265 9.673 L 3,499 1,365 2.56 5812 7,934 1,328 10,326 : 3,540 1,384 2.56 5.786 = 8,011 1,306 10,305 13.454 | 6.5% 1,491 22,956 60.0% 4.36 3,427 1,344 2.55 5,760 7,741 1.227 10,284 = 12,618 | 6.5% 1,415 21,773 60.0% 4.14 3,364 1,322 2.54 5,734 7,582 1,208 10,264 12,403 6.5% 1,389 21,373 = 60.0% 4.06 3,403 1,341 2.54 5,709 7,654 1,250 10,243 12,809 | 6.5% 1,422 21,885 60.0% 4.16 3,473 1,372 2.53 5,683 7,797 1,279 10,223. 13,075 | 6.5% 1,450 22,322 60.0% 4.24 3,559 1,409 2.53 5,657 7,973 1,304 10,202 13,307 | 6.5% 1,479 22,759 60.0% 4.33 3,616 1,436 2.52 5,632 8,085 1311 10,182 13,345 | 6.5% 1,489 22,919 60.0% 4% 3,662 1,458 2.51 5,606 = 8,172 1311 10,162 13,318 | 6.5% 1,493 22,984 60.0% 437 3,725 1,486 2.51 5,581 8.296 1,331 10,141 13,499 6.5% 1,514 23,309 = 60.0% 4.43 3,793 1,517 2.50 5,556 8,431 1,349 10,121 = 13,649 | 6.5% 1,534 23,614 60.0% 4.49 3,882 1,557 2.49 5,531 8,610 1,375 10,101 = 13,888 | 6.5% 1,563 24,062 60.0% 4.57 3,935 1,582 2.49 5,506 = 8,711 1,371 10,081 13,816 6.5% 1,565 24,092 60.0% 4.58 3,994 1,610 2.48 5,481 8,823 1,382 10,060 13,904 6.5% 1,579 24,06 = 60.0% 4.62 4,059 1,640 2.48 5.457 8,948 1,396 10,040 14,018 | 6.5% 1,596 = 24,561 60.0% 4.67 4,134 1,674 2.47 5,432 9,096 1,418 10,020 14.210 | 6.5% 1,619 24,925 60.0% 4.74 4,208 1,709 2.46 5,408 9,240 1,432 10,000 14,317 6.5% 1,637 25,194 60.0% 4.79 4,286 1,745 2.46 5,383 9,393 1,451 9,980 14,478 | 6.5% 1,659 = 25,530 60.0% 4.85 4,368 1,783 2.45 5,359 9,554 1,470 9,960 14,645 6.5% 1,681 25,880 60.0% 4.92 4,446 1,819 2.44 5,335 9,704 1,482 9,940 14,733 6.5% 1,698 26,135 0.0% 497 4,517 1,852 2.44 5.311 9,839 1,488 9,920 14,762 | 6.5% 1,709 26,309 60.0% 5.00 4,589 1,887 2.43 5.287 = 9,975 1,499 9,901 14.841 | 6.5% 1,724 26,541 60.0% 5.05 4.666 1,923 2.43 5,263 10,123 1,512 9.881 14,940 6.5% 1,741 26,804 60.0% 5.10 4,745 1,961 2.42 5,240 10,273 1,525 9,861 15,040 6.5% 1,759 27,071 60.0% 5.15 4.825 1,999 2.41 5.216 10,425 1,538 9.841 15,140 | 6.5% 1,776 27,341 60.0% 4,906 2,037 241 5,193 10,580 1,552 9,822 isa | 6.5% 1,794 27,614 60.0% 1989 - 2014 Annual Average Growth Rates: 14% 1.6% —-0.25% _-0.45% 1.2% 0.6% 0.2% 04% 0.1% 0.6% 0.7% 0.2% Speoy HuTUTH-UON ‘S3ZINSeY YsedeIOT aseD MOT - FI°E OTQeL 9g - € Mining Forecast and Totals Case: Low Onshore Gold, Summer Only Onshore Gold, Year-Round Offshore Gold, Shore Elect. Req't Mining Total ining + Non-Mining Total Net Peak Gold Total Net Peak Gold Onshore Net Peak Net Peak Net Peak Generat. Generat.| Mined Use Generat. Generat.| Mined Use — Generat. Generat.}| Gencrat. Generat.| Generat. Generat. MWh___ MW_|_,00007_MWh__MWh__MW_| _,0000z__MWh_ _MWh__ MW__ J MW MWh MW Pa-FOnae c 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0.00 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 eooscooooosoooseeocoeosesoosoceoesosoleocooosooSosS eococooecoeooeoeoeooscoeoocoosooeososooso ecooooocooeocoosooooccooeoooooooso 1989 - 2014 Annual Average Growth Rates: “34% -34% ——-3A% _—-3. -100.0% -100.0% -100.0% -100.0% STejOL pue speoy HuTUTW ‘SzINSeYyY ySeoeTOT |sed MOT - GTI°€ eTqeL Zs Non-Mining Forecast Case: High Residential Commerical, Commun. Facil., Non-Mining Totals and Street Lighting Use per Useper Total | T&D T&D Net Load People/ Custome § Indoor Employee Sales | Losses Losses Generat. . Customers Cust kWh Employe kWh MWh % MWh % if - 2.56 5,822 8,262 3,640 1,427 2.55 5831 8,322 3,793 1,491 2.54 5,840 8,706 3,913 1,542 2.54 5.850 9,018 3,966 1,567 2.53 5.859 = 9,179 4,000 1,584 2.53 5,868 9,295 4,028 1,599 2.52 5,878 9,400 4,059 1,615 2.51 5,887 9,511 4,090 1,632 2.51 5,897 9,622 4317 1,727 2.50 5,906 10,198 4,466 1,791 2.49 5,916 10,595 4,570 1,837 2.49 5,925 10,884 4,634 1,868 2.48 5,935 11,084 4,703 1,900 2.48 5,944 11,293 4,813 1,949 2.47 5,954 11,605 4,940 2,006 2.46 5,963 11,961 5,097 2,075 2.46 5,973 12,392 5,272 2,152 2.45 5,982 12,871 5,461 2,234 2.44 5,992 13,387 5,581 2,289 2.44 6,001 13,738 5,670 2,331 2.43 6011 14,014 5,807 2,394 2.43 6021 14411 5,947 2,458 2.42 6,030 14,820 6,091 2,523 241 6,040 = 15,240 6,238 2,591 241 6,049 15,672 1989 - 2014 Annual Average Growth Rates: 2.3% 2.6% __-0.25% 0.16% 2.8% 1.7% 1.0% 28% _-01% 2.7% 2.7% 0.2% speoy 5uTUuTW-UON ‘SR[NSeY Ysedet0g, eased YHTH - OT°E€ eTqeL Mining Forecast and Totals Case: High Onshore Gold, Summer Only Onshore Gold, Year-Round Offshore Gold, Shore Elect. Req’t Mining Total Mining + Non-Mining Total Gold Net Peak Gold Total Net Peak Gold Onshore Net Peak Net Peak Net Peak Mined Generat. Generat.| Mined Use Generat. Generat. Use Generat. Generat.f, Generat. Generat.) Generat. Generat. 000 oz MW_| 00002 MWh MWh | 00007 __ MWh___ MWh Mw_] MW MWh MW sclocoooooooo eloooooocoeooso 2,373 2,538 0 0 1,020 1,091 2,040 2,182 2,040 2,182 2,040 2,182 2.040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2.040 2,182 2,040 2,182 2,040 2,182 2.040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 2,040 2,182 1989 - 2014 Annual Average Growth Rates: 0.4% 0.4% 04% 0AM 0.4% 0.5% 05% 0.9% 2.4% 14% 2.6% 2.1% o o omAaam-OnW STe}IOL pue speoT HbuTuTW ‘SATNSeY YSeoSIOY oseD YHTH - LI°E€ aeTqeL 3.10 References Anonymous. March 20, 1989. "Northern Gold: Northern Gold Announces Exploration Continues at Nome." Business Wire. Anonymous. May 14, 1990. “Aspen Exploration: Aspen Exploration Announces Agreement with Tenneco Minerals." Business Wire. Berry, Kathi. June 1989. "Growing Pains Plague Placer Miners." Alaska Business Monthly, page 50. Bundtzen, T. K.; Swainbank, R. C.; Deagen, J. R.; and Moore, J. L. 1990. Alaska’s Mineral Industry. Alaska Division of Geological & Geophysical Surveys. Colt, Steve. 1989. Forecast of Electricity Demand in the Alaska Railbelt Region. Prepared for Alaska Power Authority (now Alaska Energy Authority). University of Alaska Institute of Social and Economic Research. Goldsmith, Scott. 1990. Economic Projections for Alaska 1988- 2010. Prepared for Alaska Housing Finance Corporation. Anchorage: Institute of Social and Economic Research. Griffin, James M., and Steele, Henry B. 1980. Energy Economics and Policy. New York: Academic Press, Inc. Henriques, Diana. October 12, 1987. "Prospects for Gold: Here’s How Timothy Green Assays Them." Barron’s, page 13. Impact Assessment, Inc. 1987. Institutional Change in Nome 1980- 1986. Social and Economic Studies Program Technical Report no. 127. Anchorage: Minerals Management Service. Knapp, Gunnar. 1990. Economic and Demographic Systems Analysis: Nome, Alaska. Social and Economic Studies Program Technical Report no. 144. Anchorage: Minerals Management Service. Minerals Management Service. 1990. Special Tabulations of Alaska Department .of Labor Employment Security Reports ES-202, 1980- 1989. Provided by Kevin Banks. National Academy of Sciences. 1979. Alternative Energy Demand Futures to 2010. Special Report, Washington D.C. National Academy of Sciences. 1979. US Energy Supply Prospects to 2010. Special Report, Washington D.C. Oatman, E. N. and Talbert, T. L. 1989. Assessing Supply and Demand Uncertainties. Prepared for Electric Power Research Institute, Report EPRI P-6369. Palo Alto, California. Se oo Peterson, L. A.; Todd, S. K.; Weddleton, J. A.; Hanneman, K. L. 1986. The Role of Placer Mining in the Alaska Economy. Conducted for State of Alaska Department of Commerce and Economic Development, Office of Mineral Development by L. A. Peterson & Associates. Richardson, Jeffery. September 1990. "Promise and Peril Mark the Start of a Boom Decade." Alaska Business Monthly, page 40. Ross, Marc H., and Williams, Robert H. 1981. Our Energy: Regaining Control. New York: McGraw Hill Book Company. Waring, Kevin and Associates. 1988. A Demographic and Employment Analysis of Selected Alaska Rural Communities, Volume II. Social and Economic Studies Program Technical Report no. 137. Anchorage: Minerals Management Service. Waring, Kevin and Associates. 1989. Nome Sociocultural Monitoring Study. Social and Economic Studies Program Technical Report no. 131. Anchorage: Minerals Management Service. Welling, Kathryn M. October 24, 1988. "The Prospects for Gold: Will the Precious Metal’s Glitter Ever Come Back?" Barron’s, page 8. White, Bill. September 20, 1990. "Owners Shut Down Bima Gold Dredge." Anchorage Daily News, page D-1. Yang, Xi Wei. 1989. "Regression Analysis of Rural Alaska Electric Demand." University of Alaska Institute of Social and Economic Research, unpublished spreadsheet. 3 - 60