Margaret Harper, Colin Sheppard, Charles Chamberlin, and Arne Jacobson Schatz Energy Research Center, Humboldt State University

Transcription

1 Ground Truth Analysis of Vermont s Residential Sector aeci Trend: Understanding Lack of Progress in Reducing per Capita Energy Consumption despite Aggressive Energy Efficiency Initiatives Authors: Margaret Harper, Colin Sheppard, Charles Chamberlin, and Arne Jacobson Schatz Energy Research Center, Humboldt State University Project Managers: Yerina Mugica and Dale Bryk Center for Market Innovation, Natural Resources Defense Council May 2010

2 This report is one in a series of short ground truth efforts that seek to understand historical residential sector energy consumption trends observed in state specific simulations using the proposed Performance based State Efficiency Program (PSEP) metric. 1,2 Initial simulations of the PSEP metric used historical data to estimate the number of years that the state would have made progress with respect to weather adjusted energy consumption intensity (aeci) in the residential sector, where progress is defined as a downward slope over the five year period that ends in the evaluation year. Over the period from , Vermont would have performed moderately well, with six progress years detected according to the PSEP methodology (Figure 1). Vermont has a long record of energy conservation policies dating back to the early 1970s. Key programs include a statewide building energy code, a low-income weatherization program, appliance standards, seasonal electricity pricing, and aggressive utility DSM programs. These earlier efforts were followed in 2000 by the creation of Efficiency Vermont, a statewide organization dedicated to energy efficiency. 3 Over the past seven years, Vermont has spent more money per capita on energy efficiency than any other state in the nation. Despite these proactive efforts, Vermont s per capita energy use has remained relatively constant since the mid-1980s. The following analysis investigates this counter-intuitive relationship between Vermont s aggressive energy efficiency efforts and the state s recent performance in the PSEP metric. The findings indicate that the estimated electricity savings associated with Vermont s energy efficiency efforts have been offset by increasing electricity consumption, and that the efficiency programs have done little in recent years to address the use of fuel oil, which represents the largest component of residential sector energy consumption. This analysis concludes that Vermont s lack of demonstrated aggregate-level progress in reducing energy consumption in the residential sector is due to: 1) the increasing use of air conditioning and other electric appliances, 2) a shift in the state s population toward older residents, who statistically tend to use more energy than younger residents, and 3) a lack of efficiency improvement in the use of fuel oil. 1 Exploring Strategies for Implementing a Performance Based State Efficiency Program: State Energy Consumption Metrics Residential Sector Analyses, Colin Sheppard, Charles Chamberlin and Arne Jacobson, Schatz Energy Research Center, Humboldt State University in collaboration with Yerina Mugica and Rick Duke, Center for Market Innovation, Natural Resources Defense Council. Released: May 15, For additional information on the PSEP metric: SERC webpage: NRDC webpage: 2 Unless otherwise stated, the energy data used in this report are from the Energy Information Agency of the U.S. Department of Energy s State Energy Data System (SEDS). 3 For further information on Efficiency Vermont visit: Page 2 of 18

3 Page 3 of Figure 1. Map displaying the number of years from that each state made progress in reducing aeci. Analysis of Vermont s Weather Adjusted Energy Consumption Intensity (aeci) Vermont s aeci has remained relatively constant over the period of , fluctuating around 50 MBtu/cap/year (Figure 2). All six progress years occurred before In 1999 and 2003, the five-year slope of aeci was slightly negative, but it did not pass the 80% significance test.

4 Figure 2. Adjusted ECI for Vermont from (above) and Slope of aeci vs. Year with single tailed 80% confidence intervals shown (below). In the upper graph, years without progress are indicated with gray diamonds, years with progress that is not statistically significant are indicated with blue squares, and years with statistically significant progress are indicated with green circles. In the lower graph, a progress year is indicated when the upper bar of the 80% confidence interval falls below zero. The analysis of aeci trends presented in the figure uses state-specific moving average heat rates as described in the PSEP revised methods. ENERGY EFFICIENCY INITIATIVES Vermont s energy efficiency initiatives in the residential sector started in the 1970s with Act 250, the Vermont Land Use and Development Act. This act mandated the consideration of energy conservation in the development of subdivisions; later amendments and interpretations of Act 250 required that developers incorporate the best available technology for efficient use or recovery of energy and required that new Page 4 of 18

5 subdivisions not put undue burden on utilities. 4 Under this unofficial building energy code, highly insulated 2 x 6 construction became standard practice. 5 Starting in 1997, an official statewide building energy code clarified the recommendations of Act 250 by setting baseline standards for inclusion of energy efficiency measures in residential construction. 6 Additionally, in 1976 Vermont started a low-income weatherization program that was expanded in 1990 with the establishment of the Vermont Weatherization Trust Fund to increase the capacity of the low-income weatherization program. 7 Figure 3 shows the interplay between Vermont s per capita energy consumption distributed by fuel type and the implementation of statewide energy conservation measures. 1970s: Act 250 Mandated Energy Conservation in New Developments 1980s: Seasonal Electricity Rates 1976: Low Income Weatherization Program Initiated 1990: Weatherization Program Expanded 1990s: Utility DSM Programs 1997: Statewide Building Energy Code Enacted 2000: Creation of Efficiency Vermont 2005: Appliance Standards and RPS Adopted Figure 3. Implementation dates of Vermont s energy efficiency policies overlaying the state s per capita energy consumption distributed by fuel type Utilities have also been integral in encouraging energy efficiency in the state. Starting in the 1980s, the Public Service Board instituted seasonal electricity pricing to help balance the winter and summer load profiles. Throughout the 1990s, individual utilities implemented aggressive demand side management (DSM) programs that resulted in substantial electricity and natural gas savings in the commercial and 4 The most recent version of the Act 250 statute: Articles 9F and 9J particularly address energy conservation in development. The original Act 250 criteria as presented by the Natural Resources Board: 5 Personal communication with Dave Lamont, Vermont Department of Public Service, April 21, For more information on the Residential Building Energy Standards: 7 For more information on the weatherization program: Page 5 of 18

6 residential sectors (Figure 4). 8 In 2000, management of all DSM programs in the state was transferred to Efficiency Vermont, a statewide efficiency utility contractually required to meet the state s Energy Efficiency Resource Standard (EERS). 9 Even more recent efficiency measures include the adoption of statewide appliance standards and a renewable energy portfolio standard (RPS) in Figure 4. Annualized Cumulative and Yearly Incremental DSM Energy Savings in all sectors, Adjustments for decay account for the energy savings loss, which occurs when DSM measures reach the end of their expected useful life. The incremental annual DSM savings represent approximately 1% of Vermont s total electricity consumption. (Source: Utility Facts 2008, Vermont Department of Public Service) 8 Vermont Department of Public Service, Utility Facts, July 2008, downloaded April 2010: 9 ACEEE, State Energy Efficiency Resource Standard (EERS) Activity, November 2008, downloaded March, 2010: Page 6 of 18

7 FUEL OIL ANALYSIS Distributing Vermont s energy use by fuel type shows that petroleum products dominate the state s energy profile (Figure 3). When this energy consumption distributed by fuel type is compared to the adjusted ECI displayed in Figure 2, nearly all of the fluctuations in the total residential energy use appear to result from fluctuations in the use of petroleum products. According to both the SEDS and the RECS databases, distillate fuel oil used for space heating makes up the majority of the home heating fuel in the state (Figure 5). Although fuel oil dominates Vermont s energy profile, the majority of Efficiency Vermont s DSM efforts are focused on electricity and natural gas consumption and do not appear to include measures aimed at reducing the use of fuel oil. As such, fuel oil heating has not decreased substantially over time. Figure 5. Types of heating fuel used in the New England region from 1993 to 2005 (Data source: RECS) Building Codes and Weatherization Previously, both Act 250 and the use of seasonal electricity rates may have played roles in determining fuel oil consumption; however, in recent years, the programs that have potential to address fuel oil consumption are the state s building energy code and the low-income weatherization program. 7,10 While building codes can be very effective at achieving reductions in heating fuel use, Vermont has a relatively low rate of new housing starts, which reduces the ability of a building code to induce near term reductions in energy consumption. 11,12 The low-income weatherization program can also result in reduced fuel consumption, but only a fraction of the state s population is eligible. The 2001 and 2007 evaluations of the impacts of the Weatherization Assistance Program suggest that the weatherization program realized annual savings of 10 Information on Vermont s Residential Building Energy Standards can be found at: 11 According to data from the National Association of Homebuilders, from Vermont saw approximately 4.3 new housing starts for every 1000 people, ranking 37 th in the nation. 12 Aroonruengsawat et al. suggest that state-wide building codes may reduce per capita residential energy consumption by as much as 5%. Aroonruengsawat, Anin, Maximilian Auffhammer and Alan Sanstad, The Impact of State Level Building Codes on Residential Electricity Consumption, University of California, Berkeley, November 25, 2009, Page 7 of 18

8 delivered fuels (distillate fuel oil, kerosene and propane) of 0.08 MBtu/cap and 0.04 MBtu/cap respectively. 13 Though promising, these reductions represent less than 1% of the state s fuel oil consumption. The Vermont Department of Public Service recognizes the need to expand efficiency efforts to include a stronger focus on heating fuels in their Comprehensive Energy Plan. Vermont s Comprehensive Energy Plan for 2009 states that that future plans include focusing on an All-fuels Efficiency Program, in addition to increasing code enforcement and strengthening appliance efficiency standards. FACTORS THAT EXPLAIN VARIABILITY IN FUEL OIL CONSUMPTION Seasonal Electricity Rates Following a substantial decline in fuel oil purchases in the late 70s and early 80s resulting from a combination of price shocks, limited availability of oil, and efficiency improvements, purchases of fuel oil have fluctuated between 29 MBtu/cap and 38 MBtu/cap. These wide fluctuations suggest that fuel oil consumption is likely still responding to pressures beyond energy efficiency improvements. Starting in the mid-80s, seasonal rates were applied to electricity sales in an effort to reduce the winter peak. These elevated winter rates inspired residents to weatherize their homes and lower their thermostats, while also encouraging residents to switch from electricity to other fuels, particularly fuel oil and natural gas, to meet their winter heating needs. Some of the fluctuation in fuel oil consumption may be due to this interplay between improved energy conservation and the potential increased use of fuel oil for heating due to fuel switching from electric heating. Fluctuations in fuel oil sales may also partly be explained by fluctuation in both weather and price. Heating Degree Days 14 Visually, the trends in the use of petroleum products and annual heating degree days seem to fluctuate in tandem; as would be expected, more fuel oil is used in years with more heating degree days (Figure 6). When a linear regression analysis is run to determine the dependency of fuel oil use on heating degree days, the correlation explains approximately 36% of the fluctuation in the fuel oil use (Figure 7). In reality, these data may be more strongly correlated. The data presented represent annual sales of fuel oil, and fuel oil, unlike electricity or natural gas, can be stored from year to year and is not monitored at the time of use. Therefore, correlations of annual data may be inaccurate. 13 Conversion factors from the SEDS Technical Notes and Documentation manual were used to convert reported savings of gallons of delivered fuel to savings in MBtu. These total estimates were then divided by the population of the given year to determine the annual per capita savings. 14 The petroleum products data used in this section of the report are not weather-adjusted. Page 8 of 18

9 Figure 6. Petroleum Products and Heating Degree Days, (Data source: SEDS and NCDC) Figure 7. Regression analysis of distillate fuel consumption and heating degree days, Price Fuel oil consumption is somewhat less responsive to the price of fuel oil between the years (Figure 8). A regression analysis suggests that a slightly negative linear relationship exists (fuel oil use decreases as price increases), though price fluctuations only explain approximately 10% of the fluctuations in fuel oil sales (Figure 9) Interestingly, when analysis was carried back further to 1970, a strong correlation was found between sales and price (a power relationship arose, which explained 79% of the fluctuation), largely due to the large changes in price and supply during the oil crises in the 70s and early 80s. Page 9 of 18

10 Figure 8. Petroleum products and the price of petroleum products Figure 9. Regression analysis of distillate fuel use versus distillate fuel price, Page 10 of 18

11 ELECTRICITY AND NATURAL GAS ANALYSIS According to Efficiency Vermont s Annual Reports from 2003 to 2007, efficiency efforts in the residential sector saved an average of 30.6 GWh of electricity per year, or approximately 0.17 MBtu/capita/year (Figure 10). 16 When Vermont s residential energy use is distributed by fuel type as shown earlier in Figure 5, it is apparent that these electricity savings did not result in an observable decline in the per capita trend for the primary energy associated with electricity consumption. Figure 10. Annualized Cumulative and Yearly Incremental DSM Energy Savings in the residential energy sector as reported by Efficiency Vermont, The Vermont Department of Public Service suggests that this phenomenon is explained by the underlying annual electricity load growth in Vermont that would be approximately 1.4% without energy efficiency efforts. 17 While this annual growth is estimated for all sectors together, in their 2005 Vermont Electric Plan, the Department of Public Service suggests similar predictions of load growth in the residential sector based on economic and demographic data from the REMI model. 18,19 A 1.4% annual load growth in the residential sector would result in an annual increase in electricity consumption of approximately 0.16 MBtu/cap/year. 20 Given the slightly increasing per capita residential electricity trend, there exists a slight discrepancy between the 0.16 MBtu/cap/year underlying load growth and the estimated savings of 0.17 MBtu/cap/year. This inconsistency may be explained by two factors. First, thanks to a funding increase in 2007, the estimated savings in 2007 are substantially higher than previous years and may therefore be inflating the average annual savings estimate. Additionally, as in other years, many of the DSM measures in 2007 were installed 16 Efficiency Vermont, Annual Report 200X, Downloaded March 2010: These estimates of energy savings do not account for primary energy associated with electricity generation. 17 Vermont Department of Public Service, Vermont Comprehensive Energy Plan 2009: Public Review Draft, May Downloaded March 2010: publicservice.vermont.gov/planning/cep%20%20web%20draft%20final% pdf 18 Vermont Department of Public Service, Vermont Electric Plan 2005, January Downloaded April 2010: p 3-11 (59 of pdf). 19 Details on the econometric modeling tool REMI can be found at: 20 The estimate of 0.16 MBtu/cap/year was derived by applying a 1.4% increase to the total residential electricity consumption each year from and then dividing by the total population. Page 11 of 18

12 in the latter half of the year, yet the savings were annualized, or accounted for as if they were installed for the full year, thus overstating the actual savings in that year. 21 This analysis suggests that Vermont s DSM efforts in the residential electricity sector are likely offset by the state s underlying load growth; the following sections examine potential sources to explain this load growth in Vermont. FACTORS THAT CONTRIBUTE TO ELECTRIC LOAD GROWTH AC Penetration A potential driver of electricity consumption is the adoption of air conditioning. Based on the regional data from the RECS surveys, there has been an increase in air conditioner penetration from 50.7% to 79.1% from 1984 to 2005 (Figure 11). Using a rough estimate, the increase in air conditioners alone would create an additional demand of 0.03 MBtu/cap/year, or approximately a 0.3% increase in electricity consumption annually. 22,23,24 These estimates suggest that the growth in the installation of air conditioners accounts for approximately one fifth of the potential load growth (0.16 MBtu/cap) estimated by the Vermont Department of Public Service. The Vermont Department of Public Service cites the increased use of air conditioners and other household appliances, along with the growth in residential use of electricity for information technology and entertainment, as causes of this increased electricity demand. 17 The growth in installations of air conditioners supports this theory; however, the relatively small amount of increased energy use attributed to air conditioning suggests that other factors must also be involved. Figure 11. Penetration of air conditioners in residential housing by region (Source: RECS) 21 Personal Communication with Riley Allen from RAP, April 12, Assume the increase in AC penetration in VT is in proportion to the Census division as a whole, with an average of 225,000 households in VT over the period from , an estimate of the number of AC units added to the VT housing stock would be 63,900, or ~3,050 per year over four years. According to the Energy Star AC Savings calculator, an AC unit with a Seasonal Energy Efficiency Ratio (SEER) of 10 (the low end of efficiency ratings for that period) would consume about 1638 kwh/year. The product of these figures divided by the average population of VT results in about 0.03 MBtu/cap increase in electricity consumption per year. 23 National Association of Counties. Data and Demographics: Number of Households. Downloaded March 2010: 24 EnergyStar.gov. Air Conditioning, Central Resources, Savings Calculator. Downloaded March 2010: Page 12 of 18

13 Age Distribution We have recently identified that a demographic shift, particularly an increase in the population s average age, can increase per capita residential energy use. This effect is outside the influence of energy policy makers, and it may therefore be reasonable to correct the ECI trend to account for demographic changes in age distribution within a state. An initial analysis indicated that the effect was small in most states and that the demographic age shifts were similar from state to state following a trend that is consistent with an aging baby boom generation. The effect was larger in a few states, however, particularly in Vermont where the increase in the average age in the population was larger than in any other state (Figure 12). These shifts in the age distribution of Vermont s population could account for an 8% increase in the aeci from Assuming the electricity load grows at the same rate as the total aeci, this underlying load growth would result in an increase in residential electricity consumption of approximately 0.05 MBtu/cap/year. This demographic driven growth, in combination with the increased use of electrical appliances, is large enough to cancel nearly half of the estimated 0.17 MBtu/cap/year reductions resulting from the efforts of Efficiency Vermont. A correction for the age distribution effect has not yet been incorporated into the PSEP method, but it may be included in a future version of the metric. Figure 12. Distribution of Vermont citizens by age (bottom); relationship between age and dimensionless ECI based on analysis of RECS 2005 data set (top right); impact of changes in the age distribution on ECI (top left) demonstrating that from ECI would have increased by ~8% due to shifts in the age of the state s population. Page 13 of 18

14 FACTORS THAT DO NOT CONTRIBUTE TO ELECTRIC LOAD GROWTH Vacation Homes Theoretically, if the number of vacation homes used by non-state residents in a state increased over time, the per capita energy consumption could increase because more people would be using energy than would be counted in the population estimate. Vermont has one of the highest percentages of vacation homes in the country, suggesting that this effect could significantly impact their energy consumption index. 25 According to the 1990 and 2000 census reports, the number of vacation homes in Vermont decreased modestly from 45,000 to 43, Unless this trend has reversed since 2000, this suggests that vacation homes are not a likely cause of significant demand growth for electricity. Grid Mix Additionally, if Vermont had had an increase in its average heat rate for electricity generation over time, this could have helped explain the lack of reduction in the primary energy associated with per capita electricity consumption despite the state s conservation efforts. However, the average heat rate for Vermont has remained nearly constant since 1975 (Figure 13). In the current PSEP metric, a heat rate of 1.0 is applied to electricity generated from renewable sources and nuclear power. Thanks to a production grid mix made nearly entirely of wood, nuclear, and hydroelectric power, Vermont s statewide heat rate used in the PSEP metric is approximately As a result of this low heat rate, residential electricity use makes up a relatively small fraction of the state s total residential energy use. 25 In 2000, 14.6% of Vermont s houses were reported as vacation homes. This is the second highest rate in the nation, topped only by Maine at 15.6%. Source: US Census, Historical Census of Housing Tables: Vacation Homes, Downloaded March 2010: 26 An analysis and calculation of this heat rate can be found in the ground-truth report on Washington State on the PSEP website: Page 14 of 18

15 VT Grid Mix Fraction of Grid by Energy Source Wind Elec. Waste to Elec. Wood to Elec. PV/Solar Thermal Elec. Petroleum to Elec. Nuclear Natural Gas to Elec. Hydro Elec. Geothermal Elec. Coal to Elec VT Heat Rate Dimensionless Heat Rate !!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!! Figure 13. Vermont s generation grid mix by fraction of total electricity production and the associated dimensionless heat rate Figure 13 displays the grid mix of Vermont s annual electricity production, which differs from the grid mix of electricity consumption in Vermont as displayed in Figure Due to interstate energy trading and international contracts with Hydro-Quebec, the consumption grid mix shows a substantial increase in hydroelectricity and decrease in nuclear electricity in comparison to Vermont s production grid mix. As discussed in a previous ground truth report on Washington State, because of the consideration of primary energy associated with electricity production, the PSEP metric is sensitive to differences in a state s heat rate, which can vary between a state s production grid mix and its consumption grid mix. In Vermont s case, because the PSEP metric does not apply a heat rate to hydroelectricity or nuclear electricity, the overall dimensionless heat rate applied to electricity production in Vermont would still be close to 1.0 regardless of whether the production or consumption grid mix were used. 27 Data on Vermont s consumption grid mix from were provided by Dave Lamont at the Vermont Department of Public Service. The System category represents electricity purchased by Vermont utilities on the market without a known grid mix or resource composition. The Renewable Energy Credits Sold (RECS) category was added in 2005 to properly account for electricity produced using renewable energy that was then sold as a renewable energy credit to another utility and should therefore not be counted as renewable energy for the producer. Page 15 of 18

16 Figure 14. Vermont s consumption grid mix by fraction of total electricity production. The associated dimensionless heat rate would be similar to that of the generation grid mix shown in Figure 13. Fuel Switching and Seasonal Rates In the 1980s, substantial decreases in electricity consumption may be attributed to the utilities application of seasonal rates to encourage fuel switching away from electric heaters. Until 2002, Vermont s winter peak demand consistently exceeded the summer peak demand (Figure 15). 8 In an effort to address the winter peak, the Public Service Board instituted higher electricity rates in the winter than in the summer. These seasonal rates led to a greater interest in both weatherization of houses and fuel switching away from electric heaters. 28 In their Utility Facts 2008 report, the Department of Public Service shows that Vermont s winter peak remained relatively constant over the period of while the summer peak increased substantially (Figure 15). 8 Once the winter and summer peak became nearly comparable, the seasonal rates were removed in favor of rates which accurately reflect utility costs and therefore can send clearer price signals, such as time of use rates Personal Communication Richard Cowart, RAP, April 9, Seasonal rates were eliminated for Vermont s second largest IOU, Green Mountain Power in 2001: and eliminated for the largest municipal utility in 2005: Page 16 of 18

17 Seasonal Peak Power (MW) Figure 15. Vermont Seasonal Peak Power for and all sectors (Source: Utility Facts 2008, Vermont Department of Public Service) Page 17 of 18

18 CONCLUSION Vermont s strong record of energy efficiency policies is not reflected by corresponding progress under the PSEP metric. Analysis suggests that this is due to the following key factors: 1. A lack of efficiency measures focused on fuel oil consumption 2. The increasing use of air conditioners and other electric appliances 3. A strong demographic shift in the state s population toward older residents, who statistically use more energy than younger residents Fuel oil Many of the state s energy efficiency policies target electricity and natural gas consumption, which are already relatively low. The use of fuel oil dominates Vermont s energy profile and fluctuates over time, but does not seem to be decreasing. Fluctuations in fuel oil use are shown to be correlated with changes in both weather and price. These correlations, along with the lack of targeted fuel efficiency and extensive weatherization programs, help to explain the unstable trend of fuel oil use. Because the trends of electricity and natural gas use are nearly flat, these fluctuations in fuel oil use determine the trend of the total aeci. Increased AC penetration The nearly flat trend for per capita primary energy associated with electricity consumption is explained in part by the likelihood that the proactive energy efficiency programs are offsetting potential load growth due to increased use of appliances, air conditioners, computers and entertainment systems. Demographic Shift The effect of demographic shifts on the aeci offers a final explanation for Vermont s lack of progress in reducing per capita energy use. In recent years, Vermont has seen a significant increase in its year old population segment, which signifies a shift that can lead to increased energy consumption. Factors that do not explain Vermont s performance Despite changes in the state s electrical grid mix, the heat rate associated with electricity generation has remained nearly constant and has not influenced Vermont s electricity consumption trend. The number of vacation homes in the state has decreased slightly over time; therefore, vacation homes do not seem to contribute to the underlying electrical load growth, nor do they have a substantial effect on the overall energy consumption trend. Page 18 of 18