The economic costs of reducing greenhouse gas emissions under a U.S. national renewable electricity mandate

Energy Policy 39 (2011) 2730–2739

Contents lists available at ScienceDirect

Energy Policy journal homepage: www.elsevier.com/locate/enpol

The economic costs of reducing greenhouse gas emissions under a U.S. national renewable electricity mandate Keith Crane a,n, Aimee E. Curtright b, David S. Ortiz b, Constantine Samaras b, Nicholas Burger a a b

RAND Corporation, 1200 South Hayes Street, Arlington, VA 22202, USA RAND Corporation, 4570 Fifth Avenue, Pittsburgh, PA 15213, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 28 July 2010 Accepted 4 February 2011

The electricity sector is the largest source of greenhouse gas emissions (GHGs) in the U.S. Many states have passed and Congress has considered Renewable Portfolio Standards (RPS), mandates that specific percentages of electricity be generated from renewable resources. We perform a technical and economic assessment and estimate the economic costs and net GHG reductions from a national 25 percent RPS by 2025 relative to coal-based electricity. This policy would reduce GHG emissions by about 670 million metric tons per year, 11 percent of 2008 U.S. emissions. The first 100 million metric tons could be abated for less than $36/metric ton. However, marginal costs climb to $50 for 300 million metric tons and to as much as $70/metric ton to fulfill the RPS. The total economic costs of such a policy are about $35 billion annually. We also examine the cost sensitivity to favorable and unfavorable technology development assumptions. We find that a 25 percent RPS would likely be an economically efficient method for utilities to substantially reduce GHG emissions only under the favorable scenario. These estimates can be compared with other approaches, including increased R&D funding for renewables or deployment of efficiency and/or other low-carbon generation technologies. & 2011 Elsevier Ltd. All rights reserved.

Keywords: Renewable portfolio standards Greenhouse gas emissions Technology development

1. Introduction The largest source of greenhouse gas (GHG) emissions in the U.S. is electric power generation, accounting for about 35 percent of total GHGs in 2008 (Environmental Protection Agency (EPA), 2010a). If U.S. GHGs are to fall substantially, emissions from generating electricity will need to considerably decline. One policy option for reducing GHG emissions is to impose a mandate, such as a Renewable Portfolio Standard (RPS), on electric power providers to generate a specified share of their electricity from renewable sources such as wind and biomass.1 A number of scholars have analyzed the economic costs of using an RPS to reduce GHG emissions and have arrived at differing conclusions. Wiser et al. (2007) concluded that RPSs yield ‘‘mixed-results.’’ Nogee et al. (2007) found RPSs generated ‘‘important economic and environmental benefits.’’ Michaels (2008) concluded they induce ‘‘inefficient’’ and in some cases ‘‘pernicious’’ outcomes. Some studies argue that using a diverse portfolio of low-carbon sources of energy could achieve targeted reductions in GHG emissions at lower cost than the more rigid technological mandates incorporated into RPSs (Dobesova et al., 2005;

n

Corresponding author. Tel.: +1 703 413 1100; fax: +1 703 413 8111. E-mail address: [email protected] (K. Crane). 1 For a review of RPS designs and a survey of early implementation, see Berry and Jaccard (2001). For an additional discussion of RPS design, see Espey (2001). 0301-4215/$ - see front matter & 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.enpol.2011.02.042

Apt et al., 2008). Chen et al. (2009) estimated the anticipated effects of an RPS on retail electricity rates, finding expected increases would be small. In this paper we estimate the net benefits from imposing a national RPS mandate in terms of reductions of GHGs and the corresponding economic costs relative to coal-based electricity.2 Following on an earlier study by Toman et al. (2008), we set a target of generating 25 percent of U.S. electricity from renewable sources by 2025.3 We identify which sources of renewable energy are most economical for achieving such a mandate. We also assess the economic costs and net GHG benefits of the 20 percent Renewable Electricity Standard (RES) included in the American Clean Energy and Security Act (ACES) (HR 2454; Waxman-Markey) passed by the House of Representatives in 2009 (ACES, 2009). The paper is organized as follows. In Section 2, we outline the methodology we use to estimate the costs and quantities of electricity generated by prospective renewable energy sources and associated reductions in GHG emissions. We also introduce scenarios to consider uncertainty in technological developments and costs. Section 3 presents details on the costs and emissions

2 Note that in this analysis we consider the production of electricity in pulverized coal (PC) plants without the use of carbon, capture, and sequestration (CCS). 3 We examine the GHG benefits and economic costs of a 25 percent mandate for renewable motor vehicle fuels by 2025 in another forthcoming paper.

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

associated with each of the renewable technologies we analyze: hydropower, wind, biomass, solar thermal, and geothermal. We present our results in Section 4. In Section 5, we discuss the implications of our analysis for recent proposals for legislation to reduce GHGs.

2731

decommissioning) of these technologies are taken into consideration. We calculate life cycle GHG emissions/kWh for each renewable technology and then subtract this total from life cycle GHG emissions of coal-fired electric power to derive a net reduction in GHG emissions for each technology. We then calculate the cost per metric ton of carbon dioxide equivalent (CO2e) from reduced GHG emissions using the cost and net emission estimates above.

2. Methods 2.3. Scenarios 2.1. Estimating costs and quantities We first estimate the additional cost of replacing electricity generated by coal-fired electric power plants with renewable sources of energy. For each source of renewable energy, we estimate the cost of producing electricity, factoring in the availability of sites or resources. All sources exhibit increasing marginal costs. The cost estimates are based on the full economic costs of substituting renewables for coal, and we do not incorporate subsidies into the analysis. Details of these estimates and underlying assumptions can be found in the Electronic Annex in the online version of this article. We use these estimates of quantities and costs of renewable fuels to construct supply curves for renewable electricity, employing the lowest cost sources first. To estimate the additional cost of renewable electricity, we compare the costs of renewables to the dispatch costs in 2009 of electricity generated by coal-fired power plants, electricity that we assume new renewable sources would replace. Then we compare the costs of renewables with the cost of additional coal-fired generating capacity that would be needed to meet the EIA’s Reference Case projection for U.S. demand for electricity in 2025.4 We chose to compare the cost of electricity generated by renewables with that generated from coal-fired power plants as one plausible scenario. As substantial amounts of variable renewable generation are added to the electricity grid, actual fossil resources displaced will depend on relative cost of fuels, local supply and demand characteristics, and other factors. In a recent National Renewable Energy Laboratory (NREL) analysis, they posit that baseload coal could be displaced with high penetration of wind, as more flexible generation assets are preferred to address modeled day-ahead forecast error (NREL, 2010). Wind or solar power could be dispatched to meet peak needs if storage technologies were available and affordable, but currently such technology is not widely used (see Electronic Annex). The comparison with coal-fired power plants is also useful because coalfired electric power generates the most GHG emissions per kWh relative to other technologies. As one purpose of the RPS is to reduce GHG emissions, the comparison with coal is pertinent.5 This analysis does not include externalities, such as health impacts from air pollution, in the calculation of the relative costs of coal-based versus renewable electricity production. These can be significant (NRC, 2010a; Barr and Dominici, 2010). 2.2. Estimating potential net reductions of GHG emissions We calculate the net reductions in life cycle GHG emissions that would result from substituting renewables for coal to generate electricity. All renewables generate a low level of GHG emissions when the full life cycle (i.e., constructing, operating, and 4

CCS would simultaneously lower emissions and raise the cost of coal-fired power, but the technology is still being developed and demonstrated. Consequently, we do not consider it in this analysis. 5 Note that we have not considered other factors in this analysis, such as potential siting difficulties, criteria pollutants, and water use, which may influence the type of technologies in the electricity generation portfolio in 2025.

To account for broad uncertainty in changes in technological options and costs, we employ three scenarios: (1) Reference Scenario; (2) Favorable Technology Scenario, where technological development is assumed to be relatively rapid; and (3) Unfavorable Technology Scenario, where technological development is assumed to be relatively slow. We develop the Favorable and Unfavorable Technology Scenarios by identifying the key components of each renewable technology that are most likely to determine the pace at which technologies improve or costs fall. We then evaluate the likely speed at which these technological advances could occur and the extent to which current expectations about improvements and cost reductions may fail to materialize. In these scenarios, we also take into account how rapidly the technologies are likely to be deployed. Detailed discussions of the assumptions in our technology analysis can be found in the Electronic Annex.

3. Costs and GHG emissions of renewable sources of electricity The electric power industry is capital-intensive. Generating plants are expensive and operate for decades. Obtaining site approvals, constructing a plant, and bringing it to full operation is a lengthy process, generally taking years to complete. Therefore, a substantial dramatic shift from generating electricity from fossil fuels to using renewable energy would take a considerable amount of time as renewable capacity is constructed. If the U.S. were to set a goal of generating 25 percent electricity from renewable sources by 2025, planning and construction of large amounts of renewable generating capacity would have to begin shortly to meet such a target. Consequently, we assume that only currently commercial renewable generating technologies would be available to meet such a mandate by 2025. Technologies that are not yet at this stage would be highly unlikely to be widely adopted at scale by that date. Accordingly, we confine our analysis to the following technologies: (1) Hydropower; (2) Wind; (3) Biomass; (4) Geothermal; and (5) Solar. Some renewable technologies, like hydropower, are mature; significant reductions in cost or improvements in productivity and efficiency are less likely. Some, like emerging solar thermal technologies, could see rapid improvements in efficiency or large reductions in cost. Other technologies, like wind power, fall inbetween. Here we briefly describe our estimates of the costs and GHG emissions for these five technologies. To calculate net GHG reductions, we compare emissions from electricity generated from renewable sources with that from baseload coal-fired power plants. Coal-fired power is assumed to have a life cycle emissions factor of 1.021 kg CO2e/kWh (Wang et al., 2007). For a more detailed discussion, see the Electronic Annex. 3.1. Hydropower: technologies, quantities and prices, and net GHG emissions There are two categories of hydroelectricity: conventional and emerging hydrokinetic. The deployment of hydrokinetic power technologies, including wave and tidal energy, is likely to remain

2732

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

modest for the next few decades (NRC, 2010b). Accordingly, we include only conventional resources in this study. In 2008, installed hydropower capacity in the U.S. was 78 GW and generated 297 TWh, about six percent of U.S. electricity output (Energy Information Administration (EIA), 2009b). Existing conventional hydropower capacity could potentially be expanded by 10 GW (Electric Power Research Institute (EPRI), 2007); some studies have been even more optimistic (Idaho National Laboratory (INL), 2006). However, because the potential to expand capacity and the costs of so doing are uncertain, primarily due to effects on ecosystems and problems with public acceptance, EIA assumes that the installed capacity of hydropower in 2025 will be similar to 2008 (EIA, 2009a). Accordingly, we assume no additional installed capacity or costs for hydropower by 2025 and no new net reductions in GHG emissions from hydropower; nevertheless, already-installed hydropower is assumed to contribute to the overall goal of a 25 percent RPS. We include an analysis where existing hydropower is excluded in Section 5.4, below. 3.2. Electricity from wind 3.2.1. Technologies and barriers to deployment Wind power is harnessed by using wind turbines to drive a generator. Over the past thirty years, wind turbine technologies have improved substantially. Currently, all utility-scale, commercially available wind generators use the same technology: three-bladed upwind turbines. Since the late 1990s, the average capacity of individual wind turbines has increased by 133 percent, but progress has leveled off in recent years (Wiser and Bollinger, 2009). Commercial utility-scale turbines are now considered a mature technology. They are widely deployed, accounting for an appreciable share of total capacity in some European countries (NRC, 2010b). Despite the technological maturity of the turbines, barriers such as transmission capacity, turbine siting, and wind integration may constrain substantial increases in onshore wind energy (NREL, 2010). Penetration of wind power in total generation of beyond 20 percent of total output may be very difficult without directly addressing the problem of the intermittent nature of wind by adding energy storage, additional gas-fired peaking capacity, or much more aggressive electric demand management programs, such as agreements with individual consumers to immediately reduce power consumption when the wind dies down (NRC, 2010b). These technical options for addressing the intermittency of wind are discussed in detail in the Electronic Annex. It is also unclear whether regulatory and public acceptance barriers to expanding transmission can be overcome to achieve larger scale penetration of wind power by 2025 (NREL, 2010). Other concerns include the effects of wind energy on bird and bat populations, land use, noise, and esthetics (NRC, 2007). Additionally in some instances, wind farms might affect regional weather and climate (Keith et al., 2004). All wind power installations in the U.S. are onshore (Department of Energy (DOE), 2008), but offshore wind has the potential to harness substantial wind resources. The technical challenges of operating in the harsh marine environment have already begun to be addressed with existing operating experience outside the U.S. (NRC, 2010b). However, the cost of offshore wind is substantially greater than on-shore wind because of higher capital and operating costs. In addition to economic and technological hurdles, social and regulatory challenges are increasingly common (DOE, 2008). 3.2.2. Quantities and costs of wind Wind power generated 52 TWh in the U.S. in 2008, from a cumulative installed capacity of 25,369 MW (EIA, 2009b; Wiser and Bollinger, 2009). The wind resource is abundant. Estimates

for total recoverable wind power vary, with most at two to five times the total annual current production of electricity of the U.S (Black and Veatch, 2007; NRC, 2010b; DOE, 2008). The total wind power capacity available at potential economically viable sites is much lower however, because turbines are not likely to be sited in urban areas or on protected lands and parks, or because of infrastructure issues and other barriers to deployment. Even conservative available estimates run at least 500,000 MW, an order of magnitude more than total installed U.S. wind power (Black and Veatch, 2007). To estimate total available wind power resources, the EIA characterizes potential sites by region using data generated by the national laboratories of the U.S. Department of Energy (EIA, 2009c). The EIA estimate takes into account the effects of remote and less favorable sites on capacity utilization, transmission requirements, and the costs of meeting environmental concerns by incorporating those costs into the capital costs (EIA, 2009c; Namovicz, 2009). EIA confines economically viable wind sites to onshore wind classes 4, 5, and 6 (i.e., mean wind speeds of 7.5, 8, and 8.8 m/s, respectively). For our Reference Scenario we determine the cost per kWh of various wind sites by estimating the costs of capital, operations and maintenance (O&M), and interconnection costs and capacity factors for onshore and offshore wind power, drawing on the Early Release of EIA’s Annual Energy Outlook (AEO) 2010 (EIA, 2009a) when data are available, and the AEO 2009 when 2010 data were not available (EIA, 2009d). Wind capital cost multipliers for each regional wind supply have been applied in accordance with the EIA’s National Energy Modeling System (EIA, 2009c). The Reference Scenario for wind power utilizes the quantities, capacity factors, capital, O&M, and interconnection costs of EIA’s AEO Reference Case for 2025. Further details on our modifications to EIA’s assumptions and the levelized cost of electricity (LCOE) calculation are provided in the wind section of the Electronic Annex. We also include the economic cost of integrating wind energy into the grid. Wind integration costs have been defined as ‘‘the incremental costs incurredythat can be attributed to the variability and uncertainty introduced by wind generation’’ (NREL, 2010). NREL generated estimates for these costs by simulating electricity production markets with correlated wind and load data, and estimating changes to unit commitments and capacity reserve requirements as compared to a case where only the load exhibits variability (NREL, 2010). NREL estimates integration costs at about $5/MWh for eastern wind integration, assuming large balancing areas and fully developed regional markets. Not captured in this method is that traditional units may be operated below their optimal design loads, decreasing efficiency and revenues. An assumed economic intermittency charge would also include the additional costs due to the provision of regulatory services, decreased revenues from traditional generators, and any market changes or programs required to sustain critical market services (NREL, 2010). A similar recent estimate by the Nebraska Power Association found integration costs below $3/MWh for their base case and above $9/MWh under their highest penetration case, including the costs of making it possible to export wind power when penetration rates reach 20 percent or more (Nebraska Power Association (NPA), 2009). Our Reference Scenario assumes a wind integration cost of $10/MWh to account for the economic costs associated with maintaining supply and demand in the electricity market and additional costs from unit cycling, derating, and decommitting. While individual fossil generators are typically not assigned to provide fill-in power for specific variable renewables, this $10/MWh charge is financially equivalent to: (1) backing up 15 percent of wind generation with natural gas (NG) combustion

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

turbines, (2) backing up 20 percent of generation with natural gas combined cycle (NGCC) generation, or (3) backing up 25 percent of generation with coal generation (see Electronic Annex). For comparison, a Nebraska Power Association study estimated that natural gas use would increase by 19 percent (from about 1,680 GWh to about 2,000 GWh) with 40 percent wind penetration (NPA, 2009). The Favorable Technology Scenario for wind utilizes the quantities and interconnection costs of EIA’s AEO Reference Case for 2025. Taking into account potential improvements in costs and performance, capital and O&M costs are reduced relative to the EIA Reference Case. Taking into account potential improvements in cost and performance, capital and O&M costs are reduced relative to the EIA reference case. We assume relatively higher capacity factors, following Black and Veatch (2007). Integration costs are assumed to be $5/MWh, consistent with NREL’s eastern wind integration scenario which assumes large balancing areas and fully developed regional markets (NREL, 2010). We assume these lower integration costs include all economic costs, with the lower total integration costs resulting from advanced grid technologies facilitating integration. This $5/ MWh charge is financially equivalent to backing up 5 percent of wind generation with NG combustion turbines, 10 percent of wind generation with NGCC generation, or 10 percent of wind generation with coal. The Unfavorable Technology Scenario for wind power uses the quantities, capacity factors, and interconnection costs of EIA’s 2010 AEO Reference Case for 2025. We also draw capital and O&M costs from the 2010 AEO High Cost Renewable Case for 2025. In the Unfavorable Technology Scenario, wind output is reduced by 10 percent to account for curtailment when wind power exceeds demand during off-peak times (NREL, 2010). In this scenario, integration costs are increased to $20/MWh. This $20/MWh charge is financially equivalent to backing up 25 percent of wind generation with NG combustion turbines, 35 percent of wind generation with NGCC generation, or 45 percent of wind generation with coal. 3.2.3. Potential net reductions in GHG emissions with wind power While wind power does not directly emit GHGs when generating electricity, small amounts of GHGs are produced during materials extraction, fabrication, installation and maintenance of wind turbines. Turbine production and construction account for most turbine life cycle emissions (72–90 percent); transportation, O&M, and decommissioning account for the remainder (Weisser, 2007). Since life cycle GHG emissions for wind are estimated per kWh over the turbine’s lifetime, better located turbines will have higher annual energy output and hence lower life cycle GHGs than similar turbines located at sites with poorer quality wind. Estimates of full life cycle GHG emissions range from 2 to 30 g/kWh (NRC, 2010b; Weisser, 2007). The lowest values represent wind farms with 50 or more 500 kW turbines located in good sites, while the highest values represent turbines with low capacity factors due to poor wind resources or older wind technologies (NRC, 2010b). Given the range of life cycle GHG estimates from wind turbines, and the important role of local conditions and uncertain turbine lifetimes, we assume a life cycle value of 15 g CO2e/kWh for wind power in this analysis (see Electronic Annex). For this analysis, we assume that wind (and associated measures to compensate for intermittency) will displace coal generation. Because dispatch is scheduled 24 hours prior to delivery, wind will need to be dispatched in conjunction with these other compensating measures. As noted above, natural gas turbines are likely to be used as operating reserves to accommodate wind variability and uncertainty (NREL, 2010).

2733

Katzenstein and Apt (2009) modeled the effects of fill-in power from natural gas generation and inefficiencies due to generator ramping. They found that wind variability reduces expected GHG benefits by 20–25 percent. If variable renewables are adopted extensively nationwide the system could also be partially balanced by demand response, storage, and other methods. We bound our analysis to include emissions associated with system balancing with an additional 40 g CO2e/kWh of wind power for the reference scenario and substantially lower and higher values in the favorable and unfavorable technology scenarios, respectively (see Electronic Annex). 3.3. Electricity from biomass 3.3.1. Technologies and barriers to deployment Current technologies for generating electricity from biomass involve either: (1) co-firing with coal in existing power plants or (2) burning biomass in a dedicated generating plant. The latter may entail combustion in stoker or fluidized bed boilers to produce steam. For coal, both technologies are mature. However, incremental improvements in designs and operations might improve the efficiency of using biomass. The EIA assumes that future electricity from dedicated biomass facilities will come from biomass-integrated gasification combined cycle (BIGCC) plants. A BIGCC facility has significantly improved efficiency when compared to conventional combustion facilities. However, no commercial facilities employing this technology exist today. Consequently, we assume that these plants will not be widely available commercially by 2025 and therefore do not incorporate them into this analysis. A few coal-fired power plants are experimenting with co-firing biomass. These operators have found that biomass fractions of up to 10 percent on an energy basis have almost no effect on boiler operations or emissions controls. However, plant-site modifications may be necessary, mostly related to biomass handling and processing (Van Loo and Koppejan, 2008). The principal challenge to co-firing biomass is handling and processing in preparation for combustion. Plants need to purchase, install, operate, and maintain dedicated receiving and handling equipment for biomass. In addition, the lower energy and bulk densities and increased moisture content of biomass relative to coal will influence handling and use. There are few commercial-scale electric power plants in the U.S. dedicated to burning biomass. Experience from dedicated biomass facilities has found that biomass may corrode boiler components and foul emissions control equipment, resulting in increased maintenance or higher costs for corrosion-resistant boiler components (Peltier, 2009). Herbaceous biomass tends to contain larger quantities of alkali metals and other potentially corrosive elements, especially chlorine, which can damage boiler components. Storing bales of herbaceous biomass for several months reduces the concentration of these contaminants. Clean, green wood typically does not pose these problems (McGowan, 2009). On balance, technical issues of biomass utilization should not prove to be a significant barrier to using biomass on a large scale to generate electricity. 3.3.2. Quantities and costs of biomass To conduct our analysis, we use modified estimates of available biomass resources and prices from the EIA’s 2009 AEO (LaRiviere, 2009). We use these data to create a biomass supply curve for the non-captive biomass market, i.e., biomass not already tied to a specific co-generating facility or heating plant (EIA, 2009b). EIA creates separate supply curves for four categories of biomass (1) urban wood and mill waste, (2) forestry

2734

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

residues, (3) agricultural residues, and (4) energy crops. We generate separate supply curves for each of these sources of biomass for each of the 14 coal regions in the U.S., assuming that the electric power sector can burn all four biomass types in each region. The EIA resource base estimate indicates that the amount of available urban wood waste and forestry residues is constant over the forecast period, i.e., until 2035 (LaRiviere, 2009). Cost reductions over this period for this type of biomass are assumed to be modest, since woody biomass resources (typically wastes and residues) are already available and collection systems are mature. However, since urban wood wastes, mill residues, and forestry residues are byproducts, the supply of these products face hard upper limits. Additional biomass supplies come from energy crops. EIA assumes that the amount of agricultural residues and energy crops increases over time due to improvements in collection, yields, and crop management. Most of these cost reductions in the supply of agricultural residues and energy crops are assumed to take place by 2015. A recent National Research Council study estimated the potential size of the biomass energy resource base in the U.S. (NRC, 2009). Similar to EIA, the study assumes that biomass energy resources derived from forestry residues, animal manure, and municipal solid waste are fixed. However, the study imposed two additional constraints when deriving their estimate of biomass resources from agricultural lands, namely (1) no use of land that is already producing ‘‘food, feed, or fiber’’ and (2) no greater impact on the environment than the current use of the land. These assumptions exclude existing cropland and permanent pasture from the production of biomass energy crops and limit the quantity of agricultural residues that can be sustainably collected. Under these constraints, the U.S. could potentially produce 400 million tons of dry biomass per year using today’s technologies and agricultural practices. Most of this would come from corn residue, woody residue, dedicated energy crops, and municipal solid waste. Changes in agricultural practices and improved technologies could raise this figure to 550 million dry tons per year by 2020, mostly due to improvements in the yield of energy crops. For the Reference Scenario, we construct an aggregate herbaceous and woody biomass marginal supply curve for the entire U.S. The first 39 TWh of electricity is provided from existing biomass capacity and sources; we assume current biomass generation does not affect our estimates of the economic costs from using additional sources of biomass as a result of a renewable mandate. The next 16 TWh represent electricity generated from co-firing biomass with coal. The remaining generation represents dedicated biomass facilities. The marginal cost of generating electricity by co-firing in existing facilities is estimated at 4.7 cents/kWh in the Reference Scenario. This estimate assumes a 10 percent reduction in capital costs of biomass handling and processing equipment the experience of the Chariton Valley Biomass Project (Antares, 2009), which we take to be representative of existing experience (see Electronic Annex). With the exception of a small amount of low-cost biomass, the cost of generating electricity in dedicated facilities ranges from 7.8 to 28.8 cents/kWh. As demand for additional biomass rises, prices increase steeply due to increased biomass supply costs. In the Favorable Technology Scenario, we assume that supplies of herbaceous biomass are more plentiful and cheaper than in the Reference Scenario because of advances in agricultural technologies. This change is captured by assuming prices and quantities projected by the EIA for 2035 are available in 2025. In the Unfavorable Technology Scenario, we assume that costs drop more slowly than in the Reference Scenario and quantities

of herbaceous biomass expand less rapidly than expected, reaching cost and quantity levels projected by EIA for 2015 only in 2025. The total amount of electricity that could be generated from biomass, which is constrained by the total amount of biomass available as estimated by the NRC, is approximately 450 TWh in the Unfavorable Technology Scenario and 970 TWh in the Favorable Technology Scenario. 3.3.3. Potential net reductions in GHG emissions with biomass To estimate GHG emissions from producing electricity from biomass, we use the Calculating Uncertainty in Biomass Emissions (CUBE) model, Version 1.0, (National Energy Technology Laboratory (NETL), 2010). It is described in more detail in the biomass section of the Electronic Annex. For co-fired electricity, the cost of reducing GHG emissions ranges from $16 to $22/metric ton of CO2e. However, the cost of reducing GHG emissions from dedicated biomass generating capacity rises sharply with increasing quantities. In the Reference Scenario, approximately 720 million metric tons of CO2e would be saved at prices less than $100/metric ton. In the Favorable Technology Scenario, biomass-generated power could reduce approximately 900 million metric tons of CO2e at prices less than $100/metric ton, while in the Unfavorable Technology Scenario, approximately 460 million metric tons of CO2e would be saved over coal at prices less than $100/metric ton. 3.4. Electricity from geothermal 3.4.1. Technologies and barriers to deployment Geothermal technologies use the heat in naturally occurring reservoirs of steam, hot water, or hot rocks in the Earth’s crust to generate electricity. Utility-scale geothermal technologies convert this heat into steam to run a turbine in much the same way that fossil fuels generate electricity. Conventional geothermal power generation uses hydrothermal resources, which are reservoirs of hot water (i.e., above 100–150 1C) and steam trapped in permeable rocks at depths of up to 3 km. Capacity factors for geothermal generating plants average around 90 percent, making this a good technology for baseload power. Hydrothermal is a relatively mature technology. Accordingly, the development of conventional resources is not constrained by major technological barriers but rather by physical availability of the resource. Technological advances could increase efficiencies and reduce costs, but marginal advances are unlikely to have a major impact on the rate and level of deployment or on costs. Advances in exploration and resource assessment could potentially expand the number of sites available. Because traditional geothermal resources are often located far away from populated areas, advances in transmission technologies and reduced transmission costs would improve the prospects of further exploitation of this resource (NRC, 2010b). Advanced geothermal technologies, such as enhanced (or engineered) geothermal systems (EGS), take advantage of a vast resource of heat available at many sites around the world. However, technological barriers and unfavorable economics may limit or preclude the near-term development of these sites. Potential seismic activity induced from developing these sites is a regulatory and public acceptance concern (NRC, 2010b; Patel, 2009). Consequently, we assume that these technologies will not be available at commercial scale by 2025. 3.4.2. Quantities and prices of geothermal electricity Electricity generation costs from conventional hydrothermal technologies are relatively competitive, but the resource is

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

geographically concentrated in the western U.S. and limited in quantity. A recent study of geothermal resources by the Western Governors’ Association (WGA) estimates that approximately 10 percent of the western U.S.’s electricity needs could be met with known hydrothermal resources, but less than half of that capacity was deemed viable for commercialization by 2015. In absolute terms, 13 GW of geothermal energy is available and could be developed for under $0.20/kWh; 5.6 GW of this should be commercially viable by 2015, with costs under $0.10/kWh (WGA, 2006). These resource estimates are consistent with others, including a more recent U.S. Geological Survey estimate (U.S. Geological Survey (USGS), 2008) and are widely cited as credible (NRC, 2010b). The EIA combined the site-specific data in the WGA analysis with a more detailed but geographically limited analysis from GeothermEx (2004) to construct a supply curve consisting of 88 hydrothermal sites, each with unique capital and fixed O&M costs. Based on these data and under EIA’s LCOE estimates, geothermal contributes just under 9 GW of electricity for under $0.12/kWh by 2030 (EIA, 2009c, 2009d; Smith, 2009). EIA does not include estimates of undiscovered hydrothermal or EGS resources. We therefore implicitly assume in our analysis that neither of these will contribute to electricity supply under our Reference Scenario. This is consistent with other recent analyses (NRC, 2010b; Massachusetts Institute of Technology (MIT), 2006). The Reference Scenario in our analysis relies heavily on EIA’s input values. Geothermal capital costs are based on the overnight capital cost values from NEMS for the Reference Case in 2025, scaled to relative site-specific costs. Our expected quantities for geothermal electricity production are also based primarily on EIA. Where the information was already available, assumptions were taken from the early release of AEO 2010 (EIA, 2009a). When unavailable, or where AEO 2010 does not differ from the 2009 version of the model, we use older AEO assumptions (EIA, 2009c, 2009d; Smith, 2009). In our Favorable Technology Scenario, we assume a modest increase in the resource base relative to NEMS stemming from technological advances increasing the supply of viable sites for hydrothermal geothermal electricity. In the absence of site-specific cost data for this theoretical expansion, we increase the available capacity at each individual site to modify the current cost curve proportionately at the various cost levels. Because of the relative maturity of hydrothermal technologies, even favorable technological developments are unlikely to have a dramatic effect on the costs of conventional technologies. However, such developments combined with a 25 percent mandate for the use of renewables would likely lead to modest reductions in costs beyond what EIA assumes due to increased learning and field experience. As such, assumptions about capital and O&M costs for our Favorable Technology Scenario are less than EIA’s. The Unfavorable Technology Scenario differs from the Reference Scenario in that capital costs are based on AEO values for the High Cost Renewable Case, scaled to the relative site-specific costs assumed by EIA. O&M costs are the same as in the Reference Scenario. The geothermal section of the Electronic Annex contains further details and assumptions.

3.4.3. Potential net reductions in GHG emissions The full life cycle GHG emissions associated with geothermal electricity generation include those associated with the drilling of exploration and production wells, building and installing plant components, and fixed O&M. Unfortunately, studies of GHG emissions from geothermal electricity production are relatively limited; most do not consider the full life cycle but instead only focus on reservoir emissions. Depending on reservoir gas composition and whether the system is closed-loop or vents to the atmosphere, reservoir emissions have the potential to dominate total life cycle

2735

GHG emissions. This distinguishes geothermal from many other ‘‘fuel-free’’ renewable technologies in which upstream, indirect contributions to emissions generally dominate (NRC, 2010b). Accordingly, the reservoir-only literature is relevant to this analysis. The average direct reservoir GHG emissions of all geothermal capacity in the U.S. in 2003 was 110 g CO2e/kWh (Bloomfield et al., 2003). Due to changes in technology and improvements in site management since then, this value has likely declined. However, we adopt the value of 110 g CO2e/kWh as a conservative estimate of the GHG emissions of all pre-2005 hydrothermal capacity in our supply curves. One full life cycle analysis found only 15 g CO2e/kWh in emissions (Hondo, 2005), which we have used for all post-2005 capacity in our supply curve. Given that approximately 85 percent of the resources in our post-2005 supply curve are appropriate for the development of binary plants with the potential to be closed-loop facilities (Smith, 2009), our assumed GHG-intensity should be considered a conservative estimate of the expected total GHG emissions from this source. 3.5. Solar: technologies, quantities and prices, and net GHG emissions Unlike other renewable technologies which are constrained by availability, solar energy is for all practical purposes unlimited. The National Academies recently estimated that coverage of only about 0.25 percent of the continental U.S. could provide all of the more than 4000 TWh of electricity generated annually (NRC, 2010b). The issues with the widespread use of solar technologies, therefore, are related to cost and intermittency. Two solar technologies can generate electricity from sunlight: photovoltaics (PV) and solar thermal. The first of these, PV, is expensive even relative to other renewables (Curtright et al., 2008; NRC, 2010b), and suffers from intermittency problems on par with those of wind (Curtright and Apt, 2008). PV was therefore not considered further for this study. Solar thermal technologies convert sunlight directly into useful heat. Utility-scale solar thermal technologies, collectively referred to as concentrating solar power (CSP), use concentrated solar radiation to generate high temperatures to make high-pressure steam which runs a turbine in much the same way that fossil fuel-derived steam is used to generate electricity. These technologies are likely to be less expensive than PV by 2025 and, due to their relative immaturity, have ample room for technological improvement. We therefore explicitly considered CSP in our analysis and determined the expected costs based primarily on the assumptions in AEO (EIA, 2009a).6 We additionally considered the many technological breakthroughs possible for CSP technologies in the near-term and factored this into our Favorable Technology Scenario. Despite potential for substantial cost reductions, other renewable technologies were less costly than CSP so it does not contribute to projected supply in this analysis.

4. Calculating aggregate costs and reductions in GHG emissions 4.1. Economic costs of substituting renewables for coal Electricity from some sources of renewable energy, like hydropower and the best sites for wind and geothermal, can be cheaper than electricity generated by natural gas and, in some instances, competitive with coal-fired power. However, as more capacity is added, costs rise. Additional capacity for wind power will have to 6 EIA assumes costs based on power tower CSP technology and 6 hours of thermal storage.

2736

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

be located in places where the wind blows less frequently and less strongly, or sites further from demand. Adding biomass capacity entails expanding the use of agricultural and forest residues, raising costs. Less productive geothermal sites would need to be developed. To calculate the economic costs of imposing a 25 percent renewable electricity mandate in the U.S. by 2025, we first sort the available hydropower, wind, biomass, geothermal, and solar sources of renewable energy by cost per kWh to create a supply curve, starting with the lowest cost sources and then adding increasingly expensive ones up to 1087 TWh, 25 percent of projected electric power consumption in the U.S. in 2025 of 4348 TWh under the EIA’s Reference Case (EIA, 2009a). We then estimate the additional cost of providing this electricity from renewable resources rather than from coal-fired power plants. The assumed additional cost of renewables is the difference between the cost of electricity generated from coal capacity and the cost of the renewables. Electricity from existing coal-fired power plants is assumed to be 2 to 4 cents/kWh, based on modeled dispatch curve data for the current coal fleet (Ventyx, 2009). Costs are low despite our assumption that the highest cost plants are displaced first because most coal-fired power plants in the U.S. are fully depreciated. We first compare the cost of renewables against this cost of power from existing coal-fired capacity, assuming that the addition of renewable capacity leads to the reduction of a corresponding amount of existing coal-fired generation. The highest cost coal plants tend to be older, less efficient plants located in the Northeast, Mid-Atlantic, and Midwest regions. The remaining coal plants are assumed to operate under current cost conditions. All marginal costs are assumed to include current pollutant emissions charges, if applicable; however these could increase in the future depending on emissions policies and abatement technology costs. The cost of new coal-fired generation becomes important when additional sources of renewables begin to displace new coal-fired power. EIA projects that electricity consumption in the U.S. will expand by 16 percent between 2007 and 2025 (EIA, 2009a). Retirements of older capacity and increased demand will necessitate additions to generating capacity. In its Reference Case, EIA projects gross additions of 114.6 GW in generating capacity between 2007 and 2025, of which 17.6 GW are projected to come from coal (EIA, 2010). If these new plants were to be operated at a capacity factor of 85 percent, they would generate 131 TWh a year. To induce investors to build new coal-fired power plants, prices for power will have to cover the capital costs as well as the variable costs of the new plants. As with renewable technologies, we base our parameters for capital and other costs for new energy capacity on EIA’s Annual Energy Outlook. We estimate that the LCOE of new coal-fired power would be 4.4 cents/kWh, i.e., investors would have to receive this price to recover the costs of building and operating these plants (see Electronic Annex). As with existing coal power, the costs of future coal power could be higher depending on air pollutant emissions policies and abatement costs, which could change the cost differences between renewables and coal. Under an RPS, we assume that renewables, rather than coal, provide this additional projected demand. For this segment of demand, we compare the cost of renewables against the cost of power generated from new coal-fired capacity rather than against costs from existing plants.

documented and fairly well understood for these technologies. This is not the case for biomass. Biomass is not yet commercially collected or used in generating plants on a wide scale. Until utilities have significant commercial experience in collecting and utilizing biomass, the cost estimates are subject to considerable uncertainty. In addition, the expanded collection of biomass is likely to have effects on natural habitat and the environment. We do not calculate these costs in the analysis above. For wind energy, the major uncertainties stem from intermittency and seasonality. If wind accounts for a substantial share of total generating capacity, utilities may have to adopt aggressive demand management, or deploy storage capacity or back-up generation to address these problems. Pricing and regulatory mechanisms, information technologies, and storage technologies needed to manage such large changes are not currently in use. Until these technologies are widely diffused or, in the case of some storage technologies, commercially available, wind power may face more constraints on expansion than suggested by the analysis above.

5. Results and discussion 5.1. Additional costs of electricity generated by renewable sources of energy As noted above, the EIA projects that the U.S. is likely to consume 4348 TWh of electricity in 2025. Under a 25 percent mandate, renewables would have to generate 1087 TWh. Fig. 1 shows supply curves for new renewable energy to meet this target for all three scenarios. In all cases, the first 392 TWh of electricity (36 percent) would be generated by existing sources of renewable energy, beginning with hydropower (297 TWh), followed by existing biomass, wind, and geothermal sites. Accordingly, Fig. 1 begins at 392 TWh. Additional electricity is then provided by a continuously varying mixture of new wind, geothermal, and biomass sources, with each resource deployed in order of lowest cost. As additional sources of renewables are added to the mix, costs rise, as shown by the upward sloping supply curve. We assume the most expensive coal power is displaced first, as shown by the downward sloping supply curve. Under the Reference Scenario, the cost per kWh of using new sources of renewables rises from 4.6 cents/kWh for the lowest cost biomass to 9.2 cents/kWh for more expensive biomass, with an average cost of 7.8 cents/kWh. We first assume that electricity from these sources of renewable energy replaces coal-fired electricity that ranges in cost from 2.3 to 5.4 cents/kWh,

4.2. Uncertainties Not all the estimates used in this analysis have the same level of uncertainty. Companies have substantial commercial experience with wind and geothermal power. The direct costs and potential reductions in GHG emissions are therefore well

Fig. 1. Cost of generating electricity from new renewables and coal: all scenarios.

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

averaging 2.8 cents/kWh in 2009. Electricity generated by renewable energy would cost 5.0 cents/kWh more, nearly three times the cost of the coal-fired power it would replace. The total net economic cost of substituting renewable energy for this coal-fired capacity would be $34.8 billion per year. If we assume that renewables substitute for 131 TWh of prospective new coal-fired generating capacity, the net annual cost would be $32.2 billion, as new coal-fired power is estimated to cost 4.4 cents rather than the 2.4 cents cost of the current coal-fired power at this end of the dispatch curve. Under the Favorable Technology Scenario, costs are much lower. The average cost for renewables is 4.6 cents/kWh. Assuming that electricity generated by renewable energy would substitute for current coal-fired power, the cost of the mandate would be $13.0 billion per year, 63 percent more than the cost of the coal-fired power the renewables would replace. If one assumes that 131 TWh of this electricity would substitute for new coal-fired capacity, the net annual cost falls to $9.8 billion. The Unfavorable Technology Scenario is appreciably more expensive. The additional cost of renewables, relative to coal, is $45.2 billion. In this scenario, renewables average 9.4 cents/kWh, 3.4 times the average price of the coal-fired power it would replace. Utilizing the assumption that renewables would substitute for new coalfired plants brings this estimate down to $33.9 billion per year. Fig. 2 shows the contributions to meeting the 25 percent goal by each of the renewable sources under the Reference Scenario. Together, co-fired and dedicated biomass would contribute 37 percent of electricity under a 25 percent RPS. Onshore wind provides 28 percent of renewable power and hydropower almost the same.

5.2. Reductions in GHG emissions Fig. 3 shows the reductions in GHG emissions that would result from a renewable mandate for electric power and the marginal cost per metric ton of reducing those emissions under all scenarios. The emissions reduction graphed in Fig. 3 corresponds to the output of electricity from additional renewables graphed in Fig. 1. Electricity generated from existing sources of renewables does not contribute to additional reductions in GHG emissions. Because the lowest cost biomass does not always have the lowest associated GHGs, Fig. 3 is not monotonic. Fully deployed, the renewable mandate would reduce GHG emissions by about 670 million metric tons per year in the Reference Scenario. This is equivalent to about 11 percent of U.S. GHG emissions in 2008.

2737

Fig. 3. Costs of using renewable sources of electricity to reduce GHG emissions.

In the Reference Scenario, the first 100 million metric tons of GHG emissions could be saved at a cost of less than $36/metric ton. However, marginal costs climb up to about $50 for 300 million metric ton and then to as much as $70 for the remainder. In the Unfavorable Technology Scenario costs are higher: $90 for the last metric ton of carbon dioxide avoided. In contrast, in the Favorable Technology Scenario the costs are appreciably lower, just $23 for the last metric ton of GHG emissions avoided. One alternative way to induce power companies to use renewables would be to impose a tax or charge on emissions of carbon dioxide. In light of these costs of reducing GHG emissions, in the Reference Scenario, power companies would have to face a tax of $70/metric ton CO2e, to induce them to replace coal with renewables for the last increment needed to meet the 25 percent mandate. 5.3. Other estimates Kydes (2007) modeled the effect of a 20 percent RPS requirement by 2020 using similar renewable technologies.7 That analysis used the EIA’s NEMS and a relatively more expansive modeling approach, which incorporates general equilibrium relationships between energy markets and the domestic economy. Where Kydes estimates the economy-wide impacts of an RPS, our analysis considers in greater detail the likely availability (and cost) of each renewable energy source and its potential for reducing GHG emissions. Like Kydes, we find that wind and biomass-generated power are the two largest sources of new renewable generation and account for the vast majority of nonhydro renewable power. However, Kydes finds that renewables cause a sharp reduction in natural gas use, while we assume that renewables will replace coal (as detailed in Section 2.1). Moreover, Kydes finds that the cost to the energy industry of a 20 percent RPS would be between $35 and $60 billion per year; our estimate of economic costs for a 25 percent mandate lies at the lower end of that range. Broadly, our results are consistent with Kydes, though we find slightly lower costs and higher emissions reductions. 5.4. Meeting proposed mandates in recent legislation In 2009, the House of Representatives passed the American Clean Energy and Security Act (H.R. 2454; ‘‘Waxman-Markey’’), which focuses on reducing U.S. GHG emissions. H.R. 2454

Fig. 2. Sources of renewable energy in the Reference Scenario.

7 Both analyses look at similar sets of renewable technologies, although Kydes includes solar photovoltaic and municipal solid waste biogas and does not consider hydropower. We use more recent information on energy projections and renewable capacity.

2738

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

contains a Renewable Electricity Standard (RES) similar to the renewable energy mandate analyzed in this paper. The bill would require 20 percent of the U.S. electricity to be produced from renewable sources by 2020, with up to five percentage points of the target eligible to be met by improvements in energy efficiency and excluding existing sources of hydropower.8 Using our models and assumptions of costs in 2025 and excluding existing hydropower, under our Reference Scenario the total cost of meeting the H.R. 2454 renewables requirement would be about $40 billion per year without efficiency improvements and about $26 billion assuming cost-neutral efficiency improvements. The corresponding cost of the last metric ton of GHGs eliminated would be $72 and $62 CO2e, respectively. H.R. 2454 also proposes an economy-wide cap and trade program to reduce GHGs. The program would cap total allowable emissions and allow firms to trade CO2e permits at a cost determined by the market. The EPA estimates the permit price under HR 2454 would be approximately $25/metric ton CO2e in 2025 (EPA, 2010b). At a permit price of $25/metric ton CO2e – but with no renewable energy mandate – our analysis suggests that power companies would only substitute renewable sources of power for an additional 64 TWh of electricity above and beyond the 392 TWh currently generated from existing sources of renewable energy, primarily hydropower. The renewables share in total U.S. consumption of electricity would rise from 9.0 to 10.6 percent of total U.S. consumption in 2025, 1.6 percentage points. The additional power produced through renewable energy would save approximately 64 million metric tons of CO2e annually. These results are substantially lower than the results of the 25 percent mandate considered in this report. 5.5. Policy implications Our analysis indicates that using renewable energy to generate electricity under an RPS would have net economic costs of $13–$45 billion per year. Total annual GHG reductions would be about 650–700 million metric tons. These costs and GHG savings estimates could potentially be used to design strategies for R&D, demonstration, and deployment to accelerate renewable technological development as part of a multi-criteria decision analysis. While costs are substantially reduced with further technology advances, other potential factors for such an analysis include externality values, timing of renewables deployment and associated GHG reductions, the cost of alternative technical approaches to reducing GHGs, and local economic benefits and costs. As noted above, the EPA estimates the cost of a GHG allowance under the H.R. 2454 cap and trade program would be around $25/metric ton. Using that estimate, only in the Favorable Technology Scenario would an RPS by an economically efficient means of substantially reducing GHG emissions. This highlights the value of continued R&D and strategies to accelerate renewable technological development. In the other two scenarios, electric power generators would likely find it more cost-effective to invest in low-cost efficiency options or offsets if available, as well as to purchase GHG allowance permits and continue to operate coal-fired power plants.

Acknowledgments This research was funded by the Department of Energy’s National Energy Technology Laboratory (NETL). The contents of 8 H.R. 2454 allows for up to one-quarter of the standard (five percentage points) to be met through energy efficiency improvements. If a state governor is able to petition successfully, that state may be granted an exemption that would allow up to two-fifths of the RPS obligation (eight percentage points) to be met by efficiency improvements.

this work do not necessarily reflect the opinions of the research clients and sponsors. We thank NETL staff Peter Balash, Joseph DiPietro, Kenneth Kern, Timothy Skone, and Maria Vargas for thoughtful comments. Marie LaRiviere, Thomas Lee, Laura Martin, Christopher Namovicz and Robert Smith (EIA) provided data from NEMS. Richard O’Connell (Black and Veatch) provided additional data on wind resource supply. The paper also benefited from helpful feedback from two insightful anonymous reviewers.

Appendix A. Supplementary material Supplementary material related to this article can be found online at doi:10.1016/j.enpol.2011.02.042.

References American Clean Energy and Security Act (ACES), 2009. H.R. 2454, 111th United States Congress. Apt, J., Lave, L.B, Pattanariyankool, S., 2008. A national renewable portfolio standard? Not practical. Issues in Science and Technology 25 (1), 53–59. Barr, C., Dominici, F., 2010. Cap and trade legislation for greenhouse gas emissions: public health benefits from air pollution mitigation. Journal of the American Medical Association 303 (1), 69–70. Berry, T., Jaccard, M., 2001. The renewable portfolio standard: design considerations and an implementation survey. Energy Policy 29 (4), 263–277. Black and Veatch, 2007. 20 Percent Wind Energy Penetration in the United States: A Technical Analysis of the Energy Resource, Walnut Creek, CA. Bloomfield, K.K., Moore, J.N., Neilson Jr., R.M., 2003. Geothermal energy reduces greenhouse gases. Geothermal Research Council Bulletin 32, 77–79 /http:// www.geothermal.org/articles/greenhousegases.pdfS. Chen, C., Wiser, R., Mills, A., Bolinger, M., 2009. Weighing the costs and benefits of state renewables portfolio standards in the United States: a comparative analysis of state-level policy impact projections. Renewable and Sustainable Energy Reviews 13 (3), 552–566. Curtright, A.E., Apt, J., 2008. The character of power output from utility-scale photovoltaic systems. Progress in Photovoltaics 16 (3), 241–247. Curtright, A.E., Morgan, M.G., Keith, D.W., 2008. Expert assessments of future photovoltaic technologies. Environmental Science and Technology 42 (24), 9031–9038. Department of Energy (DOE), 2008. 20% Wind Energy by 2030: Increasing Wind Energy’s Contribution to U.S. Electricity Supply (DOE/GO-102008-2567). Dobesova, K., Apt, J., Lave, L.B., 2005. Are renewable portfolio standards costeffective emission abatement policy? Environmental Science and Technology 39 (22), 8578–8583. Electric Power Research Institute (EPRI), 2007. Assessment of Waterpower Potential and Development Needs (Report 1014762). Palo Alto, California. Energy Information Administration (EIA), 2009a. Annual Energy Outlook 2010 Early Release. U.S. Department of Energy, Washington, DC /http://www.eia. doe.gov/oiaf/aeo/overview.htmlS. Energy Information Administration (EIA), 2009b. Annual Energy Review. U.S. Department of Energy, DOE/EIA-0384 (2008), Washington, DC /http://www. eia.doe.gov/aer/S. Energy Information Administration (EIA), 2010. Annual Energy Review. U.S. Department of Energy, DOE/EIA-0384 (2008), Washington, DC /http://www. eia.doe.gov/aer/S. Energy Information Administration (EIA), 2009c. Wind, Biomass and Geothermal Energy Submodules. Model Documentation: Renewable Fuels Module of the National Energy Modeling System, DOE/EIA-M069, Washington, DC /http:// tonto.eia.doe.gov/FTPROOT/modeldoc/m069(2009).pdfS. Energy Information Administration (EIA), 2009d. Annual Energy Outlook 2009. U.S. Department of Energy, Washington, DC /http://www.eia.doe.gov/oiaf/ archive/aeo09/index.htmlS. Environmental Protection Agency (EPA), 2010a. Inventory of U.S.: Greenhouse Gas Emissions and Sinks: 1990–2008 (EPA 430-R-10-006). /http://epa.gov/clima techange/emissions/usinventoryreport.htmlS. Environmental Protection Agency (EPA), 2010b. Supplemental EPA Analysis of the American Clean Energy and Security Act of 2009: H.R. 2454 in the 111th Congress /http://www.epa.gov/climatechange/economics/pdfs/HR2454_Sup plementalAnalysis.pdfS. Espey, S., 2001. Renewables portfolio standard: a means for trade with electricity from renewable energy sources? Energy Policy 29 (7), 557–566. GeothermEx, 2004. New Geothermal Site Identification and Qualification. Report to the California Energy Commission, Public Interest Energy Research Program, Sacramento, CA, /http://www.energy.ca.gov/pier/project_reports/500-04-051. htmlS. Hondo, H., 2005. Life cycle GHG emission analysis of power generation systems: Japanese case. Energy 30 (11–12), 2042–2056.

K. Crane et al. / Energy Policy 39 (2011) 2730–2739

Idaho National Laboratory (INL), 2006. Feasibility Assessment of the Water Energy Resources of the United States for New Low Power and Small Hydro Classes of Hydroelectric Plants. Idaho National Laboratory, United States Department of Energy (DOE-ID-11263). Katzenstein, W., Apt, J., 2009. Air emissions from wind and solar power. Environmental Science and Technology 43 (2), 253–258. Keith, D.W., DeCarolis, J.F., Denkenberger, D.C., Lenschow, D.H., Malyshev, S.L., Pacala, S., Rasch, P.J., 2004. The influence of large-scale wind power on global climate. Proceedings of the National Academy of Sciences 101, 16115–16120. Kydes, A.S., 2007. Impacts of a renewable portfolio generation standard on US energy markets. Energy Policy 35 (2), 809–814. LaRiviere, Marie, 2009. Energy Information Administration, Personal Communication to David Ortiz, November 30. Massachusetts Institute of Technology (MIT), 2006. The Future of Geothermal Energy: Impact of Enhanced Geothermal Systems (EGS) on the United States in the 21st Century /http://geothermal.inel.gov/publications/future_of_geother mal_energy.pdfS. McGowan, T.F. (Ed.), 2009. Biomass and Alternative Fuel Systems: An Engineering and Economic Guide. John Wiley and Sons, Inc., Hoboken, NJ. Michaels, R.J., 2008. A national renewable portfolio standard: politically correct, economically suspect. The Electricity Journal 21 (3), 9–28. Namovicz, C., 2009. Energy Information Administration, Personal Communication to C. Samaras, November 30, 2009. National Energy Technology Laboratory, 2010. Calculating Uncertainty in Biomass Emissions (CUBE) Model, Version 1.0 /http://www.netl.doe.gov/energy-ana lyses/refshelf/detail.asp?pubID=298S. National Renewable Energy Laboratory (NREL), 2010. Eastern Wind Integration and Transmission Study. U.S. Department of Energy (NREL/SR-550-47078). National Research Council (NRC), 2007. Environmental Impacts of Wind-Related Projects. National Academies Press, Washington, DC. National Research Council (NRC), 2009. Liquid Transportation Fuels from Coal and Biomass: Technological Status, Costs, and Environmental Impacts. National Academies Press, Washington, DC. National Research Council (NRC), 2010a. Hidden Costs of Energy: Unpriced Consequences of Energy Production and Use. National Academies Press, Washington, DC.

2739

National Research Council (NRC), 2010b. Electricity from Renewable Resources: Status. Prospects, and Impediments. National Academies Press, Washington, DC. Nebraska Power Association (NPA), 2009. Nebraska Statewide Wind Integration Study (AAM-9-89030-01). Nogee, A., Deyette, J., Clemmer, S., 2007. The projected impacts of a national renewable portfolio standard. The Electricity Journal 20 (4), 33–47. Patel, S., 2009. Assessing the earthquake risk of enhanced geothermal systems. Power Magazine /http://www.powermag.com/issues/features/2309.htmlS. Peltier, R., 2009. Top Plants: Rio Bravo Rocklin Power Station, Lincoln, California. Power Magazine /http://www.powermag.com/issues/cover_stories/TopPlants-Rio-Bravo-Rocklin-Power-Station-Lincoln-California_2306.htmlS. Smith, R., 2009. ‘‘Conversations and email correspondence regarding assumptions for geothermal technologies in NEMS,’’ to A.E. Curtright. Toman, M., Griffin, J., Lempert, R., 2008. Impacts on U.S. Energy Expenditures and Greenhouse-Gas Emissions of Increasing Renewable-Energy Use (TR-384-1-EFC). RAND Corporation, Santa Monica, CA. U.S. Geological Survey (USGS), 2008. Assessment of Moderate- and High-Temperature Geothermal Resources of the United States (Fact Sheet) /http://pubs. usgs.gov/fs/2008/3082/pdf/fs2008-3082.pdfS. Van Loo, S., Koppejan, J. (Eds.), 2008. The Handbook of Biomass Combustion and Co-firing. Earthscan, London. Ventyx, 2009. Velocity Suite Database /http://www.ventyx.comS. Wang, M., Wu, M.M., Elgowainy, A., 2007. GREET 1.8 Fuel-Cycle Model for Transportation Fuels and Vehicle Technologies. Argonne National Laboratory, Argonne, IL. Weisser, D., 2007. A guide to life-cycle greenhouse gas (GHG) emissions from electric supply technologies. Energy 32, 1543–1559. Western Governors’ Association (WGA), 2006. Clean and Diversified Energy Initiative: Geothermal Task Force Report /http://www.westgov.org/wga/initia tives/cdeac/Geothermal-full.pdfS. Wiser, R., Bollinger, M., 2009. 2008 Wind Technologies Market Report. U.S. Department of Energy /http://eetd.lbl.gov/ea/emp/reports/2008-windtechnologies.pdfS. Wiser, R., Namovicz, C., Gielecki, M., Smith, R., 2007. The experience with renewable portfolio standards in the United States. The Electricity Journal 20 (4), 8–20.