Article pubs.acs.org/est
Potential Impacts of Electric Power Production Utilizing Natural Gas, Renewables and Carbon Capture and Sequestration on U.S. Freshwater Resources Vincent C. Tidwell,*,† Leonard A. Malczynski,† Peter H. Kobos,† Geoffrey T. Klise,† and Erik Shuster‡ †
Sandia National Laboratories, Albuquerque, New Mexico 87123, United States National Energy Technology Laboratory Pittsburgh, Pennsylvania 15236
‡
ABSTRACT: Carbon capture and sequestration (CCS) has important implications relative to future thermoelectric water use. A bounding analysis is performed using past greenhouse gas emission policy proposals and assumes either all effected capacity retires (lower water use bound) or is retrofitted (upper bound). The analysis is performed in the context of recent trends in electric power generation expansion, namely high penetration of natural gas and renewables along with constrained cooling system options. Results indicate thermoelectric freshwater withdrawals nationwide could increase by roughly 1% or decrease by up to 60% relative to 2009 levels, while consumption could increase as much as 21% or decrease as much as 28%. To identify where changes in freshwater use might be problematic at a regional level, electric power production has been mapped onto watersheds with limited water availability (where consumption exceeds 70% of gauged streamflow). Results suggest that between 0.44 and 0.96 Mm3/d of new thermoelectric freshwater consumption could occur in watersheds with limited water availability, while power plant retirements in these watersheds could yield 0.90 to 1.0 Mm3/d of water savings.
■
INTRODUCTION Growing concern over global warming and associated climate change has prompted interest in carbon capture and sequestration (CCS) as a means of reducing greenhouse gas emissions. Electric power production is an important contributor, responsible for 39% of total CO2 emissions.1 As such, the electric power sector is a target for CCS; particularly, coal-fired plants which are responsible for ∼80% of this sector’s CO2 emissions.2 For the sake of clarity, the term “carbon” used throughout this analysis represents a shorthand term for “CO2”. While reduction of CO2 emissions from coal-fired power plants could have important financial implications for the electric power industry,3 changing demands for water associated with CCS are likely to be just as important. Using today’s technologies, for example, amine-based postcombustion carbon capture, would increase water consumption through the addition of several unit processes that are both energy and water intensive.4,5 Solvents used to capture CO2 require energy to regenerate the solvent so it can be used again and require water to cool the process. Cooling water is also required in CO2 absorbing and stripping processes. Once the CO2 is captured, it must be compressed for sequestration or beneficial reuse, with compressors usually having significant operating power and cooling requirements. The added internal energy requirements © 2013 American Chemical Society
for these processes can effectively subtract 20−30% of the energy from net plant power output6 and increase water consumption by up to 126%.7 The National Energy Technology Laboratory8 (NETL) conducted an analysis of the potential impacts of CCS on water demand by the thermoelectric power industry. Their analysis assumed all coal plants in the U.S. will be retrofitted by 2030 and that the lost generation capacity to CCS would be compensated by either nuclear, integrated gasification combined cycle (IGCC), or pulverized coal plants (the latter two cases would include CCS). They also adopted the Energy Information Administration’s (EIA’s) forecast for the nation’s future electric generation portfolio. NETL found that deploying carbon capture technologies in new and existing plants would increase water withdrawal (total water abstracted from a water body) by 2−3% and consumption (difference between withdrawal and return flow, which is the amount of water lost to the system by evaporation or direct incorporation into products) by 52−55% over 2005 levels. Chandel et al.9 Received: Revised: Accepted: Published: 8940
December 22, 2012 June 21, 2013 June 21, 2013 June 21, 2013 dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
Environmental Science & Technology
Article
and other biomass, solar thermal, solar photovoltaic, wind, and offshore wind). Siting of new power plants was accomplished so as to maintain the current ratio of electricity production at the county level, to total EMM region electricity production (by fuel type). This simple treatment assumed that future power plants would be sited so as to take advantage of renewable energy sources (wind, solar, hydropower) as well as existing fuels transportation and electricity transmission infrastructure. The thermoelectric water demand module calculated water withdrawal and consumption for the existing and future power plant fleet. For new power plants, this calculation was accomplished by multiplying the production rate by the associated water withdrawal/consumption factor (Table 1).
extended this analysis by modeling four climate policy scenarios that involve different costs for emitting CO2 and different technology options for reducing emissions. The potential impacts of the scenarios on the U.S. electric system were modeled using a modified version of the U.S. National Energy Modeling System. Results suggest freshwater withdrawals decline 2−14% while consumption increases by 24−42%. The distinctly lower water use reflects a recognition that the types of power plants built to compensate for the electricity consumed by CCS depends on the specific climate-policy regulations and the price of CO2 under that policy. Additionally, not all power plants will retrofit; rather some will retire while others will continue to operate but pay emission penalties. Here a regional analysis is performed to explore potential implications of CCS on the freshwater resources of the U.S. The analysis explores several plausible alternative futures that reflect greenhouse gas emissions policies introduced previously in the U.S. Congress and current trends toward expanding use of natural gas and renewables. Projections of changes to thermoelectric freshwater withdrawal and consumption are estimated and then compared to metrics representative of a region’s ability to accommodate the increased freshwater use. In this way the analysis identifies water resource regions where deployment of CCS technology could be particularly disruptive. This analysis is conducted for the conterminous U.S., compiled at the accounting unit level (6-digit hydrologic unit code [HUC], 329 watersheds).
Table 1. Shares of Future Capacity Expansion by Fuel Type for the Reference, $15 CO2 and $25 CO2 Scenariosa and Water Withdrawal/Consumption Factors by Fuel Typeb
coal oil and natural gas steam combined cycle combustion turbine/ diesel nuclear power conventional hydropower geothermal municipal waste wood and other biomass solar thermal solar photovoltaic wind offshore wind
■
MATERIALS AND METHODS This analysis was designed to provide a consistent view of the CCS-water nexus across the conterminous U.S. To make this problem tractable in a computational and data availability sense, the analysis was framed on an annual time step with a spatial dimension defined by watersheds at the 6-digit HUC level. The analysis framework was formulated within a system dynamics architecture10 and implemented within the commercial software package Studio Expert 2009, produced by Powersim, Inc. The analysis framework was organized according to four interacting systems, electric power production, thermoelectric freshwater demand, nonthermoelectric water demand (e.g., municipal, agricultural), and water supply. Below a brief description of the model is provided, while additional detail may be found in Tidwell et al.11 Model Description. Electric power generation was analyzed at the unit level with 18 518 individual units simulated. Units were distinguished by fuel type, geographic location, installed capacity, annual power output, build date, cooling type, and water source, as recorded in the eGRID database.12 This database reflected the 2009 fleet of power plants/units, which served as the initial conditions for the electric power sector module. Future demand for electric power was based on projections in the 2012 Annual Energy Outlook (AEO).2 New electricity demand was first satisfied with electric capacity currently under construction as reported by Electricity Market Module (EMM) region.2 Characteristics of the additional power plant capacity required to meet future demand were defined according to various projections from the 2012 AEO (see below). These projections were reported as new power plant capacity by EMM region and fuel type (coal, oil, and natural gas (steam, combined cycle, combustion turbine/diesel), nuclear power, conventional hydropower, geothermal, municipal waste, wood
reference (%)
$15 CO2 (%)
$25 CO2 (%)
withdrawal factor (m3/ MWh)
consumption factors (m3/ MWh)
0.0 0.0
0.0 0.0
0.0 0.0
3.79 4.54
2.54 3.13
45.4 20.7
36.1 6.5
30.8 1.8
0.96 0.0
0.78 0.0
5.6 1.9
20.5 2.2
32.1 1.4
4.16 0.0
2.54 0.0
2.3 0.1
1.6 0.1
1.1 0.0
1.89 0.90
1.89 0.90
0.3
0.2
0.7
3.32
2.09
0.5 4.6
0.3 5.8
0.2 11.3
3.41 0.004
3.41 0.004
18.5 0.1
26.7 0.1
20.4 0.1
0.0 0.0
0.0 0.0
a
Scenarios are taken from the 2012 AEO.2 Values are percentages of total new capacity additions. bWater withdrawal and consumption factors from Macknick et al.7 (median values assuming, “generic” technology and recirculating tower cooling).
These water use factors were unique to the type of plant and associated cooling technology.7 Estimates of water withdrawal and consumption at existing plants were developed from a variety of sources. Specifically, information available through EIA’s Form EIA-767,13 EIA-86014 and NETL’s Coal Power Plant Database15 were combined with estimates calculated using the water withdrawal/consumption factors above. Use of the different data sets was necessary because of gaps and inconsistencies in the reported data. Adjustments to the plant level data were made as needed to achieve agreement between the reported values,13−15 calculated values based on withdrawal/consumption factors,7 and aggregated use values at the county level,16,17 recognizing the USGS data did not include thermoelectric water use supplied through a municipal source. The water source, whether surface water, groundwater, or seawater, was also tracked. The nonthermoelectric water demand module projected the future demand for water according to four different use sectors: municipal (including domestic, public supply, and commercial), 8941
dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
Environmental Science & Technology
Article
the associated changes in water withdrawal/consumption and parasitic power demands. The CCS module relies on CO2 emissions projected by EIA.2 Three alternative projections are used including the EIA Reference Case as well as two others that assume different emission taxes. The latter two cases are based on pricing of carbon emissions that the U.S. Environment Protection Agency23 assessed would occur under the climate policies proposed in the American Clean Energy and Security Act (H.R. 2454, a.k.a. the former Waxman-Markey bill) and the Clean Energy Jobs and American Power Act (S. 1722, a.k.a. the former Kerry-Boxer bill). For each year of the simulation the model determines that amount of new CO2 sequestration capacity that must be added to meet the target emission rate. This required sequestration capacity is simply calculated as the difference between the current emissions and the projected EIA emissions. Current emission are calculated as the sum of emissions from plants constructed before 2010, less any sequestered carbon, plus emission from any new plant construction. Based on the sequestration capacity that must be added annually, the model determines which of the existing coal-fired power plants will be retired, retrofitted, or continue to operate while paying associated emissions tax. Only coal-fired plants are considered as they are responsible for the bulk of the carbon emissions2 and operate at the highest ratio of carbon emissions to electricity produced making them the most expensive to operate under an emission tax. The order in which coal plants are considered for retirement/retrofit is simply based on the ratio of carbon emission to the electricity produced. Plants with the highest ratio (most expensive to operate) are treated first. Rather than deciding whether a plant is likely to retire or retrofit, two alternative scenarios are consideredone in which all plants selected for treatment are retrofitted and another in which all selected plants are retired. This all or nothing analysis avoids complexities associated with each power plant’s financial reality as well as whether the plant site is capable of supporting expanded CCS operations. Although neither scenario is likely the analysis provides a reasonable upper and lower bound for the future spectrum of energy futures. Retrofitted power plants are assumed to capture 90% of their emissions. For power plants selected for retrofitting the resulting water withdrawal/consumption and parasitic power demands are calculated. Retrofitted power plants are assessed a parasitic energy loss of 30% due to implementation of CCS.6 This increase in parasitic electricity generation along with the CCS cooling/process water demand is assumed to increase water withdrawal/consumption by 81/71% in coal plants and 63/44% in IGCC plants.7 It is assumed that existing plants retrofitted for CCS maintain their current water source and type of cooling system.
industrial, agriculture, and mining. Water withdrawal and consumption were tracked separately as were the resulting return flows. Also modeled was the source of the withdrawal, whether it was surface water, groundwater, or nonpotable. Water use statistics published by the U.S. Geological Survey (USGS) served as the primary data source for the analysis,16−20 noting that reporting of consumptive use data was discontinued after 1995. Using the available data, trends in water use were estimated by sector, use type (withdrawal/consumption) and source. These trends were further related to changes in population and economic activity (as measured by gross state product), where quantifiable correlations existed. In this way water use projections were a function of population change, economic growth, and trends in historical water use rates (i.e., reflecting changing use/conservation practices).11 Projected changes in thermoelectric and nonthermoelectric water use were then compared to water supply; specifically, gauged streamflow from the USGS National Hydrographic Data set.21 These data were combined into a metric that provides a measure of physical freshwater availability, similar to that used in the Annual Water Adequacy Analysis conducted as part of the Second National Water Assessment.22 Physical water availability was represented as the ratio of freshwater demand to freshwater supply within a given watershed (modeled at the 6-digit HUC level). Specifically, freshwater availability, WA, was calculated as follows: WA =
CUw + CUu Q A + CUw + CUu
(1)
where freshwater demand is measured as the consumptive use of water both within the watershed, CUw, and that occurring upstream of the basin, CUu, whereas freshwater supply, QA, is taken as the gauged streamflow (regulated flows reflecting reservoir operations, environmental flows, etc.). As this ratio approaches one the consumptive use of freshwater approaches the physical supply, thus the availability of freshwater for further development is limited. For purposes of this analysis watersheds with limited water availability were taken as those with a WA of 0.7 or higher;22 that is, those watersheds where the demand accounts for 70% or more of the physical supply. The availability of freshwater for new development is largely limited by annual low flow conditions. To reflect this constraint the measure of physical water supply, QA, used in calculating WA was the 20th percentile flow, or that flow which is exceeded 80% of the time. This provides a conservative measure of annual low flow. These low flows are largely associated with late summer when water demand is at a maximum. While summer maximum water demand data was not available a conservative estimate (both CUw and CUu) was derived from the annual average demand data. Specifically, annual average thermoelectric water consumption was increased by 12% to reflect higher electricity demands2 and higher evaporation rates. Irrigation demands were increased by a factor of 1.5, conservatively assuming uniform irrigation over 8 rather than 12 months out of the year. Municipal consumption was increased by a factor of 1.5 to reflect that most outdoor irrigation is also limited to an 8 month window. Consumption in all other use sectors was maintained at their average levels. CCS Water Use Module. The CCS Water Use Module determines how much carbon must be captured in a given year, decides which units are retrofitted/retired and then calculates
■
RESULTS For purposes of this analysis, three alternative “energy futures,” termed scenarios, were investigated. The basis of the analysis was three out of the 47 scenarios reported in the 2012 AEO.2 Specifically considered were the Reference Case, $15 Carbon Dioxide Emission Fee ($15 CO2), and the $25 Carbon Dioxide Emission Fee ($25 CO2). The Reference case was EIA’s best guess at the future fuel mix without any significant unexpected changes in policy, fuel prices, or technology costs. The $15 and $25 CO2 cases were modeled after climate policies proposed in the American Clean Energy and Security Act and the Clean 8942
dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
Environmental Science & Technology
Article
Table 2. Change in Thermoelectric and Non-Thermoelectric Freshwater Withdrawal and Consumption (2009−2035) for the Alternative Future Energy Scenariosa withdrawal consumption
reference
$15 CO2 no retrofit
$15 CO2 full retrofit
$25 CO2 no retrofit
$25 CO2 full retrofit
non- thermoelectric
4.6 (2.5,6.6) 1.61 (1.2,2.3)
−262.9 (−264.6, −260.1) −3.3 (−3.9, −2.4)
6.0 (3.5,9.1) 3.0 (2.5,3.8)
−326.9 (−328.6, −322.7) −4.2 (−4.9, −3.2)
6.1 (3.7, 9.1) 3.4 (2.9, 4.1)
76.6 18.5
a
Values are in Mm3/d. Three values are given for each table entry reflecting the range in possible thermoelectric water use factors,7 median (top), low and high (values, respectively in parentheses below).
Figure 1. Change in nonthermoelectric (a) and thermoelectric (b−f) freshwater consumption between 2009 and 2035 in the conterminous U.S. Change in thermoelectric consumption is mapped for each of the five alternative energy futures. Data are displayed at the 6-digit HUC watershed level in units of million cubic meters per day (Mm3/d). The black outline indicates those watersheds projected to have the largest nonthermoelectric growth in freshwater consumption from 2009 to 2035 (top 25% of watersheds).
Energy Jobs and American Power Act, and were the only two scenarios that deal with a carbon constrained future. According to 2012 AEO2 electricity generation is projected to grow by 31% from 2009 to 2035 (3528−4635 terawatt-hours [TWh]) under the Reference case. Likewise, generating capacity increases from 994.9 to 1112.5 gigawatts (GW), while CO2 emissions grow by 8% from 2159 to 2330 million
metric tons (MMT). Under a carbon constrained future, electric power generation grows by only17% and 13%, for the $15 and $25 CO2 cases, respectively, due in large part to the accompanying increase in electricity prices. More importantly, carbon emissions decline by 43% and 65%, respectively, by 2035. 8943
dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
Environmental Science & Technology
Article
generation capacity which is replaced by much lower water use natural gas (combined cycle and combustion cycle) and renewables (e.g., PV and wind) as projected in the 2012 AEO.2 Thermoelectric freshwater consumption in 2009 is projected to be 14.8 Mm3/d. To put these changes in thermoelectric freshwater use in perspective, they are compared against projected changes in other use sectors; specifically, municipal, industrial, mining, and agriculture which we term nonthermoelectric use. Nonthermoelectric freshwater withdrawal is projected to increase by 76.6 Mm3/d, while consumption is estimated to increase by 18.5 Mm3/d by 2035 (Table 2). Projected changes in thermoelectric withdrawal under the Reference and Full Retrofit cases are small in comparison with that for nonthermoelectric uses, while potential declines in thermoelectric withdrawal under the No Retrofit cases could more than compensate for the projected increases in nonthermoelectric use. In terms of consumption, projected changes in thermoelectric consumption range from −23% to 18% of the projected increase in nonthermoelectric consumption. Table 2, which presents data aggregated at the national level, tells only a part of the story. In particular, future demands for freshwater are not uniformly distributed across the U.S. Figure 1 presents the projected change in freshwater consumption between 2009 and 2035 for the conterminous U.S. at the 6-digit HUC watershed level. Focus is placed on freshwater consumption (not withdrawal) as consumed water is not available for downstream use and thus impacts water availability. These maps show where “new” freshwater is needed to meet future development in the nonthermoelectric and thermoelectric sectors (five scenarios). Note that for some watersheds the net change in thermoelectric freshwater use is negative reflecting cases were water use at retired power plants exceed demand for newly constructed power plants. The most important feature of these maps is the distinct variability in new freshwater demand across the U.S. For nonthermoelectric consumption, high values are associated with densely populated watersheds reflecting projected growth in the municipal and industrial sectors. In contrast, future thermoelectric freshwater demand reflects geographic differences in electricity demand growth, primary fuel supply, and transmission capacity (among other factors). Another important difference is evident when comparing the No Retrofit and Full Retrofit cases where big changes in freshwater consumption reflect basins with significant coal-based generation (e.g., Mid-West and FourCorners region of the Southwest). Comparing the change in nonthermoelectric and thermoelectric freshwater consumption provides an idea where new development in the thermoelectric sector (Figures 1b−f) is likely to compete with other sectors for finite freshwater resources (Figure 1a). Under the Reference case 39 of the 82 watersheds with the fastest growth in thermoelectric consumptive demand (top 25% of watersheds) are associated with watersheds with the fastest nonthermoelectric growth (again defined as the top 25%). For the Full Retrofit cases the number of overlapping high growth watersheds is 39 and 41, for the $15 and $25 CO2, respectively. For the No Retrofit cases overlapping watersheds decrease to 36 in both cases. However, a net reduction in thermoelectric freshwater consumption occurs in 44 and 45 ($15 and $25 CO2) of the watersheds with the greatest growth in nonthermoelectric freshwater use due to the retirement of coal-fired power plants. Watersheds projected to have the largest growth in nonthermoelectric freshwater
The three scenarios also differ in the projected mix of fuels used in construction of the future power plant fleet.2 To give an idea of the differences between the three scenarios, shares of future capacity expansion by fuel type are given in Table 1. Although data in Table 1 are aggregated at the national level, the analysis utilized projections by EMM region. One factors kept constant across all scenarios is projected growth in nonthermoelectric water demand. It is also assumed that no new power plant will employ open-loop cooling, rather the plant will adopt a recirculating or air cooled system consistent with the mix as of 2009 (by fuel type). This assumption is consistent with recent construction trends and potential new EPA regulations restricting high capacity intake structures.24 The mix as of 2009 is 42% recirculating tower, 14% recirculating pond, 43% open-loop, and 0.9% air cooled. Also assumed is that no new thermoelectric generation will utilize seawater as its cooling source, consistent with trends over the last 20 years. Although all water sources were tracked in this analysis, only results for changes in freshwater use are presented. Freshwater Withdrawal and Consumption. Thermoelectric freshwater withdrawal and consumption aggregated at the national level are shown in Table 2. In each case three values are shown, median, low and high reflecting inherent variability in water use factors due to differences in plant equipment and site operational conditions.7 For the sake of brevity, discussions will be restricted to the median values. Table 2 compares projected thermoelectric freshwater withdrawals for the Reference case and that for the four CCS-based scenarios. The Reference and two Full Retrofit cases realize a small increase in withdrawal. Increases reflect freshwater consumption by new power plants required to meet growing demand, whereas larger increases for the Full Retrofit cases reflect additional freshwater needs for CCS operations. Nevertheless, the total change in withdrawal for any of these three cases is roughly 1% of the 2009 levels. These modest changes in withdrawal reflect the assumption that all new plant capacity will utilize recirculating or dry cooling. In contrast the two No Retrofit cases result in significant declines in withdrawal relative to 2009 levels, a 48% and 60% decrease for the $15 and $25 CO2 cases, respectively. The significant reductions are driven by the closure of numerous coal-fired plants, many of which utilize open-loop cooling, that are replaced with lower water use plants using recirculating cooling. Specifically, the $15 CO2 No Retrofit case results in the retirement of 171 GW of coal-fired capacity (81 GW with open loop cooling), whereas 248 GW (117 GW with open loop cooling) are retired under the $25 CO2 case. Thermoelectric freshwater withdrawals in 2009 are estimated to be 543.9 million cubic meters per day (Mm3/d). Similarly, thermoelectric freshwater consumption increases for the Reference and two Full Retrofit cases while decreases for the two No Retrofit cases (Table 2). Projected increases above 2009 levels for the Reference, $15, and $25 CO2 Full Retrofit cases are 11%, 20%, and 23%, respectively. Increases reflect freshwater consumption by new power plants required to meet growing demand, while larger increases for the Full Retrofit cases reflect additional water needs for CCS operations. Interestingly, for the two No Retrofit cases, thermoelectric freshwater consumption is projected to decrease relative to 2009 levels by 23% and 29% for the $15 and $25 CO2 No Retrofit cases, respectively. As with withdrawal, this decrease is due to the retirement of considerable coal-fired 8944
dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
Environmental Science & Technology
Article
Figure 2. Freshwater availability metric for 2009 (a) mapped at the 6-digit HUC level, which reflects the ratio of freshwater demand to water supply (eq 1). Index values greater than 0.7 indicate regions with limited water availability for new development. Also shown is freshwater consumption associated with new thermoelectric power plants sited in watersheds with limited water availability (b−f). This at risk water is mapped for each of the five alternative energy futures. The black outline indicates those watersheds projected to have the largest nonthermoelectric growth in freshwater consumption from 2009 to 2035 (top 25% of watersheds).
are varied over reasonable ranges the number of designated watersheds change by less than ±3%. This simply reflects that arid environments have low streamflows and corresponding high irrigation demands. This creates a distinct disparity in the water availability metric, WA, across the U.S. Additionally, projected changes in thermoelectric freshwater consumption (Figures 1b−f) are mapped onto watersheds with limited water availability (watersheds with water availability values of 0.7 or higher). Figure 2b shows the projected change in thermoelectric freshwater consumption (Reference case) to be met by a freshwater source which corresponds to a watershed with limited water availability (Figure 2a). This map shows where it is likely to be difficult or expensive to obtain a water right/permit for a new thermoelectric power plant. Alternatively, this map indicates where future power plant siting should consider alternative water sources (e.g., wastewater, brackish groundwater, seawater),27 dry/hybrid cooling, or nonthermoelectric generation. Table 3 indicates that a total of 154 TWh of new electric power production is located in
consumption (2009−2035) are outlined in black in Figures 1b−f to assist in visualizing potential cross-sectorial competition over water. Freshwater Availability and Thermoelectric Development. Of particular interest to this study is where future expansion of thermoelectric power generation may be limited by the availability of freshwater. To explore this issue the water availability metric (eq 1) was calculated and mapped for the year 2009 in Figure 2a. A review of this metric (Figure 2a) clearly reveals that values tend to be lower in the East and far Northwest, while higher values tend to dominate the West and Southwest (indicating relatively limited freshwater availability). This pattern is a result of both the aridity of the West and the higher freshwater use due to irrigated agriculture. Similar results have been achieved with other measures of surface water stress/ availability.25,26 It is worth noting that the watersheds designated as having limited water availability are relatively insensitive to assumptions on the cutoff threshold (WA ≥ 0.7), choice of the 20th percentile flow, or even the rate of growth in nonthermoelectric freshwater demand. When these parameters 8945
dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
Environmental Science & Technology
Article
demands, whereas reductions in thermoelectric consumption under the No Retrofit cases could be sufficient to offset about 19% of other sectorial demands. Ultimately the energy-water nexus must be investigated at the regional level. Between 0.44 and 0.96 Mm3/d of new thermoelectric freshwater consumption could be sited in watersheds with limited water availability. This represents new thermoelectric freshwater demands where it is likely to be difficult or expensive to obtain a water right/permit, and thus where future power plant siting should consider wastewater/ brackish groundwater, seawater, dry/hybrid cooling, or nonthermoelectric generation. On a positive note, under the No Retrofit cases between 0.90 and 1.0 Mm3/d of freshwater could be freed up for other uses through the retirement of coal-fired power plants. Depending on the alternative energy future, 36−41 of the 82 watersheds with the fastest growing nonthermoelectric demand (top 25%) overlap those watersheds with the fastest growing thermoelectric freshwater demand. Further, watersheds with the fastest growing nonthermoelectric water demand intersect with 22−23 watersheds (depending on the scenario) with limited water availability. In contrast, net retirement of thermoelectric freshwater use under the No Retrofit cases would make water available in 44−45 watersheds with rapid nonthermoelectric growth where 8−9 of these watersheds are also in basins with limited water availability. These results point to the potential for strong competition over freshwater for new development in the thermoelectric and nonthermoelectric sectors. It also suggests there is the opportunity for cooperation in which strategic retirements of power plants or construction using nontraditional water sources27 or dry/hybrid cooling could ease competition while improving drought resilience for all parties. In the West this competition is likely to drive prices for water rights up, thus challenging thermoelectric development involving freshwater sources or alternatively benefiting retiring/retrofitting plants where abandoned water rights could be sold to partially offset related costs. Presented results clearly demonstrate the geospatial heterogeneity characteristic of the energy and water sectors. Energy demand and production as well as water demand and supply vary as a function of location. This geographic variation is a function of population density, use characteristics of the population, energy fuels supply, the transmission grid, physiographic characteristics of the region, traditional practices, and many others. This variation, in turn, interacts with disparate regulation, policy, and finances at the local, state, and federal levels that drive development of water and energy resources. The complexity of this nexus is evident in the fact that thermoelectric freshwater use characteristics do not follow a simple linear scaling between the $15 and $25 CO2 scenarios. This complexity is further evidenced by the local nuances apparent when comparing the energy and water development maps in Figures 1 and 2. This speaks to the importance and need for further integrated cross-sectorial analysis.
Table 3. Thermoelectric Power Production and Associated Freshwater Consumption at Siting Risk Due to Limited Water Availability.a electricity (TWh) reference $15 CO2 $15 CO2 $25 CO2 $25 CO2
no retrofit full retrofit no retrofit full retrofit
154 135 146 113 127
water consumption (Mm3/d) 0.55 0.47 0.96 0.44 0.95
(0.53, (0.44, (0.93, (0.41, (0.92,
0.57) 0.51) 1.0) 0.48) 0.99)
a Three values are given for each water consumption entry reflecting the range in possible thermoelectric water use factors,7 median value followed by low and high values, respectively in parentheses.
basins with limited water availability. Freshwater consumption of 0.55 Mm3/d is associated with this power production. Figures 2c−f show new thermoelectric freshwater demands in watersheds with limited water availability for the four CCS alternative energy futures. Relative to the Reference case, the two Full Retrofit cases result in slightly less electric power generation at risk, 146 and 127 TWh, for $15 and $25 CO2, respectively, while associated freshwater consumption increases to 0.96 and 0.95 Mm3/d. The small decline in at risk power is due to overall reductions in power demand because of higher electricity prices2 while the increased freshwater consumption reflects that required for CCS operations. The No Retrofit cases result in small declines in both at risk power production and associated freshwater consumption (Table 3). From inspection of the associated maps (Figures 2d and f) it is apparent that numerous at risk watersheds experience a net decrease in thermoelectric freshwater consumption due to the retirement of coal-fired power plants, totaling 0.90 and 1.0 Mm3/d for the $15 and $25 CO2 No Retrofit cases, respectively.
■
DISCUSSION Results presented here for thermoelectric freshwater withdrawal and consumption in a carbon constrained world are different in many respects from past studies.8−10 The main reason is this analysis is framed in the context of recent trends in electric power generation expansion; specifically, the strong shift from coal to combustion/combined cycle natural gas and renewables,2 as well as the movement away from open-loop cooling.24 Movement away from high water use coal to low water use natural gas/renewables results in a measurable drop in projected thermoelectric freshwater demand. In cases where much of the coal fleet is retired, a decrease in the net thermoelectric water use is realized, thus making some of the retired freshwater available for other uses. While neither a complete retrofit or retirement of the fleet of impacted coal plants is likely, as explored here, the analysis provides reasonable upper and lower bounds subject to the viability of current electric sector trends into the future. Also of importance is the fact that potential changes in the thermoelectric sector are significant in relation to the growth projected in all other water use sectors. Under the Full Retrofit cases new thermoelectric freshwater withdrawal would represent about 8% of all new freshwater demands in the U.S. over the period of 2009 to 2035. In contrast, under the No Retrofit cases thermoelectric withdrawals would be reduced enough to offset new withdrawals in all other water use sectors. Under the Full Retrofit cases future thermoelectric consumption would represent about 15% of all new consumptive
■
AUTHOR INFORMATION
Corresponding Author
*Phone: (505)844-6025; fax: (505)844-8558; e-mail: vctidwe@ sandia.gov. Notes
The authors declare no competing financial interest. 8946
dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
Environmental Science & Technology
■
ACKNOWLEDGMENTS
■
REFERENCES
Article
(18) Hutson, S. S.; Barber, N. L.; Kenny, J. F.; Linsey, K. S.; Lumia, D. S.; Maupin, M. A. Estimated Use of Water in the United States in 2005; U.S. Geological Survey Circular 1268: Reston, VA, 2005. (19) Solley, W. B.; Pierce, R. R.; Perlman, H. A. Estimated Use of Water in the United States in 1990; U.S. Geological Survey Circular 1081: Reston, VA, 1990. (20) Solley, W. B.; Merk, C. F.; Pierce, R. R. Estimated Use of Water in the United States in 1985; U.S. Geological Survey Circular 1004: Reston, VA, 1988. (21) Stewart, D. W.; Rea, A.; Wolock, D. M. USGS Streamgages Linked to the Medium Resolution NHD Geospatial Data Presentation Form: Vector Digital Data; US Geological Survey DS-195: Reston, VA, 2006; http://water.usgs.gov/GIS/metadata/usgswrd/XML/ streamgages.xml#stdorder. (22) U.S. Water Resources Council. The Nation’S Water Resources: 1975−2000, Second National Water Assessment; U.S. Water Resources Council: Washington, 1978. (23) U.S. Environment Protection Agency. EPA Analysis of the American Clean Energy and Security Act of 2009; U.S. Environment Protection Agency: Washington, DC, 2009. (24) National Pollutant Discharge Elimination SystemProposed Regulations to establish requirements for cooling water intake structures at existing facilities; Notice of data availability related to impingement mortality control requirements. 40 CFR Parts 122, 123, 124, and 125, June 11, 2012, Washington, DC, Fed. Regist. 77(112), 34315−34326. (25) Roy, S. B.; Chen, L.; Girvetz, E. H.; Maurer, E. P.; Mills, W. B.; Grieb, T. M. Projecting water withdrawal and supply for future decades in the U.S. under climate change scenarios. Environ. Sci. Technol. 2012, 46, 2545−2556. (26) U.S. Bureau of Reclamation. WATER 2025 Preventing Crises and Conflict in the West; U.S. Bureau of Reclamation, 2010. (27) Li, H.; Chien, S.-H.; Hsieh, M.-K.; Dzombak, D. A.; Vidic, R. D. Escalating water demand for energy production and the potential for use of treated municipal wastewater. Environ. Sci. Technol. 2011, 45 (10), 4195−4200.
We express our appreciation to three anonymous reviews for their helpful and insightful comments. Support for this project was derived through DOE’s Office of Policy and International Affairs and through the National Energy Technology Laboratory. Sandia National Laboratories is a multiprogram laboratory managed and operated by Sandia Corporation, a wholly owned subsidiary of Lockheed Martin Corporation, for the U.S. Department of Energy’s National Nuclear Security Administration under contract DE-AC04-94AL85000.
(1) U.S. Environmental Protection Agency. Inventory of U.S. Greenhouse Gas Emissions and Sinks: 1990−2009. EPA-430-R-11-005; U.S. Environmental Protection Agency: Washington, DC, 2011. (2) U.S. Energy Information Administration. Annual Energy Outlook, U.S., 2012; U.S. Energy Information Administration: Washington, DC, 2012. (3) Zhai, H.; Rubin, E. S.; Versteeg, P. L. Water use at pulverized coal power plants with postcombustion carbon capture and storage. Environ. Sci. Technol. 2011, 45, 2479−2485. (4) Rao, A. B.; Rubin, E. S. A technical, economic, and environmental assessment of amine-based CO2 capture technology for power plant greenhouse gas control. Environ. Sci. Technol. 2002, 36 (20), 4467− 4475. (5) Zhai, H.; Rubin, E. S. Performance and cost of wet and dry cooling systems for pulverized coal power plants with and without carbon capture and storage. Energy Policy 2010, 38 (10), 5653−5660. (6) Rochelle, G. T. Amine scrubbing for CO2 capture. Science 2009, 325, 1652−1654. (7) Macknick, J.; Newmark, R.; Heath, G.; Hallett, K. C. Operational water consumption and withdrawal factors for electricity generating technologies: a review of existing literature. Environ. Res. Lett. 2012, 7 (4), 045802 DOI: 10.1088/1748-9326/7/4/045802. (8) National Energy Technology Laboratory. Estimating Freshwater Needs to Meet Future Thermoelectric Generation Requirements, 2009. DOE/NETL- 400/2009/1339; National Energy Technology Laboratory: Pittsburgh, 2009. (9) Chandel, M. K.; Pratson, L. F.; Jackson, R. B. The potential impacts of climate-change policy on freshwater use in thermoelectric power generation. Energy Policy 2011, 39, 6234−6242. (10) Sterman, J. D. Business Dynamics, Systems Thinking and Modeling for a Complex World; McGraw-Hill: Boston, 2000. (11) Tidwell, V. C.; Kobos, P. H.; Malczynski, L. A.; Klise, G.; Castillo, C. R. Exploring the water-thermoelectric power nexus. J. Water Plan. Manage. 2012, 138 (5), 491−501. (12) Emissions & Generation Resource Integrated database (eGRID), version 1.0; U.S. Environmental Protection Agency: Washington, DC, 2010. (13) U.S. Energy Information Administration. Steam-Electric Plant Operation and Design Report, Rep. EIA-767; U.S. Energy Information Administration: Washington, DC, 2009a. (14) U.S. Energy Information Administration. Annual Electric Generator Report, Rep. EIA-860; U.S. Energy Information Administration: Washington, DC, 2009b. (15) National Energy Technology Laboratory. 2007 Coal Plant Database. 2007. http://www.netl.doe.gov/energy-analyses/hold/ technology.html. (16) Kenny, R. F.; Barber, N. L.; Hutson, S. S.; Linsey, K. S., Lovelace, J. K.; Maupin, M. A. Estimated Use of Water in the United States in 2005; U.S. Geological Survey Circular 1344: Reston, VA, 2009. (17) Solley, W. B.; Pierce, R. R.; Perlman, H. A. Estimated Use of Water in the United States in 1995. U.S. Geological Survey Circular 1200: Reston, VA, 1995. 8947
dx.doi.org/10.1021/es3052284 | Environ. Sci. Technol. 2013, 47, 8940−8947
