Thermal effluent from the power sector: an analysis of once-through cooling system impacts on surface water temperature

In the United States, there are currently 425 power plants that use once-through cooling, comprising just under a third of total electricity generation (UCS 2012). Once-through cooling systems circulate water through a plant a single time to provide cooling during generation. These systems require large volumes of water, which is extracted from rivers, lakes, or the ocean, on the order of 20 000–60 000 gallon MWh⁻¹ of electricity produced (Macknick et al 2012). This water is cycled through the cooling system and then is discharged, transferring waste heat from the power plant into the discharge body of water. The discharge water body is usually also the source of the cooling water, causing a local temperature increase. Typical discharge temperatures from once-through cooling systems are 8 to 12 °C above intake temperatures, with some systems raising temperatures as much as 15 °C (Langford 2001). This heated discharge water then mixes with the receiving water body, with temperature impacts dissipating downstream through radiant transfer or evaporation into the atmosphere (Edinger et al 1968). Variability between power plant discharge temperatures can be due to differences in local climates, plant efficiencies, or volumes of water withdrawn for cooling. Under the Federal Energy Administration Act of 1974 (Public Law 93-275), the US Department of Energy (DOE) energy information administration (EIA) is required to collect data on the operations, management, and ownership of electricity generators and distribution companies in the United States. Within this process, the EIA collects environmental data such as water withdrawals, discharges and associated temperatures at the request of the US Environmental Protection Agency. This data is nominally required for all thermoelectric power plants with nameplate capacities greater than 100 MW, although reporting is voluntary (Wirman 2012). Prior to 2010, only the maximum intake and discharge temperatures were collected for the summer and winter months with peak electrical generation (EIA Form 767).

Due to the biological sensitivity of many aquatic organisms to water temperature, temperature increases caused by power plant discharges may have multiple impacts on aquatic ecosystems (Langford 2001, De Vries et al 2008, Hester and Doyle 2011). Aquatic organisms are highly dependent on specific thermal conditions in aquatic environments; water temperatures above or below optimal thermal regimes can cause stress or even death (Beitinger et al 1999, Caissie 2006). It has been concretely demonstrated that fish cannot survive at temperatures above a critical thermal maximum for a certain duration. This critical maximum rarely occurs naturally outside a power plant's effluent stream, and is high enough to kill many entrained larvae and animals (Beitinger et al 1999, Kelso and Milburn 1979). For example, a suitable habitat for rainbow trout (Oncorhynchus mykiss) has an optimal temperature range of 13–15 °C, with the lethal maximum of 24.3 °C (Bear et al 2007). Temperatures above the critical thermal maximum create uninhabitable conditions for the rainbow trout. Higher stream temperatures also decrease oxygen solubility while increasing respiration rates, both of which reduce the availability of dissolved oxygen. Lowered oxygen in aquatic environments can limit the distribution of fish and macro-invertebrates, reduce growth rates, and alter nutrient and carbon cycling (Langford 1990). Elevated temperatures can also stress organisms, increase the toxicity of chemicals, and inhibit biological processes. One review of 151 toxicology studies found that high temperatures typically increase aquatic organism vulnerability to chemicals such as ammonia, heavy metals, and pesticides (Heugens et al 2001). Another review of 48 studies found that on average, a 7 °C change in temperature (ΔT) reduced biological processes such as growth, development, and reproduction for aquatic organisms by 50%, with a 10% reduction in biological processes occurring with a temperature change of only 1 °C (Hester and Doyle 2011). While the thermal impacts of discharges dissipate downstream from power plants, the magnitude of temperature increases (ΔT) in effluent near the discharge point can be high enough to potentially impact aquatic life.

In the United States, discharges of thermal effluent are regulated by section 316(a) of the Clean Water Act, which requires states to set limits for power plant thermal effluent in order to 'assure the protection and propagation of a balanced, indigenous population of shellfish, fish, and wildlife in and on that body of water' (CWA 1972). The primary method of enforcing this law is through the National Pollution Discharge Elimination System (NPDES) permit program, which regulates point sources that discharge pollutants into waters of the United States. The NPDES program is usually administered by state environmental agencies to meet water quality standards, and is overseen at the federal level by the United States Environmental Protection Agency (EPA). In order to obtain permits to discharge heated water, power plants are either required to meet water quality temperature standards, or to obtain a temperature variance, by proving their thermal effluent does not have adverse environmental impacts. Water quality standards vary by state, but typically require surface water to remain under 32 °C (see supplementary data for further information available at stacks.iop.org/ERL/8/035006/mmedia). The duration of NPDES permits is limited by the Clean Water Act to five years, after which the permit must be renewed with the appropriate regulatory agency (EPA 2012). In practice, the stringency in enforcing the Clean Water Act varies on a case-to-case basis, and limits are not always adequately set or enforced to protect aquatic life (Duhigg 2009, GAO 2009). In 2011, the EPA found the North Carolina Division of Water Quality had issued temperature variances to power plants that did not provide enough information to determine whether the revised thermal limits harmed aquatic life (EPA 2011a).

Datasets for power plant characteristics, summer peak intake temperature (T_pin), and outflow water temperature (T_pout) from once-through freshwater cooling systems were compiled from the US Energy Information Administration form 767 for the years 1996–2005 (EIA 1996–2005). The data required that maximum peak temperature be collected at the intake station and the outflow stations for each cooling system during the winter and summer season. Plants were allowed to estimate the water temperature when readings were unavailable and collected only one data point for each season over the given time period (EIA 2005). The temperature provided indicates the maximum cooling water temperature at the intake and outflow for the 'peak load month', the month of greatest electrical generation at the plant. The seasons are defined as winter, from October to March and summer from April to September. Our study focuses on temperature data from the summer season, when electricity generation peaks and causes the largest associated thermal impacts.

This data was vetted to select for power plants with primary cooling systems designated by the EIA as once-through and using freshwater, with five years of complete, non-zero records for the periods 1996–2000 and 2001–2005. The time periods 1996–2000 and 2001–2005 were analyzed separately as the EIA did not require nuclear power plants to report to the EIA after 2000. A number of power plants did not report their cooling technologies to the EIA, so we compared the plants in our dataset to a more complete and detailed database of United States power plants in 2008 (Averyt et al 2013, UCS 2012) to determine what percentage of total fleet of once-through cooled plants were reporting temperature data. The temperature data was averaged over 1996–2000 (n = 418) and 2001–2005 (n = 403) and converted from Fahrenheit to Celsius. We also calculated ΔT_p, or the difference between peak intake and discharge temperatures, where ΔT_p = T_pout − T_pin. We used histogram analysis to evaluate the distribution of ΔT_p,T_pout, and T_pin, and gauge the normality of the data.

We outlined potential impacts of thermal discharges from power plants on aquatic species by comparing two indicators: peak discharge temperatures against counts of threatened and endangered aquatic obligate species at the watershed level. Along with temperature, power plants pose a number of adverse impacts to aquatic species, including impingement (when organisms are trapped against intake screens), entrainment (when organisms get drawn through the plant cooling system), and chemical pollution. However peak discharge temperatures are both reported nationally and regulated, making it an appropriate indicator for broad-scale analysis. While discharge temperatures do not necessarily affect endangered species counts, high discharge temperatures can lower the concentration of dissolved oxygen, inhibit biological processes, and stress sensitive aquatic species by exceeding thermal tolerances (Cairns et al 1975, Langford 1990, Heugens et al 2001, Hester and Doyle 2011). To protect aquatic ecosystems, states limit temperatures through their water quality standards, typically to 32 °C or lower (see supplementary data for further information available at stacks.iop.org/ERL/8/035006/mmedia).

Aquatic obligate species are defined as any species that belong to any taxonomic group that can be classified as aquatic in the absence of habitat data. These data were obtained from a dataset compiled from state-level natural heritage programs (NatureServe 2011) mapped at the USGS Hydrologic Unit Code-8 (HUC-8) level using a digitization of USGS watershed boundaries (Steves and Nebert 1994). Power plant locations were obtained using latitude and longitudes from the Union of Concerned Scientists Energy-Water Database (UCS 2012). At the location of each power plant, average T_pout (2001–2005) was mapped and projected onto HUC-8 watersheds with counts of aquatic obligate species that were either G1 (critically imperiled), G2 (imperiled) or federal status endangered species under the US Endangered Species Act (NatureServe 2011). We also identified watersheds where aquatic ecosystems with high aquatic biodiversity were potentially at risk from high discharge temperatures, defined as T_pout ≥ 32 °C. As a proxy for high aquatic biodiversity, we selected the 220 HUC-8 watersheds that had more than 10 G1, G2 or endangered aquatic obligate species, out of 1963 watersheds with data. To scope areas for more focused study, we identified and mapped watersheds that had both high discharge temperature and high aquatic biodiversity.

From the results of the spatial analysis, we selected one watershed, the Upper Catawba, for a more focused case study. The Upper Catawba was designated as a watershed where high biodiversity was potentially at risk from thermal pollution, as power plants in the Upper Catawba basin had peak discharge temperatures that exceeded regulatory temperature limits during the summer. The Upper Catawba also illustrates some of the regulatory issues that occur with enforcement of the Clean Water Act in regards to thermal effluent.

The Upper Catawba River basin spans both North and South Carolina, supplies millions of residents with drinking water and recreational areas, and provides five major (>100 MW) thermoelectric power plants with cooling water, generating over 30% of North Carolina's electricity (Milazi 2009, UCS 2012). Of these 5 power plants, 3 are coal-fired, 2 are nuclear, and all 5 utilize once-through cooling systems. The Upper Catawba provides aquatic habitats for numerous organisms, including thirty-nine separate species of fish (SCDNR 2008, NatureServe 2011).

In order to analyze the thermal impacts of the 5 power plants along this river, data were distilled from the EIA temperature dataset compiled from 1996–2005. Additionally, NPDES permits were obtained for each plant (NCDENR 2005, SCDHEC 2005, NCDENR 2011a, 2011b, 2011c, 2011d) in order to compare reported data with state and federal regulations for power plant thermal effluent (NCDENR 2007, SCDHEC 2008).

As table 1 indicates, average ΔT_p is between 9.5° and 10 °C. While this was calculated specifically for peak discharge temperatures in the summer, these values are generally comparable to ΔT values found by other studies (Langford 2001, Hester and Doyle 2011). In 1996–2000, 233 out of 418, or 56% of cooling systems had T_pout values that were above the benchmark temperature of 32 °C in all years. In 2001–2005, 230 out of 403, or 57% of cooling systems had peak discharge temperatures that were above the benchmark temperature of 32 °C in all years.

We also looked at how completely plants reported temperature data, both on the scale of cooling systems and individual plants. Of the cooling systems that reported temperatures, we found that 68% of individual cooling systems and 61% of all power plants had complete, non-zero temperature records for the years in our study (see supplementary data for further information available at stacks.iop.org/ERL/8/035006/mmedia). Nuclear plants had the most incomplete temperature records. From 1996–2000, 66% of nuclear power plants reported complete, non-zero intake and discharge temperatures, and nuclear power plants did not report any temperature data to the EIA after 2000.

Mapping discharge temperatures against watershed counts of G1 and G2 aquatic obligate species (as an indicator of high aquatic biodiversity) allowed us to identify watersheds that where further analysis of potential ecological risks may be useful. Figure 1 shows that the majority of once-through freshwater cooling systems are located in the eastern half of the United States (GAO 2009). In the period 2001–2005, 220 HUC-8 watersheds had 10 or G1 and G2 aquatic obligate species and 136 HUC-8 watersheds had average T_pout equal to or greater than 32 °C. 33 basins had both high temperature discharges and high biodiversity, primarily located in the South-Atlantic Gulf, Tennessee, and Ohio River Watershed Resource Regions (figure 2). We selected one of the 33 basins, the Upper Catawba River Basin, for further analysis. The Upper Catawba watershed had 11 counts of G1+G2 aquatic obligate species and an average T_pout of 37 °C in 2001–2005.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

While EIA temperature data was available for all five major power plants from 1996 to 2000, the two nuclear plants (Catawba Nuclear Station and McGuire Nuclear Station) were not required to report temperature data from 2001–2005 (EIA 2010). Temperature data was reported only as summer peak intake and discharge temperatures for the highest month of electrical generation, disguising diurnal variations, sampling rates, and the frequency of thermal discharges above thermal limits. All power plants with available data reported discharging water that exceeded state limits on ΔT and maximum discharge temperatures during the summer (figure 3). However, their NPDES permits revealed that all five power plants had been granted thermal variances that allowed them to exceed state water quality limits, instead requiring power plants to keep thermal discharges below higher monthly averages and ΔT values (NCDENR 2005, 2011b, 2011c, 2011d, SCDHEC 2005). These variances may not be sufficient for Clean Water Act compliance. In the summer of 2007, the combined effects of drought and high temperatures caused Duke Energy to scale back power production at the GG Allen Steam Station and the Riverbend Steam Station when the discharge temperatures at these two plants exceeded their permit limits (Beshears 2007). As part of an ongoing corrective action, the NPDES permits at the Marshall and McGuire power plants have been targeted for direct review by the EPA (EPA 2011b).

Zoom In Zoom Out Reset image size

Our study found that once-through cooling plants discharged freshwater with peak temperatures of 9.5–10 °C above ambient stream temperatures during the summer from 1996–2005. Other studies have indicated that substantial aquatic ecosystem impacts can occur with temperature increases of 5 °C or higher (De Vries et al 2008, Hester and Doyle 2011), which suggests that the peak temperature discharges during the summer could affect aquatic ecosystems and the organisms that inhabit them. In addition, it appears that once-through cooling systems are currently facing ambient cooling water temperatures close to known regulatory limits and discharging waters that may exceed them. Our study indicated that high peak discharge temperatures most often intersected with the highest counts of imperiled aquatic obligate species in the South-Atlantic Gulf, Tennessee, and Ohio River basins, where aquatic biodiversity is generally high and once-through cooling is more common (figure 2). This corroborates the findings of Van Vliet et al (2012), which found that climate change will raise the likelihood of plants running into conflict with regulatory limits as ambient temperatures increase.

While thermal discharges from once-through cooling systems appear likely to impact aquatic ecosystems, the current datasets available from the EIA are insufficient to evaluate watershed-scale impacts to aquatic biodiversity and ecosystems. Given the inherent problems with the availability of data, the scope of our assessment has been limited. In order to assess the impact of thermal discharges on in stream temperatures, local information on power plant and river discharge rates, mixing regime, and spatially and temporally comprehensive records of temperatures are needed. The EIA collects data on power plant discharges but it has been found that power plant reporting on the volume of water consumed and discharged may be inaccurate (Averyt et al 2013). Data reporting to the EIA has improved as of 2010 to include average and peak temperatures on a monthly basis going forward (EIA 2010). While the new reporting requirements will provide more comprehensive data in the future, it will take several years before a time series is available to assess temperature trends in cooling source waters or thermal discharges. Although the federal datasets used in this study do not provide enough detail for a full assessment on the impact of once-through cooling systems on aquatic biodiversity, peak temperature data may be comprehensive enough to at least evaluate the risk of current and future regulatory violations. Almost two thirds of US power plants using once-through cooling who are required to report peak intake and discharge temperatures did so continuously from 1996–2005, and less than 1% of these cooling systems reported obviously erroneous data.