Sunday, June 23, 2024

GAO Information on Peaking Power Plants Report – Watts Up With That?

Must read

Roger Caiazza

Environmental Justice (EJ) advocates like the PEAK coalition argue that “Fossil peaker plants in New York City are perhaps the most egregious energy-related example of what environmental injustice means today.”   This post critiques a recent General Accounting Office (GAO) report on “Information from Peak Demand Power Plants” that was prepared in response to a question about pollution from these facilities by some congressional representatives.. 

The GAO website summary for the report explains why they did the study:

Environmental advocacy groups, and some congressional leaders have expressed concerns that peakers may be less efficient than non-peakers, meaning peakers may expend more energy that is not converted into electricity than other types of plants. Further, due to the nature of their operations, peakers may also negatively affect the air quality in communities around the plants, which may be historically disadvantaged or disproportionately low-income.

GAO was asked to examine pollution from peakers across the nation. This report provides information on the number and location of peakers in the U.S., their proximity to historically disadvantaged or disproportionately low-income communities, to what extent they emit pollutants and how these pollutants affect the health of people exposed, and alternatives for replacing them. To perform this work, GAO analyzed data from EPA, the U.S. Department of Energy, and other sources, reviewed relevant literature, and interviewed federal officials and stakeholders from 19 state, industry, and nongovernmental organizations representing a diversity of perspectives about peakers.

I am unimpressed with this report.  It is not clear to me whether the political implications of this topic or the naivete of the authors was the reason for the poor quality.  Whatever the reason, the report confirmed the biased concerns of environmental advocacy groups without addressing the fundamental problematic issues associated with peaking power.  It is especially galling that the report ignored air quality protections already in place.

This topic is a particular concern of mine because issues associated with peaking power plants have been one of my responsibilities since 2000.  Initially, my concerns were associated with developing an emissions tracking system to ensure compliance with air quality requirements for peaking plants.  Later I participated in the regulatory process to develop regulations to reduce their emissions but also keep the lights on, keep the costs down, and achieve improved air quality. It took many years but New York State developed a rule that fulfilled those requirements.  I doubt that I am the only one who participated in that process who was taken aback when environmental advocacy groups started campaigning against the power plants covered by the regulations put in place to address the peaking power plant pollution.

Peaking Power Problem

I think the GAO report missed the opportunity to highlight the challenges and implications of peak power demand and how it could and should be addressed.  An Ozone Transport Commission presentation of issues associated with High Energy Demand Days (HEDD) from 2006 describes the tradeoffs.  Air quality and energy planning both prioritize energy demand peaks because the highest electric demand and worst air quality tend to coincide.  This is because the meteorological conditions that cause peak loads also exacerbate the air quality impacts of the increased emissions needed to match peak loads.  Reliably meeting the peaks results in using the dirtiest and most expensive units.  The key energy considerations slide describes the issues.

The GAO report overlooked this aspect of the peaking power plant issue.  Instead, when they explained why this issue matters, they only talked about daily peaks.  I think that devalues the criticality of the peak issue.  The real problem is that peak loads occur when customers need power the most.  If it is unavailable, then immediate acute safety and health problems occur. While the daily peak is a problem it is far less impactful than the annual peak load.

GAO Report

The GAO report determined how many peakers are in the country and where they are located.  This discussion failed to discuss another impactful nuance.  The GAO analysis determined the number of peaking power plants as a function of how much power plants ran in 2021:

For the purpose of our report, we generally define peakers as plants that use fossil fuels, including natural gas, coal, and oil; have a capacity factor (the percent of energy produced over a certain time frame, out of what could have been produced at continuous full power operation) of 15 percent or less; and have a nameplate capacity (the designed full-load sustained output of a facility) of greater than 10 megawatts (MW) of electricity.

This is a similar methodology to that used by Physicians, Scientists, and Engineers (PSE) for Healthy Energy in their report Opportunities for Replacing Peaker Plants with Energy Storage in New York State.  There is an unrecognized shortcoming to the approach.  The GAO report states that “Peakers are used to supplement other types of power plants, such as baseload plants, which run consistently throughout the day and night, and intermediate plants, which run mostly during the day and less at night”.  There is a difference between power plants designed to meet peaking applications and many facilities that now operate as intermediate or peaking units.  For example, around 1970 Consolidated Edison of New York needed peaking capacity within New York City that would only run infrequently but also needed to startup quickly.  They responded by building a fleet of around 100 simple-cycle natural gas turbines that were the cheapest capacity available.  Today many of the units that meet the GAO definition were originally designed as base-load units and cannot start up quickly.  I used to work at the Oswego Harbor generating station that had two 850 MW oil-fired units and took over a day to startup.  The units have met the capacity factor criteria for peaking units for years, but they were not designed to operate that way. 

The GAO report includes a map of plant locations but there is no capability to identify the plants on the map. If you are interested in specific facilities, the EPA Power Plants and Neighboring Communities website presents that information.  The GAO summary lists 999 peaking power plants in the following table.  The overall capacity factor of these plants is 6%. 

The GAO analysis determined how closely peakers are located to historically disadvantaged and low- income communities.  This analysis, the Physicians, Scientists, and Engineers, and even the draft New York regulations to address these facilities all use distance between the disadvantaged communities and the power plant as the metric of concern.  The GAO claims that “For example, based on our model and main definition of a peaker, a community that is 71 percent historically disadvantaged is expected to be 9 percent closer to the nearest peaker than the average community, which is 40 percent historically disadvantaged.” 

I do not think this is an unexpected result, but I also think it is meaningless.  The air quality impacts of any facility do not depend entirely upon distance between the source and a receptor location.  The stack characteristics (gas temperature, height of the stack, and stack exit dimensions) as week as meteorological conditions (wind direction, wind speed, and atmospheric stability) all affect air quality impacts.  The Con Ed turbines had exit ducts that released the pollutants that were less than 100 feet and the location of maximum downwind impact was relatively close.  On the other hand, the Oswego plant had 700’ stacks and the location of maximum downwind impact was quite a way from the plant.

The GAO report addressed the impact of emissions and the resulting health impacts: “When operating, peakers emit similar types of pollutants to other power plants that also use fossil fuels, and these pollutants are associated with various negative health effects, according to existing literature.”  The impacts section notes:

Compared to non-peakers, peakers emitted more pollutants—such as nitrogen oxides and sulfur dioxide—per unit of electricity generated, but fewer total annual pollutants in 2021, according to our analysis of EPA data (see table 2). In other words, peakers emit less in total because there are fewer peakers and they operate less frequently overall than non-peakers. However, when they do operate, they emit more pollution per unit of electricity produced. For example, the median sulfur dioxide emission rate for natural gas fueled peakers was 1.6 times more per unit of electricity generated than the median emission rate for non-peakers.

This language parrots the talking points of EJ advocates but is much ado about nothing.  It is obvious that fewer peakers that run less would have lower emissions than more numerous non-peakers that run more.  Advocates harp on the fact that emission rates are higher for peakers than non-peakers.  Highlighting the finding that “natural gas fueled peakers was 1.6 times more per unit of electricity generated than the median emission rate for non-peakers” is a naïve point because 0.008 versus 0.005 lb SO2 per MWhr is negligible for air quality impacts.  Moreover, I think there is an error in the methodology because the sulfur content in fuel determines the emissions not how it is burned, so there should be no difference in the rates.

Another issue I have with this analysis and other similar analyses is that they don’t recognize that the primary air quality issue with peaking power plants is ozone.  These units operate when energy demand is highest in the summer and those periods are typically hazy, hot, and humid.  Those conditions are conducive to the highest ozone levels (some of the haze) and there has been immense pressure to reduce their emissions to reduce ozone levels.  Ozone is a secondary pollutant produced in a photo-chemical reaction from nitrogen oxides and volatile organic compounds.  The conversion to ozone takes time and means that by the time it occurs the emissions from a power plant in a disadvantaged community have moved downwind.  For example, the location of highest downwind conditions for emissions from New York City is in Connecticut, far beyond neighboring disadvantaged communities.  Moreover the reality is that nitrogen oxides scavenge ozone so that the peaking power plants actually reduce ozone concentrations close to the facility.

I have been involved with air quality issues since I started working in 1976.  The fundamental presumption has always been that the National Ambient Air Quality Standards (NAAQS) is the health metric used to determine health impacts.  EPA explains:

The Clean Air Act, which was last amended in 1990, requires EPA to set National Ambient Air Quality Standards (40 CFR part 50) for six principal pollutants (“criteria” air pollutants) which can be harmful to public health and the environment. The Clean Air Act identifies two types of national ambient air quality standards. Primary standards provide public health protection, including protecting the health of “sensitive” populations such as asthmatics, children, and the elderly. Secondary standards provide public welfare protection, including protection against decreased visibility and damage to animals, crops, vegetation, and buildings.

As an air pollution meteorologist one of my jobs was to run air quality models to determine the air quality impacts of existing and proposed facilities.  The primary consideration was whether the modeling proved that the projected impacts were less than the NAAQS.  The basis of my work was that when I showed compliance with those standards, I proved that we were protecting the health of “sensitive” populations such as asthmatics, children, and the elderly.  Regulatory agencies are required to ensure that any facility that cannot show compliance with the NAAQS must to modify its permitted operations or it cannot be allowed to operate. 

The GAO report does not mention the NAAQS protections.  Instead, the analysts follow the lead of EJ activists and claim that there are health effects from peaking power plants.  The health effects section states: “Multiple pollutants that are emitted from peakers and other plants are associated with various negative health effects for the people exposed, according to federal agency reports we reviewed”.  What they reviewed were the EPA Integrated Science Assessments:

EPA’s Integrated Science Assessments integrate information on criteria pollutant exposures and health effects from controlled human exposure, epidemiologic, and toxicological studies to form conclusions about the causal nature of relationships between exposure and health effects. For more information, see the EPA Preamble for Integrated Science Assessments at Preamble To The Integrated Science Assessments (ISA) | ISA: Integrated Science Assessments | Environmental Assessment | US EPA (accessed 8/30/2023). 

The presumption in the report is that any level of pollution is bad: “For instance, short-term exposure to sulfur dioxide—the indicator for sulfur oxides used in EPA’s assessments—can lead to negative respiratory effects, such as decreased lung function, cough, chest tightness, and throat irritation.”  The GAO report summarizes health effects from short-term exposures.  All this is nice but it ignores the NAAQS process to determine acceptable ambient air quality levels.

The EJ activists pushing the negative impacts of the peaking power plants presume that there are alternatives.  The GAO report looked at some available alternatives that could potentially replace fossil-fueled peakers at the same high-level used throughout the analysis.  The report claims that “alternatives such as battery storage systems could potentially replace fossil-fueled peakers, according to studies we reviewed and stakeholders we interviewed”.   The report lists battery storage, pumped-hydro storage, thermal energy storage, and notes that renewable energy systems (e.g., wind and solar) may be paired with energy storage.  It claims that “adding roof-top solar and battery storage to houses could reduce the demand for peakers in adjacent areas.”  It includes two other possibilities:

Transmission and distribution infrastructure improvements: Upgrades or expansions to increase the capacity of current infrastructure that transmits and distributes electricity.  These upgrades or expansions may help enable existing underutilized plants to meet peak demand.

Efforts to decrease consumers’ use of power during peak times: Efforts to incentivize consumers to reduce or shift electricity use during times of peak use to off-peak times.

To its credit the report does address the “potential challenges of replacing peakers” including cost, reliability, and location.  In my opinion, the responses downplayed those challenges.

The report notes that “some alternatives may have higher capital and operating costs compared to current fossil-fueled peakers”:

Replacing peakers, some of which have already paid off their capital costs, will likely lead to additional up-front or operating costs compared to keeping the existing peakers. Further, the U.S. Energy Information Administration (EIA) reported that solar and wind plants had higher average construction costs compared to natural gas-fired plants in 2023.

It is remarkable that this Federal report documented that construction costs of solar and wind are greater than natural gas plants but there is a missing nuance.  While I am not an economist, I still question what kind of business model could justify developing a new resource that will operate as a peaking facility running less than 15 percent of the time.  Surely the facility will have to charge very high rates when it does operate.

Appropriately the report notes that “current alternatives may not be able to provide the same reliability of current fossil-fueled peakers”:

Similarly, some alternatives may create reliability challenges. For the grid to be reliable, the energy resources in an area need to be able to supply power to meet peak demand for as long as it lasts, according to U.S. Department of Energy (DOE) officials. Some battery storage systems provide up to 4 hours of output, but peak demand may be longer in some areas. In contrast, a fossil-fueled peaker is only limited by fuel availability—a natural gas-fueled peaker could keep operating so long as natural gas is available.

This is an important point universally ignored by the activists that want to shut down peaking power plants now.  The other nuance is that the overly broad definition of a peaking power plant covers facilities that provide different services than just peaking support.  The 1700 MW at Oswego Harbor are within ten miles of three nuclear units.  Nuclear units are required to shutdown when the grid goes down and in the 2003 blackout Oswego Harbor came on line to replace those units until the grid stabilized.

The GAO report also noted that “alternatives may not be able to be installed because of space and location concerns”:

Some alternatives may also run into space constraints or location concerns. For example, a densely populated urban community likely would not have sufficient space for a large renewable energy system paired with battery storage to help meet peak electricity demand.

I agree with this point, but it could have been expanded.  Location matters within the grid.  The transmission system is designed based on the location of the generating resources.  The requirement that energy must be available at the location of the New York City peaking power plants is not acknowledged by the EJ activists, but it is a critical reliability constraint.

Most disappointing to me is that the report does not acknowledge the following challenges to the end of the main report.  I believe that at least a hint of the following information should have been right up front:

In general, recognizing these challenges, some officials with whom we spoke identified trends that may lead to the continued use of fossil-fueled peakers. According to DOE officials, some U.S. peakers may not be able to be replaced with existing alternatives within cost, reliability, and location constraints.

Combinations of electricity generation and storage technologies, transmission and distribution improvements, and efforts to decrease consumer’s use of power during peak times may be too costly for consumers in some areas to provide an adequate level of grid reliability. Further, officials at two utilities noted that due to increased use of intermittent renewable resources on the grid (e.g., wind and solar power), the continued use of peakers to meet electricity demand may be necessary to maintain grid reliability. For example, the availability of sunlight for a solar installation may not match with peak demand in the evening when the sun goes down. Therefore, additional supplemental energy resources would be needed to fill the gaps and meet demand.

It gets worse.  Buried in the technical appendix the last sentence in the last paragraph before the end notes is the caveat that there is no basis for concern (my highlight):

Limitations. We took several steps to assess the validity and sensitivity of our models, but certain limitations remain. Importantly, our measure of distance does not include other aspects—such as stack height, wind speed, or wind direction— that play important roles in the dispersion of pollutants and potential populations exposure. In addition, although we include some variables to control for factors that could influence the findings, it is possible that other controls might be important and were not accounted for in our model. Inclusion of a state fixed- effect partially addresses this by controlling for factors that vary by state. Still, our findings of associations between distance to peakers and historically disadvantaged racial and ethnic communities does not imply any causal relationships.


The GAO was asked to respond to a question about pollution from peaking power plants by some congressional representatives.  The response is a disappointment.  The report summary found that:

  • Historically disadvantaged racial or ethnic communities tend to be closer to peakers.
  • Fossil-fueled peakers are primarily fueled by natural gas and emit air pollutants associated with various negative health effects, including on respiratory, cardiovascular, and nervous systems.
  • Alternatives are available that could potentially replace or provide similar services as peakers, but we identified challenges for their use related to costs, reliability, space, and location.

I do not dispute that disadvantaged racial or ethnic communities tend to be closer to peakers but the fact the “findings of associations between distance to peakers and historically disadvantaged racial and ethnic communities does not imply any causal relationships” indicates that the basis for the concerns is weak.  If the GAO report had evaluated the status of the communities of concern relative to the NAAQS or at least mentioned that there are standards designed to protect those communities, it would have been obvious that this is a non-problem conjured up by activists.

The second finding is another lost opportunity to inject reality into the conversation.  Obviously, fossil-fueled peakers emit air pollutants and they can be “associated with various negative health effects, including on respiratory, cardiovascular, and nervous systems.”  The complete disregard of the NAAQS protections in place unnecessarily scares the residents in the communities of concern. In addition, regarding emissions in isolation, focusing only on the negatives and disregarding any benefits creates an unnecessarily pessimistic outlook, hinders growth, and could lead to unintended consequences.

The final summary point begrudgingly admits that there are “costs, reliability, space, and location challenges” for replacements to peaking power plants.  The GAO report should have emphasized these challenges in my opinion.  The reality is that in order to deal with peaking power plants and the net-zero ambitions we do not have the generating or transmission technologies needed. 

I conclude that this analysis and the report were intended to confirm the biases of the congressional representatives that requested the report and the EJ activists who promote it.  It is only possible through a complete reading of the entire report to discover contrary evidence eviscerating this as an issue.  The bigger picture problem is the potential threat that political and activist pressure will force premature retirement of peaking power plants with a marked increase in potential reliability risks.  A blackout will have real ramifications as opposed to the over-hyped risks claimed.

Roger Caiazza blogs on New York energy and environmental issues at Pragmatic Environmentalist of New York.  This represents his opinion and not the opinion of any of his previous employers or any other company with which he has been associated.

More articles


Please enter your comment!
Please enter your name here

Latest article