United States Environmental Protection Agency
Ariel Rios Building
1200 Pennsylvania Avenue, N.W.
Mail Code: 1101A
Washington, DC 20460
RE: Public Comments on EPA-HQ-OAR-2016-0347, Section 126 Petition from Delaware and Connecticut on Brunner Island Power Plant
We urge the EPA to deny the petition of these states that would require the Brunner Island power plant to add Selective Catalytic Reduction (SCR) technology, to switch to 100% natural gas fuel to reduce nitrous oxide emissions, or to move to a 24 hour maximum average emissions regimen for the following reasons:
- The Brunner Island power plant meets the EPA mandated Pennsylvania State Implementation Plan which limits the nitrous oxide emissions rate to 0.35lbs per million British Thermal Units averaged over 30 days. Meanwhile, Delaware has met the EPA National Ambient Air Quality Standard (NAAQS) of 8 hour maximum limit of 75 parts per billion (PPB) for the fourth highest day, averaged for ozone over the last three years, and even met the new 70 PPB limit in 2015. No action is required under current regulations.
- A $100 million construction project to convert the Brunner Island power plant to co-firing potential with coal or natural gas will be complete in the spring of 2017, and will significantly lower nitrous oxide emissions in the ozone season (May 1 to September 30) reducing any impact on Delaware.
- Power plants account for only 7% of nitrogen dioxide (NO2) levels in Delaware, and totally eliminating this source would not significantly change ozone levels.
- A key assumption of the petitioner’s claims is higher NO2 levels yield higher ozone levels. An analysis of ambient levels of nitrogen dioxide (NO2) and ozone (O3) in New Castle County, Delaware, from 2013 to 2015, show no correlation between higher levels of NO2 and higher levels of ozone.
Brunner Island to Switch to Natural Gas from Coal during the Ozone Season in 2017
A $100 million construction project to convert the Brunner Island power plant to co-firing potential with coal or natural gas will be complete in the spring of 2017, and will significantly lower nitrous oxide emissions which are a precursor for ground level ozone. Natural gas inherently emits 80% less nitrous oxide during combustion compared to coal. Natural gas is used as the primary space heating fuel in 56% of homes in the United States according to the US Energy Information Agency so, prices peak in the winter when demand is high (Graph 1). This is why average generation with natural gas fired power plants peaks every summer when prices are lower (Graph 2). There is a very high probability natural gas will be the fuel used at Brunner Island during the Ozone Season (May 1 to September 30), and would have been used all year in 2015 except for the winter. The co-firing project meets the same ozone season nitrous oxide emission goals as a $600 million1 SCR project would while offering a lower cost impact for electric customers. The project also cuts carbon dioxide emissions in half and lowers sulfur dioxide emissions during the period natural gas is used. Co-firing also offers fuel diversity critical to maintaining electric grid reliability. Over dependence on natural gas could lead to blackouts during severe winters in the future when natural gas is in tight supply and pipelines might freeze as happened during the 2014 polar vortex.
Graph 1: Pennsylvania Power Plant Fuel Cost 2015 by Month – $/MMBTU
Source: US Energy Information Agency Electric Power Monthly Tables 4.10A and 4.13A
Note: Coal prices are adjusted for lower heat rates but higher operating costs for FGD use
Graph 2: Seasonal Electric Generating Unit Coal vs. Natural Gas Generation
Upwind Power Plants Account for a Small Percent of Delaware’s Ambient NO2 Levels
The Delaware petitioner claims that upwind emission sources account for 94% of Delaware’s ozone levels. By omission, the petitioner leaves the impression the transport of ozone from upwind sources is the problem. Ground level (tropospheric) ozone is actually produced locally in two general chemical reactions2:
- The energy of sunlight splits a nitrogen dioxide molecule (NO2) into a nitric oxide molecule (NO) and a free oxygen atom (O). The free oxygen atom combines with an oxygen dioxide molecule (O2) to form ozone (O3). The reaction reverses fairly quickly as nitric oxide and ozone are very reactive. Coal fired power plants emit a combination of nitrogen dioxide and nitric oxide (which quickly reacts to become nitrogen dioxide) and can be transported by the wind. NO2 also occurs naturally, and from a variety of manmade sources, primarily motor vehicles.
- A sunlight split nitrogen dioxide molecule can also react in a complex way with a wide range of Volatile Organic Compounds (VOC) to form ozone which tends to be longer lasting. VOCs can be generated by natural sources, such as vegetation, and by manmade manufacturing and combustion. Natural sources can be overwhelming in places like the Great Smokey Mountain National Park which derives its name from naturally occurring ozone in the form of smog. The VOC source is pinetene from pine trees and ozone levels exceed national air quality standards!
In both cases nitrogen dioxide is considered a catalyst, the reaction can’t occur without it but the amount of nitrogen dioxide is ultimately unchanged over a 24 hour period. Ozone is highly reactive with any physical surface and can cause injury to plants and animals, especially sensitive people with asthma or COPD. The local nature of ozone is confirmed by Graph 3 below which shows how ozone forms in the day and falls to essentially zero overnight. Ozone is also formed in the stratosphere and can be mixed by downdrafts with ground level ozone but that generally occurs in mountainous areas such as the Rockies.
Graph 3: Diurnal Ozone Production
Source: NASA Earth Observatory
There are a number of variables in ozone creation including the amount of sunlight, temperature, wind, humidity, the amount of VOCs, and even ocean breezes as ozone has a longer shelf life over water. The highest ozone days are on hot summer days grouped around the summer solstice, with low wind and moderate humidity, in urban areas with high VOCs.
The petitioner also leaves the false impression power plants are the main source of ozone precursors. Power plants account for only 7% of NO2 levels in Delaware. Electric Generating Units (EGU) emissions are overwhelmed by vehicle and other emission sources as shown in Graph 4 below. The claim the Brunner Island power plant has a significant impact on ground level ozone levels in Delaware is simply wrong.
Graph 4: NOX Contribution Sources in Delaware
Source: Delmarva Power 2012 IRP “Air Quality & Health Impacts Assessment”, ICF International using 2010 to 2013 average data
There is No Correlation between Higher Levels of Nitrogen Dioxide and Higher Levels of Ozone
The Petitioner relies on modeling from Sonoma Technology Inc. (STI). They used a computer model based on work done by the United States Environmental Protection Agency to determine downwind dispersion of various emissions, including nitrous oxide, and the chemical reactions related to the emissions. The model used is the Comprehensive Air Quality Model with Extensions (CAMx) with Ozone Source Apportionment Technology (OSAT). According to the Sonoma report, page 12, “OSAT uses reactive tracers to track the fate of these precursor emissions and the ozone formation resulting from them within a CAMx simulation”. Unfortunately, STI did no real world, on the ground verification of the validity of the tracers.
Ambient levels of NO2 are currently averaging 19 PPB daily one-hour maximum during the ozone season with a range of 5 to 50 PPB over the years 2013 to 2015 (measured at the Martin Luther King Blvd. air quality monitoring station in Wilmington, Delaware). The Wilmington site is the only air monitoring station in Delaware regularly measuring NO2. Monitoring equipment doesn’t always work so we only had data for 346 days out of a possible 459. The 19 PPB is 80% below the NAAQS 98th percentile, three year average maximum allowed of 100 PPB. When we ran a statistical correlation between NO2 levels and ozone levels we found the correlation was 0.08, so essentially there was no correlation between the two! Zero equals no correlation and 1 equals perfect correlation. The computer may have done an excellent job tracking NO2 concentrations, but the Brunner Island simulation only added 1.9 PPB average to ozone levels with a range of 0.7 to 4.9 PPB according to STI. However, real world data does not confirm the computer tracer module results. There was no impact from Brunner Island on Delaware ozone levels.
For further verification we looked at 2013 data from the Lewes, Delaware, air monitoring station which measured NO2 that year. Lewes is a rural station while Wilmington is more urban. The Lewes station showed an ozone season average ambient level of NO2 of only 3 PPB, one-sixth as much as Wilmington, with a range of 0.3 to 10 PPB. The statistical correlation was only 0.01, or once again essentially zero.
The facts do not support the Petitioner’s request to require shorter periods for nitrous oxide emission averaging, such as a 24 hour cap, or to add SCR pollution control equipment at Brunner Island. We note the link to the report is on a Sierra Club website alongside its “Beyond Coal” campaign that counts down how many coal-fired power plants have been closed on the way to zero. We suspect the Delaware and Connecticut Air Quality Divisions have responded to an advocacy group call for action to end the use of coal at Brunner Island. It would not make economic sense to spend $600 million on Brunner Island for SCR on top of the $100 million on-going project to convert to the ability to co-fire with natural gas. If the EPA grants the petitioner’s request the Brunner Island facility will simply convert 100% to natural gas and the Sierra Club can scratch off another coal plant. However, electric customers will be worse off with higher winter time electric rates, and the lost reliability of having fuel diversity. The environmental outcome for ozone comes out the same either way.
1 – Press release from Four Corners Power Plant, Fairmount, NM showing 1636 MW capacity SCR project now under way will cost $635 million. Brunner Island is about 5% smaller at 1549 MW.
2 – “Chemistry in the Sunlight”, NASA Earth Observatory, http://earthobservatory.nasa.gov/Features/ChemistrySunlight/
David T. Stevenson
Director, Center for Energy Competitiveness