Regional Modeling for Stationary Source Control Strategy Evaluation WESTAR Conference on BART Guidelines and Trading September 1, 2005 Tom Moore -

Presentation Topics Purpose is to gather input on regional modeling approach & analysis steps Review selected materials from Evaluation of the CALPUFF Chemistry Algorithms, A&WMA Annual Meeting, June 21-25, 2005, Ralph Morris, S. Lau & B. Koo, ENVIRON International Corporation Suggest process and timing for preparation and completion of Regional Modeling Protocol by the WRAP Regional Modeling Center

Challenges in Evaluating source-by-source BART as a Control Strategy Just one control strategy option, addressing just one approach to regional haze, required for analysis in isolation of other measures Reasonable progress toward RHR goal of natural conditions by 2064 will need to consist of many control strategies for different visibility precursors (e.g., ozone, PM, FP, etc.) Natural visibility conditions, especially the effect of natural fire emissions, major component even for Worst 20% days Need to design evaluation of BART as a complement, a part of an integrated point source emission reduction program

Regional Modeling Analysis When 2018 base case modeling analysis is completed in late 2005 –Shows how much reasonable progress is being made from federal mobile source programs and point/area rules on the books –Then begin to evaluate “WRAP region Point Source Control Scenarios” RMC workplan task to address BART: –Can provide met data as needed by regulatory agencies –Compile/coordinate results from individual agency analyses –Develop regional modeling protocol to support evaluation of point source control strategy options

Planning Emissions and Modeling Analyses for the “Don’t CAIR” region First need to think through questions that need to be answered Need information about agencies’ plans Need to compile sufficiently complete information and/or agreement from regulatory agencies using “source- specific BART” analysis approach –About estimates of emissions reductions –About results from CALPUFF modeling –Need information by [very] early 2006 Agencies selecting “reasonable progress” option allows RMC to apply regional model to BART and other point sources

Issues to be analyzed with modeling Class I Area AQ and AQRV Issues –PSD Pollutants (PM 10, SO 2, NO 2 ) –Visibility –Acid (S and N) Deposition Visibility is frequently the most limiting Class I Area AQ/AQRV issue in NSR/PSD permitting and is the issue associated with Regional Haze Rule (RHR) BART requirements –Visibility impairment for such sources is primarily associated with secondary sulfate (SO4) and nitrate (NO3) fine particulate matter (PM 2.5 )

Technical Assessment Options 3 Options: –Individual source assessment –Cumulative assessment of all BART-eligible sources –Assessment based on model plants

EPA’s Analysis Options Individual Source Assessment (source-by-source analysis) –Use CALPUFF or other EPA approved model –Compare to natural background –“Cause” = impact of 1.0 deciview or more –“Contribute” – 0.5 deciview (State may set lower threshold) Cumulative - Consider all eligible sources to be subject, based on an analysis of an area’s contribution to visibility impairment -- or demonstrate that no sources are subject, based on cumulative modeling. Develop model plants to exempt sources with common characteristics –BART Guidelines provide example model runs

Using CALPUFF? Sulfur and nitrogen emissions from most sources will primarily be in the form of SO 2 and NO X The SO 2 and NO X emissions are converted to sulfate (SO 4 ) and nitrate (NO 3 ) in CALPUFF CALPUFF SO 4 and NO 3 estimates drive the visibility assessment  Therefore the CALPUFF SO4 and NO3 chemical formation algorithms are critically important to the NSR/PSD and RHR BART process

Using CALPUFF? EPA 2001 draft “Guidance for Demonstrating Attainment of the Air Quality Goals for PM 2.5 and Regional Haze”: –“States should use a regional scale photochemical grid model to estimate the effects if a control strategy on secondary components of PM. Changes in primary components may be estimated using a numerical grid model (with no chemistry), a Lagrangian model, or in some cases a receptor model” (EPA, 2001, pg. 169)  CALPUFF is a Lagrangian puff model that EPA FP modeling explicitly states should not be used for SO 4 and NO 3 impacts

Using CALPUFF? Lack of temperature effects and 50°F minimum temperature used in development will overstate sulfate and nitrate formation under cold conditions –Enhanced photochemistry, radicals and sulfate and nitrate formation rates occurs on hot summer days –MESOPUFF-II transformation rates developed using temperatures of 86, 68 and 50°F –Therefore, inappropriate for use below 50 °F  MEOSPUFF-II will overstate SO 4 and NO 3 formation results as temperatures below 50°F (10°C)

Using CALPUFF? Previously evaluated CALPUFF MESOPUFF-II and RIVAD SO4 and NO3 formation rates against latest state-of-science chemistry modules developed by Carnegie Mellon University (CMU) Reported on at AWMA Specialty Conference on Air Quality Model Guidelines in Mystic, CT October 2003 –SO 4 and NO 3 formation chemistry out of date and inaccurate Developed in 1983; Overly Simplistic; Neglects Major Variables and Processes –CALPUFF should not be used for SO 4 and NO 3 Supported by EPA AQ guideline and FP Guidance Supported by chemical modeling community –Likely overestimation bias for large NOx/SO2 sources

Can you evaluate CALPUFF Chemistry using Real-World Data? Extensive PM measurement networks across US that measure real-world SO4 and NO3 –IMPROVE, CASTNet, STN, SEARCH CALPUFF typically applied to one or small group of sources – inconsistent with measurements that are due to all sources –Running CALPUFF with all sources become computationally prohibitive or must perform extensive source combination

Reduce SO2 and NOx emissions by 20 and 10 TPD, respectively Compare maximum 24-hr visibility, SO4 and NO3 impacts in rings around source Compare CMAQ- LISBON w/ MESOPUFF-II and RIVAD CALPUFF chemistry with CMAQ V4.4 9 source locations: GA; NY, OH, IL, TX, ND, WY, AZ and OR Evaluation of Single Source Impacts

large over- prediction using CALPUFF chemistry algorithms compared to CMAQ chemistry > 100% > 1000% Visibility comparisons (Mm -1 ) for January 2002 and 9 sources

CALPUFF chemistry algorithms generally smaller compared to CMAQ chemistry > 100% > 1000% Visibility comparisons (Mm -1 ) for July 2002 and 9 sources

Conclusions – Evaluation of CALPUFF Chemistry SO 4 and NO 3 formation chemistry out of date and inaccurate –Developed in 1983; Overly Simplistic; –Neglects Major Variables and Processes Greatly overstates SO4 and NO3 in winter resulting in overestimating visibility impacts by 100% to 1000% in many cases Understates SO4 in summer, overstates NO3 CALPUFF Nitrate (NO3) particularly inaccurate, overstated and unreliable

Potential process for Regional Modeling Protocol Compile 2018 BART source likely emissions reductions data – 9-12/05 Compile source-specific BART modeling data – 9-12/05 Same time - prepare master spreadsheet of individual BART sources –Emissions (projected 2018 base case and PTE) –Expected emissions reductions beyond current controls –Other supporting data Run “zero-out” simulations of NH3 and VOC at BART sources using tagged emissions in regional model – late 2005 When spreadsheet complete as possible, compile regional “BART emissions reductions” EI[s] and tag those emissions in model to estimate change from “applying” BART Conduct other regional analyses of SO2 Annex (expansion), other alternative programs, etc.

Comments/questions/next steps?