Development of Transformation Rate of SO2 to Sulfate for the Houston Ship Channel using the TexAQS 2006 Field Study Data
On June 2, 2010, EPA promulgated a new 1-hour SO2 primary NAAQS with a threshold of 75 ppb. The 1-hour SO2 NAAQS is much more stringent and replaces the old 24-hour (140 ppb) and annual (30 ppb) SO2 NAAQS. States are required to submit 1-hour SO2 State Implementation Plans (SIPs) by February 2014 that demonstrates compliance with the NAAQS by August 2017. Preliminary modeling indicates that SO2 emissions for numerous sources will result in near-by exceedances of the 1-hour SO2 NAAQS. Fossil-fueled power plants (73%) and industrial facilities (20%) are the main sources of SO2 emissions in the U.S. Photochemical oxidants will convert some SO2 to sulfate thereby reducing SO2 concentrations. However, the EPA-recommended model for near-source 1-hour SO2 modeling is the AERMOD steady-state Gaussian plume model that does not treat photochemical oxidants and has a very simple treatment of chemistry (exponential decay). EPA recommends that AERMOD be run with no SO2 conversion for addressing 1-hour SO2 NAAQS issues. This assumption may be appropriate for fossil-fueled power plants where the high NOX concentrations inhibit photochemistry and consequently SO2 oxidation near the source, but it may not be appropriate for the Houston Ship Channel where the atmosphere can be very reactive (due to HRVOC emissions) resulting in faster SO2 to sulfate conversion rates.
The goal of this project is to develop a representative SO2 transformation rate for the Houston Ship Channel area using measurements from the NOAA P-3 aircraft collected during the 2006 Texas Air Quality Study (TexAQS) that can be used with the AERMOD model to simulate 1-hour SO2 concentrations. The proposed approach uses a grid model to simulate first-order transformation of SO2 to sulfate for sources in the Houston Ship Channel. The model results with varying transformation rate are evaluated against the 2006 TexAQS P-3 aircraft measurement data to find what transformation rate best fits the observations and to determine whether one hypothetical transformation rate results in statistically better model performance than the other rates used.