 |
NARSTO Workshop on Innovative Methods for Emission Inventory Development and Evaluation University of Texas, Austin October 14-17, 2003 |
 |
The Development of a Stochastic Emissions Inventory for Industrial Emissions in Houston/Galveston Texas and Its Use in Photochemical Modeling
Harvey Jeffries, Mort Webster, Mike Symons
University of North Carolina;
Tom Tesche, Dennis McNally
Alpine Geophysics.
It has become abundantly clear since the Texas Air Quality Study of August 2000 (TEXAQS) and subsequent unprecedented analysis efforts in the last two years that the industrial point source VOC emissions in Houston are poorly represented by the standard Texas Commission on Environmental Quality (TCEQ) and EPA derived emissions inventory reporting systems. This has led to significant difficulties in achieving adequate photochemical model performance needed for State Implementation Plans for Ozone.
Hourly measurements of actual flare flows and weekly cooling tower water concentration measurements, supplied by several large industrial sources as part of an on-going study, and recent research-instrument visualization of actual fugitive emissions of olefins conducted for TCEQ show that the industrial sources are very dynamic, exhibiting a stochastic emissions pattern, but, in the case of flares, with mean annual values that comply with emissions permits. Further, the newly expanded reported upset-database has also been analyzed and characterized.
These industrial measurements have been used to create sets of probability distribution functions (PDFs) that can be applied to the standard model emissions input files produced by TCEQ for use in SIP modeling with the CAMx Photochemical Grid Model. That is, we can describe a point source stochastic emission inventory for HGA. Instances of this emissions inventory can be created with different properties and these can be used in the model to estimate the effects on ozone production. Box model and photochemical grid models differ, however, in their abilities to respond to values drawn from these stochastic emissions inventories. Process analysis studies have been conducted to explain why these models respond differently. Process analysis has also been used to connect these two modeling approaches to permit their use together to explore the consequences of stochastic variation in point source emissions inventories.
In this presentation, we will show many examples of industrial point source measurements for flares and cooling towers. We will show the PDFs that describe these and will connect the shapes of the PDFs with technology at the sources. We will show the results of applying these PDFs to TCEQ´s modeling point source inventory. We will compare selected species PDFs from the stochastic EI and from observations. And we will describe the effects that the uses of stochastic EI´s have on ozone production in box models and in photochemical grid models.