Near-Source Forward and Inverse Modeling of Reactive Chemicals from Industrial Emission Events

The HARC neighborhood air quality model is a superfine (~200 m horizontal) resolution 3D Eulerian model that can simulate rapid near-source chemistry due to large industrial emissions of nitrogen oxides, highly reactive VOCs (e.g., propylene), and radical precursors (e.g., formaldehyde). It can be run in either forward or adjoint mode, so that inverse estimation of emissions, turbulent diffusion coefficients, and other model parameters may be performed using the 4D variational (4Dvar) data assimilation technique. When used in conjunction with advanced remote sensing or real-time in situ monitoring, the HARC model is a powerful tool for quantifying large and highly transient industrial emission events, for which accurate accounting is presently very difficult. A real-world model application will be presented based on the analysis of mobile laboratory data near petrochemical facilities during the 2009 Study of Houston Atmospheric Radical Precursors (SHARP). In the example to be presented, the results of inverse modeling of formaldehyde emissions from a fluidized catalytic cracking unit and several refinery flares will be compared to direct remote sensing measurements using different varieties of the Differential Optical Absorption Spectrometry (DOAS) technique. Application of the HARC model to tomographic DOAS measurements of hazardous air pollutants (HAPs) during the upcoming Benzene and other Toxics Exposure (BEE-TEX) Study in the Houston Ship Channel will also be discussed, including inverse modeling of industrial emissions of the reactive HAPs, toluene and xylene. Finally, opportunities for application of the HARC model to upstream oil and gas issues will briefly be considered.