Evaluating and Improving Aircraft-based Emission Rate Retrieval Methodology Using Numerical Air Dispersion Modeling

Abstract:

Airborne measurement of pollutants with the aim of estimating emissions from point and area sources such as oil sands facilities, is an active area of environmental research. During the studies conducted in 2013 as part of Joint Canada Alberta implementation plan for Oil Sands Monitoring (JOSM), aircraft-based measurements of pollutants were made. The Top-down Emission Rate Retrieval Algorithm (TERRA) was developed by Environment and Climate Change Canada (ECCC) to estimate facility emission rates based on these aircraft measurements. Numerical dispersion models can be utilized in evaluating and improving aircraft-based emission rate retrieval methods and strategies. For this work, the Global Environmental Multiscale-Modelling Air-quality and Chemistry (GEM-MACH) model was utilized to create series of high resolution regional air-quality forecasts for the purpose of evaluating aircraft-based emission rate retrieval methods, such as the TERRA method. Within the GEM-MACH model, the Eulerian equations of motion and material balance are numerically solved using the semi-Lagrangian scheme. Processes such as emissions, advection, turbulence, diffusion, deposition, and gas, aqueous, and particle chemistry are simulated in GEM-MACH to generate the net temporal variations in the spatial distribution of chemicals and aerosols. Aircraft-based area source emission rate estimations can be studied by flying a virtual aircraft within the numerical model and extracting model concentrations and meteorological variables along the flight tracks. These quantities can be used in the TERRA algorithm to estimate emissions, which may then be compared to model input emissions. This comparison allows us to generate estimates of errors in retrieved emissions, and identify the sources of errors when the algorithms are used with actual aircraft data. We show that the gradient Richardson number is a useful a priori forecast variable for evaluating potential error in emission estimates. Our analysis to date suggests that the emissions error associated with the use of TERRA is in the range 5-20%. Comparisons will be made between this error range and literature-based emissions estimates using aircraft data and TERRA. The impact of averaging time in emissions retrievals will also be discussed.