Air Pollution Model Parameter Estimation Using Simulated LIDAR Data

Abstract

One way to obtain estimates of the unknown parameters in a pollution-dispersion model is to compare the model predictions with remotely sensed air-quality data. A ground-base LIDAR sensor provides relative pollution-concentration measurements as a function of space and time. The measured sensor data are compared with the dispersion-model output through a numerical estimation procedure to yield parameter estimates that best fit the data. This overall process is tested in a computer simulation to study the effects of various measurement strategies. Such a simulation is useful prior to a field measurement exercise to maximize the information content in the collected data. Parametric studies of simulated data matched to a Gaussian plumedispersion model indicate the trade offs available between estimation accuracy and data-acquisition strategy.