Use of the CAPTEX Data for Evaluations of a Long-Range Transport Numerical Model with a Four-Dimensional Data Assimilation Technique

Abstract

A ?four-dimensional data assimilation? technique is employed in a time-dependent, three-dimensional mesoscale model to simulate long-range pollutant transport and diffusion in the eastern United States using the 1983 Cross-Appalachian Tracer Experiment (CAPTEX) data. CAPTEX deployed 19 rawinsonde stations to measure upper-air meteorological conditions four times daily and 86 automatic sequential air samplers to measure tracer concentrations from a point source. The total area coverage of the data network is approximately 1000 km (east-west) ? 800 km (north-south). The assimilated wind fields and model-produced turbulence fields during a period of 2¼ days are used to simulate plume trajectories and surface concentrations through a random-particle statistical method. Two tracer releases in the CAPTEX are investigated: one was in a light-wind fair weather condition and produced a widely spread puff distribution; the other was associated with a surface cold front resulting in a rather narrow Puff distribution. The observed winds are successfully assimilated in both cases except in the period of the cold front passage, suggesting that a finer temporal resolution of the rawinsonde observations is desirable in dealing with special weather conditions. The general patterns of the puff distributions are also well simulated. Quantitatively, 57% of the modeled concentrations are within a factor of 4 in comparison with the observed concentrations in the light-wind case.