The Aerodyne Inverse Modeling System (AIMS): Source estimation applied to the FFT 07 experiment and to simulated mobile sensor data

The Aerodyne Inverse Modeling System was developed to enable location and characterization of hazardous atmospheric releases from dispersion and meteorological data. It combines an automatically-generated tangent-linear of SCIPUFF with a cost function tailored for practical applications and a minimization algorithm that can search for multiple instantaneous or continuous sources without requiring an initial guess. In this work AIMS was applied to estimate the sources in 84 FFT 07 cases that included instantaneous and continuous releases for up to four source locations. FFT 07 was a controlled short-range (～500 m) dispersion test using 100 digiPIDs evenly distributed over an area of 0.5 × 0.5 km. AIMS estimated sources were in average within 90–150 m of the real sources, with the distances from estimated to real source ranging from 0 to 510 m. AIMS performed better estimating the location of instantaneous sources than of continuous ones. It also performed better for single-source situations than for multiple source scenarios and when 16 sensors were used instead of 4. In addition to using stationary sensors, AIMS also has the capability of processing data from mobile sensors. This was applied using model-generated data in an example of a release in a setting similar to an industrial facility.