Use of evolutionary computation for isolating surface emissions from orbit

Abstract

High precision targeting of localized sources of atmospheric species outgassed from planetary surfaces remains an important goal in planetary science, but one that largely remains beyond present-day observational capabilities. For disequilibrium trace species, the detectable signature of a gas in the atmosphere often is confined to a small region (order 10-100's km) around its surface source. Identifying a plume source is challenging since the observed plume is likely many times larger than the surface source itself. Limited spatial coverage of the surface also may preclude direct detections of the plume surface source, while the large fields of view of present-day orbital instruments tend to 'blur' small-scale signals. To address these issues, we have developed a novel technique for determining the surface location of a trace gas emission source in a planetary atmosphere with an uncertainty of a few tens of km using present-day direct observational capabilities in concert with high-performance numerical modeling of atmospheric dynamics. 1,2 This paper shows how evolutionary computational models (ECMs) can be employed in the search for surface sources of trace gas plumes in the atmosphere of Mars. We present the specific approach taken in linking an ECM to a simple, idealized plume model, and gauge the overall ability of this system to return estimates of a plume source location over several generations of the model's execution. This technique can isolate trace gas source locations with orders of magnitude greater efficiency than brute-force approaches, making identification of such sources, even with limited observational data, a plausible endeavor.