[p1] High performance computing


Designing an Ensemble Data Assimilation System for the Mars Atmosphere

Steven J. Greybush (Penn State University), Eugenia Kalnay (University of Maryland), R. John Wilson (GFDL), Hartzel Gillespie (PSU), Matthew Wespetal (UMD), Thomas Nehrkorn (AER), S. Mark Leidner (AER), Ross Hoffman (AER), and Takemasa Miyoshi (RIKEN/UMD)


The Ensemble Mars Atmosphere Reanalysis System (EMARS) provides insights into Mars atmosphere weather and climate through compiling a multi-year dataset of temperatures, winds, surface pressures, and aerosols. The Local Ensemble Transform Kalman Filter (LETKF) combines spacecraft observations from the Thermal Emission Spectrometer (TES) and Mars Climate Sounder (MCS) with the GFDL Mars Global Climate Model (MGCM). Given the unique atmospheric dynamics and observing systems on Mars, the creation and ongoing development of this reanalysis requires special design considerations. These include: adaptive inflation and varying aerosols to account for regions of chaotic versus forced error growth; short assimilation windows to avoid resonance with atmospheric tides; enforcement of mass conservation; an improved forward model and a pathway for interactive retrievals; and explorations of a hybrid-ensemble data assimilation system. Implications for analyzing and predicting specific atmospheric features are discussed.