Development and Validation of the Mars Model for Prediction Across Scales (Mars MPAS)

Abstract:

The National Center for Atmospheric Research (NCAR) Model for Prediction Across Scales (MPAS) has been converted for use as a general planetary atmospheric model (planetMPAS), and here we describe the part of this work specifically focused on the model's adaptation to Mars. In this presentation we will review: 1) issues related to the use of a height coordinate model when the total mass of the atmosphere (due to the CO2 cycle) or the global mean air temperature (due to the large eccentricity) change, 2) the computational performance of the model, especially compared to the existing MarsWRF model, 3) the commonality of "physics" between WRF and MPAS, 4) the validity of MarsMPAS simulations as compared to various Martian observational datasets. The conversion of MPAS is motivated by a number of factors: 1) that it provides the ability to (almost) arbitrarily sample the globe, including the case of nearly equidistant grid points with less distortion than traditional lat/lon grids (including no need for polar filtering), yin-yang grids, or even cubed-sphere grids, 2) that the numerical approach leads to much more efficient scaling across massively parallel computers than with other grids, 3) that MPAS will become the increasing development focus for NCAR. Thankfully, MPAS and WRF share common "physics" - identical code - such that development of physics for MarsWRF over the past 15 years can be applied without modification for MarsMPAS. Concerns had been raised over the applicability of MPAS for use on Mars due to the use of a modified height coordinate, since changing density within the fixed-height domain will cause spurious contributions to the surface pressure. This has been termed the "pressure cooker effect." However, we show that for almost all model depths of any interest (i.e., that are sufficiently deep to properly simulate atmospheric circulation components on Mars), the effect is quantitatively negligible.