Evaluation of a Novel High-Resolution Climate-Scale Simulation using the Model for Prediction Across Scales - Atmosphere (MPAS-A)

We present simulations representative of the historical climate using the high-resolution (15 km) global Model for Prediction Across Scales – Atmosphere (MPAS-A) version 7.0. MPAS-A is a recently developed community atmospheric modeling platform with unique capabilities useful for weather and climate research. Our simulations span 30 years corresponding to water years 1990 through 2019 (i.e., 1 October 1989 – 30 September 2019). We analyze several atmospheric variables, including 2-meter temperature, mean sea-level pressure (MSLP), 500-hPa heights, vertically integrated water vapor transport (IVT), and 250-hPa winds. We initialize the simulations with the European Centre for Medium-Range Weather Forecasts Reanalysis v5 (ERA5) and update sea surface temperatures and sea ice daily using the fine-scale (0.05º) Operational Sea Surface Temperature and Ice Analysis (OSTIA). To our knowledge, this is one of very few long-running, high-resolution, climate-scale MPAS-A simulations to date. We analyze relationships between MPAS-A and historical records by averaging the model output and evaluating the mean fields by day, month, season, and decade using pattern correlations with the ERA5. We further evaluate the ability of MPAS-A to replicate interannual variability in atmospheric patterns using self-organizing maps (SOMs). We find that MPAS-A reasonably captures large-scale atmospheric features and their variability in the Northern and Southern Hemispheres, including maritime sea-level pressure systems, such as the subtropical highs, upper-tropospheric flow regimes, and precipitation patterns. Our results demonstrate the utility of MPAS-A for investigating climate-scale phenomena at spatial resolutions generally unachievable by general circulation models (GCMs). Furthermore, we demonstrate the applicability of these model simulations for future studies examining effects of climate change on various high-impact weather systems in a model-relative framework.