Impact of Dynamical Downscaling on Model Representation of Land-Atmosphere Coupling Strength

Abstract:

Extremes in the water cycle, such as drought and flood, threaten the sustainability of water resources and cause significant impacts on society that will likely increase due to growing populations and a changing climate. Reducing the impact of extreme events requires preparations enabled by reliable and relevant predictions of the climate. Climate predictions are made using Global Climate Models (GCMs) that typically have spatial resolutions that are too course for application at the local level where the prediction is needed to ensure a society resilient to extremes. Therefore, a common practice is to dynamically downscale results from GCM’s using a regional model that can provide predictions at scales consistent with the application. Downscaled predictions are dependent on model physics and model setup (e.g. boundary conditions, nudging) and as a result the overall validity of dynamically downscaling has not been fully demonstrated to date.

NASA has recently sponsored an intra-agency downscaling project to better understand the validity of dynamical downscaling. As part of this project, several 10-year simulations of the NASA Unified Weather Research and Forecast (NU-WRF) model that vary in resolution and large scale nudging were used to downscale MERRA-2 reanalyses over the continental U.S. This work leverages these model runs in order to understand the impact of model resolution and nudging on the representation of land-atmosphere coupling strength and its impact on downscaled predictions of the water cycle. The representation of land-atmosphere coupling strength is analyzed through a suite of local land-atmosphere coupling (LoCo) metrics that are compared across downscaling runs as well as coarse scale predictions from GEOS-5 and MERRA-2. The impact of downscaling approaches and resolution on the representation of land-atmosphere coupling is presented and the implications for future downscaling applications are discussed.