Precipitation Partitioning in Multiscale Atmospheric Simulations: Impacts of Stability Restoration Methods

Abstract

Proper simulation of high-resolution surface precipitation distribution and variability is important to local aspects of environmental pollution and climate. Global and regional climate and weather models routinely evaluate total precipitation using available measurements, but quantitative evaluation of contributions by the individual components (convective and nonconvective) to the total precipitation is not routinely performed. Wet bias in one component can alleviate dry bias in the other component, making the total precipitation look comparable to measurements, leading to an invisible bias. To study this aspect, Tropical Rainfall Measuring Mission (TRMM) measurements for precipitation components were used to quantitatively evaluate convective fractions simulated by a cumulus parameterization scheme in a regional climate simulation using 12-km grid spacing. Results indicated a wet bias in convective precipitation as compared to TRMM measurements. This wet bias helped to counter a dry bias in grid-scale precipitation and led to a total precipitation comparable to Parameter-elevation Regressions on Independent Slopes Model and TRMM measurements. A new formulation has been developed for convective cloud adjustment time scale alleviating wet bias in convective precipitation when compared to the old formulation and TRMM measurements. Results for different grid spacing also indicate that the new method produces lower subgrid-scale precipitation with overall better precipitation estimates. Our results also suggest that evaluating both components of the surface precipitation rather than just the total itself can inform a need to improve cloud formulations, as demonstrated in this study. This study calls for the development of methods to routinely produce measurements for precipitation components that help evaluating global and regional climate and weather models.