A physically based algorithm for downscaling temperature in complex terrain

Abstract

Recent improvements to an algorithm to be used operationally for downscaling screen temperatures from numerical weather prediction models are described. Testing against very high resolution dynamically downscaled screen temperatures and intensive field measurements taken during the Cold-Air Pooling Experiment (COLPEX) is performed. The improvements are based on a physical understanding of the processes involved in the formation of cold-air pools (CAPs) that is informed by recent research. The algorithm includes a parameterization of sidewall sheltering effects that lead to lower temperatures in valley-bottom CAPs on clear, calm nights. Advection and adjustment over exposed hilltops results in higher screen temperatures than on flat ground but lower temperatures relative to the free air above the valley at the same elevation, and a treatment of this effect has also been developed. These processes form the major contributions to the often dramatic small-scale variations in temperature in complex terrain in stable boundary layer conditions, even when height variation is fairly shallow. The improvements result in qualitatively better reproduction of subgrid temperature patterns in complex terrain during CAPs. Statistical forecast errors are subsequently improved.