Intercomparison of 7 Planetary Boundary-Layer/Surface-Layer Physics Schemes Over Complex Terrain for Battlefield Situational Awareness [open pdf - 5MB]
From the abstract: "This study considers the performance of 7 of the Weather Research and Forecast model boundary-layer (BL) parameterization schemes in a complex terrain environment. These BL/surface-layer options were compared and represent local, nonlocal, and hybrid approaches. Based on surface parameter root-mean-square error and bias measures, there is no clearly superior scheme; however, the BouLac, Quasi-normal Scale Elimination, and Yonsei University (YSU) schemes performed best. The surface parameters, planetary BL structure, and vertical profiles are important for US Army Research Laboratory applications, including in the formulation of some weather impact decision aids. Both the YSU and Mellor-Yamada-Janjic schemes, used in the common Weather Running Estimate-Nowcast model configurations and in operational configurations, appear to perform reasonably well in this environment. While the scale-aware scheme addresses the scale dependency problem, we have found no clear improvement when using this scheme with a 1-km inner mesh in this study. There is little variation between the schemes in the daytime BL, but at nighttime the schemes differ notably. The model's greatest difficulty is in simulating the transition between daytime and nighttime BLs. There is insufficient statistical spread in the members tested to consider a BLphysics-only-based ensemble to quantify model uncertainty."
ARL-TR-8353; Army Research Laboratory Technical Report 8353
Defense Technical Information Center (DTIC): http://www.dtic.mil/dtic/