Abstract

The lowest model level height z1 is important in atmospheric numerical models, since surface layer similarity is applied to the height in most of the models. This indicates an implicit assumption that z1 is within the surface layer. In this study, impacts of z1 on the performance of planetary bou... Show moreThe lowest model level height z1 is important in atmospheric numerical models, since surface layer similarity is applied to the height in most of the models. This indicates an implicit assumption that z1 is within the surface layer. In this study, impacts of z1 on the performance of planetary boundary layer (PBL) parameterizations are investigated. Three conceptually different schemes in the Weather Research and Forecasting (WRF) model are tested for one complete diurnal cycle: the nonlocal, first-order Yonsei University (YSU) and Asymmetric Convective Model version 2 (ACM2) schemes and the local, 1.5-order Mellor-Yamada-Janji (MYJ) scheme. Surface variables are sensitive to z1 in daytime when z1 is below 12 m, even though the height is within the surface layer. Meanwhile during nighttime, the variables are systematically altered as z1 becomes shallower from 40 m. PBL structures show the sensitivity in the similar manner, but weaker. The order of sensitivity among the three schemes is YSU, ACM2, and MYJ. The significant sensitivity of the YSU parameterization comes from the PBL height calculation. This is considerably alleviated by excluding the thermal excess term in determining the PBL height when z1 is within the surface layer. The factor that specifies the ratio of nonlocal transport to total mixing is critical to the sensitivity of the ACM2 scheme. The MYJ scheme has no systematic sensitivity, since it is a local scheme. It is also noted that a numerical instability appears accompanying the unrealistic PBL structures when the grid spacing in the surface layer suddenly jumps. Show less