Impact of Effective Roughness Length on Mesoscale Meteorological Simulations over Heterogeneous Land Surfaces in Taiwan

Abstract

The Weather Research and Forecasting (WRF) modeling system obtains the aerodynamic roughness length (z0) from a land use (LU) lookup table. The effective aerodynamic roughness length (z0eff) was estimated for the island of Taiwan by considering the individual roughness lengths (z0i) of the underlying LU types within a modeling grid box. Two z0eff datasets were prepared: one using the z0i from the default LU lookup table and the other using the observed z0i for three LU types (urban, dry cropland and pasture, and irrigated cropland and pasture). The spatial variability of the z0eff distribution was higher than that of the LU table-based z0 distribution. Three WRF sensitivity experiments were performed: (1) dominant LU table-based z0 (namely, S1), (2) z0eff estimated from the default z0i (namely, S2), and (3) z0eff estimated from the observed z0i (namely, S3). Comparisons of the thermal field, temperature, and surface sensible and latent heat fluxes revealed no significant differences among the three simulations. The wind field overestimation and surface momentum flux underestimation in S1 were reduced in S2 and S3, and these improvements were more prominent over areas with highly heterogeneous land surface conditions.