Improving the Forecasts of Surface Latent Heat Fluxes and Surface Air Temperature in the GRAPES Global Forecast System

Abstract

The GRAPES (Global/Regional Assimilation and Prediction System) global medium-range forecast system (GRAPES_GFS) is a new generation numerical weather forecast model developed by the China Meteorological Administration (CMA). However, the forecasts of surface latent heat fluxes and surface air temperature have systematic biases, which affect the forecasts of atmospheric dynamics by modifying the lower boundary conditions and degrading the application of GRAPES_GFS since the 2 m air temperature is one of the key components of weather forecast products. Here, we add a soil resistance term to reduce soil evaporation, which ultimately reduces the positive forecast bias of the land surface latent heat flux. We also reduce the positive forecast bias of the ocean surface latent heat flux by considering the effect of salinity in the calculation of the ocean surface vapor pressure and by adjusting the parameterizations of roughness length for the exchanges in momentum, heat, and moisture between the ocean surface and atmosphere. Moreover, we modify the parameterization of the roughness length for the exchanges in heat and moisture between the land surface and atmosphere to reduce the cold bias of the nighttime 2 m air temperature forecast over areas with lower vegetation height. We also consider the supercooled soil water to reduce the warm forecast bias of the 2 m air temperature over northern China during winter. These modified parameterizations are incorporated into the GRAPES_GFS and show good performance based on a set of evaluation experiments. This paper highlights the importance of the representations of the land/ocean surface and boundary layer processes in the forecasting of surface heat fluxes and 2 m air temperature.