Evaluation of an Advanced Radiative Transfer Model in US Navy's Regional Numerical Weather Prediction System

Fu-Liou's delta-4-stream radiative transfer model has been implemented into the US Navy's Coupled Ocean-Atmospheric Mesoscale Prediction System (COAMPS). It is explicitly coupled to cloud species of water and ice and an inline aerosol module. It calculates radiative fluxes in 18 solar and IR wave bands. The performance of Fu-Liou model is evaluated at high grid resolution for two periods of 8 days of clear sky and 15 days of cloudy weather using the surface flux data from Atmospheric Radiation Measurement (ARM) Program at Southern Great Plains (SGP) site of Lamont, Oklahoma. In both cases, the new radiation model outperforms COAMPS standard model developed by Harshvardhan et al. and using cloud fractions. The sensitivity of radiative fluxes on various parameterizations of cloud effective radius is conducted to find the optimal combination of water and ice parameterizations for the best performance of the new radiation. It has also been tested in COAMPS for various seasons of 2006 over the Continental US domain at nested grids of 81 and 27 km. The 3-day forecasts are verified against weather station reports and radio soundings for the whole domain. The statistical analyses show that the new radiation model improves the dynamics of COAMPS over the standard radiation package in winter season by reducing the cold bias at surface, which is mainly caused by the land surface bulk model, and reducing the warm bias in the upper air at 300 mb, which is related to the lateral boundary condition. The heating rates and flux components are examined for the new and standard models to understand their impacts on COAMPS weather forecasts.