Improving Noah land surface model performance using near real time surface albedo and green vegetation fraction

Abstract

The current operational Noah land surface model (LSM) uses multi-year climatology of monthly green vegetation fraction (GVF) and the multi-year averages of land surface albedo data for several numerical weather predictions at National Centers for Environmental Predictions of National Oceanic and Atmospheric Administration. However, these static GVF and albedo data can only prescribe the multiannual means and lack the ability to capture near real time (NRT) vegetation status and land surface condition. In this study, the impact of NRT GVF and albedo on Noah LSM (version 3.2) performances are examined against in situ measurements of surface net long wave radiation and net short wave radiation from 7 U.S. Surface Radiation Budget Network stations, and soil temperature and soil moisture from 9 USDA Soil Climate Analysis Network sites. Large differences between the NRT GVF/surface albedo and their climatological averages are found over the global, which have significant influences on Noah LSM simulations. With respect to in situ measurements, the Noah LSM simulation improvements from using the weekly GVF data are 19.3% for surface soil moisture, 9.3% for surface soil temperature. The benefits from the weekly GVF and monthly albedo can reach to 2.7Wm−2 for surface net long wave radiation and 2.6Wm−2 for surface net short wave radiation. The results suggest to Noah model developers and users that the NRT GVF and albedo should be used for better model performance.