Improvements of numerical weather predictions using a new AVHRR green vegetation fraction dataset

摘要

This study aims to analyze impacts of the NESDIS new product of green vegetation fraction (GVF) data on simulated surface air temperature and surface fluxes over the continental United States (CONUS) using the Nonhydrostatic Mesoscale Model (NMM) core of the Weather Research and Forecasting (WRF) system, i.e. WRF-NMM, coupled with the Noah land surface model (LSM). The new global 0.144 by 0.144 degree GVF dataset is an AVHHR-based, near realtime weekly dataset starting from 1982. It has better quality and a higher temporal resolution than the old monthly GVF dataset that is currently used in the NOAA operational numerical weather prediction models. The new weekly climatology GVF data shows a higher percentage of greenness fraction over most US areas than the old dataset, with the largest differences by 20-40% over the southeast U.S., the northern Middle West, and the west coast of California in summer. We have performed some case studies over CONUS during July 2006. In general, using the new GVF data cools predicted surface temperature over most regions compared to the old data, with the largest cooling over regions with the largest GVF increase. The latent heat increases significantly over most areas while the sensible heat decreases slightly. These results are physically consistent as more of the net radiation is dissipated in form of latent heat via enhanced evapotranspiration in response to increasing vegetation cover. Compared with observations, the new GVF application reduces the WRF-NMM 2-m surface air temperature warm biases, 2-m relative humidity negative biases, and their RMSEs.