Sensitivity of simulated short-range high-temperature weather to land surface schemes by WRF

The simulations of a heat wave occurring in southern Yangtze-Huaihe valley and southern China during late July, 2003 were conducted to examine the sensitivity of simulated surface air temperature (SAT) to different land surface schemes (LSSs) using the Weather Research and Forecasting Model (WRF) Version 2.2 in the short-range mode for 24-h integrations. Initial and boundary conditions employed a National Centers for Environmental Prediction (NCEP) analysis. The results showed that, overall, simulated high-temperature weather is sensitive to different LSSs. Large differences in simulated SAT intensity, threat score, and simulated error under different schemes are identified clearly. In addition, some systematic differences are also induced by the LSSs. In terms of threat score from the three LSSs, SLAB is the best, and RUC is better than NOAH. SLAB gives the lowest absolute error for area-averaged SAT, and tends to depict the western Pacific subtropical high with the easternmost position at low levels. The LSSs modify the simulated SAT, primarily via the transfer of sensible heat from the land surface to the atmosphere. The physical mechanism of the positive feedback between atmospheric circulation and the SAT is unimportant, with negative feedback over most of the simulated areas. This study emphasizes the importance of improving LSSs in SAT forecasting by numerical models.