A) The implementation of the realistic vegetation surface data improves simulations of precipitation The control experiments were performed using the 50-km RSM over the United States with the surface physics and vegetation data similar to those used in the reanalysis 2. While the model is able to capture the 1993 floods and 1988 droughts over the central United States, the Arizona New Mexico (AZNM) area is very dry. The implementation of the terrain and vegetation data from the USGS with 30 minutes resolution improves the precipitation. The improvements are from the realistic simulation of two Low Level Jets (LLJ) over the Gulf of Mexico and the Gulf of California and the increase of evaporation E. Both the positive feedback between E and P and realistic moisture divergence contribute to the improvement of precipitation.
B) Impact of soil moisture initialization Simulations were performed for the wet and dry months of both the central United States and the Southwest with soil moisture taken from the reanalysis II at the initial date. Then soil moisture from 1995(1999) was used to initialize the monthly simulation of 1999(1995) over the Southwest. Dry and wet episodes are not minor images of each other. For a dry month like 1995 over the AZNM, the impact of soil moisture is minimum. The large scale circulation pattern is more important in controlling precipitation and that is given through boundary conditions. For wet cases, wet soil increases precipitation both through the positive feedback between P and E and the improvement of the California LLJ.
(C) Impact of sea surface temperature anomalies (SSTAs) Simulations were performed using SSTAs from OI at 0.5 degree resolution and the climatological SST. Overall, the SSTAs from OI are warmer over the Gulf of California in comparison to the climatology. Warm SSTAs over the Gulf of California enhance the LLJ over California and more rainfall over AZNM.
.
Supplementary URL: