The hydrostatic version of the Regional Spectral Model (RSM) with a 10-km resolution was implemented into the operational runstream for the state of Hawaii by NCEP in early 1997. From preliminary analyses and feedback from forecasters, it is apparent that the 10-km RSM forecasts show improvement over Aviation runs. Nevertheless, orographic effects and diurnal weather patterns are not well simulated by the RSM because the complex island terrain is not adequately resolved by its 10-km horizontal grid. In addition, large variations in local climate ranging from humid tropical climate on windward slopes to hot desert over bare lava soil with different vegetation covers are typical for the Hawaiian Islands. In this work, the f97 version of the RSM and the non-hydrostatic Mesoscale Spectral Model (MSM) have been coupled with the NOAH (NCEP, Oregon State University, Air Force, and National Weather Service Office of Hydrology) land surface model (LSM) for three regions of the state of Hawaii: the Hawaii-Maui-Molokai domain at a 3-km resolution, the Oahu domain at a 1.5-km resolution, and the Kauai domain at a 1.5-km resolution. Since the summer of 2001, we have been conducting daily high resolution (1.5 km) experimental forecasts for the Oahu domain using the coupled MSM/LSM with improved surface boundary conditions.

Our results suggest that (1) over the island of Oahu, the inclusion of the LSM significantly improves the performance of the MSM in simulating 2-m temperature, 2-m dew point temperature, and 10-m wind speed and wind direction when compared to observations. The daytime cold bias and over-estimation of surface wind speed experienced by the RSM and the MSM are largely removed by the coupled MSM/LSM; and (2) the high-resolution (£ 3 km), coupled MSM/LSM shows considerable improvement over the 10-km RSM in simulating the summer trade-wind weather as well as the localized heavy rainfall and high wind events. However, the MSM/LSM tends to over estimate (under estimate) precipitation on windward (lee) side of steep terrain.