Data assimilation over the western United States

The ARPS (Advanced Regional Prediction System) Data Analysis System (ADAS), developed by the Center for Analysis and Prediction of Storms (CAPS) at the University of Oklahoma, is used for local data assimilation over the complex terrain of the western United States. The Utah version of ADAS is currently being run in several different configurations.

A fully 3-dimensional analysis is created on a 220 km square domain centered near Salt Lake City, Utah at 1 km horizontal resolution. The grid is stretched in the vertical, varying from 9 m at the surface to roughly 600 m at the domain top. There are 33 vertical levels, with a height-based, terrain-following coordinate. A terrain factor has been added to the Bratseth weights, creating anisotropic zones of influence. This factor helps to compensate for the strong terrain gradients that exist within the domain, allowing high elevations surface observations to influence data-void regions in nearby mountain ranges, while limiting their effect on the free atmosphere. The NCEP Rapid Update Cycle analyses are used as the background field with data resources including: surface observations from MesoWest; NWS rawinsondes at Salt Lake City; NWS WSR-88D velocity and reflectivity data in NIDS format; FAA TDWR velocity and reflectivity; Dugway wind profiler data; and ACARS wind and temperature data.

In addition to the local 3-dimensional analysis, an ADAS surface analysis is run every 15 minutes over the western United States at a resolution of 10 km. This analysis depends strongly upon MesoWest, which is a cooperative collection of mesonets around the western United States. Surface Aviation observations are supplemented by observations at over 2000 stations. Tests are underway to make these surface analyses available at all National Weather Service forecast offices around the West. We're conducting research to identify the tradeoffs in analysis quality as a function of horizontal resolution. Our preliminary results, based on qualitative evaluation of many analyses, is that the 10 km analysis provides considerable useful information, even though many isolated mountain ranges and valleys are not captured by the 10 km terrain background compared to the 2 km background.

Optimal configurations for data assimilation over complex terrain remain under investigation. Data from the Intermountain Precipitation Experiment (IPEX) during February 2000 are used to investigate the sensitivity of the analysis system to available data resources.