| description abstract | Tropospheric motions can be inferred from geostationary satellites by tracking clouds and water vapor in sequential imagery. These atmospheric motion vectors (AMV) have been operationally assimilated into global models for the past three decades, with positive forecast impacts. This paper presents results from a study to assess the impact of AMV derived from Geostationary Operational Environmental Satellite (GOES) imagery on mesoscale forecasts over the conventional data-poor central North Pacific region. These AMV are derived using the latest automated processing methodologies by the University of Wisconsin?Cooperative Institute for Meteorological Satellite Studies (CIMSS). For a test case, a poorly forecast subtropical cyclone (kona low) that occurred over Hawaii on 23?27 February 1997 was chosen. The Local Analysis and Prediction System (LAPS) was used to assimilate GOES-9 AMV data and to produce fifth-generation Pennsylvania State University?NCAR Mesoscale Model (MM5) initial conditions. The satellite wind assimilation is carried out on the 27-km-resolution domain covering the central Pacific area. The MM5 was run with three two-way nested domains (27, 9, and 3 km), with the innermost domain moving with the kona low. The AMV data are found to influence the cyclone?s development, improving the prediction of the cyclone?s central pressure and the track of the low?s center. Since September 2003, GOES-10 AMV data have been routinely accessed from CIMSS in real time and assimilated into the University of Hawaii (UH) LAPS, providing high-resolution initial conditions for twice-daily runs of MM5 at the Mauna Kea Weather Center collocated at the UH. It is found that the direct assimilation of AMV data into LAPS has a positive impact on the forecast accuracy of the UH LAPS/MM5 operational forecasting system when validated with observations in Hawaii. The implications of the results are discussed. | |
