P3.3
Validation of a blended-snow product using station observations and satellite images
Dorothy K. Hall, NASA, Greenbelt, MD; and J. L. Foster, P. M. Montesano, G. A. Riggs, R. E. J. Kelly, and J. B. Eylander
A joint U.S. Air Force Weather Agency (AFWA) / National Aeronautics and Space Administration (NASA) global, daily snow product has been developed. This preliminary product, called the AFWA – NASA or ANSA blended snow-cover product, was developed by blending Moderate Resolution Imaging Spectroradiometer (MODIS) standard snow-cover maps, Advanced Microwave Scanning Radiometer for Earth Observing System (AMSR-E) passive microwave standard snow-water equivalent (SWE) maps, and newly-developed algorithms to map wet and melting snow using AMSR-E and QuikSCAT scatterometer data at 25-km resolution. The ANSA product provides (or will provide) snow-cover extent, fractional-snow cover (FSC), onset of melt, and location of actively-melting areas. Specifically, the MODIS maps provide snow-cover extent, including FSC, at a resolution up to 500 m under clear skies, and the AMSR-E maps provide snow cover through clouds, albeit at a coarser resolution (25 km), and SWE. We have begun to validate the blended product, focusing at first on validating snow-cover extent. Results from analysis in the northern United States during the winters of 2002-03 and 2006-07 show that the blended product provides improved results for snow mapping relative to using either the MODIS or AMSR-E standard maps alone. National Weather Service Co-Operative Observing Network stations and Air Force Surface Observations [a compilation of synoptic, airways and METAR (aviation routine weather) reports that are used in the AFWA Snow Depth, or SNODEP, model] are used as ground truth. Methods have been tested for interpolating the station data to validate the blended-snow maps. The interpolation techniques are not ideal because the number of stations is inadequate in much of the study area to permit an optimal interpolation. In these areas, we have employed other satellite data, such as 250-m resolution images from MODIS, and 30-m resolution images from the Landsat Enhanced Thematic Mapper Plus (ETM+) to determine the presence of snow cover. Statistics are provided that demonstrate the ability of the blended product to map snow, relative to the MODIS, AMSR-E and SNODEP maps if used alone.
Poster Session 3, Validation of Hydrometeorological Observations
Wednesday, 23 January 2008, 2:30 PM-4:00 PM, Exhibit Hall B
Previous paper Next paper