Addressing the MODIS bowtie effect for snow mapping

TitleAddressing the MODIS bowtie effect for snow mapping
Publication TypeConference Proceedings
Year of Publication2001
AuthorsGomez-Landesa E., Rango A., Bleiweiss M.
Conference NameProceedings of the 5th International Workshop on Applications of Remote Sensing in Hydrology
Date PublishedOctober 2-5, 200
Conference LocationMontpellier, France
ARIS Log Number129119
AbstractA snow cover mapping and snowmelt runoff forecasting system is being operated at the U.S. Department of Agriculture, Agricultural Research Service (USDA-ARS), Jornada Experimental Range. Snow-covered area is used as input for snowmelt runoff forecasting and has been traditionally derived from National Oceanic and Atmospheric Administration-Advanced Very High Resolution Radiometer (NOAA-AVHRR) channels 1 and 2. The snow maps and snowmelt runoff forecasts are used by water resource agencies and companies in water management decision making. The quality of the snow maps is improved using the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument, but the data is complicated by the "Bowtie Effect" that causes an overlap of the satellite field of view producing data repetition. An algorithm was developed to resample the MODIS image in latitude and longitude coordinates. The algorithm reads groups of 40 rows of image pixels and groups of 10 rows of latitude and longitude values placing each image pixel in the correspondent position of an output file in latitude and longitude coordinates. The final position of each pixel is determined by 2 linear bidimensional interpolations in the latitude and longitude space, giving 2 plans approaching the local variations of latitude and longitude. The equations of these planes are derived independently for latitude and longitude; for each case it uses 3 values of latitude and longitude. A concern is the relation between rows and columns of output images which can vary strongly within one single image and near polar regions. Appearance of gaps was expected and detected at the edges of the output file due to lack of information in those areas. These problems are overcome when the algorithm is applied to small areas and far from polar regions.