|Rango, Albert - Al|
Submitted to: Canadian Journal of Remote Sensing
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 4/5/2005
Publication Date: 8/1/2004
Citation: Gomez-Landesa, E., Rango, A., Bleiweiss, M. 2004. An algorithm to address the MODIS bowtie effect. Canadian Journal of Remote Sensing. 30(4):644-650. Interpretive Summary: When using raw MODIS data on a PC-Windows computer, overlapping data occurs which causes erroneous snow cover maps, especially on the edge of the image. Because many investigators prefer to use raw satellite data for snow mapping, we developed an algorithm to remove the data repetition, or "bowtie effect" from the MODIS data. Users of the algorithm can now map the snow cover on their own machines with their preferred snow mapping techniques. The results of this algorithm compared closely with methods used on UNIX machines. Those who will profit from this algorithm are scientists and user agency investigators who do their own snow mapping on small, PC-based computers.
Technical Abstract: A snow cover mapping and snowmelt runoff forecasting system is currently being operated at the USDA-ARS Jornada Experimental Range. Snow-covered area is being used as an input for snowmelt runoff forecasting and has been traditionally derived from NOAA-AVHRR (National Oceanic and Atmospheric Administration-Advanced Very High Resolution Radiometer) channels 1 and 2. Both the snow maps and the snowmelt runoff forecasts are being sent to water resources agencies and companies for decision making in water management. The quality of the snow maps can be improved using the Moderate Resolution Imaging Spectroradiometer (MODIS) instrument on board the Terra platform. MODIS channels 1 and 2 are centered in similar wavelengths as AVHRR channels 1 and 2 but with 250-m resolution instead of 1 km off AVHRR channels. The recent implementation of MODIS data has been complicated by the so-called "bowtie effect" of the MODIS instrument, which causes an overlap of the satellite field of view producing a data repetition. This effect increases with the distance from nadir and can be especially dramatic at the edge of the image.