Water Resources Research · 2008 · 359 citations · 25 references
Natural SciencesEarth SciencesHydrologyClimatologyRemote SensingSnow HydrologyCryospheric Remote SensingRemote Sensing Image AnalysisSatellite Image InterpretationSnow Cover Remote SensingSnow MonitoringSnow ObservationAccuracyCloud CorrectionCloud RemovalFilteringFiltersSnow Cover MappingSnow Depth MeasurementSnow Mapping Accuracy
MODIS snow cover products are appealing for hydrological applications due to their good accuracy and daily availability, but their main limitation is cloud obscuration. The study evaluates simple mapping methods—temporal and spatial filters—to reduce cloud coverage by using neighboring non‑cloud pixels in time or space and combining MODIS data from Terra and Aqua. The accuracy of these filter methods is evaluated over Austria using daily snow depth observations at 754 climate stations and daily MODIS images from 2003–2005. The filtering techniques markedly reduce cloud coverage while maintaining good agreement with ground observations, with a clear seasonal trade‑off between accuracy and cloud coverage; for example, combined Aqua‑Terra images reduce 63 % cloud to 52 % with 94.9 % accuracy, spatial filter to 46 % with 94.2 %, 1‑day temporal filter to 34 % with 94.4 %, and 7‑day temporal filter to 4 % with 92.1 %.
MODIS snow cover products are appealing for hydrological applications because of their good accuracy and daily availability. Their main limitation, however, is cloud obscuration. In this study we evaluate simple mapping methods, termed temporal and spatial filters, that reduce cloud coverage by using information from neighboring non‐cloud covered pixels in time or space, and by combining MODIS data from the Terra and Aqua satellites. The accuracy of the filter methods is evaluated over Austria, using daily snow depth observations at 754 climate stations and daily MODIS images in the period 2003–2005. The results indicate that the filtering techniques are remarkably efficient in cloud reduction, and the resulting snow maps are still in good agreement with the ground snow observations. There exists a clear, seasonally dependent, trade off between accuracy and cloud coverage for the various filtering methods. An average of 63% cloud coverage of the Aqua images is reduced to 52% for combined Aqua‐Terra images, 46% for the spatial filter, 34% for the 1‐day temporal filter and 4% for the 7‐day temporal filter, and the corresponding overall accuracies are 95.5%, 94.9%, 94.2%, 94.4% and 92.1%, respectively.
25