Satellite remote sensing of snow cover and snowmelt in the Arctic

In this thesis, satellite remote sensing techniques have been used to detect snow cover and snowmelt in the Arctic from observations made by visible infrared, passive microwave, and active microwave sensors. An evaluation of the NOAA weekly snow cover dataset over the Canadian Arctic was conducted through comparisons with snow cover extent derived from the Advanced Very High Resolution Radiometer, Special Sensor Microwave/Imager, Landsat TM browse images, and surface snow depth observations. The evaluation revealed that the NOAA weekly dataset consistently overestimated snow cover extent during the spring melt period, with delays of up to 4 weeks in melt onset. The most likely causes for the delayed melt onset were frequent cloud cover in the spring melt period, and the low frequency of data coverage over higher latitudes.; The extent and duration of surface melt on the Queen Elizabeth Islands (QEI) ice caps in the Canadian high Arctic and on the Greenland ice sheet were detected from enhanced resolution QuikSCAT (QSCAT) scatterometer images. Based on the sharp changes in microwave backscatter on the appearance/disappearance of liquid water in snow, dates of melt onset and freeze-up across the ice caps and ice sheet were estimated using a dynamic threshold method. Annual melt duration and melt anomaly maps were produced for the 2000--2004 period. Over the 5-year period, the mean melt duration ranged from 30.9 to 42.6 days over the QEI, and from 14.3 to 20.5 days over Greenland. Inter-annual changes in the distribution of ice layer formation within the near surface snow and firn on the Greenland ice sheet were also investigated. There was a close correspondence between melt season duration derived from QSCAT and from temperature measurements at field sites over both the QEI ice caps and the Greenland ice sheet. Systematic mapping of melt season duration on the Arctic ice masses may therefore be an effective and rapid way of detecting the spatial pattern of temperature variability during the summer melt season. In both cases, the relationships between the derived melt patterns and the variability in the local atmospheric circulation were discussed.