Examining Sources of a Wintertime Cold Bias in Remote Sensing Temperatures from the St. Elias Range

Abstract:

Remote sensing data are crucial to relating the thinning and retreat of alpine glaciers in remote areas to climatological changes. For example, the Moderate Resolution Imaging Spectroradiometer (MODIS) land surface temperature (LST) product can be used to track glacier behavior in response to warming in the St. Elias Mountains (Yukon, Canada). However, MODIS LSTs in the Upper Kaskawulsh-Donjek region of the St. Elias have shown a cold bias relative to in situ temperature measurements, the source of which is unknown. Here we show that a wintertime cold bias in MODIS LSTs is not an artifact of MODIS’ large footprint in highly heterogenous terrain and discuss alternative sources of the bias. At study sites Icefield Divide (“Divide”, 60.68°N, 139.78°W, 2,603 m.a.s.l) and Eclipse Icefield (“Eclipse”, 60.84°N, 139.84°W, 3,017 m.a.s.l ), we found that a cold bias in remote sensing temperatures is present not only in MODIS LST products, but also in Advanced Spaceborne Thermal Emissions Radiometer (ASTER) and Landsat temperature products, both of which have a much smaller footprint (90-120 m) than MODIS (1 km). In all three datasets, the cold bias was most pronounced in the winter (mean cold bias >8°C), and least pronounced in the spring and summer (mean cold bias <2°C). Additionally, we found that the cold bias was present in brightness temperatures from each of these three instruments, as well as in their final surface temperature products. Our results demonstrate that although available remote sensing temperature products can be confidently used in the St. Elias to represent spring and summer conditions, winter conditions result in a discrepancy between calculated surface temperatures and measured air temperatures, possibly due to rapidly changing snow surface conditions (e.g. grain size, wetness) during and after winter precipitation events. We intend to test this idea using a ten-year (2003-2012) snow accumulation record from Divide and will discuss these results.