|| Checking for direct PDF access through Ovid
In order to assess glacier runoff to the Upper Columbia River Basin (UCRB) and quantify energy balance effects of tributary-trunk detachment due to recession, we used field observations to develop a distributed melt model of Shackleton Glacier, Canadian Rockies. Field data were derived from meteorological stations, ablation and snowline measurements, and weather observations between 2004 and 2010. Katabatic wind speed and direction were linked to terrain heat advection and irradiance, potentially resulting in significant cross-glacier gradients in melt. A geographic information system-based distributed melt model, using standard energy balance components, was developed for the 2010 melt season. Benchmark model parameterisations were derived for clear, cloudy and overcast days. Novel model parameterisations include terrain irradiance using a sky view factor and an albedo mask, and a katabatic wind ‘switch’ with valley temperature thresholds. Modelled energy balance components suggest significant sensitivities to terrain irradiance and katabatic wind, in part related to cloudiness. Glacier-wide melt decreased by 10–15% when katabatic wind was turned off, with an interesting spatial pattern. Longwave radiation from valley walls increased local melt up to 30%, but net glacier-wide effects were <6%. Daily glacier melt was 0.1–0.8 million m3 w.e. day−1 and peaked in early August. Net 2010 planar-area melt was 38–50 million m3 w.e., depending on cold storage, whereas slope-corrected-area melt was ˜4% higher. Our results indicate that katabatic wind and terrain are important in calculations of ablation in fragmenting glacier systems and that late-summer glacier contribution to UCRB runoff at Mica Dam is ˜25%. Copyright © 2012 John Wiley & Sons, Ltd.