Distributed energy balance modeling of South Cascade Glacier, Washington and assessment of model uncertainty

Faron S. Anslow, Steven Hostetler, William R. Bidlake, Peter U. Clark

    Research output: Contribution to journalArticlepeer-review

    62 Citations (Scopus)


    We have developed a physically based, distributed surface energy balance model to simulate glacier mass balance under meteorological and climatological forcing. Here we apply the model to estimate summer ablation on South Cascade Glacier, Washington, for the 2004 and 2005 mass balance seasons. To arrive at optimal mass balance simulations, we investigate and quantify model uncertainty associated with selecting from a range of physical parameter values that are not commonly measured in glaciological mass balance field studies. We optimize the performance of the model by varying values for atmospheric transmissivity, the albedo of surrounding topography, precipitation-elevation lapse rate, surface roughness for turbulent exchange of momentum, and snow albedo aging coefficient. Of these the snow aging parameter and precipitation lapse rates have the greatest influence on the modeled ablation. We examined model sensitivity to varying parameters by performing an additional 103 realizations with parameters randomly chosen over a ±5% range centered about the optimum values. The best fit suite of model parameters yielded a net balance of −1.69 ± 0.38 m water equivalent (WE) for the 2004 water year and −2.10 ± 0.30 m WE up to 11 September 2005. The 2004 result is within 3% of the measured value. These simulations account for 91% and 93% of the variance in measured ablation for the respective years.
    Original languageEnglish
    Pages (from-to)F02019
    JournalJournal of Geophysical Research: Earth Surface
    Issue numberF2
    Publication statusPublished (in print/issue) - 31 May 2008


    • glaciology
    • modeling
    • climate


    Dive into the research topics of 'Distributed energy balance modeling of South Cascade Glacier, Washington and assessment of model uncertainty'. Together they form a unique fingerprint.

    Cite this