Comparison of modeled and geodetically-derived glacier mass balance for Tiedemann and Klinaklini glaciers, southern Coast Mountains, British Columbia, Canada

Abstract

Predicting the fate of mountain glaciers requires reliable observational data to test models of glacier mass balance. Using glacier extents and digital elevation models (DEMs) derived from aerial photographs and ASTER satellite imagery, we calculate changes in area, elevation, and volume of Tiedemann and Klinaklini glaciers. Between 1949 and 2009, Tiedemann and Klinaklini glaciers lost approximately 10% of their area. The total area-averaged thinning of Klinaklini was 40.1 ± 1.5 m water equivalent (w.e.) and total mass loss equaled 20.24 ± 1.36 km³ w.e., whereas Tiedemann Glacier thinned by 25.7 ± 1.9 m w.e. and lost 1.69 ± 0.17 km³ w.e. of ice. We attribute lower observed rates of thinning at Tiedemann Glacier to thick debris cover in the ablation area. Both glaciers thickened at mid-elevations after the year 2000. Glacier mass balance and volume change were modeled using temperature, precipitation and evapotranspiration fields dynamically downscaled to the mesoscale (8 km resolution) using the Regional Atmospheric Modeling System (RAMS) model and further statistically downscaled to the glacier scale (100 m elevation bands) using modeled surface lapse rates. The mass balance model over-predicts total volume loss by 1.1 and 6.3 times the geodetic loss for Klinaklini and Tiedemann glaciers respectively. Differences in modeled and observed total ice loss are due to (1) the coarse resolution of the downscaled climate fields, and (2) extensive debris cover in the ablation area of Tiedemann Glacier. Future modeling efforts should dynamically downscale at resolutions that capture the topographic complexity of a region and employ strategies to account for time-evolving debris cover.

Introduction

Glaciers are integral to many natural and human systems, making them important targets for monitoring and prediction. Although mountain glaciers only constitute 3–4% of global glacierized area, their recent recession significantly contributes to sea level rise (Arendt et al., 2002, Dyurgerov, 2003, Berthier et al., 2004, Larsen et al., 2007, Meier et al., 2007, Berthier et al., 2010). Mountain glaciers are the second largest contributor to recent sea-level rise (Cazenave and Nerem, 2004), and the total volume of glaciers in western Canada and Alaska has been estimated to contain a sea-level equivalence of 5 and 68 mm, respectively (Radić and Hock, 2010). Changes in glacier thickness and volume can also influence the magnitude and timing of surface runoff, affecting water supply for agriculture, consumption, and hydropower generation (Barry, 2006, Stahl and Moore, 2006, Moore et al., 2009). Given recent trends in mean global surface temperatures, and projections of continued warming, glaciers in western North America and throughout the world are expected to continue to retreat. To estimate future changes in volume and area, methods to estimate the mass balance of glaciers under a given climate scenario are required.

Glacier mass balance models to predict the fate of glaciers vary from simple ones that use accumulated air temperature anomalies (positive degree days) to those that employ a full energy balance (Braithwaite and Zhang, 1999, Casal et al., 2004, Hock and Holmgren, 2005). Positive degree day models empirically relate glacier melt and air temperature; these empirical models assume that air temperature integrates the individual fluxes of the surface energy balance. Melt factors for snow and ice have been shown to be similar among glaciers within a region (Shea et al., 2009), but can vary temporally at inter-annual to inter-decadal time scales (Huss et al., 2009, Shea et al., 2009). Temporally varying melt factors introduce uncertainty in mass balance modeling using a PDD approach, but the magnitude of this error is difficult to quantify in data-poor regions. In the current study, the lack of available input data to drive a melt model using an energy balance approach is limited by the lack of required input data. Snow accumulation is typically calculated from the total precipitation that falls as snow, and must be melted before ice melt can occur (Braithwaite and Zhang, 1999, Shea et al., 2009).

Predicting the fate of mountain glaciers requires reliable melt modeling strategies and temporally and spatially distributed data that can be used to test these approaches. The use of remote sensing has increased the number of glaciers monitored and extended the mass balance record in regions where few traditional mass balance records exist (Berthier et al., 2004, Luthcke et al., 2008). While glacier length and area are commonly measured, changes in volume and mass balance provide a more direct, reliable indicator of climate change and can be used to verify the results of mass balance models (Kääb, 2002, Berthier et al., 2004, Barry, 2006).

The objectives of this paper are to determine change in area, elevation, and volume at Klinaklini and Tiedemann glaciers in the southern Coast Mountains British Columbia using digital elevation models (DEMs) derived from multiple sets of aerial photographs and satellite images. We expand on the results of VanLooy and Forster (2008) by incorporating DEMs that post-date the Shuttle Radar Topography Mission (SRTM), and also extend our analysis back in time to include pre-1970 data. In addition, we explore the climatic and site-specific factors that explain observed differences in area and volume change for these two glaciers. Finally, we test whether glacier mass balance estimates obtained from a hybrid modeling strategy agrees with geodetically-derived changes in glacier volume.