Calibrated Ice Thickness Estimate for All Glaciers in Austria

Abstract

Knowledge on ice thickness distribution and total ice volume is a prerequisite for computing future glacier change for both glaciological and hydrological applications. Various ice thickness estimation methods have been developed but regional differences in fundamental model parameters are substantial. Parameters calibrated with measured data at specific points in time and space can vary when glacier geometry and dynamics change. This study contributes to a better understanding of accuracies and limitations of modeled ice thicknesses by taking advantage of a comprehensive data set of in-situ ice thickness measurements from 58 glaciers in the Austrian Alps and observed glacier geometries of three Austrian glacier inventories (GI) between 1969 and 2006. The field data are used to calibrate an established ice thickness model to calculate an improved ice thickness data set for the Austrian Alps. A cross-validation between modeled and measured point ice thickness indicates a model uncertainty of 25-31% of the measured point ice thickness. The comparison of the modeled and measured average glacier ice thickness revealed an underestimation of 5% with a mean standard deviation of 15% for the glaciers with calibration data. The apparent mass balance gradient, the primary model parameter accounting for the effects of surface mass balance distribution as well as ice flux, substantially decreases over time and has to be adjusted for each temporal increment to correctly reproduce observed ice thickness. This reflects the general stagnation of glaciers in Austria. Using the calibrated parameter set, 93% of the observed ice thickness change on a glacier-specific scale could be captured for the periods between the GI. We applied optimized apparent mass balance gradients to all glaciers of the latest Austrian glacier inventory and found a volume of 15.9 km 3 for the year 2006. The ten largest glaciers account for 25% of area and 35% of total ice volume. An estimate based on mass balance measurements from nine glaciers indicates an additional volume loss of 3.5 ± 0.4 km 3 (i.e., 22 ± 2.5%) until 2016. Relative changes in area and volume were largest at glaciers smaller than 1 km 2 , and relative volume changes appear to be higher than relative area changes for all considered time periods.

Bias-corrected estimates of glacier thickness in the Columbia River Basin, Canada

“…This compensating effect between mass-balance gradient and A opt is most pronounced at the two Rocky Mountain glaciers, Haig and West Washmawapta (Supplemental S5). Like Rabatel and others (2018) and Helfricht and others (2019) we find that the use of observed mass-balance gradients yields better estimates of ice fluxes relative to model gradients. We find less improvement in model performance with observed gradients rather than modeled ones, however, indicating a relatively good performance of surface mass-balance model within OGGM.…”

Potential future lakes from continued glacier shrinkage in the Aosta Valley Region (Western Alps, Italy)

“…In the existing literature (e.g. Linsbauer et al, 2012Linsbauer et al, , 2016Farinotti et al, 2017;Helfricht et al, 2019), researchers used geophysical investigations carried out on selected glaciers to validate model results.…”

An approach to remotely monitor glacier mass balance at seasonal to annual time scales, Columbia and Rocky Mountains, Canada.

“…An average MAE of 22.1 m (<25%) for optimized ice thickness (Table 3.6) highlights the ability of OGGM to produce reliable glacier-wide estimates of ice thickness using point observations with varying levels of coverage. The range of our ME and MAE values is similar to those reported from the cross-validation applied in modeling of Austrian glaciers in Helfricht et al (2019).…”

“…Our largest study glacier, Conrad Glacier, shares the dendritic pattern and also features numerous icefalls like the Shackleton and Clemenceau glaciers (Jiskoot et al, 2009). Ice thickness of small glaciers (S <1 km 2 ) are typically underestimated by models (Helfricht et al, 2019), but the RGI does not not include glaciers smaller than 0.5 km 2 for our region (Bolch et al, 2010). The Columbia Mountains have hundreds of unmapped small glaciers, but adding these in the assessment would likely make a minor contribution to total ice volume (Leigh et al, 2019).…”

“…It is known that regional parameterization (e.g. tuning model parameters to reproduce regionally averaged mass balance gradients) introduces greater uncertainty for individual glaciers (Helfricht et al, 2019;Rabatel, Sanchez, Vincent, & Six, 2018). Tuning the model with local instead of regionally-averaged estimates of glacier mass balance is a way toward improved spatially-resolved ice thickness estimates, but is beyond the scope of this study.…”