Landslide-covered glaciers: towards a new global geodatabase

Landslides can modify the behavior of glaciers by delivering additional debris load from adjacent slopes onto the ice surface. Such debris covers may significantly reduce ablation and, hence, result in a positive glacier mass balance (e.g., at the Sherman Glacier in Alaska after a series of landslides that had slid onto it during the Good Friday Earthquake in 1964). In the longer term, this can entail glacier thickening and reduced ice mass velocity (e.g., at the Sioux Glacier in Alaska for a similar setting caused by the same earthquake). Conversely, surges with high ice mass velocities following rock avalanches onto glaciers were also documented (e.g., at the Bualtar Glacier in the Pakistani Karakoram and the Russian Geographical Society Glacier in the Tajik Pamirs).

As thermal and hydrological regime changes are widely accepted as factors influencing the kinematic behavior of glaciers, we focus on the relation of landslides and glacial processes to countervail the lack of data on that very topic. Glacial retreat and associated slope debuttressing combined with permafrost thawing are likely to increase the number of landslides onto glacier surfaces as global warming progresses. Therefore, systematic documentation of this phenomenon is necessary to fully assess the consequences for glacier dynamics.

The study aims to establish a new spatio-temporal geodatabase to determine – in the first place – worldwide distributions of glaciers covered by landslides, including potential clusters. In the second stage, spatio-temporal trends and event frequencies will be analyzed over a time frame reaching back about 50 years in time (i.e., to the launch of Landsat-1) using historical aerial photographs, and Landsat, ASTER, and Sentinel medium-resolution satellite imagery (i.e., 10-50 ground sampling distance). One of two essential aspects of the database is its planet scale, which ensures a broad spectrum of environmental conditions and possibly affected land systems such as Alaska, the European and New Zealand Alps, Iceland, the Himalayas and Pamirs, or Patagonia. Another major feature is an emphasized distinction of the type of debris on the glacier; moraine debris is not considered in the inventory. The database comprises information on topographic properties of the landslides (i.e., area, width, length, etc.), the approximate event times, prevailing geology (if available from sources), as well as the characteristics of the glaciers (i.e., area, velocity, thermal regime, etc.).

At the current stage, the geodatabase and its maps are not yet exhaustive, as we carry on our systematic quantification of landslide-covered glaciers by employing routines within the Google Earth Engine, comparison of existing inventories (e.g., GLIMS, RGI, WGI, etc.), and manual counter-checking and verification. We present the current state of our work with some speaking examples.

Research is funded by the National Science Center, Poland, via project number 2021/42/E/ST10/00186.