Complex causes of landslides after ice sheet retreat: Post-LGM mass movements in the Northern Patagonian Icefield region

Abstract

Although the dynamics of individual rock-slope failures above recently shrinking glaciers have received increasing study, less is known about the spatial distribution of landslides in paraglacial settings. Here, we present a landslide inventory for large deglaciated area (~100,000 km2) situated within the Last Glacial Maximum (LGM) limits of the Northern Patagonian Icefield (NPI). Using satellite images and the TanDEM-X digital elevation model, we mapped a total of 15,543 landslides, among which 1006 are deep-seated landslides (DSLs) with area ≥0.01 km2. The distribution of DSLs is highly asymmetric in a W-E transect of the NPI region, with pronounced clustering along the semi-arid eastern front of the Patagonian Andes. The most strongly affected domain is volcanic tablelands overlying weak Miocene sedimentary rocks, but DSLs tend to also cluster along recently deglaciated (i.e. since the end of the 19th century) eastern margin of the NPI. Compared with other high mountain regions, alpine valleys of the Patagonian Andes are affected by DSLs only in <1% of their area, an order of magnitude lower than in other reported deglaciated mountains. The modest incidence of DSLs in the Patagonian Andes is due to dominance of hard granitoid rocks and relatively weak historical seismic activity. We conclude that 1) geological conditions control the distribution of DSLs and their types in the NPI region; 2) paraglacial effects play secondary (although locally important) roles in the origin of DSLs; 3) local clusters of large DSLs originate due to specifics of the post-LGM landscape evolution, involving drawdowns of glacial lakes and incision of rivers into the unconsolidated deposits; and 4) increased abundance of landslides above the recently shrinking margin of the NPI results from the repeated Holocene fluctuations of glacier snouts around the Little Ice Age (LIA) glacier limits and the spatial coincidence of glacial debuttressing effects with the presence of active faults.