Search

Landslides and mass movements are essential geomorphological features driving landscape evolution and causing denudation at different magnitudes. Albeit the number of reported landslide inventories and their quality are continuously growing, global estimates of sediment mobilized through landslides remain poorly constrained. Here we present a first quantification of the contemporary global landslide mobilization rates (LMR). Statistical analyses of a dataset of contemporary LMR observations for 116 study sites worldwide showed that patterns of LMR are mainly controlled by topography (local relief), seismicity and lithology. Based on these results and taking into account uncertainties on the LMR observations, we calibrated and validated a robust empirical model that explains about 62% of the observed variance in LMR. By applying this model globally, we estimate that on average about 56 ± 5 billion m³ (~100 gigaton) of landslide material is annually mobilized. Intercontinental and interregional differences are large with the Central and Southeast Asian mountain ranges (a.o. the Himalaya and Karakoram, covering ~2% of the global land mass) generating 50% of this rate. Comparison with observed catchment sediment yields confirms the large importance of landsliding as a denudation process. Nonetheless, our analyses clearly indicate that a large proportion of the landslide derived sediments are stored within catchments, highlighting the important role of internal catchment buffering on sedimentary pathways. [ABSTRACT FROM AUTHOR]

Abstract: Event-based landslide maps display landslide events of different magnitudes that are associated with a single major trigger. Event-based landslide maps allow the estimation of spatio-temporal probabilities of landslide events that are required for the assessment of landslide hazard. In this paper, we discuss the generation of event-based landslide inventory maps for a region in Darjeeling Himalayas (India) with limited historical landslide information. We used eight different sets of source data, spanning the period from 1968 to 2007, which include topographic maps, satellite images and post-event field maps. For each period we mapped the various landslide types, and calculated landslide density and activity in individual terrain units to analyze changes in the spatial behavior of landslides through time. For the assessment of the temporal landslide probability, we used landslide event-days and associated daily and antecedent rainfall amounts in a multivariate statistical analysis to model the temporal relationship between landslide events and the amount of triggering rainfall. For estimating the magnitude probability, we applied magnitude–frequency analysis for the different event-based landslide maps using landslide area (m2) as a proxy. Despite the incompleteness of the source data sets used, the resulting landslide inventory maps can be used successfully in determining temporal and magnitude probabilities of future landslides. [Copyright &y& Elsevier]