Earthquake‐induced landslide (EQIL) inventories are essential tools to extend our knowledge of the relationship between earthquakes and the landslides they can trigger. Regrettably, such inventories are difficult to generate and therefore scarce, and the available ones differ in terms of their quality and level of completeness. Moreover, access to existing EQIL inventories is currently difficult because there is no centralized database. To address these issues, we compiled EQIL inventories from around the globe based on an extensive literature study. The database contains information on 363 landslide‐triggering earthquakes and includes 66 digital landslide inventories. To make these data openly available, we created a repository to host the digital inventories that we have permission to redistribute through the U.S. Geological Survey ScienceBase platform. It can grow over time as more authors contribute their inventories. We analyze the distribution of EQIL events by time period and location, more specifically breaking down the distribution by continent, country, and mountain region. Additionally, we analyze frequency distributions of EQIL characteristics, such as the approximate area affected by landslides, total number of landslides, maximum distance from fault rupture zone, and distance from epicenter when the fault plane location is unknown. For the available digital EQIL inventories, we examine the underlying characteristics of landslide size, topographic slope, roughness, local relief, distance to streams, peak ground acceleration, peak ground velocity, and Modified Mercalli Intensity. Also, we present an evaluation system to help users assess the suitability of the available inventories for different types of EQIL studies and model development.

In active mountain belts with steep terrain, bedrock landsliding is a major erosional agent. In the Himalayas, landsliding is driven by annual hydro-meteorological forcing due to the summer monsoon and by rarer, exceptional events, such as earthquakes. Independent methods yield erosion rate estimates that appear to increase with sampling time, suggesting that rare, high-magnitude erosion events dominate the erosional budget. Nevertheless, until now, neither the contribution of monsoon and earthquakes to landslide erosion nor the proportion of erosion due to rare, giant landslides have been quantified in the Himalayas. We address these challenges by combining and analysing earthquake- and monsoon-induced landslide inventories across different timescales. With time series of 5 m satellite images over four main valleys in central Nepal, we comprehensively mapped landslides caused by the monsoon from 2010 to 2018. We found no clear correlation between monsoon properties and landsliding and a similar mean landsliding rate for all valleys, except in 2015, where the valleys affected by the earthquake featured ∼5–8 times more landsliding than the pre-earthquake mean rate. The long-term size–frequency distribution of monsoon-induced landsliding (MIL) was derived from these inventories and from an inventory of landslides larger than ∼0.1 km² that occurred between 1972 and 2014. Using a published landslide inventory for the Gorkha 2015 earthquake, we derive the size–frequency distribution for earthquake-induced landsliding (EQIL). These two distributions are dominated by infrequent, large and giant landslides but under-predict an estimated Holocene frequency of giant landslides (> 1 km³) which we derived from a literature compilation. This discrepancy can be resolved when modelling the effect of a full distribution of earthquakes of variable magnitude and when considering that a shallower earthquake may cause larger landslides. In this case, EQIL and MIL contribute about equally to a total long-term erosion of $\sim \mathrm{2}\pm \mathrm{0.75}$ mm yr⁻¹ in agreement with most thermo-chronological data. Independently of the specific total and relative erosion rates, the heavy-tailed size–frequency distribution from MIL and EQIL and the very large maximal landslide size in the Himalayas indicate that mean landslide erosion rates increase with sampling time, as has been observed for independent erosion estimates. Further, we find that the sampling timescale required to adequately capture the frequency of the largest landslides, which is necessary for deriving long-term mean erosion rates, is often much longer than the averaging time of cosmogenic ¹⁰Be methods. This observation presents a strong caveat when interpreting spatial or temporal variability in erosion rates from this method. Thus, in areas where a very large, rare landslide contributes heavily to long-term erosion (as the Himalayas), we recommend ¹⁰Be sample in catchments with source areas > 10 000 km² to reduce the method mean bias to below ∼20 % of the long-term erosion.

We present a new, seismologically consistent expression for the total area and volume of populations of earthquake‐triggered landslides. This model builds on a set of scaling relationships between key parameters, such as landslide spatial density, seismic ground acceleration, fault length, earthquake source depth, and seismic moment. To assess the model we have assembled and normalized a catalog of landslide inventories for 40 shallow, continental earthquakes. Low landscape steepness causes systematic overprediction of the total area and volume of landslides. When this effect is accounted for, the model predicts the total landslide volume of 63% of 40 cases to within a factor 2 of the volume estimated from observations (R²=0.76). The prediction of total landslide area is also sensitive to the landscape steepness, but less so than the total volume, and it appears to be sensitive to controls on the landslide size‐frequency distribution, and possibly the shaking duration. Some outliers are likely associated with exceptionally strong rock mass in the epicentral area, while others may be related to seismic source complexities ignored by the model. However, the close match between prediction and estimate for about two thirds of cases in our database suggests that rock mass strength is similar in many cases and that our simple seismic model is often adequate, despite the variety of lithologies and tectonic settings covered. This makes our expression suitable for integration into landscape evolution models and application to the anticipation or rapid assessment of secondary hazards associated with earthquakes.

Rainfall-induced landslides are a common and significant source of damages and fatalities worldwide. Still, we have little understanding of the quantity and properties of landsliding that can be expected for a given storm and a given landscape, mostly because we have few inventories of rainfall-induced landslides caused by single storms. Here we present six new comprehensive landslide event inventories coincident with well identified rainfall events. Combining these datasets, with two previously published datasets, we study their statistical properties and their relations to topographic slope distribution and storm properties. Landslide metrics (such as total landsliding, peak landslide density, or landslide distribution area) vary across 2 to 3 orders of magnitude but strongly correlate with the storm total rainfall, varying over almost 2 orders of magnitude for these events. Applying a normalization on the landslide run-out distances increases these correlations and also reveals a positive influence of total rainfall on the proportion of large landslides. The nonlinear scaling of landslide density with total rainfall should be further constrained with additional cases and incorporation of landscape properties such as regolith depth, typical strength or permeability estimates. We also observe that rainfall-induced landslides do not occur preferentially on the steepest slopes of the landscape, contrary to observations from earthquake-induced landslides. This may be due to the preferential failures of larger drainage area patches with intermediate slopes or due to the lower pore-water pressure accumulation in fast-draining steep slopes. The database could be used for further comparison with spatially resolved rainfall estimates and with empirical or mechanistic landslide event modeling.

Inventories of individually delineated landslides are a key to understanding landslide physics and mitigating their impact. They permit assessment of area–frequency distributions and landslide volumes, and testing of statistical correlations between landslides and physical parameters such as topographic gradient or seismic strong motion. Amalgamation, i.e. the mapping of several adjacent landslides as a single polygon, can lead to potentially severe distortion of the statistics of these inventories. This problem can be especially severe in data sets produced by automated mapping. We present five inventories of earthquake-induced landslides mapped with different materials and techniques and affected by varying degrees of amalgamation. Errors on the total landslide volume and power-law exponent of the area–frequency distribution, resulting from amalgamation, may be up to 200 and 50%, respectively. We present an algorithm based on image and digital elevation model (DEM) analysis, for automatic identification of amalgamated polygons. On a set of about 2000 polygons larger than 1000 m2, tracing landslides triggered by the 1994 Northridge earthquake, the algorithm performs well, with only 2.7–3.6% incorrectly amalgamated landslides missed and 3.9–4.8% correct polygons incorrectly identified as amalgams. This algorithm can be used broadly to check landslide inventories and allow faster correction by automating the identification of amalgamation.