While landslides can be initiated under a range of natural triggering factors, such as seismic activity, stream cutting, and freeze/thaw processes, rainfall has been shown to be the dominant trigger in mass movement events. At the local scale, intense or prolonged rainfall can increase pore water pressures and decrease soil cohesion in the subsurface, causing the driving forces to overcome resisting forces on a hillslope and activate a landslide. Over larger areas, landslide initiation processes are often difficult to quantify due to uncertainties in rainfall intensities over areas with heterogeneous or scarce ground-based instrument coverage. Satellite precipitation estimates have provided more accessible and homogeneous coverage of rainfall over the globe; however, the resolution and accuracy often remains too coarse for landslide hazard evaluations. As a result, characterizing the spatiotemporal distribution of landslide occurrences and forecasting their potential impacts over larger areas is often limited by the accuracy and spatial coverage of precipitation information.

This research advances a prototype satellite-based global landslide hazard algorithm framework, which evaluates how multi-satellite precipitation estimates can be used to forecast potential landslide conditions. An analysis of this algorithm using a global landslide catalog (GLC) suggests that forecasting errors are geographically variable and high-intensity rainfall events can be underestimated, particularly in complex topographic settings and over short time intervals. The current study seeks to improve upon these identified challenges by considering higher spatial and temporal resolution landslide susceptibility information and testing different precipitation indices for current multi-satellite precipitation products. Satellite precipitation information from the Tropical Rainfall Measuring Mission (TRMM) Multi-satellite Precipitation Analysis (TMPA) product is currently considered within a rainfall intensity-duration triggering threshold as well as in calculations of antecedent rainfall and previous soil moisture conditions. Higher resolution, ground-based regional precipitation products are also tested within this framework and illustrate the need for globally-consistent, accurate rainfall products.

The Global Precipitation Measurement (GPM) Mission will provide near real-time, higher resolution satellite products that will be vital in improving landslide forecasts within this algorithm framework. The unified multi-satellite GPM data products offer the detailed, global, and timely observations needed to estimate, monitor, and forecast extreme rainfall that may trigger landslide events. With improved understanding of rainfall triggers within different climate regimes and topographic settings around the world, this research shows how near real-time, merged satellite GPM products can improve rainfall-triggered landslide hazard assessment and forecasting and help to make a more clear connection between local scale hydrological modeling and larger-area statistical approaches.