CAN WE INCLUDE HYDROLOGICAL INFORMATION IN REGIONAL THRESHOLDS FOR LANDSLIDE INITIATION? ANALYSIS OF THE POTENTIAL OF A ‘CAUSE-TRIGGER’ BASED HYDROLOGICAL THRESHOLDS APPROACH (Invited Presentation)

Regional landslide hazard assessment is mainly based on empirically derived precipitation-intensity-duration (PID) thresholds. Generally, two features of rainfall events are plotted and labelled with (shallow) landslide occurrence or non-occurrence. Hereafter, a separation line is drawn, mostly logarithmic scale. The conceptual idea is the assumption that precipitation information is a good proxy for both meteorological trigger and hydrological cause. Although successfully applied in some cases, it does suffer from limited predictive value as well as limited physical process insight. This is partly due to database uncertainties but partly due to missing hydrological information.

In our research, the Precipitation-Intensity-thresholds is analyzed from a hydrological perspective. Second, we analyzed the effects of database errors on the uncertainty in the thresholds. Lastly, we work on assessing what type of hydrological information has added value for the regional landslide hazard assessment in terms of predictive capacity and is required to improve landslide forecasting using regional thresholds. We do this based on case-studies having landslide inventories, meteorological information as well as hydrological information and with synthetic modelling of 1000 year of generated hillslope hydrology and landslide inventory.

The overarching aim we have with this research is to increase the representation of the underlying hydrology in landslide regional hazard assessment. We will present a theoretical framework and analysis as well as the empirical derived hydrometeorological thresholds for landslide hazards. We will discuss the important practical barriers and limitations of the proposed approach.