1) More information, such as geologic factors, could be incorporated into this global landslide
susceptibility map when they become available globally;
2) Finer resolution DEM data such as 6.1 x 6.1m LIDAR-based data can also improve the landslide
susceptibility mapping, even if only available over small areas;
3) Soil moisture conditions observed from NASA Aqua satellite with the Advanced Microwave Scanning
Radiometer-EOS (AMSR-E) instrument or an antecedent precipitation index accumulated from TRMM
will be examined for usefulness in this preliminary landslide detection/warning system; and
4) The empirical rainfall intensity-duration threshold triggering landslides may be regionalized using
mean climatic variables (e.g. mean annual rainfall).
Given the fact that landslides usually occur after a period of heavy rainfall, a real-time landslide prediction
system can be readily transformed into an early warning system by making use of the time lag between
rainfall peak and slope failure. Therefore, success of this prototype system bears promise as an early
warning system for global landslide disaster preparedness and hazard management. Additionally, it is
possible that the warning lead-time of global landslide forecasts can be extended by using rainfall forecasts
(1-10 days) from operational numerical weather forecast models. This real-time prediction system bears the
promise to extend current local landslide hazard analyses into a global decision-making support system for
landslide disaster preparedness and mitigation activities across the world.