Development of a soil moisture forecasting method for a landslide early warning system (LEWS): Pilot cases in coastal regions of Brazil

Climate change has increased the frequency of extreme weather events and, consequently, the number of occurrences of natural disasters. In Brazil, among these disasters, floods, flash floods, and landslides account for the highest number of deaths, the latter being the most lethal. Bearing in mind the importance of monitoring areas susceptible to disasters, the REMADEN/REDEGEO project of the National Center for Monitoring and Natural Disaster Alerts (Cemaden) has promoted the installation of a network of soil moisture sensors in regions with a long history of landslides. This network was used in the present paper as a base to develop a system for moisture forecasting in those critical zones. The time series of rainfall and moisture were used in an inversion algorithm to obtain the geotechnical parameters of the soil. Then the geotechnical model was used in a forward calculation with the rainfall prediction to obtain the soil moisture forecast. The landslide events of March 2020 and May 2022 in Guarujá and Recife, respectively, were used as study cases for the developed system. The obtained results indicate that the proposed methodology has the potential to be used as an important tool in the decision-making process for issuing landslide alerts.