The Development of a Landslide Alert System Based on the Prediction of Soil Moisture in Critical Cities in Brazil - Preliminary Results with the Cemaden's Observational Network

In 2019, Cemaden started the Redegeo project that aims to install 120 monitoring stations called Geotechnical Data Collection Platforms (PCDs Geo), which are composed of a rain gauge and six soil moisture sensors. This system generates accumulated rainfall (mm) and volumetric soil moisture (%) data and transmits it every 10 minutes to Cemaden's operational systems. The six moisture sensors are positioned every 50cm of depth, ranging from 0.5m to 3m. Observing soil moisture variation in the PCDs caused by rain constitutes an essential instrument for evaluating landslide risk scenarios. Several studies, based on numerical models of rainfall simulation, infiltration, and pore-pressure distribution in soil profiles, clearly point to the effect of increased moisture on the variation of the slope safety factor. This demonstrates the importance of applying PCDs data in slope stability calculation. In this project we aim to forge ahead, developing a new methodology for using this data to maximise the application of this network. Our project has two main stages, one being the development of a methodology to indirectly obtain the necessary geotechnical parameters in the soil saturation model and the other being the development of a program for the prediction of moisture based on the rainfall forecast. For the first stage, we used an adaptation of the MATLAB program "SMmodel_GA_WEB: Soil Water Balance Model" to perform the data inversion using moisture and rainfall data from the PCDs as input. For the moisture forecast stage, we adapted the direct calculation of the same program to be used with the parameters obtained with the inversion and the rainfall values of the meteorological forecast from the Global Forecast System (GFS) model. The results from the proposed method are the prediction of soil moisture from the PCDs Geo locations for each of the 6 moisture sensors. In this work, we present the results obtained for the city of Recife, Pernambuco State, which in May 2022 suffered from very serious landslide events that claimed 129 lives. Preliminary results show that the method was able to predict the increase in soil saturation at the location of the PCDs at the same time as several landslides in the vicinity.