Regressive Erosion at River Coca in Northeast Ecuador: Landslide Monitoring with Sentinel-1 to Support Disaster Risk Management

Abstract

After the collapse of the San Rafael waterfall in Northeast Ecuador on 2 February 2020, a regressive erosion started along the River Coca putting national infrastructure, the environment and indigenous communities at risk. A fast monitoring of areas exposed to landslides on local scales therefore is necessary to provide adequate risk management for the region. The study area, located in the Andean tropics close to the volcano Reventador, is characterized by steep slopes, seismic activity and high rainfall throughout the year. Sentinel-1 SAR data provide a solution for time-series monitoring in the region as imagery is available day and night and not affected by cloud cover. Landslide monitoring with Sentinel-1 SAR data was implemented using a bi-temporal change detection (BCD) with SNAP and a sequential change detection (SCD) with EESA Docker and the Google Earth Engine (GEE) aiming at the identification of a suited approach for fast disaster monitoring and management. The SCD showed an overall accuracy of 0.91 compared to 0.88 using the BCD approach validated with high-resolution imagery. Based on the landslide detection, hazard variables could be further identified to support future hazard and risk assessment. Fast processing of Sentinel-1 time-series data in a cloud-based environment allows for near real-time monitoring of ongoing erosion and provides a potential for pro-active measures to protect the national economy, the environment and the society.