Semiautomatic inventory and geomorphological characterization of mass movements using high-resolution images and open-source software in the Ribeira de Iguape Valley, Brazil

Abstract

In Brazil, as in many countries, inventories of mass movements usually exist for specific events and on a watershed scale and are predominantly produced from visual interpretation of aerial or orbital images, and those produced using automatic and semiautomatic methods are incipient in the national context. One of these events, the Ribeira de Iguape River Valley (Southeast Brazil), was severely affected by mass movements between January 13 and 14, 2014. The objective of this article was to produce a semiautomatic inventory and perform a brief geomorphological evaluation. of the mass movements triggered. A TanDEM-X digital elevation model (DEM) (12 m) was used to extract the watersheds (area: 109.7 km2) most affected by mass movements and multispectral images of the RapidEye satellites, captured before (01-10-2014) and after (01-30-2014) the event. From the multispectral images, the pre-event and post-event NDVI and the difference between the NDVI on these two dates were computed. Using the angle map, contribution area and topographic wetness index (TWI), the map of affected areas was refined into two classes: deposits of debris flows and other hydrodynamic mass movements (slope angle <6°; TWI <14; and contribution area >50,000 m2) and scars from shallow landslides (slope angle >6°; TWI <14; contribution area <50,000 m2). Approximately 5% of the total area was affected by mass movements, with 1.44 km2 of area corresponding to deposits of flows and other hydrodynamic mass movements and 4.04 km2 of area corresponding to scars from landslides. In total, 2089 landslide scars were mapped. The most affected slopes had angles between 20° and 40°, a TWI less than 9.0 and a concave curvature in profile and in planform, and approximately 90% of the scars occurred in areas covered by dense vegetation (NDVI >0.80 on 01-10-2014). The results may contribute to the expansion of the application of automatic and semiautomatic methods for the detection of mass movement scars in Brazil in a systematic manner using open source software.