Remote sensing of suspended solids concentration in a reservoir with frequent wildland fires on its watershed

Abstract

Wildland fire is an important disturbance factor that can cause severe ecological and watershed damage. Depending on fire severity and watershed extension, reservoir suspended solids concentration (SSC), which arrives through river load, is expected to increase. Satellite remote sensing is an alternative technique to measure SSC in a reservoir. In this paper we evaluate the applicability of multitemporal Landsat data for mapping and monitoring of the SSC in a reservoir whose watershed was exposed to fires. Besides, we aim to identify catchment areas that have been burned, estimating the level of burn severity that occurred because of the fire. The Landsat images were radiometrically, atmospherically and geometrically corrected. Using the differenced normalized burn ratio (dNBR) algorithm, the perimeter and severity of fire was mapped. A theoretical model to characterize the distribution of SSC using multitemporal Landsat data was developed. The relationship between burn severity maps, rainfall and SSC maps improved our understanding of management actions on a reservoir which suffers frequent wildfires on its watershed. The theoretical model here developed may be considered as a low cost measurement tool for water management authorities, particularly when in-situ data are not available.