Spatiotemporal total suspended matter estimation in Itumbiara reservoir with Landsat-8/OLI images

Abstract

The transparency of water is affected by the amount of sunlight available, suspended particles and dissolved solids such as colored dissolved organic material present in the water column. High concentrations of total suspended matter (TSM) reduce water clarity, which can affect photosynthesis of submerged aquatic vegetation, thereby affecting oxygen production which is essential to aquatic organisms at upper levels in the food chain. The aim of this work is to evaluate the use of Landsat-8 Operational Land Imager (OLI) sensor to estimate TSM concentrations in the Itumbiara hydroelectric reservoir, Midwest Brazil (18°25′ S, 49°06′ W). Concurrent proximal remote-sensing and limnological data were collected in May and September 2009, acquired between 10:00 and 14:00 (Brazil time UTC-3) to provide representative daily readings. In situ above-water radiometric data were used to simulate remote-sensing reflectance (Rrs) for the Landsat-8/OLI spectral bands. TSM empirical models were derived from Landsat-8/OLI simulated spectral bands. The data set acquired in September 2009 was used to derive the models and the data collected in May 2009 was used for validation. To assess the similarities and differences between measured and model derived TSM concentrations, two statistical indicators were calculated. The model with lowest error was applied to selected Landsat-8/OLI images. Preliminary results showed that the model with lowest error was calibrated using Rrs from bands 2 and 3 as index. The results obtained here show that Landsat-8/OLI sensor has enough sensibility to estimate TSM concentrations in inland waters in Brazil.