Estimativa da concentração de sólidos suspensos em águas opticamente complexas a partir de sensores remotos

Abstract

Currently the concentration of total suspended solids in coastal waters are often poorly controlled. This is because most of the traditional basic research are often unable to identify temporal fluctuations and are limited in their ability to control a full measure of these environments. Such problems can be, addressed using remote sensing technology; they provide general and frequent visions of synoptic coastal zone, which allows them to be more effective spatial dynamics and temporal analysis. This work aims to develop a method of predicting the concentration of total suspended solids (TSS) in Manguaba lagoon, performed by a combination of remote sensing techniques and statistical analysis. To it was necessary to perform sampling campaigns, to determine the profile of the spectral points collected and the concentration of TSS, use a set of hyperspectral radiometers to capture the spectral response of the targets and so correlate with TSS values. Our results show two models tailored MODIS images (r² 83.4%, RMSE = 3.47 mg / L) and Landsat 8 (r² 85.2%, RMSE = 3.28 mg / L). Both, the spectral region of red was the one that had better fit. This spectral range is, shown ideal for suspended sediment detection due to the high spectral sensitivity from sediments into the water column, while the spectral profile of the lagoon waters have configured a small peak around 650 nm, indicating that the reflectance it was, reduced by a prominent spectral pattern of algae. Finally, this study indicates explicit empirical models based on the band of red to predict the concentration of total suspended solids (TSS).