Identification of optically active constituents in water from the Guaíba Lake from orbital sensors data and field spectroradiometry

Abstract

The Guaíba Lake is the main source of water supply to Porto Alegre metropolitan region. Remote sensing is increasingly used for the discrimination of the different elements presented in water. This work aims to identify the active optically constituents of the water of Guaíba Lake, as regards their optical properties, from integration of data orbital remote sensing and spectroradiometry in field. To achieve this, limnologic and spectral variables were collected and related. Derivation techniques were applied in order to identify the optically active constituents, which determine the characteristics of the water. An image of the sensor LISS-III/ResourceSat-2 was used for the empirical estimation of the limnologic parameters. The results showed that empirical regressions explained that 23% of the reflectance in 708 nm is due to the presence of solids in the water, and the concentration of Chl-a did not present significant correlations (α = 0, 05) with the reflectance. From the analysis of the spectral data by the Derivative technique, was observed the increase of correlations after the derivation, highlighting the fact that this technique increases the spectral contrast and so the estimation accuracy. The first derivation in 772 nm explained 44% of the variation due to total suspended solids. Already in 690 nm, the first derivation explained 32% of the variation in the concentration of chlorophyll-a. The best results were observed in the turbidity and water transparency variables. The images LISS-III showed themselves to be potentials to map Guaíba Lake composition, by means of linear regressions.