60182

Estimativa do coeficiente de absorção total do reservatório de Ibitinga por meio de um modelo quase-analítico re-parametrizado

Favoritar este trabalho

The water represents a complex mixture of a variety of dissolved or particulate materials having different optical properties that interact with light. The water color modeling aims to express the remote sensing reflectance in terms of Inherent Optical Properties (IOPs), such as total absorption coefficient (at). Application of quasi-analytic algorithm (QAA), which derives total absorption coefficient (at) from remote sensing reflectance (Rrs), in inland waters generally returns inconsistent results and, therefore, some versions were developed aiming to retrieve more accurate values to these environments. A QAA parameterization, named QAA_BBHR, was developed for Barra Bonita reservoir, belonging to a cascade system situated at Tietê River, São Paulo State, Brazil. This paper aims to assess the applicability of QAA_BBHR to the data collected in the Ibitinga reservoir, pertaining to the same cascade system. The adjustment was assessed through the statistical analyses, Root Mean Squared Error (RMSE), Normalized Root Mean Squared Error (NMRSE), bias and the coefficient of determination (R²). Analysis of Rrs spectra, obtained from Ibitinga reservoir data, suggests high chlorophyll - a concentration (from 1.37 to 119.04 mg.m-3), indicating eutrophic condition, as well as presented in Barra Bonita reservoir. R² value was 0.74, considering all wavelengths and average NRMSE was 19.07 %. Minimum errors were observed at 665 and 681 nm, whereas the highest errors occurred at 709 nm. The greatest errors at 709 nm probably occur due to inorganic suspended matter scattering and the increasing in pure water absorption. It was possible to conclude, from this, that QAA_BBHR proved to be applicable to the study area under consideration.