Desenvolvimento de um algoritmo para estimar a dispersão da pluma do rio Fraser-Canadá, usando a banda espectral do vermelho do satélite aqua/modis

Abstract

In the present study, the spatial-temporal dynamics of the Fraser River plume, British Columbia- Canada, has been analyzed through the retrieval data of the Moderate-Resolution Imaging Spectroradiometer Sensor ( MODIS), attached to the satellite AQUA, throughout the years of 2014 and 2015. For a development of a linear model to estimate turbidity in the region, the reflectance band data of the AQUA-MODIS satellite imagery and the in situ tubidity data provided by the BC Ferries‟ spectral radiometer in the Salish Sea were used. For the algorithm developing and processing, the softwares used were the SeaDAS, to execute imagery‟s geometric and radiometric correction; ENVI for atmospheric correction and MATLAB for matching up data, statistics analysis. From the analysis of the scatterplots of the turbidity and reflectance data, a linear regression model have been developed, due to the fact that the data have shown a linear trend, resulting in a correlation coefficient r=0.57 and RMSE=8.54 NTU. The turbidity maps processed through the algorithm have shown great agreement with the in situ data and it matches with what is expected by the background bibliography studied.