Investigating empirical models to retrieve the backscattering slope from Landsat 8 Images - The case of Lago Grande Curuai

This paper aims on the investigation of well-established models to retrieve the backscattering slope γbbp from remote sensing images applied to a particular Amazon floodplain lake, named Curuai Lake. Lee, Carder e Arnone (2002) γbbp model and its derivation, Pan et al. (2015) model, were applied to measured Remote Sensing Reflectance and retrieved bbp was compared to measured bbp taken in two distinct hydrological periods: receding and rising water periods (August 2013 and April 2014). Also, an empirical model, based on Lee, Carder e Arnone (2002) approach, was derived from a nonlinear fitting to the measured dataset, considering only Landsat 8 bands combination. Lee, Carder e Arnone (2002) and Pan et al. (2015) models were not suitable to retrieve backscattering slope in Curuai Lake whist the empirical model using bands B4(655)/B5(865) presented the best result considering both hydrological periods and was therefore applied to Landsat 8 images acquired in August 2013 and April 2014. The qualitative inspection of both images reveal consistent backscattering slope retrievals which, with further analysis, could lead to a suitable proxy for particle size in Curuai Lake.