Modelagem orbital e monitoramento remoto de clorofila, sólidos suspensos e densidade celular de Ceratium na Barragem Santa Bárbara, Pelotas, RS.

Abstract

Chlorophyll-a (Chl-a), total suspended solids (SST) and cell density of organisms such as Ceratium are important indicators of reservoir water quality. Their continued monitoring is essential and remote sensing emerges as an alternative for monitoring of these indicators since existing conventional methods (e.g., in situ monitoring) are limited in time and space. The main aim of this research was to develop empirical models for the concentrations of Chl-a, SST and Ceratium, by multiple linear regressions between spectral data of Sentinel-2 images, obtained directly from the Google Earth Engine platform, and analysis of samples collected in situ from the Barragem Santa Bárbara, Pelotas/RS. A combination of 10 spectral bands, 8 band ratios and 9 spectral indices was used to generate correlations within situ analyses. As a result, the strongest correlations for Chl-a and SST were obtained with Band 5 and with the ratios between the red and near infrared bands, being chosen as the independent variables used in the regressions. Ceratium modeling was performed using a dual-model strategy. A regression was developed for Chl-a using the cell density of Ceratium as a predictor (R²=0.863 and RMSE=61.1) to be coupled with the previously generated algorithm. The bio-optical models best evaluated for Chl-a (R²=0.64, RMSE=122) and For SST (R²=0.61 and RMSE=4.33), and the combined Ceratium model (R²=0.57, RMSE=700) were used to produce spatial distribution maps. The spatial distribution maps derived from Sentinel-2 performed well where Chl-a and Ceratium were relatively high.