Calibração de equações para determinação de clorofila-a pôr Sensoriamento Remoto na Baía de Guanabara, RJ (Brasil)

The use of remote sensing through geotechnology to estimate water quality variables is a trend in current scientific circles. Therefore, the general objective of this work was to correlate the use of remote sensing based on the construction of three empirical models as a way of estimating chlorophyll-a concentrations, added to information from in situ data. These models are from the authors Mishra & Mishra, Dall'Omo et al and Kaymaz & Ates found in the literature, where they are composed of equations that have the spectral bands as variables of the Sentinel-2A satellite, serving as a basis for promoting chlorophyll maps -The. Field measurements are fundamental for this category of study because they establish the veracity of the data found in the images. In the meantime, fieldwork was carried out on June 17, 2022, starting at 10:30 am and ending at 4:30 pm, in which 70 concentrations of chlorophyll-a were measured using the Hydrolab HL7 Multiparameter Probe. With data from measured fields, statistical results were established in two stages: calibration and validation, with 57 data for the first and 13 for the second. In the Mishra & Mishra model, the calibration results were r² = 0.74 and the Root Mean Square Error (RMSE) of 6.02 μg/L, while in the validation it was r² = 0.82 and RMSE of 5. 17 μg/L. The model by Dall'Omo et al, presented in the first stage the r² of 0.77 with the RMSE of 5.90 μg/L, later in the second stage, the r² was 0.84 and RMSE of 5.17 μg/L L. The third model by Kaymaz & Ates obtained a calibration result of r² of 0.75 and RMSE of 5.88 μg/L. Validation was r² = 0.82 and RMSE of 5.16 μg/L.