Calibration and validation of algorithms for the estimation of chlorophyll-a concentration and Secchi depth in inland waters with Sentinel-2

Marcela Pereira-Sandoval, Esther Patricia Urrego, Antonio Ruiz-Verdú, Carolina Tenjo, Jesús Delegido, Xavier Soria-Perpinyà, Eduardo Vicente, Juan M. Soria
2019 Limnetica  
Calibration and validation of algorithms for the estimation of chlorophyll-a concentration and Secchi depth in inland waters with Sentinel-2 Chlorophyll-a concentration and Secchi disk depth are two of the most important biophysical parameters used to assess water quality and determine the ecological state of inland waters. The Ocean Color 2 and Dall'Olmo three-band algorithms were used to estimate chlorophyll-a concentration and the calibration of the ratio 490/705 nm was used to produce an
more » ... orithm for estimating Secchi disk depth. These algorithms have been calibrated for the Sentinel 2-Multispectral Instrument (S2-MSI) and validated using in situ measurements of chlorophyll-a, Secchi disk depth and radiometry. This data was taken in the Valencia region reservoirs as part of the project Ecological Status of Aquatic Systems with Sentinel Satellites (ESAQS). The results show that for estimating chlorophyll-a concentration, it is better to apply a prior classification based on their trophic status. For eutrophic and hypertrophic waters, the TBDO algorithm had an error of 23 mg/m 3 over a chlorophyll-a concentration range of between 10 to 169 mg/m 3 . For ultraoligotrophic to mesotrophic waters, the better algorithm was OC2_490, which resulted in an error equal to 0.9 mg/m 3 over a chlorophyll-a concentration range of between 0.54 to 5.8 mg/m 3 . For the estimation of water transparency by Secchi disk depth, we have obtained good results with the ratio 490/705 nm, with an error equal to 0.88 m over a Secchi disk depth range of between 0.26 to 8.1 m. These algorithms have been applied to S2-MSI images and satisfactory results have been obtained for different reservoirs in the Valencia region (Spain).
doi:10.23818/limn.38.27 fatcat:ldv75h5qgnhdjimylw5ci7uqqm