Hi,
I have used S3-OCLI product to extract Rrs from Barents Sea. The spectra data shows that blue band has got higher reflectance than the green band (clear waters case). However, Barents Sea is optically complicated. I am also getting -ve Rrs and estimated Ch-a value for some pixels of interets(lat/long insitu samples) . Also, the regression result aftre applying various machine learning algorithms is too bad. Any suggestions please.
Asim