RECONSTRUCTION OF CHLOROPHYLL CONCENTRATION PRODUCTS APPLYING DINEOF AND M-DINEOF VARIANCES

Resumen

Measuring chlorophyll in oceans has become one of the most important task now in our days as it is the primary indicator of phytoplankton biomass. Nevertheless, getting information from satellite images can be very difficult due to cloud contamination and other atmospheric conditions that damage most of the images with a lot of gaps. The present study proposes DINEOF method as an alternative for gap reconstruction at chlorophyll images. DINEOF (Data Interpolating Empirical Orthogonal Functions) method, allow to fill the gaps produces by clouds in an effective way, thanks to SVD decomposition technique from linear algebra. Original image is decomposed in its primary orthonormal bases and then reconstructed using fixed optimal EOFs mode. For this study 56 images of the Falkland Area – Coasts from Brazil were downloaded from MODIS and VIIRS sensors throughout the month of February 2020. All of the images were reconstructed applying DINEOF method and some variations to it, the results are exposed and validated with standard linear regressions and RMS tests between the original available data and the reconstructed data.