Algorithm Theoretical Basis Document (ATBD) for a PCA-Based Regional Chlorophyll-a Retrieval Algorithm in the Gulf of St. Lawrence and Gulf of Maine

This repository contains the Algorithm Theoretical Basis Document (ATBD) for a regional chlorophyll-a (Chla) retrieval algorithm developed within the SIMBA (Integrated Modelling System for the North Atlantic Right Whale) project, part of the smartWhales initiative funded by the Canadian Space Agency (CSA), Fisheries and Oceans Canada (DFO), and Transport Canada. The ATBD documents the theoretical background, data sources, algorithm design, implementation, and performance evaluation of a Principal Component Analysis (PCA)–based empirical algorithm specifically designed to estimate surface chlorophyll-a concentrations in the Gulf of St. Lawrence, Scotian Shelf, and Gulf of Maine (GoSLM)—a region characterized by optically complex coastal waters with high concentrations of colored dissolved organic matter (CDOM). The algorithm was developed using an extensive satellite–in situ matchup dataset (1997–2021) combining observations from multiple ocean color sensors, including MERIS, MODIS, OLCI (A/B), SeaWiFS, VIIRS-J, and VIIRS-N, processed through the Copernicus GlobColour framework. In situ chlorophyll-a measurements were compiled from Fisheries and Oceans Canada (DFO), the Bigelow Laboratory, and additional regional monitoring programs. The methodology applies PCA (also referred to as empirical orthogonal functions) to log-transformed and standardized remote sensing reflectance spectra (Rrs), followed by sensor-specific multiple linear regression models to estimate chlorophyll-a concentrations. The ATBD provides full documentation of: data selection and preprocessing, PCA computation and eigenvector derivation, regression model calibration and selection (AIC-based), algorithm implementation within the GlobColour processing chain, and quantitative performance evaluation against established algorithms (OCx, OC5, GSM). Results demonstrate that the PCA-based regional algorithm outperforms global chlorophyll-a algorithms across all evaluated sensors in the GoSLM, with reduced bias and lower RMSE when compared to in situ observations. The document also discusses known limitations, including reduced dynamic range at extreme chlorophyll concentrations, and outlines perspectives for future improvements. This ATBD supports transparent, reproducible use and reuse of the SIMBA chlorophyll-a algorithm for regional oceanographic studies, ecosystem monitoring, and species distribution modeling, particularly in coastal and high-CDOM environments.