Regional chlorophyll- a retrieval algorithm for Indian coastal waters using optical classification and band optimization

Accurate chlorophyll-a (chl-a) retrieval in coastal waters remains challenging due to optical interference from coloured dissolved organic matter (CDOM) and total suspended matter (TSM). This study developed a regional chl-a algorithm for Indian coastal waters through optical classification and derivative-based band optimization using 287 in situ remote sensing reflectance (Rrs) measurements. Four distinct optical water types were identified based on peak Rrs wavelength: Type-1 (400–450 nm), Type-2 (450–500 nm), Type-3 (500–550 nm), and Type-4 (550–600 nm). These types showed clear spatial gradients, with Type-1 waters occurring farthest from shore (38 ± 21 km), and Type-4 waters closest (6 ± 3 km), reflecting increasing influence of terrestrial inputs. Average optically active substance concentrations increased progressively from Type-1 to Type-4, with mean chl-a ranging from 0.43 to 3.68 mg m−3, CDOM from 0.95 to 1.92 ppb, and TSM from 3.64 to 11.13 mg l−1. Second-order derivative analysis identified 12 common Rrs spectral peaks across all water types (415, 440, 465, 490, 515, 530, 550, 570, 600, 650, 665, and 685 nm), representing wavelengths where chl-a maintains a diagnostic influence despite the presence of other optically active substances. Band-ratio combinations of Rrs at 415, 440, 465, 490, and 530 nm against 550 nm demonstrated the strongest correlation with in situ chl-a (R2 = 0.91) when fitted with a fourth-order polynomial regression. The validation of the developed algorithm against in situ observations showed that the proposed band-ratio algorithm substantially outperforms existing ocean-colour algorithms, reducing RMSE from 0.87–5.60 mg m− 3 to 0.52 mg m− 3 while achieving the highest correlation (r = 0.95). The algorithm’s incorporation of blue bands below 443 nm, typically excluded in global models, proved crucial for improved accuracy. These optimal spectral bands align well with the capabilities of recently launched hyperspectral missions (PACE, PRISMA, EnMAP) and future ocean colour missions, offering pathways for operational implementation and enhanced coastal water quality monitoring.