A physics-based framework for remote sensing inversion of fluorescent dissolved organic matter: incorporating...

Fluorescent dissolved organic matter (fDOM) is a vital tracer of aquatic carbon cycling and water quality dynamics, yet satellite retrievals remain largely constrained by empirical correlations and semi-analytical mappings that neglect fluorescence as an explicit inelastic radiative process, limiting physical interpretability. To address this limitation, we develop a physics-based inversion framework that treats DOM fluorescence as an inelastic source term in the radiative transfer equation, coupling excitation-emission processes with elastic scattering and constraining state vectors with apparent quantum yield priors. Implemented with synchronous field observations in the Pearl River Estuary, the framework (Route B) explicitly simulates excitation, emission, and propagation of fDOM and applies Bayesian optimization for parameter retrieval. Compared with a semi-analytical baseline (Route A) that infers fDOM indirectly from chromophoric DOM absorption, Route B reduces mean parameter uncertainty by approximately 20 % and improves fDOM retrieval accuracy in optically complex waters. These results indicate that explicitly accounting for fluorescence enhances retrieval robustness in optically complex waters and provides an observation-driven pathway for scalable regional and cross-regional fDOM monitoring.