10:45 AM - 11:00 AM
[ACG36-06] Remote estimation of global phytoplankton primary production by integrating photophysioligical model with machine learning algorithm
Keywords:Primary production, Photosynthesis , Enhanced random forest regression , Ocean-color remote sensing
Remote estimation of phytoplankton primary production (PP) has long been recognized as an important method for investigating the responses of aquatic ecosystems to global climate change. The theory-based primary production model (TPM), one of the earlier proposed photophysiological models, is potentially applicable to a variety of water bodies because of its semi-analytical nature. Its accuracy is highly dependent on whether the photophysiological response of phytoplankton is adequately parameterized, specifically the assimilation number and the light saturation parameter. The remote assignment of these parameters is acknowledged to be a challenging task, and the limited progress has impeded extensive use of the TPM. In this study, we proposed a machine learning algorithm, the enhanced random forest regression (ERFR), to retrieve the photosynthetic parameters from satellite observations. The ERFR was then integrated with the TPM (together termed as TPMERFR) to remotely estimate daily depth-integrated PP in clear to turbid waters. The ERFR was trained and validated using in situ datasets from a broad range of trophic and biogeographic conditions, covering oceanic, coastal, and inland water bodies. Validation results showed that the ERFR outperformed conventional empirical and semi-analytical algorithms. The root mean square difference (RMSD) of the satellite-based PP estimates from the TPMERFR remained within 0.27. In contrast, the benchmark models generally yielded PP estimates with RMSDs of 0.27–0.62. The TPMERFR was then implemented to climatological satellite products (2001–2022) to reassess global PP. Reasonable spatial distributions of PP were preliminarily demonstrated, especially in polar, coastal, and inland waters. These results indicated the potential utility of the TPMERFR to generate global seamless PP distributions in clear to turbid waters.