Predicting photosynthesis-irradiance relationships from satellite remote-sensing observations

Photosynthesis-irradiance (PI) relationships are important for phytoplankton ecology and for quantifying carbon fixation rates in the environment. However, the parameters of PI relationships are typically unknown across space and time. Here we use machine learning, satellite remote-sensing, and a database of in-situ PI relationships to build models that predict the seasonal cycle of PI parameters as a function of satellite-observed variables. Using only surface light, temperature, and chlorophyll, we achieve an R2 of 58% for predicting photosynthesis rates at saturating light and an R2 of 78% for predicting the light saturation parameter. Predictability is maximized when averaging environmental covariates over 30-day and 25-day timescales, respectively, indicating that environmental history and community turnover timescales are important for predicting in-situ PI relationships. These results will help improve the parameterization of satellite-based primary production models and quantify emergent environmental integration timescales in photosynthetic communities.