Lake Erie summer chlorophyll phenology: a Bayesian additive regression trees comparison of growth and decay phases

Synergistic effects of abiotic and biotic factors determine chlorophyll, a proxy for algal biomass in lakes. With harmful algal blooms (HABs) increasingly threatening water quality, it is important to determine the extent to which lake abiotic conditions contribute to chlorophyll that generally exhibits a growth phase leading to a summer peak followed by a decay phase. Here an ensemble-tree model implemented in the Bayesian additive regression trees (BART) was used to investigate the effects of seven potential drivers on chlorophyll concentrations during both growth and decay phases in western Lake Erie in the years 2012-2022. Our findings revealed that total phosphorus (TP) consistently emerged as the dominant driver, exhibiting a positive saturating relationship. During growth, interactions between TP and nitrogen forms dominated, while beam attenuation emerged as the central interacting variable during decay phase. The TP-chlorophyll relationship was similar between the growth and decay phases of the bloom. Overall, while the TP-chlorophyll relationship is well established in freshwater lakes, the fact that TP emerges as the most important factor in an exploration that includes other nutrients, temperature, and light underscores the idea that management strategies focused on phosphorus control should be effective in reducing HABs in Lake Erie.