A mechanistic population-level (i.e. differential equation) model of Microcystis growth and toxin production

Blooms of the cyanobacteria Microcystis and its toxin microcystin (MC) are an important global environmental problem. Effective management of waterbodies impaired by these cyanobacteria harmful algal blooms (CyanoHABs) requires modeling, but existing operational codes generally do not include toxins. Recently, we presented a model of Microcystis growth and toxin production, which builds on a large literature meta-analysis and can reproduce the observed succession of toxigenic and non-toxigenic genotypes and MC concentration in Lake Erie (DOI: 10.1126/science.abm6791) and other systems (DOI: 10.1016/j.hal.2024.102711). However, that model's complexity and agent-based modeling (ABM) structure make it difficult to use. Here, we present a simpler and more general population-level model (PLM) that can be integrated into existing operational Eulerian modeling frameworks. The model includes toxigenic and non-toxigenic strains and state variables for intracellular and extracellular MCs. Processes include MC synthesis as a function of temperature, light and nitrogen availability, benefit of MC (e.g. protection against H2O2 damage), excretion and degradation (extracellular). Two versions for variable (i.e. Droop) and fixed (i.e. Monod) phytoplankton nutrient composition are presented. The PLM can also reproduce the field observations, and simulations of management options are consistent with the ABM, including an increase in MC concentration under a 40 % phosphorus only reduction strategy for systems with partial N-limitation. The model constitutes a substantial step towards closing the gap between biological understanding and operational management models.