Application of a Mechanistic Model for the Prediction of Microcystin Production by Microcystis in Lab Cultures and Tropical Lake

Abstract

Microcystin is an algal toxin that is commonly found in eutrophic freshwaters throughout the world. Many studies have been conducted to elucidate the factors affecting its production, but few studies have attempted mechanistic models of its production to aid water managers in predicting its occurrence. Here, a mechanistic model was developed based on microcystin production by Microcystis spp. under laboratory culture and ambient field conditions. The model was built on STELLA, a dynamic modelling software, and is based on constitutive cell quota that varies with nitrogen, phosphorus, and temperature. In addition to these factors, varying the decay rate of microcystin according to its proportion in the intracellular and extracellular phase was important for the model's performance. With all these effects, the model predicted most of the observations with a model efficiency that was >0.72 and >0.45 for the lab and field conditions respectively. However, some large discrepancies were observed. These may have arisen from the non-constitutive microcystin production that appear to have a precondition of nitrogen abundance. Another reason for the large root mean square error is that cell quota is affected by factors differently between strains.