Using National-Scale Data To Develop Nutrient-Microcystin Relationships That Guide Management Decisions.

Area of Science:

• Environmental Science
• Ecotoxicology
• Water Resource Management

Background:

• Cyanotoxin contamination in lakes poses risks to recreation and drinking water.
• Developing predictive models is challenging due to low detection limits and high variability in cyanotoxin concentrations.

Purpose of the Study:

• To develop a national-scale hierarchical Bayesian model to predict microcystin concentrations.
• To address data limitations like below-detection-limit samples and high variability.

Main Methods:

• Utilized a hierarchical Bayesian modeling approach.
• Incorporated summer mean total nitrogen and total phosphorus concentrations as predictors.
• Analyzed data from lakes and reservoirs across the conterminous United States.

Main Results:

• The model successfully predicts mean microcystin concentrations, accounting for 69% of the variance.
• Total nitrogen showed a stronger association with microcystin concentrations than total phosphorus.
• The model effectively handles data with concentrations below detection limits.

Conclusions:

• The developed Bayesian model provides a robust tool for predicting microcystin levels in lakes and reservoirs.
• Nutrient management, particularly of total nitrogen, can be guided by these predictions to mitigate cyanotoxin risks.
• The study offers a framework for evaluating similar models for practical water management applications.