Investigating the relationship between microcystin concentrations and water quality parameters in three agricultural irrigation ponds using random forests

Authors: SMITH, JACLYN - Oak Ridge Institute For Science And Education (ORISE); Stocker, Matthew; WOLNY, JENNIFER - Food And Drug Administration(FDA); HILL, ROBERT - University Of Maryland; Pachepsky, Yakov

Submitted to: Water
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 8/2/2025
Publication Date: N/A
Citation: N/A

Interpretive Summary: Blue-green algae naturally grow in bodies of water. Their uncontrolled growth results in the release of toxic substances, posing a threat to both human and animal health. Those toxins can be transported to crops and soil during irrigation; they can remain in soil for extended periods and reach the edible parts of plants. Microcystin is one of the toxins. Assessment of microcystin pollution in agricultural water sources is challenging, as its concentrations can vary widely across different water sources, and measuring these concentrations in multiple locations requires substantial resources. One way to avoid these difficulties is to relate microcystin concentrations to easily measurable water quality parameters. Those relationships are complex, and machine learning, a branch of artificial intelligence, needs to be utilized. Random forest is one of the most popular and efficient machine learning methods, and sensing water with electrodes is a relatively straightforward method for characterizing water quality. We found that the application of random forest models with water quality sensing data could explain up to 70% of the variation in microcystin concentrations in three irrigation ponds, one in Maryland and two in Georgia. The results of this work can be of use to agricultural water managers and water quality consultants, as it suggests an efficient method for assessing the health risks associated with microcystin in waters used in agricultural production.

Technical Abstract: Cyanotoxins in agricultural waters pose a human and animal health risk. These toxins can be transported to nearby crops and soil during irrigation; they can remain in the soils for extended periods and be adsorbed by root systems. Additionally, in livestock watering ponds cyanotoxins pose a direct ingestion risk. This work evaluated the performance of the random forest algorithm in estimating microcystin concentrations from eight in situ water quality measurements at one active livestock water pond and two working irriga-tion ponds in Georgia and Maryland, USA. Measurements of microcystin along with eight in situ sensed water quality parameters were used to train and test the machine learning model. The models performed better at the Georgia ponds compared the Maryland ponds and interior models performed better than nearshore or whole pond models. The most important variables for microcystin prediction were water temperature and phytoplank-ton pigments. Overall, the random forest algorithm was able to explain 40% to 70% of the microcystin concentration variation in the three agricultural ponds. In situ sensing showed a potential to aid in the water sampling design by predicting the microcystin concentrations in the studied ponds by using readily available in situ sensing data.