Toward a Predictive Understanding of Cyanobacterial Harmful Algal Blooms through AI Integration of Physical, Chemical, and Biological Data

Abstract

Freshwater cyanobacterial harmful algal blooms (cyanoHABs) are a worldwide problem resulting in substantial economic losses, due to harm to drinking water supplies, commercial fishing, wildlife, property values, recreation, and tourism. Moreover, toxins produced from some cyanoHABs threaten human and animal health. Climate warming can affect the distribution of cyanoHABs, where rising temperatures facilitate more intense blooms and a greater distribution of cyanoHABs in inland freshwater. Nutrient runoff from adjacent watersheds is also a major driver of cyanoHAB formation. While some of the physicochemical factors behind cyanoHAB dynamics are known, there are still major gaps in our understanding of the conditions that trigger and sustain cyanoHABs over time. In this perspective, we suggest that sufficient data sets, as well as machine learning (ML) and artificial intelligence (AI) tools, are available to build a comprehensive model of cyanoHAB dynamics based on integrated environmental/climate, nutrient/water chemistry, and cyanoHAB microbiome and 'omics data to identify key factors contributing to HAB formation, intensity, and toxicity. By taking a holistic approach to the analysis of all available data, including the rapidly growing number of biological data sets, we can provide the foundational knowledge needed to address the increasing threat of cyanoHABs to the security of our water resources.

Figure 1

(top) HABs are a common occurrence in western Lake Erie (image used with permission from U.S. Geological Survey). (bottom left, right) CyanoHABs can be toxic to wildlife and to humans and pets who use the water for recreation. (Images licensed from Dreamstime (https://www.dreamstime.com/).

Figure 2

Number of publications on toxic cyanobacterial genera^, that exist in the Web of Science database as of April 2023.

Figure 3

CyanoHAB data ecosystem depicting cyanoHAB data sources and sampling approaches. These include grab samples or fixed/buoy telemetry sampling directly from the lake, weather conditions from telemetry stations on land, and satellite remote sensing of HABs. By analysis of the data collectively it will be possible to build a model that will ultimately predict cyanoHAB occurrence, persistence, and toxicity and inform decision-makers.

Figure 4

Suggested HAB ML/AI workflow for the data collection, integration, and development of a HAB ML/AI model to enable the development and understanding of an integrated picture of the cyanoHAB dynamic ecosystem. Multisource HAB data that spans the different meteorological spatial scales and temporal resolutions (as defined in section 4) should be curated and used to train, test, and validate ML models. The model should then be analyzed to enable extraction of key insights, in terms of the relative feature importance quantification and generation of design maps for cyanoHAB formation. The model development and analysis should be conducted iteratively to ensure a robust model. Insights and predictions obtained from such an adaptive framework will then be interpreted by domain experts and decision-makers.