Learning from the past

In 2018, Detroit Lake experienced numerous bloom events, and at several points the lake water contained algal toxins. Analyzing the historical data, we can see that there is a well defined HAB season:

This historical data can give us a general sense for when HABs might occur. But specific predictions are required for effective risk management. In the figure below, we show lake harmful algal concentrations (green dots) and predictions from our model (red lines: dashed lines represent the upper and lower bounds on our predictions, and the solid red line is our estimate of the expected algal concentration). As you can see, our predictions track the observations quite well, meaning that in 2018, one week in advance we could have provided advanced warning of algal blooms. Indeed, our 2018 analysis predicted algal blooms with an accuracy of over 85%.