Algal bloom forecasters blew it

SANDUSKY — Scientists are taking another look at their forecasting models after this year's prediction for Lake Erie's harmful algal bloom turned out to be substantially off. Experts issuing the annual forecast on June 30 predicted a severity of 3 on a scale of 10 with a possible range between 2.5 and 4. That would have been a mild bloom similar to the smaller bloom recorded in 2020. Instead, the bloom had a severity of 6 and covered 530 square miles of western Lake Erie at its peak. That's a "moderately severe" bloom — not as bad as 2019, 2017 or 2015, the worst on record in recent years, but significantly worse than 2020, 2018 and 2016. The scientists who issued the final email bulletin of 2021 on the Lake Erie harmful algal bloom, a publication of the National Oceanic and Atmospheric Administration, say they're trying to figure out where they went wrong. "We will continue our assessment of the models in order to make improvements to the forecast and severity metrics," the bulletin says. Dr. Richard Stumpf, an oceanographer with the National Oceanic and Atmospheric Administration and one of the scientists who issued the forecast, said it's not unusual to review their models. "Each year, we review and ask whether we were right or not, and what may be a factor," he wrote in an email to the Register. One factor that might have thrown off the forecast was that Ohio had heavy July rains arriving after the June 30 forecast, washing large amounts of phosphorus into the lake and feeding the harmful algal bloom. "The large bioavailable phosphorus load in July 2021, the third-highest since 2008, may have led to the larger than forecasted bloom. Previous years, 2015 and 2017, with large July loads also had wet springs, leading to a large input of phosphorus in the spring, reducing the relative influence of the July load. In contrast, 2021 had a drier spring than average, suggesting the relative importance of July loads on the bloom," the NOAA report says. Stumpf said he was surprised by the gap between the forecast and what actually took place. "The small amount of 'scum,' which we can see from satellite, and the relatively slow start, had led me to think we might be only slightly low," he remarked. The scientists said although the bloom covered a wide area, peaking at 530 square miles, it also seemed less thick and less toxic than in some other years. "This large area may have resulted from the heavy rain and associated high river discharge in July which may have dispersed nutrients further into the western basin," the report said. "The concentration (biomass) of algae within the bloom was moderate, resulting in less extensive scum than has been seen in previous years. Toxins (microcystins) were detected throughout the bloom, although at lower concentrations than some recent years," it said. For about 10 years now, the Lake Erie harmful algal bloom has followed an interesting pattern: The bloom is large in odd-numbered years and small in even-numbered years. The lone exception was 2014 when the bloom was larger than a 6, but, even then, it was smaller than in 2013 and 2015. Stumpf said he's stumped so far on whether that's a significant pattern. "Believe me, I have already noticed it, and I thought this year would break it," he told the Register. "My take is we are in one of those 'odd' patterns that happen, however, if this goes on another two years, I'll start some conversations with some of the regional climate experts." The odd-even variation does point up something that's important, Stumpf said. "It shows that the size of the bloom depends on the load and events of that year, not on the loads from the previous year or years," he said. "This argues strongly that if we reduce the phosphorus load, we will have an impact on the amount of bloom. 2012, 2016, 2018, 2020 were all relatively low load (2012 was a drought). Blooms were small even though following a big year," he said.