Predicting Blue-green Algae Blooms

Satellite image showing mild bloom on Lake Erie in 2004. Right – Satellite image showing 2011 blue-green algae bloom, the worst in decades. Both images courtesy of ESA/MERIS; Processed by NOAA/NCCOS.

Satellite image showing mild bloom on Lake Erie in 2004. Right – Satellite image showing 2011 blue-green algae bloom, the worst in decades. Both images courtesy of ESA/MERIS; Processed by NOAA/NCCOS.

Wouldn’t it be nice to know at the beginning of the summer how bad blue-green algae blooms were going to be on Lake Champlain? Well, researchers from the National Oceanic and Atmospheric Administration’s National Center for Coastal Sciences (NCCOS) believe they have developed a model that can make such predictions for Lake Erie. In 2012 they issued their first seasonal harmful algae bloom forecast, predicting mild blooms during the summer.

Lake Erie is famous in the annals of water quality for its deplorable condition during the 1970s. The Cuyahoga River in Cleveland caught on fire in 1969. This event spurred Time Magazine to declare, “Lake Erie is in danger of dying by suffocation.”

However, the lake rebounded. Significant improvements occurred following the banning of phosphates in laundry detergent and the passage of the Clean Water Act in 1972 which forced factories and industrial polluters to clean up their acts. The amount of phosphorus coming into Lake Erie was halved between the 1970s and the 1980s.

By the mid 1990s conditions appeared to worsen once again. A blue-green algae bloom in 2011 far exceeded anything that had been seen in recent years. On August 7th of that year the Toledo Blade said, “A huge bloom of potentially toxic microcystis algae, which has reared its ugly head almost annually since 1995 after more than a 20-year absence, has been visible from space since at least July 22.” The 2011 bloom covered 1,600 square miles at its maximum, an area the size of Long Island Sound.

The predictions of a mild bloom year in 2012 have for the most part panned out. The forecast called for a bloom one tenth the size of the 2011 bloom. As of August 16  no confirmed blooms had appeared. The modelers’ August 20 bulletin reported a relatively small bloom in the vicinity of Monroe Michigan.  

The model predicts summer bloom intensity conditions based on spring phosphorus loading from the Maumee River. When blue-green algae blooms develop on Lake Erie, they usually occur in the shallow western basin at the mouth of the Maumee River, which enters Lake Erie at Toledo. Like blooms on Lake Champlain, they are fueled by phosphorus. While total phosphorus loading from four to six months prior to summer adequately predicted bloom intensity, an even better predictor was a subset of total phosphorus, the soluble portion. They based their prediction on soluble phosphorus loading in the four to eight weeks prior to summer.

To validate the model, researchers used a 10-year data set, which included nutrients flowing into Lake Erie and the volume of harmful algae. It coupled this data set with infrared satellite imagery showing the extent of blooms for the time period modeled.

The model is useful only for predicting blooms of one particular type of algae, microcystis, the same algae that plagues Missisquoi Bay. Early season blooms of other species, such as those that have sometimes been seen in the Main Lake in late June and early July, are not captured by the forecasts.

The researchers have also decided not to update their prediction through the summer. They reason that by mid-June, vegetation has sufficiently established in fields to prevent excess runoff of soluble phosphorus. To bolster their case they point out that a wet July in 2007 did not alter bloom conditions.

The forecast can be helpful for municipal officials. The 2011 bloom affected drinking water intakes for over 2.8 million people. In many cases additional water treatment was required. Toledo spent $3,000 to $4,000 per day on carbon-activated filtration during blooms. Certainly having an advanced sense of how extensive a bloom would be helps with budgeting.

One interesting implication of the project is the focus on soluble phosphorus rather than total phosphorus as a predictor of blooms. Most of our clean-up efforts to date on Lake Champlain have not distinguished between the two forms. If indeed blue-green algae are keying in on soluble phosphorus during the spring runoff, it can lead us to more concentrated remediation efforts. For instance, tile drains have been used to increase production on many farm fields. They may decrease total phosphorus pollution, but increase dissolved phosphorus loading.

As researchers refine the model they hope to apply it to other water bodies. The model will require extensive tweaks over time. One limiting factor may be the availability of appropriate satellite data here. Yet one day we may have pre-summer predictions of how intense blue-green algae blooms will be on Missisquoi Bay. Moreover, if the model works based on just dissolved phosphorus, it will also help us better target our water quality improvement efforts.