By Andrew Erickson, PE

A 2005 study of nationwide monitoring data — “Evaluation of NPDES Phase 1 Municipal Stormwater Monitoring Data” by the University of Alabama and the Center for Watershed Protection — reports the median values of total and dissolved phosphorus as 0.27 and 0.12 milligrams/liter, respectively. With the mean fraction of dissolved phosphorus to total phosphorus at approximately 44%, lowering dissolved levels in stormwater runoff requires capturing both the particulate and dissolved fractions of total phosphorus.

Most treatment practices capture particulates through settling or filtration, but very few have a mechanism to consistently capture the chemical when dissolved. Because dissolved phosphorus is more accessible than particulate phosphorus to organisms, removing just the chemical's particulate form won't meaningfully reduce the production of algae or eutrophication. But introducing a chemical adsorption or precipitation process to the treatment train will.

In particular, adding steel wool or elemental iron to a sand filter captures a significant amount. As the elemental iron forms iron oxides — otherwise known as rust — dissolved phosphorus binds to the oxides by surface adsorption. This concept has been developed into the “Minnesota Filter.” Three applications are installed in Minnesota cities, and two more are under development at the University of Minnesota's St. Anthony Falls Laboratory.

Because the additional material is relatively inexpensive, installing the design costs little more than any filtration practice. Several examples described here are extensions of existing treatment practices, where the iron filings represent about 5% of the total cost of treatment.

A surface version of the process in Maplewood, Minn., uses approximately 5% iron filings by weight (see image on page 31). The basin area is approximately 0.27 acres and stores approximately 0.65 acre-feet of stormwater before overflow occurs. The contributing watershed is approximately 8 acres and is composed of 81% impervious surface with mostly hydrologic soil group B soils.

The average annual volume captured and treated by the basin is estimated to be approximately 5 acre-feet; therefore, the filter's annual hydraulic loading rate (total volume per surface area) is approximately 18.4 feet.

The site began treating stormwater in June 2009. Monitoring for total phosphorus, soluble reactive phosphorus, and iron is conducted during normal operating periods of April through October. Preliminary results indicate the filter is reducing the dissolved phosphorus concentration to or below the detection limit of 0.01 milligrams/liter.

The City of Prior Lake, Minn., installed four Minnesota Filter trenches along the perimeter of two wet detention basins in January and February 2010.

The trenches are designed to be below the normal water level created by the basins' outlet structure. As illustrated in the schematic on page 33, when rain flowing into the basin raises levels high enough, excess water spills over the trench surface and into the media. The overflow moves through the mix of iron and sand to a perforated pipe underdrain where it's captured and conveyed to the basin's outlet structure.

All trenches are connected to the outlet through individual underdrains that drain only the stormwater that passes through that trench. For small events, all of the rain is filtered by the trenches to capture dissolved phosphorus. During large storms — those that exceed the filter's volume — water that overflows the weir in the outlet structure bypasses the trenches. When the water level drops below the weir, the remaining stormwater is filtered by the trenches to capture dissolved phosphorus.

The filter is designed to be aerobic. This is achieved by placing the underdrain outlet above the downstream water level and separating filter material from nearby soils with an impermeable liner. In this way air can reach the filter material from the top and bottom when the filter isn't operating, allowing the material to dry between rain events.

Measurements taken between July and October 2010 showed the trenches captured 80% to 90% of incoming dissolved phosphorus. This is the capture rate during events that don't overflow the design, where the trenches treat all the rain. When overflow occurs, the load reduction is expected to be less depending on the volume of runoff the trench doesn't treat.

To truly clean up the nation's stormwater, we need to begin treating for dissolved phosphorus. Adsorption is an effective unit process to perform this task, and iron is an element that can adsorb dissolved phosphorus without toxicity concerns. The Minnesota Filter combines these two factors to cost-effectively remove dissolved phosphorus.

— Erickson ([email protected]) is a research fellow at the University of Minnesota's St. Anthony Falls Laboratory and Gulliver ([email protected]) is a civil engineering professor at the university; Weiss ([email protected]) is an associate professor of civil engineering at Indiana's Valparaiso University.