Send Letter to Editor

Wisconsin Natural Resources magazine

Wisconsin Natural Resources magazine

Workers pay out the weighted pipeline that will be sunk to the bottom of Devil's Lake to remove water with high concentrations of phosphorus. © Richard D. Lathrop
Workers pay out the weighted pipeline that will be sunk to the bottom of Devil's Lake to remove water with high concentrations of phosphorus.
© Richard D. Lathrop

June 2004

Restoring Devil's Lake
from the bottom up

A 15-year DNR project will drain phosphorus by withdrawing water from the deepest part of Devil's Lake.

Richard C. Lathrop, Timothy J. Astfalk, John C. Panuska and David W. Marshall

Phosphorus stays and cycles
Algae growth raises concerns
Developing a restoration plan
Built on a tight schedule
So far, so good

As the centerpiece of Wisconsin's most popular state park, Devil's Lake in Sauk County is a real gem – a 372-acre lake set in a spectacular setting among towering quartzite bluffs whose talus boulders tumble steeply to shimmering water. Though the lake is renowned for its naturally clear waters due to low fertility from its small forested watershed, in recent years the lake sometimes looks more emerald than diamond, especially near shore.

The park draws well more than a million visitors each year and it's been a busy place for more than a century. Since the late 1860s, Devil's Lake has been a favored playground. Over the course of 100 years up to four resorts and more than 60 private cottages surrounded the lake's shoreline. Their primitive outhouses and septic systems undoubtedly leached phosphorus into the lake. Additional phosphorus seeped in from a park sewer main that broke in the late 1970s until it was repaired in the early 1980s. And an additional amount of phosphorus entered the lake from farm runoff coming from the southwest bluff. Today, the farms, resorts and all but four of the cottages are gone as the public park expanded from its inception in 1911.

Phosphorus stays and cycles

Unfortunately, a lot of the phosphorus remains in the lake's deep-water sediments. This is because Devil's Lake is a seepage lake with no outlet to naturally dilute and flush out the phosphorus. While few nutrients from the watershed reach the lake today, the phosphorus stored in the lake's bottom sediments recycles each year making the overall fertility level of the lake abnormally high.

Studies by DNR researchers show that the phosphorus temporarily binds to iron compounds in the lake's sediments at times of the year when the overlying water is oxygen-rich. From late spring through early fall as the water warms up, the lake stratifies with a less dense warm water layer near the surface. By mid-summer, the denser colder water on the bottom loses its oxygen (anoxia). The phosphorus is then released from the sediments and builds up in the bottom waters from late summer through about mid-October. At that time the surface waters have cooled causing the lake to "turn over" and allowing the phosphorus to mix throughout the entire lake.

Algae growth raises concerns

The end result of that mixing has longer lasting effects in the lake. By spring and early summer, the phosphorus has begun fueling the growth of different forms of algae: free-floating (phytoplankton), filamentous and attached (periphyton). Since the late 1970s, the free-floating algae in late summer and early fall have become really noticeable to visitors, especially the SCUBA divers who enjoy exploring Devil's Lake's deep waters. Noxious blue-green algae blooms that accumulated on the downwind shorelines alarmed park personnel and visitors alike.

Beginning in the early 1990s, the free-floating phytoplankton blooms became less severe while the two other forms of algae became more noticeable. Filamentous algae formed thick cotton candy-like mats that draped over the bottom and smothered larger plants. Trapped air bubbles caused these mats to rise like green stalagmites toward the water surface. At the same time, the underwater boulders along the east and west shorelines were covered with a green carpet of periphyton algae. Near the beaches on the north and south shores, the bottom sediments and larger aquatic plants were also heavily coated with periphyton. This algae serves as the main food source for 17 species of snails, three of which can be intermediate hosts to a parasite that causes swimmer's itch, a painful nuisance that cuts park attendance and local tourism.

Local citizens as well as park personnel have long sought answers to the vexing swimmer's itch problem in Devil's Lake. This desire precipitated a 1999 DNR research study that found high snail densities in the lake as a result of abundant food (periphyton). The study concluded that with few natural predators such as pumpkinseed sunfish or crayfish in the lake to reduce snail densities, the only ecologically safe way to control swimmer's itch was to drastically reduce the snail populations by reducing their algae food source – effectively starving the snails. While snails can be poisoned with copper treatments, it is a short-lasting cure that isn't safe for fish and other aquatic organisms. It's also too expensive to carry out on the entire lake.

Algae also indirectly elevates mercury levels in the lake's fish populations, ultimately reaching levels of public health concern in large sport fish such as walleye. Inorganic mercury enters the lake from atmospheric pollution. Sulfate-reducing bacteria that thrive only in the oxygen-poor bottom waters in late summer convert the relatively harmless inorganic mercury to the toxic methyl-mercury form, which builds up in those waters until the lake mixes during fall turnover. At that time the methyl-mercury is readily taken up by phytoplankton and then concentrate as the methyl-mercury passes up the food chain to fish.

If researchers could find a way to reduce the buildup of methyl-mercury in the lake's bottom waters, mercury fish contamination problems should decline. While controlling atmospheric sources of mercury would reduce fish contamination problems in lakes both regionally and throughout the world, techniques to decrease the growth of sulfate-reducing bacteria in Devil's Lake should result in less methyl-mercury entering the food chain. One way of slowing that process would delay the onset of anoxia. Decreasing the amount of algae that settle to the bottom waters would slow decomposition by other bacteria that use up the available oxygen. This would give the sulfate-reducing bacteria less time to convert inorganic mercury to methyl-mercury before the bottom waters are re-oxygenated at fall turnover.

We'd expect many benefits from removing the legacy of phosphorus pollution now stored in Devil's Lake's deep-water sediments. Direct benefits would include fewer forms of algae growth and clearer water. Indirect benefits would include fewer snails less chance for swimmer's itch and lower fish mercury levels. In addition, better oxygen conditions in the bottom water would improve cold water habitat needed by brown trout being stocked in the lake.

Developing a restoration plan

A two-year study by DNR researchers from 1986-87 confirmed that algal blooms at Devil's Lake would continue unless high phosphorus concentrations could be curtailed. As in most lakes worldwide, it's easier to prevent the flow of phosphorus into lakes than to correct the situation after lakes are fertile. Techniques tried elsewhere to reduce lake fertility include dredging the nutrient-rich sediments, binding up the phosphorus in the sediments with chemical treatments, aerating the bottom waters to temporarily sustain the sediment's ability to bind phosphorus, or drawing phosphorus-rich waters from the bottom of the lake. The first three options were dismissed as unsuitable for Devil's Lake due to high cost (particularly dredging), short-term effectiveness (chemical treatment and especially aeration) and strong opposition to adding chemicals like alum or other aluminum compounds to a lake classified by the DNR as an Outstanding Resource Water.

Drawing off bottom waters (technically called "hypolimnetic withdrawal") has been used on a few moderately deep lakes and reservoirs in Europe and less often in North America. After the lakes stratify, the cold, phosphorus-rich bottom waters are removed in late summer and discharged to a receiving stream. Bottom withdrawal hadn't been tried on a large seepage lake with no outlet like Devil's Lake. However, we felt the technique was feasible because the withdrawn water could be replaced by diverting clean runoff water from a nearby intermittent stream called Babbling Brook.

In fact, an old diversion ditch had been dug in the 1890s between Babbling Brook and Devil's Lake to allow water to be periodically added to the lake when lake levels were low. In the 1960s, the open ditch (known at that time as Inman's Canal according to Ken Lange, retired park naturalist) was replaced by a 30-inch culvert and covered over. Since the early 1970s the diversion system has not been used due to generally higher lake levels. In recent years, a section of the culvert pipe nearest Devil's Lake rusted away and collapsed making the system inoperative. However, we felt the diversion system could be refurbished as part of the project.

DNR researchers, managers and park personnel discussed the potential bottom withdrawal project in the fall of 2000 at the annual meeting of the Friends of Devil's Lake State Park. This nonprofit group continues to provide active support by volunteering time and fundraising for improvements at the park. They agreed to help sponsor the project and helped obtain a small grant to develop cost estimates for potential withdrawal designs. All designs called to install a long pipe to the deepest spot in the lake to withdraw phosphorus-rich water, which would be discharged to Babbling Brook that flows to the Baraboo River. Water could either be pumped out of the lake or could be withdrawn by a passive siphon system (similar to the siphon system used to clean a home aquarium). Researchers estimated the system would need to run several weeks a year during September and early October for perhaps 15 years to reduce phosphorus levels in the lake. The siphon design was chosen for Devil's Lake because it would require no maintenance and no electricity to run it – a huge cost savings on such a long-term restoration project.

The spring of 2001 was spent doing an environmental analysis of the project and incorporating public comments. The project called for discharging water to Babbling Brook in September and early October when the intermittent stream is normally dry, so consequences for the stream's aquatic life were judged minimal. Downstream effects on the Baraboo River were also determined to be minimal. Though the discharged water would be high in phosphorus, it represented less than one-tenth of one percent of the annual phosphorus load in the river. In addition, the water from Devil's Lake would also be flushing through the Baraboo River system after the summer growing season and a long time before the next growing season.

As regulatory hurdles were being cleared, work focused on finding grant money to pay for the siphon system, estimated to cost about $300,000 if DNR staff did the land surveys, design and some of the other preparations. The project proposal submitted by the Friends of Devil's Lake garnered a $200,000 State Lake Protection Grant, EPA's Clean Lakes Grant provided $100,000, an additional $5,000 came from a Friends of Wisconsin State Parks grant that was matched by the local Friends group to provide $310,000. All the financial pieces of the puzzle came together by mid-February 2002.

Built on a tight schedule

Land surveys, contracts and the hunt for supplies began soon after to meet the project of installing and operating the siphon system by early September 2002.

The system consists of 5,500 feet of 20-inch diameter plastic pipe with 4,150 feet resting on the lake bottom. Fifty-foot pipe sections were fused together and 320-pound concrete weights attached every 12 feet as "the big straw" was pushed out from shore to the lake's deepest spot – about 47 feet depending on lake level. The weights were needed to keep the pipe on the bottom once it was filled with water and sunk in place. The intake on the far end of the pipe is a 50-foot section drilled with holes along two sides to draw water in from about eight inches above the lake's bottom sediments.

On land the pipe was trenched into the ground with a manhole at the pipe's highest point and another where the siphon ended just below the main flow valve. The difference in water levels between the terminal manhole and the lake surface creates a pressure head difference (five to nine feet depending on lake level) that determines the flow rate of the siphon. A flow meter is located at the high point manhole. A vacuum pump is connected at the same location to prime the siphon by filling it with lake water.

A private contractor (Heartland Construction, Inc.) began constructing the pipe system in July 2002 as a small work crew started fusing pipe into 450-foot segments. To accommodate the park's summer influx of visitors and vacationers and to free up as many parking spaces as possible, the work crew had to clear the work area each Friday afternoon. Once the pipe segments were finished, a bigger work crew began attaching the concrete weights as the pipe was pushed into the lake. Even with weights attached, the pipe floated and it was all a barge could do to maintain tension on the pipe. Fusing together the in-lake portions took a week and a half ending July 30th.

Early on the morning of July 31, the contractor began trenching the nearshore lakebed inside an area that had been encircled by a silt curtain to contain any sediments stirred up in the process. The intake pipe section (with extra flotation attached and the holes sealed) was then pulled to the middle of the lake and connected via a flange joint to the main pipe. While the pipe was stretched taut by a barge, fire trucks from the Baraboo Fire Department and DNR forestry programs began filling the pipe from shore with lake water as a gallery of spectators watched. Since no one had experience sinking a 4,150-foot pipe weighted by 55 tons of concrete, project personnel and the contracting crew wrestled with some unanticipated problems. A few air pockets formed that traveled like Loch Nessie along the pipe's length until the monster "burped" at the end. There were other exciting moments as the pipe developed a slight twist as it sank. The holes on each side of the pipe intake section were realigned horizontally by rotating the intake at the flange joint just before the pipe intake sank. By late afternoon the pipe lay on the bottom with the intake holes perfectly positioned eight inches above the lake sediments at the lake's deepest spot.

By mid-August the 1,350-foot land section of pipe was trenched and joined to the lake portion. On August 27th, the flow meter was installed and the siphon was primed. On August 29th, we opened the main valve and water started pouring out. The only noticeable sound was gurgling water in the stream plus a slight sulfur smell next to the siphon discharge point. Average flow rates were 5.3 cubic feet a second, a brisk 2,380 gallons a minute during the seven-week run until we shut down for the season when cooler weather naturally turned over the lake water on October 17th. By then we had removed 981 pounds of phosphorus from the lake, far exceeding our initial goal of about 350-400 pounds. Since the lake levels were really high in 2002 and the phosphorus concentrations were higher than expected, we had decided to let the system run a little longer than planned for subsequent years.

So far, so good

In 2003 the withdrawal system was used less extensively following a prolonged drought, but we still removed 377 pounds of phosphorus. One of our goals was to draw down the lake a bit by October so the contractor could re-bury the lake pipe near shore so it wouldn't be exposed on the shoreline or be an underwater obstruction during normal lake levels. The water levels were so high when the pipe was initially installed in 2002 that the contractor couldn't trench in as far offshore as originally planned. Repairs in 2003 also included replacing the small broken section of the water diversion pipe near the lake. By early November clear runoff water from Babbling Brook was diverted into Devil's Lake to begin replacing the water withdrawn in September and early October. Even greater amounts have been flowing into the lake from rainfall and snowmelt runoff during the late winter and spring months this year.

Now that the bottom withdrawal siphon system is installed, the work of restoring Devil's Lake is just beginning. DNR staff expect to operate the system each September and early October for 15 years or so. During fall through spring months, withdrawn water will be replaced each year by diverting water from Babbling Brook. The lake will be monitored for signs of improvement. With any luck the lake will experience better water clarity, fewer phytoplankton blooms, and fewer other forms of noxious algae growth, less frequent bouts of swimmer's itch, and lower fish mercury levels. Our goal is to return Devil's Lake to its original pristine state so that future generations will be able to enjoy this truly outstanding resource gem.

Richard C. Lathrop is a PhD research limnologist for the DNR and manager of this Devil's Lake restoration project.

Civil Engineer Timothy J. Astfalk of Mead and Hunt, Inc., DNR Engineer John C. Panuska and DNR Water Resources Specialist David W. Marshall assisted on the project.