The Odana Infiltration Project
A complicated way of doing a simple thing
The “Rube Goldberg” machine is an impossibly complicated device that does a simple task. Rube himself was an engineer, who celebrated the spirit of invention with his cartoons. There was always some implausible step that guaranteed failure in the real world.
Cartoon by Rube Goldberg. He was an engineer.
In Madison, we are honored to have one of the largest Rube Goldberg machines in the state, hidden away in an out-of-the way corner of Odana Golf Course. But this complex machine sends its strings, pipes, wells, and band aids all over the City and beyond.
While this machine is indeed complex--it sometimes works. And it was conceived with the best intentions--to safeguard the environment and the public water supply.
Getting down to the nuts and bolts, strings and birds
In 1990s, as the University expanded to the west, there arose a need for centralized heating and cooling for the new buildings. Because this would be more efficient if combined with electrical generation, plans for a new power plant were laid.
Although this expansion had long been planned, suddenly the power plant was put on fast track. Built by MG&E on the west side of the UW Campus, it is now known as the Cogen facility, because it combines power production with heating and cooling.
Although this expansion had long been planned, suddenly the power plant was put on fast track. Built by MG&E on the west side of the UW Campus, it is now known as the Cogen facility, because it combines power production with heating and cooling.
The Cogen plant on the west side of the UW Campus.
The Cogen plant withdraws one million gallons of water a day from Lake Mendota. Most of it evaporates from the cooling towers. You can see this during cold days in winter, as huge clouds of vapor billow above the plant. All this water into the air is lost from our watershed. Additional water, used to clean filters and purify the lakewater, goes via the sewer to the Nine Springs sewage plant, where it also leaves our watershed.
Where should the water come from? There was a lot of opposition to the plant in the neighborhood. It was decided to withdraw the water from Lake Mendota, as the University had been doing for years for the Charter Street and the Walnut Street plants.
Saving the Yahara
Experts argued that withdrawing water wouldn’t harm Lake Mendota--indeed, the lake had experienced several years of high levels. Only one major consequence of using lakewater was identified. Cogen could cause a problem for the Yahara River below Lake Waubesa, where the river can get very low during drought conditions, threatening fish and ecosystems there.
So it was decided that if the Yahara was ever threatened by low flows, resulting from drought plus water taken for the Cogen plant--then water would be pumped from the ground near the Yahara and used to augment its flow.
Sounds OK so far?
Sounds OK so far?
Fortunately, pumping water from the ground into the Yahara would need to occur only when the Yahara got really low. That hasn’t happened since the Cogen plant was built. But the groundwater is like a reservoir, and pumping for Madison’s water supply has drawn down the reservoir. Indeed, towards the center of Madison, the level of the water table has already decreased by 60 feet.
So, to compensate for any pumping into the Yahara that might be required, a “mitigation” deal was worked out between MG&E and DNR. As a result, MG&E is required to replenish the groundwater to the tune of 50 million gallons a year. That replenishment is supposed to increase to 80.4 mgy by 2012, when the Cogen plant reaches full capacity.
This requirement was put in a Water Withdrawal Permit that allowed the plant to be built and operated. This was the first of two essential permits for plant operation.
This requirement was put in a Water Withdrawal Permit that allowed the plant to be built and operated. This was the first of two essential permits for plant operation.
Are you still with me? That’s just the beginning.
The University and MG&E built the power plant together. All of us use the groundwater reservoir for drinking and cooling and industry. The Madison Metropolitan Sewerage district processes the wastewater and sends it out of the watershed, mostly into Badfish Creek. Madison's Water Utililty has an interest in keeping groundwater pure and plentiful. So many stakeholders are involved--that’s partly why this project grew to Rube Goldberg proportions.
Saving the groundwater
So, how to replenish the groundwater? Acting for MG&E, Montgomery Associates looked for a solution. They evaluated a number of alternative ways for getting water back into the ground:
- A system to inject treated sewage water into the ground
- A system to inject stormwater into the ground
- Settling ponds
- Rain gardens
MG&E was in a hurry, even though the need for a West Campus plant had been known for many years. Some say the hurry was to prevent opposition to the plant in nearby University Heights from gathering force.
Because the project was on fast track, they couldn’t dilly-dally with unproven solutions like rain gardens, which take a lot of coordination with the public. There wasn’t time to seriously consider treated water from the sewage plant. The amount of land needed for large settling basins wasn‘t available.
Because the project was on fast track, they couldn’t dilly-dally with unproven solutions like rain gardens, which take a lot of coordination with the public. There wasn’t time to seriously consider treated water from the sewage plant. The amount of land needed for large settling basins wasn‘t available.
That left the alternative of a mechanical system to inject stormwater into the subsoil. Now they had to find a location. The groundwater is a big reservoir--so it didn’t matter where the stormwater was injected--they just had to find a large supply of storm water and some unused land--with sandy soil--where the water could be injected.
The Odana Hills Golf Course turned out to be a good place--with open land and sandy soil. Nearby was Odana Pond, which receives a lot of stormwater from the Westgate Shopping center, plus neighborhoods south of the beltline.
Link to satellite photo of recharge field here.
The recharge area is the brown triangle at the center.
Odana Pond looking SW from a kite--Photo by Craig Wilson.
It was decided they could withdraw no more than a third of the water flowing through the pond, to keep it from drying out and harming habitat.
The Friends of Lake Wingra become involved
In the Odana area, groundwater flows from west to east, from the golf course towards Lake Wingra. Over the years, springs around Lake Wingra have been drying up, due to depletion of the groundwater reservoir. Storm water injected into the ground at Odana would, within a few decades, make its way towards the springs around Lake Wingra.
Before settlement, the low areas of the golf course, near Odana Pond, were a wetland. The pond and wetland didn’t have a surface outlet. Water simply trickled through the sandy soil towards Lake Wingra, as at Indian Lake today. It was like a huge rain garden.
When the golf course was built, they drained marshy soil by building a stormwater pipe leading towards Lake Wingra. Any dirty storm water dumping into Odana Pond now had a direct route to Lake Wingra. The load of phosphorus and sediment to the lake now increased.
When the golf course was built, they drained marshy soil by building a stormwater pipe leading towards Lake Wingra. Any dirty storm water dumping into Odana Pond now had a direct route to Lake Wingra. The load of phosphorus and sediment to the lake now increased.
But the MG&E plan called for taking water from Odana Pond and injecting it into the sand. Stormwater would would now flow through the ground, filtered all the way--dirty storm water converted to crystalline spring water.* This could help clean up Lake Wingra.
The Friends of Lake Wingra naturally became involved in the issue. The proposal to inject stormwater at Odana divided board members of the group. Most on the board believed that replenishing the groundwater was a good idea--that the lake would benefit. But others felt there was a potential for contamination of the groundwater.
Before they would endorse the project, the Friends negotiated with MG&E. They understood that water would not be injected in early spring, when water in Odana Pond was likely to be most contaminated with salt from roads. Moreover, the quality of groundwater around the injection site would be monitored. Most of the Friends board accepted these safeguards--but one board member could not agree that protections were adequate, and resigned in protest.
The Friends group had a number of experts in water resources on the board and among their advisers. Their approval helped in getting the whole plan to inject water approved. Now construction of the Cogen plant could proceed.
So a second key permit was written by DNR for discharging stormwater into the ground. This permit states that certain levels of contaminants like salt in the test wells around the site cannot be exceeded.
Saving the infiltration field
Saving the infiltration field
In 2005, construction began with a large square excavation in the Golf Course, not far from the beltline. The pipes would have to be surrounded by something that admits water, and also keeps the soil from closing off the holes in the injection pipes. Usually, a layer of washed gravel is used.
But MG&E was going to be pumping a lot of water, and they weren’t positive how much water the soil would accept. They were worried that even washed gravel would still have some residual rock dust that would clog up pores in the soil. So they decided instead to use plastic pellets, like the ones used in shipping boxes. The manufacturer did tests, to make sure the plastic pellets wouldn’t be squashed by the weight of the soil.
After the pipes were laid and the soil backfilled--the system didn’t work. They couldn‘t pump enough water! So they excavated the whole thing again, and found the plastic pellets squashed flat. After a lawsuit and the substitution of gravel for plastic, the system now worked. Almost.
Cleaning the filters
Cleaning the filters
The water coming to Odana Pond--like most storm water--is dirty.
Main tributary to Odana Pond emerges from under Beltline. 4/1/11.
Closeup of the water in the top photo.
The same water, further downstream, near Odana project inlet.
Nutrients in the runoff stimulates toxic algae growth. Because the sediment and algae in water from the pond could clog pores in the soil below the infiltration field, the water has to be filtered before injection.
So filters remove the sediment and algae--but they can’t filter out dissolved contaminants like salt.
The filters have to be cleaned periodically. They are flushed with water from the pond (or tapwater?)--and the resulting sludge is sent through sanitary sewers to the Nine Springs plant, and out of the watershed.
This filter cleaning uses 11% of the water withdrawn from the pond, so the sewage plant has to process about 5 mgy from the Odana facility.
Pumping into the ground is limited by two things…
- The amount of water the soil will accept.
- The amount of water that can be withdrawn from the pond without drawing it lower than a certain level.
Given these, MG&E is reluctant to accept further limitations on pumping, such as a halt when salt levels are high in springtime.
80 million becomes 60 million
80 million becomes 60 million
With 80.4 mgy as the permit requirement for groundwater recharge when Cogen reached full capacity, the project design aimed for that. But reality intervened. There were questions at Odana about the area needed and numbers of trees that would be cut. So the design objective was whittled down to 50 or 60 mgy.
For whatever reason, the project is having trouble pumping enough water into the ground. You can see the decline in the graph of daily volume below.
Pumping is trending downward. Soil pores are clogged?
Pumping cannot occur if the Pond falls below a certain level--to protect pond life.
The first permit requires they eventually reach 80.4 mgy, but they are only pumping about 60 mgy now. And that’s with pumping all the time. Those last three words are important…
Salt contamination in the groundwater
Early this winter, Jim Lorman was burning the night oil, checking some data from Lake Wingra. He’s on the faculty at Edgewood College, and a founding member of the Friends of Lake Wingra.
On the internet, he looked at data from the test wells, located around the Odana injection field. He saw, as expected, that salt levels peak in spring, when runoff carries road salt from the beltline and Westgate parking lots.
Chloride levels peak in winter/spring, when road salt runs off.
To avoid contaminating groundwater, FOLW wants pumping stopped when salt levels peak.
Regulatory limits are shown as horizontal lines--blue is the Enforcement Standard (250 mg/l), and black is the Preventative Action Limit (125 mg/l). Salt (blue) in the recharge field peaks high above the upper limit in the spring of three years. Lower levels of salt, still sometimes exceeding the PAL, are found at test wells (purple & green) beyond the infiltration zone.**
But more disturbing, he noticed that salt levels (measured as chloride) at two test wells, at some distance from the injection site, had exceeded the DNR's Enforcement Standard of 250 milligrams per liter (mg/l). In fact, one well had exceeded the Preventative Action Limit from the very beginning, in May of 2006.
Wisconsin’s Groundwater Quality regulations identify two important levels for each pollutant:
Wisconsin’s Groundwater Quality regulations identify two important levels for each pollutant:
- The Preventative Action Level (PAL). When test wells show groundwater reaches this level, DNR must be notified, and corrective action is supposed to be taken.
- The Enforcement Standard (ES). This is the level not to be exceeded. More.
For periods lasting months during the last three years, levels for salt had exceeded not only the PAL, but the ES level at which the groundwater is considered polluted. Once in the groundwater, the salt will be there for decades. Salt has widespread and harmful effects on ecosystems. For humans, low levels of salt are a problem for people on low-salt diets, and there is evidence that elevated salt in the diet leads to hypertension in certain sensitive individuals--although there is still some uncertainty. Elevated salt can also affect the taste of tap water.
There's more at stake here than whether salt is harmful to humans. For one, the salt is an indicator--a sort of canary in the mine. If salt is increasing, then other, harder-to-detect substances are increasing also. And, the aquifer is the last great free resource we received from the Winnebago. If DNR regulators allow it to become degraded even by salt--that sets a precedent.
There's more at stake here than whether salt is harmful to humans. For one, the salt is an indicator--a sort of canary in the mine. If salt is increasing, then other, harder-to-detect substances are increasing also. And, the aquifer is the last great free resource we received from the Winnebago. If DNR regulators allow it to become degraded even by salt--that sets a precedent.
Once groundwater is contaminated, it means all the springs and streams are saltier to begin with, making aquatic life more vulnerable to the “flush” of road salt that comes at first thaw.
For three years, the groundwater near the injection field had reached contaminated levels in the spring, and neither MG&E, nor the permit guardians at DNR, had said anything to Jim. Nitrates and arsenic were also briefly spiking above PAL levels, though less consistently.
Fixing the problem
Jim got the ball rolling towards a solution--the Friends met with MG&E, and then with the DNR regulators. Jim said to the DNR that MG&E was apparently in violation of both permits. Not enough water is being pumped, and it’s too contaminated with salt.
Jim feels MG&E violated things he had been told during the earlier discussions, that they wouldn’t pump in the spring, when chloride levels are high.
MG&E responds that they have to keep pumping all the time--otherwise, they can’t meet their volume targets. In addition, if they stop pumping, the infiltration bed seems to clog up, so that when they resume pumping, for a while it won’t accept as much water.
DNR has the power to require MG&E to meet their permit obligations, but they can’t dictate exactly how MG&E meets the obligations. Early in 2011, MG&E sent a draft plan (dated Jan. 21) to DNR, for how they would deal with the problem of contamination of groundwater:
- Monitor salt levels of water in the intake pipe.
- Stop pumping whenever the salt levels reach the protective action limit of 125 mg/l.
- But, pump for a minimum of an hour each day (regardless of salt levels, and even if over the PAL). This pumping of an hour every day, when pond water is salty, is presumably to help keep the pores in the soil open.
The board of Friends of Lake Wingra hasn't yet formulated an overall response to the Jan. 21 plan, except to ask for more information.
However, this writer believes the plan is inadequate:
However, this writer believes the plan is inadequate:
- It omits important details, such as why MG&E needs to pump water (regardless of salt levels) for a minimum of an hour a day.
- Once in the groundwater, the salt will be there for decades. There’s no way to remove it.
- The PAL is the level at which you take preventative action, not the OK level to pump--especially when you are pumping extremely large volumes.
- Pumping for an hour a day, no mater how contaminated, is simply a violation of regulations
- The plan does nothing to solve the root cause of the problem, which is spreading too much road salt.
MG&E has not explained why their offer to investigate a salt-reduction program was not included in the draft plan. When the Friends met with MG&E in January, officials agreed to look into a campaign to reduce salt use in the basin of Odana Pond--at places like Westgate mall.
Everyone agrees--the Friends, MG&E, and DNR--that overuse of salt on the West Side is the root cause of the problem. West Towne photo, 3/14/11.
Onwards… towards 80 million gallons a year
And--looming in the background--is the question of pumping enough water to meet the second permit--the 80.4 mgy. To reach that, MG&E will have to revisit all alternative methods to get water into the ground, including a large rain garden program in the neighborhood.
In thinking about how to get more water into the ground, do we really want more of Rube Goldberg’s approach? The more complicated a plan, the more likely it is to fail. Perhaps, in comparison, the rain garden alternative isn’t so complicated.
The first infiltration facility at Odana cost about $2 million for equipment and construction. Yearly operation costs some tens of thousands of dollars each year. Maintenance includes electric pumping costs, which are substantial and likely to rise, and includes the sewer bill (paid by who?).
A waste of energy and water
A waste of energy and water
The Odana project's basic goal is water conservation. It's a deal to prevent harmful effects to the lower Yahara River by swapping water. But the project does nothing to replace the 76 mgy*** lost from the basin as vapor. The project accepts the idea that there will be less water overall in the Yahara. It just tries to prevent very low flows by "borrowing" water from the ground during droughts. That's OK.
But the scheme uses a lot of electricity at Odana, where pumps are running most of the year to send water uphill to the infiltration zone. And if it's necessary to replenish the Yahara during a drought, the high-capacity pumps will be running there. Generating the electricity for those pumps requires more water withdrawals from Lake Mendota, and in turn, more pumping at Odana. And so on. Add to that the 11% (about 5 mgy) of water just required to clean the filters at Odana. That's water from Odana pond (or the City?) that doesn't go into the ground, but instead goes out of the watershed.
And if the goal is water conservation, why are we pumping polluted water into the ground? We've made a Faustian bargain to temporarily improve water quality in Lake Wingra, in exchange for polluting the groundwater (and the long-term health of the lake).
There are three tools available to fix the problem--to use singly or in combination:
- Create programs to reduce use of salt in the Odana basin.
- Create new infiltration areas or projects (including rain gardens) that infiltrate water from neighborhoods with less salt.
- Pump water at the Odana project only when salt levels in the pond are below the PAL.
MG&E, and the people at DNR who negotiated the "swap agreement" deserve a lot of credit. They were exploring new ways to allow development in Madison, and at the same time protect the environment. The Odana facility looked good on paper, but in the real world, it's not performing as hoped. The time has come for hard choices. Some of the remedial actions may still allow the Odana facility to continue operating.
Simply finding a "technicality" that will allow continued pumping of salty water isn't an honest or real solution. Delay isn't a solution either.
Future steps
Future steps
- DNR and MG&E discuss the plan. If DNR finds it provisionally acceptable, it’s sent to DNR permit writer, who begin the process of drafting the permit. This has already happened.
- The Friends present their position to the DNR.
- Public hearings follow. According to emails from DNR, they expect the new permit to be ready in the first quarter of 2012 (they cite a back load of permits and overload of staff)
- If there is sufficient opposition or evidence against the plan, it will be revised.
- The new permit will last for 5 years, but...
- If test wells continue to spike over the PAL (at the test wells), MG&E will have to take corrective action before 5 years are up.
If the Odana Project looks rather complicated… that’s because it is. What could be more simple than just letting rainwater from low-salt neighborhoods sink into the ground?
It's ironic. To do the job of the natural Odana Basin, essentially a giant rain garden--Madison Gas and Electric built a costly, complicated machine that wastes energy and water, and will degrade our children's drinking water.
They are going to have to build more infiltration--to reach their required goals. It's time to get back to the rain garden idea.
# # #
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
** A third test well, located next to the beltline and upstream (as groundwater flows) from the other two test wells and injection area, shows elevated levels of salt. This is to be expected, as the beltline is a major source of salt. This well shows the groundwater in this area was perhaps already partly contaminated. But this is not a good argument for the status quo. For example, you wouldn't allow Waunakee to dump sewage into Lake Mendota, if they could argue that the lake was already polluted. If MG&E were to argue (as they have implied, in conversation) that this test well makes it OK to pump salt, then they are trying to keep pumping on a technicality, and not to seriously address the cause of the problem.
*** mgy=million gallons per year. This figure of 76 mgy is for the Cogen plant. The Charter Street heating plant sends a staggering 137 mgy out of the basin, and the Walnut Street plant sends 75.5 mgy out of basin. These two older plants are covered by the license, but the Odana mitigation figures compensate just for Cogen, because the water used by the older plants was "grandfathered" into the agreement.
Further reading
"Why the Odana project is so complicated"
Documents and sources
Request for Building Commission action on Cogeneration
Study of harmful effects of salt on key aquatic species.
Risk study of salt in Australia
This powerpoint presentation by Steve Gaffield, May 11, 2011, presents data and photos about the Odana Project.
# # #
Link to the only Rube Goldberg machine more complicated than Odana.
