Update Article Reprint - Summer 2007

Not an ordinary day at the cottage:                     Article by Bill Bialkowski

How doing some routine homework for GBA’s
water levels committee led to some alarming findings

Can you believe it? It’s a beautiful warm sunny day in April at our cottage in Snug Harbour—the ice went out a few days ago—and I’m sitting inside writing an article about water levels! I hope to put recent events - the St. Clair River and the Upper Lakes Study - into perspective, but my thoughts are straying. I’ve just inspected the water line on a rock near our dock, and it tells me that the water level is exactly at Michigan-Huron (M-H) Chart Datum or 176.0 meters above sea level.

Soon it will be time to launch our sailboat, the official start of the season. The boat draws nearly four feet of water; so getting through Canoe Channel should be a breeze, since my chart shows a minimum depth of seven feet. It occurs to me that my grandchildren will not be able to do this if Environment Canada’s climate change predictions come true and the water level declines 1.5 metres. Worse still, the wetlands will turn into grass meadows and the minnows and musky will be gone. Unless of course we succeed in holding back some water.

I click on the NOAA - that is the US National Oceanographic & Atmospheric Agency website (http://www.glerl.noaa.gov) to check the current water levels. Georgian Bay is part of Lake Huron, which is joined to Lake Michigan via the 7 kilometre wide Straits of Mackinac (see the Great Lakes Basin Map), so these two really are effectively one lake and referred to as Lakes Michigan-Huron (M-H). The website confirms what my rocks tell me: the M-H level is 176.023 metres for April 19. (See Figure 1) M-H is in green, Lake St. Clair in purple and Lake Erie in blue, and all except M-H are tracking higher than last year’s numbers, which are shown in black. In January ’07, M-H was about 15 centimetres (six inches) higher than last January, but has lost ground and now it looks like it’s tracking exactly along last year’s curve.

But while M-H’s level is the same as last year, Lake St. Clair’s is higher by nearly 10 centimetres (four inches) and Erie’s is higher still by nearly 15 centimetres (six inches). I wonder why these lakes are higher than last year and Georgian Bay is not?

I serve on GBA’s water levels committee with Ed Garner, and we concentrate on keeping the numbers straight and in focus. (Other members over the years have been Mary Muter, the chair, Jane Glassco Davis and Jeremy Gawen.) Ed and I both have engineering backgrounds including some hydrodynamics and numerical modeling, useful stuff for understanding water levels.

GBA’s graph, which shows the levels of Lakes M-H, St. Clair and Erie, captures the major problem we face on Georgian Bay: our water level has been trending downwards for years in spite of normal ups and downs while the other lakes go up. But this graph hasn’t been updated since 2001; so now I’m updating it. Earlier I spent a few hours at the NOAA web site downloading monthly average water levels for all the Great Lakes onto a spreadsheet. The updating involves some careful analysis to make sure the old and new data “stitch” together properly. There are some minor discrepancies of only a few millimeters; so not to worry. See Figure 2 for the updated graph.

My new data confirm the trends: M-H has been stuck at 176.0 metres or chart datum since 2000, while both St. Clair and Erie have been climbing: Lake St. Clair by about 15 centimetres or six inches since 2000, and Erie by about 30 centimetres or one foot. This confirms what I noticed in Figure 1.

But note the historical variation range in M-H. The lowest level ever recorded was in 1964 (175.69 metres above sea level or one foot below chart datum), before it rebounded a couple of years later and then continued climbing until reaching an all-time high of 177.28 metres (4.2 feet above chart datum) in 1986.

Climatic variations drive water levels; so they can vary widely. Great Lakes water levels have a number of natural cycles, one being of thirty-three years. The 1964 low and 1986 high were part of this cycle. The Baird Report of 2004 (more on this later) identified another longer 160-year cycle, which is expected to crest around 2015, about the same time as the 33-year cycle crests. If this double-cresting happens, we’ll see high water levels for the first and only time this century before the 160-year cycle dives into its trough. But this prediction doesn’t take account of the effect of climate change; so who knows?

Major Water Level Events

Throughout the 1990’s, most people didn’t pay much attention to water levels. But back around 2000, GBA’s environment chair Mary Muter was watching wetlands drying up and headed off for the St. Clair River to have a look. What she saw alarmed her. She took pictures. She drew trend lines through water level graphs and took them to meetings, where she sounded the alarm, but many people remained skeptical. Some years later, she was speaking at one of these meetings in Pointe au Baril when she was pestered with questions from one man in particular who, as it turned out, knew something about hydrology. That man was Ed Garner.

When water levels began dropping in 1999 but didn’t rebound as they had after 1964, the water levels committee got really concerned. This coincided with several important events. In September 2003, GBA hosted Dennis Schornack, U.S. co-chair of the International Joint Commission (IJC), on a boat tour, which gave him a much better appreciation of our concerns, especially the plight of our pristine wetlands. He confirmedan outstanding IJC Order for mitigation in the St. Clair River was still valid, and that it had not been superseded by any subsequent bi-national agreement.

Then shortly after Dennis’s visit, I got an email from Ed Garner regarding some numerical modeling he was doing. By downloading water level data for M-H and Lake St. Clair, as well as flow data for the St. Clair River, Ed confirmed Mary’s suspicion that there was something really strange going on. In particular the level of M-H was dropping alarmingly relative to Lake St. Clair (Figure 3). Between 1890 and 2000 it dropped from 2.7 metres to 1.3 metres, more than 50 per cent! Wow! Something surely must be wrong!

How Lake Michigan-Huron Levels Respond to Water Supply

You have to grasp some fancy terms and a principle of physics and to understand the significance of all of this. First, Net Basin Supply (NBS). Precipitation plus the runoff from streams and tributaries, minus the evaporation, equals NBS. (Keep in mind that a basin is a landmass shaped like a saucer that includes a water body and its tributaries. All the precipitation falling inside the saucer drains towards this water body. Think of NBS as the “net” supply of water coming into the basin. Hydrologists record NBS as a flow in cubic metres per second.

The NBS for the M-H Basin averages about 3,100 cubic metres per second, to which you add the flow from Lake Superior via the St. Mary’s River, another 2,100 cubic metres, for a total of about 5,200 cubic metres per second. When everything is in a state of equilibrium, this is also the amount of water that on average flows down the St. Clair River every second.

Imagine for a moment if we could block the outflow at the St. Clair River for a full year (obviously impossible), the NBS would back-up in M-H and create a volume of 164 cubic kilometres in one year; that’s a block of water 13 kilometres square by one kilometre high (or eight miles square by 0.6 mile high.) When spread over M-H’s huge surface area of 177,400 square kilometers, the water level would go up by 1.4 metres high (4.6 feet.)

The amount of water entering the M-H basin varies considerably, season-by-season and year-by-year, and of course it is this variation which causes the lake level to go up and down. But the huge surface area of the lake causes a smoothing or averaging effect, something like the inertia of a super-tanker : it takes considerable time to register water supply changes on M-H lake levels. The 1964 lowest level ever recorded, was actually caused in 1963 by the driest year ever, in which the average water supply only averaged 4,212 cubic metres per second, or 1.1 metres of water over the whole lake.By contrast in 1986, the highest level in recent history was caused by 1985 being the wettest year, in which the average water supply was 7,560 cubic metres per second, or 2.0 metres of water over the whole lake.

So why are these NBS measurements important? Because long-term predictions are that higher evaporation and stronger winds, caused by climate change, are expected to lower the NBS for the M-H Basin.

A Physics Lesson

Now back to how the St. Clair River works and your physics lesson: as we all know, water flows downhill from a higher to a lower level. Hydrologists call the elevation differential between the two the “head”. Now energy cannot be created or destroyed (the first law of thermodynamics), and so water flows down hill because at the top of the hill it has this head or potential energy. In fact it is this energy that pushes it downhill.

The water at Sarnia and Port Huron at the top of the St. Clair River is 1.3 metres above Lake St. Clair and this represents the amount of potential energy available. As water flows downhill, it gains speed or kinetic energy but loses height or head, as the potential energy changes to kinetic. By the time the water gets under the Blue Water Bridge, the current is zipping along at three knots, evidence of lots of kinetic energy. But the kinetic energy is not all converted to speed as much of it is used up to overcome turbulence or friction as the river twists and turns and the water bangs against the banks, the bottom and weeds. It also picks up sand and drops it downstream. All this requires energy. By the time the water reaches Lake St. Clair, it has next to no kinetic energy left. The principle involved is that it takes 1.3 metres of head to push 5,200 cubic metres of water down this river (at least it did in 2000; in 1890 it took 2.8 metres of head to do the same thing).

If the water supply is greater than 5,200 cubic metres per second and the head (elevation differential) is still 1.3 metres, only 5,200 cubic metres will flow down the St. Clair River because that all the potential energy that is available. The excess water will accumulate and raise the level of M-H. As that level rises, so does the head and with it the flow down the river. This will happen until a new balance is reached.

Lake levels go up and down in accordance with the water supply, and this drives the natural variability we see in Figure 1. But the lake level is also governed by how much energy is needed to drive the water down the river; call this the flow conveyance capacity of the river.

Ed Garner’s Discovery

What Ed Garner’s e-mail said, (see Figure 3), is that the head between M-H and Lake St. Clair appeared to be dropping at an alarming rate, in spite of the fact that the river flows had maintained their historical average! Figure 3 shows a drop of over 50 per cent.

The flow conveyance capacity of the St. Clair River was obviously increasing. This was a major revelation, and GBA’s water levels committee studied it for some time. We even ran some simulations of what could be done if somehow we could artificially modify the river’s flow conveyance capacity by some form of variable control structures, like moveable dams or weirs. These simulations showed that we could restore the level of M-H to historic values by holding some water back, especially during wet years.

Attempts to Draw Attention to the St. Clair River Problem

In April 2004, Mary Muter, Jeremy Gawen and I presented our findings to Environment Canada officials. They listened patiently but they did not agree that the flow conveyance capacity of the St. Clair River had changed. They suggested that instead the relative supply of water to the Erie and M-H basins had changed; Erie was getting wetter and M-H dryer, and that is why Erie and St. Clair were going up and M-H down. But really, they said, GBA should not be spending its time and money worrying about these things, Environment Canada would look after this, for it was just an NBS matter, as they and the US Army Corps of Engineers (USACE) had determined.

We were deflated, but we didn’t quite buy into what Environment Canada was saying. If Erie was going up and M-H down because of their relative water supply, this would also lead to a decreasing head, which would in turn would cause the flow to decrease – yet since 1900 the average river flow was about 5,200 cubic metres per second, and in 1986 the flow exceeded 6,200 cubic metres per second; these are facts that needed to be explained in light of the falling head. This is a key principle, and if you don’t get it, you’re going to be drawn into an argument as to whether water flows uphill.

When informed about the seemingly illogical position of the Environment Canada officials, the GBA Foundation decided to retain the internationally respected hydrological firm of Baird and Associates to do a proper assessment. It took a lot of courage to commit a quarter of a million dollars to work that governments were supposed to be doing.

The Baird Study

The 2004 Baird Report provided evidence that the St. Clair River was eroding. This was new. Since about 1975 the relative difference in water level of M-H and Lake St. Clair (the head) had been dropping at a rate of about one centimetre per year; the drain hole was getting bigger. This one centimetre per year decline in level was equivalent to a continuous diversion from M-H of 1.2 cubic kilometres per year or 845 million gallons per day. Dr. Rob Nairn, Baird’s principal investigator, also suspected that the rate of decline had accelerated since 2003 but he didn’t have the data to prove it.

On the other hand, we were pleased to learn that the 160 and 33-year natural cycles would crest together around 2015. This would cause the water level to rise at a rate of about three centimetres per year until then. On the strength of this, we thought we stood a chance of not losing everything while the problem was being investigated.

Glory

May 30, 2005 was GBA’s moment of glory. We convened a meeting of scientists: Environment Canada, USACE, the US National Oceanic and Atmospheric Agency (NOAA), the US Geological Survey, the Great Lakes Commission and the IJC.

Dr. Rob Nairn presented the results of his study. He showed how the M-H levels had declined as a function of the declining head. He considered a range of possible causes and concluded that the St. Clair River bed’s sand and or clay particles were being swept away by the current, as a result of dredging and other events that had scooped away the upper protective layer of the river bottom. Parts of the river were becoming much deeper, thereby creating a greater flow conveyance capacity, with the result that more water was leaving M-H than was coming into it; so it was becoming lower and Lake St. Clair was becoming higher. At the end of the meeting, Dennis Schornack, the IJC co-chair said: “Ladies and gentlemen, I am convinced that this represents valid science and that we have a problem. We have no choice but to investigate the alleged erosion problem of the St. Clair River.”

With that simple statement, the mandate of the upcoming Upper Lakes Study was expanded to include a new first phase, a study of the St. Clair River erosion and potential mitigation measures. It would take about two years to complete, but GBA was ecstatic, and the U.S. and Canadian governments subsequently agreed to commit $18 million to the project.

Later we learned that, while the U.S. had allocated its portion of the funding, the Canadians were stalling. GBA met with Environment Minister Rona Ambrose and her successor John Baird to step up the pressure. The Stephen Harper government had just come to power, and this issue wasn’t a priority. It took GBA three tries to get the feds to fund the first two years of the Upper Lakes Study, and that’s where we stand now. The project will move ahead, but Environment Canada and USACE aren’t happy with the Baird Report.

Why the Controversy

Dr. Rob Nairn, had closely examined Environment Canada’s theory that differential water supplies to the Lake Erie and M-H respective basins explained the dropping head and had rejected that hypothesis. Nairn also took issue with how the Ministry and USACE calculate NBS, the so-called “residuals method.” First you figure out how much the lake level has changed over a given time period, subtract the inflow and add the outflow. Getting the levels and flows data is easy, given that these numbers are recorded hourly and saved online in numerous formats. But it’s too easy.

NOAA’s Great Lakes Environmental Research Lab. (GLERL) does it better using the “components method”. This involves using weather information to estimate the total precipitation on the basin, and temperature data to estimate the total evaporation, a complicated process, but it works.

Level measurements are taken all around the Great Lakes and along the St. Clair River, but that’s the easy part. River flows are more complicated. To start with, think about the size of the St. Clair River; at its narrowest it is about 300 metres wide and as much as 20 metres (sixty-five feet) deep in places. (Note: ships need less than 30 feet!) The flow velocity varies across a given cross-section of the river, generally faster in the middle and slower near the shores and bottom. So how do you measure flow? Traditionally on a given day you select a cross-section of the river and take lots of measurements by using flow velocity instruments in as many places as you can across your given river cross-section, both shallow and deep. You then calculate an average velocity for the whole cross-section: lots of numbers.

You determine the water level using level gauges at the nearest locations upstream and downstream of the cross-section, and subtract the two to get the head. After all, it’s the head that causes the flow to be what it is. Then you plug the average flow and the head into hydraulic equations and make them fit these data points. The result is known as a “stage-discharge-equation”, which will infer the right flow number based on measuring the head difference between the two level gauges on the river. Sounds simple enough, but there is a catch: the flow measurement will be wrong should the river contours change. And when a river is eroding, they change! But neither Environment Canada (EC) nor USACE had flow meters at the critical part of the river where the flow is fastest and the river is deepest.

Using the residuals (stage-discharge-equation) method of EC and USACE to calculate NBS, said Baird’s Rob Nairn, can lead to a circular argument. The change in level is accurate, but the flow calculation assumes the river has not changed. But if its bed has eroded, the river flow will appear less than it actually is. The result: M-H looks lower and dryer, and Lake Erie looks higher and wetter, just like EC reported in April 2004!

The Baird report compared the NBS estimates for the Erie and M-H basins using both calculation methods. Starting in about 1970 the residual method showed a wetter Erie than did the components method. The largest change in head that could be justified based on these calculations, however, was less than four centimetres. Since we were attempting to explain a head decline of over one metre, this didn’t cut it.

Refuting Validity

But EC and USACE were not convinced by Baird's conclusions and so they went back to the drawing board. In 2005 USACE conducted a new bathymetry (water depth) study of the river to investigate how the bottom contours had changed. When comparing the 2005 and 2002 data, it stated that even though some parts of the river had become deeper, others were shallower and that on balance, “because there has been no net change in the area of river flow, … these river bottom changes would not likely lead to increased river flows."

USACE did not make the 2005 bathymetry data on which they based this report available; so GBA had to make a Freedom of Information request to gain access to it, and again Baird was retained to check USACE’s findings. Baird calculated exactly how much material had been removed on a net basis from the riverbed by the scouring. From this detailed analysis, it concluded that 32,000 cubic metres of river bottom had for all intents and purposes disappeared, thus increasing the river’s flow conveyance capacity. The drain hole had gotten bigger.

As a last step of the second study, Baird entered the 2005 bathymetry data into their St. Clair River model and found that the level of M-H had dropped 2.4 centimetres (almost one inch) since 2002 based on bathymetry data.

New Head Data to 2007

So I had just downloaded the latest M-H level data to update GBA’s level trends graph from 1865 up to 2007 and thought I’d have a look at the head between M-H and St. Clair and Erie, when I hit upon a remarkable discovery.

Figure 4 shows what happened, and it’s alarming. The figure shows the head between M-H and Erie (in red) and M-H and St. Clair (in blue). The M-H to Erie head declined from 2.8 metres in 1900 down to 1.8 metres today, a total loss of one metre. The M-H to St. Clair head declined from 1.8 metres in 1900 down to 1.0 metres today, only one metre! In 2000, it was 1.3 metres, hence a drop of 30 centimetres (one foot) in only seven years, much faster than the Baird Report findings predicted!

The red plot in Figure 4 shows a decline of 23 centimetres between 1975 and 2000, about one centimetre per year. This was outlined in the findings of the Baird Report. The decline is shown as a black dashed line. Shortly after 2000 or so, erosion in the St. Clair River accelerated significantly, just as Rob Nairn suspected, and the slope now looks more like three centimetres per year. (The graph includes a slope scale; so you can judge for yourself).

If true, this would exactly cancel the predicted rise due to the cresting of the two natural cycles around 2015. Perhaps this explains why, (Figure 1), M-H has been flat-lined since 2000, while the St. Clair and Erie levels have been rising. But, since the St. Clair River erosion phase of the Upper Lakes Study is not due for completion for another two years, and even more time will be needed before anyone considers mitigation measures, let alone their implementation, it means we will lose at least another 10 centimetres (four inches) or so of head.

GBA has been using the number 845 million gallons per day to put the one centimetre per year decline in easily understood terms. Well, a loss of three centimetres per year translates into 2.5 billion gallons per day. The erosion has to stabilize because it is physically impossible for M-H to sink below Lake St. Clair. With only one meter left to go, this erosion rate must slow down and stabilizes at some point above Lake St. Clair. It can’t get much worse!

The Upper Lakes study team is coming together. We’re hoping for a highly competent team of people with the right mix of expertise. We have heard accounts of meetings being held without all the required experts, but we hope these are just normal start-up growing pains that will be quickly resolved.

Understanding how the contours of the river bottom have evolved over time and how the sediment moves along the river bottom will require a very sophisticated three-dimensional model of critical sections of the river. Such models section the river into thousands of little volumes, and in each volume the physical equations for water and sediment transport—how much potential energy is converted to kinetic energy—and then lost to the river contours can be rigorously studied. We will know that the model is accurate when it includes all these myriad details and correctly predicts everything that has actually happened to the river to date. Then we can use the model to study proposed mitigation measures. We look forward to the findings due in 2009. GBA will be watching.

Mitigation Measures

So what might these mitigation measures be? We hope that the river bottom can be stabilized to prevent further erosion by placing large rocks where the erosion is actively occurring. That would stop the bathtub drain from getting any bigger. This would not be a mega-project.

But GBA’s real hope is that some form of variable control structures will be designed that would allow some water to be held-back, allowing some of the M-H historical level loss to be recovered. Following the all time low level of 1964, the IJC had developed sketch designs using double-hinged gates to control the St. Clair River flow, but then the all time high water of 1986 arrived and nothing further was done. We hope that some form of this approach will be a part of the final mitigation measures.

In addition, we hope that our governments will ask the IJC to establish an outflow monitoring board that would decide how the river flow is to be adjusted every month or so, much as the control boards for Lakes Superior and Ontario do now. There is no doubt that the St. Clair River erosion problem and climate change represent the biggest threats to Georgian Bay water levels. But overall I am guardedly optimistic and believe that in the present environmentally sensitive political climate we stand a chance of meeting these challenges. After all, there still is an outstanding IJC order for mitigation in the St. Clair River that has never been fulfilled, surely this time it will be. I hope my grandkids will navigate our little sailboat (if it is still around) through Canoe Channel and see minnows and musky in the wetlands.