EFFECT OF PUMPING ON DIRECT AND INDIRECT GROUND-WATER LOW TO LAKE MICHIGAN AND ON GROUND-WATER FLOW BETWEEN THE MISSISSIPPI RIVER AND LAKE MICHIGAN BASINS

Regional pumping has reduced the amount of ground water discharged to Lake Michigan from shallow unlithified and Silurian deposits. Under both predevelopment and pumping conditions, the flow model shows that ground-water movement in the shallow part of the flow system along the Lake Michigan coastline is almost exclusively toward the lake. However, pumping has decreased the total amount of ground-water discharge to the lake. The decrease appears in two ways. First, pumping reduces the direct discharge of ground water beneath the coastline to deposits under the lake and ultimately up through the lakebed into the lake itself. Second, pumping reduces indirect discharge of ground water to Lake Michigan. Indirect discharge consists of ground-water discharge to streams that flow into the Lake. All streams east of the subcontinental surface-water divide flow into Lake Michigan. All the land area east of this watershed divide is part of the Great Lakes Basin.

First, consider the reduction in direct ground-water discharge to the Lake:

Model output: Map showing areas where part of ground-water circulation discharges directly to Lake Michigan rather than to streams which are tributary to the Lake: Predevelopment compared to 2000 (source: Wisconsin Geological and Natural History Survey Open-File Report 2004-01)

Note that the area where direct ground-water discharge to the Lake can occur is much smaller than the area east of the subcontinental watershed divide where indirect discharge can occur.

The model quantifies the amount of direct ground-water discharge over time. Under predevelopment conditions it simulates a rate of 13.3 million gallons per day. In year 2000, it simulates a rate of 12.2 million gallons per day, amounting to a decrease of over 8%.

Next consider the reduction in indirect discharge to Lake Michigan:

Indirect discharge can be approximated as total of all ground-water discharge to surface-water bodies within the Great Lakes Basin in southeastern Wisconsin (it is an approximation because not all water bodies necessarily route water to Lake Michigan, but it is a fair assumption that most do). The model results for predevelopment conditions simulate indirect discharge equal to 144.8 million gallons per day. The corresponding 2000 results are equal to 134.0 million gallons per day, a reduction of about 7%. Clearly the amount of indirect discharge is much larger than the amount of direct discharge before and after development.

All these results are summarized in the following table. It also includes model results for an intermediate period, 1950:

East of the subcontinental divide, pumping has changed how shallow ground water is replenished and how it is discharged. The pie charts below show the water balance for the shallow ground-water system through time for the inland part of the Lake Michigan basin east of the subcontinental divide in southeastern Wisconsin. The thumbnail charts indicate that while recharge remains the predominant source of shallow ground water through time, shallow pumping is now a major sink because it discharges more ground water within the basin than does direct outflow to the Lake. Note that the inflow and outflow units in the charts are million gallons per day (mgd).

Model output:	Model output:
Shallow Ground-Water INFLOW Rates Within Lake Michigan Basin	Shallow Ground-water OUTFLOW Rates Within Lake Michigan Basin
1864	1864
1950	1950
2000	2000
(source: D.T. Feinstein, U.S. Geological Survey)

Another way to look at the effect of pumping on ground-water flow to Lake Michigan is to consider the volumes (rather than the rates) of water discharged from the land toward the Lake over an interval of time from both the shallow and deep parts of the flow system. IN THE ABSENCE OF PUMPING, Lake Michigan and the ground-water system under the lake between 1864 and 2000 would have received the following volume of direct ground-water discharge in BILLIONS of gallons:

WITH PUMPING, Lake Michigan the ground-water system under the lake actually received or yielded (negative) the following in direct ground-water discharge between 1864 and 2000:

In other words, between 1864 and 2000 Lake Michigan and the ground-water system under the Lake along Southeastern Wisconsin has lost the following volume of water due to the effect of shallow and deep pumping:

Note: These estimates of water lost to the Lake do not take account of water added to the Lake via return flow from sewers and water-treatment plants.

Change in DEEP FLOW PATTERN below Lake Michigan

As pointed out in earlier pages of this website, pumping has reversed the pattern of deep regional ground-water flow to Lake Michigan:

Model output: Ground-water flow lines to and from Lake Michigan: A) Regional flow to the Lake under natural conditions	Model output: Ground-water flow lines to and from Lake Michigan: B) Regional flow from the Lake under 2000 conditions
NOTE: These simulated flow lines represent the flow field that would exist if 2000 pumping were to continue for hundreds of years.
(source: D.T. Feinstein, U.S. Geological Survey)

Under predevelopment conditions the model simulates that 1.9 million gallons per day of ground-water flows upward from the deep to the shallow part of the flow system and into Lake Michigan over the area of the Lake between southeastern Wisconsin and Michigan. Under 2000 conditions, the model simulates that 1.3 million gallons per day flows vertically in the opposite direction, most of which originates as Lake water. However, part of the 3.2 million gallons per day difference is due to pumping outside of southeastern Wisconsin, notably in northeastern Illinois.

SOUTHEASTERN WISCONSIN IN CONTEXT

According to the computer model, the sum of direct and indirect ground-water discharge to lake from the seven-county area is currently on the order of 145 million gallons per day or 224 cubic feet per second. This amount is a very small percentage of the total ground-water discharge to Lake Michigan, estimated to be almost 35,000 cubic feet per second (U.S. Geological Survey Water-Resources Investigations Report 00-4008). The contribution from southeastern Wisconsin is so small because the inland portion of the basin is very narrow and, therefore, receives a lower volume of precipitation than elsewhere within the Lake Michigan watershed:

Map of Great Lakes Basin with ground-water and surface-runoff components (source: U.S. Geological Survey Open-File Report 98-579)

The simulated reduction of shallow direct and indirect ground-water discharge to Lake Michigan between the late 19th century and the year 2000 is on the order of 13 million gallons per day, equivalent to 21 cubic feet per second, or 8.5% of the predevelopment shallow ground-water discharge from southeastern Wisconsin. The reduction of ground-water discharge to adjacent areas of Lake Michigan from deep rocks due to reversals of gradients is smaller - the model simulates a loss of about 3 million gallons per day or 5 cubic feet per second. The sum of these reductions, 26 cubic feet per second, is negligible when taken in context of the overall budget of the Lake. The total inflow to the Lake (precipitation + surface runoff + direct ground - water discharge + indirect ground-water discharge) is estimated in USGS publications to be about 96,000 cubic feet per second, a level nearly 4,000 times greater than the loss due to wells. In light of these numbers it is clear that pumping in southeastern Wisconsin by itself, although it has had a calculable effect on local hydrologic conditions, has had virtually no effect on the amount of water in Lake Michigan or on its water level.

In digesting all these results for the effect of pumping on ground-water discharge to Lake Michigan, it is important to keep in mind an important limitation of the modeling results. Over time, the recharge rate supplying water to the ground-water system has probably changed. However, it is very difficult to evaluate this change because competing forces are at work. Urbanization in southeastern Wisconsin has brought pavement and parking lots which tend to increase overland runoff and, consequently, reduce ground-water recharge. On the other hand, there is evidence across Wisconsin that ground-water discharge to streams is increasing in many areas, possibly because the soil is freezing later in the year and allowing more precipitation to percolate as recharge. The net effect of urbanization and climate change is unknown for any particular region such as the Great Lakes Basin in southeastern Wisconsin. In the model the recharge rate is kept constant through time, but if there has, in fact, been a significant increase or decrease in recharge, then there would have been a corresponding increase or decrease in ground-water discharge that would either oppose or magnify the simulated effect of pumping.

Another factor neglected by the model is return flow to Lake Michigan from storm and sanitary sewers and from the Deep Tunnel that underlies Milwaukee. Part of the water pumped from the shallow and deep parts of the flow system east of the subcontinental divide is discharged to sewers and routed back to the Lake. While this return flow offsets part of the decline in ground-water discharge to the Lake that occurs because of pumping, it also changes the timing, location, and quality of the discharged water relative to natural conditions.

IS THE MISSISSIPPI RIVER BASIN or THE LAKE MICHIGAN BASIN THE MAIN SOURCE OF WATER TO DEEP SANDSTONE AQUIFER WELLS IN SOUTHEASERN WISCONSIN ?

An important conclusion of the modeling study is that the deep wells in southeastern Wisconsin are located within the Lake Michigan ground-water basin (see the Case Study section called “Effect of pumping on ground-water divides in the deep sandstone aquifer in southeastern Wisconsin”). But because the shallow and deep parts of the flow system are interconnected, this fact alone does not imply that the water replenishing the wells comes completely or even mostly from the Lake Michigan Basin.

Map showing location of subcontinental divide (source: D.T. Feinstein, U.S. Geological Survey)

The majority of the water that replenishes discharge from southeastern Wisconsin wells originates as captured baseflow - that is water that would have circulated thru the shallow flow system to streams but is being diverted downward toward the deep wells. In fact, according to the model, captured baseflow constitutes 59% of the sources of water to deep wells for 2000 pumpage (= 33 mgd). Other sources include storage release and inflow from the west (e.g., Jefferson County ) and from the east (e.g., from under Lake Michigan).

The question arises - how much of that captured baseflow is captured from streams in the Mississippi River Basin and how much from streams within the Lake Michigan Basin. The answer is that ALMOST ALL diverted baseflow to streams is captured from the Mississippi River Basin - 9/10 of the captured baseflow to be precise. This is no surprise because the Maquoketa shale becomes an increasingly powerful obstacle to vertical flow downward as you move east from the Maquoketa subcrop in western Waukesha County toward Lake Michigan.

Moreover, according to the model, 18% of the 33 mgd pumped from deep wells flows into southeastern Wisconsin from the west. Virtually all that amount is due to capture of baseflow from streams located west of southeastern Wisconsin but in the Mississippi Basin.

Conclusion : 71% (0.9*59% + 18%) of the water replenishing discharge from deep wells in southeastern Wisconsin is ground water diverted from streams (captured baseflow) within the Mississippi River Basin even though the wells themselves are located in the Lake Michigan ground-water basin. The remainder, according to the model, is from captured baseflow in the Lake Michigan Basin (6%), inland storage release (3%), storage release below Lake Michigan (8%), water that once flow toward rocks under Lake Michigan (8%) and water flowing out of Lake Michigan itself (only 4%).

It follows that Lake Michigan is NOT the main source of water to deep wells – for two reasons. First, water currently discharging from wells is NOT derived from Lake Michigan but from old water that entered the ground-water system as recharge in the Mississippi Basin . Second, the source of water that is replenishing the aquifer water discharged by deep wells is mostly diverted from local streams in the Mississippi Basin.

While the diversion of streamflow is an important source of water to deep wells, it produces a small change in the streams themselves. In fact, other factors such as urbanization and climate variability are more important in determining changes in streamflow over time, and these other factors mask any measurable change due to deep pumping.

One final point – most of the water discharged from the deep well pumping centers is recirculated back to the Mississippi Basin (net a small amount of loss to evaporation) via water treatment discharge to the Fox River.

return to top