3.3 Paw Paw River, Michigan

Transcription

1 Conservancy will be applying the results of this assessment (and conducting some additional assessment using the stream power tool) to select target reaches for restoration using 1) channel restoration techniques and 2) forest and wetland restoration within catchment. Possible rules for targeting restoration could include: Subwatersheds with least hydrologic alteration where alteration is focused in a few reaches. Highly altered upstream reaches. Reaches that have species of concern or potential to support species of concern. Opportunities to implement forest or wetland restoration. The Conservancy will develop and apply several forest and wetland restoration scenarios using stream power tool to identify locations for restoration that have greatest benefit to stream power in targeted reaches. 3.3 Paw Paw River, Michigan Location: The Paw Paw River is a major westward flowing tributary of the St. Joseph River that drains approximately 445 square miles of southwestern Michigan near Benton Harbor (Figure 3.3-1). The watershed is well known for containing a diverse landscape including large fens, state designated trout streams, and several rare species including the Eastern Box Turtle. 77

2 Figure Location of the Paw Paw River watershed and USGS gage at Riverside, MI. Geology: Characterized in the headwaters by glacial end moraine and ice-contact ridges with associated kettle depressions, the Paw Paw Watershed contains a diverse surficial landscape. These coarse textured deposits result in groundwater contributions in the headwaters and along certain portions of the mainstem of the river as well as depressional storage in the form of wetlands, ponds and small lakes. Data availability: Only one USGS gage ( ) exists within the watershed that has a substantial monitoring record (Figure 3.3-1). For this study, we analyzed streamflow data using the record for the Paw Paw River at Riverside, MI ( ). Paw Paw River subwatersheds and analysis sites: The mainstem of the Paw Paw River is a coolwater stream that contains some coldwater headwater areas that support trout populations (Wesley and Duffy 1999). Within the Paw Paw River we analyzed hydrologic data and summarized anthropogenic changes for one USGS gaged location and eight other subwatersheds (Figure 3.3-2). These were of particular interest due to planning efforts underway by the Michigan Southwestern Regional Planning Authority and The Nature Conservancy due to potential changes to biological communities or habitat conditions that may be linked to hydrologic changes. 78

3 Figure Selected Subwatersheds for Data Analysis in the Paw Paw River Watershed. Land cover: Although land use within the watershed is dominated by agriculture (47%) the basin still contains a large forested component (32%). Open water and wetlands make up 15% of the catchment area with the largest fraction being wooded wetlands (12%) in the form of flood plain forest. Urban landcover currently makes up less than one tenth of the basin (7%), but development pressures are increasing, making loss of wetlands and forest lands a concern. Comparisons with presettlement landcover suggest that agriculture has replaced forested lands, and particularly wooded wetlands, in much of the watershed (Figure 3.3-3, Figure 3.3-4, Table 3.3-1). 79

4 Figure Recent land cover within the Paw Paw River watershed. Figure Presettlement land cover within the Paw Paw River watershed. 80

5 Table Presettlement and Current Land Cover for the Paw Paw River Open Water Emergent Herbaceous Wetlands Woody Wetlands Grasslands/ Herbaceous Shrubland Deciduous/ Mixed Forest Evergreen Forest Row Crops Low intensity Residential Commercial/ Industrial/ Transportation Bare Rock/ Sand/ Clay Pre- Settleme nt Current Acres 6,393 1,652 44, , % of Total Acres 5,651 1,598 34,347 26,703 3,901 47,615 4, ,156 12,010 5,320 2,291 % of Total Change % of Total 1% -5% change >5% change Acres ,112 26, ,893 4, , Results of watershed assessment tools Differences in stream power can be presented in two ways: (For a full discussion of the tool see section of this report.) 1) Using land cover from presettlement and recent time periods (and, in this case, a modeled future time period), changes in stream power can be calculated and compared using a percent power change calculation, i.e., Percent Power Change = Recent (and/or Future) Stream Power Presettlement Stream Power Presettlement Stream Power Figure illustrates percent power change from presettlement to current conditions (IFMAP) and presettlement to modeled future conditions (MI LTM 2040) for the one gaged subwatershed and other selected subwatersheds in the Paw Paw River Watershed. This analysis suggests that patterns of landcover change have led to increased stream power in many reaches of the Paw Paw River watershed. The percent change in stream power increased in all gaged watersheds from approximately 80 to slightly over 110 percent from presettlement to current conditions. A similar pattern is also expected based on the Michigan Land Transformation Model (Pijanowski et al., 1996, Pijanowski et al., 2000) although the potential increases as a result of changes in landcover are not as pronounced as changes that have already occurred (Figure 3.3-5). Estimated landcover alteration based on this scenario would lead to further increases of 80 to 100% in modeled stream power. 81

6 % Power Change from Presettlement IntPt1 IntPt2 IntPt3 IntPt4 IntPt5 IntPt6 IntPt7 IntPt8 IFMAP MI LTM 2040 Figure Percent Power Change from Presettlement in Selected Subwatersheds in the Paw Paw River Watershed Under Current Land Cover (IFMAP) and Future Land Cover (Mi LTM 2040) Scenarios. 2) Calculating the maximum stream power change possible for each stream reach and compare the actual change (from presettlement conditions) to the maximum possible change (under hypothetical conditions). This power change metric provides a quantitative way to estimate hydrologic degradation (or improvement) relative to maximum possible degradation. Power Change Metric = Recent Stream Power Presettlement Stream Power Paved Paradise Power Presettlement Power Figure illustrates the power change metric for the one gaged subwatershed and other selected subwatersheds in the Paw Paw River Watershed. This analysis suggests that patterns of landcover change have led to moderately high power change metric values in many reaches of the Paw Paw River watershed. The power change metric ranged from approximately 4 to 14. A similar pattern is also expected based on the Michigan Land Transformation Model although the potential increases as a result of changes in landcover are not as pronounced as changes that have already occurred but are significant (Figure 3.3-6). Estimated landcover alteration based on this scenario would lead to further power metric increases of 8 to 10 units bringing the total metric value of this modeled scenario from 11 to

7 25 Power Change Metric IFMAP MI LTM IntPt1 IntPt2 IntPt3 IntPt4 IntPt5 IntPt6 IntPt7 IntPt8 Figure Power Change Metric for Selected Subwatersheds in the Paw Paw River Watershed. Figure Power Change Metric by Stream Reach in the Paw Paw River Watershed - Current Conditions (IFMAP). 83

8 Flow Path Analyses: The potential effects of flow path changes on the Paw Paw watershed were evaluated by considering regional flow patterns resulting from water withdrawal, flow diversion, return flow, and the type or source and receiving waters. In the Paw Paw watershed, groundwater is the primary source of water for anthropogenic use. Surface waters are used for irrigation purposes only and represent less than 10% of the total volume of water used in the watershed. Factors considered in the pathways analysis for the Paw Paw watershed include: Source Waters and Diverted Flows Groundwater withdrawals occur from both shallow and deep aquifers. For the purpose of this project, shallow groundwater sources are defined by producing depths generally less than 60 feet with a reasonable expectation that local surfacegroundwater interaction may occur. Deep groundwater sources are defined by producing depths greater than 60 feet with a reasonable expectation that local surface-groundwater interaction will not occur. In the Paw Paw watershed there are more than 5589 producing water wells, of which 203 are publicly owned and operated. These public wells provide water to local communities where groundwater is collected, treated, stored, and then distributed through a public water supply system. There are more than 5,000 private household wells (self supply) where groundwater is withdrawn on an as needed basis. Groundwater wells are summarized below by type, producing depth, flow yield, and potential for flow path diversion (Table 3.3-2). Table Summary of Groundwater Wells for the Paw Paw Watershed. Producing Depth Average Yield (gpm) Well Type No. Wells < 60 feet > 60 feet < 60 feet > 60 feet n/a Houshld Public Type I Public Type II Public Type III Irrigation Industrial Low Flow Diversion Potential High Flow Diversion Potential Average Maximum Yield (gpm) In the Paw Paw watershed, water is withdrawn from deeper aquifers along the eastern edge and western portions of the watershed (Figure 3.3-8). Water is withdrawn from shallower aquifers within the central portion watershed. Average maximum yields are high for all types of wells. The potential for flow path diversion (flow augmentation) is highest where deep groundwater is diverted and returned to surface waters. High capacity public water wells are located along the Interstate 94 corridor and are concentrated in the northwestern portion of the watershed near Coloma, MI (Figure 3.3-9). Irrigation wells are spread uniformly across the watershed, with slightly higher concentrations in the center and southern portions of the Paw Paw watershed. Residential self-supply wells appear to be more heavily concentrated along the eastern edge of the watershed. 84

9 Figure Spatial distribution of shallow and deep groundwater wells in the Paw Paw watershed. Water is produced from deeper aquifers (orange and red symbols) along the eastern edge and western portion of the watershed. Water is produced from shallower aquifers (green symbols) in the central portions of the watershed. Figure Spatial distribution of public and self-supply groundwater wells in the Paw Paw watershed. High capacity public water supply wells (blue dots) supply water to local communities. 85

10 Annual reporting data for public water supplies and irrigation is available for the St. Joseph watershed and is provided in Table Approximately 25% of the public water supply is withdrawn from Lake Michigan and provided to the communities of Benton Harbor and St. Joseph. The remaining 75% of the public water supply is provided by groundwater sources within the St. Joseph watershed. Surface water accounts for approximately a third of the waters used for irrigation purposes, with the remaining two-thirds derived from groundwater sources. Table Reported Public Water Supply and Irrigation Withdrawals for Years for the St. Joseph (MI) watershed. Public Water Supply Year Lake (Mgd) SW (Mgd) GW (Mgd) Total (Mgd) Lake SW GW % 0.0% 74.5% % 0.0% 74.5% % 0.0% 75.6% % 0.0% 76.2% % 0.0% 75.8% % 0.0% 75.5% % 0.0% 75.2% % 0.0% 76.2% Mean: Irrigation Year Lake (Mgd) SW (Mgd) GW (Mgd) Total (Mgd) Lake SW GW % 33.9% 66.1% % 33.5% 66.5% % 33.3% 66.7% % 34.2% 65.8% % 34.2% 65.8% % 33.5% 66.5% % 32.5% 67.5% % 32.9% 67.1% Mean: The Paw Paw watershed is a subwatershed within the larger St. Joseph watershed and annual water supply and irrigation data are not available specifically for the Paw Paw watershed. Using average daily water withdrawal rates calculated by well type and producing depth in the Shiawassee watershed, it is possible to estimate the volume of groundwater withdrawn for public water supply and irrigation purposes in the Paw Paw watershed (Table 3.3-4). Water withdrawals from private household wells are typically not reported, but can be estimated based on per capita household water use for the Great Lakes region (Great Lakes Commission 2001). Using an estimated average water use of ~230 gpd per household, it is possible to estimate the volume of groundwater withdrawn for residential (self supply) use (Table 3.3-4). Based on these data, approximately 6.57 mgd of groundwater is withdrawn for residential (self supply), irrigation, and public water supply uses in the Paw Paw watershed. Using the reported ratio of surface water to groundwater (irrigation), approximately 0.93 Mgd of surface water is diverted for irrigation purposes in the Paw Paw watershed (based on the reported ratio of surface water to groundwater withdrawals for irrigation purposes in the entire St. Joseph watershed). 86

11 Table Estimated Groundwater Withdrawals by Well Type and Producing Depth for the Paw Paw watershed. Producing Depth ( # Wells) GW Withdrawl (Mgd) Well Type No. Wells < 60 feet > 60 feet n/a < 60 feet > 60 feet n/a Total Houshld Irrigation Public Supply Totals: Low Flow Diversion Potential High Flow Diversion Potential *Altered flow paths: 6.57 Mgd (0.33 Mgd Mgd) = 6.22 Mgd Approximately 6.22 Mgd* of groundwater withdrawn from the Paw Paw watershed is diverted along altered flow paths where water is returned either to a different hydrologic regime or to a location that is distinctly different (catchment or subwatershed) from where the groundwater originated. The groundwater Diversion Ratio (D GW ) for the Paw Paw watershed (V DIVg /V GW ) is 0.94 or 94%. The surface water Diversion Ratio (D SW ) is 1.0 or 100%. The Diversion Ratio (D) for all water withdrawals in the Paw Paw watershed is 0.95 or 95%. Return Flows and Receiving Waters In larger communities and in areas with public (or private) sewerage, wastewater is collected, treated, and then returned to surface waters of the Paw Paw watershed primarily the Paw Paw River or its tributaries (Figure ). In private systems (primarily households), wastewater is processed in a septic system and returned to shallow aquifers, generally less than 15 feet below the surface. Since groundwater is the primary source of water for consumptive use in the Paw Paw watershed, waters collected by public wastewater treatment plants will be diverted from a groundwater source and returned to a surface water source (the Paw Paw River). This represents a major flow path diversion where waters are being removed from one type of hydrologic regime (groundwater) and added to another hydrologic regime (surface water). The location of facilities with active NPDES wastewater discharge permits is shown in Figure Significant discharges of wastewater occur at the Paw Paw and Paw Paw Lakes waste water treatment plants (WWTPs). Virtually all groundwater withdrawn for public use in the northwestern portion of the Paw Paw watershed is treated by the Paw Paw Lakes WWTP. Table lists annualized diverted flow volumes (return flow - wastewater) for public and industrial wastewater treatment plants in the Paw Paw watershed. 87

12 Figure Location of wastewater discharges in the Paw Paw watershed. The Paw Paw and Paw Paw Lakes WWTPs are considered to be major dischargers of wastewater into the Paw Paw River. All of these waters are diverted along altered flow paths and result in base flow augmentation of the Paw Paw River. Table Annualized Diverted Flow Volumes (Return Flow - Wastewater) in the Paw Paw Watershed. Facility City County 88 Annualized Flow (Mgd) Annualized Flow (cfs) Diverted Flow Volumes (%) Lawton WWTP Lawton Van Buren % Paw Paw WWTP Paw Paw Van Buren % Coca-Cola North America Paw Paw Van Buren % Lawrence STP Lawrence Van Buren % Hartford WWTP Hartford Van Buren % Paw Paw Lakes STP Coloma Berrien % Total: % *Mean Annual Discharge for water years with complete monthly datasets Comparing tables and 3.3-5, estimated groundwater withdrawals of 3.56 Mgd for public water supply purposes compare favorably with the Mgd of wastewater that is treated and returned to the Paw Paw River watershed by local wastewater treatment plants. Given that the volume of public water supply groundwater withdrawals is an estimate based on Shiawassee usage data, the correspondence of these numbers is encouraging. It should be noted that all of

13 these waters travel along altered flow paths and the source waters are different from the receiving waters (groundwater to surface water flow augmentation). However, the volume of returned wastewater from local WWTP s is small relative to the volume of receiving waters. The primary source of return flow is the Paw Paw Lakes WWTP which is located just above the USGS gage in the lower reaches of the watershed (Figure ). Wastewater flow volumes are estimated to be 1.57% of the total volume of receiving waters in the Paw Paw watershed. In summary, the groundwater Diversion Ratio (D GW ) for the Paw Paw watershed (V DIVg /V GW ) is 0.94 or 94%. The surface water Diversion Ratio (D SW ) is 1.0 or 100%. The Diversion Ratio (D) for all water withdrawals in the Paw Paw watershed is 0.95 or 95%. The Pathway Alteration Metric (PAM) is the volume of diverted water that travels along altered flow paths relative to the total volume of the receiving waters. PAM = (V DIVg + V DIVs ) / V receiving waters For the Paw Paw watershed, the Pathway Alteration Metric is or 5.38% (Table 3.3-6). This estimate is probably low, as data are not available to assess potential flow path diversions due to agricultural drainage tile and other modifications that alter flow paths on the landscape. Table Estimated Diverted Flow Path Volumes (Total) and Pathway Alteration Metric for the Paw Paw Watershed. Flow Path Diversion Volumes GW (cfs) SW (cfs) Total (cfs) PAM Private Household Irrigation Public Supply Wastewater (wastewater public supply) Stormwater n/a n/a n/a n/a Based on sensitivity analyses of the instream hydrologic assessment tools, it is not likely that changes in hydrology due to flow path diversion will be detected by hydrologic assessment tools used in this project, especially if withdrawal or return flow volumes are relatively low compared to total flow volumes. However, flow path diversions could be locally important, especially in stream reaches that are in close proximity to (and downstream from) outfalls. Flow Duration Curve models: Multiple linear regression models can be used to predict values of specific flow exceedence frequencies (i.e., Q05, Q10, Q25, Q50, Q75, Q90, and Q95). Two separate models were developed using gage data and catchment summaries from a Michigan and from ecoregional datasets (southern Michigan northern Indiana drift plains). These models were applied to catchment summaries for the Paw Paw River at each of the sites in this study. The ecoregional model did not fit well for the Paw Paw River therefore statewide model was selected for further analysis (Figure ). 89

14 0.03 Flow (cms/da) Q5 Q10 Q25 Q50 Q75 Q90 Q95 Recent Flow Exceedence Frequency IntPt 1 IntPt 2 IntPt 3 IntPt 4 IntPt 5 IntPt 6 IntPt 7 IntPt 8 Figure Flow duration curve model results Paw Paw River (MI state models). High flow yields are more variable among sites than median or low flow yields but overall the values are still low. Modeled flow yields using the recent land cover for the Paw Paw River demonstrate a weak and negative relationship between high flow yield and catchment area (r= ). Flow estimates were also made using summaries from a presettlement land cover and a land transformation model for the Paw Paw River. The overall patterns of flow yield at different flow exceedences were similar for all applications of these models (Figures and ). However, the flow magnitudes differed under altered land cover conditions suggesting that flows have been altered since presettlement with yields decreasing at all exceedence frequencies (Figure ). Results of the land transformation model suggest that high flows will become slightly lower while low flows become substantially higher although they will still remain relatively low in magnitude (Figure ). 90

15 Flow (Q/DA) Q5 Q10 Q25 Q50 Q75 Q90 Q95 Presettlement Flow Exceedence Frequency IntPt 1 IntPt 2 IntPt 3 IntPt 4 IntPt 5 IntPt 6 IntPt 7 IntPt 8 Figure Flow duration curve model results Paw Paw River (MI state models presettlement land cover) Flow (Q/DA) Q5 Q10 Q25 Q50 Q75 Q90 Q95 LTM FLow Exceedence Frequency IntPt 1 IntPt 2 IntPt 3 IntPt 4 IntPt 5 IntPt 6 IntPt 7 IntPt 8 Figure Flow duration curve model results Paw Paw River (MI state models landcover from land transformation model). 91

16 0.00 Relative change from presettlement Q5 Q10 Q25 Q50 Q75 Q90 Q95 Flow Exceedence Frequency IntPt 1 IntPt 2 IntPt 3 IntPt 4 IntPt 5 IntPt 6 IntPt 7 IntPt 8 Figure Relative change in flow duration curve model results Paw Paw River (MI state models: recent and presettlement land cover). These models suggest that flows at all exceedence frequencies are lower under current conditions Relative change from recent Q5 Q10 Q25 Q50 Q75 Q90 Q95 Flow Exceedence Frequency IntPt 1 IntPt 2 IntPt 3 IntPt 4 IntPt 5 IntPt 6 IntPt 7 IntPt 8 Figure Relative change in flow duration curve model results Paw Paw River (MI state models: recent and land transformation land cover). Model output suggests that under the conditions of the land transformation model high flows will decrease but low flows will increase within all basins. 92

17 Dams and channel modifications: Land cover modifications are not the only type of anthropogenic change that has occurred within the watershed. Dams and channel modifications are also common within the river. There are twenty-three dams in Paw Paw River in our most downstream catchment (Figure ). Estimated extent of channel modification conducted by identifying reaches that appeared unnaturally straightened ranged from 0 63% for our study areas (Table 3.3-7). Figure Location of dams within the Paw Paw River watershed. Table Summary of dams and channel modifications within Paw Paw River subwatersheds. Paw Paw River subwatershed Watershed area (mi 2 ) Dams (total#) Channel modification (% of upstream streamlength modified) Paw Riverside, MI ( ) % IntPt % IntPt % IntPt % IntPt % IntPt % IntPt % IntPt % IntPt % 93

18 Hypotheses about affects of anthropogenic changes on flow Alterations to the ability to retain (e.g., dams, wetland removal) and transport water (i.e., channelization) have consequences to the observed hydrologic regime in rivers. Changes in catchment storage can influence the rate of change in flow (i.e., Flashiness Index) as well as most high flow metrics. Stream channelization is designed to increase the rate of movement of water and therefore is expected to increase high flow magnitudes (Table 3.3-8). Table Hypothesized responses of flow metrics due to channel modifications. Flow component Summer/Fall Magnitude Spring Magnitude High Flow (event) Magnitude Flow metric July, Aug, Sept, and Oct median flow March, April, May, and June median flow 3-Day annual high flow Anticipated change due to channelization Increase (higher flows during dry season) Increase (higher flows during wet season) Increase (higher flows) Explanation Channelization increases drainage efficiency. More precipitation is routed to the stream network, increasing dry-season flows. Channelization increases drainage efficiency. More precipitation is routed to the stream network. Efficient drainage concentrates flow and magnifies peakflow. Rate of Change Flashiness Index Increase Channelized reaches increase responsiveness of the stream network. Results of hydrologic assessment tools Indicators of Hydrologic Alteration: In the Paw Paw River, one- and two-period analyses were conducted. The one-period analysis was used to detect trends rather than to detect changes attributable to a specific event in time. The two-period analysis describe conditions before and after a discrete event (e.g., dam construction, change in dam operation) when the period of record is sufficiently long or to describe and compare historical and recent period snapshots. In the Paw Paw River only a single gage had sufficient information to run these analyses. For the Paw Paw River, the magnitude of seasonal flows and high and low flow metrics increased during the period of record. Changes to monthly flow magnitudes were categorized as increasing or decreasing using the slope of the regression line fit to the monthly summaries for each season (Table 3.3-9). 94

19 Table Summary of changes to flow metrics based on the IHA analysis for the Paw Paw River at Riverside, MI (USGS Gage ). Summary of changes to seasonal magnitude Changes to flow events High flows (3-day max) Gage ID Period of analysis Winter Spring Summer/ Fall Low flows (7-day min) Paw Paw River at Riverside, MI ( ) increase increase increase increase increase Paw Paw River at Riverside, MI ( ) increase increase increase increase increase Flow/Precipitation Ratio: The Flow/Precipitation (Q/P) ratio is a measure that normalizes watershed yield by precipitation. Changes in Q/P over time indicate that some factor other than precipitation volume is changing the amount of water running off of the landscape and/or through the river network. This analysis was performed using flow data from the Paw Paw River at Riverside (USGS gage ) for the period of 1951 to Mean annual precipitation was observed to increase only slightly during this time (r=0.12) with an average value of 37.3 inches/year (Figure ). Average Annual Precipitation (inches) Year Figure Average Annual Precipitation in Paw Paw, Michigan (from NOAA Weather Station ). A two period analysis of average monthly precipitation reveals lower precipitation in the winter months (December-March) and generally higher precipitation during most other months during the more recent time period (Figure ). 95

20 AvgP ( ) 0 AvgP ( ) JAN FEB MAR APR MAY JUN JUL AUG SEPT OCT NOV DEC Figure Two-Period Analysis of Average Monthly Precipitation in Paw Paw, Michigan (from NOAA Weather Station ). A similar analysis of water yield (Q) shows a nearly uniform increase between the early and recent periods for all months at the Paw Paw River gage. The average annual Q for the early period (1951 to 1973) was 1.06 cfs/sq mi and the average annual Q for the later period (1974 to 1996) was 1.30 cfs/sq mi (Figure ) AvgQ ( ) AvgQ ( ) JAN FEB MAR APR MAY JUN JUL AUG SEPT OCT NOV DEC Figure Two-Period Analysis of Average Monthly Yield (cfs/sq mi) for the Paw Paw River. The Q/P ratio is generally higher in the late fall and winter months of the more recent time period (0.87) than the earlier period (0.57) within the Paw Paw River (Figure ). Q/P ratios are similar or slightly higher in the recent period during all months except March and September 96

21 suggesting that the watershed has become more efficient at moving water to the channel network Avg Q/P Avg Q/P JAN FEB MAR APR MAY JUN JUL AUG SEPT OCT NOV DEC Figure Two-Period Analysis of Monthly Q/P for the Paw Paw River. Flashiness Index: The R-B Flashiness index (RBI) is a metric used to quantify the frequency and rapidity of short term changes in streamflow. Daily streamflow was used to calculate the RBI value for each year. Annual values were plotted over time to identify long term trends in flashiness (Figure ) RB Flashiness Index Figure Annual Richards-Baker Flashiness Index values at USGS gage ( ) on the Paw Paw River at Riverside, MI. 97

22 While the overall flashiness for the Paw Paw River was low (approximately 0.05) a slightly decreasing trend was observed over the period of record. Mean values for the early period ( ) were slightly more than 4% higher than those for the recent period ( ). This is partially driven by the period of extremely low flashiness. However, if we compare the periods (mean annual flashiness of ) and (mean annual flashiness of ) we observe an increase in flashiness in the more recent period of nearly 5 percent (4.8%) that is consistent with this trend. Synthesis Hydrologic summaries of low flow magnitudes in the Paw Paw River were consistent between the IHA gaged records and the modeled flow duration curves (Table ). Therefore, either of these methods could prove useful in an evaluation of low flows in this basin. However, high flow summaries were not positively correlated suggesting that the flow duration models did a poor job estimating high flow magnitudes. Our power metric was more successful at describing trends in high flows than in low flows for the Paw Paw (Table ). Therefore, it may be necessary to use multiple tools to fully describe the hydrologic regime were stream flow records are not available. By combining the low flow analysis of the MLR models and the high flow components of the Stream power metric a reasonably complete picture of the watershed can be developed. 98

23 Table Expected relationships between measured metrics (from USGS gage ) and modeled metrics for the Paw Paw River. Flow characteristic Flow metric(s) Flow exceedance frequencies Low flow magnitude Q95 Q90 Q75 High flow magnitude Q25 Q10 Q5 Change in stream power High flow magnitude Stream Power Metric Indicators of Hydrologic Alteration Other Gaged or Measured Metrics Summer/Fall Magnitude Spring Magnitude Low flow (event) magnitude High flow (event) magnitude Expected Correlation Vs. Actual Correlation July, Aug, Sept, and Oct median flow March, April, May, and June median flow Expected / Actual + / + 7- day annual low flow + / + Expected / Actual Expected / Actual + / - + / + 3-Day annual high flow + / - + / + Low Flow Magnitude Base Flow Index + / + - / + Rate of Change Flashiness Index + / + Watershed Efficiency Metrics Not Comparable Data Supports Hypothesis +/+ Yield versus Precipitation Ratio (Q v. P) Data Refutes Hypothesis "+/o" Data Refutes Hypothesis and Shows Opposite Relationship +/- + / + General Findings of our analysis include: Comparisons to presettlement landcover suggest that agriculture has replaced forested lands, and particularly wooded wetlands, in much of the watershed. Landcover change can lead to increased runoff as suggested by the stream power tool. Groundwater is the primary source of water for anthropogenic use representing a major flow path diversion where waters are being removed from one type of hydrologic regime (groundwater) and added to another hydrologic regime (surface water). The flow duration curve model analysis was not always consistent with the other types of analysis. These models are sensitive to changes in the proportion of urban and agricultural landuse in the catchment. However, alteration of wetland or forested lands do not effect the output of these models unless they are converted to urban or agriculture uses. Therefore, these models can predict decreases in high flows with altered landuse that contradict the expectations of our stream power tool or in some cases the IHA analysis. The development of models that are more sensitive to changes in wetland and/or forested lands could help alleviate this problem. 99

24 The IHA analysis demonstrates that the magnitude of seasonal flows and high and low flow metrics increased during the period of record. Mean annual precipitation was observed to increase only slightly during this time (r=0.12) with an average value of 37.3 inches/year (Figure ). We observe an increase in flashiness in the more recent period of nearly 5 percent (4.8%) when we compare the early portion of the record to the recent portion. References Wesley, Jay K., and Joan E. Duffy St. Joseph River Assessment. Michigan Department of Natural Resources, Fisheries Division, Special Report 24. Ann Arbor, Michigan. Pijanowski, B.C., S.H. Gage, D.T. Long & W. C. Cooper A Land Transformation Model: Integrating Policy, Socioeconomics and Environmental Drivers using a Geographic Information System; In Landscape Ecology: A Top Down Approach, Larry Harris and James Sanderson eds. Pijanowski, B.C., T. Machemer, S.H. Gage, D. T. Long, W. E. Cooper and T. Edens The GIS syntax of a spatial-temporal land use change model. In Proceedings of the Third Conference on GIS and Environmental Modeling. GIS World Publishers - CD-ROM. 3.4 Milwaukee River, Wisconsin Location: The Milwaukee Watershed drains approximately 883 square miles of eastern Wisconsin to Lake Michigan. Six subwatersheds drain the Milwaukee Watershed and include four Milwaukee River subwatersheds; The North Branch Milwaukee River, the East and West Branches Milwaukee River, Milwaukee River South, Cedar Creek; and also the Menomonee River subwatershed and the Kinnickinnic River subwatershed (Figure 3.4-1). 100

25 Figure Subwatersheds and long-term USGS gage locations of the Milwaukee River Watershed. Geology: The geology of the watershed is dominated by glacial deposits overlaying bedrock. The southern third of the watershed and the eastern most edge consist of clayey till from endmoraine deposits. The northern two-thirds of the watershed is dominated by loamy tills from end- and ground-moraine deposits. Interspersed in this patchwork are ice-contact and outwash sands and gravels. Elevations are highest in the north and west and slope downward to the south and east. Effects of Anthropogenic Changes on Flow Hypotheses Two types of anthropogenic modifications are present within the Milwaukee watershed: land use/land cover change and instream modification. Land cover change may be assessed by examining current and/or future land covers relative to presettlement landcover. However, land cover change only tells a portion of the story. Modern day agriculture often employs tile drainage, irrigation and different management strategies. Urban areas may have different 101

26 sources for municipal water (Lake Michigan, The Milwaukee River, groundwater, out-of-basin sources, etc.) and often combine these sources. Urban areas also employ various strategies to handle municipal waste water and stormwater, i.e., septic systems, sewered areas, and individual property owners and developments may implement practices to retain stormwater while others may not. Therefore, simple examinations of land cover change will not suffice but an analysis of land use is necessary. Further complicating the issue is that the geospatial relationships of these treatments must be considered to understand the full impact to the watershed (See Water Use / Pathways Analysis). Another important hydrologic alteration type is instream modification. These modifications include channel straightening and dredging, channel hardening (levees) that prevent the natural meander of river systems, and dams of various sizes and management purposes. For further discussion of how watershed changes affect flow hypotheses see Section 2.4. Data Availability: Six USGS gages exist within the watershed that are either currently recording daily streamflows or have recorded them until recently (Figure 3.4-1). For this study, we analyzed streamflow data from these six sites as they had more than 20 years of relatively continuous USGS approved discharge data (Table 3.4-1). Three gages record streamflows along the main stem or upper tributaries of the Milwaukee River and include , and The other three gages record streamflows along the Menomonee or tributaries of the Menomonee and include , , and All analyses done in this report are discussed in terms of the drainages, or subwatersheds, of these six gage locations. Table Milwaukee River Watershed Gages with 20+ Years of USGS Approved Daily Discharge Data. Gage ID Location Name Period of Record Cedar Creek nr Cedarburg Milwaukee River nr Cedarburg Milwaukee Milwaukee Menomonee Menomonee Falls Underwood Wauwatosa Menomonee Wauwatosa Land use/land cover change: The land use and land cover within the watershed is diverse with the northern part of the watershed consisting of predominantly agriculture land use (Dodge, Fond Du Lac and Sheboygan Counties), the middle part of the watershed is also largely agricultural but faces significant growth pressures from the growing Milwaukee metropolitan region and the southern part of the watershed is heavily urbanized with high and medium density development (Figure 3.4-2). Because this watershed s land use/land cover differs appreciably from our other demonstration watersheds it provides an interesting opportunity to examine our hypotheses in an area with different anthropogenic stressors, and potentially, different hydrologic impairments. Presettlement land cover (Figure 3.4-3) was predominantly deciduous forest (~89%) and forested wetlands or open water (~10%). Current land cover within the watershed is primarily agricultural (~39%), wetlands or open water (~14%), forest, (~12%), grassland/herbaceous (~17%), or urban (~17%) (Table 3.4-2). 102

27 Table Presettlement and Current Land Cover Open Water Emergent Herbaceous Wetlands Woody Wetlands Grasslands/ Herbaceous Shrubland Deciduous/ Mixed Forest Evergreen Forest Pre-Settlement Current Change Acres Acres Acres 5,578 8,335 2,757 2,265 23,928 21,663 49,623 47,134-2, ,749 95,260 2,558 6,193 3, ,255 55, ,441-6,081 6,081 % of Total % of Total % of Total 1.0% 1.5% 0.5% 0.4% 4.2% 3.8% 8.8% 8.3% -0.4% 0.1% 17.0% 16.9% 0.5% 1.1% 0.6% 89.2% 9.9% -79.3% 0.0% 1.1% 1.1% Row Crops Pasture/Hay Urban/ Recreational Grasses Low Intensity Residential Commercial/ Industrial/ Transportation Bare Rock/ Sand/Clay No Data Pre-Settlement Current Change Acres Acres Acres - 122, ,211-94,377 94,377-3,130 3,130-29,284 29,284-63,638 63,638-6,596 6,596-2,355 2,355 % of Total % of Total % of Total 0.0% 21.7% 21.7% 0.0% 16.8% 16.8% 0.0% 0.6% 0.6% 0.0% 5.2% 5.2% 0.0% 11.3% 11.3% 0.0% 1.2% 1.2% 0.0% 0.4% 0.4% < 1.0 % change 1.0 % to 5.0 % change > 5.0 % change Much of the presettlement deciduous forests appear to be converted to agriculture, grasslands, or urban uses. Also of note: The current amount of wetlands and open water (~14%) appears larger than that of presettlement (~10%) (Table 3.4-2). This is possibly a result of some presettlement woody wetlands being classified as deciduous forest underscoring the difficulties of comparing two different land cover datasets. Figure Current Land Cover WISCLAND

28 Figure Presettlement Land Cover WDNR Original Vegetation Cover GIS Layer, Watershed Level Assessment Stream Power Tool: Differences in stream power can be presented in two ways (For a full discussion of the tool see section of this report): 1) Using land cover from presettlement and recent time periods, change in stream power can be calculated and compared using a percent power change calculation, i.e., Percent Power Change = Recent Stream Power Presettlement Stream Power Presettlement Stream Power Figure illustrates percent power change from presettlement to current conditions for the six gaged subwatersheds in the Milwaukee watershed. In all gaged watersheds, stream power increased between approximately 130% and 180%. In both the Milwaukee and Menomonee subwatersheds, the greatest increase in stream power occurred in the subwatershed with the smallest area. 104

29 Figure Percent Change in Stream Power at Milwaukee Sites. 2) Calculate the maximum stream power change possible for each stream reach and compare the actual change (from presettlement conditions) to the maximum possible change (under hypothetical conditions). This power change metric provides a quantitative way to estimate hydrologic degradation (or improvement) relative to maximum possible degradation. Power Change Metric = Recent Stream Power Presettlement Stream Power Paved Paradise Power Presettlement Power 105

30 Figure Power Change Metric by Stream Reach Current Conditions. The power change metric increased in all catchments. These increases were higher in the Menomonee subwatersheds than the increases in the Milwaukee subwatersheds (Figure 3.4-5). This highlights the potential implications of normalizing potential hydrologic alteration against a worst-case scenario; i.e., while the Menomonee subwatersheds percent increases in stream power were similar to the Milwaukee s and, in some cases, lower, the power change metric was significantly higher indicating those subwatersheds are potentially more altered than the Milwaukee subwatersheds. The power change metric for each stream reach is shown in Figure

31 Figure Power Change Metric by Stream Reach: Current Conditions. Wetland Retention: Wetland data was gathered from the Milwaukee River Basin Wetland Assessment Project (US EPA Natural Resource Grant # ). This project used the Wisconsin Wetlands Inventory and augmented with locally specific wetlands data where available. Acreages of existing wetlands were calculated within each gaged watershed. Generally wetlands made up a larger percentage of Milwaukee subwatersheds than the Menomonee s (Table 3.4-3). Presettlement wetlands were determined using the assumption that all hydric soils that are not now wetlands were wetlands before European settlement. Thus, presettlement wetlands were determined by adding non-wetland hydric soils to current wetlands. Percentages were calculated for each gaged watershed (Table 3.4-3). 3 Potentially restorable wetlands (PRWs) are all hydric soils that are not currently wetlands and occur on agricultural or rural lands. (PRWs are a smaller subset of presettlement wetland described above. They do not include current wetlands nor do they include any non-wetland hydric soils that are on developed land. They are areas where wetlands likely were historically and hold potential restoration opportunities due to the current land use on which they occur.) 3 Note: In urban areas soil surveys are often incomplete or not conducted. There are large portions of the Milwaukee River watershed in metropolitan Milwaukee where soils data are unavailable and makes estimates of hydric soils difficult. This can lead to underestimation of presettlement wetlands in these areas and may be why some of the Menomonee River subwatersheds have lower percentages of presettlement wetlands. 107

32 For the Milwaukee River watershed they were mapped by the Milwaukee River Basin Wetland Assessment Project and were used in this project s calculations. PRWs represent a high chance for restoration since they are not located in developed areas. Table shows that better chances for PRW restoration occur in the more upstream subwatersheds of both the Milwaukee and Menomonee subwatersheds. Figure shows all PRW s in the Milwaukee Watershed with a close-up of the upper subwatersheds. Table Wetland summary of Milwaukee River Watershed gaged catchments. Current Wetlands Pre-Settlement Wetlands PRWs Gage ID Subwatershed Acres Acres % of Gaged Watershed Acres % of Gaged Watershed % Loss From Pre-Settlement % of Gaged Watershed ,813 14, % 22, % 10.6% 8.2% ,555 71, % 99, % 7.5% 6.0% ,457 74, % 105, % 7.2% 5.5% ,757 3, % 5, % 9.5% 6.3% , % 2, % 15.0% 1.4% ,287 6, % 13, % 8.3% 3.3% Figure PRWs of the Milwaukee River watershed. The team sought to not only examine acreages of current, presettlement and potentially restorable wetlands but also sought to understand the retention associated with these wetlands and what effect that may have on the hydrology of the Milwaukee River watershed. Average retention capacity is calculated by assigning water depth ranges for specific types of wetlands and/or vegetative communities as defined by the USFWS National Wetland Inventory protocols. These depths are then subtracted from the runoff grids calculated by the stream power tool, thus directly affecting the flows and stream power downstream from the wetland. The advantage of this approach is that 1) the relative importance of factors that control wetland water storage and retention in a catchment (number, type, size, and location of wetlands) can be evaluated, and 2) the effects of water storage and retention on overall watershed hydrology can be measured and quantified. 108

33 The team next desired to understand what the effect would be on hydrology from restoring PRWs. At this scale it would be time consuming, if not impossible, to understand what type of wetland each PRW was prior to settlement. Thus, a weighted average of the retention depth was done and assigned to the PRWs. Figure shows retention under existing and restoration of all PRW conditions. Figure Map of existing and possible retention (assuming restoration of all PRWs) To understand what effect restoring all PRWs might have on the hydrology of the Milwaukee River watershed PRWs were spliced into the current land cover used in the stream power alteration assessment. The stream power tool was rerun and the PRW restoration scenario power change metric was compared to current conditions power change metric. Figure summarizes changes in stream power at each gaged watershed as a result of PRW restoration. The percent decrease in stream power was greater in the Milwaukee subwatersheds than in the Menomonee subwatersheds. Decreases in stream power metric did not show a consistent trend. Changes in power generally decreased less in the downstream portion of the watershed; not surprising given fewer PRWs to restore the further downstream in the watershed (Figure 3.4-9). Figure shows changes to the power metric from PRW restoration for each stream reach. 109

34 Figure Decrease in Stream Power from Restoration of PRWs. 110

35 Figure Potentially Restorable Wetlands (PRWs) and Associated Decrease in Power Change Metric from Restoration. Water Use / Pathways Analysis: The potential effects of flow path changes on the Milwaukee watershed were evaluated by considering regional flow patterns resulting from water withdrawal, flow diversion, return flow, and the type or source and receiving waters. In the Milwaukee watershed, Lake Michigan surface water is the primary source of water for the City of Milwaukee and nearby surrounding communities. Groundwater from both deep and shallow wells is the primary source of water in rural areas and/or smaller urban communities in the northern and western portion of the watershed. High-capacity groundwater wells also supply water for irrigation and agricultural purposes. Factors considered in the pathways analysis for the Milwaukee watershed include: Source Waters and Diverted Flows For the Milwaukee metropolitan area, Lake Michigan surface waters are withdrawn and treated at two primary water treatment plants and then distributed to the City of Milwaukee and surrounding communities. Table summarizes Lake Michigan water withdrawals from Lake Michigan for the Milwaukee River watershed. 111

36 Table Summary of Lake Michigan Surface Water Withdrawals for the Milwaukee River Watershed Lake Michigan SW Withdrawl (Mgd) Location Average Maximum Capacity Reserve City of Cudahy City of Milwaukee Linwood Ave n/a n/a n/a Howard Ave n/a n/a n/a City of South Milwaukee Totals: In the Milwaukee River watershed, groundwater withdrawals occur from both shallow and deep aquifers. For the purpose of this project, shallow groundwater sources are defined by producing depths generally less than 60 feet with a reasonable expectation that local surface-groundwater interaction may occur. Deep groundwater sources are defined by producing depths greater than 60 feet with a reasonable expectation that local surface-groundwater interaction will not occur. Groundwater is withdrawn from deeper aquifers along the eastern edge and western portions of the watershed (Figure ). Water is withdrawn from shallower aquifers within the central and northern portions watershed. The potential for flow path diversion (flow augmentation) is highest where deep groundwater is diverted and returned to surface waters. In the Milwaukee River watershed there are more than 11,433 producing groundwater wells of which more than 9,000 are private household wells (self supply) where groundwater is withdrawn on an as needed basis for household use (Figure ). High capacity wells (> 70 gpm) and non-municipal private water supply wells also provide drinking water for local residential and industrial water supply systems. Flow and location data are not currently available for many these wells. The Wisconsin DNR is implementing new regulations (effective 2007) that mandate annual reporting of water withdrawals for all active groundwater wells. Flow meters will be required on all high capacity groundwater wells after October 1,

37 Figure Spatial distribution of groundwater wells by type in the Milwaukee watershed. Private wells include household (self supply) and low capacity non-potable (self supply) wells for residential or light industrial purposes. High capacity wells may supply water to private residential communities, light industry, or for irrigation or agricultural purposes. Approximately 80% of the groundwater wells in the Milwaukee River watershed are used to supply groundwater for private household (self supply) use. 113

38 Figure Spatial distribution of shallow and deep groundwater wells in the Milwaukee watershed. Water is produced from deeper aquifers (yellow, orange, and red symbols) in areas adjacent to the Lake Michigan shoreline and of the watershed. The Milwaukee metropolitan area is serviced by surface water withdrawn from Lake Michigan (area outlined in yellow). There are more than 46 high capacity municipal wells that provide drinking water to local communities where groundwater is collected, treated, stored, and then distributed through a 114

39 public water supply system. In these communities, wastewater is collected and treated in a local wastewater treatment plant before being returned to a local tributary or the Milwaukee River. These waters are diverted along an altered flow path from a groundwater source into surface receiving waters. Currently, water withdrawals from private household wells are typically not reported, but can be estimated based on per capita household water use for the Great Lakes region (Great Lakes Commission, 2001). Using an estimated average water use of ~230 gpd per household, it is possible to estimate the volume of groundwater withdrawn for domestic (self supply) purposes. The values reported for domestic (self supply) in Table are based on estimated average water use per household. There are no separate databases for industrial or irrigation wells by watershed, and the reported data are somewhat problematic due to inconsistent well classifications. As a surrogate, the 1995 USGS estimate for industrial and irrigation water use in the Milwaukee River watershed is used for this analysis. Similar problems exist for the municipal groundwater wells; data is scattered across several different databases and classifications appear to be inconsistent. As a surrogate, average wastewater return flow is used to estimate the volume of groundwater withdrawals for municipal drinking water purposes. Surprisingly, the municipal groundwater volumes are virtually identical to those estimated by the USGS in Groundwater wells are summarized below by type, estimated producing depth, and potential for flow path diversion. Table Estimated Groundwater Withdrawals by Well Type and Producing Depth for the Milwaukee watershed. Producing Depth ( # Wells) GW Withdrawl (Mgd) Well Type No. Wells < 60 feet > 60 feet n/a < 60 feet > 60 feet n/a Total Domestic (self supply) Industrial/Irrigation n/a n/a n/a n/a n/a n/a n/a 4.01* Municipal ** Totals: Low Flow Diversion Potential High Flow Diversion Potential * 1995 Industrial/Irrigation ** GW sourced wastewater Annual domestic, municipal, industrial, and irrigation withdrawals are generally reported on a county-by-county basis and are not available specifically for the Milwaukee River watershed. In 1995, the USGS published estimates of water usage not only for individual counties, but for 8- digit HUC watersheds as well. Table summarizes reported water withdrawals (by the USGS) for the Milwaukee River watershed for the year Table Reported Water Withdrawals for the Milwaukee River Watershed in 1995 (USGS). Municipal Domestic (Self Supply) Industrial Irrigation Year SW GW Total SW GW Total SW GW Total SW GW Total Total 1995 (Mgd) % by Type 91% 9% 100% 0% 100% 100% 60% 40% 100% 0% 100% 100% % Total 86% 8% 94% 0% 1% 1% 3% 2% 5% 0% 0% 0% Based on these data, water withdrawn from Lake Michigan provides ~90% of the total volume of municipal drinking water to communities in the Milwaukee River watershed. The remaining 10% is provided by groundwater sources. Domestic (self supply) groundwater accounts for less than 1% of the total volume of water withdrawn from the watershed for drinking water purposes. Industrial uses (non power generation) account for ~5% of the total volume with 60% of those 115

40 waters coming from surface water sources (primarily Lake Michigan). Irrigation is a minor component of water use in the Milwaukee River watershed. With continuing population growth and development, the proportion of source water withdrawals remains generally the same through time. Table illustrates changes in surface and groundwater usage for three Wisconsin counties every 5 years from 1979 through the year All, or a portion of these counties are within the Milwaukee River watershed. Even though there is some variability, the general surface water/groundwater proportion remains the same, even though groundwater usage appears to be decreasing in Milwaukee County and increasing slightly in Washington County. Table Reported Water Withdrawals for Milwaukee River Watershed Counties (USGS) (Mgd) 1985 (Mgd) 1990 (Mgd) 2000 (Mgd) County SW GW Total SW GW Total SW GW Total SW GW Total Milwaukee Ozaukee Washington Based on the data presented in Table 3.4-5, approximately Mgd of groundwater is withdrawn for residential (self supply), irrigation, and municipal water supply uses in the Milwaukee River watershed. Approximately Mgd of groundwater withdrawn from the Milwaukee River watershed is diverted along altered flow paths where groundwater is returned either to surface waters or to a distinctly different catchment or subwatershed from where the groundwater originated. The groundwater Diversion Ratio (D GW ) for the Milwaukee watershed (V DIVg /V GW ) is 0.93 or 93%. The surface water Diversion Ratio (D SW ) is 0.0 (V DIVs /V SW), as all withdrawn Lake Michigan waters are returned to Lake Michigan via the MMSD water treatment plants. The Diversion Ratio (D) for all water withdrawals in the Milwaukee watershed is or 12.5%. Even though 93% of the groundwater withdrawn from the watershed is returned to surface waters and are considered to be altered flows, the large Lake Michigan withdrawals dwarf total groundwater withdrawals and with groundwater representing only 13.4% of the total volume of water withdrawn in the basin. Return Flows and Receiving Waters Virtually all of the wastewater within the Milwaukee Metropolitan area is collected, treated, and returned to Lake Michigan via the MMSD South Shore and/or Jones Island wastewater treatment plants. Most of the wastewater treated by these plants is derived from Lake Michigan surface waters and/or surface runoff that would eventually flow into Lake Michigan. For the purpose of these analyses, waters withdrawn from Lake Michigan and then returned directly to Lake Michigan are not considered to be diverted waters. Lake Michigan waters are not returned to the Milwaukee River (except during major storm events via CSO s). In communities or areas with public or private sewerage not served by the MMSD (primarily watershed areas north of the Milwaukee metropolitan area), source waters are primarily from groundwater sources. In these communities, wastewater is collected, treated, and then returned to surface waters augmenting surface water flows in tributaries to the Milwaukee River. Table summarizes return flow volumes from wastewater treatment plants for the northern portion of the Milwaukee watershed and return flow volumes to Lake Michigan surface waters for the Milwaukee metropolitan area. Approximately one-third of MMSD s service area is located outside of the Milwaukee River watershed (i.e. the Menomonee Watershed - Figure

41 13). Adjusting for area, total wastewater volumes derived from the Milwaukee River watershed are estimated to be Mgd. Table Wastewater Treatment Facilities in the Milwaukee River Watershed Facility City County Existing flow (Mgd) Max Flow (Mgd) GW - SW (Milwaukee River) Campbellsport STP Campellsport Fond Du Lac Scott SD #1 Adell Sheboygan Kewaskum STP Kewaskum Washington Fredonia WWTP Fredonia Ozaukee West Bend WWTP West Bend Washington Saukville STP Saukville Ozaukee Jackson STP Jackson Washington Grafton WWTP Grafton Ozaukee Cedarburg WWTP Cedarburg Ozaukee Total: SW - SW (Lake Michigan) MMSD Jones Island Milwaukee Milwaukee MMSD South Shore Oak Creek Milwaukee Total: * * Mgd Adjusted - 33% from outside watershed Major wastewater treatment plants Figure illustrates the spatial distribution of wastewater treatment plants and NPDES industrial outfalls in the watershed. The Jones Island and South Shore WWTPs collect and treat wastewater within the general MMSD service area (black box) which is then returned to Lake Michigan. North of the MMSD service area, source water (primarily groundwater) is diverted along altered flow paths into the Milwaukee River resulting in flow augmentation. Significant sources of wastewater are the West Bend and Cedarburg WWTPs. The Milwaukee River gage near Cedarburg ( ) measures discharge in the Milwaukee River downstream from the Cedarburg WWTP before entering the MMSD service area. Average flow from 1983 through 2006 at the Milwaukee River gage near Cedarburg was cfs. Pathways metric calculations for the Milwaukee River are normalized to flows at this gage. Downstream from this gage, source waters are diverted from Lake Michigan and then returned to Lake Michigan via the MMSD South Shore and Jones Island wastewater treatment plants. Within the MMSD service area, most industrial discharges occur along the Milwaukee and Menomonee Rivers. Direct industrial discharges flowing into the Rivers are minimal as many of these discharges go through a pre-treatment process to reduce and/or eliminate pollutant loadings before entering the MMSD water treatment system. However, flow augmentation may occur due to direct discharges that occur into the Milwaukee or Menomonee Rivers during major storm events via combined sewer overflows (CSOs). 117

42 Figure Location of wastewater discharges in the Milwaukee watershed. The general MMSD service area is outlined by the black box. Waters are returned to Lake Michigan via the Jones Island and South Shore WWTPs. The USGS gage located on the Milwaukee River near Cedarburg is the master gage for these pathways analyses. In private systems (primarily households), wastewater is processed in a septic system and returned to shallow aquifers, generally less than 15 feet below the surface. 118

43 Table lists annualized diverted flow volumes (return flow - wastewater) for public and industrial wastewater treatment plants in the Milwaukee watershed. Table Annualized Diverted Flow Volumes (Return Flow - Wastewater) in the Milwaukee Watershed (not including Lake Michigan Surface Water Withdrawals) Facility City County Annualized flow (Mgd) Annualized Flow (cfs) Diverted Flow Volumes (%)* Campbellsport STP Campellsport Fond Du Lac % Scott SD #1 Adell Sheboygan % Kewaskum STP Kewaskum Washington % Fredonia WWTP Fredonia Ozaukee % West Bend WWTP West Bend Washington % Saukville STP Saukville Ozaukee % Jackson STP Jackson Washington % Grafton WWTP Grafton Ozaukee % Cedarburg WWTP Cedarburg Ozaukee % Total: % *Mean Annual Discharge for water years Comparing Tables and 3.4-9, estimated groundwater withdrawals of Mgd for municipal water supply purposes (in 1995) compare favorably with the Mgd of wastewater estimated to be treated and returned to the Milwaukee River watershed by local WWTPs. It should be noted that all of these waters travel along altered flow paths and are returned to a different hydrologic regime (groundwater to surface water flow augmentation). However, the volume of returned wastewater from local WWTPs is small relative to the volume of receiving waters. The primary source of return flow is the West Bend and Cedarburg WWTPs which are located upstream from the USGS Milwaukee stream gage near Cedarburg. Wastewater flow volumes are calculated to be 5.53% of the total volume of receiving waters for area of the Milwaukee River watershed not serviced by MMSD (Table 3.4-9). In summary, the groundwater Diversion Ratio (D GW ) for the Milwaukee watershed (V DIVg /V GW ) is 0.93 or 93%. The surface water Diversion Ratio (D SW ) is 0.0 as all Lake Michigan surface water withdrawals are returned directly to Lake Michigan by the MMSD Jones Island and South Shore wastewater treatment plants. The Diversion Ratio (D) for all water withdrawals in the Milwaukee watershed is or 12.5%. The Pathway Alteration Metric (PAM R receiving waters) is the volume of diverted water that travels along altered flow paths relative to the total volume of the receiving waters. PAM R = (V DIVg + V DIVs ) / V receiving waters For the Milwaukee watershed, the Pathway Alteration Metric (receiving waters) is or 7.17 %. Table illustrates the primary flow components that contribute the Pathway Alteration Metric. In the case of the Milwaukee River watershed, groundwater municipal water supplies are withdrawn primarily from deep aquifers and returned to tributaries of the Milwaukee River as surface water. 119

44 Table Estimated Diverted Flow Path Volumes (Total) and Pathway Alteration Metric for the Milwaukee Watershed Flow Path Diversion Volumes GW (cfs) SW (cfs) Total (cfs) PAM Domestic (self supply) Industrial/irrigation Municipal Wastewater (wastewater public supply) Stormwater n/a n/a n/a n/a Flow depletion (source water) values were not calculated for the Milwaukee River watershed due a lack of available groundwater source water data summarized by individual watersheds. In general, flow path impacts (that are measurable) typically occur at the surface and are associated with return flows and/or receiving waters. Based on sensitivity analyses of the instream hydrologic assessment tools, it is not likely that changes in hydrology due to flow path diversion will be detected by instream tools, especially if withdrawal or return flow volumes are relatively low when compared to total flow volumes. However, flow path diversions could be locally important, especially in stream reaches that are in close proximity to (and downstream from) outfalls. Multi-Linear Regression Flow Duration Curves: The multi-linear regression flow duration curve (MLR-FDC) analysis shows all high flow magnitudes are decreasing in the Milwaukee River subwatersheds from presettlement to now (Table ). It shows low and medium flows increasing, some by a large proportion, in the Menominee subwatersheds. Low flows are decreasing in USGS and , and there is no change in low flows at from presettlement to now (Figure ) Table Summary of MLR-FDC Showing Change from Presettlement to Current Conditions (Values represent proportion change relative to averaged presettlement and current flows.) exceedence flows in cms Relative Percent Change (Recent-Presettlement)/((Recent+Presettlement)/2) Summary of changes to high flow magnitude (Q10): Summary of changes to low flow magnitude (Q90): Site Q5 Q10 Q25 Q50 Q75 Q90 Q95 High flows Low flows Milwaukee River (WI) decrease decrease Milwaukee River (WI) decrease decrease Milwaukee River (WI) decrease no change Milwaukee River (WI) decrease increase Milwaukee River (WI) decrease increase Milwaukee River (WI) decrease increase 120

45 Figure Map Showing MLR-FDC Results by USGS Gaged Watershed. Dams and Channelization: Instream modifications can greatly affect the fundamental characteristics of flow. Landscape tools developed or evaluated do not consider these modifications, and the instream tools do not explicitly consider them in their analyses of gage records. The team attempted to quantify these instream modifications to improve overall understanding of the hydrologic processes occurring within the watersheds and to use this information to better explain results from other analyses. The highest percentage of instream channelization occurred in the Menominee subwatersheds with approximately 60 to 70 percent of all reaches channelized. These subwatersheds, however, had few dams. (A simple way to examine the impact of dams is to divide the total acres in a watershed by the number of dams a dam density. ) In the Menomonee subwatersheds there was approximately one dam for approximately every 20,000 acres (Table and Figure ). The Milwaukee River subwatersheds have less channelization (only 20 to 30 percent of all reaches channelized) but a higher number of dams. In these subwatersheds there was approximately one dam for approximately every six to seven thousand acres (Table and Figure ). 121

46 Table Dams (WI DNR Bureau of Watershed Management, 2002) and Channelized Reaches (Derived from visual inspection of NHD and DRGs - reaches with that appeared unnaturally straightened were coded as channelized. ) Location Name Dams (total#) In-stream (% channelized) Area (acres) Acres/Dam sm % 75,813 5, lg., 1 unknown size, 38 sm % 375,555 6, lg., 2 unknown size, 39 sm % 431,457 7, sm % 20,757 20, % 10, sm % 76,287 19,072 Figure Dams in the Milwaukee River Watershed. Assessment of Hydrologic Alterations Indicators of Hydrologic Alteration: Monthly flow magnitude trend metrics from the IHA were evaluated for the Milwaukee over the period of record for each gaged watershed. These metrics were then summarized by season; winter (November, December, January and February), spring (March, April, May, and June) and summer/fall (July, August, September, and October). 122