Project: Hydrological Impact of Beaver Habitat Restoration in the Milwaukee River Watershed

Executive Summary 
Objectives of this project are to (1) assess the potential of beaver reestablishment in the Milwaukee River watershed through GIS modeling and through habitat assessment field surveys; and (2) conduct hydrological modeling to evaluate the potential impacts of beaver constructed dams on river hydrograph processes and flood mitigation in the watershed. ​

The Beaver Restoration and Assessment Tool (BRAT) was adapted for this project to estimate the likelihood of beaver dam building activity and beaver dam capacities in the Milwaukee River watershed, based on GIS analysis of the stream network, vegetation cover, and stream power under baseflow and high-flow conditions. BRAT model simulation suggested that hydrologic conditions in the Milwaukee River watershed are favorable for beavers to establish colonies, as the landscape is generally flat and river slopes are mild throughout most of the watershed. Riparian vegetation type is the primary factor that determines the potential of beaver habitat restoration. The three northern subwatersheds including the East-West branch Milwaukee River, the North branch Milwaukee River and the Cedar Creek, are more suitable for beaver restoration. Model predicted maximum beaver dam capacities are greater than 6 dams/km on average in the three sub-basins. The dam capacities are about 5, 4, and 1 dams/km for Menomonee River, Milwaukee River South, and Kinnickinnic River subwatersheds, respectively. Model predicted dam capacity in this report should be interpreted as a measure of relative importance, since it has not been calibrated with field observations in riverscapes that are similar to the Milwaukee River watershed. 

Hydrological processes in the Milwaukee River watershed, including soil infiltration, groundwater storage, evapotranspiration, baseflow, and stream flows, are simulated by a distributed continuous hydrologic model, HEC-HMS. The model was calibrated for the watershed with stream flow data from USGS streamgages. With the calibrated HEC-HMS model, hypothetical analysis was conducted to evaluate hydrologic impacts of beaver dams. Locations of beaver dams were identified based on BRAT model results and validated through field surveys. 52 beaver restoration sites were selected representing those with the highest potential for beavers in five subwatershed (not including Kinnickinnic River). These beaver dams were included in the model with four stages with progressively increased dam numbers and dam heights, and the total potential ponding area varied between 777 acres (18 dams in Stage 1) and 3,793 acres (52 dams in Stage 4). Simulation results with and without dams were evaluated at 8 observation locations, including outlets of the five subwatersheds, and river cross sections in three urban river flood zones (in Thiensville, Brown Deer and Glendale). 

Simulation with realistic past storm events suggested that beaver dams can significantly reduce flood flows at 8 observation locations. The peak flow rates were reduced by 6% ~ 48%, and flood flow volumes were reduced by 14% ~ 48%, depending on the development stages of beaver dams, and actual storm characteristics. Two factors contribute to peak flow reduction: (1) flow interception by storage capacity of beaver dams makes the primary contribution; and (2) energy dissipation through dam overflow when the storage capacity is filled. Water evaporation from the impounded water is the primary loss that contributes to discharge volume reduction. Model simulations also indicated that most beaver dams were near their full capacity before the occurrence of major storms, due to water accumulation through prior flow events. Therefore, despite the vast disparity in potential storage among different beaver development stages, the effects the total effective storage capacity may not be significantly different before a major 2 storm. As beaver development stage changed from Stage 1 to Stage 4, the flood mitigation effects increased only slightly (about 5% for peak flow reduction, and 3% for volume reduction on average). 

Ten synthetic frequency storms were generated for simulation, they are standard 6-hour and 24-hour storms with recurrence intervals ranging from 10 years to 200 years. Total precipitation depth of these storms varied between 2.99 and 7.44 inches. Since synthetic storms were designed with a uniform spatial distribution over the entire watershed, all beaver dams were able to contribute to flow reduction at river reaches at the lower end of the watershed. Consequently, more significant flow reductions were reported at the eight observational locations. At Stage 1, average flood peak reduction ranged between 26% (24-hour 200-year storm) and 37% (6-hour 10-year storm). At the full Stage 4, the range of average peak reduction was 36% to 46%. 

Modeling analysis with both realistic past storms and synthetic frequency storms approved the hypothesis that beaver dams that are largely dispersed in the upper tributaries of the watershed can potentially mitigate flood flows in urban flood zones at the lower end of the watershed. Considering flood zones in the northern urban area of the Milwaukee County, modeled beaver dams could have reduced the peak flow by 7~40% according to the past storm simulations, and by 25~50% according to synthetic storm simulations. 

Another question this project sought to address was whether or not the Milwaukee River Watershed could reasonably support a healthy beaver population. A field team was assembled to conduct a Basin-wide habitat assessment that would indicate if there exists sufficient space and forage to support reintroduced beaver pairs and their offspring. A set of criteria for beaver reestablishment was determined based on (1) water depth; (2) access to adjoining wetlands; (3) existing forage of diverse aquatic plants and woody materials; and (4) potential flooding conflict with infrastructure. Following these criteria and BRAT model results, potential sites for restoration were identified and assessed through reviewing aerial images and follow-up field visits. 

Of the 163 sites visited throughout the Basin, 85 were ranked with moderate to high potential to support beaver reintroduction. From these sites, 52 were selected as having a high potential to reduce downstream flooding and/or lower the hydrograph of the streams during rain events if beavers were to construct dams and establish ponds at these sites. The field team also identified 14 sites that exhibit high potential to immediately support reintroduced beaver pairs. In addition to conducting field visits, the team used research established calculations to estimate the beaver carrying capacity of each of the six subwatersheds within the 89,000 acres of wetland in the Milwaukee River Basin. Based on these calculations, the Basin has the potential to support as many as 4,563 beavers in 840 colonies, indicating that the Milwaukee River Basin has sufficient wetland habitat to support the reintroduction of beavers. 

Finally, the field team put together a set of recommendations for successfully reintroducing beavers into the Basin, including policy changes, habitat enhancements, educational opportunities, land acquisition partners, conflict management opportunities, and wildlife biologist partners.

To read more, access the rest of the study below.