Purpose and Scope

The purpose of this report is to describe the hydrogeomorphic setting, hydrodynamic conditions, and results from the hydrodynamic simulations of the Sheboygan River mouth as they relate to rehabilitation associated with cleanup of the AOC. Field data were collected primarily in 2011 through 2013 during a particularly low water-level period for Lake Michigan. For the hydrogeomorphic history, the report describes the postglacial geomorphic setting, historical Lake Michigan water-level fluctuations, the interaction of lake levels with river flows, trends in Sheboygan River streamflow, and changes in the planform and sedimentation history of Wildwood Islands complex. The report includes velocity patterns and water-surface elevations from 2D simulations of interacting river flows and Lake Michigan water levels, including average water levels and river flows, high water levels and varying river flows, low water levels and varying river flows, and typical summer low and high seiche conditions for July 2012.

Description of the Study Area

The Sheboygan River empties into western Lake Michigan at the city of Sheboygan in Sheboygan County, Wisconsin. The river drains 1,603 square kilometers (619 square miles) of mainly agricultural land including pasture land, cropland, and vacant fields (Wisconsin Department of Natural Resources, 1995), with secondary land cover of nonagricultural lands consisting of open water, woodlands, wetlands, and parks. Small amounts of the drainage area have urban land cover. The topography is determined by the surrounding glacial setting of flat moraine and kettle moraine with about 30 meters (m) of relief. Clayey soils are dominant in the drainage basin and contribute to flashy runoff characteristics. The river has six impoundments greater than 4 hectares (Wisconsin Department of Natural Resources, 2001). Polychlorinated biphenyls contaminated the lower 22 km of the river and contributed to the poor water quality status in the lower Sheboygan River (Wisconsin Department of Natural Resources, 2001). The study area encompassed a 5-km reach of the lower Sheboygan River located 1.8 km upstream from Lake Michigan. The well-developed river valley at the Sheboygan River through this reach contained terraces that related to relict shorelines associated with pre-Holocene stands of glacially affected Lake Michigan (Richardson, 1991; Carlson and others, 2011).

Hydrogeomorphic History

Data used in the description of the hydrogeomorphic history of the Sheboygan River mouth included Lake Michigan water levels, previously reconstructed paleolake levels, Sheboygan River streamflows and streamflow trends analyses, analyses of overlays of historical aerial photographs to quantify planform changes in the Wildwood Islands complex, and sediment cores along transects in the Wildwood Islands complex for describing depositional history of the islands.

Hydrodynamic Conditions, 2011–13

Hydrodynamic conditions for 2011 through 2013 were examined through water-level monitoring with stage recorders and time lapse cameras. Topographic and water surface profiles were constructed. Bathymetry and velocity data were collected, along with streambed substrate particle size.

Hydrodynamic Simulations of Varying River Flows and Water Levels

Spatial variations in river velocity and water levels for the Sheboygan River in the vicinity of the Wildwood Islands complex under a range of river flows and lake levels were simulated with the USGS FaSTMECH 2D quasi-steady-state hydrodynamic model and the iRIC (International River Interface Cooperative) graphical user interface tool, version 2.3 (Nelson and others, 2006, 201634; International River Interface Cooperative, 2016). Data used for hydrodynamic simulations came from multiple sources and included the topography, bathymetry, river discharge, river water levels, and lake water levels compiled and collected for this study. Model-related data and outputs are published in Westenbroek and others (2023).

The 2D hydrodynamic model FaSTMECH was run for multiple scenarios of combinations of river discharge and lake levels. The selected combinations are examples of how lake levels, seiche magnitude, and river discharge interact to affect river currents. Like most numerical hydrodynamics models, FaSTMECH requires elevation data to define a computational grid. Data used in the elevation grid development were derived from a combination of sources discussed previously:

Despite the large number of bottom elevation measurements obtained during ADCP, RTK–GPS, and multibeam sonar surveys, the bathymetry had to be interpolated in some crucial areas (such as near the Wildwood Islands complex). In some cases, aerial photographs were used along with measured bottom elevations to estimate bottom elevations in areas of low data density. The gridding process supported within the iRIC software creates a continuous curvilinear grid of bottom elevations suitable for use with FaSTMECH; “templates” are used to control the amount of averaging, with separate rules for selecting data points depending on their location (upstream/downstream, side-to-side) relative to the point for which bottom elevation is to be estimated. The template interpolation method is based on a nearest neighbor approach that utilizes a template with a user-defined width and length, where the length follows the local curvature of the grid. Interpolated values are assigned the inverse distance weighted average of all measured points in the template. The template can be set to expand in size until points are determined.

The grid used in this work was created by defining a “grid creating condition,” which is essentially a line following the river centerline, extending from the Interstate Highway 43 bridge upstream to the 14th Street bridge downstream (fig. 1). The grid was defined by specifying a 100-m buffer on either side of the river centerline; this meandering area was then subdivided into 101 elements cross-stream and 1,602 elements stream-wise. This division made the grid resolution approximately 2 m in the cross-stream and 2.5 m in the stream-wise directions.

In addition to the computation grid, hydrodynamic models require definition of upstream and downstream boundary conditions. For this modeling effort, upstream discharge boundary conditions were generated from the discharge record at the Sheboygan River streamgage (fig. 1). Downstream boundary elevations were derived from interpolations from National Oceanic and Atmospheric Administration stage data at Kewaunee, Wis. (National Oceanic and Atmospheric Administration, 2024), water-level data at the New Jersey Avenue, and statistics associated with historical Lake Michigan water levels.

FaSTMECH allows the user to specify the hydraulic roughness in the channel by either providing a roughness length or a drag coefficient. Additionally, the roughness can be spatially distributed or assigned a constant value across the modeling domain. Streambed substrate size descriptions were initially used to calculate a spatially distributed roughness length. Because the model proved unstable with a variable roughness length, this approach was abandoned. Alternatively, the roughness was specified with a constant drag coefficient value. Varying model parameters like the drag coefficient allows the user to evaluate the sensitivity of the hydrodynamic model. The model was run for a range of drag coefficient values. Ultimately, a constant drag coefficient value of 0.0035 provided the best fit between modeled and observed water levels at the stage recorder at New Jersey Avenue downstream from the Wildwood Islands complex and was used in the subsequent model runs.

The FaSTMECH manual gives a rule-of-thumb for selection of the initial lateral eddy viscosity (LEV) using the following equation:

For the Sheboygan River, reasonable values are 0.5 meter per second (m/s) and 1 m, respectively, or an LEV of 0.005. To achieve model stability, iRIC was allowed to start the solution with an LEV value of 0.05; starting with iteration 20, this value was ramped down to a final value of 0.005 at iteration 1,000. Because the maximum number of iterations was capped at the default of 600, this amounts to use of an LEV value of about 0.023. Spot-checking of velocity magnitude and water-surface elevation at key locations for a range of LEV values indicates minimal local changes (less than 0.1 m/s and less than 0.03 m).

Similarly, modeled convergence was generally between 8 and 14 percent. Lower convergence values (on the order of 4 to 7 percent) could be obtained by increasing the number of model iterations to 2,000. However, spot-checking results at key model grid cells did not indicate significant differences, so the number of iterations was capped at the iRIC default of 600.

The modeled velocity magnitudes were validated with ADCP-measured velocities collected on October 24, 2011, during low flow, low lake-level conditions (fig. 3; Fitzpatrick and others, 2025). A mean daily value of 2.12 cubic meters per second (m³/s) at the Sheboygan River streamgage and a Lake Michigan mean daily water level of 175.1 m above IGLD 85 at the Kewaunee lake gage were used. A subset of 90 measured velocity data points was chosen for the comparison when the ADCP bottom tracking setting was used. These data have the most reliable location information. For each simulated velocity for the center point of each grid, the nearest three measured velocities were selected using an R script. If there were no ADCP matches for a grid, the simulated velocity point was dropped. Additionally, if there were widely different flow directions for the simulated compared to the measured data point, the simulated data point was dropped. Measured and simulated velocities in the comparison dataset ranged from greater than 0.1 to less than 1 m/s. The simulated velocities were generally less than the measured velocities. Velocity reversals happened quickly in the vicinity of the Wildwood Islands complex and downstream, making it difficult to compare one-on-one simulated to measured velocity data.

A combination of steady-state Sheboygan River streamflows and Lake Michigan water levels was used to simulate water levels and velocities through the Wildwood Islands complex reach for nine scenarios, with upstream and downstream boundary conditions derived from significant minimum, mean, and maximum values. Observed velocity data from July 2012 indicated frequent changes in flow velocity (on the order of 15 minutes or so) as well as frequent stage changes. Two of the nine scenarios focused on the minimum and maximum downstream boundary conditions experienced during a sequence of seiche events captured during July 2012.

Post-Glacial Geomorphic Setting

The terraces, valley character, and meanders of the Sheboygan River reflect the post-glacial drainage history and Holocene lake-level variations more so than modern alluvial processes. Understanding the extent and timing of the postglacial landforms relative to the modern river helps to describe the hydrogeomorphic history and resulting hydrodynamic and habitat features of the modern river. These landforms resulted from the interaction of retreating glacial ice with glacial meltwater and paleolake fluctuations (Richardson, 1991; Carlson and others, 2011; figs. 4 and 5). When glacial ice covered what is now the valley of the Sheboygan River, flow of meltwater off the glacier was to the south and southwest, noted by the sand and gravel outwash deposits along the valley side at the west edge of the study area. Wave-cut terraces and paleo lacustrine deposits associated with the Glenwood, Calumet, Algonquin, and Nipissing paleolakes are present along the length of the valley in the study reach. Near the Ozaukee diamicton, the ancestral Sheboygan River had a large meander to the north. This meander likely formed after the Valders readvance and before the Calumet high lake stage (Richardson, 1991). The valley narrows where it intersects the western side of the Valders diamicton but then widens downstream before entering the large expanse of glaciolacustrine deposits associated with the Glenwood high stage. The small patches of late Holocene sand and silt in the middle of the modern Sheboygan River in the vicinity of the Valders diamicton are the Wildwood Islands complex.

During successive drops in Lake Michigan water levels below 80 m above the NGVD 29 (equal to the NAVD 88 in this vicinity), at about 16,000; 14,000; and 11,500 years BP (fig. 5), the valley of the Sheboygan River would have likely been much deeper than its modern level (Kiesel, 1998; Carlson and others, 20115). After the Chippewa low stand, the lake level rose to the Nipissing level of about 184 m above NGVD 29 from 7,500 to 5,000 years BP. The lake level decreased until about 3,200 years BP, during which time the valley likely eroded (Kiesel, 1998; Kiesel and Mickelson, 2005). Following about 3,200 years BP, lake levels have varied but have been somewhat like modern levels of about 177 m. However, between about 3,000 and 2,200 years BP, a higher stand of the lake called the Algoma level (not shown) was at about 180 m (Thompson and Baedke, 1999). Deposits associated with the Algoma level blanket the valley bottom throughout most of the study reach (fig. 4). Another slightly higher stand peaked at 177.7 m about 1,600 years BP (not shown).

Historical Lake Michigan Water-Level Fluctuations

Historical Lake Michigan water-level elevations were converted from the vertical datum of IGLD 85 to NAVD 88 for this report to provide a common datum for describing relations among lake levels and Sheboygan River landforms described in later sections. Elevations with the vertical datum of NAVD 88 are 0.14 m higher than elevations with a vertical datum of IGLD 85 in the vicinity of the Sheboygan River mouth. A variety of Lake Michigan water levels, along with summary data from the Sheboygan River stage recorder at New Jersey Avenue, are shown for comparison in table 2.

From about 100 to 1800 A.D., water levels in Lake Michigan fluctuated by more than 1 m, ranging from a low of 176.7 m to a high of 178 m above the NAVD 88 (Baedke and Thompson, 2000). During the last 200 years or so, lake levels have fluctuated from the record high of 178 m in 1838 to a record low of 175.7 m in 2013 (Gronewold and others, 2013). During 1918 to 2012, the monthly average Lake Michigan water level was 176.6 m. The AEP for 0.01 for Lake Michigan was 177.9 m and the AEP for 0.10 was 177.4 m (U.S. Army Corps of Engineers, 1988). The ordinary high-water mark (OHWM) was also 177.4 m, and the LWD was 176.2 m.

Interaction of Lake Levels and River Flows

The hydrogeomorphology and hydrodynamic characteristics of the Sheboygan River AOC are affected by Lake Michigan water levels, river flows, and seiches. Comparison of water levels in the rivermouth with those in Lake Michigan (fig. 6, table 2) provided information about how the river flows and short term and seasonal lake-level variations interacted. The Kewaunee lake gage is the closest lake gage to the study area and is located 100 km north of the Sheboygan River. From March through June 2012, water levels generally climbed about 0.1 m but then fell by about 0.5 m through December 2012 (fig. 6A). Water levels rapidly varied from seiches on about 60 to 90 minutes cycle on the order of ±0.1 m; however, the magnitude of some of the fluctuations was sometimes closer to ±0.3 m, especially during November and December months.

Water levels at the Sheboygan River stage recorder at New Jersey Avenue (U.S. Geological Survey, 2024) showed similar variations in 2012 water levels in comparison with the Kewaunee lake gage except when river flows were high (figs. 6 and 7). Water levels were less than the Lake Michigan LWD of 176.2 m in the lake and the Sheboygan River for much of the time. Seiche minimums at the Sheboygan River at New Jersey Avenue were commonly 0.6 m less than the LWD in October through December 2012. Water levels in the Sheboygan Rivermouth ranged from a minimum of 175.6 to more than 176.4, including the seiche minimums and maximums from 5-minute data, which is a total range of more than 0.8 m during the summer and fall 2012 low flows.

Water-level increases at the Sheboygan River at New Jersey Avenue not mirrored at the Kewaunee lake gage were from increased river flows (figs. 6 and 7). Spring high flows in 2012 were relatively small compared to 2013 (fig. 6A) and were reflected by higher river stages at the New Jersey Ave stage recorder than the Kewaunee lake gage (fig. 6B). The summer of 2012 was droughty and river levels generally mirrored lake levels (Scott, 2012; fig. 6A).

The ever-varying combination of flows and seiches in the lower Sheboygan River result in highly fluctuating water levels that change the type and areal extent of local aquatic and shoreline habitat over very short time intervals. The result of this combination is multiple overlapping habitats within the same areal footprint. This result is observable in three example sets of time-lapse photographs from the Wildwood Islands complex when water levels varied more than 0.3 m within less than 24 hours during an early spring high flow (March 30, 2012), summer low flow (June 30, 2012), and ice jam and high spring flow (March 30 and April 1, 2013) (fig. 8; Reneau, 2012a, 2012b, 2013). At the end of March and beginning of April 2012, a high frequency seiche of about 0.2 m and with an interval of less than 60 minutes dominated the pattern of water-level variations, especially leading up to a rainfall event as a low-pressure system moved in from the west. As river flows increased during the March 30, 2012, event, water levels increased and seiche-related variation decreased (fig. 6B). During summer low flow, sand and gravel bars exposed near the islands go through several cycles of inundation and exposure. Bar areas exposed near the islands are popular for feeding birds in all seasons (figs. 8A, E).

Ice jams are common along the study reach and can quickly raise and lower water levels in the Sheboygan River near the island complex, as shown in a March 31, 2013, video (Reneau, 2013). The ice jams, combined with varying upstream and downstream river currents, transport and deposit large wood, brush, and organic debris on the islands and banks at stages not possible with one-directional downstream-only flow. The ice jams and their vegetative debris were attractive to migrating wildfowl during the 2013 spring melt (fig. 8E, F).

In the winter of 2012 and into spring of 2013, the flows and stages of the Sheboygan River were relatively high compared to most of 2012. The spikes in stage at New Jersey Avenue starting in the winter 2012–13 reflect build up and release of several ice jams that lasted through the end of April 2013, with five jams causing water levels in the vicinity of 177.2 m above NAVD 88 or near the Lake Michigan OHWM (fig. 6A, table 2). A combination of high flow and ice jams in early April resulted in a stage of 178 m above NAVD 88 at New Jersey Avenue. This stage is similar to the stage associated with the 1-percent AEP streamflow and to the record high water level for Lake Michigan (table 2). In contrast, the flows in the Sheboygan River had an AEP of about 20 percent. A time-lapse video documented the water levels nearing and exceeding most of the surface of the main island (Reneau, 2013).

Sheboygan River Streamflow Trends

The Sheboygan River had statistically significant increases in average annual and annual 7-day low flows, but significant trends in annual flood peaks were not observed during the period of record for the streamgage from 1916 to 2008 (fig. 9A). The flow at the Sheboygan River streamgage is affected by an upstream dam and is missing records from 1924 to 1950, but the trends in average annual flow and low flow still mirror trends at other agricultural rivers in southern Wisconsin for similar streamflow characteristics (Gebert and others, 2016). Gebert and others (2016) determined that increasing average annual flow and low flow were from a combination of increased annual precipitation and agricultural conservation practices. The adoption of agricultural management practices that reduce runoff and increase infiltration in the 1960s–70s helped to increase groundwater recharge and stream baseflow. Juckem and others (2008) had similar findings for the causes of increased baseflow in southwest Wisconsin streams. These same agricultural practices may have been responsible for mitigating expected increases in annual peak flows from increased precipitation and frequency of large rainfall events (Nelson Institute for Environmental Studies, 2011). The average annual flow rose from about 6 to 9 m³/s, and the average 7-day low flow from about 0.5 to 1.5 m³/s. The threefold increase in the size of low flows in the Sheboygan River likely helped to sustain aquatic organisms during dry periods, such as those experienced in 2012.

Results from the analyses of trends in AEPs for annual peak floods are supportive of the results from the Kendal tau analyses (fig. 9D–F). Low magnitude, high frequency floods with streamflows associated with an AEP of 95 percent are approximately 50 m³/s higher for 1985–2014 compared to the mean for 1916–2014. Streamflows associated with an AEP of 67 percent remained similar during most of the period of record. Streamflows associated with an AEP of 67 percent are commonly used as a surrogate for bankfull or channel-forming flows for channel design. The flood with an AEP of 1 percent, which is used for flood planning and zoning, was 280 m³/s larger for 1985–2014 period than for 1916–2014. Increases in streamflows associated with an AEP of 1 percent could affect decisions related to the design of rehabilitation features, such as additions of large wood, which would add roughness, slow velocities, and increase water levels.

Wildwood Island Planform Changes, 1938–2010

The areal extents of Wildwood Island and other smaller, nearby islands have changed with time, mainly through erosion of what used to be a much larger single island into a main island with approximately 10 adjacent islands and bars (fig. 10). River flows around the islands have changed from being mainly on the north side of the main island in 1938 to almost equal flow going around both sides of the island in 2010 (fig. 10). A second island was located to the west of the main island in 1938, but between 1938 and 1941 it became attached to the adjacent northern shoreline (fig. 10).

About 35 percent of area of the main island, mostly on the western upstream side, was eroded between 1938 and 2010 (fig. 10) The main island surface area decreased from 2.4 to 1.6 hectares. The five smaller islands or bars to the west of Wildwood island are remnants of what used to be the western side of the main island. The eastern, downstream side of the island has been growing slowly. Lake Michigan water levels in 1938 and 1941 (National Oceanic and Atmospheric Administration, 2016b) were like those in 2010 at about an average of 176.1 to 176.2 m above the NAVD 88. From 1938 to 1974, the presence of at least one structure on the main island is observable on the orthophotographs. It also appeared that woody vegetation was kept down and clearings on the island may have been mowed or farmed. By 1992, the clearings were gone, and trees had grown in over most of the upstream half of the island (fig. 10).

Wildwood Islands Complex Geomorphology and Sedimentation History

The erosion and sedimentation of bars and islands in the Wildwood Islands complex are closely linked to paleolake and modern water-level fluctuations and interactions with river flows. This relation is illustrated in the landform topography of the islands and cross-valley transects of island landforms and sedimentation patterns relative to water levels (fig. 11). Most of the sediment cores collected for this study were located on the main island, with five cores located on smaller islands and bars (fig. 2). The transects are described in the following paragraphs from upstream to downstream. Core and transect survey data are published in a USGS data release (Fitzpatrick and others, 2025).

Transect 14.5 crosses one small island and two bars on the upstream side of the main island (figs. 2, 11A). Core 14.5–2 was collected from a separate island that has been present since at least 1938 (fig. 10), whereas cores 14.5–1 and 14.5–3 are from the erosional remnant of the main island (fig. 10). The surface elevation of the small, most northwest island is about 177.4 m, which is the same as the 10-percent AEP and the OHWM for Lake Michigan. This island has a core of sand and gravel covered with about 0.8 m of layered silt and fine sand. The surface elevations of the other two bars are about 176.7 m, which are about 0.1 to 0.2 m higher than the mean lake level and the highest measured water levels in 2012. During summer low flows and minimum seiche-related levels, the river water level at this transect was 0.2 m lower than the LWD. The two north channels were dry temporarily, and the southern channel had approximately 0.3 m of water depth. The time-lapse photography documented deer and humans frequently crossing the dry and near-dry channels during low seiche water levels. The gravel on the bar near 14.5–3 is similar in size to the buried 1.5-cm diameter gravel overlain by silt and sand deposits in the cores (figs. 12A–B).

Transect 14.0 bisects one bar and the main island (fig. 11B). The bar has a similar elevation to those shown in transect 14.5, with modern silt and sand deposits overlaying alternating sand and gravel deposits. The bar is a remnant of the main island and was cut off between 1941 and 1950, and core 14–4 is located along the previous edge of the island like core 14.5–1. The two north channels went dry during the low seiches in 2012, with about 0.5 m of water depth in the south channel.

The three cores on the main island along transect 14 illustrate varying sedimentary records associated with erosion and deposition during flood flows (fig. 11B). The north side of the island is the highest and likely the oldest. Core 14–1 contained loam and sand deposits overlying a more gravelly layer at about 176 m, which is similar to the remnant deposit on the bar at core 14.5–1. Core 14–1 had abundant 2- to 4-mm ground shells interlayered with gravel from about 175.7 to 176.1 m. Core 14–2 is on a younger side of the island, near the lee side of the erosive flows. The island surfaces for the southern two-thirds of the island would have been covered by water during the 1986 high stand of Lake Michigan. The slower river flow velocities on the lee side of the island allow silt deposition. Core 14–3 is in a slight lateral build out of the island toward the south (fig. 10), as well as having vertical accretion of silt, fine sand, and loam.

Along transect 13.5 (fig. 11C), cores were taken on an island remnant of the main island (core 13.5–5) and across the highest elevations of the main island. Silt deposition on the island remnant is relatively recent (greater than 1950), likely representative of flow changes after the meander cutoff happened upstream. The high surface of the north side of the main island continues but widens and splits with erosional flood chutes. The high surface of the main island is above the water level of 177.9 m and the flood with an AEP of 0.01. Surfaces below this level and on the lee side of the island have recent silt and loam deposits. A radiocarbon age from a wood sample 1.2 m below the surface in core 13.5–2D indicated that most of the fine sand and silt above the gravel was deposited during the last 300 years or less, such as core 14.4 along transect 14 (fig. 11B).

Transect 13.5 has a somewhat continuous shell-rich zone overlying gravel at an elevation of 176 m, and a similar layer was observed in transect 14.0 (figs. 11B, C). The shell layer consisted of well-sorted rounded fragments of about 2–5 mm in diameter with little mineral matter. The gravel unit beneath is at a slightly higher elevation on the remnant of the main island to the north and about 1 m higher in general than the modern channel bed. The gravel may coincide with a coarse facies of a delta or fan when water levels in Lake Michigan were consistently higher. This gravel layer may have been deposited during the higher level of Lake Michigan between 2,200 and 1,600 years BP.

Along transects 13 and 12.5 (figs. 11D and 11E), on the downstream half of the island, the channel on the south is dry during low seiche, whereas the channel on the north has about 0.5 m of water depth. The north side of the island along the downstream portion eroded in the 1940s to 1970s but then deposited in the 1980s and 1990s to return to the 1938 form. Meanwhile, the south side of the lower portion of the island has been eroding near core 12.5–1. The surface of the downstream side of the island along transect 13 shows areas of the island that have been aggrading with silt deposition. The lee side of the island, especially when vegetated, was prone to sedimentation of silt and sand. This area of the island had a large thick growth of Phragmites sp. during the 2011–13 study. The top 1.0 to 1.5 m of sediment on the main island along transect 13 is less than 300 years old (fig. 11D). Shell-rich zones overlying gravels at 175–176 m indicate there is a continuous shell-rich gravel layer beneath the main island. In core 12.5–1, the gravel is at about 175.5 m, which is similar to the elevation of the bed of the south channel.

The longitudinal profile running lengthwise along the main island (fig. 11F) illustrates the variable sedimentary environments and different ages of the deposits. The undulating surface varies by 0.2 m and indicates local scour or narrow flood chutes that frequently cross the island from high flows. The core on the highest part of the island contained mainly sand, indicating a higher energy environment than the younger surface downstream with silty deposits.

The location of the Wildwood Islands complex suggests that it could be a remnant coarse facies of delta deposit for the Sheboygan River when the lake level was dropping from the Algoma level to more modern levels between 2,200 and 1,600 years ago. The gravels have a high abundance of ground shells, as well as gastropod and bivalve fragments. The abundance and roundness of the shell bits indicate a large somewhat local upstream source of mussels and an energy environment, such as a beach, that could sort and grind shells.

An archeological survey conducted in the area in 2011–12 found no archeological remains on the main island (Kubicek, 2012), which is not surprising because the island surface is mainly depositional, and a large part of the island has had about 1 m of sediment deposited during the last 300 years. It is possible that remnants of the older Early and Middle Woodland cultures might be present in sediment near the shell-rich zone, which would be at or in the sedimentary layers above about 176.0 m.