Author’s address: E. F. Roseman, Great Lakes Science Center, 1451 Green Road, Ann Arbor, MI 48105, USA. E-mail: firstname.lastname@example.org
Prior to the First World War, the bi-national Detroit River provided vast areas of functional fish spawning and nursery habitat. However, ongoing conflicting human uses of these waters for activities such as waste disposal, water withdrawals, shoreline development, shipping, recreation, and fishing have altered many of the chemical, physical, and biological processes of the Detroit River. Of particular interest and concern to resource managers and stakeholders is the significant loss and impairment of fish spawning and nursery habitat that led to the decline in abundance of most fish species using this ecosystem. Lake sturgeon (Acipenser fulvescens) populations for example, were nearly extirpated by the middle of the 20th century, leaving only a small fraction of their former population. Fisheries managers recognized that the loss of suitable fish spawning habitat is a limiting factor in lake sturgeon population rehabilitation in the Detroit River. In efforts to remediate this beneficial water use impairment, a reef consisting of a mixture of natural rock and limestone was constructed at the upstream end of Fighting Island in 2008. This paper focuses on the response by lake sturgeon to the different replicates of suitable natural materials used to construct the fish spawning habitat at Fighting Island in the Detroit River. Pre-construction fisheries assessment during 2006–2008 showed that along with the presence of adult lake sturgeon, spawning conditions were favorable. However, no eggs were found in assessments conducted prior to reef construction. The 3300 m2 Fighting Island reef was placed at the upstream end of the island in October of 2008. The construction design included 12 spawning beds of three replicates each consisting of either round rock, small or large (shot-rock) diameter limestone or a mixture thereof. An observed response by spawning lake sturgeon occurred the following year when spawning-ready adults (ripe), viable eggs, and larvae were collected during May and June 2009. Additional eggs and spawning-ready adults were found in 2010 (no larval sampling occurred in 2010) as well as collection of three age-0 juvenile lake sturgeon in bottom trawls fished downstream of the reef during July 2010. Spawning lake sturgeon showed no repeatable preference for any of the four particular substrate types but showed a high degree of preference for the island side of the channel, where faster water current velocities occurred. In 2009, overall lake sturgeon egg densities across all replicates averaged 102 m−2 and seven larvae were found in night drift-net samples. In 2010, average lake sturgeon egg density was 12 m−2 and three age-0 lake sturgeon averaging 120 mm TL were collected in bottom trawls in deepwater (∼8 m depth) downstream from the constructed reef. These results demonstrated successful reproduction by lake sturgeon on a man-made reef and suggested that additions and improvements to fish spawning habitat could enhance reproduction and early life history survival of lake sturgeon in the Detroit River.
Native fish populations in the Huron-Erie Corridor (HEC), including those in the Detroit River, have been greatly affected by habitat alterations (Manny et al., 1988; Manny, 2003; Roseman et al., 2007). Millions of tons of limestone bedrock, cobble, and gravel were removed from the Detroit River to build the cities of Detroit, Michigan and Windsor, Ontario and create navigation channels (Larson, 1981). These gravel and rock substrates provided spawning and nursery habitat for lithophilic broadcast spawners, such as lake whitefish (Coregonus clupeaformis), walleye (Sander vitreus), lake sturgeon (A. fulvescens), and many other native fishes (Goodyear et al., 1982). The Livingstone Channel project of the early 1900s was particularly damaging. A 19-km channel was created in the limestone bedrock sill at the mouth of the Detroit River with a minimum width and depth of 91 and 6.7 m, respectively to facilitate shipping (Larson, 1981). Although dredging had taken place in that area of the Detroit River for more than 30 years, this project greatly altered the river’s hydrology and destroyed many lake whitefish spawning grounds in the river (US Bureau of Fisheries, 1917; Manny et al., 1988).
Observed losses of fishery productivity in the Detroit River have been attributed to habitat destruction and alteration (Manny, 2003). For example, the disappearance of lake whitefish spawning runs in the Detroit River correlated in time with the Livingstone Channel construction project, which removed a limestone bedrock formation ideal for lake whitefish spawning (Roseman et al., 2007). Although there are no historical studies of lake sturgeon spawning in the Detroit River, anecdotal information suggests that the Detroit River historically supported a significant lake sturgeon population (Harkness and Dymond, 1961). As lake sturgeon eventually became more important to the economy of the Great Lakes region (Carey, 2005), some of the first attempts to artificially propagate lake sturgeon took place in the late 1880s on the Detroit River (Post, 1890; Meehan, 1909). Historically, the Detroit River and adjacent waters supported one of the largest lake sturgeon populations in the Great Lakes; however, commercial catches of lake sturgeon had collapsed by 1925 (Christie, 1974; Manny and Mohr, 2011).
For waters of the HEC, licenses were issued to commercial fishers for the capture and sale of fish, including lake sturgeon. Up until September 2009, eight such licenses were still active: seven for an annual harvest of 4240 kg of lake sturgeon from Ontario waters of southern Lake Huron (Cottrill et al., 2007) and one for an annual harvest of 2420 kg (equal to about 30 lake sturgeon; Thomas and Haas, 2002) from Ontario waters of Lake St. Clair. Currently, lake sturgeon are considered threatened or endangered by 19 of the 20 states within their original range in the US, including Michigan (Auer, 1991) and in the province of Ontario (OMNR (Ontario Ministry of Natural Resources), 2011). Under Ontario’s Endangered Species Act, the targeting and possession of lake sturgeon is prohibited in Ontario waters of the Great Lakes, including waters of the HEC (Zollweg et al., 2003a; MDNR (Michigan Dept. of Natural Resources), 2010; OMNR (Ontario Ministry of Natural Resources), 2011). Sport angling for lake sturgeon in Michigan waters of the St. Clair River is open from 16 July to 30 September and the harvest limit is one fish per year per person within the slot limit of 1067–1270 mm. In addition, a catch-and-release only season has been added that runs from October 1 to the end of November. Anyone fishing for lake sturgeon in Michigan waters is required to obtain the free lake sturgeon fishing permit that also includes a harvest tag (Thomas, M., MI Dept. Nat. Res. Mt. Clemens Fisheries Res. Office, pers. comm.).
As food, lake sturgeon is valuable. They are a high quality fish, prized for its caviar and flesh, which is often smoked. While caviar of lake sturgeon is seldom available due to restrictions on harvest and trade, other kinds of sturgeon caviar sell in North America for US $250-1800/lb (Caviarideas.com, 2008). Smoked lake sturgeon flesh, when available, sells for about CA$25/lb (Purdy’s Fishery, 2011). Lake sturgeon are also valued as an indicator of environmental health responding to human impacts on the Great Lakes, including the Detroit River and western Lake Erie, which is currently used as a source of high-quality drinking water for over five million people in southwestern Ontario and southeastern Michigan (Manny et al., 1988).
Because lake sturgeon are so valuable, restoring populations is a fisheries management priority in the Great Lakes (Holey et al., 2000; Welsh et al., 2010; Manny and Mohr, 2011) and is included in fish community goals and objectives for waters of the HEC (DesJardine et al., 1995; MacLennan et al., 2003; Ryan et al., 2003), including the Detroit River (Zollweg et al., 2003). In the 1800s, annual harvest of lake sturgeon was highest in lakes Erie and Huron, peaking in 1885 at over 11 million kg/yr in Lake Erie; over 2.3 million kg year−1 in Lake Huron; and (in 1879) at 2.4 million kg year−1 in Lake St. Clair (Baldwin et al., 1979). In the late 1800s, the HEC was, therefore, one of the most productive waters for lake sturgeon in North America, producing a combined total of over 15.7 million kg year−1 (Baldwin et al., 1979). By 1890, however, Lake sturgeon populations had declined due to overfishing and habitat loss to the point where biologists were attempting to collect gametes from wild lake sturgeon to propagate lake sturgeon in hatcheries, from which stocking programs would be undertaken (Post, 1890; Meehan, 1909). That initiative failed and lake sturgeon populations continued to dwindle. Presently in the Great Lakes, lake sturgeon have been reduced to about 1% of their former abundance by overfishing and habitat loss (Auer, 1999).
Spawning habitat for lake sturgeon in much of the HEC has been largely destroyed by creation of deep-draft shipping channels for commercial navigation (Larson, 1981). In 1999, only two of nine historic reputed lake sturgeon spawning grounds in the Detroit River were theoretically suitable for incubation of lake sturgeon eggs (McClain and Manny, 2000). Although the Detroit River had been heavily modified for urban and industrial uses (Manny et al., 1988), potential exists within the HEC for restoration of fish spawning and nursery habitat (Manny and Fiebich, 2001; Manny, 2003). Movements of lake sturgeon in the Detroit River had not been adequately quantified but ultrasonic telemetry showed that they tended to occupy a home area in the Detroit River around Fighting Island (Caswell et al., 2004). Genetically, they are the same as lake sturgeon throughout the HEC (Welsh et al., 2003; Welsh et al., 2010; Welsh and McClain, 2004); are assumed to be a remnant of the original lake sturgeon population that has inhabited this river for centuries; and are being managed as such (Hay-Chmielewski and Whelan, 1997; OMNR (Ontario Ministry of Natural Resources), 2011). Spawning by lake sturgeon on man-made habitats (e.g., cinder beds) (Nichols et al., 2003; Caswell et al., 2004) and on substrates reminiscent of historic spawning sites (Manny, 2006) also suggests that availability of suitable habitat may be limiting production and recovery of lake sturgeon in the Detroit River (Manny and Kennedy, 2002). The purpose of this habitat creation project is to restore spawning habitat for lake sturgeon. Here, we describe the decision-making process for siting and constructing a man-made spawning reef at Fighting Island and then present evidence demonstrating how lake sturgeon responded to this newly constructed habitat in the Detroit River.
Materials and methods
Fighting island reef construction
The Detroit River is the lower 51 km of a channel connecting lakes Huron and Erie. Our study area was located at the head of Fighting Island in Canadian waters of the middle and east section of the Detroit River (Fig. 1). Mean annual discharge of the Detroit River near Belle Isle is 5210 m3 s−1 (Edwards et al., 1989). Northeast Fighting Island (NEFI; Fig. 1) is a reputed historic lake sturgeon spawning ground (Goodyear et al., 1982) but no evidence of spawning by lake sturgeon had been found there recently (Boase, 2007, 2008). Studies conducted as early as 2005 showed that NEFI was used for spawning by lake whitefish (Roseman et al., in press) and walleye (USGS unpubl. data) and considered suitable for lake sturgeon reproduction in the HEC based on the following observed conditions: water velocity is at least 0.5 m s−1; the area is accessible to adult fish; water temperature is optimal for spawning (11–16°C; Bruch and Binkowski, 2002); and preferred water depths of 9–12 m exist (McClain and Manny, 2000; Manny and Kennedy, 2002). What was lacking at NEFI was at least 700 m2 of spawning area possessing at least 20 cm of interstitial space required to attract adult lake sturgeon in spawning condition and to provide adequate protection of lake sturgeon eggs from predation and dislodgement (Bruch and Binkowski, 2002). The expected likelihood was high that adult lake sturgeon would encounter the constructed spawning habitat at NEFI because recent catch data indicated their presence over the NEFI site during much of the ice-free period of the year. Bottom substrates in that area consisted of thin (<8 cm thick) patches of small-diameter gravel and sand on sculpted hard-pan clay (USGS unpubl. data) and possessed little interstitial void space for protection of deposited fish eggs. Creation of layered, larger-diameter, rock/rubble substrates at NEFI was expected to provide the particular combination of substrate and environmental conditions that are required for successful reproduction by lake sturgeon (Bruch and Binkowski, 2002). Water depths greater than five meters existed at the habitat construction site and underwater video data and diver observations at the habitat construction site indicated that low light availability did limit benthic algae (Cladophora) growth, a factor that stopped successful spawning by lake sturgeon at a constructed fish spawning site in the St. Lawrence River (Johnson et al., 2006).
Theoretically, restoration of lake sturgeon in the Detroit River could have been most easily accomplished by constructing suitable spawning substrates at NEFI directly upstream of the shallow, earthen, vegetated shorelines present at the head of the East Fighting Island Channel. This channel had not been dredged for deep-draft navigation and suitable nursery grounds for larval lake sturgeon were believed to exist within the coastal wetlands in this channel. It is believed that this continuous 10 km stretch of relatively unmodified and productive coastal wetlands is representative of those present historically along both sides of the Detroit River prior to 1900 (Manny, 2003). The downstream area of this East Fighting Island Channel is one of the few sites where age-1 to age-3 lake sturgeon were captured in the Detroit River (USFWS unpubl. data). Coastal wetlands along the East Fighting Island channel and around Turkey Island were believed to be a preferred nursery ground for age-0 lake sturgeon consisting of shallow, low-velocity waters, protected from the main-channel river currents, with some marginal, emergent aquatic vegetation and where the river bottom is mostly fine to coarse sand (Kempinger, 1988; Peake, 1999; Holtgren and Auer, 2004; Benson et al., 2005; Friday, 2006).
Our goal was not only to create habitat suitable for reproduction (spawning, egg incubation and hatch, and larval emergence) by lake sturgeon, lake whitefish and walleye at NEFI, but also to determine preferred substrate type by lake sturgeon. Materials that we selected offered a range of properties that were believed to govern egg deposition by the target fish species and represent what was historically present in the Detroit River. Each material we selected is used for spawning by lake sturgeon elsewhere in the Great Lakes basin, i.e., large-diameter (10–50 cm) broken limestone (Pringel and Wirth, 1974; Bruch and Binkowski, 2002; Kline et al., 2009); small to medium sized (5–10 cm) broken limestone (Johnson et al., 2006), and rounded rock 10–25 cm diameter (LaHaye et al., 1992; Manny and Kennedy, 2002), and a mixture of equal parts of these three materials. These materials provided a gradient of interstitial space to protect lake sturgeon eggs from predation and dislodgement by water currents and prevention of smothering by silt and sand deposits. At NEFI, four reproductive substrates that were placed in the river consisted of the above three substrates individually plus a mixture of all three of these substrates, in equal proportions. Each substrate treatment was replicated three times in the 12 beds constructed across the channel width (Fig. 2). These substrates provide a gradient in interstitial space for protection of fish eggs from predation and dislodgement by water currents in an erosion area of deflected/accelerated river velocity where river water velocity lifts and transports silt and sand that would otherwise settle on and smother fish eggs. Fines were washed from these three substrates prior to placement because use of smaller diameter, 2.5–4.0 cm rounded, igneous, rock is the preferred substrate used by sea lamprey (Petromyzon marinus) to build their spawning nest (Applegate, 1950) and we did not want to attract spawning sea lamprey to the Detroit River. Depth of substrate interstitial space on our spawning beds at NEFI needed to be at least 20 cm as recommended for adequate protection of lake sturgeon eggs from predation and dislodgement in Wisconsin (Bruch and Binkowski, 2002). At least 30 cm depth of substrate was present at each of the three active lake sturgeon spawning sites in the HEC in 2001 (Fig. 1; Manny and Kennedy, 2002). Hence, target interstitial depth of rock on rock substrates in each of our constructed beds was 30 cm minimum thickness for the smaller broken limestone and rounded-rock and 60 cm for the large broken limestone.
Economic analysis and consultation with a proposed contractor who bid to place the materials revealed that it would cost more to build 8 larger beds rather than 12 smaller beds using the same amount of materials because the present amount of material (275 m2 by 0.3 m thick or 82.5 m3) was about the maximum amount of material a standard contractor’s dump-bottom barge holds and therefore about how much material a barge can place at one spot without repositioning. Costs would have increased if more materials needed to be brought to one or more spawning beds or a crane with bucket needed to be positioned to complete one or more beds.
At the head of NEFI, in the water current deflection zone, water depth ranged from 6 to 10 m with velocities between 0.3 and 0.8 m s−1, and maintained a relatively minimal slope of river bottom (about 0–5% grade) throughout the channel. The project was placed out of main-channel navigation areas to avoid interference from shipping and propellers. The project was, by design, also upstream of coastal wetland fish nursery areas that provide food and refuge for larval and young-of-the-year fish and is within the home area of lake sturgeon in the Detroit River (Caswell et al., 2004).
The constructed habitat at NEFI was designed to be 1.3 times the size of the Algonac Reef in the North Channel of the St. Clair River Reef (one of two known spawning areas in the upper HEC, and a man-made spawning habitat created by accident in the late 1800’s from steam ship waste slag) and cover at least 3300 m2of river bottom. Lake sturgeon spawn each year at the Algonac Reef in the North Channel of the St. Clair River (Thomas and Haas, 1999; USGS unpubl. data), which is 2500 m2 and referred to as ‘small’ in size by Nichols et al. (2003). Size of the spawning beds was designed to be longer than 18 m in the axis of water movement because lake sturgeon in Wisconsin rivers did not spawn on beds less than 18 m in length (Bruch, R., Wisconsin Department of Natural Resources, Oshkosh, WI, pers. comm.). To maximize encounters between adult lake sturgeon and the reproductive substrates, the project was designed to span the entire width (180 m) of the NEFI channel at water depths of five or more meters, the depths at which lake sturgeon have been captured in the project area at lake sturgeon spawning time in 2007 and 2008 (USFWS unpubl. data). The beds were placed 4 m apart to facilitate scientific assessment of fish use of the different spawning substrates. Beds were placed beside one another’s long axis across the current so eggs deposited by fish on any bed would not be carried by water currents onto another bed and confound the analyses of fish use of the beds and the substrate most preferred by each target fish species. No materials were placed between the eastern most bed and the east shore of the NEFI channel because water depth in that area of the channel was <4 m, and would have permitted light penetration to the bottom and plant growth on the spawning bed materials. Further, lake sturgeon catch data indicated that lake sturgeon movements are restricted to waters deeper than 5 m in this channel (USFWS unpubl. data). Variation of water depth over the 12 individual beds included one set of four beds on steeply sloping (5–10 m in depth) channel bottom along the west side of the NEFI channel, one set of four beds in the deepest, (10 m or more) central part of that channel, and one set of four beds in more gradually sloping, (6–10 m) channel bottom on the east side of that channel. Bed order in the most eastern set was changed to place the mixed-material bed of that set in shallowest (5–6 m) water because the mixed-material bed in the other sets of beds were placed in deeper water (Fig. 2).
Boulders (approximately 200 boulders totaling 400 tons) were placed 10–20 m downstream of all of the beds following methods for lake sturgeon spawning reef placement in the St. Louis River in Minnesota (Cook, 2009). The boulders provide small areas of lower water velocity (current breaks) that were designed to attract spawning-ready fish to the area immediately downstream of the spawning substrates. Because it could be more difficult to add boulders after permit approvals were obtained (than to delete them) boulders were included as part of this project design from the outset. In response to concern that boulders would provide cover for fish that would prey on deposited fish eggs and larvae produced on the reproductive habitat, few boulders were placed directly downstream of the reefs. Placement provided large gaps between the boulders to ensure that potential predators would not inhabit areas immediately downstream of the reefs. Individual boulders also marked the corners of the individual spawning beds to facilitate diver assessments and egg mat placements.
Leading (upstream) edges of each bed consisted of the parent material of that bed, not armor stone like that used to protect the spawning beds constructed at Belle Isle where such armor stone slowed and disrupted water flow over the spawning bed materials, and in-filled bed void spaces with silt and sand settling out of the deflected water (USGS unpubl. data). Ideally, spawning beds could have been constructed in the shape of a wedge with thin leading (upstream) edges and the downstream part of each bed being several layers of substrate thicker than the leading edge (or constructed on a mound or foundation layer of stable material). Such a wedge shape would have elevated the downstream part of each bed and maximized hydraulic pressure on the face of each bed, causing more silt and sand to be scoured out of interstices in the bed and more self-cleaning beds. Tractive (sheer) forces were too small to affect bed stability at NEFI, owing to water depth of bed placement and a drop in water velocity (drag), near the river bottom.
Adult lake sturgeon sampling. Use of the study area by adult and subadult lake sturgeon was monitored with baited setlines following the methods of Thomas and Haas (1999, 2002). Setlines were fished at the reef construction site and at a site approximately 3.0 km downstream of the reef typically from April through June, 2007–20110 (Tables 1 and 2). Setlines consisted of a single solid braided nylon line 76 m in length rigged with 25 number 9/0 stainless steel hooks spaced at 3 m intervals along the line. Setlines were tied to an upstream and downstream anchors marked by surface buoys. Hooks were baited with frozen dead round gobies (Neogobius melanostomus). Setlines were deployed during daylight hours and fished overnight. Fishing effort is reported in hook-hour and catch is reported as numbers of adult (≥1200 mm TL) and subadult (<1200 mm TL) lake sturgeon caught per 1000 hook-h. Spawning condition was determined by visual assessment of flowing gametes. Hook-hours was the total number of baited hooks on the setline minus the number of hooks that were empty when the setline was pulled multiplied by the number of hours the setline was fished. We used the non-parametric Mann–Whitney U-test to examine the significance (α = 0.05) of differences in adult and subadult catch per unit of effort (CPUE) between pre- and post-construction years.
Table 1. Weekly catch per unit of effort (CPUE; #/1000 hook-h) for subadult and adult lake sturgeon caught on baited setlines fished at the Fighting Island reef construction site
CPUE Subadult (<1200 mm)
CPUE Adult (≥1200 mm)
Water temperature (°C)
Sampling in 2007 and 2008 are pre-construction assessments. The reef was constructed in October 2008.
Table 2. Weekly catch per unit of effort (CPUE; #/1000 hook-h) for subadult (<1200 mm TL) and adult lake sturgeon captured on baited setlines fished approximately 3.0 km downstream of the Fighting Island reef construction site during 2007–2010
Week begin date
CPUE subadult (<1200 mm)
CPUE adult (≥1200 mm)
Water temperature (°C)
Sampling in 2007 and 2008 are pre-construction assessments. The reef was constructed in October 2008.
Lake sturgeon egg deposition sampling. Egg deposition by fish was assessed with furnace-filter egg mats placed on the river bottom following methods described by Roseman et al. (In review for this volume). Egg mats were retrieved at least once weekly and inspected for the presence of eggs. For two spawning seasons (spring and fall 2007 and 2008) prior to reef construction, egg mat gangs (three mats per gang) were deployed at locations spanning the NEFI channel 0.5 km upstream of the proposed reef construction site, at the proposed reef construction site, and 0.5 km downstream of the proposed reef construction site. Following construction of the reefs in October 2008, assessment of egg deposition by fish continued in spring and fall 2009 and 2010. In addition to continued assessment at 0.5 km upstream and downstream of the reefs, one gang of egg mats was placed directly on each of the twelve reefs spanning the channel. Eggs were removed from sampling gear, placed in river water, and transported to the Great Lakes Science Center for rearing to hatch. Fish eggs collected from the mats in the field were reared and hatched in the laboratory to confirm their identity, based on subsequent identification of resultant fish larvae following dichotomous keys in Auer (1982). The density of lake sturgeon eggs was sampled on four different reef treatments (a, b, c, and d). Since eggs were removed each day they were counted, observations of egg density were treated as if they were independent. Egg density data are given in the Table 3. No sampling was conducted on 5 May 2010 and there was no reef treatment c in 2010. We fit an analysis of variance (anova) to the data, with the square root of egg density as the response and year, reef treatment, and the year*reef interaction as factors.
Table 3. Pre- and post-construction mean densities of fish eggs (#/m2) collected upstream (up), on, and downstream (dn) of the Fighting Island spawning area
NS, not sampled; LAS, Lake Sturgeon; WAE, Walleye; SUC, Sucker spp.; LWF, Lake Whitefish. LWF were collected in the fall (Oct–Dec) while other species were collected in spring (April–June). There is no fall 2008 data (LWF) because the gear was lost due to massive amounts of aquatic macrophytes moving downstream. LAS only spawned on the five western-most reefs treatments (Fig. 3), but the average presented here includes all 12 reefs.
Larval sampling. To assess larval lake sturgeon drift in the Detroit River, we used an adaptation of Auer and Baker’s (2002) D-frame drift net protocol (Roseman et al., 2011. This volume). The gear consisted of a stainless steel frame (76 cm across the base, 54 cm high) with a knotless 1600 μm mesh nylon bag 317.5 cm long and detachable cod-end. Five nets were fished on the river bottom about 5 m downstream of the NEFI reef at only reef treatments where eggs were captured 10 days earlier in 2009. Resultant fish larvae were identified according to the dichotomous keys in Auer (1982). Drift nets were fished on four nights during 2009 (19–21 and 26–27 May) with sets beginning at sunset with 90 min soak times between lifts.
Juvenile Sampling. A bottom trawl was used to explore the lower Fighting Island channels for juvenile lake sturgeon during late summer and fall, 2010. A 6.1-m, small-mesh, otter trawl (3.8-cm stretch mesh body, and 3.2-cm stretch mesh cod end; 38.1-cm tall, 76.2-cm long trawl boards) was used to assess young-of -year lake sturgeon in the Detroit River. Efforts were confined to the deep-water (7–8 m) channels beginning just below the NEFI reef and extending 5 km downstream. Effort ranged from 5 to 10 min pulling the trawl downstream while maintaining a boat speed of 1 to 1.5 km h−1 faster than the current speed. If the trawl snagged on the bottom of the river and stopped the boat the trawl doors were disconnected from the leads and the net was lifted from the cod end. Catch results were not included when the trawl was snagged. Information collected on lake sturgeon included, total length, fork length, girth, and weight. Fish were implanted with a PIT tag prior to release at the lower end of the transect where captured. Relative abundance was determined for all species and expressed as catch-per-minute-of-effort.
Lake sturgeon showed an immediate response to the newly constructed reef at NEFI by spawning there the first and second springs (2009–2010) after construction was completed in October 2008. We found adult lake sturgeon in spawning condition (ripe), fertilized lake sturgeon eggs, emerging lake sturgeon larvae on and downstream of the reef in 2009 and 2010. In 2010, we caught age-0 juvenile lake sturgeon in bottom trawls fished in the river channel 2–5 km downstream of the reef.
Preconstruction fisheries assessment during 2006–2008 showed that adult lake sturgeon inhabited waters near the reef construction site, as evidenced by the collection of dozens of adult fish on baited setlines fished in the channels on and adjacent to where the reef would be constructed (Table 1). Pre-construction assessments (2007–2008) of adult fish CPUE ranged from 0 to 2.27 fish per 1000 hook-h and the mean CPUE was 0.58/1000 hook-h. The presence of adult fish in the area influenced our decisions to place the reef in this location at the head of Fighting Island. No statistically significant difference in CPUE of adult lake sturgeon was observed between pre- (2007–2008) and post-construction (2009–2010) assessments (Mann–Whitney U = 72; P = 0.66). Catches of subadult lake sturgeon was lower than adult fish with pre-construction CPUE ranging from 0 to 0.97/1000 hook-h (mean 0.11) and post construction CPUE ranging from 0 to 0.82/1000 hook-h (mean 0.14). No statistically significant difference in subadult CPUE between pre- and post-construction assessments was detected (Mann–Whitney U = 55.5; P = 0.44).
Collections of subadult and adult lake sturgeon downstream of the NEFI reef site were of the same magnitude as CPUEs measured at the reef site (Table 2). Pre-construction CPUE of adult fish ranged from 0 to 21.7/1000 hook-h (mean 3.910/1000 hook-h) while post-construction CPUE ranged from 0 to 4.12 (mean 1.086/1000 hook-h) and did not differ significantly (Mann–Whitney U = 82; P = 0.25). Pre-construction CPUE of subadults captured at the downstream site ranged from 0 to 4.43/1000 hook-h while none were captured during post-construction assessments (Table 2). Catch-per-unit-effort of subadults was significantly higher during the pre-construction period than post-construction (Mann–Whitney U = 91; P = 0.04).
No lake sturgeon eggs were collected from the NEFI reef site in 2007 and 2008 nor any location up or downstream during pre-construction assessments in 2008. Lake sturgeon spawned on the newly constructed reefs during 2009 and 2010 with overall mean egg densities higher in 2009 (102 m−2) than 2010 (12 m−2; Table 3). In 2009, lake sturgeon eggs were also captured in very low numbers on egg mats fished approximately 75 m downstream of the constructed reefs (Table 3). All lake sturgeon eggs collected in 2009 and 2010 were captured on reef treatments A through E (Figs 2 and 3), located at the base of the slope near the Fighting Island shoreline where higher than average river velocities occurred (≥0.8 m s−1), with highest densities occurring on reef B and D (Fig. 3), sorted limestone and the mixture of materials. Analysis of variance (anova) using the square root of egg density as the response and year, reef treatment, and the year*reef interaction as factors showed that all three factors were significant at the 5% significance level (Table 4). Since the year*reef factor was significant, that means that the pattern in egg densities among reef treatments was different during 2009 and 2010 (Fig. 3). In 2009, sorted limestone and the mix of materials had significantly higher egg densities than the shot-rock limestone but in 2010, the egg densities from the four reef treatments tested were much more similar. No lake sturgeon eggs were captured on reef treatments F through L nor from sites upstream of the reef during 2009 or 2010. Eggs of walleye and Catostomid suckers were collected during spring in pre- and post-construction years at sites upstream, at, and downstream of the NEFI reef construction site (Table 3). Densities of walleye eggs increased at upstream, reef, and downstream sites in post-construction years while sucker egg densities showed no appreciable difference in density (Table 3).
Table 4. Analysis of variance (anova) results assessing interactions between year, reef treatment, and the year*reef interaction as factors using the square root of egg density as the response variable
All three factors were significant at the 5% significance level. Table headings are d.f, sum of squares (sum sq), mean sum of squares (mean sq), and P-value Pr(>F).
Seven larval lake sturgeon were captured over four nights of sampling during May 19–21 and May 26–27, 2009 demonstrating successful egg and larval survival to the swim-up stage. Six fish ranged in length from 17 to 20 mm TL and one was a sac-fry at 13 mm TL. Fish were captured between 2100 and 0030 h, but fishing effort was terminated at 0200 h on each sampling date.
A total of three young-of-year lake sturgeon were captured in trawls conducted in deep (>8 m) sections of the channel below NEFI in 2010, one on August 9 and two on August 11. Young-of year lake sturgeon captured measured 139, 146 and 150 mm total length. Relative catch-per-effort for young-of year lake sturgeon was 0.05 and 0.10 fish per minute of effort respectively for August 9 and 11. In addition to the young-of year captured, one sub-adult lake sturgeon measuring 1053 mm TL was also captured on August 11. Bottom substrates, as determined by underwater video, consisted of smooth bottom overlaid with small gravel.
Our efforts to rehabilitate lake sturgeon spawning habitat proved successful. We found an immediate response to the man-made spawning reef constructed in 2008 at NEFI as demonstrated by the successful spawning, egg incubation, hatching, and larval emergence by lake sturgeon at this site in surveys subsequent to construction.
Characterization of substrates presently used by spawning lake sturgeon is recognized as important in the restoration of lake sturgeon populations in Michigan (Hay-Chmielewski and Whelan, 1997). We postulated that protection of demersal lake sturgeon eggs and fry from predation and dislocation would be provided by 30 cm or more of interstitial space present among bottom substrates because that interstitial depth has been found to protect demersal lake trout (Salvelinus namaycush) eggs (Edsall et al., 1992) that are larger in diameter than and lack the sticky coating possessed by lake sturgeon eggs (Scott and Crossman, 1973). We further reasoned that successful development and hatch of lake sturgeon eggs and fry would be reduced in proportion to the amount of silt and decomposing organic matter present on the spawning substrates because the latter would reduce dissolved oxygen available for egg and fry survival (Manny and Edsall, 1989; Manny et al., 1995). Owing to the lack of any data on survival of lake sturgeon eggs and fry at our constructed beds, we could not relate the composition and arrangement of bottom substrates or the relative amount of silt and decomposing organic matter on such substrates to the survival of lake sturgeon eggs or fry at these three beds. Our assessments did show the presence of viable eggs in both post-construction years and emergence of larvae in 2009 suggesting that the reef treatments provided suitable conditions for embryonic survival and development.
Minimum habitat criteria of spawning lake sturgeon were recently defined by Bruch and Binkowski (2002) as: (i) clean, rocky substrates layered to provide interstitial space, (ii) water current velocity in excess of 0.5 m s−1, (iii) water temperature of 12–16°C, and (iv) accessible to adults. The multiple layers of clean, broken limestone and cobble in our constructed beds at NEFI satisfy the first criteria above and closely resemble substrates reportedly used by spawning lake sturgeon elsewhere (Scott and Crossman, 1973; Baker, 1980; Kempinger, 1988; Lane et al., 1996; Slade and Auer, 1997; Baker and Borgeson, 1999). Furthermore, cinders and till deposits near Zug Island, Detroit River closely resemble in size and arrangement the cinder substrate near Algonac, Michigan and till substrates near Port Huron, Michigan where lake sturgeon spawn in the St. Clair River (Nichols et al., 2003; Caswell et al., 2004). The areal extent and thickness of deposits of coal cinders and gravel/cobble substrates at three known spawning sites greatly exceed those present at the six and seven reputed, historic, lake sturgeon spawning sites elsewhere in the St. Clair and Detroit Rivers, respectively (Goodyear et al., 1982) that have been recently surveyed with side scan sonar and an underwater camera (USGS unpubl. data). At all three man-made sites, the composition and arrangement of substrates provided a large area of hard, clean rock in beds thick enough to possess at least 30 cm of interstitial space that is flushed by above-average water velocity. Such composition and arrangement of substrates in areas of high water velocity are present over a large area in the upper St. Clair River, at several smaller areas in the lower St. Clair River, and at a few small areas in the Detroit River (USGS unpubl. data). The extent and distribution of such substrates in areas of high enough water velocity may limit the spawning habitat available to lake sturgeon and thus their ability to reproduce in this channel. Our efforts to restore spawning habitat in the Huron-Erie corridor took into consideration these criteria as we attempted to mimic these conditions.
Water current velocity at the NEFI study area near the actual time of spawning (range: 0.3–0.8 m s−1) fell within the range that lake sturgeon were reported to deposit eggs in two Canadian rivers (0.1–1.09 m s−1; LaHaye et al., 1992) and therefore satisfied the second criteria above. Water velocity varies little from year to year at lake sturgeon spawning sites in this channel because discharge in this channel varies little, averaging 5121–5200 m3 s−1; and, ranging from 4250–4400 m3 s−1 in February to 5444–5700 m3 s−1 in August (Edsall et al., 1988; Manny et al., 1988). While our results represent only two years of post-construction monitoring, they do show a habitat preference by spawning lake sturgeon. Viable eggs were found only at the west reef treatments near the base of the shoreline slope, in higher velocity waters (≥0.8 m s−1), and egg densities were highest on the sorted limestone and mixed rock reef treatments. Therefore, future fisheries management efforts that use additions of spawning reefs as a rehabilitation strategy should consider this information in selecting sites and materials for reef construction.
While the highest densities of lake sturgeon eggs were observed on the constructed reefs, we did collect eggs downstream that were likely the result of drift from upstream spawning locations (i.e. the Fighting island reefs). Even though our assessments showed that lake sturgeon had no repeatable preference for a particular substrate type, egg survival could differ substantially from one substrate type to the next due to accessibility of predators related to the sizes of interstitial spaces in substrates. Larger interstices could allow predators easier access to eggs, thus increasing mortality due to predation (Biga et al., 1998). Experiments to assess effects of substrate type on egg predation and other mortality are needed.
Water temperature in this channel satisfies the third criteria above (range: 0.5–25.5°C); usually reaching 13–15°C in May in the Detroit River (Manny et al., 1988) and in June in the St. Clair River (Edsall et al., 1988). Waters at the NEFI study area satisfy the fourth criteria above because there are no barriers to lake sturgeon movements. Although not an important criteria for spawning lake sturgeon, water depths at the NEFI (5–8 m) exceed depths at which lake sturgeon are reported to spawn elsewhere (depth ranges: 0.6–4.6 m in Scott and Crossman, 1973; 1.8–3.6 m in Kempinger, 1988; 0.1; – 1.6 m in LaHaye et al., 1992), perhaps because, compared to smaller rivers, this channel is wider and deeper (700–1000 and 9–17 m, respectively; Edwards et al., 1989). Because light penetration, and water current velocity decreased more than threefold with distance downstream between sites where lake sturgeon spawned in the HEC, Manny and Kennedy (2002) concluded that lake sturgeon spawn over a wide range of water quality in this channel. More quantitative data on measured interstitial space present within the constructed bed substrates and on survival of lake sturgeon eggs and fry relative to the amount of silt and periphyton on different types of bottom substrates used by spawning lake sturgeon in this channel would better focus the construction or restoration of successful spawning sites for lake sturgeon throughout the Great Lakes basin.
Our egg assessment surveys on the constructed reef at NEFI also captured eggs of other fish including walleye, lake whitefish, white sucker (Catostomus commersoni), shorthead redhorse (Moxostoma macrolepidotum), and troutperch (Percopsis omyscomaycus) (USGS unpubl. data). These fishes all require similar spawning habitat as lake sturgeon and shows that habitat rehabilitation efforts targeted toward one lithophilic spawner can meet the life history requirements of other species of the similar spawning guild. Manny et al. (2007, 2010) showed evidence of spawning by these same species at the Belle Isle reef, constructed in the upper Detroit River during 2004. In that report, they conclude that additions and improvements to spawning habitat can add resilience to Great Lakes metapopulations of socially and economically important fisheries. Our results at NEFI also provide this type of positive attribute (increased spawning substrate diversity) to aquatic habitat and fish spawning populations.
With the exception of juvenile stages, our approach to assessing the success of the NEFI reef relied on intensive pre- and post-construction fisheries surveys that examined the entire breadth of lake sturgeon life history stages and their habitat requirements. First, the results of this project showed that productive fish spawning habitat could be constructed in the Detroit River to lead to successful reproduction of the state, Canadian, and provincially-threatened lake sturgeon. Second, the results demonstrated that collaboration among numerous binational partners could construct habitat that was rapidly used for spawning by a variety of native fishes. The rapid and repeated spawning by lake sturgeon and other native fishes on this constructed habitat supports the hypothesis that lack of spawning habitat at NEFI is one of the factors limiting the reproduction of lake sturgeon and other native fishes in the Detroit River. It also showed that loss of spawning habitat for native fishes, due to construction of shipping channels and gravel mining in this river, could be rehabilitated by construction of suitable spawning habitat at this location and probably other locations like this in this river and perhaps other parts of the HEC. It further showed that all of the assumptions we made about the suitability of this site, the choice of spawning substrate materials, array and depth of materials placement, thickness and size of the beds, and accelerated flow rate of water over the constructed habitat, as a result of water deflection off the head of Fighting Island, were correct. Because of the extremely long time involved from hatch of a lake sturgeon through maturity, it may take one or two decades before any noticeable improvements in recruitment are determined. In addition, our short-term (1–2 year) observations will require continued monitoring to determine if trends observed these past spawning seasons continue or are an anomaly. Only repeated monitoring of fish reproduction for a number of years at this site will show if the constructed spawning beds are in harmony with physical processes in the river that govern silting in of interstitial spaces within the matrix of fish spawning substrates, and if reproduction by lake sturgeon at this site has enhanced recruitment and numbers of juvenile and adult lake sturgeon in the Detroit River.
The capture of three age-0 lake sturgeon a short distance downstream from the constructed spawning habitat at NEFI in 2010 may suggest that the constructed spawning habitat was directly connected by river water flows to nursery habitat, although these juvenile fish were collected from deeper depths than we initially anticipated. Previous studies examining the distribution and survival of juvenile Great Lake’s lake sturgeon cited shallow nearshore waters as primary nursery areas (Holtgren and Auer, 2004; Benson et al., 2005; Friday, 2006). Our discovery that age-0 lake sturgeon used deeper portions (>8 m) of Detroit River channels is novel for Great Lakes populations. However, Barth et al. (2009) collected juvenile lake sturgeon in the Winnipeg River, Manitoba at depths exceeding 13 m where velocities were greater than 0.20 m s−1 and substrates were variable, similar to habitat conditions in our study. Collections of juveniles also suggests that lake sturgeon larvae from the constructed spawning habitat could survive transport/outmigration from the spawning ground and reach the nursery habitat and that lake sturgeon larvae found enough food in the river during out-migration and in the nursery habitat to grow rapidly. Continued research efforts and the development of a comprehensive recovery and management plan for lake sturgeon in the HEC will facilitate the maintenance of self-sustaining lake sturgeon populations at levels to serve as a potential source of fish and/or eggs for future rehabilitation efforts in other water bodies.
The successful reproduction of lake sturgeon in the Detroit River is likely, in part, a result of 40 years of pollution prevention and control activities in the Detroit/Windsor metropolitan areas and serves as a testament to the effectiveness of past environmental policy and binational planning to remediate pollution and habitat loss in this Great Lakes connecting channel ecosystem. These findings and the discovery of spawning activity by lake whitefish (Roseman et al., 2007; Roseman et al., in press) and walleye (Manny et al., 2010) show promise that progress is being made toward achieving fish community objectives directed toward improving fish habitat, restoring native fish stocks, and adding to the fishery productivity, biodiversity and ecological resilience of the HEC (MacLennan et al., 2003) and ecosystem-based fishery management of the HEC (Ecosystem Principles Advisory Panel, 1998).
This project was an excellent example of strengthening the science-management linkage. Lake sturgeon research and monitoring data were summarized by researchers for fishery and natural resource managers in both the US and Canada. Based on the weight of evidence of this science, researchers and managers from 17 different Canadian and US organizations agreed to collaborate on the design and construction of a lake sturgeon spawning reef off Fighting Island. The group functioned on a consensus basis and developed a concept proposal for the NEFI reef. This concept proposal was taken by different organizations and agencies in both Canada and the US and expanded into full proposals that were submitted to both US and Canadian nonprofit organizations, foundations, governmental agencies, and corporations. In total, over $250 000 was raised in Canada and the US for NEFI reef construction. The money was pooled and transferred to Essex Region Conservation Authority for reef construction in 2008, representing the first ever fish habitat restoration project in the Great Lakes funded with both US and Canadian funds.
Funds for construction of the NEFI reef were provided by Environment Canada-Great Lakes Sustainability Fund; Province of Ontario, through the Canada-Ontario Agreement; National Fish and Wildlife Foundation-Bring Back the Natives; Michigan Wildlife Conservancy; US Fish and Wildlife Service Coastal Program Grant; DTE Energy, and BASF Corporation. Funding for fisheries assessment of the spawning reef was provided by Essex Region Conservation Authority; Ontario Ministry of Natural Resources; Michigan Department of Natural Resources and Environment; US Fish and Wildlife Service-Detroit River International Wildlife Refuge; US Fish and Wildlife Service Cost Share Challenge Grant Program; US Fish and Wildlife Service-Alpena Fishery Resources Office; US Geological Survey-Great Lakes Science Center including the USGS Science Support Program Project 10-R3-04; and the Great Lakes Restoration Initiative Project 70. Engineering services were provided by Landmark Engineers. Project coordination was provided by Michigan Sea Grant. Grant funds were managed by the Detroit River International Wildlife Refuge Alliance. Ashley Horne, Emily Bouckaert, Justin Chiotti, John Hartig, Kelsey Lincoln, Patricia Thompson, Stacey Wade, and two anonymous reviewers provided critical reviews of earlier drafts of this report. Jean V. Adams and Justin Chiotti provided statistical assistance. Jeff Allen, M. Glen Black, and Marc Blouin provided SCUBA diving assistance. Linda Begnoche, David Bennion, Chantel Caldwell, Steve Farha, Sarah Friedl, Margaret Hutton, Christina Jovanovic, Kelsey Lincoln, Richard Quintal, Kim Smith, and Beth Stockdale provided technical assistance in field sampling. This is contribution number 1640 of the Great Lakes Science Center.