A Fish Community Comparison Upstream and Downstream a High-Head Dam on the Upper Mississippi River Free

Long-term standardized monitoring is useful for detecting spatial and temporal trends in fish communities in large rivers caused from anthropogenic impacts. Dam construction is one of the most dramatic and widespread anthropogenic impacts on natural river systems (Dynesius & Nilsson 1994). Dams eliminate floodplain habitats (Koel 2004), alter flow regimes (Galat & Lipkin 2000), widen portions of the river, and alter rates of sediment transport and volume of bedload (Bhowmik & Adams 1989). Dams result in an unnatural series of lotic and lentic habitats and inhibit the longitudinal movements of river biota (Fullerton et al. 2010).

The Upper Mississippi River (UMR) extends 1,070 km from Minneapolis, Minnesota, downstream to St. Louis, Missouri, and is a biologically productive and economically important large floodplain river system in the United States (Koel 2004). Over the last 100 years, 29 lock and dams were constructed on the UMR mainstem to create a waterway for commercial and recreational traffic (Anderson et al. 2019). The dams have impounded the natural flow of the river to create a series of 28 navigational “pools,” a term used to define the held water between the dam structures. The pools are numbered consecutively north to south: 1‒5, 5a, 6‒22, 24‒26, with Pool 1 starting immediately downstream of Lock and Dam 1 in Minnesota and ending with Pool 26, immediately upstream of Lock and Dam 26 in Missouri. The lock and dam system has reduced habitat heterogeneity and longitudinal connectivity critical to aquatic community diversification and the life histories of species in the UMR (Fremling 2005).

The U.S. Army Corps of Engineer’s Upper Mississippi River Restoration Program’s Long Term Resource Monitoring element (LTRM) aims to provide resource managers with information necessary to maintain the UMR as a sustainable multiple-use large scale river ecosystem. Long Term Resource Monitoring has used standardized long-term monitoring to evaluate fish communities in the UMR since 1994 to present (15 June to 31 October annually) from four pools (Pools 4, 8, 13, 26; Ratcliff et al. 2014; Figure 1). To resolve the 452 km monitoring gap between Pools 13 and 26, in 2009 the Illinois Natural History Survey’s Long Term Electrofishing (LTEF) program, designed to evaluate spatial and temporal trends in fish populations in the Illinois Waterway and provides fishery managers with essential information to effectively manage the fishery within Illinois boundary rivers, was expanded to include seven pools in the UMR (Pools 16 to 21 and 25; Figure 1). Both LTRM and LTEF monitoring follow procedures originally developed for the LTRM program (Ratcliff et al. 2014), although LTEF sampling is limited to main channel border sites (Fritts et al. 2017).

Lock and Dam 19 (LD 19) is unusual among UMR dams as it is a high-head, run-of-river hydroelectric dam possessing a mean hydraulic head of 11.64 meters (m) and is second in height only to the Upper Saint Anthony Falls Dam in Minneapolis, Minnesota. Completed in 1913, LD 19 separates Pool 19 (upstream) from Pool 20 (downstream) and created the first artificial impoundment on the UMR (Fremling 2005). Lock and Dam 19 has inundated 10,100 hectares of floodplain and trapped sediment loads resulting in an estimated 55% reduction in pool volume directly above the dam (Bhowmik & Adams 1986; Galtsoff 1924). Macrophyte beds have developed in resulting shallow water areas outside of the navigation channel margin, accounting for 27% of total estimated coverage for the UMR (Day 1984).

Lock and Dam 19 has formed two vastly different hydrologic regimes in Pool 19 and Pool 20. Pool 19 [74.5 kilometers (km)] is considered geomorphically diverse. The lower half of the pool (20 km) is wide (over 2.5 km across downstream of Fort Madison, Iowa), with lacustrine conditions, low current velocities, static water levels, and extensive (6,800 hectares) shallow-water areas with floating-leaf vegetation, a direct result of the construction of LD 19. The upper portion of Pool 19 is riverine with extensive side channels and shallow backwaters. This diverse geomorphology and floating-leaf vegetation in Pool 19 result in diverse habitat for river flora and fauna. Pool 20 (35.2 km) is characterized as straight, narrow, and riverine with high current velocities, lotic conditions, sparse vegetation, and limited off-channel or lacustrine habitat for river flora and fauna. Proportionately, Pool 20 consists of mostly main channel (80.4%), followed by side channels (19%), and a small percentage of backwaters (0.6%). Pool 19 has less main channel (29.3%) and side channel (17.4%) areas, and more backwater (5.6%) and impounded areas (47.7%) than Pool 20 (Anderson et al. 2019; Figure 2; Figure 3).

Lock and Dam 19 serves as a substantial barrier to upstream migration for aquatic organisms, prohibiting movement except through the lock chamber (Coker et al. 1921; Nielsen et al. 1986; Kelner & Sietman 2000). The geographic range of invasive Bighead Hypophthalmichthys nobilis and Silver Carp H. molotrix, also known as Bigheaded carp (used hereafter), has quickly expanded since their introduction to the Mississippi River in the 1970s (Kolar et al. 2007) and populations are abundant downstream of LD19 (Irons et al. 2009). These carps raise concern of the ecological impact they have on invaded systems due to their fast growth, high fecundity, and early maturation (Schrank et al. 2003, Williamson & Garvey 2005). Additionally, Bigheaded carp are efficient filter feeders that forage at multi-trophic levels causing a substantial threat to many aquatic ecosystems (Love et al. 2018). Detections of Bigheaded carps are important to suppress their expansion and focus management efforts upstream of LD19 in the UMR. Although densities are low, Bigheaded carp have been detected upstream of LD19, albeit scarcely with any standardized long-term collection method (Larson et al. 2017). It is important to note that at the time this research was facilitated in 2013‒2014 no Bigheaded carp had been detected upstream of LD 19 with a standardized long-term collection method. However, since then few Bigheaded carp have been detected upstream of LD 19 with standardized long-term collection methods (DeBoer & Lamer 2021).

Long Term Resource Monitoring procedures are useful for detecting and monitoring the establishment and spread of invasive species within the UMR. Populations of Bigheaded carp upstream LD 19 generally inhabit off-channel areas (DeBoer & Lamer 2021), and because LTEF sampling is limited to main channel border sites, detection probability may be lower than if off-channel areas were included in the sampling regime. Expanded spatial coverage of fish monitoring to off-channel areas using LTRM procedures is necessary between Pools 13 and 26 to pinpoint the Bigheaded carp invasion front in the UMR. Additionally, there is limited standardized fisheries data for Pools 19 and 20, particularly in off-channel areas. Due to vastly different hydrologic regimes above and below LD 19, fish community differences are expected between Pool 19 and Pool 20.

Previous non-standardized fish assemblage data in Pools 19 and 20 has been valuable (Coker 1929; Dunham 1970, 1971; Bertrand & Russell 1973; Jahn et al. 1986), but does not allow for significant comparisons between pools. In 2000, the Upper Mississippi River Conservation Committee recognized the need for fisheries data within this stretch of river and initiated limited (over a two day period) LTRM electrofishing to assess fish assemblages in this area (Chick et al. 2006). The study resolved that further concentrated sampling is needed to adequately define the structuring of fish communities immediately upstream and downstream of LD19. To address these needs, we conducted two years (2013–2014) of standardized LTRM electrofishing (Ratcliff et al. 2014) in Pools 19 and 20 to 1) detect a presence of adult or juvenile Bigheaded carp, 2) quantify differences in fish community structure (i.e., relative abundance of species) between Pools 19 and 20 and among strata [(i.e., backwater (BWS), side channel border (SCB), main channel border (MCB), and impounded reaches (IMP)], and 3) quantify hydrological differences (i.e., water quality variables) between Pools 19 and 20 and among strata.

Long Term Resource Monitoring procedures use a random sampling design stratified by aquatic area types also known as ‘strata’ (Wilcox 1993). Strata are intended to represent reasonably permanent and enduring aquatic geomorphic features within the UMR (Ratcliff et al. 2014). We randomly selected sample sites in Pools 19 and 20 from four strata including: MCB, SCB, BWS, and IMP shorelines (Figure 2) according to LTRM procedures (Ratcliff et al. 2014). The main channel conveyed the majority of the river discharge and in most reaches included the navigation channel. Main channel borders were the apparent shorelines, straight lines across the mouths of side channels, and along the top of inundated portions of the natural bank line. Side channels were large channels that carry less flow than the main channel. Side channel borders were the apparent shorelines, straight lines across the mouths of tertiary channels, and straight lines at the upstream and downstream limits of the apparent shorelines where side channels connected with the main channel. Backwaters were areas that are beyond the banks of the main and side channels. Backwaters included a variety of alluvial floodplain waterbodies. Tertiary and smaller tributary channels were included in backwater areas. Impounded areas were large, mostly open water areas located in the downstream portions of the navigation pools.

We randomly selected primary and secondary sites from a Geographic Information System (GIS) database. Secondary sample sites were utilized as alternative sites to primary sites where sampling gear was unable to be deployed due to insufficient water depth or obstructions. In these cases, the nearest secondary site (within the same strata) to the primary site was selected for sampling. We collected fish at three discrete time intervals: time period 1: 15 June to 31 July, time period 2: 1 August to 15 September, and time period 3: 16 September to 31 October of 2013 and 2014. We used unique sampling sites each time interval and year. We selected 32 primary sites for Pool 19 per time interval per year, totaling 192 unique sampling sites. We selected 24 primary sites for Pool 20 per time interval per year, totaling 144 unique sites (altogether 336 unique sites between Pools 19 and 20 for years 2013 and 2014). Sampling sites were divided evenly among habitat strata (i.e., eight sites per strata per time interval). Because Pool 20 does not have an impounded reach, fewer sampling sites were in Pool 20 than in Pool 19. We indexed and referenced sampling sites by strata in Pool 19 and Pool 20 as Universal Transverse Mercator coordinates. We omitted inaccessible areas, including those with private or no physical access from the sampling frame.

Our fish collection methodology generally followed LTRM standardized electrofishing methods described in Ratcliff et al. (2014; Supplemental Reference S5). Each electrofishing run lasted 15 minutes and spanned a 200-m stretch of shoreline, which was consistent for all strata. A pilot operated the electrofishing boat, while two dip netters collected fish as they appeared, regardless of size or species. We placed fish in a holding tank until the run was completed, and then enumerated, recorded, and released them back into the river. We used Pulse Direct Current boat daytime electroshocking to sample fish using LTRM standardized electrofishing procedures. Power goals from the electrofishing boat and sampling design emulated LTRM procedures to achieve comparable fish catch rates and standardization among strata and between pools. We measured and recorded a suite of water quality and environmental variables (i.e., water temperature, dissolved oxygen, Secchi disk transparency, conductivity, water depth, river stage height, percent coverage and density of aquatic vegetation, and river bottom substrate) at each sample site.

We performed statistical analyses in R (R Core Team 2023). We defined ‘community structure of fishes’ as relative abundance of a species, measured as catch-per-unit-effort (CPUE = number of fish per 15 minutes). We used the Bray-Curtis Dissimilarity index as implemented in the vegan R package (version 2.6-4, Oksanen et al. 2022; herein vegan) to identify differences in fish community structure between Pools 19 and 20. In Bray-Curtis Dissimilarity index, higher values indicated more dissimilarity (i.e., that two sites are less alike). For visualization purposes, we used nonmetric multidimensional scaling (NMDS) in vegan, which attempted to condense the matrix of Bray-Curtis dissimilarities into two variables that preserve as much of the original rank order as possible. The two variables were used as axes to plot the position of the sites and the fish species. We used permutational multivariate analysis of variance using distance matrices (adonis in vegan) to determine if fish community structure differed significantly (α = 0.05) between Pools 19 and 20. We determined the specific species that drove the fish community structure differences (α = 0.05) between Pools 19 and 20 using permutation-based statistical tests (envfit in vegan) and displayed these results with NMDS. We color coded the species drivers in the NMDS plot by habitat preference (lentic vs lotic) from Frimpong and Angermeier’s (2009) life-history traits database to identify if species’ habitat preferences grouped together.

We used the Bray-Curtis Dissimilarity index as implemented in vegan to identify differences in fish community structure among the four geomorphic strata: MCB, SCB, BWS, and IMP. We used NMDS to visually display the Bray-Curtis dissimilarities. We used permutational multivariate analysis of variance using distance matrices (adonis in vegan) to determine if fish community structure differed significantly (α = 0.05) among habitat strata. We used adonis in vegan to view significant (α = 0.05) pairwise comparisons between strata. We did not analyze fish community differences between Pools 19 and 20 strata. Because strata proportions differed between Pools 19 and 20, this analysis was helpful to explain if strata were driving potential fish community differences between the pools.

We selected dissolved oxygen [milligram/Liter (mg/L)], water temperature [°Celsius (°C)], Secchi disk Transparency [centimeters (cm)], and current velocity [meter/second (m/s)] as the water quality variables to quantify hydrological differences between Pools 19 and 20 and among strata because they were continuous data that were quantifiable. We used the Shapiro-Wilk test in rstatix (R package version 0.7.2 Kassambara 2023; herein rstatix) to determine if water quality variables were normally distributed (α = 0.05). As a non-parametric alternative to the standard t-test, we ran Mann-Whitney U tests in rstatix for each of the four water quality variables to determine if significant (α = 0.05) differences occurred between Pools 19 and 20. As a non-parametric alternative to analysis of variance (ANOVA), we ran the Kruskal-Wallis test in rstatix to determine if water quality variables differed significantly (α = 0.05) among strata. We used the non-parametric post-hoc Dunn’s test in rstatix with the Bonferroni method to adjust the p-value for multiple comparisons to view significant (α = 0.05) pairwise comparisons between strata. We did not analyze water quality variables between Pool 19 and 20 strata. We concluded water quality variables were likely to differ among strata, which was helpful to explain the driving factors in potential fish community differences between the pools.

We collected two adult Silver Carp in Pool 19 between years 2013 and 2014. These were the first documented Bigheaded carp upstream of LD19 captured using a standardized long-term collection method. In Pool 20, we collected 89 adult Silver Carp and 1 adult Bighead Carp between years 2013 and 2014 (Table 1). We collected or observed no juvenile Bigheaded carp.

We collected 45,799 total fishes at 283 sites between Pool 19 and Pool 20, years 2013 and 2014 (Table 2). Fifty-three sites were not sampled because of time constraints or sampling inaccessibility despite alternates. We documented relative abundance of fish species measured as mean CPUE and the associated standard error values divided among strata for Pools 19 and 20 in Table 3. We determined that fish community structure differed significantly between Pool 19 and Pool 20 (P = 0.001; Figure 4). We determined 32 species drove the fish community structure differences between Pools 19 and 20 (Table 4; Figure 5). The species NMDS plot showed a cluster of 23 species that had a positive correlation with Pool 19. All species classified as lentic by Frimpong and Angermeier (2009) were included in this cluster. Of the 23 species, those with the strongest relationship to Pool 19 were Largemouth Bass Micropterus salmoides, Bluegill Lepomis macrochirus, Orangespotted Sunfish Lepomis humilis, Green Sunfish Lepomis cyanellus, and Warmouth Lepomis gulosus, all of which are in the Centrarchidae family. Some predominately lotic fishes (classified by Frimpong and Angermeier 2009) were dissimilar to the Pool 19 cluster (i.e., Spotfin Shiner Cyprinella spiloptera, River Shiner Notropis blennius, Mississippi Silvery Minnow Hybognathus nuchalis, Emerald Shiner Notropis atherinoides, and Blue Sucker Cycleptus elongatus) and correlated with Pool 20 (Table 4; Figure 5). We determined that fish community structure differed significantly among strata (P = 0.001; Figure 6). Of the six strata pairwise comparisons, MCB vs. SCB was the only comparison that did not show a significant difference (P = 0.35) between strata. Backwater vs IMP (P = 0.003), BWS vs MCB (P = 0.001), BWS vs SCB (P = 0.001), IMP vs MCB (P = 0.002), and IMP vs. SCB (P = 0.001) showed significant differences between strata.

We determined each of the four water quality variable data sets not to be normally distributed. Of the four water quality variables analyzed to determine significance in hydrological differences between Pool 19 and 20, we determined dissolved oxygen was the only variable that was not significantly different between the pools (P = 0.95). We determined current velocity (P < 0.02), Secchi disk transparency (P = 0.02), and water temperature (P < 0.01) were significantly different between Pools 19 and 20 (Table 5). Of the four water quality variables analyzed to determine significance in hydrological differences among strata, we determined dissolved oxygen (P = 0.71) and water temperature (P = 0.60) were not significantly different among strata. We determined current velocity (P ≤ 0.007) and Secchi disk transparency (P < 0.004) were significantly different among strata (Table 5). Using the Dunn’s Test for Secchi disk transparency, we determined there were significant pairwise comparisons between BWS and MCB (P = 0.002) and BWS and SCB (P < 0.003; Table 6). Using the Dunn’s test for current velocity, we determined there were significant pairwise comparisons between BWS and MCB (P < 0.001), BWS and SCB (P < 0.005), IMP and MCB (P < 0.005), and IMP and SCB (P < 0.004; Table 6). We archived the raw data to replicate these results under the Supplemental Materials section of this manuscript (Table S1).

Our findings suggested fish community structure differences between Pool 19 and 20 are likely caused from LD 19’s influence on hydrology, sedimentation, macrophyte abundance (Pool 19), and upstream fish passage. We detected two Bigheaded carp upstream Lock and Dam 19, the first to be documented using a standardized collection method. We determined fish community structure differed significantly between Pool 19 and Pool 20, and that fish community structure differed significantly among most strata. We determined some water quality variables differed significantly between Pool 19 and Pool 20, and among strata. Prior to 2013, limited fisheries assemblage data existed for Pools 19 and 20, particularly in off-channel areas. Our results presented the first comprehensive investigation comparing fish assemblages between these pools.

We collected two adult Silver Carp in Pool 19 during our sampling timeframe (2013-2014). Both detections occurred in MCB strata with habitat features suitable for Bigheaded carp (i.e., along a large island complex and near the mouth of a large tributary). Although no Bigheaded carp were detected in off-channel areas during our sampling timeframe, research indicates adult and juvenile Bigheaded carp frequent off-channel areas (Calkins et al. 2012, Prechtel et al. 2017) and expanding sampling regimes to include these areas will be valuable for increased detection probability to inform management efforts upstream LD 19.

Larson et al. (2017) suggested established reproducing Bigheaded carp exist upstream of LD19 (i.e., between Pools 14 and 16), although spawning and recruitment are variable. Bigheaded carp spawning events have been observed in 2013 and 2016, with a large year class occurring in 2016 (La Hood et al. 2021; Williams et al. 2021), but no major spawning events have been observed in other years upstream of the dam. Although populations are less dense than immediately downstream in Pool 20, contracted removal efforts have been implemented to decrease densities of Bigheaded carp upstream of the dam, removing 100,000 to 200,000 pounds of Bigheaded carp annually in Pools 14 to 19 (Supplemental Reference S7).

Based upon the knowledge of Bigheaded carp relative abundance upstream of LD19 and the limited Bigheaded carp captures using LTRM electrofishing procedures upstream LD 19, electrofishing may not be the most effective gear type at capturing Bigheaded carp, but still is the most effective LTRM gear currently used. Bigheaded carp exhibit highly complex swimming behavior and heightened sensitivity to disturbance (i.e., electrical current, generator noise, propeller vibration) that contribute to their ability to avoid capture (Bouska et al. 2017). Despite this, the LTRM and LTEF standardized electrofishing programs have been successful at documenting the expansion of Bigheaded carp throughout the UMR over time (i.e., Chick and Pegg 2001; Irons et al. 2007, 2011; DeBoer & Lamer 2021). For example, data showed high catch rates of Bigheaded carp in LTRM and LTEF sampling pools where Bigheaded carp are abundant (i.e., Pools 20, 21, 25 and 26), and showed little to no catch rates of Bigheaded carp in pools where Bigheaded carp densities are low (Zero detection in Pools 4, 8, 13 and two detections in Pool 19). It is important to note that the two Bigheaded carp detections in Pool 19 mentioned here had occurred in more recent years, after our sampling timeframe (2013–2014).

The leading edge of the Bigheaded carp population is between Pools 14 and 16 (Larson et al. 2017) where limited long-term standardized sampling occurs. The LTEF program does not sample off-channel areas, nor at all above Pool 16. In 2021, the Illinois Department of Natural Resources initiated LTRM fish sampling in Pools 12 to 22 in off-channel areas (i.e., SCB, BWS, and IMP strata) occurring only during the LTRM period two timeframe (1 August to 15 September) and is ongoing. Since 2021, these efforts have produced two Silver Carp captures in off-channel areas above LD 19, one in Pool 12 and one in Pool 19. These monitoring efforts emphasize the need for expanded long-term standardized sampling to off-channel areas that would allow enhanced monitoring to a key choke point on the UMR.

Lock and Dam 19 has shaped vastly different hydrological landscapes immediately upstream of and downstream of its structure which likely contributed to the fish structure differences observed between Pool 19 and Pool 20. Due to the high lift of LD19, Pool 19 has an extensive impounded area characterized by abundant aquatic vegetation, increased sediment loads, and low current velocities. Consequently, species that were present in Pool 19 (and absent in Pool 20) generally prefer these conditions. For instance, Spotted Sucker Minytrema melanops, Northern Pike Esox lucius, Pugnose Minnow Opsopoeodus emiliae, Pumpkin Seed Lepomis gibbosus, Redear Sunfish Lepomis microlophus, and Rock Bass Ambloplites rupestris prefer impounded and backwater areas with vegetation and slow water velocities (Pflieger 1997; Rasmussen & Pitlo 2004). The southernmost range of Northern Pike ends just downstream of Pool 19, as this species prefers the colder temperatures in northern North America (Smith 2002). Species that were present in Pool 20 (and absent in Pool 19) were mostly migratory species that may not be able to surpass LD19, despite passage through the navigational lock (i.e., Skipjack Herring Alosa chrysochloris, Blue Sucker, and Blue Catfish Ictalurus furcatus; Smith 2002). The Illinois state-threatened Banded Killifish Fundulus diaphanus presence in Pool 20 is likely the result of its apparent dramatic recovery and range expansion in recent years (Willink et al. 2018). Although undetected in Pool 19 during our sampling timeframe, Banded Killifish detections are commonly reported in LTEF sampling events.

Centrarchids species correlated the most of all species with Pool 19. This is likely because of the abundant aquatic vegetation, virtually absent downstream of LD19 (Day 1984; Tazik et al. 1993). Centrarchids prefer littoral, highly vegetated systems (Savino & Stein 1982; Aday et al. 2009; Warren 2009) and occupy these areas to reduce the risk of predation and maximize foraging benefits (Werner & Hall 1988). Lake-like areas with vast and diverse beds of aquatic macrophytes serve as spawning and nursery habitat for Centrarchid populations (Holland 1986; Dewey et al. 1997).

The high abundance of vegetation in Pool 19 may explain why we observed the high abundance of Centrarchids in this pool. The high-head structure of LD19 has caused deposition of more than 10 meters of sediment behind the dam since its completion in 1913 (Bhowmik & Adams 1989). Deposition of sediment has reduced water depth in the lower half of Pool 19, creating a largely impounded aquatic area. Immediately upstream of the dam, to as far north as Nauvoo, Illinois (24 km upstream), shallow depths and still waters provide ideal habitat for macrophyte colonization. Aerial surveys have shown macrophyte expansion overtime since 1966 (Thompson 1973; Tazik et al. 1993). The most dominant species within these expansive beds is American lotus Nelumbo lutea, (Bhowmik & Adams 1989). As sedimentation continues, macrophyte expansion will likely increase, thus yielding even higher abundances of Centrarchids over time. However, this trend may be reversed over time with increased sediment deposition. Bhowmik and Adams (1986, 1989) predicted Pool 19 will reach a dynamic equilibrium by the year 2050 in which the pool volume will be 20% of its initial post-impounded volume. Most of the river will revert to floodplain forests dominated by moist-soil shrubs and trees (Bhowmik & Adams 1986, 1989), thus eliminating potential Centrarchid habitat.

Seven species that grouped with Pool 19 were classified as lotic species by Frimpong and Angermeier (2009) but could arguably be considered lentic species. Tadpole Madtom Noturus gyrinus, Smallmouth Buffalo Ictiobus bubalus, Mud Darter Etheostoma asprigene, Blackstripe Topminnow Fundulus notatus, Golden Redhorse Moxostoma erythrurum, Green Sunfish, and Pumpkinseed are commonly found in UMR backwater and impounded reaches. These species generally prefer slow current velocities and calm waters, defining Pool 19 (Pflieger 1997; Rasmussen & Pitlo 2004). Pool 20 correlated with fishes that prefer lotic habitat (i.e., high current velocity; deep, narrow, channels, riverine conditions) defining this reach (Pflieger 1997; Rasmussen & Pitlo 2004). Pool 20 is leveed for agricultural and navigation purposes. Levees restrict the river from its natural floodplain, narrowing the river channel, reducing fish habitat diversity, and increasing current velocities (Grubaugh & Anderson 1988).

Our data showed that fish community structure in main and side channels were not significantly different from each other, but backwater and impounded reaches were significantly different from all other strata. Pool 20 has greater than 99% main channel and side channel strata, whereas Pool 19 has greater than 50% impounded and backwater strata. Koel (2004) sampled fishes from LTRM reaches to examine potential relationships between fish species richness and fish habitat diversity among UMR reaches. Upper reaches (i.e., Pools 4, 8, and 13) had the highest species richness whereas lower reaches (i.e., Pool 26, Open River Reach, and the La Grange Reach of the Illinois River) had significantly lower (p = 0.0001) species richness. Our results were comparable with Koel (2004), with greater species diversity in Pool 19 (diverse fish habitat, abundant backwaters) and limited species diversity in Pool 20 (restricted fish habitat selection, extensive main channels). We have demonstrated that the strata proportional differences between Pool 19 and Pool 20 are likely driving the fish community structure differences between the pools.

Our results showed Secchi disk transparency and water temperature were significantly higher, and current velocity was significantly lower in Pool 19. Secchi disk transparency is a measure of water clarity, therefore higher values indicated higher water clarity. We expected to find higher water clarity in Pool 19 because of the abundant aquatic vegetation in the impounded reach. Aquatic vegetation can reduce current velocity and wave energy (from wind exposure or boat traffic) thus increasing sedimentation and water clarity (Barko & James 1998; Schulz et al. 2003). Water temperatures were likely higher in Pool 19 due to the shallow depths of the backwater and impounded reaches allowing sunlight to warm the waters more efficiently. The construction of LD 19 has influenced the hydrology of Pool 19 (i.e., shallow water depths, slow current velocities, and abundant aquatic vegetation) by inundating greater than 10,000 hectares of floodplain and trapped sediment load which resulted in a greater than 50% reduction in pool volume (Bhowmik & Adams 1986; Galtsoff 1924). Among strata, Secchi disk transparency was significantly lower in backwaters than in side channels and main channels, and current velocity was significantly lower in backwaters and impounded reaches than side channels and main channels. Turbidity in backwaters is likely the result of suspended sediment inputs from main and side channels, resuspension of bottom sediments, and phytoplankton (Barko & James 1998; Schulz et al. 2003). We expected to find higher current velocities in main and side channels. As mentioned previously, Pool 20 is mostly main and side channel strata and is characterized as straight, narrow, and riverine with high current velocities.

Continued monitoring upstream and downstream of LD19 is important to detect changes in shifting baselines because of human activity, climate change, and introduced species invasions. In future years, Centrarchid abundance in Pool 19 may increase as the dam continues to impound sediment and macrophyte beds expand (Thompson 1973; Tazik et al. 1993), or Centrarchid abundance may decrease if the impounded reach of Pool 19 reverts to floodplain forest dominated by moist soil shrubs and trees (Bhowmik & Adams 1986, 1989). Range expansion of Bigheaded carp may occur because of climate change, particularly increased rainfall causing flood events. The flood event of 2019 caused an increased presence of Bigheaded Carp in pools where Bigheaded Carp sightings are considered uncommon (i.e., above Pool 14; Turney et al. 2022). Utilizing standardized long-term monitoring to identify evolving trends of Bigheaded carp populations dynamics, movement, and spawning and recruitment is important to inform effective management for this species. We have demonstrated that expanding long-term standardized sampling efforts to Pools 19 and 20, and the reaches upstream of LD 19 is important to understand long-term changes in fish assemblages and the effects of LD19 on the ecology of the Upper Mississippi River.

Please note: The Journal of Fish and Wildlife Management is not responsible for the content or functionality of any supplemental material. Queries should be directed to the corresponding author for the article.

Table S1. Fish raw data file used to analyze our research questions in this paper using Long Term Resource Monitoring electrofishing in Pool 19 and Pool 20 of the Upper Mississippi River, U.S.A. years 2013 and 2014.

Available: https://doi.org/10.3996/JFWM-23-033.S1 (6.27 MB XLSX)

Table S2. Header definitions and descriptions of the fish raw data file (Supplemental Material Table S1) from Pool 19 and Pool 20 of the Upper Mississippi River, U.S.A. years 2013 and 2014.

Available: https://doi.org/10.3996/JFWM-23-033.S2 (12.1 KB XLSX)

Reference S1. Coker RE, Shira AF, Clark HW, Howard AD. 1921. Natural history and propagation of freshwater mussels. Washington DC: US Department of Commerce. 136 pages.

Available: https://doi.org/10.3996/JFWM-23-033.S3 (13,476 KB PDF)

Reference S2. Coker RE. 1929. Keokuk Dam and the fisheries of the Upper Mississippi River. Washington DC: US Department of Commerce. 63 pages.

Available: https://doi.org/10.3996/JFWM-23-033.S4 (5,626 KB PDF)

Reference S3. Irons KS, DeLain SA, Gittinger E, Ickes BS, Kolar CS, Ostendorf D, Ratcliff RN, Benson AJ. 2009. Nonnative fishes in the Upper Mississippi River system: Scientific Investigations Report 2009–5176. Reston, Virginia: US Department of Interior, US Geological Survey. 82 pages.

Available: https://doi.org/10.3996/JFWM-23-033.S5 (17,966 KB PDF)

Reference S4. Jahn LA, Anderson RV, Duffy WG. 1986. The ecology of Pools 19 and 20, Upper Mississippi River: a community profile. Washington DC: National Wetlands Research Center, Research and Development, U.S. Department of the Interior. 162 pages.

Available: https://doi.org/10.3996/JFWM-23-033.S6 (8,939 KB PDF)

Reference S5. Ratcliff EN, Gittinger EJ, O’Hara MT, Ickes BS. 2014. Long Term Resource Monitoring program procedures: fish monitoring, 2^nd edition. US Army Corps of Engineers’ Upper Mississippi River Restoration- Environmental Management Program. Program Report LTRMP 2014-P001. 88 pages.

Available: https://doi.org/10.3996/JFWM-23-033.S7 (15,709 KB PDF)

Reference S6. Wilcox DB. 1993. An aquatic habitat classification system for the Upper Mississippi River System. Long Term Resource Monitoring Technical Report 93-T003. Onalaska, Wisconsin: U.S. Fish and Wildlife Service, Environmental Management Technical Center. 9 pages + Appendix A.

Available: https://doi.org/10.3996/JFWM-23-033.S8 (3,884 KB PDF)

Reference S7. Western Illinois University. 2018. Bigheaded carp monitoring and removal 2018 report. Mississippi Interstate Cooperative Resource Association, Upper Mississippi Sub-Basin Annual Summary Reports. 38 pages.

Available: https://micrarivers.org/upper-mississippi-river-sub-basin-annual-summary-reports/

This project was partially funded by the Department of Interior, U.S. Geological Survey (USGS), LaCrosse, Wisconsin, USA., Cooperative Agreement Award number: G13AC00069, and the Kibbe Field Station at Western Illinois University, Macomb, Illinois, USA. Thanks to USGS Upper Midwest Environmental Sciences Center (UMESC; LaCrosse, Wisconsin, USA.) personnel Brian Ickes for facilitating database and Geographic Information Systems (GIS) strata layers set-up and assuring consistency with LTRM methodology, James Rogala for creating GIS coordinates and strata data layers for sample site locations, and Ben Schlifer for assisting with LTRM database usage. We thank Karen Rivera from the Illinois Department of Natural Resources for her assistance with minnow identification, and also Eli Lampo, Katie Mainor, Brooke Bryant, Kyra Koehler, Erik Carlson, Nick Anderson, Neil Gillespie, among many others from the Kibbe Field Station and Fisheries Laboratory at Western Illinois University who assisted with electrofishing sampling. We thank the Associate Editor and the anonymous reviewers for their suggestions that improved the quality of this manuscript.

Any use of trade, product, website, or firm names in this publication is for descriptive purposes only and does not imply endorsement by the U.S. Government.