Northeastern Naturalist Vol. 21, No. 3 M.Pyron, L. Etchison, and J. Backus 2014 419 2014 NORTHEASTERN NATURALIST 21(3):419–430 Fish Assemblages of Floodplain Lakes in the Ohio River Basin Mark Pyron1,*, Luke Etchison1, and Julia Backus1 Abstract - We sampled fish assemblages in 41 floodplain lakes in the Ohio River Basin in the summer of 2012. We collected 2427 individual fishes in 70 species. Mean abundance of individuals at sites was 66, and mean species richness per site was 8.1. We used two multivariate procedures to predict fish-assemblage variation from habitat and environmental variables: an indirect gradient approach (reciprocal averaging [RA]) and a direct gradient approach (canonical correspondence analysis [CCA]). When we applied a forward selection process in the CCA, the habitat and environmental variables that contributed significantly to explaining variation in fishes were mean elevation, latitude, maximum depth, conductivity, longitude, dissolved oxygen, cobble and sand substrates, and lake-surface area. RA provided different results that suggested the presence of additional environmental gradients we did not quantify. Our results show that floodplain lakes in the Ohio River basin contain high species richness and are important habitats to conserve because they have the potential to act as source pools for river fish populations. Introduction Lowland rivers are dynamic ecosystems consisting of main channels and broad floodplains that contain aquatic off-channel habitats including sloughs, oxbow lakes, and wetlands that are collectively described as floodplain lakes. These floodplain features extend river ecosystems into terrestrial environments and provide important habitats for many aquatic organisms including fishes. Fishes may require floodplain-lake habitat as adults, or as spawning and nursery sites (Scheimer 2000, Winemiller et al. 2000). Lateral connections between a river and its floodplain provide a means for fishes and other aquatic organisms to move between the two, and they help to maintain habitats by facilitating sediment movement (Amoros and Bornette 2002, Junk et al. 1989). Maintenance of these lateral connections is contingent on hydrology and processes of sediment erosion and deposition (Sullivan and Watzin 2009). Floodplain lakes are biodiversity hotspots that can provide source populations of fish and other organisms to streams (Copp 1989, Sullivan and Watzin 2009). Winemiller et al. (2000) suggested that these habitats serve as source populations for recruitment of certain fishes. Their example was for periodic-strategist fishes that may have good recruitment during years with favorable spring discharge followed by flooding, allowing young-of-the-year fishes connections to the main river channel. Environmental variables strongly influence fish assemblages in floodplain lakes (Lubinski et al. 2008, Miyazono et al. 2010). The degree of connectivity, 1Aquatic Biology and Fisheries Center, Department of Biology, Ball State University, Muncie, IN 47306. *Corresponding author - mpyron@bsu.edu. Manuscript Editor: David B. Halliwell Northeastern Naturalist 420 M.Pyron, L. Etchison, and J. Backus 2014 Vol. 21, No. 3 and lake size and volume variables tend to be correlated, and they explain most fish assemblage variation (Miranda 2005, Miyazono et al. 2010). Connectivity also tends to influence local habitat variables such as turbidity and dissolved oxygen as isolated lakes fill with sediments (Miranda 2005). Fish-occurrence patterns and assemblage structure are well described for many locations in the Mississippi River basin (Dembkowski and Miranda 2012, Miranda 2005, Miranda and Lucas 2004, Miyazono et al. 2010) and elsewhere in North America (Sullivan and Watzin 2009, Winemiller et al. 2000). Our goal was to quantify fish biodiversity and describe relationships between environmental variables and fish assemblages in floodplain lakes of the Ohio River Basin. Field-Site Description Floodplain rivers in the Ohio River watershed are impaired from a multitude of anthropogenic influences including urban point-source pollution, dam operations, agriculture, channelization, and dredging (Pyron and Neumann 2008, White et al. 2005). These impairments have created hydrologically altered ecosystems with losses of riparian vegetation, excessive streambank erosion, increased turbidity, altered temperature regimes, and loss of natural connectivity to floodplain lakes. Prior to our study, floodplain-lake fish assemblages in the Ohio River basin had not been examined. We identified 115 floodplain-like sites in the Ohio River basin in Google Earth and sampled fishes at 41 of the sites that were accessible and not dry during our visit in summer 2012 (Fig. 1). The drought of 2012 was the most severe since 1895 (Hoerling et al. 2013) and caused the majority of sites we visited to be too dry to sample. Methods Our sites varied widely in water depth and habitat complexity (thick macrophytes, trees, and rootwads), which made it impossible for us to use the same sampling approach for all of them. We sampled fishes with a backpack electrofisher (ETS Electrofishing Model ABP-3, Middletown, WI) for 30 min (35 sites), a boat electrofisher (Midwest Lake Electrofishing Systems Infinity, Polo, MO) for 30 min (1 site), or at least 3 hauls with a 10-m x 2-m x 10-mm-mesh seine (5 sites). We used 7mm-mesh dipnets for electrofishing collections and released fishes after we identified them. At each site, we recorded latitude and longitude with a GPS unit and quantified habitat and environmental variables as follows: water temperature (°C), pH, dissolved oxygen (mg/L), and conductivity (μmhos) with a Hydrolab portable meter; maximum water depth; dominant substrate type (boulder, cobble, gravel, sand, silt, hardpan); and presence of woody debris. The following variables were obtained using GIS ArcMap 10 software and a Bing maps base-layer: surface-water area (m2), elevation of water body, elevation difference to closest river (m), and distance to closest river (m). To avoid effects of rare species on multivariate analyses (Gauch 1982), we included only species with abundances higher than 0.1% of total fishes collected, Northeastern Naturalist Vol. 21, No. 3 M.Pyron, L. Etchison, and J. Backus 2014 421 and abundances were log (x + 1) transformed. We analyzed fish and habitat data using two ordination approaches—an indirect gradient analysis based on an underlying unimodal model of species distributions, and a direct gradient analysis that constrained the results using the environmental variables (Palmer 1993). Our purpose in using two analyses was to identify associations between the fish assemblages and environmental variables. We used an indirect gradient method— reciprocal averaging (RA) in Canoco 5 (Ter Braak and Smilauer 2012)—to examine the distribution of species among sites and subsequent correlations with environmental variables. We employed a constrained multivariate analysis—canonical correspondence analysis (CCA) in Canoco 5 (Ter Braak and Smilauer 2012)—with a stepwise-regression approach to predict species-abundance patterns among sites based on environmental variables. We included the forward selection option (P ≤ 0.05) to select habitat variables that were significant contributors to variation in fish abundance, with 499 permutations to test significance (Miyazono et al. 2010). Both multivariate analyses were repeated without the single boat-electrofishing site, to test whether the site provided biased or dif ferent responses. Results Floodplain lakes are unevenly distributed across the Ohio River watershed. Because the gradient of rivers decreased in the western portion of the watershed, there was a greater number of sites in the Wabash River watershed (Fig. 1) than in Figure 1. Collection sites in the Ohio River basin. Northeastern Naturalist 422 M.Pyron, L. Etchison, and J. Backus 2014 Vol. 21, No. 3 Table 1. Ranked abundance and number of sites where fishes were captured. Code is abbreviation used for species inluded in Figures 2 and 3. # of Common name Code Scientific name Abundance sites Bluegill BLGI Lepomis macrochirus Rafinesque 689 29 Brook Silverside BRSI Labidesthes sicculus (Cope) 210 7 Gizzard Shad GISH Dorosoma cepedianum (Lesueur) 194 10 Western Mosquitofish MOSQ Gambusia affinis (Baird & Girard) 171 14 Bluntnose Minnow BLMI Pimephales notatus (Rafinesque) 125 11 Central Stoneroller CEST Campostoma anomalum (Rafinesque) 91 5 Lepomis hybrid LESP Lepomis spp. 65 7 LargemouthBass LABA Micropterus salmoides (Lacepéde) 65 15 Creek Chub CRCH Semotilus atromaculatus (Mitchill) 63 8 Warmouth WAMO Lepomis gulosus (Cuvier) 52 12 Southern Redbelly Dace SRBD Phoxinus erythrogaster (Rafinesque) 50 2 Longear Sunfish LESF Lepomis megalotis (Rafinesque) 49 10 Spotfin Shiner SPSH Cyprinella spiloptera (Cope) 49 7 Green Sunfish GRSF Lepomis cyanellus Rafinesque 48 14 White Sucker WHSU Catostomus commersonii (Lacepéde) 48 4 Eastern Blacknose Dace BLDA Rhinichthys atratulus (Hermann) 45 3 Sand sShiner SASH Notropis stramineus (Cope) 33 3 Steelcolor Shiner STSH Cyprinella whipplei Girard 29 5 White Crappie WHCR Pomoxis annularis Rafinesque 27 9 Black Bullhead BLBU Ameiurus melas (Rafinesque) 25 3 Common Carp COCA Cyprinus carpio L. 22 5 Northern Hog Sucker NOHS Hypentelium nigricans (Lesueur) 22 3 Rock Bass RB Ambloplites rupestris(Rafinesque) 21 4 Blackspotted Topminnow BLTM Fundulus olivaceus (Storer) 19 3 Blackstripe Topminnow BSTM Fundulus notatus (Rafinesque) 17 5 Pumpkinseed PUSF Lepomis gibbosus (L.) 17 5 Mottled Sculpin MOSC Cottus bairdii Girard 14 3 Shortnose Gar SNGA Lepisosteus platostomus Rafinesque 13 4 Yellow Bullhead YEBH Ameiurus natalis (Lesueur) 12 6 Spotted Sucker SPSU Minytrema melanops (Rafinesque) 11 6 Black Crappie BLCR Pomoxis nigromaculatus (Lesueur) 9 5 Goldfish GOFI Carassius auratus (L.) 9 5 Golden Shiner GOSH Notemigonus crysoleucas (Mitchill) 7 3 Redfin Shiner RFSH Lythrurus umbratilis (Girard) 7 4 Smallmouth Bass SMBA Micropterus dolomieu Lacepéde 7 5 Channel Catfish CHCA Ictalurus punctatus (Rafinesque) 6 3 Silverjaw Minnow SJMI Notropis buccata (Cope) 6 4 Striped Shiner STRS Luxilus chrysocephalus Rafinesque 6 4 Bowfin BOFI Amia calva Linnaeus 5 4 Bullhead Minnow BHMI Pimephales vigilax (Baird & Girard) 4 2 Greenside Darter GRDA Etheostoma blennioides Rafinesque 4 4 Mississippi silvery Minnow MSMI Hybognathus nuchalis Agassiz 4 2 Redfin Pickerel RDPI Esox americanus Gmelin 4 3 Smallmouth Buffalo SMBU Ictiobus bubalus (Rafinesque) 4 4 Black Buffalo BLBU Ictiobus niger (Rafinesque) 3 2 Grass Carp Ctenopharyngodon idella (Valenciennes) 3 3 Johnny Darter Etheostoma nigrum Rafinesque 3 2 Mud Darter Etheostoma asprigene (Forbes) 3 2 Northeastern Naturalist Vol. 21, No. 3 M.Pyron, L. Etchison, and J. Backus 2014 423 the subwatersheds east of it. We collected 2427 individual fishes in 69 species and 1 hybrid at 34 sites (Table 1; 7 sites contained no fishes). Mean abundance of individuals at sites was 66 (range = 0–1506), and mean species richness per site was 8.1 (range = 2–21). Mean Shannon-Weiner diversity for sites was 1.2 (range = 0–2.3), mean Jaccard evenness index was 0.64 (range = 0–1), and mean Simpson dominance score was 0.43 (range = 0.13–1). Mean water temperature was 21 °C (range 5–35 °C), mean dissolved oxygen was 5.7 mg/L (range = 0.5–12 mg/L), mean pH was 7.5 (range = 5.5–8), and mean conductivity was 508 μmhos (range = 35–1090 μmhos). Mean surface area was 12,000 m2 (SD = 19,000), mean maximum depth was 0.27 m (SD = 0.1), mean elevation difference from the site to nearest river was 9 m (SD = 10), and mean distance from the nearest river was 289 m (SD = 491). The first and second RA axes explained 19.6 and 9.0% of variation, respectively (Fig. 2), and gradient lengths of these axes were 4.3 and 2.9, respectively. The first RA axis was negatively correlated with latitude (r = - 0.40, P = 0.017) and positively correlated with surface area (r = 0.44, P = 0.007). Lepomis gulosus (Warmouth) and Ameiurus melas (Black Bullhead) were abundant at southern lakes with large surface areas (Fig. 2). Phoxinus erythrogaster (Southern Redbelly Dace) and Rhinichthys atratulus (Eastern Blacknose Dace) were abundant at northern lakes with small surface areas. The second RA axis was negatively correlated with conductivity ( r = -0.33, P = 0.048). Sites with lower conductivity had higher abundance of Labidesthes sicculus (Brook Silverside) and Pomoxis annularis (White Crappie) (Fig. 2). Sites with higher conductivity had increased abundance of Black Bullhead. Table 1, continued. # of Common name Code Scientific name Abundance sites Orangespotted Sunfish Lepomis humilis (Girard) 3 4 Redear Sunfish Lepomis microlophus (Günther) 3 3 Silver Carp Hypophthalmichthys molatrix (Valenciennes) 3 4 Bigeye Chub Hybopsis amblops (Rafinesque) 2 2 Blackside Darter Percina maculate (Girard) 2 2 Freshwater Drum Aplodinotus grunniens Rafinesque 2 3 Logperch Percina caprodes (Rafinesque) 2 2 Mimic Shiner Notropis volucellus (Cope) 2 2 Pirate Perch Aphredoderus sayanus (Gilliams) 2 3 Silver Shiner Notropis photogenis (Cope) 2 2 Bigeye Shiner Notropis boops Gilbert 1 2 Bigmouth Buffalo Ictiobus cyprinellus (Valenciennes) 1 2 Brindled Madtom Noturus miurus Jordan 1 2 Brown Bullhead Ameiurus nebulosus (Lesueur) 1 2 Flier Centrarchus macropterus (Lacepéde) 1 2 Longnose Gar Lepisosteus osseus (L.) 1 2 Quillback Carpiodes cyprinus (Lesueur) 1 2 Rainbow Darter Etheostoma caeruleum Storer 1 2 River Carpsucker Carpiodes carpio (Rafinesque) 1 2 Shortnead Redhorse Moxostoma macrolepidotum (Lesueur) 1 2 Slough Darter Etheostoma gracile (Girard) 1 2 Tadpole Madtom Noturus gyrinus (Mitchill) 1 2 Northeastern Naturalist 424 M.Pyron, L. Etchison, and J. Backus 2014 Vol. 21, No. 3 The first 2 axes of the CCA explained 17% and 11% of variation, respectively (Fig. 3). Habitat and environmental variables—mean elevation, distance to river, surface area, sand substrate, and dissolved oxygen—accounted for 44.7% of fish variation in a forward selection process (Table 2). Sites with lower elevation difference, further distance from the adjacent river, smaller surface area, and low dissolved oxygen had higher abundances of Carassius auratus (Goldfish) and Gambusa affinis (Western Mosquitofish) than other sites (Fig. 3). Sites nearer the adjacent river with greater surface area, higher dissolved oxygen, and lower frequency of sand substrates had higher abundances of Dorsoma cepedianum (Gizzard Shad) and Pomoxis nigromaculatus (Black Crappie) than sites farther from the river with different conditions. Sites with lower surface area, higher elevation difference, and sand substrates tended to have high species richness (Fig. 3). Figure 2. Biplot for first and second axes of a reciprocal averaging analysis. Closed circles represent fish species and open circles are sites. Significant environmental correlations are listed along axes. See Table 1 for species codes. Table 2. Significant environmental variables from a forward selection procedure in canonical correspondence analysis (CCA). Variable Percent variation P Mean elevation (m) 11.3 0.034 Distance to river (km) 10.2 0.038 Surface area (m2) 10.3 0.002 Sand substrate 7.0 0.010 Dissolved oxygen (mg/L) 5.9 0.044 Total variation 44.7 Northeastern Naturalist Vol. 21, No. 3 M.Pyron, L. Etchison, and J. Backus 2014 425 Discussion Floodplain river ecosystems are maintained by predictable seasonal flood pulses that add and distribute nutrients and sediments (Sparks 1995). Scheimer (2000) defined the ecological integrity of a large river and its floodplain habitats Figure 3. First two axes of a canonical correspondence analysis (CCA) ordination. The top plot contains species, and vectors represent significant habitat/environmental predictors of fish abundances. The bottom plot represents sites, and circles are scaled to species richness at sites. See Table 1 for species codes. Northeastern Naturalist 426 M.Pyron, L. Etchison, and J. Backus 2014 Vol. 21, No. 3 according to hydrological connectivity, flux of nutrients and organic matter, and habitat connectivity for fishes. Connectivity of a river with off-channel habitats varies with water level (Lyon et al. 2010), and is likely the strongest explanation of fish occurrence in off-channel habitats. Miyazono et al. (2010) found that fish species with periodic life-history strategies (e.g., Lepisosetus spp. [gar] , Ictiobus spp. [buffalo fish]; see Winemiller and Rose 1992) tended to occur in floodplain lakes with higher connectivity-index scores than other species. These authors also reported that fishes with opportunistic strategies (minnows, topminnows, and poeciliids) tended to occur in floodplain lakes with lower connectivity; their connectivity index increased with increasing distance of lakes to rivers, outlets, and other nearby lakes (Miyazono et al. 2010). We found that 2 connectivity variables were significant predictors of assemblage structure: mean elevation difference and distance to the nearest river, but they were at opposite ends of the CCA ordination. In our study, the floodplain lakes that were isolated by a higher elevation difference contained higher abundances of Hypentelium nigricans (Northern Hog Sucker), Cottus bairdii (Mottled Sculpin), and several minnows, a pattern that fits the opportunistic life-history strategy of Winemiller and Rose (1992). However, floodplain lakes that were isolated by distance contained invasive Goldfish and Western Mosquitofish. Isolation of floodplain lakes has a strong influence on fish-assemblage attributes. Shoup and Wahl (2009) suggested that lakes with sufficient depth to avoid dessication that were farther from a main river channel were more stable because they were less affected by flood events than shallower water bodies located closer to main channels. Schomaker and Wolter (2011) suggested an alternative interpretation of the influence of isolation on fish occurrences using a generalist-specialist categorization: generalist species of fishes tend to occupy water bodies in river floodplains near a river channel, and specialist species tend to occupy water bodies farther away from rivers. We found a generalist group of cyprinids (Notropis stramineus [Sand Shiner], Eastern Blacknose Dace, Semotilus atromaculatus [Creek Chub]; Fig. 2) and Catostomus commersonii [White Sucker] at sites with high connectivity (low elevation difference from river to a floodplain lake). However, we did not find a specialist group of fishes in floodplain lakes at the opposite end of this connectivity gradient. In addition, we did not find a strong pattern of species richness with the maximum depth gradient. Our findings were likely influenced by conditions during the drought year in which we made our collections when isolated sites with the potential to contain specialist species were dry. Sampling floodplain lakes for multiple years would likely result in different patterns (Shoup and Wahl 2009). Fish species richness is higher in assemblages that occur in floodplain lakes where water depth and surface area are higher and where habitat diversity may be greater (Dembkowski and Miranda 2012). Dembkowski and Miranda (2012) predicted that shallow lakes that are likely to experience desiccation during drought will have depauperate fish assemblages that are limited to species with the ability to colonize rapidly. Deeper floodplain lakes with more stable water levels are predicted to contain higher species richness and sensitive species (Dembkowski and Northeastern Naturalist Vol. 21, No. 3 M.Pyron, L. Etchison, and J. Backus 2014 427 Miranda 2012). We found a strong surface-area gradient for the second CCA axis, but sites with the highest species richness tended to have a smaller surface area. The species that we collected in floodplain lakes with smaller surface areas included rapid colonizers (Western Mosquitofish and cyprinids). Although our collections resulted in high overall species richness (70), the distribution of species among lakes varied widely. Only 1 species occurred in more than half of lakes— Lepomis macrochirus (Bluegill)—and only 7 additional species occurred in one third of lakes. The majority of species occurred in less than 10 lakes, so each was relatively rare in our collections. We found 4 non-native species, and only Cyprinus carpio (Common Carp) occurred in higher abundance (22) than the other exotics (maximum of 9). However, we observed multiple dry floodplain lakes that appeared to contain hundreds of dead Hypophthalmichthys molatrix (Silver Carp) or H. nobilis (Richardson) (Bighead Carp); flooded backwater areas are likely highquality nursery habitats for larvae of these two species (Garve y 2008). Floodplain habitats in the Ohio River Basin contain high fish-species richness, with the potential to act as source pools to repopulate rivers following disturbances. Several studies have provided evidence that floodplain lakes contribute to fish assemblages in adjacent streams (Copp 1989, Lyon et al. 2010, Sullivan and Watzin 2009). Sullivan and Watzin (2009) interpreted the widespread presence of habitat opportunists in floodplain habitats as evidence that these habitats are important refuges under conditions stressful to fish, including high flows, drought, and temperature extremes. Zeug and Winemiller (2008) found that fish recruitment in floodplain lakes occurred primarily during low-flow periods, resulting in important contributions to river-channel populations when flows increased. Floodplain lakes with seasonal connections to rivers are spawning and nursery habitats for fishes and contribute to main river populations (Sabo et al. 1991, Shoup and Wahl 2009, Turner et al. 1994). Our use of multiple gears and collection techniques likely biased our results. Boat electrofishing is biased towards larger-bodied species, and seine collection is biased towards smaller-bodied species. In smaller water bodies with low habitat complexity where we effectively sampled all habitats, we were confident that our collections were representative of the species present. In water bodies with complex habitats and deeper water, our collections were likely not representative because we could not effectively sample all habitats. We suggest that deleting rare species prior to ordinations partially addressed these issues; given the habitat complexity in floodplain lakes, we might also have eliminated some problematic sites from our analyses or, for species that we consider to have been undersampled, we might have interpreted the results differently. Our indirect and direct ordination approaches to analyze these fish-assemblage data had different results. Although the results of direct gradient ordination (CCA) showed significant patterns explained by sampled environmental variables, indirect gradient ordination (RA) resulted in different patterns for sites and species. Outcomes of both analyses suggested that surface area was a significant predictor (or correlate) of fish-occurrence. The indirect gradient analysis did not show significant gradients for elevation difference, sand substrate, Northeastern Naturalist 428 M.Pyron, L. Etchison, and J. Backus 2014 Vol. 21, No. 3 or maximum depth. This finding implies the presence of other unknown environmental gradients we did not quantify. Historically, these floodplain habitats have been degraded by levee and dam construction, bank stabilization, and agricultural activities (including draining). These modifications block natural flooding, eliminate floodplain connections, and prevent the flood-pulse hydrologic regime that many organisms require (Sparks 1995). The natural hydrologic regime controls sediment accumulation and water depth in floodplain lakes (Miranda 2011). Anthropogenic disturbances cause increased sediment accumulation and subsequent loss of depth in floodplain lakes. Agricultural activities in the watershed likely contribute the most sediments and have the greatest impact on water depth (Dembkowski and Miranda 2012, Wren et al. 2008). Miranda (2011) posited that decreased depth in floodplain lakes results in the presence of fewer available habitats and decreased biodiversity. Management of floodplain lakes requires increased awareness of the diverse habitats they support and maintenance of natural flow regimes and connectivity (Sparks 1995). A natural flow regime can only be restored in these floodplain ecosystems if anthropogenic flow and connectivity modifications (i.e., dams and levees) are removed or mitigated (Bayley 1991, Gergel et al. 2002). River-ecosystem improvement by flow restoration can be accomplished through modification of dam operations (Bednarek and Hart 2005) including flow experiments (Konrad et al. 2011), removal or repositioning of levees (Opperman et al. 2009), and removal of modifications that alter natural-flow regimes (Poff et al. 1997). Acknowledgments Funding for collections was from the Ohio River Basin Fish Habitat Partnership, Bloomington, MN. We are grateful to R. Durtsche and D. Etchison for lodging durin g sampling. Literature Cited Amoros, C., and G. Bornette. 2002. Connectivity and biocomplexity in waterbodies of riverine floodplains. Freshwater Biology 47:761–776. Bayley, P.B. 1991. The flood-pulse advantage and the restoration of river-floodplain systems. Regulated Rivers: Research and Management 6:75–86. Bednarek, A.T., and D.D. Hart. 2005. Modifying dam operations to restore rivers: Ecological responses to Tennessee River dam mitigation. Ecological Applications 15:997–1008. Copp, G.H. 1989. The habitat diversity and fish reproductive function of floodplain ecosystems. Environmental Biology of Fishes 26:1–27. Dembkowski, D.J., and L.E. Miranda. 2012. Hierarchy in factors affecting fish biodiversity in floodplain lakes of the Mississippi Alluvial Valley. Environmental Biology of Fishes 93:357–368. Garvey, J.E. 2009. Placing the North American invasion of Asian carp in a spatially explicit context. Biological Invasions 11:905–216. Gauch, H.G. 1982. Multivariate Analysis in Community Ecology. Cambridge University Press, Cambridge, UK. 312 pp. Gergel, S.E., M.D. Dixon, and M.G. Turner. 2002. Consequences of human-altered floods: Levees, floods, and floodplain forests along the Wisconsin River. Ecological Applications 12:1755–1770. Northeastern Naturalist Vol. 21, No. 3 M.Pyron, L. Etchison, and J. Backus 2014 429 Hoerling, M., J. Eischeid, A. Kumar, R. Leung, A. Mariotti, K. Mo, S. Schubert, and R. Seager. 2013. Causes and predictability of the 2012 Great Plains drought. Bulletin of the American Meteorological Society doi: http://dx.doi.org/10.1175/BAMS-D-13-00055.1. Junk, W.J., P.B. Bayley, and R.E. Sparks. 1989. The flood-pulse concept in river-floodplain systems. Canadian Special Publication of Fisheries and Aquatic Sciences 106:110–127. Konrad, C.P., J.D. Olden, K.B. Gido, N. Hemphill, M.J. Kennard, D.A. Lytle, T.S. Melis, C.T. Robinson, J.C. Schmidt, E. Bray, M.C. Freeman, L. McMullen, M.C. Mims, M. Pyron, and J.G. Williams. 2011. Large-scale flow experiments for managing rivers. Bioscience 61:948–959. Lubinski, B.J., J.R. Jackson, and M.A. Eggleton. 2008. Relationships between floodplainlake fish communities and environmental variables in a large river-floodplain ecosystem. Transactions of the American Fisheries Society 137:895–908. Lyon, J., I. Stuart, D. Ramsey, and J. O’Mahony. 2010. The effect of water level on lateral movements of fish between river and off-channel habitats and implications for management. Marine and Freshwater Research 61:271–278. Miranda, L.E. 2005. Fish assemblages in oxbow lakes relative to connectivity with the Mississippi River. Transactions of the American Fisheries Society 134:1480–1489. Miranda, L.E. 2011. Depth as an organizer of fish assemblages in floodplain lakes. Aquatic Sciences 73:211–221. Miranda, L.E., and G.M. Lucas. 2004. Determinism in fish assemblages of floodplain lakes of the vastly disturbed Mississippi alluvial valley. Transactions of the American Fisheries Society 133:358–370. Miyazono, S., J.N. Aycock, L.E. Miranda, and T.E. Tietjen. 2010. Assemblage patterns of fish functional-groups relative to habitat connectivity and conditions in floodplain lakes. Ecology of Freshwater Fish 19:578–585. Opperman, J.J., G.E. Galloway, J. Fargione, J.F. Mount, B.D. Richter, and S. Secchi. 2009. Sustainable floodplains through large-scale reconnection to rivers. Science 326:1487–1488. Palmer, M.W. 1993. Putting things in even better order: The advantages of canonical correspondence analysis. Ecology 74:2215–2230. Poff, N.L., J.D. Allan, M.D. Bain, J.R. Karr, K.L. Prestegaard, B.D. Richter, R.E. Sparks, and J.C. Strombert. 1997. The natural flow regime. Bioscience 47:769–784. Pyron, M., and K. Neumann. 2008. Hydrologic alterations in the Wabash River watershed. River Research and Applications 24:1175–1184. Sabo, M.J., W.E. Kelso, C.F. Bryan, and D.A. Rutherford. 1991. Physicochemical factors affecting larval-fish densities in Mississippi River floodplain ponds, Louisiana (USA). Regulated Rivers: Research and Management 6:109–116. Scheimer, F. 2000. Fish as indicators for the assessment of the ecological integrity of large rivers. Hydrobiologia 422/423:271–278. Schomaker, C., and C. Wolter. 2011. The contribution of long-term isolated water bodies to floddplain fish diversity. Freshwater Biology 56:1469–1480. Shoup, D.E., and D.H. Wahl. 2009. Fish diversity and abundance in relation to interannual and lake-specific variation in abiotic characteristics of floodplain lakes of the Lower Kaskaskia River, Illinois. Transactions of the American Fisheries Society 138:1076–1092. Sparks, R.E. 1995. Need for ecosystem management of large rivers and their floodplains. Bioscience 45:168–182. Sullivan, S.M.P., and M.C. Watzin. 2009. Stream-floodplain connectivity and fish-assemblage diversity in the Champlain Valley, Vermont, USA. Journal of Fish Biology 74:1394–1418. Northeastern Naturalist 430 M.Pyron, L. Etchison, and J. Backus 2014 Vol. 21, No. 3 Ter Braak, C.J.F., and P. Šmilauer. 2012. Canoco reference manual and user’s guide: Software for ordination, version 5.0. Microcomputer Power, Ithaca, NY. 118 pp. Turner, T.F., J.C. Trexler, G.L. Miller, and K.E. Toyer. 1994. Temporal and spatial dynamics of larval and juvenile fish abundance in a temperate floodplain river. Copeia 1994:174–183. White, D., K. Johnston, and M. Miller. 2005. Ohio River Basin. Pp. 375–424, In A.C. Benke and C.E. Cushing (Eds.). Rivers of North America. Elsevier Academic Press, Burlington, MA. [# PP.?] Winemiller, K.O., and K.A. Rose. 1992. Patterns of life-history diversification in North American fishes: Implications for population regulation. Canadian Journal of Fisheries and Aquatic Sciences 49:2196–2218. Winemiller, K.O., S. Tarim, D. Shormann, and J.B. Cotner. 2000. Fish-assemblage structure in relation to environmental variation among Brazos River oxbow lakes. Transactions of the American Fisheries Society 129:451–468. Wren, D.G., G.R. Davidson, W.G. Walker, and S.J. Galicki. 2008. The evolution of an oxbow lake in the Mississippi alluvial floodplain. Journal of Soil and Water Conservation 63:129–135. Zeug, S.C., and K.O. Winemiller. 2008. Relationships between hydrology, spatial heterogeneity, and fish-recruitment dynamics in a temperate floodplain river. River Research and Applicatons 24:90–102.