2009 SOUTHEASTERN NATURALIST 8(1):23–36 Infl uence of Small Impoundments on Habitat and Fish Communities in Headwater Streams Michael T. Kashiwagi1 and Leandro E. Miranda1,* Abstract - We surveyed the habitat and fish assemblages of four impounded and three unimpounded neighboring headwater streams, separated longitudinally into multiple upstream and downstream reaches. Instream habitat characteristics were similar between reaches of unimpounded streams and reaches above impoundments, and differed significantly from reaches below impoundments that included deeper water and more stable fl ows. Species richness was similar above and below impoundments, and between impounded and unimpounded streams, but fish assemblage composition and structure differed. Stream reaches above impoundments supported higher percentages of centrarchids compared to upper reaches of unimpounded streams, which had more obligate stream cyprinids. Reaches below impoundments supported mainly centrarchid species whereas lower reaches of unimpounded streams supported a balanced mix of cyprinids and centrarchids. Percina maculata (Blackside Darter) occurred throughout the study area except in upper reaches of impounded streams, illustrating how stream fragmentation can lead to localized extirpations. Changes to the fish assemblages in reaches above impoundments were due to the loss of downstream connectivity, and changes to the fish assemblages below impoundments were due to alterations of instream habitat caused by the impoundments. Small impoundments can have important effects on fish faunas of small geographical areas, but also potentially large cumulative effects if distribution of impoundments is not administered strategically at the scale of the river basin. We caution that continual population increase, recent droughts, and projected changes in climate patterns are prompting a renewed interest in impoundment construction, and urge close regulatory oversight over such projects. Introduction Low-order (order 1–2) headwater streams account for the majority of watercourses traversing terrestrial landscapes, and provide essential habitat for lotic fish communities. In the United States, low-order streams represent about 85% of all streams (Leopold et al. 1964). Relative to the more hydrologically stable high-order streams, low-order streams generally have fl ashy fl ows and may dry up seasonally. The fish communities in headwater streams refl ect this environmental variability (Meffe and Berra 1988, Schlosser 1995), supporting simple fish assemblages with few species (Horwitz 1978, Schlosser 1987). Nevertheless, the large number of streams combines with spatial habitat heterogeneity to sustain high gamma diversity and overall ecosystem biodiversity (Gomi et al. 2002). According to the National Inventory of Dams (2007), there are over 20,000 impoundments ranging between 10 and 200 ha in the United States. 1Mississippi Cooperative Fish and Wildlife Research Unit, US Geological Survey, PO Box 9691, Mississippi State, MS 39762. *Corresponding author - smiranda@ cfr.msstate.edu. 24 Southeastern Naturalist Vol. 8, No. 1 These small impoundments have generally been constructed on low-order headwater streams. Proliferation of these small impoundments has generally received little or no scrutiny by regulatory agencies, which have focused mainly on large impoundments that dam higher order streams and large volumes of water. Large impoundments are usually more visible and interrupt the movement of large migratory fish that attract public attention because of their commercial, recreational, or cultural value. Nevertheless, small impoundments fragment headwater streams and can also disrupt fish communities. Distresses occur both upstream and downstream of impoundments, through several mechanisms including isolation of upstream tributaries from their downstream reaches, alteration of seasonal fl ow patterns below the impoundment, and modification of habitat characteristics both above and below the impoundment (Yeager 1993). These environmental changes can affect fish communities upstream by preventing recolonization after droughts, resulting in the extirpation of species unable to find refuge in the impoundment (Reyes-Gavilan et al. 1996, Winston et al. 1991), by changing fish abundances (Erman 1973), and by shifting assemblage composition (Pyron et al. 1998), reportedly from fl uvial specialists to macrohabitat generalists (Herbert and Gelwick 2003). Downstream effects on fish communities include reduced species richness and diversity (Edwards 1978), increased species richness and habitat alteration (Taylor et al. 2001), and the establishment of reservoir-adapted species (Swink and Jacobs 1983). We surveyed the fish assemblages of several impounded and unimpounded headwater streams. We predicted that some fish species would be extirpated from streams above impoundments due to the loss of downstream connectivity, that fish species richness would be reduced, and that fish assemblages would be altered. We also predicted that in reaches below impoundments, the stream would regain their original richness and natural fish assemblage. To test these predictions, we surveyed four impounded and three unimpounded reference headwater streams situated within a confined 170-km2 geographical area within the same drainage, separated the streams longitudinally into upstream and downstream reaches, and compared fish assemblages among reaches to reveal the potential effects of the dams. We included surveys and analyses of selected physical habitat conditions to account for infl uences they might have that could potentially obscure the effects of the impoundments. Methods Study site Study streams were located either in or adjacent to the Tombigbee National Forest in northeast Mississippi (Fig. 1). Streams were classified into impounded and unimpounded. Four impounded streams were considered, including Redland (11 km long, 28 km2 drainage area), Mill (7 km, 14 km2), Goodfood (8 km, 22 km2), and Owl (7 km, 13 km2) creeks, each with a single impoundment with surface area of 70, 25, 33, and 40 ha, respectively. These impoundments have variable surface-release discharge that allow relatively 2009 M.T. Kashiwagi and L.E. Miranda 25 stable releases throughout the year, and were built in the mid-1960s for fl ood control. Unimpounded streams included Dicks (12 km long, 27 km2 drainage area), Cane (9 km, 13 km2), and Soctahoma (14 km, 48 km2) creeks. The watersheds for these streams included a mixture of forests, pastures, and agricultural lands in descending order of prevalence. Sampling sites Sampling reaches were established in the upper and lower sections of each stream. In the impounded streams, the upper and lower sections were separated by the impoundments. In the unimpounded streams, the length of the stream was separated into approximately equal thirds, and sampling conducted in the upper and lower thirds. Either two or three 100-m long reaches, each reach length representing roughly 20 times the mean wetted width, were sampled in the upper and in the lower sections of each stream to characterize instream habitat and the fish community. In all, 39 reaches were sampled from the seven study streams. Instream habitat Habitat measurements were based on the methods of Gorman and Karr (1978) with minor modifications. Stream wetted width, bankful width, depth, Figure 1. Location of the study streams in Pontotoc and Chickasaw counties, MS, including four impounded and three unimpounded streams, position of the impoundments, and the upper (U) and lower (L) sections of each stream. 26 Southeastern Naturalist Vol. 8, No. 1 and current velocity were measured along horizontal transects. Four transects spaced at 25-m intervals were included in each 100-m stream reach. Measurements of depth and current velocity at 0.6 times the depth were obtained every 1 m along each transect. Water velocity measurements were made using a Flow-Mate® Model 2000 digital fl ow meter (Marsh-McBirney Inc., Frederick, MD). Transect measurements were averaged to determine mean habitat characteristics for each 100-m stream reach. All instream habitat measurements were made in August–September 2003 and 2004. Selected measurements were taken within the 25-m stream sections between each transect. Percentage composition of channel type was estimated according to three categories of depth and velocity (Aadland 1993, Gorman and Karr 1978), including riffl es (2–20 cm deep, high velocity), runs (2–50 cm deep, medium velocity), and pools (>50 cm deep, low velocity). Instream cover was considered as any permanent (stumps, root wads, large woody debris) and semi-permanent (debris dams, small woody debris) structure. Measurement of structure length and width were made to determine total cover area for each 25-m stream section. Also, the deepest point in a section was determined by walking the section to identify and record maximum depth. Measurements taken in each section were averaged to characterize a reach. Fish collections The sampling protocol was designed to quantitatively assess fish community composition with electrofishing, which has been effectively used to collect fish in wadeable streams (Reynolds 1996, Wiley and Tsai 1983). Each 100-m stream reach was blocked at both ends with a 6-mm mesh seine and electrofished with a single pass using a Smith-Root Model 15-D backpack electrofisher (Smith-Root, Inc., Vancouver, WA). A standard setting of 110 Hz pulsed DC and 4–6 A was used in water that ranged in conductivity from 40 to 60 μS/cm. Captured fish were identified to species before returning them to the water. Any fish that could not be identified in the field was preserved in 10% formalin and identified in the laboratory with the aid of taxonomic keys (Page and Burr 1991, Ross 2001). Fish were collected in May–September 2003 and May–September 2004. Data analyses Mean values were calculated for each stream reach to create a habitat (n = 9 habitat variables) by reach (n = 39 reaches) matrix. We compared single habitat variables with analysis of similarities (ANOSIM; Clarke and Gorley 2006), a permutation-based nonparametric procedure, equivalent to analysis of variance, that analyzes variability among ranked similarity coefficients. Unlike analysis of variance, ANOSIM makes no assumptions about the distribution of the data. Similarity coefficients between every pair of reaches were calculated with the Euclidean distance coefficient, to create a triangular similarity matrix from the original rectangular habitat-by-reach matrix for each habitat variable. This analysis involved multiple two-way ANOSIMs, with location (i.e., upper and lower) and impoundment status (i.e., impounded and unimpounded) as factors, and the ranked similarity coefficient as the test variable. Pairwise tests were conducted among all factor groups by the ANOSIM procedure. 2009 M.T. Kashiwagi and L.E. Miranda 27 We graphically explored the multivariate among-reaches habitat similarity with non-metric multidimensional scaling (NMS; Clarke and Gorley 2006). The NMS procedure represents the reaches as points in two- or threedimensional space such that the relative distances separating all points in the plot are in the same rank order as the similarity coefficients in the similarity matrix (i.e., reaches that have greater similarity considering all nine habitat variables were plotted closer in the graph than reaches that are more dissimilar). Euclidean distance similarities were computed based on all nine habitat characteristics combined to create the among-reach habitat similarity matrix. The large number of variables (i.e., species) in the fish dataset precluded single species comparisons, and thus analyses included only fish assemblage comparisons. A matrix of species densities (average number of fish of each species/100-m reach) by stream reach (n = 39) was constructed. All cells in the matrix were loge transformed to reduce the infl uence of abundant species. Similarity between every pair of reaches was calculated with the Bray-Curtis similarity coefficient to create an among-reaches similarity matrix, to which the NMS procedure was applied. After visually examining the NMS plots of habitat and fish similarities, we tested for statistical differences among sets of reaches with ANOSIM. Firstly, similarities among the upper and lower reaches of the three unimpounded streams were contrasted to test if differences, possibly refl ecting natural gradients, existed. This comparison required a two-way design with location as treatment (i.e., upper and lower), stream as block (three unimpounded streams), and the ranked similarities among habitat measures or among species densities taken in multiple reaches in each stream as replicates. Secondly, the upper reaches of the impounded streams were compared to the upper reaches of the unimpounded streams to test for potential effects of impoundments on the upper reaches. Correspondingly, the lower reaches of the impounded streams were compared to the lower reaches of the unimpounded streams to test for potential effects of impoundments on the stream below the dam. Both of these comparisons also required a two-way design with impoundment as treatment (i.e., impounded and unimpounded), stream as block (seven streams), and the habitat or species ranked similarities among multiple reaches in each stream as replicates. Additionally, the fish communities were described in terms of species richness, defined as the number of species collected in a reach. A two-way ANOSIM design was also applied to this analysis, repeating the interreaches comparisons listed in the previous paragraph. As for the habitat data set, the among-reaches similarity matrix was constructed using Euclidean distance, and the ANOSIM analysis applied to the ranked similarity matrix. Results Instream habitat Headwater streams in the study area were generally small and relatively homogeneous across this confined geographical area. Bankful width and wet width averaged 7.4 m and 4.5 m, respectively. Mean depth and deepest points 28 Southeastern Naturalist Vol. 8, No. 1 averaged 0.17 m and 0.38 m, respectively. Current velocity at 0.6 times the depth averaged 0.07 m/s. Percentage composition of channel type within a reach in terms of riffl e, run, and pool averaged 10, 81, and 8%, respectively. Percentage cover within a reach averaged 10%. Seven of the nine habitat variables showed statistical differences among the test reaches (Table 1). Lower reaches of impounded streams were grouped in the left-central areas of the NMS plot (i.e., low axis 1 scores and low-intermediate axis 2 scores), whereas the remaining reaches were mixed throughout the upper two-thirds of axis 1 and all of axis 2 (Fig. 2). This pattern suggested a greater similarity among sites in the lower reaches of impounded streams, but no differences among the other sites that were associated with a broad range of habitat conditions. The ANOSIM procedure showed that habitats in the lower reaches of impounded streams were indeed different from habitats in other reaches (P < 0.01), but that the other three groups of reaches did not differ among themselves (P > 0.43). A cursory examination of Table 1 indicated that lower reaches in impounded streams contained deeper habitat than upper reaches in impounded streams, or than in reaches in unimpounded streams. Lower reaches in impounded streams had greater average depth (0.24 vs. 0.13 m), mean deepest point (0.49 vs. 0.32 m), and percentage pool (15 vs. 6%) than other reaches. Fish communities In all, 8440 fish representing 28 species in nine families were collected. All species were native to the drainage. Cyprinids accounted for 41% of the collection, with Pimephales notatus (Rafinesque) (Fathead Minnow), Semotilus atromaculatus (Mitchill) (Creek Chub) 8%, and Lythrurus bellus (Hay) (Pretty Shiner) representing 21, 8, and 7% of the collection respectively. Centrarchids represented 48% of fishes caught, with Lepomis macrochirus Rafinesque (Bluegill) and L. cyanellus Rafinesque (Green Sunfish) representing 23% and 21% respectively. The remaining 23 species combined represented 11% of all the fish collected and included the families Clupeidae (n =1 species), Cyprinidae (n = 3), Catostomidae (n = 2), Ictaluridae (n = 3), Table 1. Mean values of habitat variables in upper and lower reaches of impounded (n =4) and unimpounded (n =3) streams. Two-way analyses of similarities (ANOSIM) indicated two variables did not differ among reaches (P > 0.1). Pairwise tests identified various differences among the means of variables that did show differences (means in the same row and followed by the same letter are not significantly different, P < 0.05). Impounded Unimpounded ANOSIM Lower Upper Lower Upper P Bankful width (m) 7.2a 6.7a 9.1b 6.7a <0.01 Wet width (m) 4.8a 4.3b 4.0b 4.1b 0.04 Current velocity (m/s) 0.01a 0.09b 0.07b 0.05b 0.02 Depth (m) 0.24a 0.15b 0.14b 0.12b <0.01 Deepest (m) 0.49a 0.38b 0.38b 0.31b <0.01 Cover (m²) 11.9a 10.1a 9.4a 6.6a 0.73 Percent pool (%) 15.2a 6.8b 7.4b 1.3b 0.01 Percent riffl e (%) 4.0a 10.7b 13.6b 15.6b 0.02 Percent run (%) 80.8a 82.5a 78.9a 83.1a 0.21 2009 M.T. Kashiwagi and L.E. Miranda 29 Aphredoderidae (n = 1), Fundulidae (n = 2), Poeciliidae (n = 1), Centrarchidae (n = 6), and Percidae (n = 4). Certain species were either missing from, or were unique to, impounded streams. The Blackside Darter was found in upper and lower unimpounded reaches and lower reaches of impounded streams, but was not collected in upper reaches of impounded streams. Dorosoma cepedianum (Lesueur) (Gizzard Shad) were collected only below impoundments. Ameiurus natalis (Lesueur) (Yellow Bullhead), which was completely absent from unimpounded streams, was collected from all upper and lower reaches of impounded streams. Examination of family composition over different stream types and locations showed shifts in community assemblages (Fig. 3). In reaches above impoundments, centrarchids comprised 50% of the fish collected, whereas in reaches of unimpounded streams, they accounted for less than 25%, with cyprinids becoming dominant. Similarly, lower reaches in impounded streams were dominated by centrarchids that made up over 75% of the community, whereas in lower reaches of unimpounded streams, cyprinids and centrarchids abundances were more balanced, with these two groups being about equally abundant. Species composition showed strong organization relative to position of the reach (Fig. 4). Two significant ordination axes were identified by NMS, with a stress value of 0.09 (scale = 0–1, with 0 representing an absolute fit, and values <0.1 representing a good fit). Lower reaches Figure 2. Fish habitat similarities (NMS axis scores) in the 39 study reaches classified according to presence or absence of impoundments and position in the stream. Reaches closer together in ordination space have habitat characteristics that are more similar than do those that are farther apart. 30 Southeastern Naturalist Vol. 8, No. 1 of impounded streams had high scores on axis 1 and 2, whereas upper reaches of impounded streams had intermediate-high scores on axis 2 and low scores on axis 1. Scores of unimpounded streams were generally low on axis 1 and axis 2. While overlapping, scores of lower reaches of unimpounded streams were higher on axis 1 and 2 than those of upper reaches in unimpounded streams. The ANOSIM procedure supported the informal examination of the NMS plot. Ranked similarity coefficients of upper reaches relative to lower reaches in the unimpounded streams did not differ significantly (P = 0.34), suggesting that community compositions were similar. However, differences in similarity coefficients existed between the impounded and unimpounded streams, suggesting dissimilar fish communities. Ranked similarity coeffi- cients differed significantly between upper reaches (P = 0.04) of impounded and unimpounded streams, and between lower reaches (P < 0.01) of impounded and unimpounded streams. Figure 3. Relative abundance of fish families in the study streams. The four pie charts represent means for the two stream types (impounded and unimpounded) and reach locations (upper and lower). 2009 M.T. Kashiwagi and L.E. Miranda 31 Despite differences in community composition, species richness was not different among stream reaches. Richness averaged 11.5 species in the upper reaches of unimpounded streams and 12.0 species in lower reaches, and did not differ significantly between the two (P = 0.66). In the upper reaches, richness averaged 12.0 species in impounded streams and 11.5 species in unimpounded streams, and these also did not differ significantly (P = 0.54). Likewise, in the lower reaches, richness averaged 10.4 species in impounded streams and 12.0 species in unimpounded streams, and did not differ signifi- cantly between the two (P = 0.26). Discussion Droughts are major disturbance events in stream environments (Horwitz 1978, Schlosser 1982). During periods of drought, headwater streams experience severely reduced surface fl ows and may dry up. Following a drought disturbance, the stream is recolonized by individuals from the downstream community (Larimore et al. 1959). Experimental and observational studies have demonstrated that fish can rapidly colonize these areas if they have unrestricted access (Lonzarich et al. 1998, Peterson and Bayley 1993). Gauging stations were not available within the study area, but contacts with landowners and conservation officers suggested that surface fl ows stopped approximately 1–3 times per decade. Given these reported drought events, impoundments were expected to hamper species richness. Yet, richness was Figure 4. Fish a s s e m b l a g e s i m i l a r i t i e s (NMS axis scores) in the 39 study reaches classified according to presence or absence of impoundments and position in the stream. Reaches closer together in ordination space have fish assemblages that are more similar than do those that are farther apart. 32 Southeastern Naturalist Vol. 8, No. 1 similar above and below impoundments, and between impounded and unimpounded streams. Although counterintuitive, a constant or an increase in species richness may not be uncommon. Indeed, Guenther and Spacie (2006) reported that part-time use of upstream reaches by reservoir species actually boosted estimates of species richness. Fish assemblage composition differed among stream reaches. Fragmented stream reaches above impoundments supported higher percentages of centrarchids compared to unfragmented upper reaches of unimpounded streams, which had higher percentages of obligate stream cyprinids. The reduced cyprinids assemblages occupying reaches above impoundments were represented by tolerant, generalist species, such as Fathead Minnow and Creek Chub, capable of surviving severe environmental conditions in streams, and unfavorable conditions in upper reaches of impoundments (Smogor and Angermeier 1998). Similarly, Herbert and Gelwick (2003) and Guenther and Spacie (2006) reported a decline of fluvial specialists and an increase in macrohabitat generalists in streams fragmented by impoundments. During drought periods when stream flows are reduced in the study streams, fish in reaches above impoundments are apparently forced downstream to seek shelter in the impoundment, or survive in wetland areas that provide temporary refuge. After resumption of normal stream flows, reaches above impoundments are recolonized by tolerant cyprinids and centrarchid species from the impoundments. In contrast, following droughts in unimpounded streams, the recolonizers of the upper reaches are mainly cyprinid species representative of the downstream fish community. Despite annual fluctuations, fish communities of unimpounded headwater streams can remain fairly stable over time (Moyle and Vondracek 1985, Ross et al. 1985), but require connectivity with the downstream community to preserve their integrity. The absence of Blackside Darter in upper reaches of impounded streams illustrates how localized extirpation might occur due to the loss of downstream connectivity. Blackside Darter is a common species with a widespread distribution throughout Mississippi (Ross 2001). Although collected in unimpounded and lower impounded reaches, the species was absent from upper impounded reaches. Localized extirpation of this species from streams above impoundments most likely results from the species’ inability to survive in the stream or in the adjacent impoundments during drought periods, or from predation by pool-dwelling fish or reptiles. Following resumption of fl ow, the upstream movement of recolonizers is blocked by the impoundments, effectively preventing reestablishment of populations in upper reaches of impounded streams. By comparison, in unimpounded streams, recolonizers were apparently able to reestablish populations of this species along the entire length of the streams. Similar extirpations caused by the construction of small impoundments have been reported for several minnow species in Oklahoma by Winston et al. (1991), and for Notropis topeka (Gilbert) (Topeka Shiner) in Kansas by Schrank et al. (2001). Local extirpations have also been reported in larger streams with major reservoirs (e.g., Guenther and Spacie 2006, Luttrel et al. 1999), but such losses are less common 2009 M.T. Kashiwagi and L.E. Miranda 33 (e.g., Herbert and Gelwick 2003). The fish assemblages in reaches below impoundments were shaped by the alterations to the instream habitat caused by the impoundments. These reaches were deeper and contained a higher percentage of pool habitat, alterations resulting from changes to stream hydrology caused by the impoundments. The natural hydrology marked by periods of extreme high fl ows is restrained by the storage capacity of fl oodcontrol impoundments, by reducing the short- and long-term fl ow variability of downstream reaches (Petts 1984). The four impoundments were designed to gradually release the stored water, resulting in moderately elevated discharge levels over longer periods of time. This release mode likely reduced short-term bankful-width and out-of-bank fl ows, but prolonged wet-width fl ows as suggested by our results. Moreover, reaches below the dam showed physical signs of incision, which, according to Grams et al. (2007), produce narrower and deeper channels. The overall result is an increase in water levels above historic basefl ow conditions, producing the deeper water and pool conditions observed in reaches below impoundments. The fish assemblages recorded in reaches below impoundments refl ect the changes in habitat conditions. These reaches supported mainly centrarchid species whereas lower reaches of unimpounded streams supported a more balanced mix of cyprinids and centrarchids. Stream centrarchids are often associated with increasing average depth (Aadland 1993). Deep pools provide areas of refuge and environments with greater temporal stability (Schlosser 1987). Also, the increased volume of the deeper habitat provides more physical space, and therefore supports larger fish (Harvey and Stewart 1991). Thus, besides centrarchids, the pools in reaches below impoundments supported large-sized species such as Moxostoma poecilurum (Jordan) (Blacktail Redhorse) and Ictalurus punctatus (Rafinesque) (Channel Catfish), which, while not collected in large numbers, were found primarily in these belowimpoundment reaches. Other species present in pools below impoundments may be escapees from the impoundments, such as Gizzard Shad and Yellow Bullhead that may be able to survive in the deeper habitat conditions provided by pools below impoundments but not in shallower unimpounded streams. However, informal observations indicated that the infl uence of impoundments on downstream habitats extended no more than 5–10 km, after which habitat conditions seemed to return to their natural state. Construction of impoundments in the United States has slowed signifi- cantly since the 1970s, owing to escalating construction costs, scarcity of suitable sites, and increased concern over environmental effects (Miranda 1996). Although the intensity of construction observed in the twentieth century is not likely to return, continued population increase, recent droughts, and projected changes in climate patterns are prompting a renewed interest on impoundment construction. Global warming scenarios predict a possible decrease in precipitation and increase in evaporation (Jacobs et al. 2000), which together with sediment accumulations in our aging reservoirs is likely to propel new constructions. This trend is most visible in large southern metropolitan regions such as Atlanta and Dallas, where water development boards are actively considering new reservoirs (Sabine River Authority 34 Southeastern Naturalist Vol. 8, No. 1 1999, Sutherland 2003), although new constructions are being considered and pursued throughout the Southeastern region. Our results indicate that small impoundments can have important effects on the fish faunas of small geographical areas, but potentially large cumulative effects if distribution of impoundments is not administered strategically at the scale of the river basin. Thus, there is a need for increased environmental analysis and close regulatory oversight of such projects. Impoundments should be avoided in streams with headwaters having particularly diverse habitats and/or rich faunas, or that include species of special concern. Literature Cited Aadland, L.P. 1993. Stream habitat types: Their fish assemblages and relationship to fl ow. North American Journal of Fisheries Management 13:790–806. Clarke, K.R., and R.N. Gorley. 2006. PRIMER version 6: Users manual/tutorial. PRIMER-E, Plymouth, UK. 145 pp. Edwards, R.J. 1978. The effect of hypolimnion reservoir release on fish distribution and species diversity. Transactions of the American Fisheries Society 107:71–77. Erman, D.C. 1973. Upstream changes in fish populations following impoundment of Sagehen Creek, California. Transactions of the American Fisheries Society 102:626–628. Gomi, T., R.C. Sidle, and J.S. Richardson. 2002. Understanding processes and downstream linkages of headwater systems. BioScience 52:905–916. Gorman, O.T., and J.R. Karr. 1978. Habitat structure and stream fish communities. Ecology 59:507–515. Grams, P.E., J.C. Schmidt, and D.J. Topping. 2007. The rate and pattern of bed incision and bank adjustment on the Colorado River in Glen Canyon downstream from Glen Canyon Dam, 1956–2000. Geological Society of America Bulletin 119:556–575. Guenther, C.B., and A. Spacie. 2006. Changes in fish assemblage structure upstream of impoundments within the Upper Wabash River Basin. Transactions of the American Fisheries Society 135:570–583. Harvey, B.C., and A.J. Stewart. 1991. Fish size and habitat depth relationship in headwater stream. Oecologia 87:336–342. Herbert, M.E., and F.P. Gelwick. 2003. Spatial variation of headwater fish assemblages explained by hydrologic variability and upstream effects of impoundment. Copeia 2003:273–284. Horwitz, R.J. 1978. Temporal variability patterns and the distributional patterns of stream fishes. Ecological Monographs 48:307–321. Jacobs, K., D. Adams, and P. Gleick. 2000. Potential consequences of climate variability and changes for the water resources of the United States. Report by the National Assessment and Synthesis Team, US Global Change Research. Available online at http://www.usgcrp.gov/usgcrp/Library/nationalassessment/14Water. pdf. Accessed January 15, 2007. Larimore, R.W., W.F. Childers, and C. Heckrotte. 1959. Destruction and re-establishment of stream fish and invertebrates affected by drought. Transactions of the American Fisheries Society 88:26–285. Leopold, L.B., M.G. Wolman, and J.P. Miller. 1964. Fluvial processes in geomorphology. Freeman, San Francisco, CA. 522pp. 2009 M.T. Kashiwagi and L.E. Miranda 35 Lonzarich, D.G., M.L. Warren, Jr., and M.R.E. Lonzarich. 1998. Effects of habitat isolation on the recovery of fish assemblages in experimentally defaunated stream pools in Arkansas. Canadian Journal of Fisheries and Aquatic Science 55:2141–2149. Luttrell, G.R., A.A. Echelle, W.L. Fisher, and D.J. Eisenhour. 1999. Declining status of two species of the Macrhybopsis aestivalis complex (Teleostei:Cyprinidae) in the Arkansas River basin and related effects of reservoirs as barriers to dispersal. Copeia 1999:981–989. Meffe, G.K., and T.M. Berra. 1988. Temporal characteristics of fish assemblage structure in an Ohio stream. Copeia 1988:684–690. Miranda, L.E. 1996. Development of reservoir fisheries management paradigms in the twentieth century. Pp. 3–11, In L.E. Miranda and D.R. DeVries (Eds.). Multidimensional Approaches to Reservoir Fisheries Management. American Fisheries Society Symposium 16, Bethesda, MD. 463 pp. Moyle, P.B., and B. Vondracek. 1985. Persistence and structure of the fish assemblage in a small California stream. Ecology 66:1–13. National Inventory of Dams. 2007. National Inventory of Dams. Engineer Research and Development Center, US Army Corps of Engineers. Available online at http://www.crunch.tec.army.mil/nid/webpages/nid.cfm. Accessed December 12, 2007. Page, L.M., and B.M. Burr. 1991. A Field Guide to Freshwater Fishes. Houghton Miffl in, New York, NY. 432 pp. Peterson, J.T., and P.B. Bayley. 1993. Colonization rates of fishes in experimentally defaunated warmwater streams. Transactions of the American Fisheries Society 122:199–207 Petts, G.E. 1984. Impounded rivers: Perspectives for ecological management. Wiley, New York, NY. 326 pp. Pyron, M., C.C. Vaughn, M.R. Winston, and J. Pigg. 1998. Fish assemblage structure from 20 years of collections in the Kiamichi River, Oklahoma. Southwestern Naturalist 43:336–343. Reyes-Gavilan, F.G., R. Garrido, A G. Nicieza, M.M. Toledo, and F. Brana. 1996. Fish community variation along physical gradients in short streams of northern Spain and the disruptive effect of dams. Hydrobiologia 321:155–163. Reynolds, J.B. 1996. Electrofishing. Pp. 238–240, In B.R. Murphy, and D.W. Willis (Eds). Fisheries Techniques, 2nd Edition. American Fisheries Society, Bethesda, MD. 732 pp. Ross, S.T. 2001. Inland fishes of Mississippi. Mississippi Department of Wildlife, Fisheries, and Parks, Jackson, MS. 548 pp. Ross, S.T., W.J. Matthews, and A.A. Echelle. 1985. Persistence of stream fish assemblages: Effects of environmental change. The American Naturalist 126:24–40. Sabine River Authority. 1999. Comprehensive Sabine watershed management plan. Prepared by Freese and Nichols, Inc. for the Sabine River Authority of Texas in Conjunction with the Texas Water Development Board. Available online at http://www.sratx.org/srwmp/comprehensive_plan/final_report/pdf/CSWMP.pdf. Accessed December 11, 2007. Schlosser, I.J. 1982. Fish community structure and function along two habitat gradients in a headwater stream. Ecological Monographs 52:395–414. Schlosser, I.J. 1987. A conceptual framework for fish communities in small warmwater streams. Pp. 17–24, In W.J. Matthews and D.C. Heins (Eds.). Community and Evolutionary Ecology of North American Stream Fishes. University of Oklahoma Press, Norman, OK. 310 pp. 36 Southeastern Naturalist Vol. 8, No. 1 Schlosser, I.J. 1995. Critical landscape attributes that infl uence fish population dynamics in headwater streams. Hydrobiologia 303:71–81. Schrank, S.J., C.S. Guy, M.R. Whiles, and B.L. Brock. 2001. Infl uence of instream and landscape-level factors on the distribution of Topeka Shiners Notropis topeka in Kansas streams. Copeia 2001:413–421. Smogor, R.A., and P.L. Angermeier. 1998. Effects of drainage basin and anthropogenic disturbances on relations between stream size and IBI metrics in Virginia. Pp. 257–260, In T.P., Simon, editor. Assessing the Sustainability and Biological Integrity of Water Resources using Fish Communities. CRC Press, New York, NY. 671 pp. Sutherland, E. 2003. Water supply reservoirs in Georgia: An update on Henry County’s proposed Tussahaw Creek reservoir and the importance of thorough environmental impacts analyses.Pp. 318–320, In K.J. Hatcher (Ed). Proceedings of the Georgia Water Resources Conference, The University of Georgia, Athens, GA. Swink W.D., and K.E. Jacobs. 1983. Infl uence of a Kentucky fl ood-control reservoir on the tailwater and headwater fish populations. North American Journal of Fisheries Management 3:197–203. Taylor, C.A., J.H. Knouft, and T.M. Hiland. 2001. Consequences of stream impoundment on fish communities in a small North American drainage. Regulated Rivers:Research and Management 17:687–698. Wiley, M.L., and C.F. Tsai. 1983. The relative efficiencies of electrofishing vs. seines in piedmont streams of Maryland. North American Journal of Fisheries Management 3:243–253. Winston, M.R., C.M. Taylor, and J. Pigg. 1991. Upstream extirpation of four minnow species due to damming of a prairie stream. Transactions of the American Fisheries Society 120:98–105. Yeager, B.L. 1993. Dams. Pp. 57–113, In C.F. Bryan, and D.A. Rutherford (Eds). Impacts on Warmwater Streams: Guidelines for Evaluation. American Fisheries Society, Bethesda, MD. 285 pp.