Abstract
Using seasonally collected data (2009-2010) from 15 sampling sites that represent first- to fifth-order streams within the Qingyi watershed, we examined the spatio-temporal patterns of fish assemblages along two longitudinal gradients to explore the effects of a large dam on fish assemblages at the watershed scale. No significant variation was observed in either species richness or assemblage structure across seasons. Species richness significantly varied according to stream order and gradient. Dam construction appeared to decrease species richness upstream substantially, while a significant decrease between gradients only occurred within fourth-order streams. Along the gradient without the large dam, fish assemblage structures presented distinct separation between two neighboring stream orders, with the exception of fourth-order versus fifth-order streams. However, the gradient disrupted by a large dam displayed the opposite pattern in the spatial variation of fish assemblages related with stream orders. Significant between-gradient differences in fish assemblage structures were only observed within fourth-order streams. Species distributions were determined by local habitat environmental factors, including elevation, substrate, water depth, current discharge, wetted width, and conductivity. Our results suggested that dam construction might alter the longitudinal pattern in fish species richness and assemblage structure in Qingyi Stream, despite the localized nature of the ecological effect of dams.
Keywords: Fish assemblage, Species richness, Spatiotemporal pattern, Longitudinal gradient, Dam building
Stream fish assemblages are structured by abiotic factors, biotic interactions, and historical processes (Hoeinghaus et al, 2007). The physicochemical stream environment exhibits spatial heterogeneity and temporal variability, and may cause spatiotemporal variations in fish assemblages (Meador & Matthews, 1992). Typically, among the multiple spatial factors affecting fish assemblages, longitudinal-gradient variations in environmental conditions substantially influence the distribution and abundance of stream fishes (Araújo et al, 2009; Inoue & Nunokawa, 2002; Taniguchi et al, 1998; Vannote et al, 1980). From headwaters to downstream, fish species richness generally increases due to increases in stream size (Matthews, 1986), habitat diversity (Gorman & Karr, 1978), and shelter availability (da Silva Abes & Agostinho, 2001), and the different rates of fish immigration and extinction (Power et al, 1988). Concurrently, fish species composition may vary longitudinally by species addition and/orspecies replacement (Boys & Thoms, 2006; Gorman & Karr, 1978; Roberts & Hitt, 2010). The River Continuum Concept (RCC) describes the changes in community structure and species richness of organisms from headwaters to mouth waters and relates these changes to flow regime, water temperature, food availability, and substrate conditions (Vannote et al, 1980). However, due to various human activities (e.g., damming, pollution, erosion), few riverine ecosystems remain free-flowing over their entire course. Overlaying this pattern displayed by the RCC, the Serial Discontinuity Concept (SDC) claims that regulation by dams disrupts the underlying continuum and produces a series of lentic and lotic reaches, thereafter causing longitudinal shifts in abiotic and biotic parameters and processes (Ward & Stanford, 1983).
Dams impact stream fishes and invertebrates in diverse ways, such as blocking migratory pathways (March et al, 2003), fragmentizing habitats (Travnichek et al, 1993), altering natural flow regimes (Bonner & Wilde, 2000) and food webs (Power et al, 1996), decreasing water temperature downstream (Clarkson & Childs, 2000) and current velocity upstream (Bennett et al, 2002), disrupting riparian plant communities (Nilsson et al, 1997), and shifting water chemistry (Humborg et al, 1997). The extent to which stream fishes and invertebrates are affected by dams may be associated with the characteristics of dams (e.g., location, purpose, and management) and fauna (Cumming, 2004; March et al, 2003). For example, large dams without spillways are impermeable barriers for migratory organisms and may extirpate all native migratory fish from upstream habitat (Holmquist et al, 1998). In contrast, large dams with spillways provide a possible passage for some native fishes. To date, investigations have primarily identified the effects of dams on lotic reaches directly below dams, mainstream reservoirs directly above dams, and lotic reaches upstream of impoundments (Clarkson & Child, 2000; Cumming, 2004; Santucci et al, 2005; Travnichek et al, 1993). However, little attention has been paid to the effects of dams at the watershed scale.
Stream ecosystems in China are experiencing massive ecological perturbation due to diverse human activities, such as extensive agricultural and industrial production, urbanization, development, and hydraulic engineering construction, all of which threaten freshwater fish (Chen, 2005). Based on data collected from first- to fifth-order streams, representing two longitudinal gradients (one disrupted by a large dam and the other not) within a tributary (Qingyi Stream) of the lower reaches of the Yangtze River, the spatial and temporal patterns of fish assemblages were examined. Totally, our three key goals were to: (1) to determine how fish assemblage patterns vary longitudinally from first- to fifth-order streams; (2) to identify the correlations between local habitat and fish assemblages; and (3) to assess how a large dam affects the longitudinal patterns of fish assemblages at the watershed scale.
MATERIALS AND METHODS
Study area
Qingyi Stream originates from the northern portion of Huangshan Mountain, and flows northeast toward its confluence with the lower reaches of the Yangtze River. This watershed is approximately 309 km long and covers an area of 7 195 km2. Due to the subtropical monsoon climate, temperature and precipitation in this area are quite asymmetric across seasons. Annual air temperature ranges from -2.1 °C (January) to 27.5 °C (July), with a mean of 17.8 °C. Annual precipitation is high (approximating 2 000 mm/year), with most rainfall (79%) occurring in spring and summer (April to September) and only minor rainfall (less than 5%) occurring in the cold, dry winter (December to February) (Yan et al, 2011).
Chencun Hydropower Station was constructed at the mid-reaches of the Qingyi mainstream in the 1950s. It is the largest station along the lower reaches of the Yangtze River, with 119 m deep impoundment and covering an area of 88.6 km2. Derived from Anhui Province topographic maps (1:300 000 scale), the Qingyi Stream is categorized as fifth-order (Strahler, 1957) at its mainstream. A total of 75 first-order, 34 second-order, 12 third-order, 4 fourth-order, and 1 fifth-order stream are included in this watershed. Among the fourth-order streams, the Qingxi, Shuxi, and Machuan streams flow into an artificial reservoir (i.e., Taiping Lake) formed by the impoundment of Chencun Hydropower Station, while Huishui Stream directly flows into the mainstream with its confluence downstream of the station (Figure 1).
Figure 1.
Map of the Qingyi Stream watershed in Anhui Province, China
Solid circles mark the 15 sampling sites. Sites 1-3, 4-6, 7-9, 10-12, and 13-15 represent first-, second-, third, fourth-, and fifth-order streams, respectively. Sites 1, 2, and 3 are located at Qingxi (QS), Shuxi (SS), and Huishui Streams (HS), respectively. >-< shows Chencun Hydropower Station.
Fish sampling
A total of 15 segments representing first- to fifth-order streams were surveyed seasonally during April, August, and October 2009, and January 2010. The sampling segments were derived from two longitudinal gradients. One gradient (A) was set from the headwaters of Qingxi and Shuxi streams to the lower reaches of the mainstream, representing the gradient disrupted by Chencun Hydropower Station, and the other gradient (B) was set from the headwaters of Huishui Stream to the mainstream, representing the gradient not disrupted by a station. Each surveyed segment was set on the topographic map based on their stream orders. Four segments ranging from first- to fourth-order streams were selected within Qingxi, Shuxi, and Huishui streams, respectively. Three other segments representing fifth-order streams were set at the mainstream (Figure 1). Sampling sites were selected in the field based on habitat representativeness and accessibility. Each shallow-water (<1.0 m depth) site of first-, second-, and third-order streams was sampled using backpack electro-fishing by wading in two passes. Each site encompassed at least two mesohabitat units (pool and riffle). Each deep-water (>1.0 m depth) site of fourth- and fifth-order streams was sampled twice using boat electro-fishing along both riversides. Sampling sites were far (>1.0 km distance) from conspicuous human disturbance, such as dam, farmland, and urban land, and each sampling was conducted with comparable effort (i.e., 100 m long and approximately 40 min). Fish were identified to species level (except for Ctenogobius due to the deficiency in identifying tools), counted, and returned to sampling sites if alive. Voucher specimens were placed in 8% formaldehyde solution for further identification.
Environmental survey
Each sampling site was characterized using 10 variables to describe local habitat: elevation (m), wetted width (m), water depth (m), water temperature (°C), pH, conductivity (mS/cm), dissolved oxygen (mg/L), current velocity (m/s), discharge (m3/s), and substrate. Elevation was determined by a portable GPS receiver. Wetted width was measured along five transects, regularly spaced across the stream channel. Water depth, temperature, dissolved oxygen, pH, and conductivity were surveyed at four equal-interval points along each transect. Water depths <1.0 m were measured using a graduated wading rod, and depths >1.0 m were measured using a supersonic echo sounder. Current velocity was taken at 60% of water depth at each point of each transect. Discharge of each channel was determined at the transect that yielded the most accurate measurement (smooth bottom and laminar flow). Along each transect, the proportion of substrate categories (particle size 1= 0-1 mm; 2=1-5 mm; 3=5-25 mm; 4=25-50 mm; 5=50-100 mm; 6=100-500 mm; 7=500-1 000 mm; 8=>1 000 mm) was visually estimated, and an index of substrate coarseness ranging from 1 to 8 was derived for each site following Bain et al (1985).
Data analysis
Each fish species collected in this study was examined for frequency of occurrence (F) and relative abundance (RA), estimated from Fi=100(Si/S)% and Pi=100(ni/n)%, respectively, where Si and S are the abundance of the samplings of which species i were collected and of total samplings, respectively, and ni and n are the individual numbers of species i and total fish species, respectively. The above F and P values were determined independently for three groups of sampling segments, i.e., first- to fourth-order streams along gradient A and B and fifth-order streams, respectively. Species richness and fish abundance were analyzed using a separate three-way ANOVA model with stream order (first- to fifth-order), season (January, April, August, October), and dam (gradient A and B) as factors. One-way ANOVA was used to test differences in species richness across stream orders along gradients A and B, respectively. The Newman-Keuls test was used for post-hoc comparisons after ANOVA. Independent sample t-tests were used to compare species richness within the same ordered streams between the two gradients. All data were log-transformed to meet the assumptions of normality and homogeneity of variances.
Discrete spatial patterns in fish assemblages were identified using PRIMER 5 (Primer-E Ltd., 2001). Following a Bray-Curtis similarity matrix calculation, an analysis of similarity (ANOSIM) was used to test variations in fish assemblages across stream orders, seasons, and gradients. Firstly, variations associated with stream orders and seasons were analyzed separately for gradients A (two-way crossed ANOSIM) and B (one-way ANOSIM due to no replication), respectively. Secondly, variations associated with gradients and seasons were analyzed separately for each stream order using two-way crossed ANOSIM. The relationships among assemblages from each site and season were graphically presented using non-metric multi-dimensional scaling (NMS) analysis. The contribution of each species to differences among assemblage groups was identified using similarity percentages analysis (SIMPER) (Clarke & Warwick, 2001). The variability in fish assemblages in relation to local habitat environment was evaluated by canonical correspondence analysis (CCA) using CANOCA 4.5. All variables entered the CCA after a forward-selection procedure, showing their importance in explaining total variability in species composition. The significance (P<0.05) of the CCA gradients was assessed by Monte Carlo permutation tests, and their importance measured by the eigenvalues of the first two axes (ter Braak & Verdonschot, 1995). All variables of fish assemblages (i.e., species richness and abundance) and habitat environment were log(X+1) transformed to meet assumptions of multivariate normality and to moderate the influence of extreme data. Species that occurred in less than two sites were excluded from the above analysis to avoid negligible weighting (Gauch, 1982).
RESULTS
Overview of species diversity
A total of 13 647 fish were captured throughout this study, representing 57 species, 15 families, and 5 orders. Species of family Cyprinidae comprised on average 59.6% of total species richness Species richness per sampling site amounted to 11.9±6.6 (mean±SD) species, and abundance per site was 227.6±205.5 specimens. A total of 16, 24, 24, 47, and 44 species were collected in first- to fifth-order streams, respectively. Cyprinus carpio, Mylopharyngodon piceus, Hemiculter leucisculus, Sarcocheilichthys sinensis, Abbottina obtusirostris, Macropodus chinensis, Mystus macropterus and Hyporhamphus intermedius were only collected within fifth-order streams. Within the first- to fourth-order streams, 18 species were only collected from streams along gradient B, including Distoechodon tumirostris, Elopichthys bambusa, and Spinibarbus hollandi. Saurogobio dabryi was a unique species occurring in gradient A but not gradient B. Zacco platypus, Acrossocheilus fasciatus, Pseudogobio vaillanti, Vanmanenia stenosoma and Ctenogobius sp. were common (>40% of F) and relatively abundant (>1% of P) within both gradients. In addition, Acheilognathus barbatulus, A. chankaensis, Rhodeus ocellatus, Squalidus argentatus and Misgurnus anguillicaudatus were more frequent and abundant within gradient A, while Opsariichthys bidens, Hemiculter bleekeri, Sarcocheilichthys parvus, Parabotia fasciata, Siniperca chuatsi, Odontobutis obscurus, Silurus asotus and Pseudobagrus albomarginatus were more frequent and abundant in gradient B (Table 1).
Table 1.
Frequency of occurrence (%; left of dash) and relative abundance (%; right of dash) of fish collected within Qingyi Stream
| Species | Code | First to fourth-order | Fifth-order | |
|---|---|---|---|---|
| Gradient A | Gradient B | |||
| Cypriniformes | ||||
| Abbottina rivularis | ABR | 50.0/1.7 | 31.3/2.3 | 83.3/2.9 |
| Abbottina tafangensis | ABT* | 3.1/<0.1 | 6.3/0.1 | |
| Abbottina obtusirostris | ABO* | 16.7/0.2 | ||
| Acheilognathus barbatulus | ACB | 53.1/3.9 | 12.5/0.3 | 58.3/0.8 |
| Acheilognathus chankaensis | ACC | 18.8/0.9 | 6.3/0.3 | 41.7/0.5 |
| Acrossocheilus fasciatus | ACF | 68.8/7.7 | 87.5/10.1 | 25.0/0.3 |
| Carassius auratus | CAA | 28.1/1.8 | 31.3/2.1 | 91.7/5.8 |
| Cobitis rarus | COR | 53.1/5.6 | 37.5/2.2 | |
| Cobitis sinensis | COS | 28.1/0.6 | 18.8/1.3 | 8.3/<0.1 |
| Cyprinus carpio | CYC* | 16.7/0.1 | ||
| Culter erythropterus | CUE* | 6.3/<0.1 | ||
| Distoechodon tumirostris | DIT* | 6.3/0.1 | ||
| Elopichthys bambusa | ELB* | 6.3/<0.1 | ||
| Erythroculter ilishaeformis | ERI | 6.3/0.2 | 41.7/2.3 | |
| Gnathopogon taeniellus | GNT | 18.8/0.6 | 25.0/0.4 | |
| Gobiobotia tungi | GOT* | 12.5/0.1 | ||
| Hemiculter leucisculus | HEL | 75.0/7.9 | ||
| Hemiculter bleekeri | HEB | 6.3/0.2 | 31.3/4.3 | 58.3/1.2 |
| Hemibarbus labeo | HEL | 18.8/0.4 | ||
| Hemibarbus maculatus | HEM | 31.3/0.7 | 16.7/0.1 | |
| Misgurnus anguillicaudatus | MIA | 37.5/1.1 | 18.8/0.4 | 8.3/<0.1 |
| Mylopharyngodon piceus | MYP | 41.7/0.8 | ||
| Opsariichthys bidens | OPB | 21.9/1.0 | 56.3/1.6 | 66.7/2.1 |
| Phoxinus oxycephalus | PHO | 12.5/0.2 | 18.8/3.8 | |
| Pseudobrama simoni | PSS | 18.8/1.6 | 58.3/3.8 | |
| Rhodeus ocellatus | RHO | 65.6/13.9 | 31.3/2.3 | 66.7/3.8 |
| Spinibarbus hollandi | SPH* | 12.5/0.1 | ||
| Zacco platypus | ZAP | 100/34.7 | 93.8/26.0 | 58.3/5.5 |
| Parabramis pekinensis | PAP | 6.3/0.1 | 16.7/0.1 | |
| Parabotia fasciata | PAF | 18.8/0.6 | 31.3/2.0 | 33.3/5.0 |
| Pseudorasbora parva | PSP | 15.6/0.6 | 25.0/2.7 | 50.0/2.2 |
| Pseudogobio vaillanti | PSV | 43.8/2.5 | 43.8/4.8 | 33.3/0.5 |
| Sarcocheilichthys sinensis | SAS* | 8.3/<0.1 | ||
| Sarcocheilichthys parvus | SAP | 3.1/0.1 | 25.0/2.9 | 66.7/2.8 |
| Sarcocheilichthys nigripinnis | SAN | 25.0/4.0 | 75.0/2.3 | |
| Squalidus argentatus | SQA | 34.4/1.2 | 12.5/0.7 | 33.3/1.0 |
| Saurogobio dabryi | SAD | 6.3/0.1 | 8.3/<0.1 | |
| Varicorhinus barbatulus | VAB* | 12.5/0.2 | ||
| Vanmanenia stenosoma | VAS | 65.6/5.2 | 50.0/3.9 | 25.0/1.1 |
| Perciformes | ||||
| Channa argus | CHA | 18.8/1.4 | 91.7/6.1 | |
| Channa asiatica | CHA* | 6.3/0.1 | 8.3/0.1 | |
| Siniperea roulei | COR | 12.5/0.3 | 33.3/0.3 | |
| Ctenogobius sp. | CTS | 93.8/12.6 | 68.8/4.4 | 25.0/2.0 |
| Hypseleotris swinhonis | HYS | 12.5/0.2 | 9.3/0.1 | |
| Mastacembelus aculeatus | MAA | 21.9/0.6 | 6.3/0.2 | |
| Macropodus chinensis | MAC* | 16.7/<0.1 | ||
| Odontobutis obscurus | ODO | 12.5/0.9 | 37.5/2.7 | 91.7/7.9 |
| Siniperca chuatsi | SIC | 3.1/<0.1 | 25.0/1.2 | 75.0/2.4 |
| Siniperca obscura | SIO | 6.3/0.2 | 50.0/0.6 | |
| Siluriformes | ||||
| Glyptothorax fukiensis | GLF* | 6.3/0.9 | ||
| Liobagrus styani | LIS | 31.3/1.0 | 31.3/0.6 | |
| Mystus macropterus | MYM* | 8.3/0.1 | ||
| Pelteobagrus fulvidraco | PEF | 25.0/3.9 | 100.0/21.1 | |
| Pseudobagrus albomarginatus | PSA | 3.1/<0.1 | 37.5/0.9 | 66.7/4.1 |
| Silurus asotus | SIA | 3.1/0.1 | 25.0/0.5 | 41.7/0.7 |
| Synbranchiformes | ||||
| Monopterus albus | MOA | 31.3/0.5 | 18.8/0.4 | 33.3/0.9 |
| Beloniformes | ||||
| Hyporhamphus intermedius | HEI* | 8.3/0.9 | ||
*: Rare species occurring in less than two sites, not included in statistical analysis.
Local species richness
Fish species richness significantly varied according to stream order and gradient, but not season (Table 2).
Table 2.
Three-way ANOVA results of the spatial and temporal variations in fish species richness within Qingyi Stream
| Factors | df | SS | MS | F | P | Student-Newman-Keuls |
|---|---|---|---|---|---|---|
| Stream orders | 4 | 2.3 | 0.6 | 38.9 | ** | 1st<2nd=3rd<4th<5th |
| Seasons | 3 | 0.1 | 0.02 | 1.9 | ns | |
| Gradients | 1 | 0.1 | 0.1 | 4.0 | * | A<B |
| Orders×seasons | 12 | 0.2 | 0.01 | 0.9 | ns | |
| Seasons×gradients | 3 | 0.02 | 0.01 | 0.5 | ns | |
| Orders×gradients | 4 | 0.6 | 0.1 | 9.5 | ** | |
| Orders×seasons×gradients | 12 | 0.1 | 0.01 | 0.8 | ns |
Factors are stream orders (first-fifth), seasons (January, April, July, October) and gradient (A, B; with and without the effect of a large dam, respectively).
*: P<0.05, **: P<0.01, ns P>0.05. Student-Newman-Keuls test: P<0.05.
Species richness significantly increased with stream orders, with the exception of no significant difference between second- and third-order streams. Gradient B had more species than that of Gradient A. A significant interaction effect was observed for stream order by gradient (Table 2). Within Gradient A, the lowest and highest species richness occurred in first-order and fifth-order streams, respectively, and species richness did not vary significantly among second-, third- and fourth-order streams. Within Gradient B, however, the highest species richness was observed in fourth-order streams, while first-, second- and third-order streams were not significantly different in species richness (Figure 2). When comparing species richness within the same-order streams between the two gradients, significant differences were only observed in fourth-order streams (t-test, F=0.23, P<0.01), but not in first- (F=0.4, P=0.54), second- (F=1.5, P=0.21), or third-order streams (F=7.2, P=0.19) streams.
Figure 2.
Fish species richness along two longitudinal gradients by stream orders within Qingyi Stream (upper: gradient A; lower: gradient B)
Different lowercases represent significant differences in species richness (One-way ANOVA, P<0.05) .
Fish assemblages
The two-way crossed ANOSIM results suggested that fish assemblages significantly differed across stream orders but not seasons, and were uniform for both longitudinal gradients of A and B (Table 3). Within gradient B, significant differences in inter-order fish assemblages were almost exclusively observed, with the exception of fourth- vs. fifth-order. Whereas, fish assemblages in fourth-order streams were similar with those in second- and third-order but not fifth-order streams within gradient A (Table 3). For fish assemblages within the same-order streams, significant between-gradient differences were observed in fourth-order streams, but not in first-, second, and third-order streams (Table 4).
Table 3.
R-values and their significance levels (P) for comparisons of fish assemblages among stream orders and seasons using ANOSIM
| Gradient A | Gradient B | |||
|---|---|---|---|---|
| R | P | R | P | |
| Order | 0.50 | 0.001 | 0.66 | 0.001 |
| 1st vs. 2nd* | 0.06 | 0.440 | 0.60 | 0.029 |
| 1st vs. 3rd | 0.38 | 0.012 | 0.42 | 0.036 |
| 1st vs. 4th | 0.29 | 0.049 | 1.00 | 0.029 |
| 1st vs. 5th | 1.00 | 0.001 | 0.99 | 0.001 |
| 2nd vs. 3rd* | 0.06 | 0.420 | 0.80 | 0.029 |
| 2nd vs. 4th* | 0.06 | 0.370 | 1.00 | 0.029 |
| 2nd vs. 5th | 0.98 | 0.001 | 0.96 | 0.002 |
| 3rd vs. 4th* | 0.19 | 0.296 | 1.00 | 0.029 |
| 3rd vs. 5th | 0.90 | 0.001 | 0.93 | 0.002 |
| 4th vs. 5th* | 0.85 | 0.002 | 0.06 | 0.310 |
| Season | -0.25 | 0.988 | -0.10 | 0.971 |
One-way ANOSIM was conducted for gradient (B) without the effect of Chencun Hydropower Station, two-way crossed ANOSIM was conducted for gradient (A) disrupted by the station and both gradients (Both). Significance level was accepted at P<0.05 (in bold). *: opposite significance level occurring between the two gradients.
Table 4.
R-values and their significance levels for comparisons of fish assemblages among seasons and between two gradients disrupted by a large dam or not using ANOSIM
| R | P | |||
|---|---|---|---|---|
| Season | Gradient | Season | Gradient | |
| 1st-order | -0.33 | -0.25 | 0.876 | 0.815 |
| 2nd-order | -0.04 | 0.00 | 0.562 | 0.617 |
| 3rd-order | -0.31 | -0.73 | 0.800 | 0.988 |
| 4th-order | -0.38 | 0.75 | 0.790 | 0.020 |
| 5th-order | -0.19 | 0.872 | ||
Two-way crossed ANOSIM was conducted for first- to fourth-order streams, one-way ANOSIM was conducted for fifth-order streams. Significance level was accepted at P<0.05 (in bold).
Based on NMS analysis, fish assemblages within gradient B varied substantially from headwaters to mainstream, despite the overlap between fourth- and fifth-order streams (Figure 3). In contrast, within gradient A, substantial overlaps were observed between neighboring orders from first- to fourth-order streams, while fish assemblages in fourth-order streams were distinct from those in fifth-order streams.
Figure 3.
Non-metric multi-dimensional scaling (NMS) by stream orders and dams for fish assemblage data in Qingyi Stream
Black symbols, gradient disrupted by Chencun Hydropower Station; Open symbols, gradient without damming; Grey symbols, mutual part of both gradients. Normal triangle, reverse triangle, square, diamond, and rotundity, fish assemblages from first- to fifth-order streams, respectively.
SIMPER analysis revealed that Z. platypus was more abundant in first-, second- and third-order streams and less in fourth- and fifth-order streams within gradient B. The occurrence and abundance of C. rarus, P. fasciata, R. ocellatus, and S. nigripinnis determined how fish assemblages varied substantially from first- to fourth-order streams. Ctenogobius sp. and C. argus contributed most to the divergence in fish assemblages between fourth- and fifth-order streams within Gradient A (Table 5).
Table 5.
Diagnostic species by SIMPER analysis for first- to fifth-order streams along gradients A and B within Qingyi Streams
| Gradient A | Gradient B | |||||
|---|---|---|---|---|---|---|
| AvAbu. | AvSim. | Contri (%) | AvAbu. | AvSim. | Contri (%) | |
| 1st-order | AvSim., 46.3% | AvSim., 44.1% | ||||
| Zacco platypus | 60.9 | 18.5 | 40.0 | 167.3 | 19.4 | 30.2 |
| Acrossocheilus fasciatus | 39.6 | 10.2 | 22.1 | 17.3 | 11.5 | 17.9 |
| Vanmanenia stenosoma | 16.3 | 5.1 | 11.0 | |||
| Cobitis rarus | 23.5 | 12.4 | 19.3 | |||
| 2nd-order | AvSim., 51.6% | AvSim., 52.9% | ||||
| Zacco platypus | 63.1 | 16.8 | 32.5 | 37.0 | 17.7 | 33.5 |
| Acrossocheilus fasciatus | 15.3 | 9.7 | 18.9 | 15.0 | 13.1 | 24.7 |
| Ctenogobius sp. | 14.4 | 7.4 | 14.4 | |||
| Parabotia fasciata | 16.0 | 7.6 | 14.4 | |||
| 3rd-order | AvSim., 48.6% | AvSim., 57.8% | ||||
| Zacco platypus | 44.8 | 11.9 | 24.6 | 95.8 | 15.9 | 27.6 |
| Rhodeus ocellatus | 22.1 | 9.9 | 20.5 | 69.7 | 11.4 | 19.7 |
| Ctenogobius sp. | 28.5 | 6.6 | 13.6 | 28.8 | 8.4 | 14.6 |
| 4th-order | AvSim., 51.0% | AvSim., 64.3% | ||||
| Ctenogobius sp. | 24.6 | 11.9 | 23.4 | |||
| Zacco platypus | 46.1 | 8.5 | 16.8 | |||
| Acrossocheilus fasciatus | 12.6 | 6.1 | 11.9 | 51.8 | 5.2 | 8.1 |
| Sarcocheilichthys nigripinnis | 36.5 | 5.7 | 8.8 | |||
| Abbottina rivularis | 20.3 | 4.8 | 7.5 | |||
| 5th-order | AvSim., 52.9% | AvSim., 52.9% | ||||
| Pseudobagrus albomarginatus | 58.0 | 7.0 | 15.1 | 58.0 | 7.0 | 15.1 |
| Channa argus | 16.8 | 4.5 | 9.2 | 16.8 | 4.5 | 9.2 |
| Carassius auratus | 15.8 | 3.4 | 7.3 | 15.8 | 3.4 | 7.3 |
The first three species contributing most to the average similarity within each order are shown. AvSim., average similarity (%); AvAbu., average abundance; Contri., contribution.
Relationships between species and environment
The first and second axes of the CCA ordination accounted for 49.7% and 12.4% of total variance in species richness and abundance among sampling sites, and for 54.7% and 13.6% of variance in species data and environmental relationship, respectively. From CCA, gradient A separated fish assemblages based on elevation, substrate, and flow velocity on the right, and wet width, water depth, discharge, and conductivity on the left. The species related to this gradient were E. ilishaeformis, P. pekinensis, S. nigripinnis, H. labeo, C. rarus, C. argus, S. chuatsi, P. parva, C. auratus, A. chankaensis, R. ocellatus, C. sinensis, Ctenogobius sp., A. fasciatus, and Z. platypus (left to right). Gradient B was caused by water temperature and dissolved oxygen on the top and pH on the bottom, along which P. oxycephalus, P. fasciata, A. chankaensis, P. vaillanti, M. albus, G. taeniellus, and S. argentatus were distributed from top to bottom (Figure 4).
Figure 4.
Canonical correspondence analysis (CCA) diagrams for environmental variables and fish assemblages in Qingyi Stream
Black arrows indicate the most important factors and grey arrows show the other factors. Species codes as in Table 1.
DISCUSSION
Our results showed that along a longitudinal gradient without the direct effect of a large dam, fish species richness was highest in fourth-order streams and showed no significant difference among first- to third-order streams. Following the increase in stream size and habitat diversity from headwaters to downstream, fish species richness generally increases downstream (Gorman & Karr, 1978; Matthews, 1986). Based on fish specimens collected from 89 shallow (first- to third-order) tributaries within the Qingyi watershed, Yan et al. (2011) discovered that local species richness was determined by local habitat variables (e.g., wetted width and water temperature) but not tributary spatial variables (e.g., stream order and link magnitude). Yan et al (2010) also revealed that stream order was not the optimal framework explaining spatial variation in fish assemblages within Puxi Stream, a tributary above Chencun Reservoir in the Qingyi watershed. The results in our study that fish species richness showed no substantial variation across first- to third-order streams appears to support fish species responding to local habitat features but not to stream classification schemes of humans (Matthews, 1998). In addition, our findings that the highest species richness occurred in fourth-order, not fifth-order, streams is consistent with the river continuum concept (RCC) claiming that maximum species diversity often occurs in mid-sized, not large, streams. Oberdorff et al (1993) also found a decline in fish species richness in the lower reaches of the Seine River, which was explained by anthropogenic disturbances decreasing habitat diversity.
In addition to species richness, fish also exhibit an alteration in zonation along the upstream-downstream gradient (Lasne et al, 2007; Miranda & Raborn, 2000). One of the most recognized theoretical classifications dividing running waters based on fish is longitudinal fish zonation proposed by Huet (1959); the brown trout, grayling, barbel and bream zones exhibit alteration, in turn, from headwaters to lower reaches. However, associated with continuous, not abrupt, variation in the natural environments of many streams, fish assemblages vary by “addition” rather than by “replacement” (Matthews, 1998), which suggests a deficiency in the application of this zonation concept within these streams. In our results, despite some species only being collected in one ordered stream (e.g., P. oxycephalus and C. molitorella in first-order streams, C. carpio, M. piceus, and H. leucisculus in fifth-order streams), most fish species were distributed in at least two different ordered streams. Based on our SIMPER analysis, along the longitudinal gradient without the effects of a large dam, changes in the relative abundances of Z. platypus, C. rarus, P. fasciata, R. ocellatus, and S. nigripinnis contributed most to longitudinal species replacement. Our results suggest that “addition”, not “replacement”, was the main underlying mechanism explaining species distributions from upstream to downstream in Qingyi Stream. This spatial variation in species distributions could be explained by the relationship between species and the local habitat environment based on CCA. The spatial distributions of fish species in this study were affected by a series of local habitat environmental factors, such as elevation, substrate, water depth, current discharge, wetted width, and conductivity. At the watershed scale, Yan et al. (2011) reported that fish assemblage structure in this study area was related to the combined effects of local habitat (i.e., elevation, substrate, and water depth) and spatial stream position (i.e., magnitude link and confluence link), suggesting that fish species distribution was associated with their ecological requirements, such as habitat preference and trophic ecology.
In Qingyi Stream, fish assemblages presented different patterns in their spatial variations between the two longitudinal gradients, which could be viewed as the effect of Chencun Hydropower Station. Fish showed significantly lower species richness and different species composition within the gradient disrupted by the dam compared with the gradient without the large dam. However, the substantial variations in fish assemblages between gradients were only observed in fourth-order streams. Numerous studies have revealed that dams may impact stream fish by blocking migratory pathways (March et al, 2003), decreasing current velocity upstream (Bennett et al, 2002), altering natural flow regime (Bonner & Wilde, 2000), and decreasing water temperature downstream (Clarkson & Childs, 2000). In this study, when considering first- to fourth-order streams within the two gradients independently, a total of 18 fish species (e.g., D. tumirostris, E. bambusa, S. hollandi) were collected in gradient A only, and eight species (e.g., O. bidens, H. bleekeri, S. parvus) were more frequent and abundant in gradient A than in gradient B. This may be explained by two causes resulting from the effects of large dams on fish assemblages. Firstly, most of the 18 species only occurring in gradient A, including H. bleekeri, E. ilishaeformis, H. maculatus, and S. roulei, inhabited the lower reaches of the streams. The Chencun Hydropower Station may play a role in constraining their upstream movements, resulting in the local extinction of these fish in areas above the station. Secondly, dams may modify local habitat conditions upstream and downstream (Bennett et al, 2002; Bonner & Wilde, 2000; Clarkson & Childs, 2000), which may decrease the fitness of endemic fish to naturally adapt to lotic conditions, evemtually resulting in a decline in their abundance (Scott & Helfman, 2001). This may explain why some fish (e.g., O. bidens, H. bleekeri, S. parvus) showed lower abundance within the gradient disrupted by the dam. However, the influence of impoundment by dams on stream fish assemblage structure upstream was highly localized, which means that this influence was partially determined by proximity to impoundments (Falke & Gido, 2006a, b). Our results that significant between-gradient differences in fish assemblages only occurred in fourth-order, not lower-order, streams may support the findings of Falke & Gido (2006a, b), though some additional analyses may be necessary.
Funding Statement
This study was financially supported by the National Basic Research Program of China (2009CB119200) and the Natural Science Foundation of China (31071900, 31172120)
Footnotes
The authors have declared that no competing interests exist.
References
- [1].Araújo FG, Pinto BCT, Teixeira TP. 2009. Longitudinal patterns of fish assemblages in a large tropical river in southeastern Brazil: evaluating environmental influences and some concepts in river ecology. Hydrobiologia, 618(1): 89-107. [Google Scholar]
- [2].Bain MB, Finn JT, Booke HE. 1985. Quantifying stream substrate for habitat analysis studies. North American Journal of Fisheries Management, 5(3B): 499-500. [Google Scholar]
- [3].Bennett SJ, Cooper CM, Ritchie JC, Dunbar JA, Allen PM, Caldwell LW, Mcgee TM. 2002. Assessing sedimentation issues within aging flood control reservoirs in Oklahoma. Journal of the American Water Resources Association, 38(5): 1307-22. [Google Scholar]
- [4].Bonner TH, Wilde GR. 2000. Changes in the Canadian River fish assemblage associated with reservoir construction. Journal of Freshwater Ecology, 15(2): 189-98. [Google Scholar]
- [5].Boys CA, Thoms MC. 2006. A large-scale, hierarchical approach for assessing habitat associations of fish assemblages in large dryland rivers. Hydrobiologia, 572(1): 11-31. [Google Scholar]
- [6].Chen YY. 2005. Giving promotions to integrated management of river basins and protecting the life river of the Yangtze. China Water Resources, (8): 10-2. (in Chinese) [Google Scholar]
- [7].Clarke KR, Warwick RM. 2001. Change in Marine Communities: an Approach to Statistical Analysis and Interpretation. 2nd ed Plymouth: PRIMER-E. [Google Scholar]
- [8].Clarkson RW, Childs MR. 2000. Temperature effects of hypolimnial-release dams on early life stages of Colorado River Basin big-river fishes. Copeia, 2000(2): 402-12. [Google Scholar]
- [9].Cumming GS. 2004. The impact of low-head dams on fish species richness in Wisconsin, USA. Ecological Applications, 14(5): 1495-506. [Google Scholar]
- [10].Da Silva Abes S, Agostinho AA. 2001. Spatial patterns in fish distributions and structure of the ichthyocenosis in the Águs Nanci stream, upper Paraná River basin, Brazil. Hydrobiologia, 445(1-3): 217-27. [Google Scholar]
- [11].Falke JA, Gido KB. 2006. a. Effects of reservoir connectivity on stream fish assemblages in the Great Plains. Canadian Journal of Fisheries and Aquatic Sciences, 63(3): 480-93. [Google Scholar]
- [12].Falke JA, Gido KB. 2006. b. Spatial effects of reservoirs on fish assemblages in Great Plains streams in Kansas, USA. River Research and Applications, 22(1): 55-68. [Google Scholar]
- [13].Gauch HG. 1982. Multivariate Analysis in Community Ecology. New York: Cambridge University Press. [Google Scholar]
- [14].Gorman OT, Karr JR. 1978. Habitat structure and stream fish communities. Ecology, 59(3): 507-15. [Google Scholar]
- [15].Hoeinghaus DJ, Winemiller KO, Birnbaum JS. 2007. Local and regional determinants of stream fish assemblage structure: inferences based on taxonomic vs. functional groups. Journal of Biogeography, 34(2): 324-38. [Google Scholar]
- [16].Holmquist JG, Schmidt-Gengenbach JM, Yoshioka BB. 1998. High dams and marine-freshwater linkages: effects on native and introduced fauna in the Caribbean. Conservation Biology, 12(3): 621-30. [Google Scholar]
- [17].Huet M. 1959. Profiles and biology of western European streams as related to fish management. Transactions of the American Fisheries Society, 88(3): 155-63. [Google Scholar]
- [18].Humborg C, Ittekkot V, Cociasu A, Bodungen BV. 1997. Effect of Danube River dam on Black Sea biogeochemistry and ecosystem structure. Nature, 386(6623): 385-8. [Google Scholar]
- [19].Inoue M, Nunokawa M. 2002. Effects of longitudinal variations in stream habitat structure on fish abundance: an analysis based on subunit-scale habitat classification. Freshwater Biology, 47(9): 1594-607. [Google Scholar]
- [20].Lasne E, Bergerot B, Lek S, Laffaille P. 2007. Fish zonation and indicator species for the evaluation of the ecological status of rivers: example of the Loire basin (France). River Research and Applications, 23(8): 877-90. [Google Scholar]
- [21].March JG, Benstead JP, Pringle CM, Scatena FN. 2003. Damming tropical island streams: problems, solutions, and alternatives. Bioscience, 53(11): 1069-78. [Google Scholar]
- [22].Matthews WJ. 1986. Fish faunal “breaks” and stream order in the eastern and central United States. Environmental Biology of Fishes, 17(2): 81-92. [Google Scholar]
- [23].Matthews WJ. 1998. Patterns in Freshwater Fish Ecology. New York: Chapman & Hall. [Google Scholar]
- [24].Meador MR, Matthews WJ. 1992. Spatial and temporal patterns in fish assemblage structure of an intermittent Texas stream. American Midland Naturalist, 127(1): 106-14. [Google Scholar]
- [25].Miranda LE, Raborn SW. 2000. From zonation to connectivity: fluvial ecology paradigms of the 20th century. Polskie Archiwum Hydrobiologii, 47(1): 5-19. [Google Scholar]
- [26].Nilsson C, Jansson R, Zinko U. 1997. Long-term responses of river-margin vegetation to water-level regulation. Science, 276(5313): 798-800. [DOI] [PubMed] [Google Scholar]
- [27].Oberdorff T, Guilbert E, Lucchetta JC. 1993. Patterns of fish species richness in the Seine River basin, France. Hydrobiologia, 259(3): 157-67. [Google Scholar]
- [28].Power ME, Dietrich WE, Finlay JC. 1996. Dams and downstream aquatic biodiversity: potential food web consequences of hydrologic and geomorphic change. Environmental Management, 20(6): 887-95. [DOI] [PubMed] [Google Scholar]
- [29].Power ME, Stout RJ, Cushing CE, Harper PP, Hauer FR, Matthews WJ, Moyle PB, Statzner B, De Badgen IRW. 1988. Biotic and abiotic controls in river and stream communities. Journal of the North American Benthological Society, 7(4): 456-79. [Google Scholar]
- [30].Roberts JH, Hitt NP. 2010. Longitudinal structure in temperate stream fish communities: evaluating conceptual models with temporal data. American Fisheries Society Symposium, 73: 281-302. [Google Scholar]
- [31].Santucci VJ Jr, Gephard SR, Pescitelli SM. 2005. Effects of multiple low-head dams on fish, macroinvertebrates, habitat, and water quality in the Fox River, Illinois. North American Journal of Fisheries Management, 25(3): 975-92. [Google Scholar]
- [32].Scott MC, Helfman GS. 2001. Native invasions, homogenization, and the mismeasure of integrity of fish assemblages. Fisheries, 26(11): 6-15. [Google Scholar]
- [33].Strahler AN. 1957. Quantitative analysis of watershed geomorphology. Transactions, American Geophysical Union, 38(6): 913-20. [Google Scholar]
- [34].ter Braak CJF, Verdonschot PFM. 1995. Canonical correspondence analysis and related multivariate methods in aquatic ecology. Aquatic Sciences, 57(3): 255-89. [Google Scholar]
- [35].Taniguchi Y, Rahel FJ, Novinger DC, Gerow KG. 1998. Temperature mediation of competitive interactions among three fish species that replace each other along longitudinal stream gradients. Canadian Journal of Fisheries and Aquatic Sciences, 55(8): 1894-7901. [Google Scholar]
- [36].Travnichek VH, Zale AV, Fisher WL. 1993. Entrainment of ichthyoplankton by a warm water hydroelectric facility. Transactions of the American Fisheries Society, 122(5): 709-16. [Google Scholar]
- [37].Vannote RL, Minshall GW, Cummins KW, Sedell JR, Cushing CE. 1980. The river continuum concept. Canadian Journal of Fisheries and Aquatic Sciences, 37(1): 130-7. [Google Scholar]
- [38].Ward JV, Stanford JA. 1983. Serial discontinuity concept of lotic ecosystems. In: Fontaine TD, Bartell SM. Dynamics of Lotic Ecosystems. MI, Ann Arbor: Ann Arbor Science, 29-42. [Google Scholar]
- [39].Yan YZ, Xiang XY, Chu L, Zhan YJ, Fu CZ. 2011. Influences of local habitat and stream spatial position on fish assemblages in a dammed watershed, the Qingyi Stream, China. Ecology of Freshwater Fish, 20(2): 199-208. [Google Scholar]
- [40].Yan YZ, He S, Chu L, Xiang XY, Jia YJ, Tao J, Chen YF. 2010. Spatial and temporal variation of fish assemblages in a subtropical small stream of the Huangshan Mountain. Current Zoology, 56(6): 670-7. [Google Scholar]