INTRODUCTION
There are many types of man-made reservoirs or hydraulic structures of various sizes from small-sized reservoirs to large-sized multipurpose dams. These water bodies are used for drinking and various purposes (flood control, electricity generation, water supply), and they are usually constructed in areas with low rainfall or areas that may otherwise lack water resources (Benke, 1990; Dynesius and Nilsson, 1994). Among man-made reservoirs, those where organisms are well preserved and systematically managed not only do they improve biodiversity, but also provide aesthetic and economic value. They also serve an ecological function as habitats to protect diverse living things (Barbier, 1994). However, they have been even more valued when managing man-made reservoirs suit people's needs, from a direct and economic viewpoint, such as flood control, securing a water supply for drinking and farming and recreationt. Such reservoirs provide various benefits to people, but also act as the biggest artificial factors that disrupt freshwater ecosystems (Ward and Stanford, 1979; Ricciardi and Rasmussen, 1999; Baron et al., 2002). In addition, since the beginning of the millennium, water pollution, increases in land use and climate change have become the greatest threats to aquatic ecosystems. Among them, reservoirs and marshes, lentic ecosystems, are seriously experiencing eutrophication due to farmland areas in the neighborhood and human activities (Finlayson and Moser, 1992; Dugan, 1994).
In the past in Korea, research was conducted mostly on the physicochemical water-quality of reservoirs. However, the importance of inhabitants is being emphasized due to the recently growing interest in living organisms, and the development and application of an index which assesses the soundness of reservoir ecosystems by using inhabitants (Center For Aquatic Ecosystem Restoration, 2010). Among them, freshwater fishes, which are a high-trophic level species in aquatic ecosystems, are used as an important indices to assess the soundness of lentic ecosystems. This is because they are directly or indirectly substantially influenced by internal and external changes in reservoir ecosystems and various environmental variables (Matthews and Robison, 1988; Matthews et al., 1992; Moyle and Cech, 2000; Kouamélan et al., 2003; Buisson et al., 2008). There are multiple environmental factors that could affect the fish community structures within the reservoir ecosystem: (1) the effects of parasites, viruses and dissolved oxygen; (2) the introduction of foreign species; (3) biological disturbance factors include fishing activities by humans (Byeon et al., 1997; Shin et al., 2008: Lee et al., 2012); (4) physical disturbance factors including the influx of turbid waters due to mudslides within a basin and habitat destruction due to construction at riverbeds and riversides (Steffe et a1., 2007); and (5) physicochemical disturbance factors including the influx of nutrients due to human activities within a basin (Zalidis et al., 2002). Among them, there is the physicochemical disturbance due to water pollution is known to have a linear functional relationship with fishes' productivity or biomass not only in man-made reservoirs in Korea, but in that of North America and European countries (Ryder, 1965; Ryder et al., 1974; Matuszek, 1978; Kerr and Ryder, 1988): A lot of studies have been conducted regarding what influence the eutrophication of reservoir ecosystems, which is due to nutrients such as phosphorus (P) and nitrogen (N), has on fish communities. Such studies explain by connecting with fish communities and nutrient loading or the productivity of phytoplankton, the first producer (Melack, 1976; McConnell et al., 1977). In recent studies, they stated that such results are the consequences of human activities, and inquired into the relationship between human activities and fish communities.
Unlike Gucheon Reservoir, Yeoncho Reservoir, which is located at the upper part of Geoje Island, has a shallow water depth. Also, aquatic vegetation, where communities of aquatic plants such as emerged plants; floating plants; floating-leaved plants; and submerged plants are formed, is well developed at the shore embankment. Such aquatic vegetation areas are used as the spawning grounds for many types of fishes. And because they could be also used as shelters for fry to hide from carnivorous fish species (Gulati et al., 1990; Han et al., 2009), unlike the Gucheon Reservoir, Yeoncho Reservoir environment is an attractive habitat for various fishes (Park et al., 2009). The Reservoir is designated and managed as a water supply source protection area, and fishing and water activities are strictly restricted. However, eutrophication has been observed due to high population density, and the inflow of nutrients from point and non-point pollutant sources such as barns and farmland. Meanwhile, Gucheon Reservoir, which is located at the lower part of Geoje Island, is used for drinking water, and no eutrophication has been reported due to the low population density and human activities. Both man-made Reservoirs that are used for drinking water, appear to be similar in climate conditions and morphological characteristics, but may show slight differences due to human activities within the basins. In particular, the first research on the fish fauna of these two reservoirs was first reported in literature when an investigation was carried out along the entire river areas of Geoje Island in 1998 (Son and Song, 1998). Since then, no research has been conducted on freshwater fishes. Identifying the geographical distribution and habitats of fishes provides fundamental data for a relevant study. It is also substantially important for the efficient use and protection of water resources. Therefore, in this study, by comparing the physicochemical water-quality characteristics of two major man-made reservoirs in Geoje area, it aims to identify how fish communities are affected by water quality. It also intends to investigate habitat conditions of fish fauna and introduced species of each reservoir, and provides useful fundamental data to enable the stable fishery resource management of reservoirs, and comparison with other regional reservoirs water systems.
MATERIALS AND METHODS
1.Description of Sampling Sites
The Yeoncho Reservoir is a dam for water supply that was built in 1979 to supply water for living and industrial purposes. Along with the Gucheon Reservoir, it has been an important water supply source in Geoje. Yeoncho Reservoir, which is relatively small in size, has the following specifications: the basin area of 11.7 km2; the reservoir area of 60 ha; and a maximum water storage of 4,590 × 103 tons. To take a look at the characteristics of Yeoncho Reservoir's basin area, there are five tributary streams: Myeongdong Stream, Myeongha Stream, Juryeong Stream, Inam Stream and Imok Stream. Among them, Myeongdong Stream and Juryeong Stream make up the largest share in the basin area. They also have the highest population density and largest ranch area (Korea Water Resources Corporation, 2004). With regards to wastewater discharge facilities, there were three village-unit sewage treatment facilities, but there are no other excretion treatment facilities nor wastewater discharging firms (Seo et al., 2004). Gucheon Reservoir is situated in the mountainous region of Gucheon-ri, Gangbu-myeon, Geoje Gyeongsangnam-do, and was built in 1987 for the purpose of providing drinking water. Its specification is as follows: the basin area of 12.7 km2; the reservoir area of 50 ha; and a maximum water storage of 10,000 × 103 tons. In terms of its basin's characteristics, Guryong Stream is its only tributary stream, which flows in from the upper stretches of the reservoir. At present, the Gucheon Reservoir is not designated as a water supply source protection area. However, due to a relatively small population (around 250 in total population), some pollution from domestic sewage is expected as there are no sewage treatment facilities, but other than this, there were no critical pollutants observed (Korea Water Resources Corporation, 2004).
2.The Study Period and site Descriptions
In this study, the Gucheon Reservoir and Yeoncho Reservoir were selected, which have the most similar environmental conditions regarding the basin area; reservoir area; purpose of use, among man-made reservoirs located in Geoje Island. Also, four areas of each reservoir were chosen, and fisheries research was conducted twice a year in 2012: before (May ~ June) and after (September - October) the rainy season, the periods when water bodies are stable. With regards to the research areas of each reservoir, longitudinal gradient characteristics were taken into account, and three littoral zones, the lacustrine zone (S1); transitional zone (S2); riverine zone (S3), were selected (US EPA, 1998). Also, this study added a pelagic zone (S4), which could represent each reservoir. Three littoral zones with high accessibility from each selected section were chosen and fisheries research were carried out (Figure 1).
3.Sampling Gears and Methods
Six types of nets were used for fish collection: fyke net (FN), gill net (GN), trammel net (TN), casting net (CN), kick net (KN) and minnow trap (MT). A fyke net (mesh: 5 mm; one bag net), gill net (mesh: 4.5 cm), trammel net (mesh: 1.2 cm) and minnow trap net (mesh: 4 mm) was used at the pelagic zone (S4). And at the littoral zones, a casting net (mesh: 7 × 7 mm) and a kick net (mesh: 4 × 4 mm) were used for fisheries collection. Four types of nets, a fyke net, gill net, trammel net and minnow trapt, were installed at the sites and collected after 24 hours. As for the casting net, an investigation was conducted at each littoral zone for an hour, and it was thrown considering the characteristics of the riverbed structure (sand, gravel, mud); littoral zones (emerged plant zone, non-emrged plant zone); and water depth (shallow, deep) to find the representative value of homogeneous fish community. The kick net was used together with the casting net in places where it was difficult to perform research using the casting net due to the shallow water depth, vegetation zone and the shore line. With regard to researchers, more than one expert who has under gone professional training or has more than three years of field experience was included. Also, in case of research at the littoral zones and the pelagic zone, a group of three was deployed to each zone. Fishes that were caught at certain sites were set free after identification and counting in accordance with Kim and Park (2002) and Son and Song (2006). However, species that were difficult to identify were fixed with an aqueous solution of 10 % formalin, and observed it at a research laboratory. Among the collected fishes, fry below 20 mm in length, which were impossible to identify, were excluded from the calculation of the total population. When abnormal species were caught, abnormality types, deformity (DE), erosion (EF), lesions (LE), and tumors (TU), were determined according to the method of Sanders et al. (1999), and recorded in a fisheries field sheet.
4.The Analysis of Fish Community and Ecological Indicator Characteristic
Classifications of fish trophic guilds and tolerance guilds followed the approach of the US EPA (1993) and Barbour et al. (1999). Analysis of region-specific endemic species was based on previous classifications (An et al., 2001). In this study, tolerance guilds were classified as sensitive species (SS), intermediate species (IS), and tolerant species (TS), whereas trophic guilds were divided into four categories: insectivores (I), omnivores (O), carnivores (C), and herbivores (H). This approach is based on the principle that increased numbers of species and individuals in the first two categories indicate better ecosystem health, whereas an increase in omnivores indicates a degradation of ecosystem health (US EPA, 1993). In case of criteria for the tolerance characteristics and trophic level by fish species, this study conducted an analysis based on classification criteria of MOE/NIER (2008). The length-weight relationship (LWR) and the condition factor (K) of species, which were most often caught among collected fishes, were calculated. This was conducted in order to surmise the health conditions of fishes living in the two reservoirs. The analysis of fisheries growth and the condition factor (K) allows us to identify health conditions and fecundity. Moreover, analysis information is used as an index as it provides various data including habitat quality, water quality and prey availability (Anderson and Gutreuter, 1983; Busacker et al., 1990; Ney, 1993). The length-weight relationship was calculated as follows (Seo, 2005):
TW = aTLb
TW: total weight, TL: total length(a, b: parameters)
A condition factor (K), along with the length-weight relationship, is one of the methods to measure an individual fish's health. It is also an index that reflects the feeding conditions of a fishes' food sources, and can explain energy accumulation in fish population according to changes in the condition factor (K) (Seo, 2005). Fulton’s condition factor (K) and relative condition factor (Kn) can be more easily explained than the parameters a and b of the length-weight relationship, and the equation is as follows:
K = TW / TL3 × 10n(Anderson and Neumann, 1996)
Kn = TW / aTLb(LeCren, 1951)
5.The Analysis of Physicochemical Water-Quality
To analyze the physicochemical water-quality of the Gucheon Reservoir and the Yeoncho Reservoir at Geoje Island, this study used data analyzed for five years from 2008 to 2013 of which were from the Water Environmental Information System (WEIS; http://water.nier.go.kr) of the Ministry of Environment. To analyze seven water-quality variables, biological oxygen demand (BOD); chemical oxygen demand (COD); total nitrogen (TN); total phosphorus (TP); electric conductivity (EC, 25°C); chlorophyll-a, (Chl-a); and total suspended solids (TSS), and nutrients that affect the primary productivity of reservoirs, this study used TN/TP ratios to analyze long-term changes trophic state in reservoirs. In order to analyze precipitation characteristics at the research sites, this study used the Korea Meteorological Administration's precipitation data on the Geoje area. Also, the Carlson (1977)'s Trophic State Index (TSI) was used to quantify the trophic states of the two reservoirs: among them, this study used TP and Chl-a to assess the eutrophication level of those reservoirs. The assessment of eutrophication levels using the Trophic State Index (TSI) is simpler than other methods, and it is known to enable the comprehensive consideration of water-quality characteristics. This study conducted a t-test to identify whether there were differences between the two reservoirs regarding water-quality variables. SPSS (Version 20.0 IBM SPSS Statistics for windows, 2011) was used for statistics analysis.
RESULTS AND DISCUSSION
1.The Comparison of Physicochemical Water Quality Characteristics
According to the 5-year analysis of precipitation patterns and the water quality of Geoje Island, the annual precipitation was 1,777 mm, and the heaviest precipitation was recorded in 2012 (2,037 mm) (Figure 2a). The monthly average precipitation was 148 mm, and an average of more than 200 mm rainfall was observed during May and August (Figure 2b). Supposing that the climate conditions at the research sites are same as average figures stated in KMA's data, it was found that the annual range of those sites was narrow and rainfall was abundant (Ji et al., 2006). Such concentrated rainfall during the summer season has been reported to affect the trophic state characteristics of water bodies and water quality (An, 2000; An and Jones, 2000; An et al., 2001). Based on the 5-year data, average biochemical oxygen demand (BOD), an indicators of organic matter pollution, of the Gucheon Reservoir and the Yeoncho Reservoir was 0.8 mg L-1 and 1.45 mg L-1, respectively. The results were classified as a Ia-class (very good) and a Ib-class (good), respectively, according to the criteria set by the Ministry of Environment, indicating that Gucheon Reservoir has lower organic pollution levels. The BOD concentration temporarily watered down from April to August when rainfall was relatively abundant, but the BOD increased again when water bodies stabilized after heavy rainfall (Figure 2c). The average COD of the Gucheon Reservoir was 1.94 mg L-1, and 3.72 mg L-1 for the Yeoncho Reservoir, both displaying same level as BOD (Figure 2d). The average concentration of total nitrogen (TN), which is one of the nutrients that is widely used as a trophic state index, of Gucheon and Yeoncho was low, at 0.86 mg L-1 and 0.95 mg L-1, respectively. The TN concentration was observed to increase in the Yeoncho Reservoir after the rainy season in September to an average of 1.5 mg L-1 (Figure 2e). An average total phosphorus (TP) concentration of each Reservoir was 14.1 μg L-1 (Gucheon) and 34.1 μg L-1 (Yeoncho), and the concentration increased in July when rainfall was heavy. However, it was analyzed that the Gucheon Reservoir was not affected by the inflow of TP due to rainfall (Figure 2f). Monthly changes in electrical conductivity (EC) were fewer than other variables, and the average EC of Gucheon and Yeoncho was 65.3 μS cm-1 and 108.1 μS cm-1, respectively, Yeoncho showing a relatively higher figure (Figure 2g). The average total suspended solids (TSS), a water pollution index of Gucheon and Yeoncho was 2.82 mg L-1 and 6.71 mg L-1, respectively, Yeoncho recording a higher concentration. Also, it sharply increased after lots of precipitation in July in Yeoncho, but Gucheon was not much affected by the rise in rainfall (Figure 2h). In particular, it is surmised that nutrients emitted from farmland and stock farms within the basin are causing the eutrophication of the Yeoncho Reservoir during the rainy season. Despite a similarity in basin environment, Yeoncho and Gucheon had different water quality. Besides the TN and CHL, the average concentration of the Yeoncho Reservoir's BOD, COD, TP, TSS and EC was more than two times higher than (p < 0.001) that of the Gucheon Reservoir (Figure 3). Such differences were due to land use within the basins of both reservoirs: in case of the Yeoncho Reservoir, about 20% of the agricultural land, whereas only about 5% of the Gucheon Reservoir's were agricultural land. In particular, because organic fertilizers (eg. manure) are used a lot as nutrients for crops at rice paddies and farms, BOD, COD, TP, TSS and EC levels tend to rise when rainfall causes soil mixed with the nutrients to get discharged into water bodies. This analysis coincides with that of Han et al. (2010), which stated that, with regard to the Geum River's water system, reservoirs close to farmland showed higher organic pollution levels and nutrient concentrations than those close to mountains and cities. Also, the results of Han et al. (2010)'s paper showed similar results of this paper regarding the analysis of how farmland and stock farms within reservoir basins cause a large quantity of nutrients (TN, TP) and TSS to flow into reservoirs in case of rainfall, leading to the eutrophication of water bodies. Likewise, because a farmland area within a basin has a strong correlation with a man-made reservoir's water quality (Mouri et al., 2011), it is judged that the management of land within the Yeoncho Reservoir's basin is thought to be necessary for the supply of clean water in the long-term. Also, the Gucheon Reservoir where water quality is good, should be designated as a protected area as soon as possible. Meanwhile, the average N/P ratios of the Gucheon Reservoir and the Yeoncho Reservoir were 77 and 43, respectively. Both Reservoirs' primary productivity (Chlorophyll-a, CHL) was analyzed to be limited by phosphorus (P) (Figure 3), and this feature is similar to other man-made reservoirs that supply drinking water in Korea (An et al., 2001; Kim et al., 2012). According to the TSI (TP) and TSI (Chl-a) values that were used to assess the eutrophication level of the two reservoirs, TSI (TP) of the Gucheon Reservoir and the Yeoncho Reservoir was 47.2 and 63.2, respectively; and the TSI (Chl-a) was 51.6 and 53.8, respectively. On the basis of the criteria Carlson (1977) used to classify eutrophication level, the Gucheon Reservoir was classified as a mesotrophic reservoir (TSI index: 40 ~ 50), and the Yeoncho Reservoir was a eutrophic reservoir (TSI index: 50 ~ 70). This indicates that nutritive conditions differ by reservoir.
2.The Comparison of the Fish Fauna and Fish Community between the Gucheon Reservoir and the Yeoncho Reservoir
The research of fish fauna was conducted before and after the rainy season, when water bodies are stable, by using a fyke net (FN), gill net (GN), trammel net (TN) casting net (CN), kick net (KN) and minnow trap (MT). During the study, fishes of 18 species and 3,626 individuals were sampled from the Gucheon Reservoir. Among these, cyprinidae families made up the largest share with 8 species (55.2 %) (Table 1). According to the fish fauna analysis, more species were identified in the Yeoncho Reservoir: fishes of 16 species 1,362 individuals (13.6 kg) were identified at the four zones in the Gucheon Reservoir; and 12 species 2,264 individuals (15.1 kg), were identified at the four zones in the Yeoncho Reservoir. However, both Gucheon and Yeoncho estimated low biodiversity as biodiversity indices (Shannon and Weaver, 1963) were 0.57 and 0.55, respectively. From the two reservoirs, only one exotic species, Cyprinus carpio, was found in Yeoncho. Korean endemic species, three (16.7 %) were identified in both Reservoirs, showing a very low rate (Table 1). Among fishes with greater than 5 % of relative dominance in Gucheon Reservoir, Zacco temminckii (67.5 %) made up the largest share in the total individuals, followed by Pseudorasbora parva (8.3 %) and Carassius auratus (7 %). Chaenogobius urotaenia, which showed the highest relative dominance, occupied 56.5% of total individuals identified in the Yeoncho Reservoir, followed by Carassius auratus (25.6 %) and Zacco temminckii (7 %). There were differences in species that dominate the two reservoirs. This is thought to be due to their differences in nutritive and physical conditions. The Gucheon Reservoir, a mesotrophic reservoir, was dominated by Zacco temminckii, whereas Yeoncho Reservoir, a eutrophic reservoir, was dominated by Chaenogobius urotaenia. A total of 11 species that were identified in the last research (Son and Song, 1998; MOE/NIER, 2008) were not found this time: six endemic species, Coreoleuciscus splendidus, Squalidus gracilis majimae, Zacco koreanus, Silurus microdorsalis, Liobagrus mediadiposalis and Odontobutis platycephala; one diadromous species, Plecoglossus altivelis altivelis; and four other species, Rhynchocypris oxycephalus, Lefua costata, Oryzias latipes and Rhinogobius giurinus. Meanwhile, a total of seven new species were identified in the two reservoirs: a class 1 endangered species, Odontobutis obscura; two endemic species, Rhodeus uyekii and Iksookimia yongdokensis; and four other species, Zacco platypus, Hypomesus nipponensis, Tridentiger brevispinis and Cyprinus carpio. This is surmised to be due to a decrease in the number of riffle-benthic species and a relative increase in lentic species due to a disturbance that occurred at riverbeds when the dams were constructed. Such a construction changed the reservoirs' environment from lotic to lentic. In particular, Coreoleuciscus splendidus, a riffle-benthic species, is thought have died out due to a maintenance construction at the upper Gucheon. Silurus microdorsalis and Liobagrus mediadiposalis, which live in the upper stream, were also not identified due to a low possibility of their existence. Regarding other species, it is thought to be due to a difference in research zones.
3.The Analysis of Fish Tolerance Guilds and Trophic Guilds
According to the analysis of tolerance guilds of fishes collected from the two man-made reservoirs, sensitive species (SS) occupied 67.5 % of total individuals in the Gucheon Reservoir, a mesotrophic reservoir, whereas they made up a mere 7% in the Yeoncho Reservoir. This is because sensitive species are vulnerable to the physical disturbance of the water environment and pollution due to a decrease in physicochemical water-quality. On the other hand, when water environment is more exposed to pollution, the number of tolerant species (TS) tends to increase: The ratio was only 19.6 % in the Gucheon Reservoir, whereas it was 30.7 % in the Yeoncho Reservoir. Meanwhile, the portion of intermediate species (IS) in the Yeoncho Reservoir was 62.3 %, which was considerable high (Figure 4). According to a trophic guild analysis, the ratio of omnivore (O) in the Gucheon Reservoir and the Yeoncho Reservoir was 18.1% and 30.7 %, respectively; and the ratio of insectivore (I) in Gucheon and Yeoncho was 81.6% and 69.3 %, respectively. Both reservoirs' insectivore (I) ratios were analyzed to be high. On the other hand, with regard to carnivores (C), Odontobutis obscura accounted for 0.22 % in Gucheon Reservoir, insectivore species were observed to dominate the area. Only a small number of Odontobutis obscura, which were identified in the Gucheon Reservoir, are known to live in only on Geoje Island (Chae, 1999), making them a very important species from an ecological and geographical standpoint. Accordingly, it is judged that they should be provided with continuous protection. The ratio of tolerant species and omnivore species was relatively high in the Yeoncho Reservoir compared to in the Gucheon Reservoir. This result coincides with Karr (1981)'s paper and US EPA (1991)'s report. According to those papers, an increase in the eutrophication level of a reservoir will result in physicochemical-quality degradation in a habitat such as organic pollution and habitat destruction; leading to an increase in tolerant species and omnivore species.
4.The Analysis of Length-Weight Relationship of Fish Individuals
According to the analysis of the length-weight relationship of fish population (Zacco temminckii, Carassius auratus, Pseudorasbora parva, Rhinogobius brunneus), that were commonly identified in the two man-made reservoirs where research was conducted, the value of the regression coefficient b of Gucheon (mesotrophic) was in the range of 2.15 ~ 2.40, and that of Yeoncho (eutrophic) was in the range of 2.59 ~ 3.14. The result shows that eutrophic Yeoncho Reservoir's value of regression coefficient b, which indicates a growth in fish population, is higher than that of the Gucheon Reservoir. In general, the value of regression coefficient b is a slope that shows the obesity of the fish population. In case of normal fishes, the range is in between 2.7 ~ 3.2, and when the value goes beyond 3.2, then the relevant fishes are obese. Also, when the value is less than 2.7, then fishes are known to be undersized (Seo, 2005). In case of Zacco temminckii of which a relatively many number were identified, the value of regression coefficient b was 2.15 in Gucheon Reservoir, whereas it was higher in the Yeoncho Reservoir (2.59) (Figure 5a). Also in case of Carassius auratus, the regression coefficient b of Gucheon Reservoir and Yeoncho Reservoir was 2.4 and 3.14, respectively (Figure 5b). Also, the regression coefficient values of Pseudorasbora parva in the Gucheon Reservoir and the Yeoncho Reservoir were 2.39 and 2.59, respectively, Yeoncho displaying higher figure (Figure 5c); and the values of Rhinogobius brunneus were 2.35 and 3.14, respectively, showing a big difference where the figure was again higher in Yeoncho (Figure 5d). Such results indicate that as the level of eutrophication escalates, a habitat becomes more suitable for fishes to grow. This also matches the results shown in Ko et al. (2012)'s paper, where it states that an influx of nutrients to a reservoir has positive effects on the growth of a fish population.
5.The Analysis of Condition Factor (K) of Fish Population
The condition factor (K) of fish population can also be expressed with the Fulton’s condition factor (K). A relationship between the condition factor and the length of the fish population has a positive or a negative slope. A positive slope indicates that fishes are obese, and a negative slope means fishes are undersized (Seo, 2005; Park et al., 2013). After analyzing a slope value driven from the condition factor-length relationship, it was in the range of -0.0075 ~ 0.0016. In case of the mesotrophic Gucheon Reservoir, four fish species had negative slope values, and regarding the eutrophic Yeoncho Reservoir, besides Carassius auratus and Rhinogobius, brunneus Zacco temminckii and Pseudorasbora parva populations had negative slope values (Figure 6). A high condition factor (K) can be explained with the accumulation of high energy due to abundant food sources for fishes. In particular, this phenomenon can be frequently observed in those fishes that live in water bodies polluted with domestic sewage and organic pollution (Adams et al., 1992). Meanwhile, in case of water pollution due to disease or heavy metals, it can directly and indirectly affect fishes and lead to a loss in weight (Adams, 2002). In particular, diseases due to pollutants disturb fishes' feeding activities, causing them to eat less. As a consequence, fishes will have low condition factors, because the additional consumption of energy within the immune system will lead to a loss in weight (Sindermann, 1990). According to the result of the analysis of the length-weight relationship and condition factor, the health of fishes living in the Gucheon Reservoir was poor condition level, while those living in Yeoncho Reservoir showed a normal health level.
6.Discussion
The purpose of this study was to identify the effects of chemical water quality and the trophic state on fish species composition and community structure. This research was conducted at the two man-made reservoirs, Gucheon Reservoir and Yeoncho Reservoir in Geoje Island. The two reservoirs were constructed to supply drinking water to people living on the island, and both have similar environmental conditions. Lake morphometry and land-use within the basin areas showed some distinction. With regard to the Gucheon Reservoir, its water level is deep; there are barely any emerged plants nor pollutants within its basin. On the other hand, in case of the Yeoncho Reservoir, its water level is low; aquatic plants are widely distributed; and nutrients are steadily loaded into its water body due to human activities within its basin area (Cheong et al., 2004). Such differences directly influence the physicochemical water-quality of man-made reservoirs, and are known to affect fish populations and community structure (Gburek and Folmar, 1999; Zalidis et al., 2002; Han et al., 2010). The two reservoirs, which are used for drinking water, are in a mesotrophic ~ eutrophic state, and Yeoncho Reservoir's water quality was analyzed to be worse. The degradation in water quality is due to a large amount of phosphorus (P) brought into the reservoir during the rainy season. And the main culprits for such nutrients are farmland, livestock wastewater and domestic wastewater within the reservoir's basin area (Jung et al., 2006; Yoon et al., 2007; An and Yang, 2007; Kim et al., 2012).
In the ecological analysis of tolerance guilds and trophic guilds, the ratios of sensitive species (SS) and insectivore (I) were relatively low in the Yeoncho Reservoir compared to the Gucheon Reservoir, while the ratios of tolerant species (TS) and omnivores (O) were relatively high. These results coincide with that of preceding research of Karr (1981)US EPA (1991). According to those papers, with regard to the number of individuals and species of tolerant species, species richness tends to increase in proportion with a decrease in a habitat physicochemical water-quality due to organic pollution and habitat destruction. In case of the Yeoncho Reservoir, diverse fish species were expected to dwell there as there were well-developed emerged plants, but its species diversity was lower than the Gucheon Reservoir. In this way, the low species diversity of the Yeoncho Reservoir was surmised to be the result of the degradation in habitat quality due to deteriorated chemical water-quality, rather than physical habitat destruction (Han et al., 2009).
Also, by assessing the health of fish species that commonly live in both reservoirs with the length-weight relationship (LWR) and the condition factor (K), major species living in both reservoirs showed poor health. Based on the foregoing results, the trophic state of the Gucheon Reservoir and the Yeoncho Reservoir, which are used for drinking water, is surmised to be closely related to growth and the community structure of fish populations living in each ecosystem. Also, those living in a poor trophic state were observed to provide a stable environment to a community structure (Ko et al., 2012). Moreover, the trophic state of these man-made reservoirs reflects growth, condition factor, fish tolerance characteristics and feeding guild characteristics. In conclusion, the Gucheon Reservoir and the Yeoncho Reservoir, which were made for the supply of drinking water, are currently in a mesotrophic ~ eutrophic state. However, in case of the Yeoncho Reservoir, although it was designated as the water supply source protection area, wastewater and sewage is still discharged from farmland and residential areas within its basin area. Therefore, it will become too contaminated to be used as drinking water (Seo et al., 2004) in the future unless the inflow of polluted water to the reservoir is not prohibited. To prevent this in advance, it is necessary to prepare the proper countermeasures.
