ZHANG Ying-Xue ， WANG Rui ， QU Xiao ， XIA Wen-Tong ， XIN WeiGUO Chuan-Bo and Chen Yu-Shun

(1. State Key Laboratory of Freshwater Ecology and Biotechnology， Institute of Hydrobiology， Chinese Academy of Sciences，Wuhan 430072， China; 2. University of Chinese Academy of Sciences， Beijing 100049， China)

Abstract: There is limited information available on effects of aquaculture on lake ecosystems. Community structure and biodiversity of phytoplankton were investigated from July to September， 2015 in 23 lakes in the middle reach of the Yangtze River Basin， which include four groups: reservoir group (A)， aquaculture ban group (B)， low stocking aquaculture group (C) and high stocking aquaculture group (D). The analysis showed that dominant genera of the four groups were different. Group A was dominated by Pseudanabaena(Y=0.642) and Planktolyngbya (Y=0.064). Group B was dominated by Merismopedia (Y=0.428)， Planktolyngbya (Y=0.118)， Pseudanabaena (Y=0.133)， and Scenedesmus (Y=0.066). Group C was dominated by Pseudanabaena (Y=0.395)， Merismopedia (Y=0.097)， and Planktolyngbya (Y=0.122). Group D was dominated by Merismopedia (Y=0.308)， Microcystis (Y=0.118)， and Pseudanabaena (Y=0.077). The phytoplankton abundance in group A was significantly lower than those in other lake groups (P＜0.05). There was no significant difference in Shannon-Wiener index， Margalef index and Pielou index of phytoplankton among the four lake groups. The study indicated that fish culture could affect the abundance and dominant genera of phytoplankton， which may provide some implications for lake ecosystem management.

Key words: Lake; Fish aquaculture; Phytoplankton community

Aquaculture can not only provide aquatic food with high quality， but also provide employment opportunities and promote economic development[1—3].However， the unregulated and intensive aquaculture practices may have a profound impact on aquatic environment， affect structure and function of ecosystems， and eventually degrade integrity of natural ecosystems. Some aquaculture systems can cause the reduction of wild fish supplies[4]， habitant modification[5]，exotic species invasion[6，7]， eutrophication of lakes and coastal areas[8，9]， and changes of benthic macroinvertebrate communities[10]. Uneaten food and fecal waste materials generated from aquaculture systems can make the receiving water body rich in nutrients and then possible algal-blooms[11]. Also， species cultured in the system can affect phytoplankton community by directly or indirectly grazing algae and disturbing water column， which accelerate nutrient cycles[12，13].

Phytoplankton are primary producers of lake ecosystems and play important roles in nutrient cycles， energy exchanges and function transfer of lake ecosystems[14—17]. Phytoplankton are extremely sensitive to changes in water quality and environment， and often used as important biological indicators for evaluation of trophic levels in lakes[18—20].

Effects of aquaculture on phytoplankton communities have been reported in different systems and from many aspects， most of them focused only on part of the systems without considering the overall spatial scale of each system. For instance， some studies compared phytoplankton communities between cage cultured and non-cultured areas in brackish water， reservoirs and lakes[21—25]. Some studies assessed effects of species cultured， stocking density and culture modes on phytoplankton communities in enclosures of a river or lake， ponds， and cage farms of sea water[26—30]. Others tried to detect the effects of longterm aquaculture on phytoplankton communities by comparing data in different years[31，32]. However， information on effects of aquaculture on overall lake phytoplankton communities is still limited.

The middle reach of the Yangtze River Basin is one of the most productive areas for fisheries in China. According to the China Fisheries Yearbook 2016[33]， freshwater aquaculture production of three provinces in this region， including Hubei， Hunan and Jiangxi Province， accounted for 30.14% of the total inland fish yield in China. In Hubei province， about 12% of the freshwater fishery production came from lakes and reservoirs.

The aim of the present study was to detect effects of aquaculture on overall phytoplankton communities in lakes. The research question was: were phytoplankton composition， abundance， dominant taxa and diversity index different among four fish culture groups of lakes in the middle reach of the Yangtze River Basin? Research results could provide suggestions for management， utilization and protection of lakes under different aquaculture practices in the middle reach of the Yangtze River Basin.

1 Materials and Methods

1.1 Study area and sampling lakes

We selected a total of 23 lakes in Hubei Province，the middle reach of the Yangtze River Basin， China，and these lakes were mainly used for fish production and away from other human related activities (Fig. 1，Tab. 1). According to their aquaculture practices，these lakes were divided into 4 groups: (1) group A(n=7) or reservoir group， mainly natural reservoirs，used as backup drinking water source and for irriga-2tion. These lakes were stocked with about 9000 kg/km of carps (mainly Silver CarpHypophthalmichthys molitrixand Bighead CarpHypophthalmichthys nobilis)and harvested every year; (2) group B (n=2) or aquaculture ban group， lakes that were banned for aquaculture for 2 years. No fish was stocked in these lakes; (3) group C (n=6) or low stocking aquaculture group， lakes that were s2tocked with the above carps for around 35000 kg/km. No external feed was input into these lakes and fish were harvested every year;and (4) group D (n=8) or high stocking aquaculture group， lakes that were stocked with the above carps for around 120000 kg/ km2. Fish in these lakes were fed with commercial feed for about 550000 kg/km2and harvested every year.

1.2 Field sampling and lab analysis

Samplings of phytoplankton were performed at 0.5 m below the surface in the pelagic zone of the lakes from July to September 2015. To better represent the whole lake， three to six sampling stations were established according to the area of each lake.At each station， one 1 L water sample was collected and fixed by 1% neutral Lugol’s solution， and allowed to settle for 48h. The supernatant was then siphoned off with a small-diameter silicone tube to a final volume of 30 mL[34]. After mixing， 0.1 mL of phytoplankton sample were identified to genera[35—39]and counted in a counting chamber under a microscope (CX41， Olympus， Tokyo， Japan) at ×400 magnifications. Algae in clusters， such asMicrocystis，their individual cells were counted directly.

1.3 Data analyses

We calculated the mean values from the abovementioned three to six stations for each lake. The mean value for each group was calculated from all the lakes in that group. Then， the following analysis was conducted.

First， the dominant taxa were determined by the dominance value [Yi， Eq. (1)]. Species diversity indices were calculated by Shannon-Wiener index [H′，Eq. (2)]， Margalef index [DMa， Eq. (3)] and Pielou index [J， Eq. (4)][14];

Fig. 1 The location of sampling lakes

whereniis the total cell number of each genus in a sample，Nis the total phytoplankton cell number in a sample，fiis the occurrence frequency of each genus，andSis the total number of genus. Taxa withYi＞0.02 were selected as dominant genera[17].

One-way ANOVA was used to test significant differences in the phytoplankton abundance and diversity indices among groups. Prior to ANOVAs， all variables were tested for normality (Shapiro-Wilk test) and homogeneity (Levene test)， and phytoplankton abundance were log10(X+1) transformed to satisfy the presuppositions of ANOVAs. Statistically significant differences (P＜0.05) among groups were further assessed using the LSD (the least significant difference) test. All statistical analyses were conducted with software SPSS 20 and Primer 6.

2 Results

2.1 Phytoplankton community composition

Eight phytoplankton phyla were identified in the 23 lakes: Chlorophyta， Cyanophyta， Bacillariophyta，Pyrrophyta， Euglenophyta， Cryptophyta， Chrysophyta and Xanthophyta. The highest taxa number was observed in Chlorophyta (65 genera)， followed by Cyanobacteria (19)， Bacillariophyta (16)， Pyrrophyta(4)， Euglenophyta (4) and Cryptophyta (2)， while Chrysophyta and Xanthophyta were represented by only one taxa.

Common genera in the lake groups were shown in Tab. 2. The four lake groups shared similar phytoplankton taxa. Chlorophytes and cyanobacteria were the dominant phyla of all the four lake groups (Fig. 2).The average number of genera in groups A， B， C and D were 39， 51， 52 and 50， respectively (Fig. 2).

2.2 Phytoplankton abundance

Tab. 1 Basic information of sampling lakes in the middle reach of the Yangtze River Basin

The lowest total phytoplankton abundance[(5.50×107±1.38×108) cells/L] was observed in group A， which was significantly (P＜0.05) lower than groups B [(1.57×108±4.95×106) cells/L]， C [(3.25×108±2.07×108) cells/L]， and D [(1.41×108±9.78×107) cells/

L] (Fig. 3). Cyanobacteria were the first abundant group in all the lake groups. Relative abundance of eight phytoplankton phyla differed among lake groups B， C and D. Relative abundance of cyanobacteria in group D was the lowest， followed by groups B and C， while relative abundance of chlorophytes in group D was the highest， followed by groups B and C.

Tab. 2 Dominance value (Yi) of common phytoplankton genera in water samples of lakes in the middle reach of the Yangtze River Basin

2.3 Dominant genera and diversity indices

The dominant taxa were different among the lake groups (Tab. 2). The dominant genera in group A werePseudanabaena(Y=0.642) andPlanktolyngbya(Y=0.064) (Tab. 2). The dominant genera in group C werePseudanabaena(Y=0.395)，Merismopedia(Y=0.097)，Planktolyngbya(Y=0.122)，Cylindrospermopsis(Y=0.083)，Raphidiopsis(Y=0.060)，Limno-thrix(Y=0.058) andDolichospermum(Y=0.022) (Tab. 2).The dominant genera in both groups A and C belonged only to cyanobacteria， while the dominant genera in groups B and D were more diverse. The first dominant genus in group B wasMerismopedia(Y=0.428)， followed byPlanktolyngbya(Y=0.118)，Pseudanabaena(Y=0.133)，Scenedesmus(Y=0.066)，Dolichospermum(Y=0.052) andMicrocystis(Y=0.047) (Tab. 2). Group D was dominated byMerismopedia(Y=0.308)，Microcystis(Y=0.118)，Pseudanabaena(Y=0.077)，Raphidiopsis(Y=0.072)，Limnothrix(Y=0.066)，Scenedesmus(Y=0.058)，Crucigenia(Y=0.029) andCyclotella(Y=0.025) (Tab. 2).

Fig. 2 Phytoplankton composition in water samples of lakes in the middle reach of the Yangtze River Basin

Fig. 3 Phytoplankton abundance in water samples of lakes in the middle reach of the Yangtze River BasinDifferent lower-case letters indicate sifnificant differences

Tab. 3 Diversity indices of phytoplankton community in water samples of lakes in the middle reach of the Yangtze River Basin

IndexH′ in the four groups A， B， C and D were 2.91±1.46， 2.61±0.53， 2.50±0.92， and 2.95±0.59， respectively (Tab. 3). IndexDMain the four groups A，B， C and D were 2.45±0.92， 2.62±0.18， 2.58±0.38，and 2.64±0.32， respectively (Tab. 3). IndexJin the four groups A， B， C and D were 0.54±0.25， 0.46±0.10， 0.44±0.15， and 0.52±0.11， respectively (Tab. 3).None of these indices was significantly different(P＞0.05) among the four lake groups.

3 Discussion

3.1 Phytoplankton community composition

Feed and fertilizers are often used in intensive fish farming systems. Materials from uneaten food，faeces， excretion， fertilizers and aquatic drugs can change the concentrations of organic and inorganic nutrients， which in turn affect growth and composition of phytoplankton[40]. Plankton feeding or filterfeeding fish may affect the diversity of phytoplankton and the effect varies with the species stocked. In general， stocking plankton feeding fish can enrich the diversity of phytoplankton in the system. Only a few fish species stocked may reduce the diversity of phytoplankton community[41]. In the current study， the average number of genera in groups B， C and D was higher than that in the group A. Mercurio，et al.[25]found that 36 genera was observed from the aquaculture sites while 30 genera was observed from the nonaquaculture areas. In the research of Bartozek，et al.[22]， no significant discrepancy among the composition， abundance and diversity of phytoplankton was observed when comparing the cage cultured and non-cultured areas. According to their analysis， this no discrepancy was probably due to two reasons: one was few cage and fish used， and the other was the hydrobiology affected by the upper region of the river[22].

3.2 Phytoplankton abundance

The phytoplankton abundance in the reservoir group was significantly (P＜0.05) lower than other three lake groups， in which lots of fish was either being cultured or once cultured. Previous studies[25，42，43]found similar results when comparing phytoplankton in aquaculture and non-aquaculture areas. Large amounts of artificial feed are usually used in aquaculture but aquatic animals only absorb 25% to 35% of these feed[44]， and more uneaten left in the water could accelerate the development of phytoplankton. What is more， fertilizers used can also increase the abundance and biomass of phytoplankton[45]. In the current study，the lowest phytoplankton abundance was observed in group A， which was consistent with the lowest nutrient concentration in the water [Total nitrogen， TN=(0.42±0.24) mg/L; Total phosphorus， TP=(0.02±0.01)mg/L][46]. Besides， the water depth of reservoirs in group A is ranging from 5.0—29.2 m， deeper than lakes in other three groups， which may have enhanced the vertical loss of phytoplankton cells driven by sinking and P-translocation through sedimentation，or migration[47]. Deep lakes with similar levels of nutrients tend to have lower phytoplankton biomass than shallow lakes that without thermal stratification[48].Though there was no statistical difference among the groups B， C and D， the total phytoplankton abundance in group B was lower than group C but higher than group D. The dominance value ofMerismopediain group B was higher than that in group D，which partly accounted for the higher abundance of B. Another likely cause of high abundance in group B was the comparable nutrient concentrations [group B:TN=(1.36±0.71) mg/L， TP=(0.22±0.00) mg/L; group D: TN=(2.27±0.77) mg/L， TP=(0.44±0.38) mg/L][46]in water due to past fish culture activities. One point we possibly can confirm from this is that it is difficult to restore a lake just by virtue of banning fish culture in a short time[49—51]. Both increased input of nutrients and morphology of lakes could contribute to cyanobacterial development[52]. High abundance of cyanobacteria in group C was possibly related to rich nutrient [TN=(0.95±0.53) mg/L， TP=(0.11±0.08)mg/L][46]and dominance by filamentous species (Tab. 2)was probably related to large area and shallow depth of group C (Tab. 1)[52]. In addition， the species and densities of fish cultured may influence the phytoplankton abundance. Currently， there are two common explanations related to this: (1) omnivorous fish can increase phytoplankton as its predation on zooplankton that reduced the grazing pressure from zooplankton; (2) fish can change the concentration， proportion and cycling rate of nutrients because of excretion and sediment resuspension[13，53]. Silver and bighead carps are filter-feeding fishes; they share a similar trophic niche and compete for natural resources[54].Higher densities and fewer food resources increased diet breadths but decreased the diet overlap in both types of carps[55]. Silver and bighead carps were probably be released from diet competition and shift to feed mainly on zooplankton in group C with lower densities than group D， decreasing the efficiency of controlling cyanobacterial blooms and resulting the highest abundance of cyanobacteria in group C. While grazing pressure from fish in the group D was smaller than group C and nutrient concentration was higher due to the large use of commercial feed， and thus abundance of cyanobacteria in group D was less than that in group C. Liang，et al.[56]observed that abundance of all phytoplankton phyla in a pond with high density of silver carp were lower than those in the pond with lower density of silver carp at the later phase of experiment. However， according to Jiang，et al.[57]， fish farming did not significantly increase the phytoplankton abundance. And they argued that there was not a simple correlation between the nutrient load of fish farming and the phytoplankton abundance as the phytoplankton community at the cultured area is more controlled by both top-down and bottom-up effects.

3.3 Dominant genus and diversity indices

Clear dominance of cyanobacteria in all the four lake groups during our study period is consistent with the preference of cyanobacteria for higher temperatures[47]. Rich nutrient and fish can promote the growth of cyanobacteria and green algae[58]. Jensen，et al.[59]found that cyanobacteria and chlorophytes were the predominant groups when nutrient concentrations increased in Danish shallow lakes， and argued that chlorophytes will outcompete cyanobacteria in hypertrophic conditions (＞1 mg/L TP)， owing to faster growth of chlorophytes under continuous input of nutrients from external and internal sources. Liang，et al.[56]found that at the later phase of the experiment，the dominant taxa in the pond under high density of silver carp tend to shift from cyanobacteria to other algae groups. This pattern was partially consistent with what we found in the present study: the dominant taxa in groups A and C with low fish stocking density and nutrient concentration was cyanobacteria，the dominant taxa in group B with rich nutrient were both chlorophytes and cyanobacteria， while the dominant taxa in group D with highest fish stocking den-sity and nutrient concentration were three classes，chlorophytes， cyanobacteria and diatoms. Along with nutrient and predation， physical features such as lake morphology and water depth might contribute to between-lake and group differences in phytoplankton community[52，60，61]， which had been partly referred to during the discussion of discrepancy in phytoplankton abundance. Large standard deviation of phytoplankton abundance within lake groups was possibly jointly caused by variance of nutrient， predation and hydrological environment (Fig. 3).

The phytoplankton biodiversity index is the most commonly used indicator to evaluate the trophic state of a lake[17]. In this study，H′ values in the four lake groups were not significantly different (P＞0.05) and ranged from 2.50 to 2.95， which indicated a moderately polluted state of water in the four lake groups during the study period[62]. However， the trophic state of group A was different from that of groups B， C and D in the view of water quality[46]. The Shannon-Wiener index is the overall evaluation of richness and evenness. TheH′ value in community with high richness and low evenness can be the same with that in community with low richness and high evenness. No significant difference ofH′ value among the four lake groups could be the result of different richness and evenness. Any changes of the environment can usually cause variances of species diversity. The biodiversity index should be used only as a reference index in the evaluation of trophic state and water quality monitoring[62].

One limitation of the current study is that we only had two lakes available for the aquaculture ban group (B). This type of lake management is still new and at the testing stage. But this group does play an important role in helping us understand the aquaculture ban on lake ecosystems. Further， we had unequal numbers of lakes in other groups. The current study was designed to catch the overall condition for each group， and then we tried to reach this goal to sample as more lakes as possible for each group.

4 Conclusions

In this study， effects of aquaculture on phytoplankton of lakes in the middle reach of the Yangtze River Basin were studied. The results showed that the dominant phytoplankton genera of the four groups were different. The dominant genera of group D were cyanobacteria， chlorophytes and diatoms. The total abundance of phytoplankton in group A was significantly lower than that of the other lake groups. No significant difference in the Shannon-Wiener index，Margalef index and Pielou index was observed among the four lake groups. The results indicated that fish culture could affect the abundance and dominant genera of phytoplankton in lakes.

Acknowledgements:

Many thanks to Zhao Liang， Cai Fang-Fang，Wang Jing-Ya and Zheng Ling-Ling from algae labs at Institute of Hydrobiology， Chinese Academy of Sciences， for their assistance in the identification of phytoplankton samples.