Temporal and spatial changes of macrobenthos community in the regions frequently occurring black water aggregation in Lake Taihu

Seasonal survey was performed from August 2015 to May 2016 at 50 sampling sites in Lake Taihu to determine the spatial and temporal changes in macrobenthos community and their relationships with environmental variables. A total of 58 macrobenthos species were collected and identified, including 28 species of annelids, 17 species of molluscs, and 12 species of arthropods. Both the community composition and the dominant species changed temporally and spatially. Correspondingly, the macrobenthos biodiversity differed among regions and seasons. The macrobenthos density decreased with increased sediment depth, which is the first report about the vertical distribution of macrobenthos in Lake Taihu. The majority of benthic animals were located within the sediment depth of 0–5 cm and 5–10 cm, accounting for 39.25% and 24.87% of the total abundance respectively. Redundancy discriminate analysis revealed that the main environmental factors affecting the most contributing macrobenthos species were temperature in summer, transparency, dissolved oxygen and pH in autumn, and water depth and dissolved oxygen in winter. Particularly, salinity and conductivity showed high correlation with the macrobenthos community through the whole sampling period. The investigation reveals the inherent spatiotemporal variation of macrobenthos community, and provides references for the biological assessment of water quality in Lake Taihu.


Dominant species.
The dominant macrobenthos species in black water aggregation regions in Lake Taihu varied depending on season and location. There were total 8 dominant species, among which, T. tubife, L. hoffmeistteri, Bellamya aeruginosa, Corbicula fluminea, and T. punctipennis were detected during the whole sampling period, and L. grandisetosus and Propsilocerus akamusi were dominated only in winter, with B. purificata only in summer (Table 2). In the spatial scale, the dominant species varied among sampling sections (Table 3). In particular, L. hoffmeistteri is always the dominant species in all the nine sampling sections, and T. punctipennis dominated only in the sections adjacent to Liangxi River and Xiaoxi harbor. Propsilocerus akamusi was only observed in the section adjacent to Liangxi River. The detailed dominance and horizontal or seasonal variations of the dominated species in different seasons and sampling locations can be found in Tables 2 and 3.
There were 22 macrobenthos collected during the vertical sampling in Moon bay in spring 2015, of which 19 species were located within 0-5 cm of the sediment (Fig. 3B). The macrobenthos density generally decreased with increased sediment depth. The benthic animals located within the sediment depth of 0-5 cm and 5-10 cm accounted for 39.25% and 24.87% of the total abundance respectively. The vertical distribution of macrobenthos     Table 4. The water depth was relatively shallow in winter and spring. The water pH, L and MD at different sampling sites were comparable in the temporal scale. The WT in spring and summer (23.81-28.69 °C) was significantly higher than that in autumn and winter (10.16-13.72 °C). The Eh reached peak (247.74 mv) in spring with the lowest value of 70.82 mv in autumn. The average DO decreased by ~30% in summer compared with the constant DO in other seasons. The SD in autumn and winter was significantly higher than that in spring and summer. Both the COND and SAL reached peak values (668.88 μS·cm −1 for COND and 0.33 mg L −1 for SAL) in winter with the low values in autumn. The TDS fluctuated from 294.24 mg L −1 in autumn to 456.38 mg L −1 in winter. According to Pearson's correlation test, different relationships between the macrobenthos species number, density, biomass and environmental factors in the four seasons were observed (Table 5): (1) the animal species number was positively correlated with the L in summer, with T and WD in winter, and with Eh in spring, but negatively correlated with SD in autumn; (2) the animals density was negatively related with the WT, COND, SAL, and TDS in summer, with SAL in autumn, and positively correlated with T and WD in winter. In spring, the macrobenthos density was negatively related with COND and SAL, but positively related with L; (3) No closely relationship between macrobenthos biomass and environmental factors was found in summer and spring. The biomass was negatively correlated with Eh, DO, pH and SD in autumn, but positively correlated with COND, SAL and TDS in winter.

Correlation of environmental factors with the dominant species.
Based on the length <4 of the DCA axis, RDA was used to analyze the relations between dominant macrobenthos species and environmental factors (Fig. 4). In summer, the densities of dominated B. aeruginosa, T. punctipennis, B. purificata were positively related with SAL, COND, TDS, WT, WD, and MD, but was negatively related with Eh, DO, and T. The dominated C. fluminea and T. tubifex were positively related with L, but negatively related with SD. The dominated L. hoffmeisteri was positively related with L, MD, and WD, but negatively related with pH, Eh, and SD. In autumn, the dominated T. tubifex and C. fluminea were positively related with L and DO, and negatively related with COND, TDS, SAL, and SD. The dominated L. hoffmeisteri, B. aeruginosa and T. punctipennis were positively related with DO, MD and WD, and negatively related with WT, Eh, and pH.
In winter, the dominated B. aeruginosa, C. fluminea and L. grandisetosus were positively related with T, WT, SAL, TDS, COND, Eh and pH, and negatively related with SD and DO. The dominated T. tubifex and L. hoffmeisteri were positively related with MD and pH, and negatively related with SD, DO and L. The dominated T.    Table 3. The dominance (IRI) and horizontal distribution of the dominant macrozoobenthos in black water aggregation areas in Lake Taihu.

Discussion
The present investigation showed that the number of macrobenthos species, density and biomass changed among seasons, during which, the number of species reached the peak in autumn and the density and biomass were highest in summer. The macrobenthos in black water aggregation areas in Lake Taihu were dominated by the Oligochaeta and Chironomus larva, consisting of 93.83% of the total macrobenthos density. This is comparable to the finding by Cai et al. 20 based on the quarterly investigation on macrozoobenthos between February 2007 and November 2008. The dominance by the resistant Oligochaeta and Chironomidae generally suggests the water deterioration [29][30][31] , which is in accordance with the declining water quality when black water aggregation occurs. While large amounts of cyanobacteria accumulated and died under certain conditions like high temperature, slow wind, and weak reoxygenation capacity, thioethers substances such as volatile sulfide were released after the decomposition of dead cyanobacteria. These substances could chemically combine other materials like the heavy metals from sediment, favoring the formation of black water aggregation and water deterioration 27,32,33 . In spite of high density of macrobenthos in autumn, the animals' biomass was relatively low because of the small sizes of Oligochaeta and Chironomus larva. The more frequency of mollusks with high biomass resulted in the relatively high macrobenthos biomasses in total in other seasons (Figs 2 and 3). The macrobenthos community also varied spatially. Among the nine sampling sections, Wuxi harbor had the maximum 42 macrobento species, contrasting to Xiaoxi harbor with the minimum 19 species. In particular, the sampling sections of Taige canal, Wuxi sewage treatment plant and Wangyu River had relatively higher biomasses although the macrobenthos density was low. This is attributed to the high abundance of mollusks in these sections (Fig. 1). Qiu et al. 34 studied that the gastropod like the B. aeruginosa shifted its diets from the planktic to benthic  Table 4. Environmental factors of black water aggregation areas in Lake Taihu. Note: T-climate temperate; WDwater depth; WT-water temperature; Eh-oxidation-reduction potential; DO-dissolved oxygen; SD-secchi depth; COND-conductivity; SAL-salinity; TDS-total dissolved solid; L-the distance from the shore; MD-sediment thickness.  Table 5. Correlation analysis of environmental factors with macrozoobenthos numbers, density and biomass in black water aggregation regions in Lake Taihu.
materials under toxic cyanobacterial bloom. Thus, sandy silt and aquatic plants benefit the growth and reproduction of mollusk 35,36 , which was further evidenced by the silty sand substrate and several macrophyte distributed of  was the lowest with the highest abundance of Chironomidae larvae. This is in accordance with the investigation by Qin et al. 18 and Cai et al. 20 . Because of the deep sediment (averaged > 1.5 m), the organic substances were high in Meiliang Bay. In addition, the cyanobacteria, together with other phytoplankton such as the diatom, tend to accumulate in this region with the decreased dissolution oxygen due to the southeast monsoon in summer. These environments benefit the growth of Chironomidae larvae. The dominance by B. aeruginosa in Gonghu Bay in 2010 20 was not observed in present investigation, which was replaced by T. tubife and L. hoffmeisteri. The macrobenthos density was generally decreased from the offshore towards the centers of the Lake Taihu, which was comparable to the founding by Xu et al. 37 .
There were a few studies reporting the vertical distributions of macrobenthos in freshwater lakes except for these in intertidal zones or rivers with different altitudinal gradients 38-40 . Our study indicated that the macrobenthos in Lake Taihu distributed as deep as 45 cm in the sediment, which was significantly deeper than the 25 cm in Lake Donghu in China. Due to the high demands for dissolved oxygen and their filter-feeding behavior, Polychaeta and Mollusk dominated the upper sediment (0-20 cm). Chironomidae larvae and Oligochaete vertically distributed as deep as 45 cm in sediment, which may be highly related with their low dependence on oxygen and diving behavior to escape from the surface predation. The majority distribution of macrobenthos (88.28%) in upper sediments (0-30 cm) indicated that the oxygen and food resources may be the main environmental factors affecting the vertical distribution of macrobenthos in our study. Some other factors, such as organic matter and grain size, are also studied to regulate the vertical heterogeneous distribution of macrobenthic community 41,42 .
The main environmental factors affecting the macrobenthos community changed seasonally, which were water temperature in summer, transparency, dissolved oxygen, and pH in autumn, and water depth and dissolved oxygen in winter ( Table 5). The varied sensitivities of benthic animals to physiochemical variables contributed to the temporal changes of main environmental factors. For example, the somatic growth and survival of C. fluminea were primarily determined by water temperature 43 . In particular, salinity and conductivity showed highly correlation with the benthic animals through the four seasons. The structuring factors of the macrobenthos community varied based on the numerous studies, for example, the depth in estuary and the sediment quality in deep sea [44][45][46] . Gao et al. 47 studied that water conductivity, along with total nitrogen, were the main environmental factors affecting the distribution of macrobenthos in Lake Taihu from August 2009 to May 2010. The conductivity tended to influence the Oligochaeta and Mollusks mostly 48 , which was in accordance with the present study. Change in water quantity is one major reason affecting the salinity fluctuations. Because of the low rainfall and runoff in winter 49 , Lake Taihu had the highest salinity in winter. Rainfall generally increased from Spring. Nonetheless, evaporation also increased corresponding to the increased temperature, which reached peak in summer 50 . This may contributed to the relatively higher salinity in summer than those in spring and autumn. Different freshwater species showed unequal resistance to salinity fluctuation 51 . Besides, the adult species have generally greater osmoregulatory capability than their larvaes do 52 . The varied salinity-tolerances facilitated the seasonal changes of dominant macrobenthos species corresponding to the salinity fluctuation. Besides, the macrobenthos distribution was highly affected by many other abiotic variables like the nutrient-related factors. The diversity of macrobenthos was generally negatively related with the total nitrogen and total phosphorus. Moreover, Lake Taihu is kind of lakes for fish farming. The predation risk from fishes also influence the macrobenthos community 53 . Thus, more parameters need to be included in future investigations to better understand the biological-environmental relationships.
The habitat heterogeneity is traditionally considered to affect the biological structure 54,55 . Researchers have concluded that habitat complexity is the key variable determining the diversity of zoobenthos community 54,56 . Compared with the present macrobenthos structure in Lake Taihu including 58 species dominated by annelida, mollusks and arthropods, different benthic community structures also located in Yangtze River basin were observed. The Qingjiang River, which is located in the middle reaches of the Yangtze River, has as many as 82 zoobenthos species with the Shannon-Wiener index of 4.36 57 . In the lower reaches of Yangtze River, the water quality of Liangtang River and Dianshan Lake fell in the moderate to seriously polluted status with as little as 10 benthic species and the Shannon-Wiener index of 0.32 58 . In these waters, the benthic community was dominated by the pollution-tolerant species like the L. hoffmeistteri and B. aeruginosa. Although the water quality was not detected in our investigation, the similar dominated species suggest that Lake Taihu was still with the deterioration of water quality.
SCiEntifiC REPORTS | (2018) 8:5712 | DOI:10.1038/s41598-018-24058-y horizontal and vertical samples were sieved through 60 unit mesh sieve at the site and preserved by adding 75% ethanol in 500-mL plastic bottles. The biological samples were then brought back to the laboratory and further fixed by adding 4% formaldehyde solutions 59 , which were sorted, enumerated and identified. The identification of individual specimens was preformed referring to the literatures [60][61][62][63] .
The physico-chemical parameters while sampling sediments were measured, including dissolved oxygen (DO) and water depth (WD) detected by portable dissolved oxygen meter (USA YSI-550A), and water temperature (WT), salinity (SAL), total dissolved solid (TDS), conductivity (COND), oxidation-reduction potential (ORP or Eh) and pH measured by water quality analyzer (China Y2001). The water transparency (indicated by secchi depth (SD)) was measured using a Secchi disk. The distance from the shore (L) was determined using Hand-held Laser Distance Meters.
Data processing. The dominant species was determined based on the relative importance index (IRI), which was calculated as IRI = (W + N) × F, where W and N represents the biomass percent and abundance percent of one species, and F represents the frequency of occurrence percentage 64 . The species with IRI > 1000 was defined as the dominant species.
The diversity of macrobenthos was evaluated by Shannon-Wiener diversity index (H′), Margalef richness index (D), and Pielou evenness index (J′) as H′ = −Σ(Pi)(log 2 Pi); D = (S − 1) ln N; J′ = H′/log 2 S. Pi represents the abundance percent of the species i; S represents the species number, and N represents the abundance of total species.
The analysis of variance (ANOVA) or Kruskal-Wallis H was used to compare the difference in terms of species number, density, biomass and diversity among different sampling sections or among seasons. Kolmogorov-Smirnov was used to validate the normality. ANOVA was conducted if the normality was satisfied. Kruskal-Wallis H test was conducted instead if the normality was violated. Differences were determined using