INTRODUCTION
Planktonic protists are the main components of the microplankton community and play an important role in the functioning of mirobial food webs, especially in terms of energy flow and element cycling in aquatic ecosystems (Finlay & Esteban, Reference Finlay and Esteban1998). Autotrophic protists are responsible for the bulk of primary production in most aquatic habitats; protozoan grazers transfer the production of algae and of the bacteria that grow on algal exudates to higher trophic levels in the food chain. Structural changes of protist communities may significantly affect other components of the aquatic food web, and thus may influence the distribution and abundance of both lower and higher organisms (Finlay & Esteban, Reference Finlay and Esteban1998). Some protists can tolerate extreme environmental conditions and inhabit biotopes that are unfavourable to most metazoans (Patterson et al., Reference Patterson, Larsen and Corliss1989). Furthermore, with their rapid growth and delicate external membranes, protists may react more quickly to environmental change than most other eukaryotic organisms and thus serve as bioindicators of water pollution (Shen et al., Reference Shen, Zhang, Gong, Gu, Shi and Wei1990).
Semi-enclosed mariculture waters are typically characterized by their small size, poor exchange of water with the sea, heavy disturbance from the introduction of cultured animals, and high nutrient and/or contaminant inputs either from mariculture sources or of autochthonous origin. This commonly results in eutrophic or hypertrophic environments that are subject to recurrent eutrophication events. Furthermore, environmental conditions (e.g. water temperature, salinity, pH, nutrients, etc.) are often highly variable on short spatial and/or temporal scales resulting in significant changes in the abundance, biomass, diversity and community structure of microorganisms (Nuccio et al., Reference Nuccio, Melillo, Massi and Innamorati2003). Although there have been a number of reports of protist community dynamics in mariculture waters and in semi-enclosed marine habitats such as lagoons (Pitta et al., Reference Pitta, Karakassis, Tsapakis and Zivanovic1998; Gilabert, Reference Gilabert2001; Cytryn et al., Reference Cytryn, Gelfand, Barak, van Rijn and Minz2003), such studies have yet to be carried out in semi-enclosed mariculture waters.
A six-month baseline survey of planktonic protists was carried out in a semi-enclosed shrimp-farming pond near Qingdao, China. The farming of shrimp was responsible for great variations of environmental factors in the pond thus offering the opportunity for us to analyse the relationships between planktonic protists and a range of physico-chemical and biological parameters. It should be noted, however, that because of the constraints of the methods used not all protist groups could be investigated so no data are available for the picoplanktonic forms (0.2–2 µm), for example. The aims of this study were to document the taxonomic composition and the temporal variation of planktonic protists in a semi-enclosed mariculture pond, to monitor the community structure of the planktonic protists and investigate relationships with a range of environmental factors.
MATERIALS AND METHODS
Study site
The shrimp-farming pond is located on the Laoshan Bay coast near Qingdao, China (Figure 1). It is a shallow marine pond, maximum depth about 1.2 m, with a mud–sandy bottom and covers an area of about 800 m2. It is connected to the sea via a long and narrow canal that may be closed by means of a sluice gate. The shrimp juveniles were introduced on 15 June 2002 and fed with artificial granular foodstuff after two weeks. During the period of study (May to October), the water depth in the pond increased by 1.1 to 1.2 m about every two weeks due to seawater influx.
Fig. 1. Map showing the location of the semi-enclosed shrimp-farming pond on the Laoshan Bay coast.
Sampling
Fifteen samples (referred to as 22 May etc) were collected every ten days from May to October 2002. All water samples were collected at a depth of 0.5 m. For quantitative studies and for the identification of ciliate, 1000 ml water samples were fixed with Bouin’s solution to a final concentration of 10%. For the identification of autotrophs, 1000 ml of water was concentrated to 50 ml by filtering through a 20 µm-mesh plankton net in situ (Shen et al., Reference Shen, Zhang, Gong, Gu, Shi and Wei1990).
For the measurement of concentrations of dissolved inorganic nitrogen (DIN, sum of NO3-N, NO2-N and NH3-N) and soluble reactive phosphate (SRP), 1000 ml seawater was collected and analysed following standard methods (APHA, 1989). For the measurement of chlorophyll-a (chl a), a further 500 ml water sample was filtered through Whatman 25 mm GF/F filters by gentle vacuum filtration; following extraction of the filter paper in 90% DMF (N, N-dimethyl foramide) for 24 hours at 4°C, the concentration of chl a in the supernatant was determined using a spectrophotometer (UV-1601, Shimadzu) (Talling & Driver, Reference Talling, Driver and Oi1961). For enumeration of bacteria, glutaraldehyde was added to 100 ml seawater to give a final concentration of 1.25%.
Water temperature, salinity and pH were recorded with appropriate sensors (WTW) at a depth of 0.5 m, and turbidity was measured in situ using a turbidimeter (Hach 2100P, Hach).
Identification and enumeration
For identification of ciliates, the quantitative protargol staining (QPS) method (Montagnes & Taylor, Reference Montagnes and Taylor1994) and the identification guides of Song et al. (Reference Song, Zhao, Xu, Hu and Gong2003) were used. Identification of other protists was performed following Steidinger & Tangen (Reference Steidinger, Tangen and Tomas1997).
For protist enumeration (except for ciliates), 800 ml of Bouin’s-fixed sample was settled for 48 hours resulting in 30 ml of concentrated sediment. A 0.1 ml sub-sample of this sediment was placed in a Perspex counting chamber and enumerated under a light microscope at a magnification 400× (Shen et al., Reference Shen, Zhang, Gong, Gu, Shi and Wei1990). For ciliate enumeration, 200 ml of Bouin’s-fixed sample was directly used for enumeration of ciliated protozoa by the QPS method (Montagnes & Taylor, Reference Montagnes and Taylor1994). All counts were repeated three times, and three sub-samples from each sample yielded a SE of <8% of the mean values of counts.
Biovolume estimates were determined for nanoplankton (2–20 µm) from microscopical measurement of cell dimensions and assuming spherical or ellipsoidal shape. Microplankton (20–200 µm) biovolumes were determined from measurements of their linear dimensions and using volume equations of appropriate geometric shape (Winberg, Reference Winberg1971). Phototrophic and heterotrophic nanoplankton was converted to carbon biomass using a conversion factor of 183 fg C µm−3 (Dennett et al., Reference Dennett, Caron, Murzov, Polikarpov, Gavriliva, Georgieva and Kuzmenko1999). Diatom biovolumes were converted to carbon values using the modified Strathmann equation (Smayda, Reference Smayda and Sournia1978). For other microplankton, conversion factors were 140 fg C µm−3 for dinoflagellates and non-loricate ciliates, and 53 fg C µm−3 for tintinnid ciliates (Putt & Stoecker, Reference Putt and Stoecker1989; Stoecker et al., Reference Stoecker, Sieracki, Verity, Michaels, Haugen, Burkill and Edwards1994).
Bacteria in seawater were counted by epifluorescence microscopy. Cells were stained by adding the DNA specific fluorochrome, 4′6-diamidino-2-phenylindol (DAPI; Sigma) to a final concentration of 0.12 mg ml−1 and collected on a black nucleopore filter (0.2 µm pore size) supported by a 0.8 µm backing filter. At least 1000 bacteria were counted in each sample. In all cases, examinations were made at 1000× magnification using UV light excitation (Sherr et al., Reference Sherr, Sherr and Fallon1987).
Data analysis
Species diversity (H′), evenness (J) and species richness (d) of samples were calculated as follows:
where H ′ = observed diversity index; Pi = proportion of the total count arising from the i th species; S = total number of species; and N = total number of individuals.
The community structures of samples were analysed using the PRIMER (Plymouth Routines in Multivariate Ecological Research) package (Clarke & Warwick, Reference Clarke and Warwick1994). A Bray–Curtis similarity coefficient matrix was calculated on root-transformed data, and the separate clusters were identified by hierarchical clustering (CLUSTER) and multidimensional scaling (MDS) ordination. Differences between groups of community samples were tested by the PRIMER program ANOSIM. The contribution of each species to the average Bray–Curtis dissimilarity between groups of samples and to similarity within a group was examined by the program SIMPER analysis.
The multivariate biota-environment (BIOENV) procedure (Clarke & Warwick, Reference Clarke and Warwick1994) was used to explore the potential relationships between the abiotic features of the water and the similarity patterns among biological samples. BIOENV functions within the PRIMER program and allows either a full search of all abiotic variable combinations or of specific subsets, e.g. all combinations containing certain variables or containing a fixed number of variables. Chl a was omitted from the environmental matrix due to its collinearity with temperature. Data for NO3-N, NO2-N NH3-N and SRP were standardized by logarithmic transformation before analysis. Analysis of univariate analysis was carried out using the software SPSS (version 11).
RESULTS AND DISCUSSION
Environmental variables
The values for 10 environmental variables in each of 15 samples are shown in Table 1. Water temperature ranged from 19.5°C to 31°C levelling off steadily from May to June, increasing slowly and then dropping sharply after peaking in late August.
Table 1. Environmental variables in mariculture pond water samples between May and October 2002.
chl a, chlorophyll-a; NTU, nephelometric turbidity units; S, salinity; SRP, soluble reactive phosphate; T, temperature; Tur, turbidity.
Salinity averaged around 28.7 psu, maintaining high levels from May to the middle of July, but dropping sharply to the lowest level (8.1 psu) in end of July (22 July sample) due to the heavy rainfall, and reverting to its original levels from the end of July to the beginning of September.
The pH values remained relatively stable ranging from 7.24 to 8.05 while turbidities exhibited an increasing trend and peaked at the end of September.
Concentrations of chl a were characterized by double peaks, the first of which (60.39 µg l−1) occurred at the end of June and the second in mid-September (77.58 µg l−1).
The turbidity averaged around 8.38 ntu, maintaining low levels from May until mid-September and increasing to a maximum value of 20.3 ntu on 24 September.
The average value of DIN over the whole sampling period was 5.5 mg l−1. There was an initial decline in DIN between May and mid-June followed by an increasing trend after the introduction of shrimp juveniles. NH3-N (mean 3.10 mg l−1) represented 56% of total DIN and exhibited an increasing trend, whereas NO3-N (mean 0.89 mg l−1) and NO2-N (mean 1.31 mg l−1) levelled off steadily reaching maximum values of only 1.8 and 2.4 mg l−1 respectively.
The concentrations of SRP ranged from 0.01 to 6.24 mg l−1, and were much higher in the period after the introduction of shrimp juveniles (13 June) than in the period before.
Densities of bacteria (mean 4.03 × 106 ml−1) ranged from 1.89 × 106 to 7.25 × 106 ml−1 (Table 1), the lowest value being on 22 May (i.e. before the shrimp juveniles were introduced) and the highest on 5 August.
Taxonomic composition and taxa distribution
A total of 54 protists were identified during the six-month survey comprising 4 chlorophyceans, 2 chrysophyceans, 5 cryptophyceans, 10 dinoflagellates, 3 euglenophyceans, 10 diatoms, 18 ciliates and 2 sarcodines (Table 2). Ciliates, dinoflagellates and diatoms were the most common forms, accounting for 32%, 19% and 19% respectively of the species recorded. The other 6 groups were represented by comparatively few species (Table 2; Figures 2 & 3a).
Fig. 2. Composition of planktonic protist communities; the percentage of the total number of species recorded throughout the period of sampling is shown for each group.
Fig. 3. Temporal variations of species richness (A), abundance (B) and biomass (C) of protists obtained.
Table 2. List of the protist species found in 15 samples, including body size, mean abundances and biomass.
Body size (μm): length × width; abundance (ind ml−1): + = 0–50, ++ = 50–500, +++= 500–5000, ++++ = over 5000; biomass (μg l−1): + = 0–10; ++ = 10–100; +++ = 100–1000; ++++ = over 1000.
The species number of protists in the 15 samples varied considerably with respect to the shrimp-farming cycle. The temporal variation of species number showed a bimodal distribution during the six-month period with peaks in August and October. The maximum values were 21 species in August and 20 species in October. Ciliates, dinoflagellates and diatoms were primarily responsible for the two peaks. The lowest species number (6 species) was found in the 26 June sample, the first sample after the introduction of shrimp juveniles (Figure 3a).
Protist species diversity is generally lower in the semi-enclosed biotopes than in more open-water sites. Investigations of temperate coastal, inshore and estuarine sites have revealed over 30 ciliate taxa in the Jiaozhou Bay of Qingdao, China (Gong et al., Reference Gong, Song and Warren2005), compared with the 18 taxa in the shrimp-farming biotope. Diversities of dinoflagellates, diatoms, chlorophyceans, cryptophyceans and chrysophyceans in the pond were also low, but their abundances were very high, so they are therefore likely to make a significant contribution to microplankton dynamics (Gilabert, Reference Gilabert2001; Nuccio et al., 2003).
Abundance and temporal variation
The temporal variation of the protist abundance had unimodal distribution (Figure 3b). The abundances maintained relatively low values (mean 1.35 × 103 ind ml−1) from May to September, followed by a peak in October when maximum cell densities reached 1.23 × 105 ind ml−1. Chlorophyceans (e.g. Mamiella sp.), chrysophyceans (e.g. Chromulina sp.) and diatoms (e.g. Ephenera sp.) were primarily responsible for the October peak reaching abundances of 5.44 × 104 ind ml−1, 3.48 × 104 ind ml−1 and 3.24 × 105 ind ml−1 respectively. Of the total protist abundance for the six-month period, chlorophyceans accounted for 47.71%, chrysophyceans 25.43%, and diatoms 19.45% compared to dinoflagellates (3.13%), cryptophyceans (2.93%), ciliates (0.98%), euglenophyceans (0.35%) and sarcodines (0.02%) (Figure 4).
Fig. 4. Proportions of cumulative abundances (A) and biomasses (B) of protists throughout the period of sampling.
There were 10 dominant species, the individual abundance of which exceeded 30% of the total at some point during sampling period: Chromulina sp., Hillea fusiformis, Hillea marina, Teleaulax acuta, Peridinium sp.2, Peridinium sp.4, Mamiella sp., Thalassiosira sp.1, Ephenera sp., and Pseudoscourfieldia marina. The abundance of five of these (Hillea marina, Hillea fusiformis, Pseudoscourfieldia marina, Teleaulax acuta and Peridinium sp.4) had one high peak and at least one other smaller peak whereas the other five (Chromulina sp., Peridinium sp.2, Mamiella sp., Thalassiosira sp.1 and Ephenera sp.) occurred in significant abundances on only one occasion (Figure 5). These ten dominant species showed a clear succession from May to October (Figure 5).
Fig. 5. Abundance (ind ml−1) and temporal succession of the 10 dominant protist species.
At least one previous investigation has demonstrated the numerical dominance of flagellates and cryptophyceans, with relatively low numbers of other protist groups such as diatoms, in a marine lagoon (Nuccio et al., Reference Nuccio, Melillo, Massi and Innamorati2003). Moreover, the studies on other semi-enclosed marine waters, for example various Mediterranean lagoons, the Varano Lagoon near the Adriatic Sea, and the Center-Western Sardinia lagoon, have also revealed that protist abundances show distinct peaks in abundances of a limited number of species in the summer and occasionally in the winter, mainly flagellates (e.g. chlorophyceans, cryptophyceans and euglenophyceans) and diatoms (Gilabert, Reference Gilabert2001). There are, however, no previous studies of protist communities in semi-enclosed shrimp-farming ponds with which to compare our data.
In the present study, the autotrophic groups (cryptophyceans, chlorophyceans and other small flagellates) were the most abundant mainly due to their ability to bloom rapidly. Similar findings have previously been reported for marine lagoons (Gilabert, Reference Gilabert2001; Nuccio et al., Reference Nuccio, Melillo, Massi and Innamorati2003).
Compared with lower latitude fjords, the protist abundance in the semi-enclosed pond is very high. Ciliate abundance in our sampling pond, for example, ranged between 1.6 to 7.8 × 105 ind l−1, whereas in Ellis Fjord, maximum abundance only reached 2.2 × 102 ind l−1 (Grey et al., Reference Grey, Laybourn-Parry, Leakey and McMinn1997). Compared with coastal waters around the Antarctic, the ciliate abundances in the mariculture pond were also high. In Admiralty Bay, for example, ciliate numbers remained little more than 3–4 × 103 ind l−1 (Brandini, Reference Brandini1993).
Biomass and temporal variation
The temporal variation of biomass exhibited a clear bimodal distribution with one peak in July and another in October, but only the latter corresponded to the abundance peak (Figure 3c). The maximum values were 12.14 mg l−1 in July and 16.74 mg l−1 in October. Dinoflagellates (e.g. Peridinium sp.4) and ciliates (e.g. Urotricha venatrix) were responsible for the July peak when their biomasses reached 8.43 and 3.37 mg l−1 respectively, while diatoms, dinoflagellates and ciliates were the major contributors to the October peak with biomass values of 7.52, 5.96 and 1.79 mg l−1 respectively. Ciliates, dinoflagellates and diatoms accounted for 42.68%, 39.04% and 12.45% respectively of the total protist biomass compared to chrysophyceans (2.23%), cryptophyceans (1.93%), sarcodines (0.65%), euglenophyceans (0.60%) and chlorophyceans (0.35%) (Figure 4).
Temporal patterns of community structure
Although autotrophic and heterotrophic species appeared in almost all samples, the patterns of protist communities in 15 samples exhibited a clear temporal succession in relative species composition, abundance and biomass (Figure 6). In terms of the relative abundances, the patterns of protist communities might be distinguished as five structural types each of which dominated at different times during the period of study: (1) cryptophyceans (maximum 96.18%) dominated the protist communities from May to June; (2) dinoflagellates (maximum 56.73%) in June and July; (3) diatoms (maximum 71.44%) in August; (4) chrysophyceans (maximum 83.05%) in September; and (5) chlorophyceans (maximum 96.54%) in October (Figure 6b). The temporal variation of relative biomass showed an alternate functional change of protist communities: the heterotrophic protists, mainly represented by the ciliates, gave rise to four peaks, one each in June, August, September and October, followed by three peaks of autotrophic forms in July, August and October, respectively (Figure 6c).
Fig. 6. Temporal succession of protist communities during the period of study.
The variation of protist community structure showed high frequency oscillations with rapidly increasing and decreasing blooms. The lowest densities occurred from May to August and highest ones during September and October 2002. Despite the variability in the protist abundance and biomass over short time scales, a recurrent pattern was apparent. During the sampling period, the autotrophic protists were always the dominant group, mainly cryptophyceans during May and June, dinoflagellates from July to August, chrysophyceans and chlorophyceans during September and October 2002, and diatoms in August and October. The protist community structure appeared to be more diverse from early July to the middle of September with varying contributions of ciliates, dinoflagellates, diatoms, cryptophyceans and chlorophyceans. To some extent, these findings are consistent with previous findings, e.g. the elevated number of cryptophyceans when temperatures are lower, the high abundances of certain flagellate groups such as the chlorophyceans and cryptophyceans associated with poor abundances of diatoms, and the co-dominance of diatoms and dinoflagellates in the summer period (Dennett et al., Reference Dennett, Mathot, Caron, Smith and Lonsdal2001; Nuccio et al., Reference Nuccio, Melillo, Massi and Innamorati2003). Pitta et al. (Reference Pitta, Karakassis, Tsapakis and Zivanovic1998) also found evidence that seasonal factors were more important than the effects of fish-farming in influencing planktonic protist communities, although this study was carried out in an open, rather than a semi-enclosed, mariculture system.
A dendrogram of the 15 samples was plotted using group-average clustering from Bray–Curtis similarities on square root transformed abundances (Figure 7a). The cluster analysis resulted in the 15 samples falling into 2 groups at a 25% similarity level (P < 0.001): group I was composed of the first eight samples (May to beginning of August), and group II the last seven samples (from August to October). The MDS ordination shows a temporal distribution of samples in agreement with the dendrogram with 2 groups appearing at separated locations on the plot (Figure 7b).
Analysis of similarities (ANOSIM) revealed that the two groups are significantly different at the 85.86% dissimilarity level (P < 0.001). Similarity percentage (SIMPER) analysis showed that the cryptophyceans Teleaulax acuta and Hillea fusiformis and chlorophycean Pseudoscourfieldia marina dominated group I, while the chrysophycean Chromulina sp. and the dinoflagellates Gyrodinium spirale and Prorocentrum rostratum dominated group II. Furthermore, ANOSIM indicated that group I was clustered into three subgroups (dissimilarity 74.15%; Ia, Ib and Ic) and group II into two subgroups (dissimilarity 73.81%; IIa and IIb) at P < 0.01 level (Figure 7). SIMPER analysis revealed that: subgroup Ia was characterized by the cryptophyceans Hillea fusiformis and H. marina; subgroup Ib by the chlorophycean Pseudoscourfieldia marina and the cryptophycean Teleaulax acuta; subgroup Ic, the cryptophycean Teleaulax acuta and the dinoflagellate Peridinium sp.4; subgroup IIa, the dinoflagellate Peridinium sp.3 and the cryptophycean Teleaulax acuta; and subgroup IIb, the chrysophycean Chromulina sp., the chlorophycean Mamiella sp. and the dinoflagellate Prorocentrum rostratum.
Fig. 7. Cluster analysis (A) and MDS ordination (B) of protist communities of 15 samples. 1, 22 May; 2, 2 June; 3, 13 June; 4, 26 June; 5, 8 July; 6, 18 July; 7, 26 July; 8, 5 August; 9, 15 August; 10, 25 August; 11, 4 September; 12, 14 September; 13, 24 September; 14, 4 October; 15, 14 October; I, group I; II, group II; Ia, Ib, Ic, subgroups in group I; IIa, IIb, subgroups in group II.
Interaction between biota and environmental variables
The correlations (Spearman, SPSS) between the various environmental variables and abundance, biomass, biomass/abundance ratios (B/A), species diversity, species evenness and species richness in 15 protist samples are shown in Table 3. The protist abundance exhibited significant positive correlations with turbidity, chl a, DIN, DIN + SRP, NH3-N and NO2-N, and a significant negative correlation with water temperature. For protist biomass, significant positive correlations were found with DIN, DIN + SRP, NH3-N, NO2-N and pH. By contrast, the biomass/abundance ratio (B/A) was positively correlated with water temperature, but was negatively correlated with chl a and the concentration of bacteria. Both diversity (H ′) and evenness (J) indices of protist species showed significant positive correlations with water temperature and were negatively correlated with salinity, while the species richness values only exhibited significant negative correlation with salinity (Table 4).
Table 3. Correlations between environmental variables and species diversity (H′), species evenness (J), species richness (d), abundance, biomass and biomass/abundance ratio (B/A) of the protist community.
* P < 0.05; **P < 0.01; DIN, dissolved inorganic nitrogen; see Table 1 for other abbreviations.
Table 4. Correlations between abundance of 10 dominant protists and environmental variables (see Table 1 for abbreviations).
* P < 0.05; ** P < 0.01.
Table 4 summarizes the correlations between abundance of dominant species and environmental variables. The chlorophycean Pseudoscourfieldia marina was negatively correlated with DIN (r = –0.681, P < 0.01); Teleaulax acuta was positively correlated with water temperature (r = 0.521, P < 0.05), but was negatively correlated with NO2-N and chl a; Mamiella sp. and Chromulina sp. were positively correlated with turbidity (r = 0.684, 0.858, P < 0.01), and also with nutrients such as DIN and NH3-N (Table 4); there was a strong positive correlation between Peridinium sp.4 and salinity (r = 0.541, P < 0.05) and total nutrients (r = 0.603, P < 0.05); a significant positive correlation between Ephenera sp. and pH was also noted (r = 0.612, P < 0.05).
For all 15 samples collected during the six-month period, the top 6 correlations between biota and environmental variables, established by BIOENV analysis, are dominated by nutrients, salinity and water temperature (Table 5). The highest correlation occurred with the combination of NO3-N and SRP. It was also found that the nutrient SRP is the only one variable that was in each of the top six correlations with both abundance and biomass (Table 5).
Table 5. Summary of results from biota-environment (BIOENV) analysis, with the top 6 correlations corresponding to different variables.
r, Spearman correlation coefficient; for abbreviations see Table 1.
Species diversity, evenness and richness indices are commonly employed in community investigations and are amenable to simple statistical analysis (Ismael & Dorgham, Reference Ismael and Dorgham2003). In our case, however, univariate correlation analysis showed that these indices only correlated at a high level of significance with water temperature and salinity, but failed to show significant correlations with nutrients. All three indices sharply increased in the end of June and exhibited a peak in the middle of August. This might have been due to low salinity and high water temperature during this period.
Univariate correlation analysis demonstrated that there were significant correlations between the protists (abundance and biomass) and environmental variables, such as water temperature, turbidity, chl a and nutrients. The chrysophycean Chromulina sp. and the chlorophycean Mamiella sp. showed significant correlations with nutrients, which might be the reason that these two species dominated the communities and gave rise to the October peaks of protist abundance. Furthermore, their strong correlation with turbidity suggests that Chromulina sp. and Mamiella sp. were largely responsible for the elevated levels of turbidity in the water during this period. The cryptophycean Teleaulax acuta bloomed in the warmest months (from May to August) and was positively correlated with water temperature, while the chlorophycean Mamiella sp., which dominated the protist communities in the autumn when the water was cooler, was negatively correlated with water temperature.
The biomass/abundance (B/A) ratio of the community, i.e. the mean protist body-size, showed a strong positive correlation with water temperature and negative correlations with chl a and the concentration of bacteria. That is to say, the higher the water temperature, the more large-sized species were present, and the higher the abundance of phototrophic microorganisms or bacteria, the more small-sized forms dominate. This is consistent with the use of abundance/biomass comparison (ABC) plots to determine levels of disturbance (Warwick, Reference Warwick1986). This method, which is usually used in benthic macrofauna studies, might thus also be suitable for biomonitoring levels of organic pollution using planktonic protist communities.
Multivariate analyses were more sensitive than univariate ones for detecting changes in community structure. Cluster analysis revealed the difference of protist communities between summer (group I) and autumn (group II). Furthermore, MDS ordination analysis clearly showed the succession sequence of 15 communities due to the disturbance from shrimp-farming activities: subgroup Ia, which was composed of the first three samples, represented the community structure before the introduction of shrimp juveniles; subgroup Ib, which comprised only the first sample after shrimp-farming started, was significantly different in terms of community structure, almost certainly due to disturbance caused by shrimp-farming; subgroup Ic (8 July, 18 July, 26 July and 5 August), IIa (15 August, 25 August and 4 September) and IIb (14 September, 24 September, 4 October and 14 October) exhibited variations in the structural patterns of protist communities during the shrimp-farming period.
The BIOENV analysis demonstrated that NO3-N and SRP were the most important factors influencing the structure of the planktonic protist community, based on all 15 samples. Moreover, nutrients were always among the top combinations of variables along with salinity and water temperature, suggesting that the succession of protist communities is significantly related to these parameters in the semi-enclosed shrimp-farming waters. It should be noted, however, that the present study was restricted to the nano- and microplanktonic protists. Other methods, such as denaturing gradient gel electophoresis (DGGE) and real-time PCR which have previously been used for analysing prokaryote communities in mariculture ponds (Cytryn et al., Reference Cytryn, Gelfand, Barak, van Rijn and Minz2003), might usefully be employed in order to expand the range of the protist data to include, for example, the picoplanktonic forms.
In conclusion, the results of this study demonstrate that planktonic protists are abundant and diverse in the semi-enclosed mariculture pond near Qingdao and that they are correlated with various environmental parameters including nutrients such as nitrogen (NO3-N) and phosphate (SRP), both individually and in combination. This suggests that planktonic protists might be useful bioindicators of water quality in such systems. However, further studies are needed on a range of semi-enclosed mariculture ponds and over extended time periods in order to verify this conclusion.
ACKNOWLEDGEMENTS
This work was supported by ‘The Natural Science Foundation of China’ (project numbers: 40676076; 30500057; 30570236), the Darwin Initiative Programme (Project No. 14-015) which is funded by the UK Department for Environment, Food and Rural Affairs, and a grant from the Center of Excellence in Biodiversity, King Saud University.
