Spatial and seasonal shifts in bloomforming cyanobacteria

Spatial and seasonal shifts in bloom-forming cyanobacteriain Lake Chaohu: Patterns and driving factorsMin Zhang,1* Yuchao Zhang,1 Zhen Yang,1 Lijun Wei,1,2 Wenbin Yang,1,2 Chao Chen1,2 and Fanxiang Kong1*1State Key Laboratory of Lake Science and Environment, Nanjing Institute of Geography and Limnology, CAS, Nanjing210008, China and 2University of Chinese Academic of Science, Beijing 100049, China

SUMMARYThe patterns of spatial and temporal shifts in bloom-formingcyanobacteria and the driving factors for these patternswere determined by analyzing the distribution of thesecyanobacteria in Lake Chaohu using data from satelliteimages and field samples collected during 2012 and 2013.The cyanobacterial blooms primarily occupied the westernregion of Lake Chaohu, and the direction and speed of theprevailing wind determined the spatial distribution of theseblooms. The cyanobacteria in Lake Chaohu were dominated byspecies of Microcystis and Anabaena. Microcystis reached itspeak in June, and Anabaena had peaks in May and November,with an overall biomass that was higher than that ofMicrocystis. Microcystis generally occupied the western regionof the lake in summer, whereas Anabaena dominated in otherregions and seasons. Temperature may be responsible forthese seasonal shifts. However, total phosphorus (TP), pH,temperature, turbidity and nitrate/nitrite nitrogen determinedthe coexistence of the two genera in different regions insummer. TP was correlated with Microcystis dominance, andpH and light availability were correlated with Anabaena domi-nance. Our results contribute to the understanding of shifts inbloom-forming cyanobacteria and are important for the controlof cyanobacterial blooms.

Key words: Anabaena, bloom-forming cyanobacteria, LakeChaohu, Microcystis, spatial and temporal distribution.

INTRODUCTION

Blooms of harmful cyanobacteria pose a threat to the fresh-water ecosystems used for recreation and drinking watersupplies because they can synthesize toxic secondary metabo-lites, such as cyanotoxins. The cyanobacteria responsible forforming blooms are gas-vacuolate species. These species aredistributed across a number of genera and vary in form andsize from small filaments, such as Anabaena minutissimaLemmermann or Oscillatoria agardhii Gomont, to large globu-lar colonies, such as Microcystis aeruginosa Kützing.

Cyanobacterial blooms have been found in lakes, estuariesand coastal waters, and the dominant cyanobacteria maydiffer depending on the water body. In some waters,cyanobacterial blooms are dominated by one genus, albeitwith the occurrence of other genera. For example, Microcystissp. is dominant in Lake Taihu, China (Chen et al. 2003; Denget al. 2014) and in Lake Erie, USA (Michalak et al. 2013),and Planktothrix rubescens De Candolle ex Gomont is domi-nant in Lake Zurich, Switzerland (Posch et al. 2012). Otherwaters may contain different dominant cyanobacteria. For

example, Aphanizomenon flos-aquae Lemmermann andNodularia spumigena Bornet & Flahault are dominant in theBaltic Sea (Kanoshina et al. 2003), and Aphanizomenon sp.and Microcystis sp. are dominant in Lake Dianchi, China(Zhang et al. 2012; Paerl & Otten 2013). To a large degree,these differences in dominance can be explained by climaticand meteorological conditions, which influence the degree ofstratification and mixing, and by light and nutrient availability(Huisman et al. 2005). These physical and chemical factorsplay key roles in mediating competitive interactions amongspecies.

Interactive physical, chemical and biotic factors areinvolved in controlling the growth and dominance of bloom-forming cyanobacteria (Paerl & Fulton 2006). The distinctmorphological, physiological and ecological characteristics ofindividual species suggest that the factors that promote onespecies will not necessarily promote another in the same way.For example, Microcystis aeruginosa may co-exist withAnabaena flos-aquae but does not co-exist with P. rubescens(as Oscillatoria rubescens; Oliver & Ganf 2002). Differentdominant species may co-exist in the same water body,although they seldom make similar contributions to the totalbiomass in one column, or they may dominate differentregions of the water body and shift seasonally according to ourobservations in Lake Dianchi, Lake Taihu and Lake Chaohu.However, on small spatial and time scales, the factors deter-mining the distribution of the dominant cyanobacteria arehighly dependent on the specific water body. Understandingthe dynamics of bloom-forming cyanobacteria and the drivingfactors of these dynamics will be helpful for developing man-agement strategies to reduce the economic and healthimpacts of blooms.

In the present study, the spatial and temporal distributionsof bloom-forming cyanobacteria in Lake Chaohu were analyzedaccording to cyanobacterial blooms distribution (satelliteimage data) and to pigment and species composition(field investigation data) conducted during 2012 and 2013.In Lake Chaohu, cyanobacterial blooms showed a decreasingtrend from west to east, with spatial and temporal shifts inspecies composition. The primary factors affecting thedistribution of cyanobacterial blooms and bloom-formingcyanobacteria were identified and analyzed using canonicalcorrespondence analysis. The results indicated that wind was

*To whom correspondence should be addressed.Email: [email protected]; Email: [email protected] editor: Mitsunobu KamiyaReceived 30 May 2015; accepted 7 October 2015.

Phycological Research 2016; 64: 44–55 doi: 10.1111/pre.12112

bs_bs_banner

© 2015 Japanese Society of Phycology

mailto:[email protected]

mailto:[email protected]

the primary factor determining the distribution of blooms,whereas temperature, phosphorus, pH and light may be cor-related with shifts in the dominant species.

METHODS

Study lakes

Lake Chaohu, the fifth largest freshwater lake in China,is located in the middle of Anhui Province, China(117°116′46″–117°151′54″E, 30°143′28″– 31°125′28″N,Fig. 1) and is an important fishery and drinking water resourcefor more than 9.66 million people in the city of Hefei(Falconer 2005; Krüger et al. 2010, 2012). The surface areaof the lake is approximately 750 km2, and the catchment areais approximately 9200 km2. Lake Chaohu has a volume of32.3 × 108 m3 in the rainy season, with a residence time ofapproximately 150 days, but has a volume of only17.2 × 108 m3 in the dry season, with a residence time ofapproximately 210 days (Tu et al. 1990). Lake depth is typi-cally less than 4 m but varies according to the hydrologicconditions (maximum depth, 6 m; mean depth, 3 m). LakeChaohu, which receives inflow and pollutants from nearbyriverine networks, including over 30 streams, canals andrivers, has been a eutrophic lake since the 1960s.Cyanobacterial blooms (primarily Microcystis and Anabaena)have dominated this lake over the past few decades (Jianget al. 2010, 2014).

Sampling and environmental analyses

To investigate the distribution of bloom-forming cyanobacteria,we performed a 2-year field investigation of 17 sites. Samplingwas performed once each month from 2012–2013. To explain

the co-existence of Microcystis and Anabaena during the sameperiod, we also performed a supplementary investigation of 43sites in June 2012. The water temperature, pH, dissolvedoxygen (DO), turbidity and conductivity were monitored at eachsample site using a multiparameter meter (model 6600;Yellow Spring Instruments, OH, USA). Integrated watersamples were collected by mixing the surface (50 cm belowthe surface), middle, and bottom (50 cm above the bottom)samples. These samples were taken to the laboratory within4 h for analysis of chlorophyll a (as total phytoplanktonbiomass), phycocyanin (as cyanobacterial biomass, there wereno data for January and February in 2012) (Izydorczyk et al.2005; Cao et al. 2006) and water chemistry, including totalphosphorus (TP), total nitrogen (TN), ammonium (NH4-N) andchemical oxygen demand (CODMn). Chlorophyll a was extractedwith 90% acetone, and the concentration was measured usinga spectrofluorophotometer at a scan speed of 60 nm min−1, aband pass of 5 nm, a response time of 2 s, and a low pulsemodulation gain. A synchronous scan with a wavelength dif-ference Δλ = 258 nm was conducted from 608 nm to 708 nmwith an excitation wavelength of 350 nm and the maximumfluorescence peak at 674 nm. Phycocyanin was extractedusing 0.05 M Tris buffer (pH 7.0), and its concentrationwas determined using a spectrofluorophotometer at anexcitation wavelength of 620 nm and an emissionwavelength of 647 nm (Abalde et al. 1998; Yan et al. 2004).TP and TN were quantified according to Valderrama (1981)and APHA (1985). Ammonium was measured using acontinuous flow analyzer (Skalar SA 1000, Breda, the Neth-erlands). CODMn was measured using the permanganatemethod in acid medium using standard procedures (APHA1985). Meteorological data, including wind speed and winddirection were obtained for 2012 and 2013 from meteorologystations #58321 and #58326 of the China MeteorologicalAdministration.

Fig. 1. Map of Lake Chaohu showing

the locations of sampling sites. The solid

circles indicate 17 routine sampling

sites, and the stars indicate 43 sup-

plementary sampling sites.


45Shifts in bloom-forming cyanobacteria

Phytoplankton analyses

Phytoplankton sampling was performed in 2012. Integrated500-mL samples were collected at each site by mixing thesurface (50 cm below the surface), middle, and bottom(50 cm above the bottom) samples taken with a Ruttnersampler. The samples were fixed with acid Lugol’s solution.Phytoplankton were identified to species based on the mostrecent literature (Hu & Wei 2006). Counts were made inrandom fields in sedimentation chambers (30 mL) using aninverted microscope following the criteria of Utermöhl (1958).The biovolume of each species at each site was calculatedfrom the measurements of 30 cells according to Hillebrandet al. (1999). The biomass was determined as the algalvolume for each site and converted to fresh weight, assuminga specific gravity of 1 g cm−3.

Acquisition of satellite images forblooms dynamics

The areas covered by cyanobacterial blooms were obtainedfrom moderate-resolution imaging spectroradiometer (MODIS)satellite images as follows: 41 scenes in 2012 and 29 scenesin 2013. The MODIS images were downloaded from the NASAEOS Data Gateway and used to derive the spatial-temporalinformation regarding blooms. The MODIS data (dimension-less) were georeferenced to UTM projections with an error ofless than 0.5 pixels. The 500 m resolution data at 1240 nmwere resampled to 250 m resolution (to match the resolutionat 645 nm). The MODIS data were corrected by removingthe molecular (Rayleigh) scattering effects and then convertedto Rayleigh-corrected reflectance (Rrc) following Hu et al.(2004). Cyanobacterial bloom coverage was calculated to thesubpixel level according to the algae pixel-growing algorithm(Zhang et al. 2014).

Data analyses

The spatial and temporal distributions of nutrients,cyanobacterial blooms, pigments and dominant cyanobacteriawere described using the R package (R-Core-Team 2014)(http://www.r-project.org) and ArcGIS Desktop 9.3 (Environ-mental Systems Research Institute, Redlands, CA, USA).Ordinations were performed using Canoco for Windows version4.5 (Ter Braak & Smilauer 2002). The relationship betweenpigments, dominant cyanobacteria and environmental vari-ables, including TN, TP, TN/TP ratio, CODMn, NH4-N, NO3-N,NO2-N, PO4-P, pH, DO, conductivity, turbidity, temperatureand photosynthetically active radiation (PAR), were assessedby Pearson correlations. All statistical analyses were consid-ered significant at P < 0.05.

Ordinations were used to identify the factors driving dif-ferences in the spatial distribution of Microcystis andAnabaena in summer. The log-transformed phytoplankton datawere analyzed using detrended correspondence analysis (DCA)by segments (Hill & Gauch 1980) to determine the gradientlengths for the first two axes. The DCA indicated that the firstgradient length (2.505) was smaller than 3 standard devia-tions and explained 27.0% of the total species variability;therefore, the use of linear ordination techniques would be

appropriate (Ter Braak 1987). Consequently, a constrainedordination method, redundancy analysis (RDA), was used torelate the phytoplankton structure to all predictor environmen-tal variables and to explore the relationships among andbetween species and the environment (Ter Braak &Verdonschot 1995). According to a preliminary RDA, we iden-tified collinear variables and selected a subset of variablesbased on inspection of the variance inflation factors (VIFs).Those variables that exhibited a VIF > 10 (i.e. multiplecollinear variables) were removed one at a time (Wilson et al.1994). After each removal, another RDA was run, and theVIFs were reexamined until no extreme values were observed.Stepwise forward selection and a Monte Carlo permutationtest were used to further reduce the environmental variables tothose that correlated significantly with the derived axes (Lepš& Šmilauer 2003) at a cutoff point of P = 0.05. Only thosetaxa that were observed in more than 3% of the samples wereincluded in the analyses of taxa abundance to minimize theinfluence of rare taxa. Taxa biomass was logarithmically trans-formed [log10 (x+1)] in all analyses to reduce the effect ofhighly variable population densities on ordination scores.Environmental data (except pH) were transformed logarithmi-cally before analysis to reduce distributional skewness.

RESULTS

Characteristics of the water environmentand meteorology

The means and standard deviations of environmental param-eters, nutrients, pigments and cyanobacterial biomassfrom monthly water samples collected at the 17 samplingsites during 2012–2013 are shown in Table 1. Thespatial and temporal distributions of TN, TP, NH4-N andCODMn are presented in Figure 2, with average concen-trations of 2.56 mg L−1 (0.56–31.69 mg L−1), 0.14 mg L−1

(0.03–1.68 mg L−1), 0.73 mg L−1 (0.02–23.55 mg L−1) and

Table 1. Environmental variables, nutrients, pigments and

cyanobacterial biomass summarised as the mean values and

standard deviation in Lake Chaohu during 2012–2013

Parameter Mean ± standard deviation

TN (mg L−1) 2.56 ± 2.49NH4-N (mg L−1) 0.73 ± 1.66NO3-N (mg L−1) 0.45 ± 0.52NO2-N (mg L−1) 0.05 ± 0.11TP (mg L−1) 0.14 ± 0.16PO4-P (mg L−1) 0.03 ± 0.10TN/TP 21.81 ± 11.17CODMn (mg L−1) 4.95 ± 1.80DO (mg L−1) 10.39 ± 3.41pH 7.52 ± 1.14Turbidity (NTU) 21.05 ± 156.07Conductivity (mS cm−1) 0.48 ± 1.91Temperature (°C) 18.68 ± 8.48PAR (μmol/(m2s)) 722.01 ± 601.43Chlorophyll a (μg L−1) 28.29 ± 39.34Phycocyanin (μg L−1) 45.11 ± 119.62Microcystis biomass (mg L−1) 9.36 ± 25.14Anabaena biomass (mg L−1) 36.97 ± 60.07


46 M. Zhang et al.

http://www.r-project.org

4.95 mg L−1 (2.45–16.99 mg L−1), respectively. These param-eters tended to decrease from west to east. Nutrients levelswere higher in the river mouths (site 14: Nanfei River mouth;site 15: Tangxi River mouth; site 16: Pai River mouth; site 17:Shiwuli River mouth) than in other regions of the lake andreached 31.69 mg L−1 TN, 23.55 mg L−1 NH4-N and1.68 mg L−1 TP in the Tangxi River mouth in 2012. Theseparameters showed obvious seasonal variations. Moreover, theseasonal variations were more intense in high nutrient regionsthan in low nutrient regions (test for homogeneity of variances,PTN, PTP and PNH4-N < 0.001, PCODMn = 0.002). Maximum valuesof TN, TP and NH4-N normally occurred in winter and spring,

whereas minimum values were normally confined to summerand autumn. However, the dynamics of CODMn were typicallythe opposite of those for the other three parameters andreached maximum values in summer and autumn, withminimum values in winter and spring.

The Pearson correlation coefficients (r) and probabilityvalues (P) for the relationships between environmental vari-ables, pigment concentrations, and the biomass of Microcystisand Anabaena are shown in Table 2. Chlorophyll a,phycocyanin and Microcystis biomass were significantly cor-related with TN, TP, the TN/TP ratio, CODMn and temperature(P < 0.05). Anabaena biomass was positively correlated with

Fig. 2. Observed monthly variations in the mean values of total nitrogen (TN), total phosphorus (TP), ammonium (NH4-N) and chemical

oxygen demand (CODMn) at seventeen sites in Lake Chaohu in 2012 (top) and 2013 (bottom). , 1; , 2; , 3; , 4; , 5; ,

6; , 7; , 8; , 9; , 10; , 11; , 12; , 13; , 14; , 15; , 16; , 17.

Table 2. Pearson correlation coefficients (r), probability values (P) and degree of freedom (df) for correlations between environmental

factors, nutrients and Chlorophyll a (Chl a), phycocyanin (PC), and biomass of Microcystis and Anabaena. Significant values (P < 0.05) are

in bold type

Chl a PC Microcystis Anabaenar P df r P df r P df r P df

TN 0.283 <0.001 395 0.331 <0.001 363 0.265 <0.001 196 0.022 0.756 196NH4-N 0.012 0.813 395 0.021 0.692 363 -0.011 0.883 196 0.014 0.845 196NO3-N -0.071 0.158 395 0.030 0.565 363 -0.064 0.371 196 -0.138 0.053 196NO2-N 0.057 0.256 395 0.251 <0.001 363 0.054 0.450 196 0.020 0.775 196TP 0.505 <0.001 395 0.472 <0.001 363 0.428 <0.001 196 0.048 0.502 196PO4-P 0.054 0.283 395 0.052 0.321 363 0.042 0.555 196 -0.042 0.557 196TN/TP -0.299 <0.001 395 -0.200 <0.001 363 -0.246 <0.001 196 -0.179 0.012 196CODMn 0.846 <0.001 395 0.701 <0.001 363 0.699 <0.001 196 0.122 0.088 196DO 0.080 0.156 310 0.082 0.172 278 0.188 0.023 145 0.169 0.040 145pH 0.119 0.027 395 0.046 0.419 363 0.070 0.400 196 -0.139 0.094 196Turbidity -0.174 0.120 395 -0.209 0.150 363 -0.086 0.742 196 0.339 0.183 196Conductivity -0.045 0.407 344 -0.026 0.650 312 -0.044 0.594 145 0.018 0.825 145Temperature 0.204 <0.001 344 0.152 0.007 312 0.293 <0.001 145 -0.053 0.524 145PAR 0.029 0.617 293 0.067 0.276 261 0.127 0.182 111 0.319 0.001 111



PAR (r = 0.319, P = 0.001) and negatively correlated with theTN/TP ratio (r = –0.179, P = 0.012).

Figure 3 shows the wind patterns in the western andeastern regions of Lake Chaohu; these patterns were producedafter constructing the climatology for the period from April–October of 2012 and 2013. In the eastern region of LakeChaohu, the prevailing winds for the 2012 and 2013 werefrom the east or east-southeast. Wind speeds of 1.2–3.6 m s−1

occurred at a frequency greater than 40 % in 2012, butoccurred at a frequency less than 30 % in 2013. Additionally,the frequency of wind speeds > 3.6 m s−1 increased remark-ably. In the western region of Lake Chaohu, the prevailingwinds during 2012 were from the east, south, or southeast.Wind speeds were typically < 3.6 m s−1 at close to 30% fre-quency for each of the prevailing wind directions. The prevail-ing winds during 2013 were primarily from the northeast orsouth-southwest. Wind speeds of 1.2–3.6 m s−1 comprisedapproximately 25% of the total wind speed; windspeeds > 3.6 m s−1, comprised approximately 10% of thetotal.

Distribution of cyanobacterial blooms

Cyanobacterial blooms occurred primarily in western LakeChaohu according to the analysis of the remote sensingimages; blooms occurred primarily in the littoral zone of thenorthwestern part of the lake. Blooms also occurred occasion-ally in eastern Lake Chaohu (Fig. 4). The coverage ofcyanobacterial blooms was significantly higher in 2012 thanthat in 2013 (Fig. 4). Moreover, the area of cyanobacterialblooms was significantly higher in 2012 than in 2013(P < 0.05, Figure 5). The average area of cyanobacterialblooms in 2012 was 104.8 km2, and the maximum area,579 km2, occurred on 8 June 2012. The average area ofcyanobacterial blooms in 2013 was 63.4 km2, and themaximum area, 175 km2, occurred on 9 October 2013.

Distribution of bloom-forming cyanobacteria

The spatial and temporal distributions of pigments are shownin Figure 6. There were no significant temporal changes in the

Fig. 3. The wind roses in the eastern and western end of Lake Chaohu in 2012 and 2013. , ≥ 7.2; , 6.0–7.2; , 4.8–6.0; , 3.6–4.8;

, 2.4–3.6; , 1.2–2.4; , 0.0–1.2.


48 M. Zhang et al.

average concentrations of chlorophyll a and phycocyaninbetween 2012 and 2013 (P > 0.05). The average concentra-tions of chlorophyll a and phycocyanin were 25.71 ±36.47 μg L−1 and 45.27 ± 119.92 μg L−1 in 2012, 30.73 ±26.30 μg L−1 and 44.98 ± 17.99 μg L−1 in 2013, respec-tively. Chlorophyll a concentrations reached their peaks in July2012 and October 2013. Phycocyanin concentrationsreached maximum values in July–August 2012, with twopeaks occurring in 2013, including one in July and one inSeptember–October 2013.

Spatially, the concentrations of chlorophyll a andphycocyanin in 2012 and 2013 showed an obvious decreas-ing trend from the western to the eastern region of LakeChaohu. The northwest coastline was the primary region ofcyanobacteria accumulation. Variations in pigment concentra-tions among sites were greater in 2012 than in 2013, with thegreatest values occurring in the river mouths of the NanfeiShiwuli and Tangxi Rivers.

Six common cyanobacterial genera were found in LakeChaohu, including Microcystis, Anabaena, Aphanizomenon,Dactylococcopsis, Pseudanabaena, and Planktothrix. Othercyanobacteria were rare, occurring at frequencies of less than10%. Among the six genera, Microcystis was the mostcommon, occurring at frequencies up to 90%, followed byAnabaena and Aphanizomenon, at frequencies of approxi-mately 80% and 60%, respectively. However, the biomass ofAnabaena was greater than that of the other species at 64% ofthe total cyanobacterial biomass. The biomasses ofMicrocystis and Aphanizomenon contributed approximately27% and 6% to the total cyanobacterial biomass, respectively.

Microcystis and Anabaena were the dominant generaof bloom-forming cyanobacteria in Lake Chaohu, withAphanizomenon occasionally being the dominant genus at afew sites. Significant seasonal and spatial shifts in the primarycyanobacterial genera occurred (Figs 7–10). Microcystis andAnabaena began to bloom in April. However, the biomass ofAnabaena was clearly higher than that of Microcystis. Thebiomass of Anabaena peaked in May and June. During sub-sequent months, the biomass of Anabaena decreased sharplyfrom July to September but peaked again in November. Thebiomass of Microcystis decreased slowly and disappearedgradually from November onward (Figs 7, 8). The two generabegan their growth in the western region of Lake Chaohu, withAnabaena as the dominant species. When Microcystis becamethe dominant genera in the region, Anabaena disappearedfrom this area and gradually came to dominate the middle andeastern regions gradually. However, when Microcystis declinedin the western region, Anabaena occupied the region again(Figs 9, 10).

Factors driving the distribution ofbloom-forming cyanobacteria

Temperature appeared to be responsible for the seasonalshift of cyanobacteria. The average temperatures at whichMicrocystis dominated (20–34°C) were significantly higherthan the temperatures at which Anabaena (13–30°C) andAphanizomenon (13–25°C) were most abundant (P < 0.05,Fig. 11). No significant differences in thermal optima wereobserved between Anabaena and Aphanizomenon (P > 0.05).

Fig. 4. The mean coverage of cyanobacterial blooms in Lake Chaohu in 2012 and 2013.

Fig. 5. Boxplot of cyanobacterial blooms areas in Lake Chaohu

in 2012 and 2013. Whiskers represent the minimum and

maximum values. Boxes symbolize the 25th and 75th percentiles.

Lines inside the boxes show the median values, and the squares

inside the boxes show the mean values.



The canonical correspondence analysis (CCA) includedphysical, chemical and biological variables. In total, 91 taxawere identified, although only nine taxa were observed in morethan 3% of the samples. The 14 environmental variables wereused together to explain the total variation within the speciesdata. A CCA with forward selection of the species data indi-cated that water temperature, TP, turbidity, pH, and nitrite/nitrate (NOx-N) were the environmental variables thataccounted for significant portions (P < 0.05) of the total vari-ance in phytoplankton composition (Fig. 12). The CCA ordi-nation showed that 48.3% of the variance in phytoplanktonassemblage was explained by axis 1 and that a further 13.2%was explained by axis 2. NOx-N was the primary explanatoryvariable in the positive direction of axis 1, and pH was theprimary explanatory variable in the negative direction of axis

1. TP, turbidity and temperature were the primary explanatoryvariables in the positive direction of axis 2. Sites in thepositive direction of axis 2 displayed Microcystis dominance,and sites in the negative direction of axis 1 displayedAnabaena dominance.

DISCUSSION

The cyanobacterial blooms primarily occupied in the westernregion of Lake Chaohu and tended to decrease from thewestern to the eastern region according to the remote sensingimages, which is consistent with the distribution of pigmentsand nutrients determined from field samples. The decreasingtrends showed interannual differences between 2012 and

Fig. 6. Observed monthly variations in

the mean values of chlorophyll a and

phycocyanin at seventeen sites in Lake

Chaohu in 2012 (left) and 2013 (right).

, 1; , 2; , 3; , 4; ,

5; , 6; , 7; , 8; , 9; ,

10; , 11; , 12; , 13; , 14;

, 15; , 16; , 17.

Fig. 7. Observed monthly variations in

Microcystis and Anabaena biomass at

seventeen sites in Lake Chaohu in 2012.

, 1; , 2; , 3; , 4; ,

5; , 6; , 7; , 8; , 9; ,

10; , 11; , 12; , 13; , 14;

, 15; , 16; , 17.


50 M. Zhang et al.

2013. In 2012, the annual average bloom coverage wassignificantly higher than in 2013 in spite of similar pigmentconcentrations. The drift of cyanobacteria was affected bywind speed and direction, which may be the primary reasonfor the interannual difference (Wu et al. 2010). In 2012, thewind speeds were typically < 3.6 m s−1 with approximatelyequal representation for each of the prevailing wind

directions. The wind speed was favorable for cyanobacteriafloating to the water surface and drifting horizontally (Wu &Kong 2009). Winds from the east, south, or southeast willdrive cyanobacteria accumulation into the western region ofLake Chaohu. In 2013, wind speeds were 1.2–3.6 m s−1

approximately 25% of the time, and speeds of > 3.6 m s−1

occurred approximately 10% of the time. The mean prevailingwinds were primarily from the northeast or south-southwest.These wind speeds and directions were adverse for the floatingand accumulating of cyanobacteria compared with 2012.South and southwest winds could prevent cyanobacterialblooms from expanding to the middle and eastern regions ofthe lake, resulting in cyanobacterial scum dispersed through-out the lake in a thin layer making detection by remotesensing difficult. These conditions may explain why theannual average bloom coverage in 2013 was lower than that in2012.

The dominance of Microcystis and Anabaena showed sig-nificant seasonal and spatial variability in Lake Chaohu. Gen-erally, Anabaena was considered dominant when nitrogen waslimited or the TN/TP ratio was low. This finding may bebecause Anabaena is able to fix N2 when nitrogen is in shortsupply (Reynolds 2006). Here, the TN/TP ratio showedobvious seasonal variations following the seasonal dynamics ofTN but no significant spatial variations. The TN/TP ratio inspring was higher than 22:1, which is the key threshold valuefor nitrogen fixation by nitrogen-fixing cyanobacteria (Smithet al. 1995). The ratios were lower than 22:1 in other

Fig. 8. Area graph of average Microcystis and Anabaena biomass

in Lake Chaohu in 2012. , Microcystis; , Anabaena.

Fig. 9. The spatial distribution of Microcystis biomass in Lake Chaohu in 2012. , 0–0.2; , 0.2–0.5; , 0.5–1; , 1–2; , 2–3; ,

3–6; , 6–12; , 12–24; , 24–48; , 48–96; , 96–180.



seasons. However, Anabaena blooms also occurred in the lakewhen the TN/TP ratio was over 22:1. A negative correlationbetween Anabaena biomass and the TN/TP ratio was observed,although the relationship was weak (r = –0.179), suggestingthat the TN/TP ratio may make limited contributions toAnabaena biomass. In addition, few heterocysts were observed

in the sample dominated by Anabaena, suggesting that nitro-gen availability may be sufficiently high, so that nitrogenfixation by heterocysts was unnecessary. Therefore, the vari-ations in the TN/TP ratio may be not the primary factor drivingthe shifts in the two genera in Lake Chaohu.

Fig. 10. The spatial distribution of Anabaena biomass in Lake Chaohu in 2012. , 0–4; , 4–8; , 8–15; , 15–30; , 30–50;

, 50–80; , 80–110; , 110–150; , 150–200; , 200–300; , 300–480.

Fig. 11. Boxplot of the temperature corresponding to the time

period at which Aphanizomenon, Anabaena or Microcystis domi-

nated. See the caption of Figure 5 for other information. Fig. 12. Redundancy analysis biplot showing different sites in

relation to environmental factors.


52 M. Zhang et al.

The seasonal shifts may be attributed to temperature adap-tations (Takano & Hino 1997). Although the populations ofthe two genera in question began to grow in April, the biomassof Anabaena reached its peak at relatively low temperatures inMay (22.7°C) and in November (9.7°C), whereas that ofMicrocystis reached its peak at a relatively high temperaturein June (27.5°C). Moreover, the biomass of Anabaena washigher than that of Microcystis, suggesting that Anabaena hada faster growth rate than Microcystis at a relatively low tem-perature. Our results were consistent with previous findingsthat Microcystis was limited by low temperature more severelythan Anabaena (Robarts & Zohary 1987; Ohkubo et al. 1993).Furthermore, Li and Watanabe (2001) found that 45 of 50strains of Anabaena had high optimal growth above 20°C andthat both Microcystis and Anabaena grew optimally at28–32°C under optimal nutrient conditions but differed intheir maximal growth rates. The maximal growth rate ofMicrocystis is generally higher than that of Anabaena(Nalewajko & Murphy 2001; Crawford 2008), which maypartially explain the high biomass of Microcystis observed insummer and the decrease in biomass of Anabaena from Julyto September. Bormans et al. (2005) found that Anabaenapopulations collapsed when the temperatures were greaterthan 26°C. However, here, Anabaena did not collapse, even inAugust when the surface water temperature was over 30°C,but maintained a high biomass. However, this peak biomasswas lost. The fact that the peak biomass of Microcystisoccurred in summer, whereas that of Anabaena occurred inspring and autumn, may also indicate that changing tempera-tures regulated the seasonal shifts.

Anabaena dominance in the western region of Lake Chaohuwas overcome by increasing Microcystis dominance insummer but was reinstated after the Microcystis biomassdecreased. The microcystin produced by Microcystis isconsidered the primary factor that controls the growth ofAnabaena (Li & Li 2012). Increasing temperatures promotedthe remarkable accumulation of Microcystis biomass and theproduction of microcystin (Yu et al. 2014), which may con-tribute to the competitive advantage Microcystis had overAnabaena and the collapse of Anabaena in the western regionof Lake Chaohu in summer.

Although Anabaena collapsed in the western region of thelake in summer, it occurred in the middle of the lake andspread toward the eastern region. Temperature changes couldexplain the shift from Anabaena to Microcystis. However, thisfactor has limitedly power to explain the difference in thespatial distribution of the two genera. These genera coexistedin the lake in summer and were distributed in different regionsof the lake, with Microcystis dominating the western region,and Anabaena dominating the middle and eastern regions ofthe lake. The ordination results showed a good positive rela-tionship between TP and Microcystis dominance. Microcystishas a low minimum phosphorus content and a large capacityto accumulate phosphorus (Kromkamp et al. 1989). More-over, it is able to survive, albeit with poor growth, inphosphorus-limited waters by processing organic phosphorus(Shi et al. 2011). Microcystis usually forms blooms whenthere is high availability of phosphorus and nitrogen(Nalewajko & Murphy 2001; Xu et al. 2010) due to the largerequirement for and rapid uptake of phosphorus (Takano &Hino 2000). In the western region of Lake Chaohu, the avail-

able phosphorus and nitrogen were higher than in otherregions and were sufficient for the development of Microcystisblooms. Although Anabaena has an efficient phosphorus-uptake system, the primary strategy for growth of this speciesin low-phosphorus environments may depend on phosphorusstorage during periods of abundant phosphorus supply(Nalewajko & Murphy 2001). In spring, Anabaena in thewestern region could store phosphorus, then migrate to themiddle and eastern regions following the current, and grow insummer by using intercellular and extracellular phosphorus.In addition, a positive relationship between pH and Anabaenadominance was observed. The results were consistent with theprevious observation that high pH favored Anabaena in labexperiments (Chang et al. 2009; Chaudhary et al. 2013).However, in the field, pH is affected by the uptake of carbondioxide during cyanobacterial photosynthesis, which increasesthe pH surrounding the cyanobacterial cell (Planavsky et al.2009). The significance of pH as a cause or consequence ofcyanobacterial blooms remains poorly described (Soares et al.2009). In Lake Chaohu, the pH in the western region wasgenerally lower than that of other regions due to the inflowfrom the rivers, and cyanobacterial blooms concentrated pri-marily in this region in summer. Thus, it appears that theeffect of pH on cyanobacterial blooms was limited in LakeChaohu. However, the high pH levels in the middle andeastern regions had some effect on the increase in Anabaenablooms.

Light availability may also be an important factor affect-ing the spatial shift of Microcystis and Anabaena prevalence.In Lake Chaohu, Secchi depths (28.54 cm) in the westernregion were lower than those in the middle (35.73 cm) andeastern (52.33 cm) regions due to suspended solids fromthe inflowing rivers, especially in summer (Yu et al. 2014).The difference in Secchi depths indicated that the watercolumn in the western region had lower light levels than thatin other regions. N2 fixation and nitrate assimilation ofdiazotrophs Anabaena is a very energy-demanding process(Stewart 1974), as indicated by the positive correlationbetween PAR and Anabaena biomass. Therefore, whenMicrocystis began to bloom in the western region, whicheffectively shaded the underlying water column, Anabaenawould have to migrate towards the middle and easternregion to maintain high photosynthetic rates. Dominance ofMicrocystis over N2-fixing genera in light limited conditionswas also observed in some other studies (Havens et al.2003; Paerl et al. 2014).

In conclusion, our study described patterns of spatial andseasonal shifts in bloom-forming cyanobacteria in LakeChaohu and identified the driving factors for these patterns.The results of this study improve our understanding ofcyanobacterial blooms distributions in large eutrophic lakes,which will be helpful for monitoring and predicting the poten-tial health risks associated with toxic cyanobacterial blooms.The study also highlights the need to treat cyanobacterialblooms following shifts in the dominant species of bloom-forming cyanobacteria.

ACKNOWLEDGMENTS

We thank Dr. Haisheng Lu for the assistance with identifyingthe phytoplankton samples and Dr. Straile for reviewing the



manuscript. This work was supported by grants from theNational Natural Science Foundation of China (31200353,31470520, 31570457), National High Technology Researchand Development Program of China (2014AA06A509) andthe Frontier Project of the Nanjing Institute of Geography andLimnology, CAS (NIGLAS2012135010).

REFERENCESAbalde, J., Betancourt, L., Torres, E., Cid, A. and Barwell, C. 1998.

Purification and characterization of phycocyanin from the marinecyanobacterium Synechococcus sp. IO9201. Plant Sci. 136: 109–20.

APHA. 1985. Standard Methods for the Examination of Water andWastewater, APHA-AWWA-WPCF, Washington, U. S. A.

Bormans, M., Ford, P. W. and Fabbro, L. 2005. Spatial and temporalvariability in cyanobacterial populations controlled by physicalprocesses. J. Plankton Res. 27: 61–70.

Cao, H. S., Kong, F. X., Luo, L. C. et al. 2006. Effects of wind andwind-induced waves on vertical phytoplankton distribution andsurface blooms of Microcystis aeruginosa in Lake Taihu. J. Fresh-water Ecol. 21: 231–8.

Chang, F., Pan, X., Kang, L., Shen, Y., Li, D. and Liu, Y. 2009. A studyon the growth and some physiological characteristics of AnabaenaSpiroides isolated from Lake Erhai. Acta Hydrobiol. Sin. 33:385–90 (in Chinese).

Chaudhary, V., Prasanna, R. and Bhatnagar, A. K. 2013. Influence ofphosphorus and pH on the fungicidal potential of Anabaenastrains. J. Basic Microbiol. 53: 201–13.

Chen, Y., Qin, B., Teubner, K. and Dokulil, M. T. 2003. Long-termdynamics of phytoplankton assemblages: Microcystis-dominationin Lake Taihu, a large shallow lake in China. J. Plankton Res. 25:445–53.

Crawford, K. A. 2008. The effects of nutrient ratios and forms on thegrowth of Microcystis aeruginosa and Anabaena flos-aquae. Gradu-ate College Dissertations and Theses. Paper 59. University ofVermont.

Deng, J., Qin, B., Paerl, H. W., Zhang, Y., Ma, J. and Chen, Y. 2014.Earlier and warmer springs increase cyanobacterial (Microcystisspp.) blooms in subtropical Lake Taihu, China. Freshwater Biol.59: 1076–85.

Falconer, I. R. 2005. Is there a human health hazard frommicrocystins in the drinking water supply? Acta Hydrochim.Hydrobiol. 33: 64–71.

Havens, K. E., James, R. T., East, T. L. and Smith, V. H. 2003. N : Pratios, light limitation, and cyanobacterial dominance in a sub-tropical lake impacted by non-point source nutrient pollution.Environ. Pollut. 122: 379–90.

Hill, M. O. and Gauch, H. G. 1980. Detrended correspondenceanalysis: an improved ordination technique. Plant Ecol. 42:47–58.

Hillebrand, H., Dürselen, C. D., Kirschtel, D., Pollingher, U. andZohary, T. 1999. Biovolume calculation for pelagic and benthicmicroalgae. J. Phycol. 35: 403–24.

Hu, C., Chen, Z., Clayton, T. D., Swarzenski, P., Brock, J. C. andMuller-Karger, F. E. 2004. Assessment of estuarine water-qualityindicators using MODIS medium-resolution bands: initial resultsfrom Tampa Bay, FL. Remote Sens. Environ. 93: 423–41.

Hu, H. and Wei, Y. 2006. The Freshwater Algae of China: Systematics,Taxonomy and Ecology. Science Press, Beijing, China.

Huisman, J., Matthijs, H. C. P. and Visser, P. M. 2005. HarmfulCyanobacteria. Springer, Dordrecht, Netherlands.

Izydorczyk, K., Tarczynska, M., Jurczak, T., Mrowczynski, J. andZalewski, M. 2005. Measurement of phycocyanin fluorescenceasan online early warning system for cyanobacteria in reservoir intakewater. Environ. Toxicol. 20: 425–30.

Jiang, X., Wang, S., Zhong, L., Jin, X. and Sun, S. 2010. Seasonalvariation characteristics of algae biomass in Chaohu Lake. Environ.Sci. 31: 2056–62 (in Chinese).

Jiang, Y.-J., He, W., Liu, W.-X. et al. 2014. The seasonal and spatialvariations of phytoplankton community and their correlation withenvironmental factors in a large eutrophic Chinese lake (LakeChaohu). Ecol. Indic. 40: 58–67.

Kanoshina, I., Lips, U. and Leppänen, J.-M. 2003. The influence ofweather conditions (temperature and wind) on cyanobacterialbloom development in the Gulf of Finland (Baltic Sea). HarmfulAlgae 2: 29–41.

Kromkamp, J., Van Den Heuvel, A. and Mur, L. R. 1989. Phosphorusuptake and photosynthesis by phosphate-limited cultures of thecyanobacterium Microcystis aeruginosa. Br. Phycol. J. 24: 347–55.

Krüger, T., Hölzel, N. and Luckas, B. 2012. Influence of cultivationparameters on growth and microcystin production of Microcystisaeruginosa (Cyanophyceae) isolated from Lake Chao (China).Microb. Ecol. 63: 199–209.

Krüger, T., Wiegand, C., Kun, L., Luckas, B. and Pflugmacher, S.2010. More and more toxins around–analysis of cyanobacterialstrains isolated from Lake Chao (Anhui Province, China). Toxicon56: 1520–4.

Lepš, J. and Šmilauer, P. 2003. Multivariate Analysis of EcologicalData using CANOCO. Cambridge University Press, Cambridge,U. K.

Li, R. and Watanabe, M. M. 2001. Physiological properties of plankticspecies of Anabaena (Cyanobacteria) and their taxonomic value atspecies level. Arch Hydrobiol Suppl Algol. Stud. 140: 31–45.

Li, Y. and Li, D. 2012. Competition between toxic Microcystisaeruginosa and nontoxic Microcystis wesenbergii with AnabaenaPCC7120. J. Appl. Phycol. 24: 69–78.

Michalak, A. M., Anderson, E. J., Beletsky, D. et al. 2013. Record-setting algal bloom in Lake Erie caused by agricultural andmeteorological trends consistent with expected future conditions.Proc. Natl. Acad. Sci. USA 110: 6448–52.

Nalewajko, C. and Murphy, T. P. 2001. Effects of temperature, andavailability of nitrogen and phosphorus on the abundance ofAnabaena and Microcystis in Lake Biwa, Japan: an experimentalapproach. Limnology 2: 45–8.

Ohkubo, N., Yagi, O. and Okada, M. 1993. Studies on the successionof blue-green algae, Microcystis, Anabaena, Oscillatoria andPhormidium in Lake Kasumigaura. Environ. Technol. 14: 433–42.

Oliver, R. and Ganf, G. 2002. Freshwater blooms. In Whitton, B. andPotts, M. (Eds) The Ecology of Cyanobacteria. Springer, Nether-lands, pp. 149–94.

Paerl, H. and Otten, T. 2013. Harmful cyanobacterial blooms: causes,consequences, and controls. Microb. Ecol. 65: 995–1010.

Paerl, H. W. and Fulton, R. S. 2006. Ecology of harmfulcyanobacteria. In Granéli, E. and Turner, J. T. (Eds) Ecology ofHarmful Algae. Springer, Berlin Heidelberg, pp. 95–109.

Paerl, H. W., Xu, H., Hall, N. S. et al. 2014. Controlling cyanobacterialblooms in Hypertrophic Lake Taihu, China: will nitrogen reductionscause replacement of Non-N2 Fixing by N2 Fixing Taxa? PLoS ONE9: e113123.

Planavsky, N., Reid, R. P., Lyons, T. W., Myshrall, K. L. and Visscher,P. T. 2009. Formation and diagenesis of modern marine calcifiedcyanobacteria. Geobiology 7: 566–76.

Posch, T., Koster, O., Salcher, M. M. and Pernthaler, J. 2012. Harmfulfilamentous cyanobacteria favoured by reduced water turnover withlake warming. Nat. Clim. Chang. 2: 809–13.

R-Core-Team. 2014. R: A Language and Environment for StatisticalComputing. R foundation for statistical computing, Vienna,Austria.

Reynolds, C. S. 2006. Ecology of Phytoplankton (Ecology, Biodiversityand Conservation). Cambridge University Press, Cambridge, UK.


54 M. Zhang et al.

Robarts, R. D. and Zohary, T. 1987. Temperature effects on photo-synthetic capacity, respiration, and growth rates of bloom-formingcyanobacteria. New Zeal J. Mar. Fresh. 21: 391–9.

Shi, X. L., Qian, S. Q., Kong, F. X., Zhang, M. and Yu, Y. 2011.Differences in growth and alkaline phosphatase activity betweenMicrocystis aeruginosa and Chlorella pyrenoidosa in response tomedia with different organic phosphorus. J. Limnol. 70: 21–5.

Smith, V. H., Bierman, V. Jr., Jones, B. and Havens, K. 1995. His-torical trends in the Lake Okeechobee ecosystem. 4. Nitrogen:phosphorus ratios, cyanobacterial dominance, and nitrogen fixa-tion potential. Arch Hydrobiol Suppl Algol. Stud. 107: 71–88.

Soares, M. C. S., De a Rocha, M., Marinho, M. M., Azevedo, S.,Branco, C. W. and Huszar, V. L. 2009. Changes in species com-position during annual cyanobacterial dominance in a tropicalreservoir: physical factors, nutrients and grazing effects. Aquat.Microb. Ecol. 57: 137–149.

Stewart, W. D. P. 1974. Algal Physiology and Biochemistry. Univ ofCalifornia Press, Los Angeles.

Takano, K. and Hino, S. 1997. Effect of temperature on the succes-sion of planktonic algae in hypertrophic Lake Barato, Hokkaido.Jpn. J. Phycol. 45: 89–93.

Takano, K. and Hino, S. 2000. Effect of temperature and solublereactive phosphorus on abundance of Aphanizomenon flos-aquae(Cyanophyceae). Phycol. Res. 48: 9–13.

Ter Braak, C. J. F. 1987. The analysis of vegetation-environmentrelationships by canonical correspondence analysis. Plant Ecol.69: 69–77.

Ter Braak, C. J. F. and Smilauer, P. 2002. Canoco for Windows Version4.5, Centre for Biometry Wageningen, Wageningen, Netherlands.

Ter Braak, C. J. F. and Verdonschot, P. F. M. 1995. Canonical corre-spondence analysis and related multivariate methods in aquaticecology. Aquat. Sci. 57: 255–89.

Tu, Q. Y., Gu, D. X., Yi, C. Q., Xu, Z. R. and Han, G. Z. 1990. TheResearches on the Lake Chaohu Eutrophication. Publisher of Uni-versity of Science and Technology of China, Hefei.

Utermöhl, H. 1958. Zur Vervollkommnung der quantitativen Erfassungdes Planktons (mit besonderer Berücksichtigung desUltraplanktons). Mitt. Int. Ver. Theor. Angew. Limnol. 9: 1–38.

Valderrama, J. 1981. The simultaneous analysis of total nitrogenand total phosphorus in natural waters. Mar. Chem. 10: 109–22.

Wilson, S. E., Cumming, B. F. and Smol, J. P. 1994. Diatom-salinityrelationships in 111 lakes from the Interior Plateau of BritishColumbia, Canada: the development of diatom-based models forpaleosalinity reconstructions. J. Paleolimnol. 12: 197–221.

Wu, X. D. and Kong, F. X. 2009. Effects of light and wind speed on thevertical distribution of Microcystis aeruginosa colonies ofdifferent sizes during a summer bloom. Int. Rev. Hydrobiol. 94:258–66.

Wu, X. D., Kong, F. X., Chen, Y. W. et al. 2010. Horizontal distributionand transport processes of bloom-forming Microcystis in a largeshallow lake (Taihu, China). Limnologica 40: 8–15.

Xu, H., Paerl, H. W., Qin, B., Zhu, G. and Gao, G. 2010. Nitrogen andphosphorus inputs control phytoplankton growth in eutrophic LakeTaihu, China. Limnol. Oceanogr. 55: 420–32.

Yan, R., Kong, F. and Han, X. 2004. Analysis of the recruitment of thewinter survival algae on the sediments of Lake Taihu byfluorometry. J. Lake Sci. 16: 163–8 (in Chinese).

Yu, L., Kong, F., Zhang, M., Yang, Z., Shi, X. and Du, M. 2014. Thedynamics of Microcystis genotypes and microcystin production andassociations with environmental factors during blooms in LakeChaohu, China. Toxins 6: 3238–57.

Zhang, M., Yu, Y., Yang, Z., Shi, X. and Kong, F. 2012. The distributionof phytoplankton along trophic gradients and its mediation byavailable light in the pelagic zone of large eutrophic lakes. Can. J.Fish. Aquat. Sci. 69: 1935–46.

Zhang, Y., Ma, R., Duan, H., Loiselle, S., Xu, J. and Ma, M. 2014. Anovel algorithm to estimate algal bloom coverage to subpixel reso-lution in Lake Taihu. IEEE J. Sel. Top. Appl. Earth Obs. RemoteSens. 7: 3060–8.



Documents

Spatial and seasonal shifts in bloomforming cyanobacteria