Environment International 62 (2014) 113–118

Contents lists available at ScienceDirect

Environment International

j ourna l homepage: www.e lsev ie r .com/ locate /env int

Learning from 8years of regional cyanobacteria observation in Brittany inview of sanitary survey improvement

Frédéric Pitois a, Olivier Thomas b,c, Isabelle Thoraval a, Estelle Baurès b,c,⁎a Limnology, 16 rue Paul Langevin, 35200 Rennes, Franceb EHESP Rennes, Sorbonne Paris Cité, Avenue du Professeur Léon Bernard, CS 74312, 35043 Rennes Cedex, Francec Inserm, U 1085-IRSET, LERES, France

⁎ Corresponding author at: EHESP Rennes, Sorbonne PLéon Bernard, CS 74312, 35043 Rennes Cedex, France. Tel

E-mail address: estelle.baures@ehesp.fr (E. Baurès).

0160-4120/$ – see front matter © 2013 Elsevier Ltd. All rihttp://dx.doi.org/10.1016/j.envint.2013.09.018

a b s t r a c t

a r t i c l e i n f o

Article history:Received 15 July 2013Accepted 30 September 2013Available online 7 November 2013

Keywords:CyanobacteriaMonitoringWater qualityMicrocystinsRegional scale

Cyanobacteria and cyanotoxins have been more and more studied during the last decades with regard to envi-ronment and health issues. More recently the consequences of climate change reinforced the need for researchin viewof a bettermanagement of cyanobacteria blooms. In this context the exploitation of thewater quality sur-vey of 26 recreational lakes in Brittany (north western France) between 2004 and 2011 is reported in this paperin order to encompass spatial and interannual patterns of cyanobacteria development at a regional scale. Startingfromweekly data principally acquired during summertime, the links between cyanobacteria cell densities, toxinoccurrences and interannualmeteorological factors can give insights on the potential evolution of cyanobacterialcrisis in the future. This study is part of a project aiming at a better understanding of potentially toxiccyanobacteria crisis occurrences in recreational waters, in order to improve predictive monitoring routines.

© 2013 Elsevier Ltd. All rights reserved.

1. Introduction

Since their introduction in the public health debate (Chorus andBartram, 1999; Falconer, 1996), cyanobacteria and their toxins havegradually become a major concern for public health authorities. Severalliterature reviews focusing on toxin production were regularly pub-lished for the last 10years, showing that biosynthetic pathways and tox-icological issues are relatively well known (Araoz et al., 2010; Leflaiveand Ten-Hage, 2007; Pearson et al., 2010; Smith et al., 2008). Amongscientific questions remaining unsolved, there is no evidence of anydirect relation between cyanobacterial biomass and toxin concentra-tion, nor any consensus on toxin ecological role in the cyanobacterialcell life cycle. This lack of knowledge of toxin causality and of under-standing of toxin purpose prevents any attempt in the predictionof algae bloom risk occurrence. The relations between cyanobacterialbiomass and trophic enrichment (eutrophication processes) are widelyadmitted,whereas the influence ofmeteorological parameters on in situalgal growth is less discussed. Some abiotic parameters have never-theless been studied in vitro in terms of selection and growth rates ofpotentially toxic species (Baker and Bellifemine, 2000; De Nobel et al.,1998; Hadas et al., 2002; Imai et al., 2009; McCausland et al., 2005;Yamamoto and Nakahara, 2005) and through their influence on toxinbiosynthesis (Briand et al., 2008; Preußel et al., 2009; Rapala andSivonen, 1998; Tonk et al., 2005; Wiedner et al., 2003). Reports ofin vitro influence of light and temperature on cyanobacterial growth

aris Cité, Avenue du Professeur.: +33 2 99 02 29 19.

ghts reserved.

and toxin production emphasize the possible consequences of climatechange on public health concerns regarding bathing andwater resourcesfor drinking water production. Exploring the relations between in situabiotic factors, climate change consequences, cyanobacteria distributionand public health parameters at a regional scale is the main objective ofthis research, based on the analysis of monitoring data collected by localhealth authorities since 2004 in Brittany (France). This study intends toenlighten some aspects of the interannual dynamics of cyanobacterialdevelopment in 26 lakes used as recreative waters, and is directedtoward the development of risk indices predictive of cyanobacteria-associated crisis.

2. Material and method

Our study is based on public health weekly survey data from 26recreational lakes monitored every year from 2004 to 2011 in Brittany(north-western France, Fig. 1).

Table 1 gives some characteristics of these 26 sites.Cyanobacteria data, i.e. cell densities, species composition and

microcystin concentrations, were collected from the regional publichealth authorities (French Agence Régionale de Santé (ARS) of Brittany).Hydrology and lakemorphology parameters (i.e. depth, lake volume,wa-tershed area…) were gathered from dedicated institutional databanksheld by the Ministry for the Environment (BDHydro) and Agencede l'Eau Loire-Bretagne (OsurWeb), or directly from lake owners.

In addition, monthly meteorological data (precipitation rates,temperature, sunshine…) were collected from Meteo France data base(Climathèque) between 2001 and 2011.

0 100 km

Fig. 1. Spatial distribution of the 26 investigated sites in Brittany.

114 F. Pitois et al. / Environment International 62 (2014) 113–118

For this study, correlations between parameters were calculatedusing the Pearson test.

3. Results

3.1. Meteorological context

In the context of the last ten years, the oceanic climate appearedunstable during summer in Brittany. From 2001 to 2011, in spite ofhigh interannual variability, available data (Fig. 2) show a common ten-dency to decrease for mean temperatures (with a loss of nearly 1 °C),sunshine hours (−7%) and global radiations (−4%), whereas cumulatedpluviometry increased by +5%.

3.2. Cyanobacteria cell densities

First it can be noted that from 2004 to 2011 (i.e. the monitoringtimespan), weekly maximal cyanobacteria cell densities tended todecrease (Fig. 3a), although the opposite could be expected in a contextof global warming. In the same time however, alert levels exceedingfrequencies (Fig. 3b) increased from 31 to 55% for WHO alert level 2(20000cell/ml) and 13 to 27% for WHO alert level 3 (100000cell/ml).

This evolution seems to follow a gradual shift of frequency distribu-tion (Fig. 4) characterized on one hand, by a slow increase of backgroundnoise cell densities, as the 10th percentile (Q10) increases by a factor 2.5,and, on the other hand, by a faster increase of median cell densities,as the median (Q50) increases by 5.5. In the same time, maximal celldensities (90th percentile: Q90) only increased by 1.5.

For the same period, (from 2004 to 2011, June to September) WHOlevel 2 exceeding frequency followed a continuous +2 to +4 week/year increase (Fig. 5a). In the same time, WHO level 3 exceedingfrequency tended to increase by +1 to +2 week/year from June toJuly, moving to +3week/year in August and September. Cell densitiesappeared to follow the same evolution as WHO level frequencies

Table 1Characteristics of the 26 investigated sites.

Minimum Maximum Median

Watershed area (km2) 0.2 676.0 19.6Lake volume (m3) 20000 51000000 815700Maximal depth 2.0 45.0 4.1

(Fig. 5b), with at first a slow increase rate (+2000 cell·ml−1·yr−1)in June and July, followed by a fast increase rate (+4000 to+8000 cell·ml−1·yr−1) in August and September. This fast increasecan explain why cell densities exceeding theWHO level 3 are observedearlier with time.

3.3. Microcystins

In spite of the high observation frequency of cell densities higherthan the level 3 WHO thresholds, only 30% of the analyzed sampleshosted quantifiable microcystin concentrations, MCs (Fig. 6a). Itshould be noted that detection frequency was inversely related tocell density, with a maximum of 33% for cell densities rangingfrom 10 to 20000 cell/ml, compared to 15% for cell densities higherthan 1000000 cell/ml (Fig. 6b).

MC detection frequencies could not be correlated to lakemorphology,neither to watershed characteristics (data not shown). Detection fre-quencies, on the other hand, appeared to be correlated to summermeteorological parameters such as mean temperatures (r = 0.52) orcumulated sunshine hours and global radiation (r= 0.68 and 0.69 re-spectively, pb0.05). Surprisingly, frequencies could be negatively corre-lated to sunshine and global radiation of the previouswinter (r=−0.54to−0.61), but with a lower significance (pb0.1). This could mean thatMCs are more likely to be detected at the regional scale during yearscombining low winter sunshine and a high summer sunshine (it mustbe noted that winter and summer sunshine are not correlated).

3.4. Regional patterns

For each site, annual maximal cell densities and exposure duration(expressed as week number with cell density N 100000 cell/ml) werepositively linked (Fig. 7).

The cell density-exposure duration slope varies on an annual basis,and ranges from 98000 to 280000 cell/ml per exceeding week. Theannual slope values appear correlated with mean annual summer tem-perature (r=0.86),mean cumulated summer sunshine hours (r=0.71)or cumulated global radiation (r=0.69). Surprisingly the slope is alsonegatively correlated with previous winter conditions such as sunshinehours (r = −0.68), global radiation (r = −0.58) and temperature(r=−0.55). The winter sunshine/radiation parameters are not corre-lated with the next summer sunshine/radiation parameters. At theregional scale, these parameters combination can be translated intoyears with higher cell densities (weak winter light and high summer

mea

n te

mpe

ratu

re 0

6-09

(°C

)

12

14

16

18

20

22

24

pluv

iom

etry

06-

09 (

mm

)

50

100

150

200

250

300

350

400

450su

nshi

ne 0

6-09

(ho

urs)

35000

40000

45000

50000

55000

60000

65000

2000 2002 2004 2006 2008 2010 2012

2000 2002 2004 2006 2008 2010 20122000 2002 2004 2006 2008 2010 2012

2000 2002 2004 2006 2008 2010 2012ra

diat

ion

(W/m

²)1.75e+5

2.00e+5

2.25e+5

2.50e+5

Fig. 2.Meteorological data in Brittany, from2001 to 2011, June to September:mean temperatures (n=33 sites), cumulated pluviometry (n=34 sites), cumulated sunshine hours (n=18sites) and cumulated global radiation (n=21 sites).

115F. Pitois et al. / Environment International 62 (2014) 113–118

light) or longer exposure durations (highwinter light andweak summerlight).

Whatever the meteorological variations, an interannual recurrencerelation can be highlighted. For each site, the annual exposure duration(i.e. number ofweeksN level 3WHO threshold) increaseswith the inter-annual frequency of cell density aboveWHO level 3 observations, what-ever the duration of these exceedance episodes (Fig. 8). This increase isobserved for all sites where exceedance episodes were recorded at least5 years out of 8, corresponding to 50% of the survey period. The meanannual exposure duration increases from 2+/−1.6weeks if recurrenceis less than 50% to a maximum of 7.5+/− 4.4weeks for the sites withrecurrence higher than 80%. By definition, a higher recurrence is inde-pendent from meteorological interannual variations. Recurrence indexthen appears correlated with lake morphological parameters such asmaximal depth (r = −0.62) and watershed area (r =−0.56). Theserelations are statistically significant (p b 0.01) but correlation coeffi-cients are relatively low, indicating that supplementary parameters,still not identified, should be considered. The addition of both negativerelations with depth and watershed area points out that shallower

cyan

obac

teria

cel

l den

sity

(ce

ll.m

l-1)

0

2000000

4000000

6000000

8000000

2004 2005 2006 2007 2008 2009 2010 2011

Fig. 3. (a, left) Weekly maximal cell density evolution, from 2004 to 2011, June to SeptemberJune to September.

lakes with relatively smaller watersheds are more likely to experienceinterannually frequent, longer-lasting blooms. In this case, highercyanobacteria cell densities (as duration and density parametersare correlated) will be recorded whereas MC detection should be lessfrequent.

4. Discussion

Many studies hypothesized that in the current climate changecontext, cyanobacteria should be more competitive than other species,leading to better dominance opportunities (Carey et al., 2012; Kostenet al., 2012; O'Neil et al., 2012; Paerl and Otten, 2013). The effect ofclimatic parameters however can be difficult to generalize, especiallyin lakes covering different ranges of depth, volume and nutrient loads,but it is likely that shallow unstratified lakes (the most common laketype in Brittany) should be the most responsive to global warmingand the most affected by cyanobacteria extension (Malmaeus et al.,2006; Mooij et al., 2007, 2009). It must be noted that since the secondhalf of 20th century, western France is experiencing a slight rise of

2002 2004 2006 2008 2010 2011

% s

ampl

es e

xcee

ding

WH

O a

lert

leve

l

0

20

40

60

80> WHO lev. 2

> WHO lev. 3

(1 dot = 1week), (b, right) annual WHO level exceeding frequency, from 2004 to 2011,

cyan

obac

teria

cel

l den

sity

(10

% q

uant

ile, c

ell.m

l-1)

0

2000

4000

6000

8000

cyan

obac

teria

cel

l den

sity

(50

% q

uant

ile, c

ell.m

l-1)

0

20000

40000

60000

80000

100000

120000

2004 2005 2006 2007 2008 2009 2010 2011 2004 2005 2006 2007 2008 2009 2010 2011 2004 2005 2006 2007 2008 2009 2010 2011

cyan

obac

teria

cel

l den

sity

(90

% q

uant

ile, c

ell.m

l-1)

0

200000

400000

600000

800000

1000000

1200000

Fig. 4. Cyanobacteria cell density percentile (Q10, Q50 and Q90) evolution from 2004 to 2011, June to September.

116F.Pitois

etal./EnvironmentInternational62

(2014)113

–118

< MQL

MC

s oc

cure

nces

(%

ana

lysi

s)

0

20

40

60

80

cell density (cell.ml-1)

< 1µg/l < 10µg/l > 10µg/l 0 1x106 2x106 3x106 4x106 5x106 6x106

p

erce

nt p

ositi

ve

dete

ctio

n (>

0.2

µg.l-

1)

5

10

15

20

25

30

35

Fig. 6. (a, left) Microcystin concentration (MC) frequency distribution, (b, right) detection frequency vs. cell density. From 2004 to 2011, June to September.

Annual monitoring duration (weeks)

Med

ian

cell

dens

ity e

volu

tion

rat

e (c

ell.m

l-1.y

ear-1

)

0

2000

4000

6000

8000

10000

Annual monitoring duration (weeks)0 2 4 6 8 10 12 14 16 180 2 4 6 8 10 12 14 16 18

WH

O le

vel e

xced

ance

rate

(w

eeks

.yea

r-1)

0

1

2

3

4

5

WHO level 2WHO level 3

Fig. 5. Cyanobacteria density evolution from 2004 to 2011, June to September, for weeks 1 to 16 of the monitoring course: (a, left) WHO levels 2 and 3 frequency increase rate, (b, right)Median cell density increase rate.

)

117F. Pitois et al. / Environment International 62 (2014) 113–118

annual mean temperatures and precipitations, with projected summertemperature increase and summer precipitation decrease (Terray andBoé, 2013). In the case of our study however, at the regional scaleof Brittany, climate parameters follow a tendency to a slight decreaseof summer temperature and sunshine conjugated with an increase ofprecipitations.

Long term studies of field data are still scarce in the literature. They alltendhowever to point out the strong influence of summermeteorologicalparameters on cyanobacterial populations, through light availability(Noges et al., 2003), water temperatures (Liu et al., 2011; Wiedneret al., 2007), or a combination of water temperature and sunshine hours(Zhang et al., 2012). While species-specific data cannot be easily extrap-olated to population level, this could be expected in view of the manystudies exploring temperature and light influence on cyanobacterialgrowth rate (Baker and Bellifemine, 2000; De Nobel et al., 1998; Hadaset al., 2002; Imai et al., 2009; McCausland et al., 2005; Yamamoto andNakahara, 2005).

Annual exposure duration (weeks, cumulated)

0 3 6 9 12 15 18

Cya

noba

cter

ia m

ax. c

ell

d

ensi

ty (

cell.

ml-1

)

100 000

1 000 000

10 000 000

Fig. 7.Relation between annualmaximal cyanobacteria cell density and exposure duration(number of weeks N level 3 WHO threshold, 1 dot = 1 lake), from 2004 to 2011, June toSeptember.

From 2004 to 2011, meteorology in Brittany followed an evolutionopposite to predictions, but the analysis of survey data shows neverthe-less a strong tendency to a continuous increase of cyanobacterialpresence at the regional scale. This can be related to an increase ofbackground noise more than maximal biomasses, leading to a longerlasting presence of relatively lower cell densities, rather than a higherfrequency of ‘classical’ bloom episodes with high turbidity and floatingcyanobacterial scum.

Our results can be compared to the conclusions of former studiescarried out byWiedner et al. (2007) and Rücker et al. (2009) emphasiz-ing the influence of increasing winter temperatures on earlier akinetegermination, this earlier germination leading to a faster growth ofCylindrospermopsis raciborskii populations in spring. This principle of alarger inoculum influence on growth parameter seems to be relevantat the population level and regional scale, and can be illustrated by

Years with observed WHO lev.3 occurence

2 4 6 8 10Cel

l den

sity

> W

HO

lev.

3 (

wee

ks/y

ear

0

2

4

6

8

10

12

14

Fig. 8. Relation between years with level 3 WHO threshold observation and annualexposure duration, from 2004 to 2011, June to September.

Table 2Meteorological influence on cyanobacteria parameters.

Winter Temperature + Exposure duration favored vs. cell densityTemperature − Cell density favored vs. exposure durationSunshine + Exposure duration favored vs. cell densitySunshine − Cell density favored vs. exposure duration

Summer Temperature + Cell density favored vs. exposure durationTemperature − Exposure duration favored vs. cell densitySunshine + Cell density favored vs. exposure durationSunshine − Exposure duration favored vs. cell density

118 F. Pitois et al. / Environment International 62 (2014) 113–118

the correlations between winter meteorology (namely temperaturesand sunshine hours) and cell density vs. exposure duration slopesummarized in Table 2. At regional scale, higher winter temperaturescan lead to a larger spring inoculum and high median cell densitieseven if summer temperature and sunshine are too low to favorhigher biomasses.

5. Conclusion

Thisfirst study of eight years (from 2004 to 2011) of sanitary survey,annually during 16 weeks, shows that there is a strong homogeneity,at the regional scale, between parameters related to cyanobacteria celldensities and toxin detection frequency:

− The exposition pathways to cyanobacteria (cell density, WHO thresh-olds exceeding duration) and the exposition frequency (interannualrecurrence of WHO thresholds exceeding episodes) are correlatedand sites with higher cell density will tend to suffer longer prolifera-tion episodes with a higher interannual frequency.

− Microcystin (MC) detection is inversely correlated with exposuredurations and cell densities: MC detection frequency is higher whencell densities are lower, i.e. namely in the lakes where proliferationepisodes are less frequent. As toxin monitoring was carried out withdifferent analyticalmethods, it is impossible to correlateMC detectionfrequencies and concentrations.

− The factors controlling user exposure pathways to cyanobacterialbiomass and toxins appear to derive in the same time from lakemorphology (interannual recurrence) and regional-scalemeteorology(exposition duration, cell density, MC detection frequency). Howeverthe lack of nutrient data for most lakes prevented us from assessingthe trophic status influence on cyanobacterial parameters.

− The influence of lakemorphology (volume, depth,watershed surfaceare interannual constants) and of winter meteorological parameters(sunshine, radiation) allows to consider the development of a simplemodel for annual risk characterization. In this case, potential riskwould be evaluated before the beginning of summer sanitarysurveys, indicating which exposure pathway should be emphasized(biomass and/or MCs).

− MC detection frequencies are higher when cell densities are lower.This indicates that MC monitoring should be extended to lakeswhere cyanobacteria are less visible, while toxin analysis is currentlyfocused on higher cell density episodes.

Acknowledgments

This work is part of the ORACLE project funded by the FrenchNational Agency of Sanitary Security (ANSES) under the nationalprogram of Environment-Health-Work research (PNRSET 2012). Theauthors sincerely thank the French Agence Régionale de Santé (ARS)of Brittany and the French Agence de l'Eau Loire-Bretagne (AELB) forthe data provided and are grateful to Caroline Farenc, student at theSchool of Environmental Engineering (EME), for the data exploitation.

References

Aráoz R, Molgó J, Tandeau de Marsac N. Neurotoxic cyanobacterial toxins. Toxicon2010;56(5):813–28.

Baker PD, Bellifemine D. Environmental influences on akinete germination ofAnabaena circinalis and implications for management of cyanobacterial blooms.Hydrobiologia 2000;427(1):65–73.

Briand E, Yéprémian C, Humbert J-F, Quiblier C. Competition between microcystin-and non-microcystin-producing Planktothrix agardhii (cyanobacteria) strainsunder different environmental conditions. Environ Microbiol 2008;10(12):3337–48.

Carey CC, Ibelings BW, Hoffmann EP, Hamilton DP, Brookes JD. Eco-physiologicaladaptations that favour freshwater cyanobacteria in a changing climate. Water Res2012;46(5):1394–407.

Chorus I, Bartram J. Toxic cyanobacteria in water: a guide to their public healthconsequences, monitoring and management. London & New York: Spon Press; 1999.

De Nobel WT, Matthijs HCP, Von Elert E, Mur LR. Comparison of the light-limited growthof the nitrogen-fixing cyanobacteria Anabaena and Aphanizomenon. New Phytol1998;138:579–87.

Falconer IR. Potential impact on human health of toxic cyanobacteria. Phycologia1996;35(6S):6–11.

Hadas O, Pinkas R, Malinsky-Rushansky N, Shalev-Alon G, Delphine E, Berner T, SukenikA, Kaplan A. Physiological variables determined under laboratory conditions mayexplain the bloom of Aphanizomenon ovalisporum in Lake Kinneret. Eur J Phycol2002;37:259–67.

Imai H, Chang K-H, Kusaba M, Nakano S-I. Temperature-dependent dominanceof Microcystis (Cyanophyceae) species: M. aeruginosa and M. weisenbergii. J PlanktonRes 2009;31:171–8.

Kosten S, Huszar VL, Becares E, Costa LS, Donk E, Hansson LA, Scheffer M. Warmerclimates boost cyanobacterial dominance in shallow lakes. Glob Chang Biol2012;18(1):118–26.

Leflaive J, Ten‐Hage, L. Algal and cyanobacterial secondary metabolites in freshwaters:a comparison of allelopathic compounds and toxins. Freshw Biol 2007;52(2):199–214.

Liu X, Lu X, Chen Y. The effects of temperature and nutrient ratios on Microcystis bloomsin Lake Taihu, China: an 11-year investigation. Harmful Algae 2011;10:337–43.

Malmaeus JM, Blenckner T, Markensten H, Persson I. Lake phosphorus dynamics andclimate warming: a mechanistic model approach. Ecol Model 2006;190(1):1–14.

McCausland M, Thompson P, Blackburn SI. Ecophysiological influence of lightand mixing on Anabaena circinalis (Nostocales, Cyanobacteria). Eur J Phycol2005;40(1):9–20.

Mooij WM, Janse JH, De Senerponts Domis LN, Hülsmann S, Ibelings BW. Predictingthe effect of climate change on temperate shallow lakes with the ecosystem modelPCLake. Hydrobiologia 2007;584:443–54.

Mooij WM, De Senerpont Domis LN, Janse JH. Linking species- and ecosystem-levelimpacts of climate change in lakes with a complex and a minimal model. Ecol Model2009;220(21):3011–20.

Nõges P, Tuvikene L, Feldmann T, Tõnno I, Künnap H, Luup H, Nõges T. The role ofcharophytes in increasing water transparency: a case study of two shallow lakes inEstonia. Hydrobiologia 2003;506(1–3):567–73.

O'Neil JM, Davis TW, BurfordMA, Gobler CJ. The rise of harmful cyanobacteria blooms: thepotential roles of eutrophication and climate change. Harmful Algae 2012;14:313–34.

Paerl HW, Otten TG. Harmful cyanobacterial blooms: causes, consequences, and controls.Microb Ecol 2013;65(4):1–16.

Pearson L, Mihali T, Moffitt M, Kellmann R, Neilan B. On the chemistry, toxicologyand genetics of the cyanobacterial toxins, microcystin, nodularin, saxitoxinand cylindrospermopsin. Mar Drugs 2010;8(5):1650–80.

Preußel K, Wessel G, Fastner J, Chorus I. Response of cylindrospermopsin production andrelease in Aphanizomenon flos-aquae (cyanobacteria) to varying light and tempera-ture conditions. Harmful Algae 2009;8(5):645–50.

Rapala J, Sivonen K. Assessment of environmental conditions that favor hepatotoxicand neurotoxic Anabaena spp. strains cultured under light limitation at differenttemperatures. Microb Ecol 1998;36(2):181–92.

Rücker J, Tingwey EI, Wiedner C, Anu CM, Nixdorf B. Impact of the inoculum sizeon the population of Nostocales cyanobacteria in a temperate lake. J Plankton Res2009;31(10):1151–9.

Smith JL, Boyer GL, Zimba PV. A review of Cyanobacterial odorous and bioactivemetabolites: impacts and management alternatives in aquaculture. Aquaculture2008;280:5–20.

Terray L, Boé J. Quantifying 21st-century France climate change and related uncertainties.Geoscience 2013;345(3):136–49.

Tonk L, Visser PM, Christiansen G, Dittmann E, Snelder EO, Wiedner C, Huisman J.The microcystin composition of the cyanobacterium Planktothrix agardhii changestoward a more toxic variant with increasing light intensity. Appl Environ Microbiol2005;71(9):5177–81.

Wiedner C, Visser PM, Fastner J, Metcalf JS, Codd GA, Mur LR. Effects of light on themicrocystin content of Microcystis strain PCC 7806. Appl Environ Microbiol2003;69(3):1475–81.

Wiedner C, Rücker J, Brüggemann R, Nixdorf B. Climate change affects timing and sizeof populations of an invasive cyanobacterium in temperate regions. Oecologia2007;152(3):473–84.

Yamamoto Y, Nakahara H. The formation and degradation of cyanobacteriumAphanizomenon flos-aquae blooms: the importance of pH, water temperature, andday length. Limnol. 2005;6(1):1–6.

Zhang M, Duan H, Shi X, Yu Y, Kong F. Contributions of meteorology to the phenology ofcyanobacterial blooms: implications for future climate change. Water Res 2012;46(2):442–52.