Journal of Sea Research 85 (2014) 359–364

Contents lists available at ScienceDirect

Journal of Sea Research

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

Temperature effects on phytoplankton diversity — The zooplankton link

Aleksandra M. Lewandowska a,⁎, Helmut Hillebrand b, Kathrin Lengfellner c, Ulrich Sommer a

a Helmholtz Centre for Ocean Research Kiel (GEOMAR), Düsternbrooker Weg 20, 24105 Kiel, Germanyb Institute for Chemistry and Biology of the Marine Environment (ICBM), Carl von Ossietzky University of Oldenburg, Schleusenstrasse 1, 26382 Wilhelmshaven, Germanyc Department of Ecology and Environmental Sciences, Umeå University, 90187 Umeå, Sweden

⁎ Corresponding author. Tel: +494316004411; fax: +4E-mail address: alewandowska@geomar.de (A.M. Lew

1385-1101/$ – see front matter © 2013 Elsevier B.V. All rihttp://dx.doi.org/10.1016/j.seares.2013.07.003

a b s t r a c t

a r t i c l e i n f o

Article history:Received 21 September 2012Received in revised form 5 July 2013Accepted 10 July 2013Available online 17 July 2013

Keywords:Climate warmingMesocosmsPlanktonDiversity

Recent climate warming is expected to affect phytoplankton biomass and diversity in marine ecosystems. Tem-perature can act directly on phytoplankton (e.g. rendering physiological processes) or indirectly due to changesin zooplankton grazing activity. We tested experimentally the impact of increased temperature on natural phy-toplankton and zooplankton communities using indoor mesocosms and combined the results from different ex-perimental years applying a meta-analytic approach. We divided our analysis into three bloom phases to definethe strength of temperature and zooplankton impacts on phytoplankton in different stages of bloom develop-ment. Within the constraints of an experiment, our results suggest that increased temperature and zooplanktongrazing have similar effects on phytoplankton diversity, which aremost apparent in the post-bloomphase, whenzooplankton abundances reach the highest values. Moreover, we observed changes in zooplankton compositionin response to warming and initial conditions, which can additionally affect phytoplankton diversity, becausechanging feeding preferences of zooplankton can affect phytoplankton community structure. We concludethat phytoplankton diversity is indirectly affected by temperature in the post-bloom phase through changingzooplankton composition and grazing activities. Before and during the bloom, however, these effects seem tobe overruled by temperature enhanced bottom-up processes such as phytoplankton nutrient uptake.

© 2013 Elsevier B.V. All rights reserved.

1. Introduction

Current globalwarming pits uswith the necessity to understand andpredict the impact of rising temperatures on ecosystems. During thelast two decades, ecological sciences have therefore put more focus onthis important issue, for example to gain insight into how temperatureaffects properties and functioning of food webs. Recent marine studiesrevealed that temperature impacts marine organisms on different tro-phic levels and alters species interactions between and within trophiclevels (Kordas et al., 2011; O'Connor, 2009; O'Connor et al., 2011).

Phytoplankton generally forms the base of the pelagic food web andhence merits special attention. Although increased temperature speedsup metabolic processes of phytoplankton and might increase primaryproduction in certain regions (Chavez et al., 2011; Doney, 2006), globalphytoplankton declinewith climate warming has been reported (Boyceet al., 2010;Moran et al., 2010). Twomajor processeswere defined to beresponsible for this decline: i) increasing resource limitations as a con-sequence of stronger water column stratification in the warmed oceanand ii) increasing top-down control of phytoplankton by zooplanktonwith rising temperature.

Phytoplankton diversity is also expected to be altered by climatechange but this link is less well understood. Recent studies draw dif-ferent pictures: whereas controlled laboratory experiments reported

9 4316004402.andowska).

ghts reserved.

more rapid competitive exclusion resulting in a loss of species richnessat higher temperature (Burgmer et al., 2011), field studies found an in-creasing number of species (richness) by immigrating warm-adaptedspecies (Beaugrand et al., 2010). It seems, however, that irrespectiveof thenet effect on richness, higher temperatures are strongly associatedto higher species turnover (Hillebrand et al., 2012). Before speciesgo extinct, rising temperatures will alter species dominance. Therefore,phytoplankton evenness (a measure of how equitable species are dis-tributedwithin the community) is expected to be evenmore responsiveto rising temperatures than richness (Hillebrand et al., 2008).

Mesozooplankton can strongly reduce the biomass of microalgaeand affect phytoplankton diversity (richness and evenness). Generallyzooplankton can reduce the number of phytoplankton species by in-creasing phytoplankton mortality or can increase richness by feedingon dominant algae taxa and thus releasing rare species from interspecificcompetition. With respect to evenness, consumers predominantly havea positive effect as they reduce the proportion of the dominant species(Hillebrandet al., 2007). However, zooplankton can also decrease phyto-plankton evenness if the dominant algal species are not included in thezooplankton food spectrum. Thus, the consumer effect on phytoplank-ton evenness depends on consumer quantity as well as on its identityand feeding preferences. Moreover, it depends on the quality of phyto-plankton itself in terms of nutritional value and essential compounds(Hall et al., 2007).

The impact of temperature on phytoplankton depends on its succes-sional stage (Thackeray et al., 2008). Prior the peak in biomass, positive

360 A.M. Lewandowska et al. / Journal of Sea Research 85 (2014) 359–364

effects of temperature on the phytoplankton growth rates prevail(Reynolds, 2006). After the peak in biomass, loss rates exceed thegrowth rates and temperature acts on phytoplankton mainly indirectlymodifying grazing activity of consumers. One can suspect that the phy-toplankton diversity will change over entire time of the phytoplanktonbloom and its response to increased temperature might also vary withthe bloom development.

In this study we hypothesised i) that phytoplankton biomass anddiversity responses to increased temperature differ depending on thebloom phase and ii) that the temperature effects are mediated byincreased zooplankton grazing activity under warmer conditions.Previously published results have shown a decline of phytoplanktonbiomass and size at the bloommaximum in response to increased tem-perature and copepod density (Sommer and Lewandowska, 2011).However, the periods before and after the bloom maximum havenot been analysed yet. Thus, our present analysis extends previousanalyses of the mesocosm experiments not only by its topical focus(on diversity) but also by paying attention to different phases of thebloom development.

2. Material and methods

2.1. Experimental setup and laboratory techniques

Eight (experiments 2006 and 2007) or twelve (experiments 2008and 2009)mesocosms (1400 L volume, 1 m depth)were set up in tem-perature regulated climate rooms. Sea water containing the natural latewinter plankton community (phytoplankton, bacteria and protozoa)from the Kiel Fjord, Baltic Sea, was pumped into a distributiontank and gravitationally transferred to the experimental units. Themesocosms were filled simultaneously to assure homogenous distribu-tion of plankton. Mesozooplankton was added from net catches atappropriate concentrations for each experiment (Table 1) as it didnot pass through the pumping system. The water column was gentlymixed by a propeller. Temperature and light conditions simulated natu-ral daily and seasonal patterns. There were two temperature scenariostested in the experiments 2008 and 2009: a baseline corresponding tothe decadal mean (1993–2002) of sea surface temperature in KielFjord starting from 15th of February (ΔT = 0 °C) and a warming sce-nario where the temperature was elevated 6 °C above the baseline(ΔT = 6 °C) according to the most drastic warming scenario predictedby the Intergovernmental Panel on Climate Change (IPCC, 2007). Inthe experiment 2008 the factor temperature was combined with thefactor light intensity with three levels of the initial surface irradiance(4.8, 5.7 and 6.5 mol quanta m−2 d−1) and in the experiment 2009the factor temperature was combined with the factor initial copepoddensity with three start abundances (1.5, 4 and 10 ind. L−1), resultingin two replicates of each factor combination in each experiment(Lewandowska and Sommer, 2010; Sommer and Lewandowska,2011). In the experiments 2006 and 2007 four temperature regimes:ΔT = 0 °C,ΔT = 2 °C,ΔT = 4 °C andΔT = 6 °Cwere tested wherebyeach regime was replicated twice (Sommer and Lengfellner, 2008).

Table 1Experimental design of mesocosm experiments. Temperature elevation (ΔT), initial light inten

Experiment ΔT (°C) I(mol quantam−2 d−1)

ICD (ind. L−1)

2009 0, 6 5.7 1.5, 4, 10(Acartia)

2008 0, 6 4.8, 5.7, 6.5 8(Oithona)

2007 0, 2, 4, 6 1.9 4.5(Pseudocalanus)

2006 0, 2, 4, 6 3.9 8.5(Pseudocalanus)

For straight comparison between experiments we used only data forΔT = 0 °C and ΔT = 6 °C.

Phytoplankton was sampled three times per week, preserved withLugol's iodine and counted using the inverted microscope techniqueaccording to Utermöhl (1958) for species N5 μm. Flow cytometry tech-nique (FACScalibur, Becton Dickinson)was used to count smaller species(Sommer and Lengfellner, 2008). Phytoplankton biomass was definedas carbon content calculated from cell volumes (Menden-Deuer et al.,2000) after an approximation of cell volumes to geometric standards(Hillebrand et al., 1999). Zooplankton was sampled once a week with anet (12 cm diameter, 64 μm mesh size), shock frozen with liquid nitro-gen (experiments 2006 and 2007) or fixed with Lugol's iodine (experi-ments 2008 and 2009) and counted with a stereomicroscope. Copepodswere specified to the genus level, Temora sp. and rare Eurytemora sp., aswell as Pseudocalanus sp. and rare Paracalanus sp. were paired together,because their early copepodid stages are difficult to distinguish. Copepodbiomass was estimated as a carbon content using species and stagespecific conversion factors (Lengfellner, 2008). More details on theexperimental setup and sampling procedure for each experiment canbe found elsewhere (Lewandowska and Sommer, 2010; Sommer andLengfellner, 2008; Sommer and Lewandowska, 2011).

2.2. Diversity parameters and statistics

Wedefined three phases of the phytoplankton bloomand performedseparate analyses for each of them. The period before the bloom wascharacterised by themean biomass values from the beginning of the ex-periment to the phytoplankton total biomass maximum and representsthe phase of exponential growth. Bloom period was characterised as apoint of the phytoplankton total biomassmaximum (a proxy for phyto-plankton carrying capacity). The post-bloomphasewas characterised bythemean values from the phytoplankton total biomassmaximum to theend of the experiment and represents the phase, inwhich loss processesoverrule phytoplankton growth.

Phytoplankton species richness (S) was calculated as the total num-ber of species, phytoplankton evenness ( J) was calculated according tothe equation:

J ¼ H0

ln S

where H' is the Shannon diversity index (Shannon et al., 1949), whichwe based on biomass proportions, and S is the phytoplankton richness.

We calculated themagnitude of the effect of temperature on phyto-plankton richness, evenness and biomass for each experiment, using logresponse ratios (LRR):

LRR ¼ lnX6�C

X0�C

where X6 °C is phytoplankton richness, evenness or biomass under hightemperature (ΔT = 6 °C) and X0 °C is phytoplankton richness, evennessor biomass under low temperature (ΔT = 0 °C) accordingly.

sities (I), initial copepod densities (ICD) and dominant copepod species.

Bloom forming algae(% phytoplankton biomass)

References

Diatoms(93 ± 6% SD)

Sommer and Lewandowska, 2011

Diatoms(97 ± 6% SD)

Lewandowska and Sommer, 2010

Silicoflagellate(42 ± 38% SD)

Sommer and Lengfellner, 2008diatoms(95 ± 2% SD)

Table 2Effect sizes of increased temperature on phytoplankton diversity (richness and evenness)and biomass for three bloom phases.

Pre-bloom

Experiment Richness Evenness Biomass

Effectsize

Variance Effectsize

Variance Effectsize

Variance

2006 0.0585 0.0018 −0.0771 0.0029 0.2908 0.01322007 −0.0263 0.0011 0.1017 0.0081 −1.5219 0.33762008 −0.0203 0.0004 −0.1560 0.0008 −0.3181 0.09752009 −0.0073 0.0001 −0.0151 0.0008 0.4984 0.2781

Bloom

Experiment Richness Evenness Biomass

Effectsize

Variance Effectsize

Variance Effectsize

Variance

2006 0.0445 0.0019 −0.3596 0.0132 −0.6774 0.05142007 −0.2461 0.0035 0.5558 0.1664 −1.8984 0.16782008 −0.0711 0.0006 −0.0575 0.0016 −0.6136 0.01282009 0.0290 0.0002 0.0530 0.0060 −0.4289 0.0637

Post-bloom

Experiment Richness Evenness Biomass

Effectsize

Variance Effectsize

Variance Effectsize

Variance

2006 −0.0301 0.0001 −0.0030 0.0059 −1.1534 0.00852007 −0.3455 0.0016 0.3859 0.0762 −2.3453 0.48532008 −0.2043 0.0005 0.3241 0.0026 −1.1281 0.03572009 −0.0487 0.0002 0.2823 0.0081 0.3038 0.1510

Fig. 1. Overall effect sizes (±95% confidence intervals, CI) of increased temperature onphytoplankton diversity (richness and evenness) and biomass in three bloom phases(pre-bloom, bloom, post-bloom).

361A.M. Lewandowska et al. / Journal of Sea Research 85 (2014) 359–364

The magnitude of the overall effect across the studies was thencalculated as:

Εþ ¼ Σi¼ni¼1:Εi

Σi¼ni¼1wi

wherewi is the inverse of variance and Ei is the effect size for study i. 95%confidence intervals (95% CI) were used to test, if the effects are signif-icantly different from zero.

The effect of copepods on phytoplankton richness, evenness and bio-mass for each experimental year were calculated as:

Ε¼12

ln1þ r1−r

� �

where r is the Fisher-z-transformed correlation coefficient betweenvariables. We used the biomass of copepods (copepodites and adultsonly) as a proxy to copepods in calculating the effects. The overall effectsacross the experiments were calculated as previously stated (Eq. (3)).Because of a limited number of statistical units, the effects are saddledwith large errors. Nevertheless, the general trend is still evident and im-portant for further interpretation of our results.

Taxonomic compositions of copepodswere compared by conductinga nested analysis of similarities (ANOSIM), in which the factor temper-aturewas nestedwithin the factor year of the experiment to account fordifferences between the experimental years (Clarke, 1993). R statisticwas applied to test for differences between groups (global R = 0 indi-cates completely random grouping). ANOSIMs based on the Bray–Curtisdissimilarity coefficients and multidimensional scaling plots (MDS)were used for graphical representation. All ANOSIM calculations wereperformed in Primer 5 (PRIMER-E Ltd).

Redundancy analysis was performed to estimate the proportionof variance in phytoplankton biomass and diversity (richness andevenness) explained by experimental year, temperature, initial copepodabundance and light intensity. The analysiswas performed using vegan-package in R (Development Core Team, version 2.15.2).

3. Results

3.1. Course of phytoplankton biomass

In our experiments phytoplankton biomass was dominated bySkeletonema costatum. Phytoplankton biomass increased by 100–200%within the first 3 weeks and declined afterwards. An exception to thiswas the experiment 2007 where an initial stagnation of phytoplanktonbiomass occurred and the bloom peakwas observed after 6 weeks fromthe start of the experiment. Phytoplankton community during thebloom in 2007was dominated by the silicoflagellateDictyocha speculum(Table 1). In all experiments phytoplankton biomass decreased inresponse to warming (for further details on changes in phytoplank-ton biomass at the bloom peak in particular years see Lewandowskaand Sommer, 2010; Sommer and Lengfellner, 2008; Sommer andLewandowska, 2011).

3.2. Temperature impacts on phytoplankton biomass and diversity

The effects of temperature on phytoplankton biomass varied be-tween the individual experiments, especially before the bloom, whenalgae were still growing (Table 2). The largest effects of temperatureon phytoplankton biomass were observed in 2007. Total biomass ofphytoplankton initially increased in response to warming, but duringthe bloom and especially after the bloom, a negative effect of increasedtemperature was observed.

Phytoplankton species richness did not respond to increased tem-perature before and during the bloom (Fig. 1). However, richnessdeclined with temperature after the bloom. Phytoplankton evenness

showed a negative response to increased temperature before thebloom and no response during the bloom. After the bloom, phytoplank-ton evenness increased with warming (Fig. 1). Again, the largest effectsof temperature on phytoplankton diversity were observed in 2007.

Redundancy analysis showed that temperature explained 27% of thevariance in the data set in the pre-bloom phase (experimental year andinitial copepod abundance explained 7% and 1%, respectively), 33% ofthe variance in the bloom phase (year explained 15% and copepodsexplained 2%) and 31% of the variance in the post-bloom phase (yearand copepods explained 12% and 1%, respectively).

3.3. Copepod impacts on phytoplankton biomass and diversity

The effects of copepods on phytoplankton biomass and diversitydepended on experimental years (Table 3), which differed in initialcopepodnumbers and community composition. The overall effects of co-pepods on phytoplankton biomass and diversity were non-significant inmost cases, except a strong positive effect on phytoplankton evenness

Table 3Effect sizes of copepod abundance on phytoplankton diversity (richness and evenness)and biomass for three bloom phases.

Pre-bloom

Experiment Richness Evenness Biomass

Effectsize

Variance Effectsize

Variance Effectsize

Variance

2006 −0.4053 1.0000 −0.0658 1.0000 −0.1791 1.00002007 −0.7083 1.0000 0.7175 1.0000 −0.8054 1.00002008 −0.0781 0.1111 −0.4457 0.1111 −0.3825 0.11112009 0.6267 0.1111 0.7207 0.1111 −0.5151 0.1111

Bloom

Experiment Richness Evenness Biomass

Effectsize

Variance Effectsize

Variance Effectsize

Variance

2006 −1.2783 1.0000 0.7884 1.0000 1.6038 1.00002007 −0.9099 1.0000 0.9873 1.0000 −0.8534 1.00002008 −0.3987 0.1111 −0.5621 0.1111 −0.4144 0.11112009 0.7664 0.1111 0.4939 0.1111 −0.8520 0.1111

Post-bloom

Experiment Richness Evenness Biomass

Effectsize

Variance Effectsize

Variance Effectsize

Variance

2006 1.3099 1.0000 −0.0212 1.0000 2.2846 1.00002007 −0.8667 1.0000 2.9685 1.0000 −2.2388 1.00002008 −0.9374 0.1111 0.7060 0.1111 −0.8226 0.11112009 −0.4625 0.1111 0.8152 0.1111 0.3038 0.1111

362 A.M. Lewandowska et al. / Journal of Sea Research 85 (2014) 359–364

during the post-bloom phase (Fig. 2). The overall effects of copepods onphytoplankton diversity and biomass had the same directions as theeffects of increased temperature (Fig. 1).

3.4. Variability of copepod composition between the years and bloomphases

Initial composition of zooplankton varied between the years(ANOSIM between years for the pre-bloom phase: global R = 0.896,p = 0.01, 105 permutations, Fig. 3). The experiment 2006 wasdominated by Pseudocalanus over the whole experimental period. Thestarting density of copepods was 8.5 ind. L−1. The copepods in the ex-periment 2007 were initially dominated by Pseudocalanus and Oithona,but the dominance slightly shifted towards Centropages at the endof theexperiment in the warm treatments (ΔT = 6 °C). Overall, the initialdensity of copepods (4.5 ind. L−1) was lower than in the previous ex-periment. In the experiment 2008, a shift in copepod composition was

Fig. 2. Overall effect sizes (±95% confidence intervals, CI) of copepod abundance onphytoplankton diversity (richness and evenness) and biomass in three bloom phases(pre-bloom, bloom, post-bloom).

observed. All mesocosms were initially dominated by Oithona (initialcopepod density: 8 ind. L−1), but Temora and Centropages becamemore abundant in the warmer mesocosms (ΔT = 6 °C) after thebloom and started to dominate. The experiment 2009 was dominatedby Acartia during the entire experiment and no differences in copepodcomposition could be noticed between colder (ΔT = 0 °C) andwarmer(ΔT = 6 °C) treatments.

Increased temperature altered the composition of copepods afterthe phytoplankton bloom (ANOSIM between temperature groups,global R = 0.426, p = 0.001, 999 permutations, Fig. 3), but not beforeand during the bloom (ANOSIM between temperature groups for thepre-bloom phase: global R = 0.186, p = 0.054, 999 permutations;ANOSIM between temperature groups for the bloom phase: globalR = 0.096, p = 0.132, 999 permutations).

The difference in copepod composition between colder (ΔT = 0 °C)and warmer (ΔT = 6 °C) mesocosms in the experiment 2006 couldalready be observed during the pre-bloomphase. In contrast, separationbetween mesocosms by temperature was not evident at the pre-bloomand bloom phases in the other experiments (2007–2009) (Fig. 3), butapparent during the post-bloom phase. The experiment 2009 wascharacterised by the strong variability within the year caused by thedifferent copepod density treatments (1.5, 4 and 10 ind. L−1).

4. Discussion

Consumer activity tends to increase in response to increased tem-perature (Lopez-Urrutia et al., 2006; O'Connor, 2009). For marine phy-toplankton this means that with rising sea surface temperature top-down effects of their consumers will become stronger. Results of ourmesocosm experiments confirm the field observations on decliningphytoplankton biomass in response to climate warming and increasinggrazing pressure (Figs. 1 and 2, see also Sommer and Lewandowska,2011). In this article, we show that zooplankton not only reduces thebiomass of phytoplankton, but also has a strong impact on phytoplank-ton diversity. However, the top-down control of algal consumers isoften counteracted by bottom-up processes and a direct positive effectof warming on phytoplankton metabolism (Eppley, 1972). It shouldbe emphasised that our experimental design had many limitations(Sommer and Lengfellner, 2008) and the differences in initial conditionsbetween experimental years might be substantial. Although experi-mental year explained a large part of the variance in our data, tempera-ture was the best predictor of changes in phytoplankton biomass anddiversity. Moreover, most of the field studies, which address the impactof warming on phytoplankton, are confronted with potentially greatereffects of confounding factors associated with seasonal and specialvariations. Thus, our conclusions are valid for the typical spring bloomconditions in temperate waters.

While warming enhances grazing activity of zooplankton and thusleads to the reduction of the phytoplankton biomass, increased temper-ature has been shown to have a positive effect on algae's growth due toa faster nutrient uptake (O'Connor et al., 2009). Thus initially, when theloss of producer biomass due to grazing was balanced by increasingmetabolic rates under enhanced temperature, we observed a slightlypositive, however non-significant, effect ofwarming on the phytoplank-ton biomass (Fig. 1). Later on, when the top-down control prevailedand nutrients were depleted a strong reduction of the phytoplanktonbiomass could be observed in all experiments. This became mostevident in 2007 where biomass was dominated by the silicoflagellateD. speculum in contrast to diatoms, whichdominated in the other exper-imental years. An exception to this was the experiment 2009, duringwhich we observed slightly positive effect of temperature on phyto-plankton biomass after the bloom (Table 2). This unexpected responsewas probably caused by a trophic cascade effect, which is expected tobecome stronger under warmer conditions (O'Connor, 2009). Afterthe decline of well edible (medium and moderately large) phytoplank-ton species copepods seem to have switched to feeding on ciliates,

Fig. 3. Multidimensional scaling plot of variation in copepod compositions among experimental years (2006–triangles, 2007–squares, 2008–circles, 2009–diamonds) and temperaturetreatments (ambient temperature – open symbols; warming scenario – filled symbols).

363A.M. Lewandowska et al. / Journal of Sea Research 85 (2014) 359–364

thereby releasing small phytoplankton from ciliate grazing pressure.In consequence, small algae would have been able to build up morebiomass, resulting in the observed differences between warmer andcolder mesocosms in our study. Another explanation of the increasedphytoplankton biomass after the bloom in 2009 might be a strongergrowth of benthic algae on the walls of the mesocosms under elevatedtemperature and faster remineralisation processes, which enhancedthe available nutrient pool for phytoplankton growth. Although some“wall growth” in themesocosms could be observed at the end of the ex-periments, therewas no visible evidence of temperature driven changesbetween the mesocosms and enhanced nutrient concentrations in thepost-bloom phase.

During pre-bloom and bloom phases there were no significantchanges in phytoplankton richness because phytoplankton mortalitydue to copepod grazing was well balanced by a positive phytoplanktongrowth and no new phytoplankton species were introduced to the sys-tem. After the bloom phytoplankton richness decreased because con-sumers reduced the number of edible phytoplankton species (Fig. 2)which in consequence fell below the detection limit and reduced ap-parent richness. Nevertheless, it should be noticed that other loss pro-cesses (e.g. sinking) should also be considered as possible reasons forthe loss of species at the end of the experimental period. This is, how-ever, less probable scenario as the phytoplankton communities afterthe bloom were dominated by motile species, which can counteractsinking. Interestingly, if copepods feed on the dominant phytoplank-ton species, as it happened in our experiments, they might have apositive effect on phytoplankton richness (Table 3, experiment 2006),because their impact on the dominant competitor is disproportionallygreater and species below the detection limit might be released fromcompetition and become detectable. If this is not the case, it would sug-gest that rare phytoplankton species avoided by grazers were not ableto compensate the loss of dominant species under depleted nutrientconditions after the bloom.

Overall, phytoplankton evenness negatively responded to increasedtemperature throughout the pre-bloom phase (Fig. 1). As the copepod

grazing seemed to have no significant effect on thephytoplankton even-ness before and during the bloom (Fig. 2) we hypothesise that high ini-tial nutrient contents promoted growth of a few dominant diatoms(mostly S. costatum) which “outgrew” the other phytoplankton groupsand strongly dominated the blooms. Only in 2007, D. speculum domi-nated the phytoplankton community instead of diatoms, and this wasalso the only year when we observed a positive impact of temperatureon phytoplankton evenness (Table 2). This suggests that the initial phy-toplankton composition, especially the presence of good competitorsand fast growing species, strongly affects the response of the wholephytoplankton community to environmental changes. In fact, the ex-periment 2007was characterised by the slowest growth and the largestdecrease of phytoplankton in response to warming.

As soon as copepods started to play a major role in controllingthe bloom, they increased phytoplankton evenness by feeding onthe most dominant species, which is often observed in consumer-producer relationships (Hillebrand et al., 2007). Thus we observed apositive effect of copepods (Fig. 2) and temperature (the effect drivenby the higher grazing activity under warmer conditions, Fig. 1) on phy-toplankton evenness throughout the post-bloom phase. The year 2006was an exception (Table 2) as the copepods abundance dropped downafter the bloom (Lengfellner, 2008). The opposite effect of copepodgrazing activity on phytoplankton evenness would be suspected if thebloom had been dominated by inedible phytoplankton (too smallor too big species, toxic algae). In such a case we would suspect rathera decrease of phytoplankton evenness with increasing grazing pressurebecause copepodswould feedmostly on the rare edible species therebyincreasing phytoplankton dominance.

Besides grazing pressure, which is related to enhanced zooplanktonactivity or abundance, zooplankton feeding preferences and thuscommunity composition can also potentially influence phytoplanktondiversity. Here, the role of microzooplankton, especially ciliates, asconsumers of small phytoplankton species might be very important. Inour experiments microzooplankton did not respond to temperaturechanges in terms of biomass (Aberle et al., 2012). Thus, we focused on

364 A.M. Lewandowska et al. / Journal of Sea Research 85 (2014) 359–364

copepods as the major phytoplankton consumers. Although copepodsare expected to feed on the same phytoplankton size spectrum (N500to 1000 μm3 cell volume, Sommer and Sommer, 2006), they havedifferent feeding behaviours and some can switch between suspensionfeeding and raptorial feeding (Tiselius et al., 1990). Thus, copepod effec-tiveness in catching their prey might vary between species, therebyaffecting phytoplankton diversity. It has been shown that rising seasurface temperatures provoke changes in zooplankton communitycomposition towardswarmwater species (e.g. Centropages and Temora)similar to those observed in our mesocosms (Möllmann et al., 2000).Temperature induced responses in copepods such as taxonomic shifts(Fig. 3) and enhanced grazing activities, are possible explanations forthe strong changes in phytoplankton diversity that we could observein the post-bloom phase of our experiments. Similarly, changes inthe initial zooplankton composition between the experimental yearsmight explain part of the interannual variability in phytoplankton diver-sity responses to warming. Using our experimental setup, we were notable to directly disentangle the effects of different copepod species onphytoplankton diversity. The limited number of experiments includedin our analysis, low statistical power due to replication of only twomesocosms per treatment combination or potential confounding effectsdue to differences in initial conditions should be also considered. Never-theless our results suggest that the changes in zooplankton communitycomposition might be crucial to understand the effects of warming onaquatic ecosystems and there is a need for further, well designed studieson consumer impact on phytoplankton diversity. This future workshould be improved by sufficient treatment replication, longer experi-ment duration (to capture all development stages of zooplankton),repeated experiments with different initial communities (e.g. with thesame experimental setup in different locations) and by parallel grazingexperiments to better understand zooplankton feeding behaviour.

In conclusion, the impact of temperature on phytoplankton diversityseems to be mostly mediated via changes in zooplankton activity andcommunity structure, but the strength of the zooplankton impact onphytoplankton varies with the bloom development. There is still a lackof knowledge about how thenaturally variable zooplankton communityaffects phytoplankton diversity and how this relationship changes in re-sponse to climate warming. Thus, there is a need of complex ecosystemstudies, where community interactions could be fully represented.

Acknowledgements

This studywas funded by DFG (German Research Foundation)with-in the priority programme 1162 “AQUASHIFT”. T. Hansen, H. Tomanetzand C.Meyer are acknowledged for their technical assistance.We thankN. Aberle for the microzooplankton data.

References

Aberle, N., Bauer, B., Lewandowska, A., Gaedke, U., Sommer, U., 2012. Warming inducesshifts inmicrozooplankton phenology and reduces time-lags between phytoplanktonand protozoan production. Marine Biology 159, 2441–2453.

Beaugrand, G., Edwards,M., Legendre, L., 2010.Marine biodiversity, ecosystem functioning,and carbon cycles. Proceedings of the National Academy of Sciences of the UnitedStates of America 107 (22), 10120–10124.

Boyce, D.G., Lewis, M.R., Worm, B., 2010. Global phytoplankton decline over the pastcentury. Nature 466, 591–596.

Burgmer, T., Hillebrand, H., 2011. Temperature mean and variance alter phytoplanktonbiomass and biodiversity in a long-term microcosm experiment. Oikos 120 (6),922–933.

Chavez, F.P., Messie, M., Pennington, J.T., 2011. Marine primary production in relation toclimate variability and change. Annual Review of Marine Science 3, 227–260.

Clarke, K.R., 1993. Non-parametric multivariate analyses of changes in community struc-ture. Australian Journal of Ecology 18, 117–143.

Doney, S.C., 2006. Phytoplankton in a warmer world. Nature 444, 695–696.Eppley, R.W., 1972. Temperature and phytoplankton growth in the sea. Fishery Bulletin

70 (4), 1063–1085.Hall, R.S., Shurin, J.B., Diehl, S., Nisbet, R.M., 2007. Food quality, nutrient limitation of

secondary production, and the strength of trophic cascades. Oikos 116, 1128–1143.Hillebrand, H., Dürselen, C.D., Kirschtel, D., Pollingher, U., Zohary, T., 1999. Biovolume

calculation for pelagic and benthic microalgae. Journal of Phycology 35, 403–424.Hillebrand, H., Gruner, D.S., Borer, E.T., Bracken, M.E.S., Cleland, E.E., Elser, J.J., Harpole,

W.S., Ngai, J.T., Seabloom, E.W., Shurin, J.B., Smith, J.E., 2007. Consumer versus re-source control of producer diversity depends on ecosystem type and producer com-munity structure. Proceedings of the National Academy of Sciences of the UnitedStates of America 104 (26), 10904–10909.

Hillebrand, H., Bennett, D.M., Cadotte, M.W., 2008. Consequences of dominance: a reviewof evenness effects on local and regional ecosystem processes. Ecology 89 (6),1510–1520.

Hillebrand, H., Burgmer, T., Biermann, E., 2012. Running to stand still: temperature effectson species richness, species turnover, and functional community dynamics. MarineBiology 159 (11), 2415–2422.

IPCC, 2007. Climate Change 2007: The Physical Science Basis. IPCC, Geneva.Kordas, R.L., Harley, C.D.G., O'Connor, M.I., 2011. Community ecology in a warmingworld:

the influence of temperature on interspecific interactions in marine systems. Journalof Experimental Marine Biology and Ecology 400 (1–2), 218–226.

Lengfellner, K., 2008. The Impact of Climate Warming on Plankton Spring Succession: AMesocosm Study. Christian Albrechts University of Kiel, Kiel 110 (PhD Dissertation).

Lewandowska, A., Sommer, U., 2010. Climate change and the spring bloom: amesocosm study on the influence of light and temperature on phytoplankton andmesozooplankton. Marine Ecology Progress Series 405, 101–111.

Lopez-Urrutia, A., San, Martin E., Harris, R.P., Irigoien, X., 2006. Scaling the metabolic bal-ance of the oceans. Proceedings of the National Academy of Sciences of the UnitedStates of America 103, 8739–8744.

Menden-Deuer, S., Lessard, E.J., 2000. Carbon to volume relationships for dinoflagellates,diatoms, and other protist plankton. Limnology and Oceanography 45 (3), 569–579.

Möllmann, C., Kornilovs, G., Sidrevics, L., 2000. Long-term dynamics of mainmesozooplankton species in the central Baltic Sea. Journal of Plankton Research22 (11), 2015–2038.

Moran, X.A.G., Lopez-Urrutia, A., Calvo-Diaz, A., Li,W.K.W., 2010. Increasing importance ofsmall phytoplankton in a warmer ocean. Global Change Biology 16 (3), 1137–1144.

O'Connor, M.I., 2009. Warming strengthens an herbivore-plant interaction. Ecology 90(2), 388–398.

O'Connor, M.I., Piehler, M.F., Leech, D.M., Anton, A., Bruno, J.F., 2009. Warming andresource availability shift food web structure and metabolism. PLoS Biology 7 (8).

O'Connor, M.I., Gilbert, B., Brown, C.J., 2011. Theoretical predictions for how temperatureaffects the dynamics of interacting herbivores and plants. American Naturalist 178(5), 626–638.

Reynolds, C.S., 2006. The Ecology of Phytoplankton. Cambridge University Press,Cambridge.

Shannon, C.E., Weaver, W., 1949. TheMathematical Theory of Communication. Universityof Illinois Press, Urbana, IL.

Sommer, U., Lengfellner, K., 2008. Climate change and timing,magnitude, and compositionof the phytoplankton spring bloom. Global Change Biology 14, 1–10.

Sommer, U., Lewandowska, A., 2011. Climate change and the phytoplankton spring bloom:warming and overwintering zooplankton have similar effects on phytoplankton.Global Change Biology 17 (1), 154–162.

Sommer, U., Sommer, F., 2006. Cladocerans versus copepods: the cause of contrasting top-down controls on freshwater and marine phytoplankton. Oecologia 147, 183–194.

Thackeray, S.J., Jones, I.D., Frederiksen, M., Burthe, S., Bacon, P.J., Bell, J.R., Botham, M.S.,Brereton, T.M., Bright, P.W., Carvalho, L., Clutton-Brock, T., Dawson, A., Edwards, M.,Elliott, J.M., Harrington, R., Johns, D., Jones, I.D., Jones, J.T., Leech, D.I., Roy, D.B.,Scott, W.A., Smith, M., Smithers, R.J., Winfield, I.J., Wanless, S., 2008. Long-termchange in the phenology of spring phytoplankton: species-specific responses to nu-trient enrichment and climatic change. Journal of Ecology 96 (3), 523–535.

Tiselius, P., Jonsson, P.R., 1990. Foraging behaviour of six calanoid copepods: observationsand hydrodynamic analysis. Marine Ecology Progress Series 66, 23–33.

Utermöhl, H., 1958. Zur Vervollkommung der quantitativen Phytoplankton-Methodik.Mitteilung InternationaleVereinigung fuer TheoretischeundeAmgewandte Limnologie9, 263–272.