Patterns of thermal limits of phytoplankton

Transcription

1 Journal of Plankton Research plankt.oxfordjournals.org J. Plankton Res. (2015) 37(2): First published online February 27, 2015 doi: /plankt/fbv009 Patterns of thermal limits of phytoplankton BINGZHANG CHEN* STATE KEY LABORATORY OF MARINE ENVIRONMENTAL SCIENCE AND FUJIAN PROVINCIAL JOINT KEY LABORATORY FOR COASTAL ECOLOGY AND ENVIRONMENTAL STUDIES, XIAMEN UNIVERSITY, XIAMEN, FUJIAN, P. R. CHINA *CORRESPONDING AUTHOR: Received October 31, 2014; accepted February 8, 2015 Corresponding editor: John Dolan The latitudinal patterns of the optimal, minimal and maximal growth temperatures of phytoplankton are analyzed using linear mixed-effect models, and whether environmental temperature plays a role in affecting these thermal traits is tested. The optimal, minimal and maximal growth temperatures of phytoplankton decrease with latitude for marine taxa; whereas the minimal and maximal growth temperatures are relatively invariant with latitude for freshwater phytoplankton. The thermal breadth, defined as the range between the maximal and minimal growth temperature, is larger for freshwater than marine phytoplankton. In contrast to Jenzen s rule, there is no increasing trend of thermal breadth with increasing latitude. For most phytoplankton, the minimal growth temperatures are lower than the lowest environmental temperatures and the maximal growth temperatures are higher than the highest environmental temperatures. In marine phytoplankton, there is a strong phylogenetic signal in the minimal growth temperature and hence the thermal breadth. After controlling other variables (i.e. latitude and maximal growth rate) constant, the minimal growth temperatures of cyanobacteria and dinoflagellates are significantly higher than that of diatoms. The thermal breadths of cyanobacteria and dinoflagellates are narrower than of diatoms. The maximal growth rate is positively correlated with thermal breadth for marine but not freshwater phytoplankton. KEYWORDS: phytoplankton; temperature; macroecology; Janzen s rule Featured Article INTRODUCTION Temperature is an important factor affecting distribution and productivity of phytoplankton (Behrenfeld et al., 2006; Finkel et al., 2010). Species distributions often track the changes of isotherms (Walther et al., 2002; Gregory et al., 2009). The limits of species distribution may be determined by physiological tolerance to temperature extremes or by indirect factors such as solar radiation, nutrients or biotic interaction that covary with temperature in aquatic ecosystems (Brown, 1995). Knowledge of the temperature tolerance curves or thermal reaction norms is essential for predicting how available online at # The Author Published by Oxford University Press. All rights reserved. For permissions, please journals.permissions@oup.com

2 JOURNAL OF PLANKTON RESEARCH j VOLUME 37 j NUMBER 2 j PAGES j 2015 organisms and ecosystems will respond to future climate change (Follows et al., 2007; Thomas et al., 2012). While there have been numerous studies on the effect of temperature on phytoplankton growth rate, information on temperature limits has not been adequately synthesized. Classic functions describing the dependence of phytoplankton growth rate on temperature are often monotonous (e.g. exponential or power law) (Eppley, 1972; Brown et al., 2004), without taking into account the effect of temperature extremes. While some studies focus on the optimal growth temperature of phytoplankton (Thomas et al., 2012), there is little information on the temperature limits of phytoplankton although its importance has long been recognized (Li, 1980, 1985). Analogous to optimal growth temperatures, temperature limits can also be viewed as species functional traits that might be useful in understanding ecological responses to environmental factors. In this paper, I focus on the latitudinal distributions of temperature limits of phytoplankton, by analyzing a dataset compiled from laboratory culture experiments using linear mixed-effect models. I evaluate three hypotheses related to the patterns of phytoplankton thermal limits. The first is whether the environmental temperatures affect phytoplankton thermal limits. Secondly, Janzen s rule states that the thermal breadth often increases with latitude owing to the fact that the minimal growth temperature decreases with latitude and upper temperature limits are relatively constant with latitude (Addo-Bediako et al., 2000; Deutsch et al., 2008; Gaston et al., 2009). Li (Li, 1985) used a small dataset to show that the thermal breadth of phytoplankton photosynthetic rate seems to increase with increasing latitude. It remains unknown whether phytoplankton will follow similar patterns in larger, more comprehensive datasets. The third question is whether there is a tradeoff between maximal growth rate and thermal breadth. A reasonable hypothesis, based on the supposed tradeoff between the evolution of an optimal enzyme structure under the optimal temperature and enzyme functions under non-optimal temperatures,isthatageneralistthatcansurviveunderawide range of temperatures should grow slower than a specialist attheoptimaltemperatureofthespecialist(levins, 1964). However, Huey and Hertz (Huey and Hertz, 1984) show a counter-example in lizards whereby the fastest species can also tolerate the broadest thermal ranges. What patterns exist for phytoplankton remains an interesting problem. METHOD I compiled from the literature the data of phytoplankton growth rates cultured at different temperatures under optimal light and nutrient conditions (see Supplementary Appendix). This dataset merged four previous data compilations (López-Urrutia et al., 2006; Rose and Caron, 2007; Bissinger et al., 2008; Thomas et al., 2012), with additional data from recently published literature. The optimal growth temperature was roughly determined as the temperature corresponding to the highest growth rate when there is a unimodal relationship between temperature and growth rate. The minimal growth temperature (i.e. the lower end of the temperature limit) was determined as the highest temperature above which the growth rate was positive and the maximal growth temperature (i.e. the upper temperature limit) was determined as the lowest temperature below which the growth rate was positive. This definition minimizes the problem that there could be a sharp increase (or decrease) of growth rate from zero (or a significant positive value) to a significant positive value (or zero) between two adjacent experimental temperatures. The thermal breadth of phytoplankton was determined as the range between maximal and minimal growth temperature. I also recorded the coordinates of the location where the species were isolated (Fig. 1) and estimated the annual mean temperature as well as minimum and maximum temperatures of these locations based on the World Ocean Atlas 2009 (http :// with the method of k-nearest neighbor classification using the function knn in the R package class. Cell volume was also recorded. When cell volume was not reported in the primary literature, the volume of the same species was taken from the secondary literature. I used the latitude, cell volume and phylum as fixedeffect predictors for the optimal, minimal and maximal growth temperatures, collectively referred to as thermal traits in the following text. Because the within-species variations of the thermal traits (i.e. multiple measurements for one species are common) are significantly lower than among-species based on F-tests (P, 0.05), mixed-effect linear regression models were applied to allow random variations among species, which were implemented with the function lme in the R package nlme (Pinheiro and Bates, 2000). I used a second-order polynomial to fit the relationship between the optimal, minimal and maximal temperatures and latitude, in addition to the linear effects of log-transformed cell volume and phylum (Table I). Higher orders of polynomials were also tried but were found unnecessary. I used a mixed-effect generalized additive model with a penalized cubic regression spline to fit the latitudinal distribution of thermal breadth that was more complex than the simple unimodal pattern. This was implemented with the function gamm in the R package mgcv (Wood, 2006). I also tried a fourth-order polynomial, which gave similar results. The maximal growth rate was also considered 286

3 B. CHEN j THERMAL LIMITS OF PHYTOPLANKTON Fig. 1. A map showing the locations where the phytoplankton taxa were originally isolated. The filled circles are for marine species and the open triangles are for freshwater species. The contours indicate the annual mean sea surface temperature (8C) estimated from the World Ocean Atlas (WOA) 2009 climatology. as a predictor for thermal breadth based on the consideration of the potential tradeoff between maximal growth rate and thermal breadth mentioned above. All the analysis and graph plotting were conducted with R (R Core Team, 2014). RESULTS Table I: Regression models and relevant statistics for the optimal (T op, 8C), minimal (T min, 8C) and maximal (T max, 8C) growth temperatures and derived thermal breadths ( T b ¼ T max 2 T min, 8C). Models N SD R Optimal temperature Marine: T op (i,j) ¼ 20.7 þ 0.077Lat Lat 2 þ 0.66 ln(vol) þ Phylum j þ 1 i,sp Marine: T op (i,j) ¼ 1.1 þ 2.51T a 0.118T 2 a þ 0.002T 3 a þ 0.64 ln(vol) þ Phylum j þ 1 i,sp Freshwater: T op ¼ Lat Lat 2 64 Freshwater: T op ( j) ¼ 13.6 þ 0.46T a þ Phylum j 64 Minimal temperature Marine: T min (i,j) ¼ 9.5 þ 0.035Lat 0.003Lat 2 þ Phylum j þ 1 i,sp Marine: T min (i,j) ¼ 0.35 þ 0.40T l þ Phylum j þ 1 i,sp Maximal temperature Marine: T max (i,j) ¼ 33.6 þ 0.068Lat 0.005Lat 2 þ Phylum j þ 1 i,sp Marine: T max (i) ¼ 8.7 þ 1.6T h 0.028T 2 h þ 1 i,sp Freshwater: T max ¼ 26.9 þ 0.21T h þ Phylum j 45 Thermal breadth Marine: T b (i,j) ¼ s(lat) þ Phylum j þ 1.49 ln(m m ) þ 1 i,sp Only three significant regression models (two for optimal growth temperature and one for maximal temperature) are given for freshwater phytoplankton. N, number of observations; Sd r, standard deviation of the random effect; T a, environmental annual mean temperature; T l, the lowest environmental temperature; T h, the highest environmental temperature; T op (i,j), the optimal growth temperature of a taxon belonging to the ith species of the jth phylum. Similar meanings apply to T min (i,j); T max (i,j), and T b (i,j). 1 i,sp, the random deviation from the mean of the ith species; Vol, cell volume (mm 3 ). In total, I compiled 339 phytoplankton taxa including 275 marine and 64 freshwater. The marine taxa include 116 diatoms, 56 dinoflagellates, 44 cyanobacteria, 28 haptophytes, 14 raphidophytes, 10 chlorophytes and a few others. The freshwater ones were dominated by cyanobacteria (44 taxa). The optimal growth temperatures were highly correlated with maximal growth temperatures (Pearson r ¼ 0.56 and 0.92, respectively, for freshwater and marine taxa, P, ), but were only weakly or insignificantly correlated with minimal growth temperature (Pearson r ¼ 0.06 and 0.18 and P ¼ 0.76 and 0.05, respectively, for freshwater and marine taxa). In marine taxa, the thermal breadth was more strongly correlated with minimal growth temperature (Pearson r ¼ -0.72, P, ) than with maximal growth temperature (Pearson r ¼ 0.29, P, ), implying that it is the minimal temperature rather than the maximal temperature that more likely determines thermal breadth. In marine phytoplankton, thermal breadth was highly positively correlated with maximal growth rate (r ¼ 0.68, P, ). Cell size was not correlated with any of the thermal traits. The optimal growth temperatures of both marine and freshwater phytoplankton decreased from the equator to high latitudes and this decreasing trend of marine 287

4 JOURNAL OF PLANKTON RESEARCH j VOLUME 37 j NUMBER 2 j PAGES j 2015 phytoplankton was more drastic than that of freshwater phytoplankton (Fig. 2A). There was a positive trend between optimal temperature and cell volume for marine taxa when controlling latitude constant (Table I; P, 0.01). For marine phytoplankton, after controlling latitude and cell volume constant, the average optimal temperature of cyanobacteria was significantly higher than that of diatoms (t ¼ 2.77, P, 0.05), while the optimal temperatures of other phyla were not significantly different from that of diatoms. In freshwater phytoplankton, no significant effects of cell size or phylum on optimal temperature could be found. The likelihood ratio test suggested that the random variations of optimal temperature were significant for marine taxa (likelihood ratio ¼ 25.8, P, 0.001), but not for freshwater taxa (P. 0.05). The latitudinal patterns of minimal and maximal growth temperatures of marine phytoplankton were similar to those of optimal temperatures (Fig. 2B and D). Cell size did not play a role in affecting minimal and maximal growth temperatures of marine or freshwater phytoplankton. The signal of phyla was evident for minimal growth temperatures of marine phytoplankton (Table I). After controlling latitude constant, the minimal growth temperatures of dinoflagellates and cyanobacteria were significantly higher than those of diatoms (P, 0.001; Fig. 2C). The maximal growth temperature of chlorophytes was also significantly higher than that of diatoms (P, 0.001). For freshwater phytoplankton, the patterns were not evident. Visual inspection suggested that the maximal growth temperatures of freshwater phytoplankton were often higher than those of marine ones (Fig. 2D). The minimal and maximal growth temperatures of freshwater phytoplankton were relatively invariant with latitude (Fig. 2B and D). The thermal breadth of marine phytoplankton did not evidently increase with latitude as Jenzen s rule predicts (Fig. 2E). When thermal breadth was calculated as the difference between minimal and maximal temperatures that were predicted based on fixed effects of the linear mixed-effect models in Table I, there was even a trend that the thermal breadth peaked at the equator and decreased to high latitudes (Fig. 2F). Maximal growth rate increased with thermal breadth irrespective of whether latitude and phylum were controlled (Fig. 2G). Holding latitude and maximal growth rate constant, the thermal breadths of dinoflagellates, cyanobacteria and haptophytes were significantly higher than those of diatoms (P, 0.001; Fig. 2H). Freshwater phytoplankton could tolerate wider temperature ranges than marine taxa (Fig. 2E). The positive relationship between maximal growth rate and thermal breadth did not exist in freshwater phytoplankton (Fig. 2F). I also examined the relationships between environmental temperature extremes and thermal limits of phytoplankton growth to assess to what extent the thermal limits could be affected by environmental temperature extremes (Fig. 3). The optimal growth temperatures of both marine and freshwater phytoplankton increased with environmental annual mean temperature (Table I; Fig. 3A) with the trend more pronounced for marine phytoplankton. The relationship between optimal temperature of marine phytoplankton and environmental mean temperature was well fit by a three-order polynomial, confirming the conclusion that species optimal growth temperatures could be lower than annual mean temperature in the tropics (Thomas et al., 2012) (Fig. 3A). The minimal growth temperatures of marine taxa increased linearly with, and were usually lower than, environmental lowest temperatures. The phylogenetic signal was still evident in the minimal growth temperatures. When controlling environmental lowest temperatures constant, the average minimal growth temperatures of marine chlorophytes, cyanobacteria, dinoflagellates and raphidophytes were all significantly higher than those of diatoms (P, 0.01). The maximal growth temperatures of both marine and freshwater phytoplankton increased significantly with environmental highest temperature, with the trend of marine phytoplankton much more drastic than of freshwater ones (Fig. 3C). The maximal growth temperatures of freshwater phytoplankton were significantly higher than marine ones and were much higher than environmental highest temperatures, particularly for polar taxa (Fig. 3C). There were no clear trends of thermal breadths of phytoplankton with environmental temperature ranges (Fig. 3D). Again, the phylogenetic differences were pronounced in the thermal breadth of marine phytoplankton. The thermal breadths of cyanobacteria, dinoflagellates, and haptophytes were significantly lower than those of diatoms (P, 0.001). DISCUSSION This study provides a macroscopic pattern of thermal limits of phytoplankton growth, which has not been thoroughly investigated previously. Although this dataset is compiled from the datasets of four comprehensive studies, there are still limitations associated with this dataset. There are insufficient observations of freshwater taxa, particularly in the tropics. And the freshwater taxa are over-represented by cyanobacteria. Therefore, the analysis and discussion mainly focuses on marine taxa. The null hypothesis of the first hypothesis is that the environment does not affect the thermal traits of phytoplankton. The fact that the optimal, minimal and maximal growth temperatures are highly correlated with 288

5 B. CHEN j THERMAL LIMITS OF PHYTOPLANKTON Fig. 2. (A) Optimal growth temperatures versus latitude. The solid line represents the regression line for a marine diatom with a cell volume of 500 mm 3 and the dashed line is for freshwater cyanobacteria. The thin dotted lines represent 95% confidence intervals. (B) Minimal growth temperatures versus latitude. (C) Taxonomic differences of minimal growth temperatures at 08N. The circles represent average values for each group and the horizontal lines represent 95% confidence intervals. Chloro, chlorophytes. Cyano, cyanobacteria; Dino, dinoflagellates; Hapto, haptophytes; Raphi, raphidophytes. (D) Maximal growth temperatures versus latitude. (E) Thermal breadth versus latitude. The thick solid line represents a cubic spline fitted for a marine diatom with maximal growth rate of 1 d 21 (F) Thermal breadth versus maximal growth rate. The fitted line is for a marine diatom at 08N. (G) Taxonomic differences of thermal breadth when fixing latitude at 08N. (H) Estimated latitudinal patterns of thermal breadth based on modeled minimal and maximal growth temperatures for three marine phyla. 289

6 JOURNAL OF PLANKTON RESEARCH j VOLUME 37 j NUMBER 2 j PAGES j 2015 Fig. 3. Scatterplots of (A) optimal growth temperature versus environmental annual mean temperature, (B) minimal growth temperature versus environmental lowest temperature, (C) maximal growth temperature versus environmental highest temperature, and (D) thermal breadth versus environmental temperature range. Other symbols similar to Fig. 2. latitude (Fig. 2) rejects the null hypothesis and suggests that the environmental temperature of the isolation locale should be an important factor shaping phytoplankton physiology. It is noteworthy that this evident pattern exists for phytoplankton taxa that have been possibly cultivated in the laboratory after multiple generations, suggesting that the thermal characteristics can be inherited for a relatively long period. We need to recognize that the thermal limits obtained from temperature tolerance curves represent the most extreme temperatures that phytoplankton species can tolerate; while in nature, phytoplankton experience high mortality due to grazing, sinking and viral attack (Lehman and Sandgren, 1985; Calbet and Landry, 2004), which suggests that, even at the lowest or highest environmental temperatures, the intrinsic growth rate of phytoplankton should be positive to balance mortality. This can well explain why we observe that the thermal breadths tend to encompass the environmental temperature range (Fig. 3). It is unclear why phytoplankton do not follow Jenzen s rule in the same way as terrestrial organisms (Addo- Bediako et al., 2000; Deutsch et al., 2008; Gaston et al., 2009). One speculation might be because the highest temperature latitude relationship differs between the land and the ocean. The maximal environmental temperatures are relatively invariant with latitude on the land so that environmental temperature ranges broaden at high latitudes (Addo-Bediako et al., 2000). However, this is not the case for the ocean, which has large heat storage, and the temperature variations are less abrupt than on the land. Another factor might be also related with the high mortality rate of phytoplankton in the ocean, in which environmental temperature extremes are only correlated with, but do not directly determine, the thermal limits of phytoplankton. We can see from Fig. 2 that the maximal growth temperatures seem to decrease more rapidly than minimal temperatures with increasing latitude, leading to the plausible pattern that thermal breadth of marine phytoplankton 290

7 B. CHEN j THERMAL LIMITS OF PHYTOPLANKTON decreases with latitude (Fig. 2F). However, the random effects and errors cannot be overlooked (Table I). This pattern needs to be corroborated by more data and requires a reasonable explanation if it really exists. The positive correlation between maximal growth rate and thermal breadth of marine phytoplankton, which holds both within and among phyla, suggests that fast-growing species have the capacity to tolerate a wider range of temperatures. One possible explanation might be that the growth rate is determined by the concentrations of the ratelimiting enzymes [e.g. the ribulose-1,5-bisphosphate carboxylase/oxygenase (RubisCO) enzyme]. A fast-grower tends to have higher concentrations of the key enzyme. Compared with the slower growers, even if the same proportion of the enzyme is denatured at extreme temperatures, it still has more enzymes to support positive growth. Significant differences of minimal growth temperature and thermal breadth are found between diatoms and other algae including cyanobacteria and dinoflagellates (see also Moore et al., 1995; Staal et al., 2003). It is well known that diatoms thrive in cold, turbulent and nutrient-rich environments, and dinoflagellates and cyanobacteria are well adapted to warm and stratified oligotrophic conditions (Margalef, 1978; Li, 2002, 2009). While it is often believed that resource availability plus grazing impact play the dominant role in determining phytoplankton community structure and the effects of temperature are marginal or even nonexistent, sometimes the effect of temperature in modulating the interactions between species cannot be overlooked (Tilman et al., 1981). It is also possible that, for diatoms, the ability to tolerate cold environments is an evolutionary outcome of long-term adaptation to the nutrient-rich waters that are also cold at the same time. There are remarkable differences between freshwater and marine taxa. Much of the difference relates to one study (Tang et al., 1997), which has already noted the differences of thermal tolerance between polar freshwater cyanobacteria and marine diatoms. They attribute this difference to the relative stable marine environment and variable freshwater environment. They also speculate that the polar freshwater algae might have been dispersed into the polar areas so that their thermal characteristics are similar to the temperate taxa. This remains to be further investigated. The patterns reported in this study are useful for modeling phytoplankton diversity and functional groups (Smith et al., 2014). Many ecosystem models do not take into account the thermal limits of phytoplankton growth. In a global modeling project, Follows et al. (Follows et al., 2007) generate random temperature growth curves for a number of phytoplankton taxa. This approach is powerful but requires high computing resources. The patterns found in this study can be used to simplify the model, with the implicit assumption that phytoplankton can acclimate/adapt to the thermal environment. The adaptive capacity of phytoplankton is not limited to temperature. For example, the positive correlations between half-saturation constants of nitrate uptake and ambient nitrate concentrations suggest phytoplankton adaptation to the nutrient conditions (Collos et al., 2005; Smith et al., 2009). Therefore, modelers should take into account the adaptive capacities of organisms and be careful of assigning constants to represent traits of organisms across large scales (Smith et al., 2011). One important difference between laboratory data and the field situation is that the growth of phytoplankton in situ is often nutrient-limited. While the complex nutrient effects on the temperature sensitivity of phytoplankton growth are beyond the scope of this study that mainly focuses on the thermal limits of phytoplankton growth, it is noteworthy that there are very few data of thermal limits of in situ phytoplankton assemblages. Hence, at present we have to rely on the laboratory data to obtain the patterns of phytoplankton thermal limits. More studies are needed to relate the species temperature tolerance curves with in situ distributional ranges. ACKNOWLEDGEMENTS I sincerely thank one anonymous reviewer and W. K. W. Li for constructive comments on an earlier version of the manuscript. FUNDING This work was supported by National Basic Research Program ( 973 Program) of China (2015CB954003), National Science Foundation of China ( and ) and Program for New Century Excellent Talents of Xiamen University. REFERENCES Addo-Bediako, A., Chown, S. L. and Gaston, K. J. (2000) Thermal tolerance, climatic variability and latitude. Proc. R. Soc. B Biol. Sci., 267, Behrenfeld, M. J., O malley, R. T., Siegel, D. A. et al. (2006) Climate-driven trends in contemporary ocean productivity. Nature, 444, Bissinger, J. E., Montagnes, D. J., Sharples, J. et al. (2008) Predicting marine phytoplankton maximum growth rates from temperature: Improving on the Eppley curve using quantile regression. Limnol. Oceanogr., 53, 487. Brown, J. H. (1995) Macroecology, The University of Chicago Press, Chicago. 291

8 JOURNAL OF PLANKTON RESEARCH j VOLUME 37 j NUMBER 2 j PAGES j 2015 Brown, J. H., Gillooly, J. F., Allen, A. P. et al. (2004) Toward a metabolic theory of ecology. Ecology, 85, Calbet, A. and Landry, M. R. (2004) Phytoplankton growth, microzooplankton grazing, and carbon cycling in marine systems. Limnol. Oceanogr., 49, Collos, Y., Vaquer, A. and Souchu, P. (2005) Acclimation of nitrate uptake by phytoplankton to high substrate levels. J. Phycol., 41, Deutsch, C. A., Tewksbury, J. J., Huey, R. B. et al. (2008) Impacts of climate warming on terrestrial ectotherms across latitude. Proc. Natl Acad. Sci. USA, 105, Eppley, R. W. (1972) Temperature and phytoplankton growth in the sea. Fishery Bull., 70, Finkel, Z. V., Beardall, J., Flynn, K. J. et al. (2010) Phytoplankton in a changing world: cell size and elemental stoichiometry. J. Plankton Res., 32, Follows, M. J., Dutkiewicz, S., Grant, S. et al. (2007) Emergent biogeography of microbial communities in a model ocean. Science, 315, Gaston, K. J., Chown, S. L., Calosi, P. et al. (2009) Macrophysiology: a conceptual reunification. Am. Nat., 174, Gregory, B., Christophe, L. and Martin, E. (2009) Rapid biogeographical plankton shifts in the North Atlantic Ocean. Global Change Biol., 15, Huey, R. B. and Hertz, P. E. (1984) Is a jack-of-all-temperatures a master of none? Evolution, 38, Lehman, J. T. and Sandgren, C. D. (1985) Species-specific rates of growth and grazing loss among freshwater algae. Limnol. Oceanogr., 30, Levins, R. (1964) Evolution in Changing Environments. Princeton University Press, Princeton. Li, W. K. W. (1980) Temperature adaptation in phytoplankton: cellular and photosynthetic characteristics. In Primary Productivity in the Sea. Springer, USA, pp Li, W. K. W. (1985) Photosynthetic response to temperature of marine phytoplankton along a latitudinal gradient (168N to748n). Deep-Sea Res I, 32, Li, W. K. W. (2002) Macroecological patterns of phytoplankton in the northwestern North Atlantic Ocean. Nature, 419, Li, W. K. W. (2009) From cytometry to macroecology: a quarter century quest in microbial oceanography. Aquat. Microb. Ecol., 57, López-Urrutia, Á., San Martin, E., Harris, R. P. et al. (2006) Scaling the metabolic balance of the oceans. Proc. Natl Acad. Sci. USA, 103, Margalef, R. (1978) Life-forms of phytoplankton as survival alternatives in an unstable environment. Oceanologica Acta, 1, Moore, L. R., Goericke, R. and Chisholm, S. W. (1995) Comparative physiology of Synechococcus and Prochlorococcus: influence of light and temperature on growth, pigments, fluorescence and absorptive properties. Mar. Ecol. Prog. Ser., 116, Pinheiro, J. C. and Bates, D. M. (2000) Mixed-effects Models in S and S-PLUS. Springer, New York, USA. R Core Team. (2014) R: A language and environment for statistical computing. R Foundation for Statistical Computing, Vienna, Austria. URL Rose, J. M. and Caron, D. A. (2007) Does low temperature constrain the growth rates of heterotrophic protists? Evidence and implications for algal blooms in cold waters. Limnol. Oceanogr., 52, Smith, S. L., Merico, A., Wirtz, K. W. et al. (2014) Leaving misleading legacies behind in plankton ecosystem modeling. J. Plankton Res., 36, Smith, S. L., Pahlow, M., Merico, A. et al. (2011) Optimality-based modeling of planktonic organisms. Limnol. Oceanogr., 56, Smith, S. L., Yamanaka, Y., Pahlow, M. et al. (2009) Optimal uptake kinetics: physiological acclimation explains the pattern of nitrate uptake by phytoplankton in the ocean. Mar. Ecol. Prog. Ser., 384, Staal, M., Meysman, F. J. and Stal, L. J. (2003) Temperature excludes N 2 -fixing heterocystous cyanobacteria in the tropical oceans. Nature, 425, Tang, E. P. Y., Tremblay, R. and Vincent, W. F. (1997) Cyanobacterial dominance of polar freshwater ecosystems: are high-latitude matformers adapted to low temperature? J. Phycol., 33, Thomas, M. K., Kremer, C. T., Klausmeier, C. A. et al. (2012) A global pattern of thermal adaptation in marine phytoplankton. Science, 338, Tilman, D., Mattson, M. and Langer, S. (1981) Competition and nutrient kinetics along a temperature gradient: An experimental test of a mechanistic approach to niche theory. Limnol. Oceanogr., 26, Walther, G.-R., Post, E., Convey, P. et al. (2002) Ecological responses to recent climate change. Nature, 416, Wood, S. N. (2006) Generalized Additive Models: An Introduction with R. Chapman and Hall, Boca Raton, Florida, USA. 292