Warming and eutrophication combine to restructure diatoms and dinoflagellates

Accepted Manuscript Warming and eutrophication combine to restructure diatoms and dinoflagellates Wupeng Xiao, Xin Liu, Andrew J. Irwin, Edward Laws, Lei Wang, Bingzhang Chen, Yang Zeng, Bangqin Huang PII:

S0043-1354(17)30885-0

DOI:

10.1016/j.watres.2017.10.051

Reference:

WR 13308

To appear in:

Water Research

Received Date: 14 June 2017 Revised Date:

27 September 2017

Accepted Date: 23 October 2017

Please cite this article as: Xiao, W., Liu, X., Irwin, A.J., Laws, E., Wang, L., Chen, B., Zeng, Y., Huang, B., Warming and eutrophication combine to restructure diatoms and dinoflagellates, Water Research (2017), doi: 10.1016/j.watres.2017.10.051. This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT

(a) 60% stations show an decrease (blue)

(b) 70% stations show an increase (red)

120

100

28

0

50

25

27

200 150 000 e 1.5 1 200 2000 3.5 e 50 6.5 00 e 2 ng/L 500 25 00 500 0 2000 5 0 300 121 122 4123 000 124 125 126 127 Longitude (°E)

0

25

10

10

25

AC C

200 150 000 e 1.5 1 200 2000 3.5 e 50 6.5 00 e 2 ng/L 500 25 00 500 0 2000 5 0 300 121 122 4123 000 124 125 126 127 Longitude (°E)

0

26 25

50

150

26

0

50

EP

27

150

25

100

TE 25 D

31

50

Latitude (°N) 28 29 30 25

M 29 30 31 32 AN US CR IP T

Peridinin

32

Fucoxanthin

25

120

ACCEPTED MANUSCRIPT 1

Warming and eutrophication combine to restructure

2

diatoms and dinoflagellates Wupeng Xiao,ab†

3

Xin Liu,ab†

Andrew J. Irwin,c

Lei Wang,ab

Bangqin Huangab*

RI PT

Bingzhang Chen,e Yang Zeng,ab

4

Edward Laws,d

5 6

a

7

Xiamen, China.

8

b

9

Provincial Key Laboratory of Coastal Ecology and Environmental Studies, Xiamen

SC

State Key Laboratory of Marine Environmental Science, Xiamen University,

M AN U

Key Laboratory of Coastal and Wetland Ecosystems, Ministry of Education/Fujian

10

University, Xiamen, China.

11

c

12

Sackville, New Brunswick, Canada.

13

d

14

Louisiana State University, Baton Rouge, Louisiana 70803, USA.

15

e

16

Global Change, Japan Agency for Marine-Earth Science and Technology, Yokohama,

17

Japan.

TE D

Department of Environmental Sciences, School of the Coast & Environment,

EP

Ecosystem Dynamics Research Group, Research and Development Center for

AC C

18

Department of Mathematics & Computer Science, Mount Allison University,

19

†

20

* Corresponding author: Dr. Bangqin Huang, tel: 0592-2187783,

21

email: [email protected]

These authors contributed equally to this work.

1

ACCEPTED MANUSCRIPT Abstract

23

Temperature change and eutrophication are known to affect phytoplankton

24

communities, but relatively little is known about the effects of interactions between

25

simultaneous changes of temperature and nutrient loading in coastal ecosystems. Here

26

we show that such interaction is key in driving diatom-dinoflagellate dynamics in the

27

East China Sea. Diatoms and dinoflagellates responded differently to temperature,

28

nutrient concentrations and ratios, and their interactions. Diatoms preferred lower

29

temperature and higher nutrient concentrations, while dinoflagellates were less

30

sensitive to temperature and nutrient concentrations, but tended to prevail at low

31

phosphorus and high N:P ratio conditions. These different traits of diatoms and

32

dinoflagellates resulted in the fact that both the effect of warming resulting in

33

nutrients decline as a consequence of increasing stratification and the effect of

34

increasing terrestrial nutrient input as a result of eutrophication might promote

35

dinoflagellates over diatoms. We predict that conservative forecasts of environmental

36

change by the year 2100 are likely to result in the decrease of diatoms in 60% and the

37

increase of dinoflagellates in 70% of the surface water of the East China Sea, and

38

project that mean diatoms should decrease by 19% while mean dinoflagellates should

39

increase by 60% in the surface water of the coastal East China Sea. This analysis is

40

based on a series of statistical niche models of the consequences of multiple

41

environmental changes on diatom and dinoflagellate biomass in the East China Sea

42

based on 2815 samples randomly collected from 23 cruises spanning 14 years (2002–

43

2015). Our findings reveal that dinoflagellate blooms will be more frequent and

AC C

EP

TE D

M AN U

SC

RI PT

22

2

ACCEPTED MANUSCRIPT 44

intense, which will affect coastal ecosystem functioning.

45

Key words: Diatoms; dinoflagellates; warming; eutrophication; generalized additive

46

mixed models; the East China Sea

48

RI PT

47 1. Introduction

Anthropogenic eutrophication and global warming have dramatic impacts on

50

marine phytoplankton and are likely to continue for many centuries (Doney et al.

51

2012, Hallegraeff 2010, Karl and Trenberth 2003). Eutrophication is a major problem

52

in many coastal ecosystems and has resulted in large phytoplankton blooms and the

53

expansion of low-oxygen dead zones (Anderson et al. 2002, Edwards et al. 2006).

54

Warming affects phytoplankton in two fundamentally different ways: directly through

55

the effect of warming on metabolic rates and indirectly through physical mixing,

56

which affects nutrient availability (Lewandowska et al. 2014). Simultaneous warming

57

and eutrophication will have complex consequences because of interactions between

58

the processes that are driven by nutrient supply and metabolic rates. There is

59

accumulating evidence that the intensity of the response differs considerably across

60

taxa; these changes may therefore cause a shift in phytoplankton community structure

61

(Barton et al. 2016, Edwards and Richardson 2004, Irwin et al. 2012, Wells et al.

62

2015, Xie et al. 2015).

AC C

EP

TE D

M AN U

SC

49

63

Within the phytoplankton community, the consequences for diatoms and

64

dinoflagellates are a major concern because these taxa play key roles in ecosystem

65

processes and form the basis of many aquatic food webs (Agusti et al. 2014, 3

ACCEPTED MANUSCRIPT Menden-Deuer and Lessard 2000). Diatoms and dinoflagellates are two typical groups

67

that form harmful phytoplankton blooms, which can adversely affect human health as

68

well as marine fisheries and aquaculture (Anderson et al. 2002, Moore et al. 2008). In

69

fact, dinoflagellates and diatoms account for 75% and 5% of all harmful

70

phytoplankton species, respectively (Smayda and Reynolds 2003). Various theoretical

71

arguments based on laboratory experiments and field studies suggest that

72

phytoplankton bloom dynamics and habitat preferences can be described in terms of

73

nutrient availability, degree of mixing, and/or nutrient/resource ratios (Margalef 1977,

74

Reynolds 1987, Smayda and Reynolds 2001, 2003, Tilman 1977). All of these factors

75

are closely correlated with coastal eutrophication and surface layer warming. While

76

numerous studies have shown that the relative abundance of diatoms and

77

dinoflagellates varies in response to regional climate warming, contrasting results can

78

be obtained (Hinder et al. 2012, Leterme et al. 2005, 2006, Xie et al. 2015), because

79

the net effect of warming depends on other environmental changes. Quantitative

80

predictive models that consider the effects of temperature and nutrient interactions on

81

the concentrations of diatoms and dinoflagellates are still in their infancy for the lack

82

of sufficient real-time data (Glibert et al. 2010, Moore et al. 2008, Wells et al. 2015).

83

In this study we developed a series of statistical niche models to explore the existing

84

relationships between the diatom and dinoflagellate biomass and environmental

85

factors, and to project the biomass of diatoms and dinoflagellates in response to future

86

environmental changes using an extensive dataset of coastal phytoplankton

87

community composition and water quality from the entire euphotic zone in the East

AC C

EP

TE D

M AN U

SC

RI PT

66

4

ACCEPTED MANUSCRIPT 88

China Sea (ECS), a region that is affected by both eutrophication and warming and

89

that is important to biogeochemical cycles and fisheries. Previous studies have demonstrated that diatoms are dominant during much of the

91

year in the ECS, but dinoflagellates become important during late spring blooms (Guo

92

et al. 2014). Long-term studies have found that the relative biomass of diatoms

93

compared to dinoflagellates has decreased during a time of increasing coastal

94

eutrophication (Jiang et al. 2014, Tang et al. 2006, Zhou et al. 2008). Our earlier work

95

revealed that differences in the responses of diatoms and dinoflagellates to

96

temperature may result in the earlier occurrence of dinoflagellate blooms (Liu et al.

97

2016). Here, we further propose that warming combined with eutrophication may

98

have a great impact on the dynamics of diatoms and dinoflagellates in the ECS.

M AN U

SC

RI PT

90

TE D

99 100

2. Materials and methods

101

2.1. Data sources

Available material makes the ECS an ideal region for studying the effects of

103

environmental changes on phytoplankton community dynamics (see SI 1). We

104

conducted 23 cruises spanning 14 years (2002–2015) and obtained 2815 pigment

105

samples in the ECS (Fig. 1, Table S1). The cruises took place in all four seasons and

106

during most months of the year. Stations were located irregularly in space and were

107

not all the same between cruises. Data density was greatest near the Changjiang River

108

estuary and adjacent coastal waters (Fig. 1).

109

AC C

EP

102

Temperature, salinity and water pressure were determined at every station from 5

ACCEPTED MANUSCRIPT casts of a Sea Bird CTD fitted with a Sea Tech fluorometer. Using Niskin bottles

111

attached on the CTD rosette system, seawater samples were taken in the euphotic

112

zone (≤ 150 m) from at least three layers: the surface water (≤ 5 m), the subsurface

113

chlorophyll a (Chla) maximum, and the bottom of the euphotic zone. The euphotic

114

zone was defined as the water column down to the depth at which the downward

115

photosynthetically active radiation (PAR) was 1% of the value just below the surface.

116

PAR was made using a free-fall spectroradiometer (SPMR, Satlantic) just prior to

117

water sampling. Nutrient concentrations were determined by professional laboratories

118

(Han et al. 2012, Liu et al. 2016). To avoid issues with detection limits, nutrient

119

concentrations below the detection limit were deemed as the value of detection limit.

120

Samples for phytoplankton pigment analysis were collected at every station on all

121

cruises. Pigment concentrations were measured by high performance liquid

122

chromatography (HPLC) following the method presented by Furuya et al. (2003) and

123

are reported in detail by Liu et al. (2012).

TE D

M AN U

SC

RI PT

110

The long-term sea surface Chla (mg m–3) from August 1997 to January 2015 was

125

extracted from multiple-satellite products (AVHRR, QuikSCAT, SeaWiFS and

126

MODIS/Aqua). The resolutions of these four databases were 0.044, 0.25, 0.1 and 0.05

127

degrees respectively.

128

2.2. Generalized additive mixed models

AC C

EP

124

129

To explore how temperature and nutrients influence the dynamics of diatoms and

130

dinoflagellates, we developed a series of generalized additive mixed models

131

(GAMMs), which are an extension of GAMs (see SI 2) (Wood 2006, Zuur et al. 2009). 6

ACCEPTED MANUSCRIPT 132

Fucoxanthin (Cf) and peridinin (Cp) concentration were used as response variables.

133

Fucoxanthin and peridinin are the diagnostic photosynthetic pigments of diatoms and

134

autotrophic peridinin-containing dinoflagellates, respectively (Roy et al. 2011).

135

Phytoplankton

136

concentration is widely used because it is easier to obtain high resolution data relative

137

to microscopic and molecular methods (Peloquin et al. 2013).

structure

determined

from

diagnostic

pigment

RI PT

community

Temperature (°C), salinity (psu), nitrate+nitrite (NOx, µmol L–1), soluble reactive

139

phosphorus (PO4, µmol L–1) and the ratio of NOx and PO4 (N:P) were taken as

140

potential environmental explanatory variables. Silicate (SiO3, µmol L–1) was not

141

incorporated into the models for reasons that silicate is usually not limiting in the ECS

142

(Guo et al. 2014, Zhou et al. 2008) and that its effect co-varied with NOx and/or PO4

143

during our initial test. TChla (the sum of Chla plus divinyl chlorophyll a, ng L–1),

144

geographic and temporal parameters (longitude, latitude, depth, month) were

145

incorporated into the models as proxies for some missing parameters driving

146

phytoplankton dynamics (e.g., grazing, light). In order to account for the potential

147

influence of the different measuring instruments, we added a nominal variable

148

representing the measurement effect for each group of cruises (Table S1). Of all these

149

variables, we focus on the results of temperature and nutrient parameters, which are

150

most likely to change in the future.

AC C

EP

TE D

M AN U

SC

138

151

Before modeling, we used Cleveland dot plots and boxplots to check for potential

152

outliers in both response and explanatory variables and used pairplots to check for

153

collinearity between the explanatory variables. After quality control we had 2809 7

ACCEPTED MANUSCRIPT 154

observations of Cf and 1998 observations of Cp remaining. Cf, Cp, TChla, NOx, PO4

155

and N:P ratio were ln transformed (Figs. S1 and S2, see SI 2). To reduce collinearity, we built a set of models (Models F1–F3, P1–P3) that use

157

NOx as the metric of nutrient availability, a similar set of models (Models F4–F5, P4–

158

P5) that use PO4 for comparison, and a set of models that use the N:P ratio instead

159

(Models F6–F7, P6–P7), among which a set of models (Models F3, P3, F5, P5, F6, P6)

160

treated nutrient parameters as independent variables and another set of models

161

(Models F1–F2, P1–P2, F4, P4, F7, P7) assumed interactions between temperature

162

and nutrients (Table S2).

SC

M AN U

163

RI PT

156

The model without interaction between temperature and nutrient (NOx, PO4, or N:P) was:

165

lnY =  +  + s longitude, latitude + s temperature + s nutrient +

166

s TChl + s salinity + s depth + s month + 

167

while the model with temperature and nutrient interaction was:

168

lnY =  +  + s longitude, latitude + te temperature, nutrient + s TChl +

169

s salinity + s depth + s month + 

170

where Y represents Cf or Cp and i denotes measuring instrument. The term α is a grand

171

mean. The term ɑi is a random intercept assumed to be normally distributed with

172

mean 0 and variance σɑ2, which allows for random variation around the intercept and

173

reflects between measurement variability. The term s(﹒) indicates a one-dimensional

174

nonlinear function based on cubic regression spline and te(﹒) a two dimensional

175

effect based on tensor product spline. The two models are nested because

EP

TE D

164

(1)

AC C

(2)

8

ACCEPTED MANUSCRIPT te(temperature, nutrient) contains s(temperature), s(nutrient), and their interaction.

177

The term ε represents within measurement variability, and it is assumed to

178

independently, normally distributed with variance σ2. Furthermore, the σ2 is assumed

179

allowing heterogeneity between months as the homogeneity assumption was violated

180

along the month effect (Fig. S3, see SI 3).

181

2.3. Generalized linear models

RI PT

176

To simulate the effect of interactions of temperature and nutrients on the relative

183

biomass of diatoms and dinoflagellates, we formulated a more robust generalized

184

linear model (GLM). In this model, only samples of the upper 15 m were used. The

185

samples were divided into 64 bins on the basis of every 2.5°C in temperature and

186

every 1 interval in ln transformed PO4 concentrations. The medians of each parameter

187

in each bin were extracted instead of the raw data. After removing those bins with less

188

than 3 samples, we had 37 available bins remaining.

TE D

M AN U

SC

182

We used the ratio of peridinin/fucoxanthin (Cp/Cf) instead of their absolute

190

concentrations as the response variable. Temperature, NOx and PO4 concentrations

191

were taken as the explanatory variables. The model was formulated as follows:

192

 ⁄ = ∑&#'( ∑&$'( ∑&%'( "#$% temperature# NO+ $ PO& % ,  + - + . ≤ 4

193

which produced 25 independent variables (Table S4). The term "#$% is the

194

coefficient of each variable. It will be the intercept when a, b and c equate zero at the

195

same time. We assumed a gamma distribution with log-link function to keep the fitted

196

ratio of Cp/Cf more than zero. Model optimization was based on the AIC value and the

197

t-statistic, i.e., dropping the variables that did not reduce the AIC value significantly

AC C

EP

189

9

(3)

ACCEPTED MANUSCRIPT 198

step by step, and dropping those variables that did not significantly correlated with

199

Cp/Cf judged by the t-statistic. Model simulations were conducted by assuming possible changes in temperature

201

and nutrients concentration. The actual data showed the NOx concentration was

202

linearly correlated with the PO4 concentration, but the slope was dependent on the

203

PO4 concentration (Fig. S4a). In addition, the N:P ratio was nonlinearly correlated

204

with temperature (Fig. S4b). We took this into account by linking NOx as the function

205

of PO4 and temperature. The model formation was built as follows:

206

ln NO+  =  + "1 × ln PO&  + s temperature

207

where α is the intercept and β the slope of ln(PO4). The s(temperature) term means a

208

cubic regression spline for the temperature. The model accounted for 93% of the

209

variance of logarithms of NOx concentration. Then we calculate NOx from PO4 and

210

temperature based on this equation.

(4)

TE D

M AN U

SC

RI PT

200

All analyses were done using R in version 3.2.2 (R Development Core Team 2016).

212

The GAMMs and GAMs were fitted with the ‘mgcv’ (GAMs with GCV smoothness

213

estimation and GAMMs by REML/ML (Wood 2006)) package.

AC C

214

EP

211

215

3. Results

216

3.1. Long-term variations of phytoplankton in the ECS

217

Long-term sea surface Chla concentration in our study area from multiple-satellite

218

products displayed a clear seasonal cycle in each year (Fig. 2a). The maximum in

219

Chla occurred in spring followed by another peak in summer. The magnitude of the 10

ACCEPTED MANUSCRIPT spring algal bloom appeared to increase since 2000, except the last two years (Fig. 2a).

221

Monthly averaged TChla concentrations collected from our field observations

222

matched well with satellite derived Chla (r = 0.55, p < 0.01 after removing four

223

observations with TChla concentration greater than 2.5 mg m–3) given the uncertainty

224

of satellite data and irregularity of our sampling areas. The ratio of surface

225

peridinin/fucoxanthin increased during the last 10 years (Fig. 2b). High fucoxanthin

226

occurred in early spring and summer (Fig. S3a); high peridinin values were found in

227

late spring and summer (Fig. S3b).

228

3.2. Partial effects of temperature and nutrients on Cf and Cp

M AN U

SC

RI PT

220

Temperature was significantly correlated with both Cf and Cp with different

230

relationships (Fig. 3a,b). After keeping other parameters constant, the partial effect of

231

temperature on Cf was significantly positive below ~24°C and strongly negative

232

above ~24°C. There was a plateau of high Cf at temperatures of 20–24°C (Fig. 3a). Cp

233

was less sensitive to temperature than Cf (Fig. 3b). Cp was positively correlated with

234

temperature in the approximate range 13–26°C, insensitive to temperature at

235

temperatures of 10–13°C, and negatively correlated with temperature at temperatures

236

of 26–30°C (Fig. 3b). The optimum temperature for Cf was roughly 24°C versus 26°C

237

for Cp, and the decline of Cf at temperatures exceeding 24°C was much more

238

precipitous than the analogous decline of Cp at temperatures exceeding 26°C (Fig.

239

3a,b). These results suggest that peridinin-containing dinoflagellates are better able to

240

tolerate high temperatures than fucoxanthin-containing phytoplankton.

241

AC C

EP

TE D

229

Both Cf and Cp responded nonlinearly to nutrient concentrations (Fig. 3c–f). NOx 11

ACCEPTED MANUSCRIPT was positively correlated with Cf at NOx concentrations less than ~6 µmol L–1 (Fig.

243

3c). Cp was positively correlated with NOx at NOx concentrations less than ~1 µmol

244

L–1 but negatively correlated thereafter (Fig. 3d). When PO4 was used as the

245

independent variable instead of NOx, its effect on Cf was qualitatively similar to that

246

of NOx; Cf and PO4 were positively correlated at PO4 concentrations below ~0.3 µmol

247

L–1, but there was no correlation at higher PO4 concentrations (Fig. 3e). There was no

248

highly significant correlation (t-test, p = 0.2) between PO4 and Cp across the range of

249

PO4 concentrations in the database, ~0.01 µmol L–1 to ~2.7 µmol L–1 (Fig. 3f), the

250

indication being that Cp may not be sensitive to phosphate concentrations. These

251

trends also show that optimal NOx and PO4 concentrations for Cf are much higher

252

than those for Cp, the clear indication being that peridinin-containing dinoflagellates

253

tend to dominate at nutrient concentrations lower than those associated with

254

fucoxanthin-containing phytoplankton dominance.

TE D

M AN U

SC

RI PT

242

Cf and Cp have an opposite relationship with N:P ratio (Fig. 3g,h). The relationship

256

between N:P and Cf was unimodal, with maximum Cf values at N:P ratio of ~16 (Fig.

257

3g). The correlation between N:P and Cp was positive across the observed range of

258

N:P ratios (0.2 to 300), but there was a local minimum at a N:P of ~16 (Fig. 3h).

259

3.3. Combined effects of temperature and nutrients on Cf and Cp

AC C

260

EP

255

The interactions between temperature and

nutrient concentrations were

261

significantly correlated with both Cf and Cp (p < 0.0001, Table S3) and inclusion of

262

these interactions produced statistically better results than the models assuming no

263

interaction (Table S2, see F2 (P2) versus F3 (P3), F4 (P4) versus F5 (P5), and F6 (P6) 12

ACCEPTED MANUSCRIPT versus F7 (P7)). The nature of the interaction between temperature and NOx was clear

265

in the case of Cf (Fig. 4a). The negative correlation between Cf and temperature

266

estimated from the partial temperature effect (Fig. 3a) for temperatures above 20°C

267

was apparent, but the magnitude of the slope was negatively correlated with NOx

268

concentrations

269

phytoplankton are better able to tolerate supraoptimal temperatures at relatively high

270

NOx concentrations. The response of Cp to the interaction of temperature and NOx,

271

however, was rather complex (Fig. 4d). Cp remained relatively low but with two

272

obvious peaks at temperatures of ~18°C and ~25°C, and the corresponding NOx

273

concentrations were both around 1 µmol L–1, which is consistent with the partial NOx

274

effect (Fig. 3d).

4a).

The

implication

is

that

fucoxanthin-containing

M AN U

SC

(Fig.

RI PT

264

The interactive model of the effects of temperature and PO4 on Cf produced results

276

qualitatively similar to the results of the interactive model with temperature and NOx,

277

although the response surface was more complex (Fig. 4b). The effect of interactions

278

between temperature and PO4 on Cp (Fig. 4e) was quite different from that of

279

temperature and NOx. The model results showed high Cp values in some regions

280

where Cf values were low (Fig. 4b,e). At the lowest PO4 concentrations, Cp increased

281

dramatically with increasing temperature. At intermediate PO4 concentrations, there

282

was a maximum Cp at a temperature of ~18°C; at higher PO4 concentrations, Cp was

283

relatively insensitive to temperature. The results of these models (Models P2, P4 and

284

P6, Table S2) imply that the relationship between Cp and temperature depends on the

285

PO4 concentration (Fig. 4e).

AC C

EP

TE D

275

13

ACCEPTED MANUSCRIPT Use of N:P and temperature as independent variables revealed relationships with Cf

287

and Cp that were less complex than was the case when nutrient availability was

288

quantified in terms of nutrient concentrations (Fig. 4c,f). The temperature-N:P models

289

suggested that Cf would exceed e4 = 55 under most conditions (Fig. 4c), whereas Cp

290

exceeded e3 = 20 mostly at temperatures above 22°C and N:P higher than 100 (Fig.

291

4f). The highest Cf values were associated with N:P ratios of roughly 16 and

292

temperatures of 20–24°C (Fig. 4c). In contrast, relatively high Cp values were

293

estimated at around 18°C and between 25–28°C when N:P ratios are low and at

294

temperatures upon 18°C when N:P ratios are high (Fig. 4f). The implication of the

295

model is that the interaction of temperature and the relative concentration of nutrients

296

plays

297

fucoxanthin-containing phytoplankton and peridinin-containing dinoflagellates in the

298

ECS.

299

3.4. Relations of the Cp/Cf ratio to interactions of temperature and nutrient

300

stoichiometry

role

in

the

outcome

of

competition

between

TE D

important

EP

an

M AN U

SC

RI PT

286

The Cp/Cf ratio showed higher values at ~18°C and at above 26°C (Fig. 5). There

302

was a clear interaction between temperature and PO4 concentration (Fig. 5a). A

303

negative correlation was found between temperature and the Cp/Cf ratio at PO4

304

concentrations greater than 0.8 µmol L–1. At PO4 concentrations between 0.4 µmol L–1

305

and 0.8 µmol L–1 there was a positive correlation between temperature and the Cp/Cf

306

ratio. At PO4 concentrations less than 0.4 µmol L–1 there was a minimum of the Cp/Cf

307

ratio at intermediate temperatures. The ratio became higher at lower/higher

AC C

301

14

ACCEPTED MANUSCRIPT temperature end members. The Cp/Cf ratio was also correlated with the interaction of

309

temperature and NOx concentration. It is clear that the higher ratios at ~18°C occurred

310

at NOx concentrations ranging from 1–5 µmol L–1, and the higher ratios at higher

311

extreme temperatures occurred at NOx concentrations less than 5 µmol L–1 (Fig. 5b).

312

The implication is that the succession of fucoxanthin-containing phytoplankton and

313

peridinin-containing dinoflagellates do depend on the absolute concentration of the

314

limiting nutrient, while which nutrient is limiting depends on the ambient N:P ratio.

SC

RI PT

308

Based on the relationships between the Cp/Cf ratio and the interactions of

316

temperature and nutrient concentrations, we formulated a GLM model (equation (3),

317

Table S4). This model accounted for 97% of the variance of the Cp/Cf ratio, with the

318

residuals showing no heterogeneity with respect to temperature (Fig. S5). Then we

319

simulated the relationship of the Cp/Cf ratio and temperature at three constant PO4

320

concentrations, 0.05 µmol L–1, 0.25 µmol L–1, and 0.6 µmol L–1, based on this model.

321

When we assumed the molar N:P ratio to be constant and equal to 16:1, the

322

simulations presumably reflected the trends in the actual data at PO4 concentrations

323

less than 0.25 µmol L–1, but deviated from the actual data at higher PO4

324

concentrations (Fig. 5a and Fig. S6a). We therefore assumed that the N:P ratio varied

325

from 8 to 32. Comparing with the actual data, we found that different temperatures

326

selected for different optimum N:P ratios; lower temperature selected for a higher N:P

327

ratio, whereas higher temperature selected for a lower N:P ratio (Fig. S6b). The

328

explanation is that in the ocean the dissolved N:P ratio is not constant. The data

329

showed that the NOx concentration was positively correlated with the PO4

AC C

EP

TE D

M AN U

315

15

ACCEPTED MANUSCRIPT concentration, and N:P ratio variability was dependent on both the PO4 concentration

331

and the temperature (Fig. S4a,b). We therefore calculated NOx from PO4 and

332

temperature through equation (4) and repeated the simulations (Fig. 6).

333

3.5. Future responses of Cf and Cp to warming and eutrophication

RI PT

330

Future changes in the Cf and Cp as a result of warming and eutrophication are not

335

straightforward to anticipate because the effects of temperature, nutrients, and their

336

interaction are all significant. Because we obtained clear relationships of the

337

interaction of temperature and N:P ratio to Cf and Cp, using GAMM models F6 and P6

338

(Table S2), we simulated surface Cf and Cp concentrations based on assumed surface

339

temperatures increases of 2°C and N:P ratio increases of e2 = 7.4 while leaving other

340

parameters constant. The predicted Cf concentrations decreased significantly (t-test, p

341

< 0.05), with 60% stations showing decreased concentrations (Fig. 7a). In contrast,

342

the predicted Cp increased significantly (t-test, p < 0.05), with 70% of the stations

343

showing an increase (Fig. 7b). The magnitude of these changes was a function of

344

latitude and longitude. Most of the big changes occurred in coastal regions, especially

345

at the mouth of the Changjiang River estuary and adjacent coastal waters. The

346

changes decreased in magnitude in a southeast offshore direction (Fig. 7).

M AN U

TE D

EP

AC C

347

SC

334

We then focused on the surface waters of coastal regions with salinities less than 31.

348

We predicted a significant decrease of Cf (Fig. 8a) and a significant increase of Cp

349

(Fig. 8b) when temperature and ln N:P ratio were simultaneously increased by 0.25°C

350

and 0.25, respectively. The present surface median temperature and N:P ratio in these

351

regions are 23.0°C and e3.19 = 24, respectively. Given an increase of 2 units, the values 16

ACCEPTED MANUSCRIPT would increase to 25°C and e5.19 = 180 (~7 fold increase), respectively.

353

Correspondingly, we predicted that the mean Cf should decrease by 19% (from 508 ng

354

L–1 to 412 ng L–1) while the mean Cp should increase by 60% (from 334 ng L–1 to 535

355

ng L–1).

RI PT

352

356 357

4. Discussion

Phytoplankton that contain peridinin are found only in the class Dinophyceae (Roy

359

et al. 2011). Although not all autotrophic Dinophyceae contain peridinin (Zapata et al.

360

2012), the dominant species, Prorocentrum donghaiense, Alexandrium catenella,

361

Alexandrium minutum and Gymnodinium catenatum, in the ECS do contain (Guo et al.

362

2014, Zapata et al. 2012). Fucoxanthin (fuco), however, is found in phytoplankton

363

belonging to at least some species in five phytoplankton classes: Chrysophyceae,

364

Pelagophyceae, Prymnesiophyceae, Bacillariophyceae, and Dinophyceae (Roy et al.

365

2011). Fucoxanthin is commonly associated with diatoms (i.e., Bacillariophyceae)

366

(Letelier et al. 1993), but the presence in some of our samples of pigments such as

367

19´-butanoyloxyfucoxanthin (but) and 19´-hexanoyloxyfucoxanthin (hex), which are

368

not associated with diatoms but are found in some species of Pelagophyceae,

369

Prymnesiophyceae, and Dinophyceae, indicates that diatoms did not account for all

370

the fucoxanthin in our samples. However, the fuco/but and fuco/hex ratios in our

371

samples averaged 15 and 6, respectively, whereas the fuco/but and fuco/hex ratios in

372

Dinophyceae and Prymnesiophyceae, for example, are only about 0.5 and 0.02,

373

respectively (Chang and Gall 2013, Letelier et al. 1993), thus it seems unlikely that

AC C

EP

TE D

M AN U

SC

358

17

ACCEPTED MANUSCRIPT the fucoxanthin in our samples reflects high biomass of Dinophyceae and

375

Prymnesiophyceae. We do have some cases where the concentrations of but and hex

376

exceed the fucoxanthin concentrations. Those cases presumably are Chrysophyceae,

377

Prymnesiophyceae, Pelagophyceae, and maybe some of the Dinophyceae that contain

378

no peridinin. However, those are all cases where the fucoxanthin concentrations are

379

low (Fig. S7). Another possible source of fucoxanthin may be Phaeophyta, but such

380

macroalgae are usually benthic and unlikely to be contained in our water samples. In

381

our previous study, diagnostic pigment concentration in the ECS was comparable to

382

the microscopy results based on morphology (Liu et al. 2016). Thus we feel confident

383

in assuming that most of the fucoxanthin and peridinin in our samples was associated

384

with diatoms and dinoflagellates, respectively, and that all of the high fucoxanthin

385

concentrations were the result of diatom blooms.

TE D

M AN U

SC

RI PT

374

Based on the raw data, our GAMM model revealed that Cf exceeded Cp in the ECS

387

when NOx and PO4 concentrations were relatively high (Fig. 3c,e). The Cp

388

concentrations were relatively high when PO4 concentrations were less than 0.1 µmol

389

L–1 but NOx concentrations were higher than 0.5 µmol L–1 (Fig. 3d,f). These patterns

390

in fact seem reasonable if we assume that most of the fucoxanthin was associated with

391

diatoms. Diatoms tend to have a competitive advantage when nutrient concentrations

392

are high for several reasons. Diatoms are typically R-strategists (species tolerant of

393

shear/stress forces in physically disturbed water masses) (Alves-de-Souza et al. 2008)

394

and tend to have high maximum carbon-specific nutrient uptake rates and

395

disproportionately large storage vacuoles (Edwards et al. 2012, Litchman and

AC C

EP

386

18

ACCEPTED MANUSCRIPT Klausmeier 2008). Dinoflagellates, however, may replace diatoms by using multiple

397

growth strategies (C, colonist species which predominate in chemically disturbed

398

habitats; S, nutrient stress tolerant species; and R) to exploit abiotic habitats that

399

support them (Smayda and Reynolds 2003). Dinoflagellate have an intrinsically high

400

affinity for phosphate and the ability to utilize organic phosphorus compounds via

401

alkaline phosphatase (e.g., P. donghaiense) (Huang et al. 2005), as well as

402

mixotrophic and vertical migration capabilities. As a group, dinoflagellates are not

403

very sensitive to nutrient conditions.

M AN U

SC

RI PT

396

We found that lower temperatures favored diatoms and that higher temperatures

405

inhibited the growth of diatoms. That dinoflagellates were less sensitive to

406

temperature than diatoms (Fig. 3a,b) is inconsistent with a previous found that the

407

adaptive temperature range of diatoms (Skeletonema costatum, 15–25°C) is wider

408

than that of dinoflagellates (P. donghaiense, 20–25°C) (Wang et al. 2006) and the

409

global scenario that diatoms have a wider thermal tolerance than dinoflagellates

410

(Chen 2015). However, we found that the temperature effect was very dependent on

411

the nutrient regime, a consequence of either species specific variations or an interact

412

effect of the temperature and the nutrient regime on the same dominant species. The

413

result showed that diatoms can tolerate higher temperatures at high nutrient (NOx and

414

PO4) concentrations than at low nutrient concentrations (Fig. 4a,c). The implication is

415

that the response of diatoms to high temperature may not be a consequence of an

416

inherent metabolic intolerance to supraoptimal temperatures as long as the

417

temperature stress is not accompanied by the additional stress of nutrient limitation.

AC C

EP

TE D

404

19

ACCEPTED MANUSCRIPT Thermal stratification of marine waters will increase as a consequence of global

419

warming, which will inhibit nutrient inputs from waters below the mixed layer. In the

420

offshore waters of the ECS, a reduced degree of mixing in the summer results in low

421

nutrient availability, which inhibits the growth of diatoms but has no apparent

422

negative effect on S-strategist dinoflagellate species. Such species include the harmful

423

unarmored dinoflagellate Cochlodinium polykrikoides, which prefers a temperature of

424

~27°C (Matsuoka et al. 2010). In the coastal waters, there is a continuous input of

425

terrestrial nutrients from the Changjiang River, even when the seawater is stratified,

426

and that input benefits diatoms (Furuya et al. 2003). High nutrient concentrations that

427

benefit diatoms also occur in the Changjiang Diluted Water front, where it mixes with

428

the Taiwan Warm Current Water, and in the upwelling zones in the coastal waters of

429

Zhejiang and Fujian provinces (Tang et al. 2006), where diatom blooms have been

430

reported. However, dinoflagellates may replace diatoms in any of these areas if there

431

is a relaxation of the processes that deliver the allochthonous nutrients (Smayda 2010).

432

For example, well-known late spring dinoflagellate blooms (Gymnodinium and

433

Prorocentrum species) occur at temperatures of ~18°C when nutrient concentrations

434

are low, after the phosphate has been consumed by diatoms and there is still some

435

nitrogen remaining. This phenomenon has been reported in the ECS (Guo et al. 2014)

436

and in other coastal and shelf areas (Anderson et al. 2002, Heisler et al. 2008).

437

Similarly, we found that during the summer Changjiang Diluted Water, characterized

438

by both high temperatures and high nutrient concentrations, can be a site of

439

dinoflagellate blooms (Fig. 4d–f), perhaps due to the combined ability of R- and

AC C

EP

TE D

M AN U

SC

RI PT

418

20

ACCEPTED MANUSCRIPT S-strategists to tolerate high turbulence and phosphate depletion, but otherwise

441

eutrophic conditions. The biomass of diatoms relative to that of dinoflagellates

442

decreased from 1959 to 2009 in the Changjiang River estuary and adjacent coastal

443

waters in summer (Jiang et al. 2014). Our data (Fig. 2) and model results indicate that

444

this pattern may have been enhanced in the past several years.

RI PT

440

Using surface microscopy data (dominated by Pseudo-nitzschia delicatissima,

446

Thalassionema nitzschioides, Paralia sulcata, and Skeletonema sp. for diatoms, and P.

447

donghainese, Scrippsiella trochoidea, and G. lohmanni for dinoflagellates) collected

448

from cruises C1–C4 (Table S1), Guo et al. (2014) has obtained a decreasing trend for

449

diatoms against N:P ratio and an opposite trend for dinoflagellates against N:P ratio.

450

Our analysis of temperature and N:P ratios provides additional insight into the

451

dynamics of diatoms and dinoflagellates in the ECS. An analysis of the GAMM

452

results (Fig. 4) and an analysis of the medians of the bins of the GLM results (Figs. 5

453

and 6) revealed that the N:P ratio may be a factor that drives the succession of

454

diatoms and dinoflagellates, and that the effect of the N:P ratio is a function of

455

temperature. The N:P ratio in algal biomass is positively correlated with temperature

456

at the global scale (Yvon-Durocher et al. 2015), and the external N:P ratio is

457

negatively correlated with temperature in the ECS (Fig. S4b). The implication is that

458

the algal internal N:P ratio will exceed the external N:P ratio when the temperature is

459

higher than a threshold. Because the internal N:P ratio of dinoflagellates is generally

460

higher than that of diatoms (John and Flynn 2000, Rhee and Gotham 1981), the high

461

N:P ratios of particulate matter at high temperatures was probably accounted for by

AC C

EP

TE D

M AN U

SC

445

21

ACCEPTED MANUSCRIPT the greater biomass of dinoflagellates, the indication being that high temperatures

463

favor dinoflagellates in the ECS. In addition, according to the resource ratio theory

464

(Rhee and Gotham 1981, Tilman 1977), if the maximum growth rates of diatoms and

465

dinoflagellates are similar, diatoms will competitively eliminate dinoflagellates when

466

external N:P ratios are low, whereas mixotrophic dinoflagellates will prevail when the

467

external N:P ratios are high. Given the fact that dinoflagellates typically have lower

468

growth rates than do diatoms of a comparable size (Finkel et al., 2010), the internal

469

N:P ratio must be lower for diatoms than dinoflagellates (Klausmeier et al. 2004a,

470

2004b). The winner of the competition will therefore tend to be mixotrophic

471

dinoflagellates when the external N:P ratios are high, a conclusion consistent with our

472

observations that high N:P ratios favor dinoflagellates in the ECS and a previous

473

report that a high N:P condition is more likely to promote species that have slow

474

growth rates (Glibert and Burkholder, 2011). The N:P ratio (26 ± 20) at PO4

475

concentrations less than 0.08 µmol L–1 was obviously higher than the N:P ratio (17 ±

476

8) at higher PO4 concentrations. The ratio of Cp/Cf was significantly higher for PO4 <

477

0.08 µmol L–1 than at the latter conditions (Kruskal-Wallis rank sum test, p < 0.001,

478

Fig. S4c). Because there is evidence that the internal N:P ratio is highest under

479

phosphorus limited conditions compared to the condition of abundant resources and

480

conditions limited by light or nitrogen (Klausmeier et al. 2004a), the dominance of

481

dinoflagellate under high N:P conditions is expected in the ECS.

AC C

EP

TE D

M AN U

SC

RI PT

462

482

The GLM model reflected the fact that the spring dinoflagellate blooms became

483

weaker with increasing PO4 concentrations but became stronger with increasing N:P 22

ACCEPTED MANUSCRIPT ratios (Fig. 6 and Fig. S6b). Projecting future changes of these blooms requires

485

knowing what will happen to nutrient concentrations as a result of eutrophication.

486

Assuming a C/Chla ratio of 50 by weight and a C/N ratio of 5.68 by weight (Laws

487

1991, Laws and Redalje 1979, Laws and Wong 1979), we estimated that the

488

percentage of nitrogen in phytoplankton was negatively correlated with nutrient

489

concentrations, and this relationship was rather insensitive to the assumed C/Chla and

490

C/N ratio (Fig. S8). Therefore, low ambient nutrient concentrations were the result of

491

phytoplankton taking up a large fraction of the nutrients. In accord with this logic, we

492

hypothesized that there would be an increase in phytoplankton as a result of

493

eutrophication instead of an increase of inorganic nutrient concentrations. Previous

494

studies have demonstrated that the nitrogen in the ECS has been increasing while the

495

phosphorus has been decreasing slightly as a result of the long-term increase in the

496

Changjiang River discharge and an increase in nitrogen fertilizer use in China (Jiang

497

et al. 2014, Zhou et al. 2008). The regression lines between NOx and PO4 would

498

therefore move toward the left and up (Fig. S4a). The result would be more

499

phosphate-limited situations, leading to higher N:P ratios, and such conditions would

500

favor dinoflagellates.

SC

M AN U

TE D

EP

AC C

501

RI PT

484

To predict the future responses of diatom and dinoflagellate biomass to combined

502

effects of warming and eutrophication, we assumed a 2°C increase in water

503

temperatures and ~7.4-fold increase in the N:P ratio and focused on the surface water

504

of coastal regions with salinity less than 31 (Gong et al. 1996). The surface waters of

505

the ECS underwent a significant warming at a rate of 0.09°C per decade during 1870– 23

ACCEPTED MANUSCRIPT 2011 (Bao and Ren 2014). The largest warming trend occurred in the mouth of the

507

Changjiang River, where the annual mean warming rate exceeded 2.4°C per century

508

(Liu and Zhang 2013) and exceeded 0.49°C per decade in the summer (Jiang et al.

509

2014). If this warming trend continues, it is reasonable to assume that there will be a

510

temperature rise of at least 2°C by 2100 in the surface waters of much of the ECS. In

511

addition to warming, anthropogenic activities have led to a shift in the nutrient regime

512

in the ECS, especially in the Changjiang River estuary and adjacent coastal waters,

513

where the N:P ratio has been steadily increasing (Zhou et al. 2008), and the observed

514

average molar N:P ratio in 2010 reached about 200 (Jiang et al. 2014). As a result of

515

the increase of nitrogen fertilizer use in China, our presumptive future N:P ratio of

516

about 180 will therefore likely be an underestimate, at least near the Changjiang River

517

estuary. We predict that the mean Cf should decrease by 19% and the mean Cp should

518

increase by 60% as the temperature increases by 2°C. These decrease in Cf and

519

increase in Cp reflect a shift from diatoms to dinoflagellates over the next century, as

520

proposed previously (Jiang et al. 2014, Zhou et al. 2008). We did not project changes

521

of temperature beyond the year 2100 because it is unclear whether phytoplankton will

522

adapt to future niches over time (Irwin et al. 2015). Changes beyond 2100 could have

523

even more serious implications than the changes we have projected. We did not

524

project changes of N:P ratio over 180 because such situations may be more

525

complicated than we assumed here. For instance, although silicate is not limiting in

526

the ECS at present, increasing nitrogen input and diatom consumption will result in a

AC C

EP

TE D

M AN U

SC

RI PT

506

24

ACCEPTED MANUSCRIPT 527

decrease of Si:N ratio (Jiang et al. 2014), and silicate may be the limiting factor for

528

diatoms someday. It is important to realize that the influence of changes in irradiance may alter

530

phytoplankton community composition (Barton et al. 2015, Edwards et al. 2015), but

531

we do not feel that the predictions of our model are seriously compromised by the

532

absence of irradiance because the indirect effect of this factor is accounted for

533

approximately by the variables we considered in this study; temperature, irradiance,

534

nutrient concentrations and ratios and changes in these variables co-vary (Irwin et al.

535

2012, Litchman and Klausmeier 2008). An important caveat is that we did not

536

consider the effect of grazing because high-resolution sampling of processes such as

537

grazing is difficult. Because observations in the China seas have shown that

538

microzooplankton grazing is enhanced in eutrophic ecosystems (Chen et al. 2012,

539

Zheng et al. 2015), further study is needed to explore the effect of grazing to make

540

more realistic predictions about the impact of future climate scenarios.

543

SC

M AN U

TE D

EP

542

5. Conclusions

AC C

541

RI PT

529

We determined a series of quantitative functional responses of diatom and

544

dinoflagellate biomass (Cf and Cp) resulting from the interaction of temperature and

545

nutrient concentrations or ratios. Such interaction is key in driving the succession of

546

diatoms and dinoflagellates in the ECS. We project that decreases in diatom biomass

547

(Cf) of 19% should occur in 60% of stations while increases in dinoflagellate biomass

548

(Cp) of 60% should occur in 70% of stations as a result of anthropogenic changes in 25

ACCEPTED MANUSCRIPT temperature and nutrient inputs. Such a shift projects a more important role of

550

dinoflagellates in this coastal ecosystem, including a change in the timing and

551

increase in the spatial extent and frequency of dinoflagellate blooms. These changes

552

may be responsible for a decrease of fish production and an increase of gelatinous

553

organisms, such as jellyfish in the ECS (Jiang et al. 2008). Because the sinking rate of

554

dinoflagellates is lower than that of diatoms (Guo et al. 2015), these changes may also

555

reduce the efficiency of the biological pump in the ECS. Such a shift in phytoplankton

556

community structure from diatoms to dinoflagellates should be taken into

557

consideration in future studies of global climate models, carbon cycling and fishery

558

yields.

AC C

EP

TE D

M AN U

SC

RI PT

549

26

ACCEPTED MANUSCRIPT 559

Acknowledgements This work was supported by grants from the National Key Research and

561

Development Program of China (No.2016YFA0601201), the China NSF (Nos.

562

41406143, 41330961, U1406403), the Chinese Academy of Science Project

563

(XDA11020103) and China Postdoctoral Science Foundation funded project

564

(2014M551839). We thank Profs. Sumei Liu, Minhan Dai, Zhigang Yu and Zhiming

565

Yu for nutrient data, and Profs. Daji Huang, Jianyu Hu, Hao Wei and Fei Yu for

566

hydrographic data. We also thank Michael R. Landry and Paul J. Harrision for their

567

helpful comments and suggestions on this manuscript.

568

M AN U

SC

RI PT

560

Supporting Information captions.

570

SI 1. Site description of the East China Sea.

571

SI 2. Data exploration.

572

SI 3. GAMMs model optimization.

573

SI 4. Supplementary figures and tables.

AC C

EP

TE D

569

27

ACCEPTED MANUSCRIPT References

575 576 577 578 579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614

Agusti, S., Vaqué, D., Estrada, M., Cerezo, M.I., Salazar, G., Gasol, J.M. and Duarte, C.M. (2014) Ubiquitous healthy diatoms in the deep sea confirm deep carbon injection by the biological pump. Nature Communications 6, 6-7608. Alves-de-Souza, C., González, M.T. and Iriarte, J.L. (2008) Functional groups in marine phytoplankton assemblages dominated by diatoms in fjords of southern Chile. Journal of Plankton Research 30(11), 1233-1243. Anderson, D.M., Glibert, P.M. and Burkholder, J.M. (2002) Harmful algal blooms and eutrophication: nutrient sources, composition, and consequences. Estuaries 25(4), 704-726. Bao, B. and Ren, G. (2014) Climatological characteristics and long-term change of SST over the marginal seas of China. Continental Shelf Research 77, 96-106. Barton, A.D., Irwin, A.J., Finkel, Z.V. and Stock, C.A. (2016) Anthropogenic climate change drives shift and shuffle in North Atlantic phytoplankton communities. Proceedings of the National Academy of Sciences 113(11), 2964-2969. Barton, A.D., Lozier, M.S. and Williams, R.G. (2015) Physical controls of variability in North Atlantic phytoplankton communities. Limnology and Oceanography 60(1), 181–197. Chang, F.H. and Gall, M. (2013) Pigment compositions and toxic effects of three harmful Karenia species, Karenia concordia, Karenia brevisulcata and Karenia mikimotoi (Gymnodiniales, Dinophyceae), on rotifers and brine shrimps. Harmful Algae 27(3), 113-120. Chen, B. (2015) Patterns of thermal limits of phytoplankton. Journal of Plankton Research 37(2), 285-292. Chen, B., Landry, M.R., Huang, B. and Liu, H. (2012) Does warming enhance the effect of microzooplankton grazing on marine phytoplankton in the ocean? Limnology and Oceanography 57(2), 519-526. Doney, S.C., Ruckelshaus, M., Duffy, J.E., Barry, J.P., Chan, F., English, C.A., Galindo, H.M., Grebmeier, J.M., Hollowed, A.B. and Knowlton, N. (2012) Climate change impacts on marine ecosystems. Annual Review of Marine Science 4(1), 11-37. Edwards, K.F., Thomas, M.K., Klausmeier, C.A. and Litchman, E. (2012) Allometric scaling and taxonomic variation in nutrient utilization traits and maximum growth rate of phytoplankton. Limnology and Oceanography 57(2), 554-566. Edwards, K.F., Thomas, M.K., Klausmeier, C.A. and Litchman, E. (2015) Light and growth in marine phytoplankton: allometric, taxonomic, and environmental variation. Limnology and Oceanography 60(2), 540-552. Edwards, M., Johns, D.G., Leterme, S.C., Svendsen, E. and Richardson, A.J. (2006) Regional climate change and harmful algal blooms in the northeast Atlantic. Limnology and Oceanography 51(2), 820-829. Edwards, M. and Richardson, A.J. (2004) Impact of climate change on marine pelagic phenology and trophic mismatch. Nature 430(7002), 881-884.

AC C

EP

TE D

M AN U

SC

RI PT

574

28

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Finkel, Z.V., Beardall, J., Flynn, K.J., Quigg, A., Rees, T.A.V. and Raven, J.A. (2010) Phytoplankton in a changing world: cell size and elemental stoichiometry. Journal of Plankton Research 32(1), 119-137. Furuya, K., Hayashi, M., Yabushita, Y. and Ishikawa, A. (2003) Phytoplankton dynamics in the East China Sea in spring and summer as revealed by HPLC-derived pigment signatures. Deep Sea Research Part II: Topical Studies in Oceanography 50(2), 367-387. Glibert, P.M., Allen, J.I., Bouwman, A.F., Brown, C.W., Flynn, K.J., Lewitus, A.J. and Madden, C.J. (2010) Modeling of HABs and eutrophication: status, advances, challenges. Journal of Marine Systems 83(3), 262-275. Glibert, P.M. and Burkholder, J.A.M. (2011) Harmful algal blooms and eutrophication: “strategies” for nutrient uptake and growth outside the Redfield comfort zone. Chinese Journal of Oceanology and Limnology 29(4), 724-738. Gong, G., Lee Chen, Y. and Liu, K. (1996) Chemical hydrography and chlorophyll a distribution in the East China Sea in summer: implications in nutrient dynamics. Continental Shelf Research 16(12), 1561-1590. Guo, S., Feng, Y., Wang, L., Dai, M., Liu, Z., Bai, Y. and Sun, J. (2014) Seasonal variation in the phytoplankton community of a continental-shelf sea: the East China Sea. Marine Ecology Progress Series 516, 103-126. Guo, S., Sun, J., Zhao, Q., Feng, Y., Huang, D. and Liu, S. (2015) Sinking rates of phytoplankton in the Changjiang (Yangtze River) estuary: A comparative study between Prorocentrum dentatum and Skeletonema dorhnii bloom. Journal of Marine Systems 154, 5-14. Hallegraeff, G.M. (2010) Ocean climate change, phytoplankton community responses, and harmful algal blooms: a formidable predictive challenge. Journal of Phycology 46(2), 220-235. Han, A., Dai, M., Kao, S., Gan, J., Li, Q., Wang, L., Zhai, W. and Wang, L. (2012) Nutrient dynamics and biological consumption in a large continental shelf system under the influence of both a river plume and coastal upwelling. Limnology and Oceanography 57(2), 486-502. Heisler, J., Glibert, P.M., Burkholder, J.M., Anderson, D.M., Cochlan, W., Dennison, W.C., Dortch, Q., Gobler, C.J., Heil, C.A. and Humphries, E. (2008) Eutrophication and harmful algal blooms: a scientific consensus. Harmful Algae 8(1), 3-13. Hinder, S.L., Hays, G.C., Edwards, M., Roberts, E.C., Walne, A.W. and Gravenor, M.B. (2012) Changes in marine dinoflagellate and diatom abundance under climate change. Nature Climate Change 2(4), 271-275. Huang, B., Ou, L., Hong, H., Luo, H. and Wang, D.(2005) Bioavailability of dissolved organic phosphorus compounds to typical harmful dinoflagellate Prorocentrum donghaiense Lu. Marine Pollution Bulletin 51(8), 838-844. Irwin, A.J., Finkel, Z.V., Müller-Karger, F.E. and Ghinaglia, L.T. (2015) Phytoplankton adapt to changing ocean environments. Proceedings of the National Academy of Sciences 112(18), 5762-5766. Irwin, A.J., Nelles, A.M. and Finkel, Z.V. (2012) Phytoplankton niches estimated from field data. Limnology and Oceanography 57(3), 787-797.

AC C

615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658

29

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Jiang, H., Cheng, H.-Q., Xu, H.-G., Arreguín-Sánchez, F., Zetina-Rejón, M.J., Luna, P.D.M. and Le Quesne, W.J.F. (2008) Trophic controls of jellyfish blooms and links with fisheries in the East China Sea. Ecological Modelling 212(3), 492-503. Jiang, Z., Liu, J., Chen, J., Chen, Q., Yan, X., Xuan, J. and Zeng, J. (2014) Responses of summer phytoplankton community to drastic environmental changes in the Changjiang (Yangtze River) estuary during the past 50 years. Water Research 54, 1-11. John, E.H. and Flynn, K.J. (2000) Growth dynamics and toxicity of Alexandrium fundyense (Dinophyceae): the effect of changing N:P supply ratios on internal toxin and nutrient levels. European Journal of Phycology 35(1), 11-23. Karl, T.R. and Trenberth, K.E. (2003) Modern global climate change. Science 302(5651), 1719-1723. Klausmeier, C.A., Litchman, E., Daufresne, T. and Levin, S.A. (2004a) Optimal nitrogen-to-phosphorus stoichiometry of phytoplankton. Nature 429(6988), 171-174. Klausmeier, C.A., Litchman, E. and Levin, S.A. (2004b) Phytoplankton growth and stoichiometry under multiple nutrient limitation. Limnology and Oceanography 49(4), 1463-1470. Laws, E.A. (1991) Photosynthetic quotients, new production and net community production in the open ocean. Deep Sea Research Part A Oceanographic Research Papers 38(1), 143-167. Laws, E.A. and Redalje, D.G. (1979) Effect of sewage enrichment on the phytoplankton population of a subtropical estuary. Pacificence 33(2), 129-144. Laws, E.A. and Wong, D.C.L. (1979) Studies of carbon and nitrogen metabolism by three marine phytoplankton species in nitrate-limited continous culture. Journal of Phycology 14(4), 406-416. Letelier, R.M., Bidigare, R.R., Hebel, D.V., Ondrusek, M., Winn, C.D.and Karl, M. (1993) Temporal variability of phytoplankton community structure based on pigment analysis. Limnology and Oceanography 38(7), 1420–1437. Leterme, S.C., Edwards, M., Seuront, L., Attrill, M.J., Reid, P.C. and John, A.W.G. (2005) Decadal basin-scale changes in diatoms, dinoflagellates, and phytoplankton color across the North Atlantic. Limnology and Oceanography 50(4), 1244-1253. Leterme, S.C., Seuront, L. and Edwards, M. (2006) Differential contribution of diatoms and dinoflagellates to phytoplankton biomass in the NE Atlantic Ocean and the North Sea. Marine Ecology Progress Series 312(8), 57-65. Lewandowska, A.M., Boyce, D.G., Hofmann, M., Matthiessen, B., Sommer, U. and Worm, B. (2014) Effects of sea surface warming on marine plankton. Ecology Letters 17(5), 614-623. Litchman, E. and Klausmeier, C.A. (2008) Trait-based community ecology of phytoplankton. Annual Review of Ecology, Evolution, and Systematics, 615-639. Liu, Q. and Zhang, Q. (2013) Analysis on long-term change of sea surface temperature in the China Seas. Journal of Ocean University of China 12(2), 295-300. Liu, X., Huang, B., Liu, Z., Wang, L., Wei, H., Li, C. and Huang, Q. (2012) High-resolution phytoplankton diel variations in the summer stratified central Yellow Sea. Journal of Oceanography 68(6), 913-927.

AC C

659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678 679 680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702

30

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

Liu, X., Xiao, W., Landry, M.R., Chiang, K.-P., Wang, L. and Huang, B. (2016) Responses of phytoplankton communities to environmental variability in the East China Sea. Ecosystems, 1-18. Margalef, R. (1977) Life-Forms of phytoplankton as survival alternatives in an unstable environment. Oceanologia 1(4), 493-509. Matsuoka, K., Mizuno, A., Iwataki, M., Takano, Y., Yamatogi, T., Yoon, Y.H. and Lee, J.-B. (2010) Seed populations of a harmful unarmored dinoflagellate Cochlodinium polykrikoides Margalef in the East China Sea. Harmful Algae 9(6), 548-556. Menden-Deuer, S. and Lessard, E.J. (2000) Carbon to volume relationships for dinoflagellates, diatoms, and other protist plankton. Limnology and Oceanography 45(3), 569-579. Moore, S.K., Trainer, V.L., Mantua, N.J., Parker, M.S., Laws, E.A., Backer, L.C. and Fleming, L.E. (2008) Impacts of climate variability and future climate change on harmful algal blooms and human health. Environmental Health 7 Suppl 2(Suppl2), S4-S4. Peloquin, J., Swan, C., Gruber, N., Vogt, M., Claustre, H., Ras, J., Uitz, J., Barlow, R., Behrenfeld, M. and Bidigare, R. (2013) The MAREDAT global database of high performance liquid chromatography marine pigment measurements. Earth System Science Data 5(1), 109-123. Reynolds, C.S. (1987) The response of phytoplankton communities to changing lake environments. Aquatic Sciences 49(2), 220-236. R Development Core Team (2016) R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing. Open access available at: http://cran. r-project. org. Rhee, G.Y. and Gotham, I.J. (1981) Optimum N:P ratios and coexistence of planktonic algae. Journal of Phycology 16(4), 486-489. Roy, S., Llewellyn, C.A., Egeland, E.S. and Johnsen, G. (2011) Phytoplankton pigments: characterization, chemotaxonomy and applications in oceanography, Cambridge University Press. Smayda, T. (2010) Adaptations and selection of harmful and other dinoflagellate species in upwelling systems. 2. Motility and migratory behaviour. Progress in Oceanography 85(1), 71-91. Smayda, T.J. and Reynolds, C.S. (2001) Community assembly in marine phytoplankton: application of recent models to harmful dinoflagellate blooms. Journal of Plankton research 23(5), 447-461. Smayda, T.J. and Reynolds, C.S. (2003) Strategies of marine dinoflagellate survival and some rules of assembly. Journal of Sea Research 49(2), 95-106. Tang, D., Di, B., Wei, G., Ni, I.-H. and Wang, S. (2006) Spatial, seasonal and species variations of harmful algal blooms in the South Yellow Sea and East China Sea. Hydrobiologia 568(1), 245-253. Tilman, D.(1977) Resource Competition between Plankton Algae: An Experimental and Theoretical Approach. Ecology 58(2), 338-348. Wang, Z., Li, R. and Zhu, M. (2006) Study on population growth processes and interspecific competition of Prorocentrum donghaiense and Skeletonema costatum in

AC C

703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734 735 736 737 738 739 740 741 742 743 744 745 746

31

ACCEPTED MANUSCRIPT

EP

TE D

M AN U

SC

RI PT

semi-continuous dilution experiments. Advances in Marine Science 24: 496-503. doi:10.3969/j.issn.1671-6647.2006.04.011(Article in Chinese with English abstract). Wells, M.L., Trainer, V.L., Smayda, T.J., Karlson, B.S., Trick, C.G., Kudela, R.M., Ishikawa, A., Bernard, S., Wulff, A. and Anderson, D.M. (2015) Harmful algal blooms and climate change: Learning from the past and present to forecast the future. Harmful Algae 49, 68-93. Wood, S. (2006) Generalized additive models: an introduction with R, Chapman and Hall. Xie, Y., Tilstone, G.H., Widdicombe, C.L., Woodward, E.M.S., Harris, C. and Barnes, M.K.S. (2015) Effects of increases in temperature and nutrients on phytoplankton community structure and photosynthesis in the Western English Channel. Marine Ecology Progress Series 519, 61-73. Yvon-Durocher, G., Dossena, M., Trimmer, M., Woodward, G. and Allen, A.P. (2015) Temperature and the biogeography of algal stoichiometry. Global Ecology and Biogeography 24, 562-570. Zapata, M., Fraga, S., Rodríguez, F. and Garrido, J.L. (2012) Pigment-based chloroplast types in dinoflagellates. Marine Ecology Progress Series 465, 33-52. Zheng, L., Chen, B., Liu, X., Huang, B., Liu, H. and Song, S. (2015) Seasonal variations in the effect of microzooplankton grazing on phytoplankton in the East China Sea. Continental Shelf Research 111, 304-315. Zhou, M., Shen, Z. and Yu, R. (2008) Responses of a coastal phytoplankton community to increased nutrient input from the Changjiang (Yangtze) River. Continental Shelf Research 28(12), 1483-1489. Zuur, A.F., Ieno, E.N., Walker, N.J., Saveliev, A.A. and Smith, G.M. (2009) Mixed effect models and extensions in ecology with R, Springer New York.

AC C

747 748 749 750 751 752 753 754 755 756 757 758 759 760 761 762 763 764 765 766 767 768 769 770 771

32

ACCEPTED MANUSCRIPT Figure captions

773

Fig. 1. Sampling stations from 23 cruises in the East China Sea from 2002 to 2015.

774

Symbol color represents the number of overlapping stations between the cruises.

775

Fig. 2. Long-term variations of phytoplankton pigments. a. Long-term variations of

776

surface Chla concentration for an 8°×6° region (120–128°N by 26–32°E) in the ECS.

777

The grey line is the weekly averaged data obtained from the satellite; the red solid line

778

produced by loess smoothing and the red dashed lines represent 95% confidence

779

intervals; the blue diamonds are monthly averaged surface TChla determined by

780

HPLC during our field cruises. b. Ratio of monthly median surface peridinin and

781

fucoxanthin concentrations.

782

Fig. 3. Partial effects of temperature and nutrient concentrations or ratios on

783

logarithms of fucoxanthin and peridinin. (a, c), (b, d), (e), (f), (g) and (h) are results

784

produced by models F3, P3, F5, P5, F7 and P7, respectively (Table S2). The solid line

785

is the smoother and the dashed lines 95% confidence bands. The partial temperature

786

effect on fucoxanthin and peridinin was qualitatively independent of whether NOx,

787

PO4, or N:P ratio was used as the metric of nutrient availability in the models.

788

Fig. 4. Combined effects of temperature and nutrient concentrations or ratios on

789

fucoxanthin and peridinin. a–f are results produced by models F2, F4, F6, P2, P4 and

790

P6, respectively (Table S2). White areas in the lower-left corner of each panel

791

represent there are few observations to support credible estimations.

792

Fig. 5. Relationships of the peridinin/fucoxanthin ratio and temperature at different

AC C

EP

TE D

M AN U

SC

RI PT

772

33

ACCEPTED MANUSCRIPT nutrient concentrations. The colored lines are simulated using the loess method (span

794

= 0.7). a. Data were grouped by PO4 concentration. b. Data were grouped by NOx

795

concentration.

796

Fig. 6. Fitted ratio of peridinin/fucoxanthin on the basis of equation (3) at constant

797

PO4 concentrations. NOx concentrations were calculated from PO4 and temperature

798

according to equation (4).

799

Fig. 7. Distribution of predicted changes in pigment concentrations assuming surface

800

temperatures increases of 2°C and N:P ratio increases of e2 = 7.4. Red circles

801

represent increases and blue circles decreases compared to present concentrations. a.

802

fucoxanthin. b. peridinin.

803

Fig. 8. Predicted concentrations of pigments displayed as a boxplot for a sequence of

804

increases in both temperature and ln transformed N:P ratio. The white horizontal line

805

and the white dot inside each box are the median and the mean respectively. The solid

806

line across the boxes represents GAM smoothing and shaded areas denote 95%

807

confidence bands. a. fucoxanthin. b. peridinin.

AC C

EP

TE D

M AN U

SC

RI PT

793

34

ACCEPTED MANUSCRIPT

●2

●●●● ●● ●●●●● ● ● ● ● ●● ● ● ● ●● ● ● ●● ● ● ● ● ● ● ● ● ● ● ● ● ● ●4 ● ● ●● ● ● 50 ● ● ● ● ● ●● ●●●●● ● ● ● ● ● ●5 ● ●● ● ● ● ●● ● ● ●●●● ●6 ●● ● ● ● ●● ● ● ●● ● ● ● ●● ●● ●● ●7 ● ● ● ● ● ●● ●●●● ● ● ● ●8 ● ●●●●●●● ●● ● ●● ●● ● ●● ●9 ● ● ● ●●●●●● ●●● ● ● ● ● ● ● ● ●●● ● ●● ● ● 10 ● ●● ● ● ●●● ●● ● ● ● ● ● ● ● ● ● 11 ● ●● 0 0 ● ● ● 1 ● ● ● ● ● ●● ● ●● ● ● ●● ● 12 ● ●●● ● ● ● ●● ● ●● ● ● ● 13 ●● ● ●● ● 150 ● ●● 0 ● ● ● 50 ● ● ● ● ● 14 ● ● ● ● ● ● ● ● ● ● ●●●● 200 ●●●● ●● ● 150 ● ●● ● ● ● 1000 200 ●● ● 0 0 0 2 ●● ●● ● ● 50

AC C

0

25

25

EP

TE D

M AN U

SC

RI PT

●3

500

10

25

26

Latitude (°N) 27 28 29 30

31

32

●1

25

120

121

122

0

500

25

2000 0 0 0 3 124 125 126

123 Longitude (°E) 4000

20

0

0 50

127

25

ACCEPTED MANUSCRIPT

0 10 5025 500 0

25 10025

50

24 26 28 30 32 34

25

5 5002 5 5002 200 2005000 0 50 1 00025 20 25 000 2000 4030500 50 0 0 125 119 121 123 127 129

SC

RI PT

10

0

150

0

M AN U

1

Chla (mg/m3) 2 3

4

(a)

TE D EP AC C

0.0

Peridinin/fucoxanthin 0.2 0.4 0.6 0.8

1.0

(b)

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015

1998 1999 2000 2001 2002 2003 2004 2005 2006 2007 2008 2009 2010 2011 2012 2013 2014 2015 Year

Fucoxanthin

(b)

(e)

ACCEPTED MANUSCRIPT

Peridinin

Fucoxanthin

(f)

Peridinin

0.6

0.0

0.2

−0.5

0.0 −1.0

−0.2

RI PT

0.4

SC

Partial residuals

(a)

15

25

(d)

20

25

Peridinin

0.0

−0.5

0.1

30

EP

Fucoxanthin

15 30 10 Temperature (°C)

AC C

Partial residuals

(c)

20

TE D

10

M AN U

−0.4

(g)

0.4

1 2 PO4 (μmol/L) (h) Fucoxanthin

0.1

0.4

1

2

Peridinin

1.0 0.5 0.0 −0.5 −1.0

0.1

0.1 0.5 1 2 5 1020 50 NOx (μmol/L)

0.5 1 2

5 1020 50

0.2

2

8 16 32 100 300 0.2 N:P ratio

2

8 16 32 100 300

1 0.5

20 10 5

1

0.1 (d)

15

20

25

30

EP

0.5 0.2 0.1

30

10

e3.0

10 300

15

20

25

30

e3.0

32 16 8

e2.5 e2.0 (f)

15 20 25 Temperature (°C)

30

e2.5 e3.5

2

(e) 15 20 25 Temperature (°C)

2

100

AC C

2 1 0.5

10

10 2

N:P

30

e3.5 (c)

TE D

25

PO4 (µmol/L)

NOx (µmol/L)

20

e4.0

R I N:P PT SC

0.1 (b)

15

e4.5

32 16 8

0.2

0.1 (a) 10 50

100

M AN U

2 1 0.5

e5.0

10

Fucoxanthin (ng/L)

20 10 5

300

ACCEPTED MANUSCRIPT

15 20 25 Temperature (°C)

30

e1.5

Peridinin (ng/L)

2

PO4 (µmol/L)

NOx (µmol/L)

50

(a) PO4 (µmol/L)

●

(0,0.08)

10.0

●

[0.08,0.4)

●

ACCEPTED MANUSCRIPT

[0.4,0.8)

●

[0.8,1.5)

(b)

NOx (µmol/L)

●

(0,1.0)

10.0

[1.0,5.0)

●

[5.0,10)

●

[10,50)

●

SC ●

1.0 ●

●

0.1

●

● ●

●

● ● ●

●

●

● ● ●

●● ●

●

●

●

●

●

● ●

●

●

● ●

● ●

●

●

● ●

●

●

● ●

● ●● ●

●

●

●

●

● ●

● ●

12

●

●

●

●

10

●●

●

0.1

● ●

●

●

●

EP

● ●

●

1.0

TE D

●●

●

●

M AN U

●

AC C

Peridinin/fucoxanthin

RI PT

●

●

14

16

18 20 22 24 Temperature (°C)

●

26

28

30

10

12

14

16

18 20 22 24 Temperature (°C)

26

28

30

ACCEPTED MANUSCRIPT

0.25

0.60

RI PT

0.05

SC

20

M AN U

Fitted (peridinin/fucoxanthin)

PO4 (µmol/L)

EP

TE D

1.0

AC C

0.1

10

15

20 Temperature (°C)

25

30

(a)

(b)

ACCEPTED MANUSCRIPT

Peridinin

120

M 28 AN 29 U

30 27

150

0

25

50

25

0

26 25

200 000 e 1.5 1 200 2000 3.5 e 50 6.5 00 e 2 ng/L 500 25 0 0 00 5 0 2000 5 0 0 0 3 121 122 4123 000 124 125 126 127 Longitude (°E) 150

10

10

0

26

25

AC C

200 000 e 1.5 1 200 2000 3.5 e 50 6.5 00 e 2 ng/L 500 25 0 0 00 5 0 2000 5 0 0 0 3 121 122 4123 000 124 125 126 127 Longitude (°E) 150

25

0

50

100

25

27

150

TE D

100

EP

Latitude (°N) 28 29 30 25

50

SC

31

31

50

RI PT

32

32

Fucoxanthin

25

120

(a)

ACCEPTED MANUSCRIPT

•

•

•

•

•

•

•

•

M Predicted peridinin AN (ng/L) US CR IP T

•

TE D

e6

e8

EP

e4

e6

•

•

•

•

•

•

•

•

•

e4

AC C

Predicted fucoxanthin (ng/L)

e8

(b)

0

0.25 0.5 0.75 1

1.25 1.5 1.75 2

0

0.25 0.5 0.75 1

Increases in temperature and ln(N:P)

1.25 1.5 1.75 2

ACCEPTED MANUSCRIPT

Highlights We established niche models for fucoxanthin and peridinin in the East China Sea.




Fucoxanthin and peridinin responded differently to temperature and nutrients.




Warming and eutrophication combined might promote dinoflagellates over

RI PT




diatoms.


We predict changes of –19% in diatoms and +60% in dinoflagellates by the year

EP

TE D

M AN U

Dinoflagellate blooms will be more frequent and intense in the East China Sea.

AC C




SC

2100.