Large-scale season-dependent effects of temperature and zooplankton on phytoplankton in the North Atlantic

Transcription

1 This supplement accompanies the article Large-scale season-dependent effects of temperature and zooplankton on phytoplankton in the North Atlantic Jianfeng Feng 1, *, Leif Chr. Stige 2, Joël Marcel Durant 2, Dag Olav Hessen 2, Lin Zhu 1, Dag Øystein Hjermann 2,3, Marcos Llope 4, Nils Chr. Stenseth 2,5,6 1 Key Laboratory of Pollution Processes and Environmental Criteria at Ministry of Education,College of Environmental Science and Engineering, Nankai University, Tianjin , PR China 2 Centre of Ecological and Evolutionary Synthesis (CEES), Department of Biosciences, University of Oslo, PO Box 1066 Blindern, 0316 Oslo, Norway 3 Norwegian Institute for Water Research (NIVA), Gaustadalléen 21, 0349 Oslo, Norway 4 Instituto Español de Oceanografía, Centro Oceanográfico de Cádiz, Puerto Pesquero, Muelle de Levante s/n, PO Box 2609, Cádiz, Andalucía, Spain 5 Institute of Marine Research, Department of Coastal Zone Studies, Flødevigen Research Station, 4817 His, Norway 6 University of Agder, PO Box 422, 4604 Kristiansand, Norway *Corresponding author: fengjf@nankai.edu.cn Marine Ecology Progress Series 502: Supplement. Additional results of statistical analysis Table S1. Granger causality in the relationships of sea-surface temperature (SST) on chlorophyll (CHL). Lag timescale represents months (lag times of 1 to 10 mo considered). Final model with the best lag length was selected with the lowest Akaike's information criterion value.*: significant at p < See Fig. 1 in the main article for location of regions SST CHL Region χ 2 p Lag B e-05* 5 B e-06* 3 B e-07* 3 B e-07* 3 B * 5 C e-07* 5 D e-12* 3 Table S2. Granger causality in the relationships between nauplii and chlorophyll (CHL). Lag timescale represents months (lag times of 1 to 10 mo considered). Final model with the best lag length was selected with the lowest Akaike's information criterion value. *: significant at p < 0.05 Nauplii CHL CHL Nauplii Region χ 2 p Lag χ 2 p Lag B * 4 B B B * B C * 3 D * * 3 1

2 Fig. S1. Monthly anomalies of phytoplankton (log 10 [N + 1], where N is chlorophyll concentration [mg m 3 ], bars) and total copepodite abundance (log 10 [N + 1], where N is zooplankton abundance [ind. m 3 ], circles) for the 7 study regions in the North Atlantic from 1998 to

3 Fig. S2. Monthly anomalies of phytoplankton (log 10 [N + 1], with N being chlorophyll concentration [mg m 3 ], bars) and Calanus I IV abundance (log 10 N + 1], where N is zooplankton abundance [ind. m 3 ], circles) for the 7 study regions in the North Atlantic from 1998 to

4 Fig. S3. Monthly anomalies of phytoplankton (log10[n + 1], where N is chlorophyll concentration [mg m 3 ], bars) and Calanus V VI abundance (log10[n + 1], with N being zooplankton abundance [ind. m 3 ], circles) for the 7 study regions in the North Atlantic from 1998 to

5 Fig. S4. Residual diagnostics of the final seasonal generalized additive models. Normal quantile-quantile plot (Q-Q) for each region (left panels), autocorrelation functions (ACF) of residuals for each region (middle panels) and partial autocorrelation functions (PACF) of residuals for each region (right panels) D7 C7 B7 B6 B5 B4 B2 C7 D Theoretical Quantiles Theoretical Quantiles ACF ACF Series c7m5$res Series d7m5$res 20 Lag 20 Lag Partial ACF Partial ACF Series c7m5$res Series d7m5$res 20 Lag 20 Normal Q-Q Plot ACF(lag:month) PACF(lag:month) Lag 5

6 Fig. S5. Partial effect of month (f 1 term in Eq. 1 in the main article) on phytoplankton (CHL) in 7 study areas. For each plot, the x-axes show the month, and the y-axes, the partial effect that seasonal change has on the phytoplankton. The line is the smooth term effect of month on phytoplankton, with the pointwise 95% confidence interval around the mean prediction (greyshaded area). The estimated degrees of freedom of the smoothing term for every region are listed in Table 1 in the main article 6

7 Fig. S6. Partial effect of the seasonal coefficient for the linear effect of CHL t 1 (chlorophyll concentration at time t 1, mg m 3 ) on CHL t, (chlorophyll concentration at time t, mg m 3 ) (see model formulation in the 'Materials and methods' section, Eq. 1). For each plot, the x- axes show the month, and the y-axes, the partial effect that CHL t 1 has on CHL t. The line is the smooth term effect of the CHL t 1 on the CHL t with the pointwise 95% confidence interval around the mean prediction (grey-shaded area). The estimated degrees of freedom of the smoothing term for every region are listed in Table 1 in the main article Fig. S7. Monthly average values of sea-surface temperature (SST, C) for the 7 study regions in the North Atlantic: (a) Regions B7 D7 with a latitudinal gradient and (b) Regions B2 B6 with a longitudinal gradient. Solid lines and points indicate means; error bars indicate standard errors of the mean. SST data were obtained from the NOAA Extended Reconstructed SST V3b data set ( and averaged spatially in the 7 standard regions a 12 B7SST C7SST D7SST b B2SST B4SST B5SST B6SST Sea surface temperature ( C) Sea surface temperature ( C) Month Month 7

8 Fig. S8. Monthly long-term means for mixed layer depths (NODC World Ocean Atlas 1994, NODC_WOA94) derived from potential density for the 7 study areas. Solid lines and points indicate the means; error bars indicate standard errors of the mean. NODC_WOA94 data provided by the NOAA/OAR/ESRL PSD, Boulder, Colorado, USA, from their web site at 8