全球变化知识资源中心

Primary production in the Arctic Ocean supports significant fisheries⁶ and renders it an important sink for anthropogenic carbon²; however, climate change has the potential to alter these capacities. Accelerated ice loss is opening surface area across the Arctic, resulting in observations of increased rates of primary production⁷. The reduced salinity caused by melting ice, combined with increasing temperatures, however, increases stratification, restricting turbulent nutrient supply to surface layers⁸. Ice loss also increases surface area for air–sea CO₂ exchange, causing an uptake from the atmosphere into surface waters with already low p_CO2 (ref. 9), and ice melt introduces freshwater with low alkalinity and dissolved inorganic carbon, further lowering the carbon content of surface waters¹⁰. The surface waters of the Arctic Ocean are largely undersaturated with respect to CO₂ throughout spring and summer². In the European sector of the Arctic Ocean (Barents–Greenland Sea/Fram Strait), p_CO2 varies seasonally by more than 200 μatm, with values as low as 100 μatm in spring months¹¹ owing to strong net community production associated with the spring bloom of ice algae followed by that of planktonic algae in open waters^{12, 13}. Hence, increased CO₂ may stimulate primary production during spring and favour a greater CO₂ sinking capacity in the future^{2, 9}, resulting in a feedback between increased CO₂ and primary production, which biogeochemical models do not consider at present (for example, refs 3, 14).

Predicting future primary production in a changing Arctic is not straightforward; models diverge strongly in their predictions depending on the region and drivers for change (that is, sea ice, light, nutrients, warming, and so on)¹⁵, and modelling studies including rising CO₂ concentrations are rare¹⁵. Experimental research from the European Arctic suggests that increasing CO₂ concentrations enhance primary production in nutrient-replete conditions¹⁶, although this response is possibly species-specific owing to varying efficiencies of the mechanisms for concentrating cellular carbon¹⁷. However, the response to increased CO₂ when combined with warming may deviate from the expected additive effect.

Here we seek to determine if there is an interaction of increased CO₂ concentration and temperature on planktonic GPP throughout the spring and summer in the European Arctic region. On the basis of metabolic theory, we would expect a positive effect of both warming and higher CO₂ (a main substrate for autotrophic growth) on GPP rates^{5, 18}. Although previous studies have not found a strong effect of warming on GPP rates in the European Arctic^{13, 19}, as such the effects of warming and increased CO₂ on primary production could cancel each other, leading to no increase in GPP in warmer, high-CO₂ conditions, signalling a temperature dependence for CO₂ fertilization in Arctic planktonic autotrophs. Nevertheless, the effect of enhanced CO₂ on primary production is probably dependent on the availability of nutrients²⁰.

To test our hypotheses, we examined in situ relationships of GPP, p_CO2 and nutrients using data from four oceanographic cruises in the European sector of the Arctic Ocean. We exposed a spring bloom and a summer post-bloom plankton community (inorganic nitrogen: 0.71 and 0.04 μmol N l⁻¹ respectively) to increased CO₂ concentrations. In the latter we bubbled CO₂ at concentrations ranging from 145 to 2,099 μatm in three controlled temperature treatments (1, 6 and 10 °C). We exposed the spring community to five fixed CO₂ treatments ranging from 143 to 1,097 μatm over 24 h. We did not include temperature treatments in the spring experiment as temperatures in the spring are not expected to change with climate warming as long as sea ice is present. Over the course of the experiments we monitored the evolution of GPP, chlorophyll a, nutrients and carbonate system parameters (see Supplementary Table 2).

Examination of in situ data revealed that GPP and p_CO2 are positively related, with GPP increasing as the 1.50  ±  0.46 power of p_CO2 (Fig. 1a and Supplementary Table 1). However, temperature is also strongly positively related with p_CO2 (Fig. 1b and Supplementary Table 1), as gases expand with increasing temperature, confounding the relationship of GPP and CO₂ in situ. To test for an interaction with temperature we standardized p_CO2 to 1 °C, the approximate mean temperature in the data set, so as to remove the thermodynamic effect of temperature from p_CO2. We found a stronger relationship of GPP with p_CO2 at 1 °C—increasing as the 1.83 ± 0.54 power of p_CO2 (Fig. 1c and Supplementary Table 1)—suggesting that an interaction with temperature blurs the relationship between GPP and p_CO2 in situ. Whereas GPP and chlorophyll a concentration were independent of nutrient concentration (p > 0.05, Supplementary Fig. 2), p_CO2 showed a strong positive relationship with nutrient concentration (Supplementary Fig. 3), indicating that CO₂ drawdown is directly connected with nutrient uptake. The intercepts of the p_CO2–nutrient relationships (141.9 ± 8.9 and 157.9 ± 8.2 μatm p_CO2 for p_CO2–phosphate and p_CO2–nitrate, respectively, Supplementary Fig. 3) indicate a threshold p_CO2 of about 150 μatm below which nutrient limitation will preclude GPP from responding to an increase in CO₂.

a, GPP increases with p_CO2. b, However, p_CO2 and temperature (°C) are strongly related in a half-logarithmic relationship. c, When p_CO2 is standardized to 1 °C (see Supplementary Methods), the power relationship between GPP and p_CO2 steepens. In a–c, black lines represent significant regression relationships (Supplementary Table 2).

Corrected online 08 September 2015

In the version of this Letter originally published online, the following should have been included in the Acknowledgements: 'M.S.-M. was funded by Fundación 'La Caixa' PhD grants (Spain).' This error has been corrected in all versions of the Letter.

1. Trenberth, K. E. et al. in IPCC Climate Change 2007: The Physical Science Basis (eds Solomon, S. et al.) 236–336 (Cambridge Univ. Press, 2007).
2. Bates, N. R. & Mathis, J. T. The Arctic Ocean marine carbon cycle: Evaluation of air–sea CO₂ exchanges, ocean acidification impacts and potential feedbacks. Biogeosciences 6, 2433–2459 (2009).
   • CAS
   • Article
3. Steinacher, M., Joos, F., Frölicher, T. L., Plattner, G.-K. & Doney, S. C. Imminent ocean acidification in the Arctic projected with the NCAR global coupled carbon cycle-climate model. Biogeosciences 6, 515–533 (2009).
   • CAS
   • ISI
   • Article
4. Hein, M. & Sand-Jensen, K. CO₂ increases oceanic primary production. Nature 388, 526–527 (1997).
   • CAS
   • Article
5. Brown, J. H., Gillooly, J. F., Allen, A. P., Savage, V. M. & West, G. B. Towards a metabolic theory of ecology. Ecology 85, 1771–1789 (2004).
   • ISI
   • Article
6. Watson, R., Zeller, D. & Pauly, D. Primary productivity demands of global fishing fleets. Fish Fish. 15, 231–241 (2013).
   • Article
7. Arrigo, K. R., van Dijken, G. & Pabi, S. Impact of a shrinking Arctic ice cover on marine primary production. Geophys. Res. Lett. 35, L19603 (2008).
   • Article
8. Wassmann, P., Carroll, J. & Bellerby, R. Carbon flux and ecosystem feedback in the northern Barents Sea in an era of climate change: An introduction. Deep Sea Res. II 55, 2143–2153 (2008).
   • Article
9. Bates, N. R., Moran, S. B., Hansell, D. A. & Mathis, J. T. An increasing CO₂ sink in the Arctic Ocean due to sea-ice loss. Geophys. Res. Lett. 33, L23609 (2006).
   • CAS
   • Article
10. Rysgaard, S., URL: