Skip to main content

Thank you for visiting You are using a browser version with limited support for CSS. To obtain the best experience, we recommend you use a more up to date browser (or turn off compatibility mode in Internet Explorer). In the meantime, to ensure continued support, we are displaying the site without styles and JavaScript.

Rapid emergence of climate change in environmental drivers of marine ecosystems



Climate change is expected to modify ecological responses in the ocean, with the potential for important effects on the ecosystem services provided to humankind. Here we address the question of how rapidly multiple drivers of marine ecosystem change develop in the future ocean. By analysing an ensemble of models we find that, within the next 15 years, the climate change-driven trends in multiple ecosystem drivers emerge from the background of natural variability in 55% of the ocean and propagate rapidly to encompass 86% of the ocean by 2050 under a ‘business-as-usual’ scenario. However, we also demonstrate that the exposure of marine ecosystems to climate change-induced stress can be drastically reduced via climate mitigation measures; with mitigation, the proportion of ocean susceptible to multiple drivers within the next 15 years is reduced to 34%. Mitigation slows the pace at which multiple drivers emerge, allowing an additional 20 years for adaptation in marine ecological and socio-economic systems alike.


Marine ecosystems provide services of high socio-economic value1, including the primary protein source for one in seven of the world’s population2 and regulation of Earth’s climate via the uptake and storage of atmospheric carbon dioxide3,4. However, climate change is predicted to have profound consequences for marine ecosystems, affecting both their structure and functioning, for example, refs 5, 6, 7. The Intergovernmental Panel on Climate Change (IPCC) identifies four principal climate drivers that affect marine ecosystem structure, functioning and adaptive capacity8: pH, temperature, oxygen concentration and food availability. All four are subject to substantial perturbations in projections of future climate change scenarios.

Research has generally focused on the potential negative effects of climate change on marine ecosystems. For example, ocean pH is reduced by increasing atmospheric CO2 concentration, which may result in reduced viability of calcareous organisms, among other effects5. Warming ocean temperatures are associated with increased ocean stratification, which restricts nutrient supply to photosynthetic organisms in surface waters9. The solubility of oxygen and exchange of subsurface waters with the atmosphere will also be reduced with warmer temperatures, driving lower oceanic oxygen concentrations with potentially negative effects on marine organisms10. Although regionally variable, the combined effect of these changes is predicted to be an overall global decrease in primary production (PP), which is the ultimate determinant of food availability to marine ecosystems11. However, in some cases positive (or neutral) responses to potential marine stressors have been observed12,13, implying that uncertainty surrounding the future of the marine ecosystem is large. Here we adopt the terminology that a change in a potential stressor where the ecosystem response is unknown (and may not necessarily be negative) is termed a ‘driver’.

Although the projected climate change response over the coming century in these environmental drivers is large, so is the natural variability encountered by marine organisms, suggesting that some species have the capacity to adapt or acclimate to change5,14,15. Importantly though, natural variability occurs on timescales of a few years (interannual variability) to millennia (glacial–interglacial cycles), whereas climate change is essentially a one-way street, so that associated changes in the marine environment are unlikely to be reversed. Hence, climate change will eventually push marine ecosystem drivers beyond the range of natural variability, potentially resulting in migration of species16, reorganization of ecological niches, the establishment of novel climates14,17 and the requirement for socio-economic systems to adjust to these changes so that livelihoods and human well-being are protected.

In addition, marine ecosystem drivers rarely vary in isolation and multiple factors may act additively or synergistically to increase the impact of a single driver18,19,20. For example, ocean acidification may alter the carbon to nitrogen ratio of sinking organic material so that more oxygen is required for remineralisation21,22. Ocean acidification may act in concert with rising temperatures to reduce coccolithophore abundance or calcite production23. In turn, lower pH and oxygen concentration can enhance temperature sensitivity in corals24,25 and crustaceans26. The synergistic effects of multiple drivers are challenging to investigate in field or lab studies, due to the difficulties of undertaking multi-factorial experiments over multiple generations. However, existing observational evidence tends to suggest that co-occurring stress can add to or amplify the effect of a single stressor18, as seen in a microalgal species grown in 96 experiments where population growth declined with the number of environmental drivers27. Other studies have shown, however, that multiple stressors can interact in unexpected ways, sometimes resulting in a positive or neutral effect12,28. For example, warming and ocean acidification have antagonistic effects on sea urchin larval growth, resulting in minimal overall impact (although both have a negative effect on larval abnormality29).

Although the response of the marine ecosystem to changing drivers is not yet clear, it is nevertheless important to quantify when, where and which combinations of stressors are likely to occur. Of particular relevance for an organism’s ability to adapt to a changing climate is the speed with which drivers of ecosystem stress emerge from the background of natural variability30,31,32. The environmental niche that organisms occupy roughly matches the ambient conditions that they experience33,34, that is, the organisms must be resilient to the range in natural variability. Organisms must be adapted to at least the annual extrema in conditions during winter and summer (the seasonal variability), which in almost all locations exceeds interannual variability. The more rapidly the system is pushed out of its natural range of variability, the less time the organisms will have to adapt or acclimate to the new conditions or migrate to more suitable areas. Where rapid emergence of multiple drivers occurs, marine ecosystems and dependent socio-economic systems may be unable to adjust sufficiently quickly to avert disruption.

Here we examine how rapidly climate change signals in the annual extrema of the drivers of ecosystem stress emerge from the background of natural variability in several ocean properties (sea surface temperature (SST), pH, PP and interior oxygen content), focusing on when and where multiple drivers occur simultaneously. Further, we investigate whether mitigation efforts are able to slow the pace at which multiple drivers emerge. We quantify when the climate change signal exceeds natural variability in the principal drivers of marine ecosystem stress identified by the IPCC under two climate change scenarios35: a ‘business-as-usual’ scenario (RCP8.5)36,37 and a mitigation scenario (RCP4.5). The mitigation scenario was chosen as representative of a possible post-COP21 situation, as a result of which nations submitted Intended Nationally Determined Contributions to indicate their intended level of emissions post 2020. Even if achieved, these intended cumulative emissions still imply a median warming of 2.6–3.1 °C by 2100 (ref. 38), which is within the range of RCP4.5. Currently, global emissions are tracking along the upper end of the IPCC scenarios, as represented by RCP8.5 (refs 39, 40).


Time of emergence

Using a multi-model ensemble from the CMIP5 archive (Supplementary Table 1), we construct time series for each variable of annual maxima (for SST) or minima (for pH, PP and oxygen) for the period 1860–2100 using a combination of the historical runs (1860–2005) and future scenarios (2006–2100)41. The start of the climate change signal is defined as the year when conditions become persistently uni-directional, and the time of emergence (ToE) is defined as the year when annual extrema exceed the long-term trend persistently for the remainder of the time series. The long-term trend is presented in Supplementary Note 1; the ‘Methods’ section and Supplementary Fig. 1 provide more details on ToE definition. The ToE presented here thus represents the exposure of the ecosystem to conditions outside the range of previously experienced seasonal variability. In contrast to previous work on emergence of marine stressors, for example, refs 11, 42, 43, 44, we define natural variability using the seasonal amplitude derived from monthly model output. (The same analysis performed on annual mean model output results in earlier ToE and more rapid pace of climate change; Supplementary Note 2 and Supplementary Figs 2–6). We recognize that variability on shorter timescales, such as diurnal or extreme event-related (for example, heatwaves, storms), can also be pronounced and important45, but we are constrained here by the temporal resolution of the available model output. In addition, the coarse model grid (1° × 1°) eliminates small-scale spatial variability. If natural variability is higher than we estimate (but the trend is of similar magnitude) then ToE will be later than we calculate.

The ToE is illustrated in Fig. 1a–d. The climate change signals of pH and SST emerge very rapidly (global median of 1924 (±4.7 years) for pH and 2034 (±8.8 years) for SST; ±=inter-model s.d. on the global median). Indeed, the climate change signal in pH already exceeds the bounds of natural variability (ToE<2016) in 99% (±0.5%) of the open ocean. The climate change signal in SST has also already emerged in the subtropics and the Arctic. The climate change trends in PP and interior oxygen content emerge later (global median of 2052 (±12 years) for oxygen and 2070 (±8.3 years) for PP).

Figure 1: ToE and pace of climate change in ecosystem drivers.
figure 1

Multi-model median of the year when annual extrema exceed the climate change trend (see ‘Methods’ section) for (a) SST, (b) PP, (c) pH and (d) interior oxygen content in the ‘business-as-usual’ scenario (RCP8.5). Note the different colour scales for each variable. (eh) The pace of climate change: the number of years between the start of climate change and the signal emerging (see ‘Methods’ section). White areas indicate where ecosystem stress does not emerge above the range of variability for that parameter by 2100.

The pace of climate change is represented here by the length of time (in years) between the start of the climate change signal and its emergence from the background natural variability (Fig. 1e–h). The pace of change is uniformly very rapid in pH, occurring in 25 years almost everywhere (see also Supplementary Fig. 7). pH undergoes both a rapid pace of change and emerges early, due to its very small interannual variability, and because a large fraction of anthropogenic CO2 has been absorbed by the ocean46. (Note however that in coastal waters, which are not resolved by these global models, variability in pH is substantially greater47). For SST, the pace is most rapid (<25 years) in subtropical and Arctic regions, and considerably slower in subpolar areas (>50 years). Interior oxygen content evinces a slower pace of change, with the exception of the Northeast Pacific, where models likely underestimate the large natural variability associated with major climate modes48. PP shows the slowest pace of change, potentially because PP acts as an integrator of changes in light, temperature and nutrients. The later ToE in PP is consistent with previous observations that trends are more rapidly detectable in parameters such as SST and pH than in PP, primarily due to the large natural variability in the latter49,50.

Emergence of multiple drivers

Where the rapid emergence of multiple drivers co-occurs is where ecosystems are likely to experience the greatest exposure to climate change20,51. Here we demonstrate that by 2030 regions encountering multiple drivers, principally by pH and SST, dominate the ocean (Fig. 2a–d). Species sensitive to change in pH and/or temperature will likely need to adapt to these new conditions very rapidly. In particular, regions of low variability (for example, subtropics) will potentially be more vulnerable to climate change as these areas emerge both earliest and most rapidly. However, although ocean acidification and warming are widespread, the climate change signal in PP and oxygen remains smaller than the substantial natural variability in most regions in the near future. The exception is the Arctic, which appears to be a hotspot of change with rapid emergence of the climate change response in pH, SST and oxygen content. (However, note that the Arctic Ocean is also a region of high model uncertainty in projections of PP44,52).

Figure 2: Emergence of multiple drivers.
figure 2

Combination of stressors that have emerged above the background of seasonal variability by (a) 2010, (b) 2030, (c) 2050 and (d) 2100 for a ‘business-as-usual’ scenario (RCP8.5), based on the model mean ToE estimates shown in Fig. 1. (eh) same, but for the mitigation scenario (RCP4.5). In the legend, T refers to SST, PP to primary production and O2 to interior oxygen concentration.

By 2050, a mosaic of multiple environmental drivers has emerged (Fig. 2c). Where previously climate change trends in pH and SST dominated the patterns of environmental change, a more heterogeneous patchwork of multiple drivers now develops. In 86% (±10%) of the ocean, multiple drivers have emerged, implying the potential for widespread disruption to marine ecosystems. By 2100 an ocean environment has developed in which climate change trends have emerged in all four marine drivers of ecosystem stress in 62% (±5%) of the ocean (Fig. 2d), and a further 37% (±4%) of the ocean is experiencing a novel combination of conditions in pH, oxygen content and SST.

The effect of mitigation

As we have shown, following a ‘business-as-usual’ emissions pathway (RCP8.5) results in rapid emergence of multiple drivers of the marine environment. The aim of mitigation activities is to reduce the magnitude and rate of climate change impacts. To what extent then can mitigation limit or slow the pace of ecosystem stress emergence? Projections with the IPCC RCP4.5 mitigation scenario demonstrate that the climate change trend in marine ecosystem drivers is reduced (Supplementary Figs 8 and 9). As a result, mitigation slows and delays the emergence of the climate change signal by several years (Fig. 3 and Supplementary Fig. 6); for example, the SST trend emerges in 2050 (±10.5 years). For PP, mitigation results in only small regions of the ocean experiencing climate change-driven conditions outside the range of natural variability before 2100 (Fig. 3).

Figure 3: ToE and pace of climate change in ecosystem drivers under a mitigation scenario.
figure 3

Multi-model median of the year when climate change trend exceeds the range of natural seasonal variability (see ‘Methods’ section) for (a) SST, (b) PP, (c) pH and (d) interior oxygen content in a mitigation scenario (RCP4.5). Note the different colour scales for each variable. (eh) Number of years between the start of climate change and the signal emerging (see ‘Methods’ section). White areas indicate where ecosystem stress does not emerge above the range of seasonal variability for that parameter by 2100.

The effect of mitigation on the emergence of multiple ecosystem drivers is clear (Fig. 4). Within the next 15 years, only 34% (±9%) of the ocean becomes susceptible to multiple drivers in a mitigation scenario41, compared with 55% (±12%) under a ‘business-as-usual’ scenario (Fig. 4). Note, however, that ocean acidification is unavoidable; although reduced in a mitigation scenario, it has already emerged as a driver across the entire open ocean. The pronounced reduction in multiple stress conditions due to mitigation persists through 2050 with only 69% (±12%) of the ocean exposed, compared with 86% (±10%) under the ‘business-as-usual’ scenario. Even by 2100, in the mitigation scenario only 30% (±4%) of the ocean is affected by all four ecosystem drivers, in contrast to the 62% (±5%) in the ‘business-as-usual’ scenario, and there exist large areas where PP remains within the range of natural variability. Mitigation slows the pace of potential ecosystem exposure substantially, delaying the widespread emergence of the ‘quadruple whammy’20 by 25 years (Fig. 4). The exception is the Arctic, where mitigation does little to slow the emergence of multiple drivers (Fig. 2).

Figure 4: Effect of mitigation on the global emergence in drivers of ecosystem stress.
figure 4

The proportion of the ocean in each year 1900–2100 affected by multiple stress (>1 driver) and quadruple stress (all 4 drivers) in the ‘business-as-usual’ scenario (RCP8.5) and a mitigation scenario (RCP4.5). Shaded areas represent ±1 inter-model s.d.


The vulnerability of the marine ecosystem to climate change is considered to be a function of exposure and sensitivity to stressors, combined with adaptive capacity53,54. In this study, we quantify one part of this equation—the potential for exposure to environmental drivers. Marine ecosystems, and the individual organisms that make up those ecosystems, are adapted to the range of conditions they experience33,34. When the environment changes sufficiently that new conditions, or a new combination of conditions, emerge and persist, the organisms must adapt, migrate to more favourable areas, or face extinction. In this way, climate forcing can induce changes at the organism level, which result in changes to ecosystem structure, species interactions and food web dynamics55,56. If novel conditions emerge rapidly, species have less time to either adapt or migrate, potentially increasing the probability that disruption to the ecosystem will occur.

We find that the climate change-driven trend in pH already exceeds the range in natural seasonal variability over most of the ocean, as does SST in the subtropics and Arctic. For PP and interior oxygen content, although the trend is large, the natural variability is also large, resulting in later ToE. Species adapted to living in regions of low variability are likely to have relatively narrow environmental niches8 and may be living close to their maximum tolerance57. Subtropical and tropical species are therefore likely to be more sensitive to the rapid emergence of climate change trends. Polar species are also particularly vulnerable as they cannot shift their geographical range northward in response to emerging drivers and so must either adapt to changing conditions or go extinct58. However, new ecological niches may open for species resilient to ocean acidification and warming waters and inured to the large natural variability that occurs in PP and oxygen.

Our model results suggest that seasonal minimum pH levels have been lower than the previously experienced natural range for >90 years already. This is consistent with analyses of (shorter) observational records, which suggest that trends in ocean acidification are likely to be anthropogenically driven50,59. However, evidence that this change has had a significant or lasting impact on marine organisms is scarce, except perhaps for warm-water coral communities60. Does this imply that the marine ecosystem is actually rather resilient to climate change? The key factor may be the speed with which climate change emerges in marine ecosystem drivers relative to the speed with which organisms can adapt. Individual species can seemingly adapt relatively rapidly (compared with the timescales of climate change) to new conditions. For example, a tropical reef fish was found to acclimate to acute exposure to warmer temperatures within two generations (damselfish have lifespans of >5 years61). Organisms with shorter lifespans, such as phytoplankton (few days), adapt correspondingly more rapidly. Coccolithophores, for example, adapted to a large degree to more acidic conditions within 500 generations62. The fossil record can also provide some insight into the possible impacts of climate change on marine ecosystems (although the paleoclimate changed at slower rates than projected for anthropogenic climate change). During the Paleocene–Eocene Thermal Maximum for example, up to 50% of benthic foraminifera went extinct63 and warm-water species expanded their ranges northward64,65. Although the limits of adaptation capacity are presently very poorly known, past extinctions at slower rates of climate change suggest that adaptation rates in some organisms are unlikely to be fast enough to keep pace, ultimately implying extinction.

We demonstrate that a heterogeneous mosaic of multiple environmental drivers develops in the next 50 years (Fig. 2c). This mosaic suggests that species resilient to change in one driver but negatively affected by another may be able to migrate to newly formed suitable habitats, provided the velocity of climate change (sensu16) does not outpace migration speed. In addition, some species may be able to alter their depth range so that they avoid decreases in thermocline oxygen whilst still remaining within their thermal niche (note that here we only assess changes in surface temperature). Again, little is known about the potential migration speed of marine organisms. Large motile species such as fishes and mammals are likely to be able to migrate rapidly to more favourable conditions (although locating these refugia may not be as simple as tracking a northward-moving isotherm). Smaller motile species, such as zooplankton, have also been observed to migrate in response to climate trends or variability, as in the North Atlantic where the distribution of warm-water copepod species has shifted northward in recent decades66. Planktonic species may be able to rely on rapid dispersal to maintain populations, whereas sessile species may not be able to migrate sufficiently rapidly to keep pace with future climate change. Returning to the Paleocene–Eocene Thermal Maximum as an analogy, mobile crustaceans avoided significant community changes67, whereas sedentary sediment dwellers were heavily impacted63.

Translating the emergence of climate drivers that we present here into an understanding of how the structure and functioning of the marine ecosystem may respond is an extremely challenging task. The interplay between biogeochemical stressors, including synergistic effects, adaptation and migration potential and speed, the bioclimatic envelope, organisms’ climate sensitivity, non-linear responses to changing conditions, short-term acclimatization and extinction risk is complex in the extreme. Even the first step, of connecting the emergence of climate change in ocean conditions with potential stress in marine organisms, requires the implicit assumption that the niche width of an organism scales with the local variability it experiences, and if conditions exceed that variability then a response (whether negative or positive) may occur. However, many species appear to thrive in environments that are less than optimal, for example, some marine fish and invertebrates have warmer or cooler temperature optima than the environment in which they are found54. In addition, the emergence of persistent, anomalously low oxygen (compared to previous seasonal variability) may not be of relevance to individual organisms if concentrations still remain above hypoxic levels, although oxygen has been found to limit animal life even at higher concentrations8.

Currently, it is not clear whether ecosystem-wide adaptation or migration can outpace the speed at which multiple drivers emerge (Fig. 1). How individual species will fare, or how the ecosystem as the sum of its parts will fare, is poorly understood. What is clear however is that there are likely to be winners and losers in the future ocean68. However, lacking the ability to predict the future impact of drivers on marine ecosystems creates significant challenges to determining an appropriate course for sustainable management of the ecosystem services, such as fisheries, that they provide. The quantification of the ToE of multiple drivers presented here is an important first step in achieving an understanding of the response of marine ecosystems to future climate change.

Slowing the pace of climate change could give species more time to adapt to changing conditions or migrate to more suitable areas, potentially reducing extinction risk. Although most ecosystems have the capacity to adapt to changing conditions to some extent14,69, for species endemic to the Arctic or sensitive to ocean acidification, there may be no refuge from climate change impacts, regardless of mitigation efforts. For other species, the timescales over which threshold changes in environmental stressors emerge will be key to determining the degree of disruption17,30,69. Our results demonstrate that mitigation measures substantially slow the pace of climate change (Fig. 4), and so likely ameliorate its impacts; it would also allow time for additional conservation planning and adaptation of fisheries and aquaculture, likely resulting in reduced species extinction risk17,69,70,71. Given that global CO2 emissions are currently tracking the IPCC’s business-as-usual scenario39,40, timely implementation of the reductions pledged under COP21 is needed to slow the rapid development of ubiquitous multiple ecosystem stress.


Model output

Output from 12 IPCC-class Earth system models run for the CMIP5 exercise driven by greenhouse gas, aerosol and ozone concentrations prescribed in a specific representative concentration pathway scenario of climate change were downloaded from the archive at The models and institutions providing the output are listed in Supplementary Table 1. For all models the monthly output for 100-year sections taken from the end of the preindustrial control run, the historical run (1860–2005) and for the RCP8.5 and RCP4.5 scenarios for 2006–2100 were downloaded. The variables SST (‘tos’ in the CMIP archive), surface pH (‘ph’), total annual vertically integrated PP (‘intpp’) and dissolved oxygen (‘o2’) in the thermocline (200–600 m depth range, as in refs 11, 20) were used. All data were re-gridded onto a regular 1° × 1° grid using linear nearest-neighbour interpolation. In the case of depth-resolved dissolved oxygen concentrations, monthly output is not a standard CMIP5 output. Data were therefore sourced from the model progenitors directly; however, output could not be obtained for HadGEM2-CC, HadGEM2-ES or IPSL-CM5B-LR models and so only nine models were used for the monthly oxygen analysis. Preindustrial control run output was detrended, if necessary, using a linear regression prior to further analysis. For each year of model output, the annual maxima (for SST) and minima (pH, thermocline oxygen and PP) were extracted. Supplementary Note 3 and Supplementary Fig. 10 contain a comparison of model output with observations.

Regression analysis

A generalized least-squares model with a first-order autoregressive error term (AR(1)) was applied to time series of the annual maxima (SST) or minima (other variables) using R package ‘nlme’72:

where Yt is the annual extreme in the variable, μ is a constant term (the intercept), t is the linear trend function (here time in years), ω is the magnitude of the trend (the slope) and Nt is the unexplained portion of the data, which is modelled as an AR(1) process. Trends for SST are reported as °C per decade; for PP, pH and oxygen as the % change per decade with respect to the mean of 1986–2005. The assumption of normality in the annual extrema was checked using the Lilliefors test. Approximately 80% of pixels pass the test for pH and SST, and 65% for PP and oxygen. The majority of pixels failing these tests are in the Arctic.

Definition of ToE

All analyses are performed at every model grid cell, separately for each model. To calculate the ToE, first the time series of annual extrema in the conjoined historical and warming scenario runs is created. Then, an inflection point is located by calculating the cumulative sum of the gradient in YtYt) and identifying the year when it exceeds zero (for a positive trend) or drops below zero (for a negative trend) for the remainder of the time series—we refer to this as the start of the climate change signal (see Supplementary Fig. 1). The trend in Yt is then calculated from that start point forward to 2100 using equation (1). The natural variability (or noise) is defined as one standard deviation in annual extrema of a 100-year section of the model’s control run. As a consequence, the natural variability as defined here accounts for only the unforced natural variability, excluding the influence of changing incoming solar radiation or volcanoes. The ToE is then defined as:

This criterion ensures that the trend exceeds 95% of the values in the noise, assuming it is normally distributed, and therefore that the emerging signal is highly unusual. Any values of ToE exceeding 2100 are excluded from further analysis. Assumptions of normality and heteroskedasticity in the residuals from the trend regression were checked using the Lilliefors and Breusch–Pagan test, respectively. Approximately 5–30% of pixels fail one of the tests and, for most models and variables, these pixels are predominantly in the Arctic. Assuming a normal distribution for trends in annual extrema is reasonable given the low percentage (5–30%) of pixels where the assumption is not met. Where the assumptions are not met, the ToE criterion will not necessarily indicate that the trend exceeds 95% of the noise values, and may instead be lower. The ToE is calculated for each individual model and the median year for each pixel is then used as the ensemble mean. The inter-model difference in ToE is expressed as ±1 s.d. The pace of climate change is described here by the difference (in years) between the ToE and the start of the climate change signal. All analyses were performed in Matlab (Mathworks).

Data availability

The model output analysed here is publically available via the CMIP5 archive. The data used in the figures and the code used to generate them will be made available via request to the corresponding author.

Additional information

How to cite this article: Henson, S. A. et al. Rapid emergence of climate change in environmental drivers of marine ecosystems. Nat. Commun. 8, 14682 doi: 10.1038/ncomms14682 (2017).

Publisher’s note: Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.


  1. 1

    Rogers, A. D., Sumaila, U. R., Hussain, S. S. & Baulcomb, C. The High Seas and Us: Understanding the Value of High-Seas Ecosystems Global Ocean Commission (2014).

  2. 2

    Food and Agricultural Organisation of the United Nations. The State of World Fisheries and Aquaculture Fisheries and Aquaculture Department (2012).

  3. 3

    Kwon, E. Y., Primeau, F. & Sarmiento, J. L. The impact of remineralization depth on the air-sea carbon balance. Nat. Geosci. 2, 630–635 (2009).

    CAS  ADS  Article  Google Scholar 

  4. 4

    Parekh, P., Dutkiewicz, S., Follows, M. J. & Ito, T. Atmospheric carbon dioxide in a less dusty world. Geophys. Res. Lett. 33, L03610 (2006).

    ADS  Article  Google Scholar 

  5. 5

    Doney, S. C. et al. Climate change impacts on marine ecosystems. Ann. Rev. Mar. Sci. 4, 11–37 (2012).

    Article  Google Scholar 

  6. 6

    Gattuso, J. P. et al. Contrasting futures for ocean and society from different anthropogenic CO2 emissions scenarios. Science 349, aac4722 (2015).

    Article  Google Scholar 

  7. 7

    Kappel, C. V. Losing pieces of the puzzle: threats to marine, estuarine and diadromous species. Front. Ecol. Environ. 3, 275–282 (2005).

    Article  Google Scholar 

  8. 8

    Portner, H. O. et al. in Climate Change 2014: Impacts, Adaptation and Vulnerability. Part A: Global and Sectoral Aspects. Contribution of Working Group II to the Fifth Assessment Report of the Intergovernmental Panel on Climate Change (eds C. B. Field et al.) 411–484 (Cambridge University Press, 2014).

  9. 9

    Steinacher, M. et al. Projected 21st century decrease in marine productivity: a multi-model analysis. Biogeosciences 7, 979–1005 (2010).

    CAS  ADS  Article  Google Scholar 

  10. 10

    Vaquer-Sunyer, R. & Duarte, C. M. Thresholds of hypoxia for marine biodiversity. Proc. Natl Acad. Sci. USA 105, 15452–15457 (2008).

    CAS  ADS  Article  Google Scholar 

  11. 11

    Bopp, L. et al. Multiple stressors of ocean ecosystems in the 21st century: projections with CMIP5 models. Biogeosciences 10, 6225–6245 (2013).

    ADS  Article  Google Scholar 

  12. 12

    Boyd, P. W. et al. Physiological responses of a Southern Ocean diatom to complex future ocean conditions. Nat. Clim. Change 6, 207–213 (2015).

    ADS  Article  Google Scholar 

  13. 13

    Ries, J. Biodiversity and ecosystems: acid ocean cover up. Nat. Clim. Change 1, 294–295 (2013).

    ADS  Article  Google Scholar 

  14. 14

    Williams, J. W., Jackson, S. T. & Kutzbacht, J. E. Projected distributions of novel and disappearing climates by 2100 AD. Proc. Natl Acad. Sci. USA 104, 5738–5742 (2007).

    CAS  ADS  Article  Google Scholar 

  15. 15

    Boyd, P. W. et al. Biological responses to environmental heterogeneity under future ocean conditions. Global Change Biol. 22, 2633–2650 (2016).

    ADS  Article  Google Scholar 

  16. 16

    Burrows, M. T. et al. Geographical limits to species-range shifts are suggested by climate velocity. Nature 507, 492–495 (2014).

    CAS  ADS  Article  Google Scholar 

  17. 17

    Williams, J. W. & Jackson, S. T. Novel climates, no-analog communities, and ecological surprises. Front. Ecol. Environ. 5, 475–482 (2007).

    Article  Google Scholar 

  18. 18

    Crain, C. M., Kroeker, K. & Halpern, B. S. Interactive and cumulative effects of multiple human stressors in marine systems. Ecol. Lett. 11, 1304–1315 (2008).

    Article  Google Scholar 

  19. 19

    Portner, H. O. Oxygen- and capacity-limitation of thermal tolerance: a matrix for integrating climate-related stressor effects in marine ecosystems. J. Exp. Biol. 213, 881–893 (2010).

    Article  Google Scholar 

  20. 20

    Gruber, N. Warming up, turning sour, losing breath: ocean biogeochemistry under global change. Philos. Trans. A Math. Phys. Eng. Sci. 369, 1980–1996 (2011).

    CAS  ADS  Article  Google Scholar 

  21. 21

    Riebesell, U. et al. Enhanced biological carbon consumption in a high CO2 ocean. Nature 450, 545–548 (2007).

    CAS  ADS  Article  Google Scholar 

  22. 22

    Oschlies, A., Schulz, K. G., Riebesell, U. & Schmittner, A. Simulated 21st century's increase in oceanic suboxia by CO2-enhanced biotic carbon export. Global. Biogeochem. Cycles. 22, GB4008 (2008).

    ADS  Article  Google Scholar 

  23. 23

    Feng, Y. Y. et al. Effects of increased pCO2 and temperature on the North Atlantic spring bloom. I. The phytoplankton community and biogeochemical response. Mar. Ecol. Prog. Ser. 388, 13–25 (2009).

    CAS  ADS  Article  Google Scholar 

  24. 24

    Anthony, K. R. N., Kline, D. I., Diaz-Pulido, G., Dove, S. & Hoegh-Guldberg, O. Ocean acidification causes bleaching and productivity loss in coral reef builders. Proc. Natl Acad. Sci. USA 105, 17442–17446 (2008).

    CAS  ADS  Article  Google Scholar 

  25. 25

    Reynaud, S. et al. Interacting effects of CO2 partial pressure and temperature on photosynthesis and calcification in a scleractinian coral. Global Change Biol. 9, 1660–1668 (2003).

    ADS  Article  Google Scholar 

  26. 26

    Findlay, H. S., Kendall, M. A., Spicer, J. I. & Widdicombe, S. Post-larval development of two intertidal barnacles at elevated CO2 and temperature. Mar. Biol. 157, 725–735 (2010).

    Article  Google Scholar 

  27. 27

    Brennan, G. & Collins, S. Growth responses of a green alga to multiple environmental drivers. Nat. Clim. Change 5, 892–897 (2015).

    ADS  Article  Google Scholar 

  28. 28

    Hoppe, C. J. et al. Iron limitation modulates ocean acidification effects on Southern Ocean phytoplankton communities. PLoS ONE 8, e79890 (2013).

    ADS  Article  Google Scholar 

  29. 29

    Byrne, M. et al. Ocean warming will mitigate the effects of acidification on calcifying sea urchin larvae (Heliocidaris tuberculata) from the Australian global warming hot spot. J. Exp. Mar. Bio. Ecol. 448, 250–257 (2013).

    CAS  Article  Google Scholar 

  30. 30

    Hawkins, E. & Sutton, R. Time of emergence of climate signals. Geophys. Res. Lett. 39, L01702 (2012).

    ADS  Article  Google Scholar 

  31. 31

    Mahlstein, I., Hegerl, G. & Solomon, S. Emerging local warming signals in observational data. Geophys. Res. Lett. 39, L21711 (2012).

    ADS  Google Scholar 

  32. 32

    Mora, C. et al. The projected timing of climate departure from recent variability. Nature 502, 183–187 (2013).

    CAS  ADS  Article  Google Scholar 

  33. 33

    Portner, H. O. & Peck, M. A. Climate change effects on fishes and fisheries: towards a cause-and-effect understanding. J. Fish. Biol. 77, 1745–1779 (2010).

    CAS  Article  Google Scholar 

  34. 34

    Sunday, J. M., Bates, A. E. & Dulvy, N. K. Thermal tolerance and the global redistribution of animals. Nat. Clim. Change 2, 686–690 (2012).

    ADS  Article  Google Scholar 

  35. 35

    Van Vuuren, D. P. et al. The representative concentration pathways: an overview. Clim. Change 109, 5–31 (2011).

    Article  Google Scholar 

  36. 36

    Riahi, K. et al. RCP8.5 - a scenario of comparatively high greenhouse gas emissions. Clim. Change. 109, 33–57 (2011).

    CAS  Article  Google Scholar 

  37. 37

    Magnan, A. et al. Implications of the Paris agreement for the ocean. Nat. Clim. Change 6, 732–735 (2016).

    ADS  Article  Google Scholar 

  38. 38

    Rogelj, J. et al. Paris agreement climate proposals need a boost to keep warming well below 2°C. Nature 534, 631–639 (2016).

    CAS  ADS  Article  Google Scholar 

  39. 39

    Peters, G. et al. The challenge to keep global warming below 2°C. Nat. Clim. Change 3, 4–6 (2013).

    ADS  Article  Google Scholar 

  40. 40

    Sanford, T., Frumhoff, P. C., Luers, A. & Gulledge, J. The climate policy narrative for a dangerously warming world. Nat. Clim. Change 4, 164–166 (2014).

    ADS  Article  Google Scholar 

  41. 41

    Moss, R. H. et al. The next generation of scenarios for climate change research and assessment. Nature 463, 747–756 (2010).

    CAS  ADS  Article  Google Scholar 

  42. 42

    Keller, K. M., Joos, F. & Raible, C. C. Time of emergence of trends in ocean biogeochemistry. Biogeosciences 11, 3647–3659 (2014).

    ADS  Article  Google Scholar 

  43. 43

    Rodgers, K., Lin, J. & Frolicher, T. L. Emergence of multiple ocean ecosystem drivers in a large ensemble suite with an Earth system model. Biogeosciences 12, 3301–3320 (2015).

    ADS  Article  Google Scholar 

  44. 44

    Frolicher, T. L., Rodgers, K., Stock, C. & Cheung, W. W. L. Sources of uncertainties in 21st century projections of potential ocean ecosystem stressors. Global Biogeochem. Cycles 30, 1224–1243 (2016).

    ADS  Article  Google Scholar 

  45. 45

    Wernberg, T. et al. An extreme climatic event alters marine ecosystem structure in a global biodiversity hotspot. Nat. Clim. Change 3, 78–82 (2013).

    ADS  Article  Google Scholar 

  46. 46

    Le Quere, C. et al. Global carbon budget 2015. Earth Syst. Sci. Data 7, 349–396 (2015).

    ADS  Article  Google Scholar 

  47. 47

    Duarte, C. M. et al. Is ocean acidification an open-ocean syndrome? Understanding anthropogenic impacts on seawater pH. Estuaries Coasts 36, 221–236 (2013).

    CAS  Article  Google Scholar 

  48. 48

    Andrews, O., Bindoff, N., Halloran, P., Ilyina, T. & Le Quere, C. Detecting an external influence on recent changes in oceanic oxygen using an optimal fingerprinting method. Biogeosciences 10, 1799–1813 (2013).

    CAS  ADS  Article  Google Scholar 

  49. 49

    Henson, S. et al. Detection of anthropogenic climate change in satellite records of ocean chlorophyll and productivity. Biogeosciences 7, 621–640 (2010).

    CAS  ADS  Article  Google Scholar 

  50. 50

    Henson, S., Beaulieu, C. & Lampitt, R. Observing climate change trends in ocean biogeochemistry: when and where. Global Change Biol. 22, 1561–1571 (2016).

    ADS  Article  Google Scholar 

  51. 51

    Boyd, P. W., Lennartz, S. T., Glover, D. M. & Doney, S. C. Biological ramifications of climate-change-mediated oceanic multi-stressors. Nat. Clim. Change 5, 71–79 (2015).

    ADS  Article  Google Scholar 

  52. 52

    Vancoppenolle, M. et al. Future Arctic Ocean primary productivity from CMIP5 simulations: uncertain outcome, but consistent mechanisms. Global Biogeochem. Cycles 27, 605–619 (2013).

    CAS  ADS  Article  Google Scholar 

  53. 53

    Williams, S. E., Shoo, L. P., Isaac, J. L., Hoffmann, A. A. & Langham, G. Towards an integrated framework for assessing the vulnerability of species to climate change. PLoS Biol. 6, 2621–2626 (2008).

    CAS  Article  Google Scholar 

  54. 54

    Stuart-Smith, R. D., Edgar, G. J., Barrett, N. S., Kininmonth, S. J. & Bates, A. E. Thermal biases and vulnerability to warming in the world's marine fauna. Nature 528, 88–92 (2015).

    CAS  ADS  Article  Google Scholar 

  55. 55

    Boyd, P. W., Strzepek, R., Fu, F. X. & Hutchins, D. A. Environmental control of open-ocean phytoplankton groups: now and in the future. Limnol. Oceanogr. 55, 1353–1376 (2010).

    CAS  ADS  Article  Google Scholar 

  56. 56

    Ottersen, G., Kim, S., Huse, G., Polovina, J. J. & Stenseth, N. C. Major pathways by which climate may force marine fish populations. J. Mar. Syst. 79, 343–360 (2010).

    ADS  Article  Google Scholar 

  57. 57

    Tewksbury, J. J., Huey, R. B. & Deutsch, C. A. Ecology - putting the heat on tropical animals. Science 320, 1296–1297 (2008).

    CAS  Article  Google Scholar 

  58. 58

    McCarty, J. P. Ecological consequences of recent climate change. Conserv. Biol. 15, 320–331 (2001).

    Article  Google Scholar 

  59. 59

    Friedrich, T. et al. Detecting regional anthropogenic trends in ocean acidification against natural variability. Nat. Clim. Change 2, 167–171 (2012).

    CAS  ADS  Article  Google Scholar 

  60. 60

    Bruno, J. F. & Selig, E. R. Regional decline of coral cover in the indo-pacific: timing, extent, and subregional comparisons. PLoS ONE 2, e711 (2007).

    ADS  Article  Google Scholar 

  61. 61

    Donelson, J. M., Munday, P. L., McCormick, M. I. & Pitcher, C. R. Rapid transgenerational acclimation of a tropical reef fish to climate change. Nat. Clim. Change 2, 30–32 (2012).

    ADS  Article  Google Scholar 

  62. 62

    Lohbeck, K. T., Riebesell, U. & Reusch, T. B. H. Adaptive evolution of a key phytoplankton species to ocean acidification. Nat. Geosci. 5, 346–351 (2012).

    CAS  ADS  Article  Google Scholar 

  63. 63

    Thomas, E. in Large Scale Ecosystem Perturbation: Causes and Consequences Vol. Special Paper 424 (eds S. Monechi, R. Coccioni & M. R. Rampino) 1–23 (Geological Society of America, 2007).

  64. 64

    Gibbs, S. J., Bown, P. R., Sessa, J. A., Bralower, T. J. & Wilson, P. A. Nannoplankton extinction and origination across the Paleocene-Eocene thermal maximum. Science 314, 1770–1773 (2006).

    CAS  ADS  Article  Google Scholar 

  65. 65

    Bralower, T. J. Evidence of surface water oligotrophy during the Paleocene-Eocene thermal maximum: nannofossil assemblage data from Ocean Drilling Program Site 690, Maud Rise, Weddell Sea. Paleoceanography 17, 13-1–13-12 (2002).

    Article  Google Scholar 

  66. 66

    Beaugrand, G., Luczak, C. & Edwards, M. Rapid biogeographical plankton shifts in the North Atlantic Ocean. Global Change Biol. 15, 1790–1803 (2009).

    Article  Google Scholar 

  67. 67

    Webb, A. E., Leighton, L. R., Schellenberg, S. A., Landau, E. A. & Thomas, E. Impact of the Paleocene-Eocene thermal maximum on deep-ocean microbenthic community structure: using rank-abundance curves to quantify paleoecological response. Geology 37, 783–786 (2009).

    ADS  Article  Google Scholar 

  68. 68

    Dutkiewicz, S., Scott, J. R. & Follows, M. J. Winners and losers: ecological and biogeochemical changes in a warming ocean. Global Biogeochem. Cycles 27, 463–477 (2013).

    CAS  ADS  Article  Google Scholar 

  69. 69

    Visser, M. E. Keeping up with a warming world; assessing the rate of adaptation to climate change. Proc. R Soc. B Biol. Sci. 275, 649–659 (2008).

    Article  Google Scholar 

  70. 70

    Thomas, C. D. et al. Extinction risk from climate change. Nature 427, 145–148 (2004).

    CAS  ADS  Article  Google Scholar 

  71. 71

    McLaughlin, J. F., Hellmann, J. J., Boggs, C. L. & Ehrlich, P. R. Climate change hastens population extinctions. Proc. Natl Acad. Sci. USA 99, 6070–6074 (2002).

    CAS  ADS  Article  Google Scholar 

  72. 72

    Pinheiro, J., Bates, D., DebRoy, S. & Sarkar, D. nlme: Linear and Nonlinear Mixed Effects Models. (2016).

Download references


We acknowledge the World Climate Research Programme's Working Group on Coupled Modelling, which is responsible for CMIP, and we thank the climate modelling groups (listed in Supplementary Table 1 of this paper) for producing and making available their model output. This work was supported by Natural Environment Research Council National Capability funding to S.A.H. Contributions of T.I. and R.S. were supported by EU H2020 project CRESCENDO (grant 641816). Contribution of J.L.S. was supported by BP under the Princeton Carbon Mitigation Initiative. Contribution of JT was supported by Bjerknes Centre (BIGCHANGE) and RCN (no. 239965).

Author information




S.A.H. conceived the study; S.A.H., C.B. and J.L.S. undertook the analyses; T.I., J.G.J., M.L., R.S. and J.T. contributed bespoke model output. All authors contributed to writing the manuscript.

Corresponding author

Correspondence to Stephanie A. Henson.

Ethics declarations

Competing interests

The authors declare no competing financial interests.

Supplementary information

Supplementary Information

Supplementary Notes, Supplementary Table, Supplementary Figures and Supplementary References (PDF 2423 kb)

Peer Review File (PDF 209 kb)

Rights and permissions

This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Henson, S., Beaulieu, C., Ilyina, T. et al. Rapid emergence of climate change in environmental drivers of marine ecosystems. Nat Commun 8, 14682 (2017).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI:

Further reading


By submitting a comment you agree to abide by our Terms and Community Guidelines. If you find something abusive or that does not comply with our terms or guidelines please flag it as inappropriate.


Quick links

Nature Briefing

Sign up for the Nature Briefing newsletter — what matters in science, free to your inbox daily.

Get the most important science stories of the day, free in your inbox. Sign up for Nature Briefing