Review

# Contrasting futures for ocean and society from different anthropogenic CO2 emissions scenarios

See allHide authors and affiliations

Science  03 Jul 2015:
Vol. 349, Issue 6243, aac4722
DOI: 10.1126/science.aac4722

## Carbon emissions and their ocean impacts

Anthropogenic CO2 emissions directly affect atmospheric chemistry but also have a strong influence on the oceans. Gattuso et al. review how the physics, chemistry, and ecology of the oceans might be affected based on two CO2 emission trajectories: one business as usual and one with aggressive reductions. Ocean warming, acidification, sea-level rise, and the expansion of oxygen minimum zones will continue to have distinct impacts on marine communities and ecosystems. The path that humanity takes regarding CO2 emissions will largely determine the severity of these phenomena.

Science, this issue 10.1126/science.aac4722

## Structured Abstract

### BACKGROUND

Although the ocean moderates anthropogenic climate change, this has great impacts on its fundamental physics and chemistry, with important consequences for ecosystems and people. Yet, despite the ocean’s critical role in regulating climate—and providing food security and livelihoods for millions of people—international climate negotiations have only minimally considered impacts on the ocean. Here, we evaluate changes to the ocean and its ecosystems, as well as to the goods and services they provide, under two contrasting CO2 scenarios: the current high-emissions trajectory (Representative Concentration Pathway 8.5, RCP8.5) and a stringent emissions scenario (RCP2.6) consistent with the Copenhagen Accord of keeping mean global temperature increase below 2°C in the 21st century. To do this, we draw on the consensus science in the latest assessment report of the Intergovernmental Panel on Climate Change and papers published since the assessment.

Warming and acidification of surface ocean waters will increase proportionately with cumulative CO2 emissions (see figure). Warm-water corals have already been affected, as have mid-latitude seagrass, high-latitude pteropods and krill, mid-latitude bivalves, and fin fishes. Even under the stringent emissions scenario (RCP2.6), warm-water corals and mid-latitude bivalves will be at high risk by 2100. Under our current rate of emissions, most marine organisms evaluated will have very high risk of impacts by 2100 and many by 2050. These results—derived from experiments, field observations, and modeling—are consistent with evidence from high-CO2 periods in the paleorecord.

Impacts to the ocean’s ecosystem services follow a parallel trajectory. Services such as coastal protection and capture fisheries are already affected by ocean warming and acidification. The risks of impacts to these services increase with continued emissions: They are predicted to remain moderate for the next 85 years for most services under stringent emission reductions, but the business-as-usual scenario (RCP8.5) would put all ecosystem services we considered at high or very high risk over the same time frame. These impacts will be cumulative or synergistic with other human impacts, such as overexploitation of living resources, habitat destruction, and pollution. Fin fisheries at low latitudes, which are a key source of protein and income for millions of people, will be at high risk.

### OUTLOOK

Four key messages emerge. First, the ocean strongly influences the climate system and provides important services to humans. Second, impacts on key marine and coastal organisms, ecosystems, and services are already detectable, and several will face high risk of impacts well before 2100, even under the low-emissions scenario (RCP2.6). These impacts will occur across all latitudes, making this a global concern beyond the north/south divide. Third, immediate and substantial reduction of CO2 emissions is required to prevent the massive and mostly irreversible impacts on ocean ecosystems and their services that are projected with emissions greater than those in RCP2.6. Limiting emissions to this level is necessary to meet stated objectives of the United Nations Framework Convention on Climate Change; a substantially different ocean would result from any less-stringent emissions scenario. Fourth, as atmospheric CO2 increases, protection, adaptation, and repair options for the ocean become fewer and less effective.

The ocean provides compelling arguments for rapid reductions in CO2 emissions and eventually atmospheric CO2 drawdown. Hence, any new global climate agreement that does not minimize the impacts on the ocean will be inadequate.

## Abstract

The ocean moderates anthropogenic climate change at the cost of profound alterations of its physics, chemistry, ecology, and services. Here, we evaluate and compare the risks of impacts on marine and coastal ecosystems—and the goods and services they provide—for growing cumulative carbon emissions under two contrasting emissions scenarios. The current emissions trajectory would rapidly and significantly alter many ecosystems and the associated services on which humans heavily depend. A reduced emissions scenario—consistent with the Copenhagen Accord’s goal of a global temperature increase of less than 2°C—is much more favorable to the ocean but still substantially alters important marine ecosystems and associated goods and services. The management options to address ocean impacts narrow as the ocean warms and acidifies. Consequently, any new climate regime that fails to minimize ocean impacts would be incomplete and inadequate.

Atmospheric carbon dioxide (CO2) has increased from 278 to 400 parts per million (ppm) over the industrial period and, together with the increase of other greenhouse gases, has driven a series of major environmental changes. The global ocean (including enclosed seas) acts as a climate integrator that (i) absorbed 93% of Earth’s additional heat since the 1970s, offsetting much atmospheric warming but increasing ocean temperature and sea level; (ii) captured 28% of anthropogenic CO2 emissions since 1750, leading to ocean acidification; and (iii) accumulated nearly all water resulting from melting glaciers and ice sheets, hence furthering the rise in sea level. Thus, the ocean moderates anthropogenic climate change at the cost of major changes in its fundamental chemistry and physics. These changes in ocean properties profoundly affect species’ biogeography and phenology, as well as ecosystem dynamics and biogeochemical cycling (13). Such changes inevitably affect the ecosystem services on which humans depend. The ocean represents more than 90% of Earth’s habitable space, hosts 25% of eukaryotic species (4), provides 11% of global animal protein consumed by humans (5), protects coastlines, and more. Simply put, the ocean plays a particularly important role in the livelihood and food security of hundreds of millions of people.

The United Nations Framework Convention on Climate Change (UNFCCC) aims to stabilize atmospheric greenhouse gas concentrations “at a level that would prevent dangerous anthropogenic interference with the climate system ... within a time-frame sufficient to allow ecosystems to adapt naturally to climate change, to ensure that food production is not threatened, and to enable economic development to proceed in a sustainable manner” (6). According to the Copenhagen Accord (7), meeting these goals requires that the increase in average global surface temperature be less than 2°C over the preindustrial average. However, despite the ocean’s critical role in global ecosystem goods and services, international climate negotiations have only minimally considered ocean impacts, especially those related to ocean acidification (8). Accordingly, highlighting ocean-related issues is now crucial, given that even achieving the +2°C target (set on global temperature) would not prevent many climate-related impacts upon the ocean (9).

This paper first summarizes the key findings of the Fifth Assessment Report (AR5) of the Intergovernmental Panel on Climate Change (IPCC) and, given the ongoing acceleration of climate change research, adds newer literature to assess the impacts of global change—including ocean warming, acidification, deoxygenation, and sea level rise—linking ocean physics and chemistry to biological processes, ecosystem functions, and human activities. Second, it builds on scenarios based on the range of cumulative fossil carbon emissions and the IPCC Representative Concentration Pathways (RCP) RCP2.6 and RCP8.5, contrasting two potential futures. RCP2.6 reflects the UNFCCC target of global temperature staying below +2°C, whereas RCP8.5 reflects the current trajectory of business-as-usual CO2 emissions. Third, this paper provides a broad discussion of the options society has for addressing ocean impacts and ends with key messages that provide further compelling arguments for ambitious CO2 emissions reduction pathways.

## Changes in ocean physics and chemistry

Ocean changes resulting from anthropogenic emissions include long-term increase in temperature down to at least 700 m, increased sea level, and a decrease in Arctic summer sea ice (Fig. 1 and Table 1) (10). Other radiatively active agents—such as ozone, methane, nitrous oxide, and aerosols—do not affect the ocean as much as CO2. Setting it apart, CO2 accounts for two or more times the warming attributed to the non-CO2 greenhouse gases by 2100 (11) and causes ocean acidification. The uptake of excess anthropogenic CO2 by the ocean increases the partial pressure of carbon dioxide (Pco2) and dissolved inorganic carbon while decreasing pH and the saturation state of seawater with respect to the calcium carbonate minerals aragonite and calcite, both being critical drivers of solubility of shells and skeletons (12). Rising global CO2 also further exacerbates the nearshore biogeochemical changes associated with land use change, nutrient inputs, aquaculture, and fishing (13).

Table 1 Changes in SST, pH, oxygen content, sea level, and ocean volume with respect to aragonite for CMIP5 models and several RCP emissions scenarios.

After Bopp et al. (23) except sea level rise (28).

View this table:

Both the magnitude and rate of the anthropogenic carbon perturbation exceed the extent of natural variation over the last millennium and over glacial-interglacial time scales (1416). Variability of pH in coastal waters is considerably larger than that in the open ocean, partly driven by upwelling (17), freshwater input (18), eutrophication (19) and biogeochemical processes (20). Anthropogenic trends in biogeochemical variables—notably in pH, Pco2, and the saturation of calcite and aragonite—emerge from the noise of natural variability much faster than sea surface temperature (SST) (21). The combined changes in these parameters will be distinguishable from natural fluctuations in 41% of the global ocean within a decade (22), and the change in aragonite saturation over the industrial period has been more than five times greater than natural variability over the past millennium in many regions (15).

The condition of the future ocean depends on the amount of carbon emitted in the coming decades (Figs. 1 and 2A). The current suite of earth system models illustrate the contrast between future oceans under the high-carbon-emission, business-as-usual RCP8.5 versus the stringent emission-mitigation RCP2.6 (23, 24). The more stringent scenario allows less than one-sixth of 21st century emissions expected under business-as-usual. Between 2012 and 2100, compatible cumulative carbon emissions from fossil fuel use are 1685 gigatons of carbon (GtC) and 270 GtC for the two RCPs, respectively (10, 25). This is in addition to the 375 and 180 GtC already emitted by 2011 by fossil fuel and land use, respectively (25). Because carbon emissions were 10 GtC in 2013 (26), fast and massive emission reductions are required to keep global surface temperature below the 2°C target of the Copenhagen Accord. Carbon emissions would need to be even lower if the ocean absorbs less excess CO2 than is currently predicted. Indeed, the ocean’s effectiveness in absorbing CO2 decreases with increasing emissions: the fraction of anthropogenic emissions absorbed by the ocean in the 21st century is projected to decline from 56% for RCP2.6 to 22% for RCP8.5 (27).

Ocean physics and chemistry will be quite different under these two emissions scenarios, although differences between the two trajectories will not be apparent until 2035. In 2100, the ocean will be much warmer and have a lower pH under RCP8.5 than under RCP2.6 (Fig. 1): The 21st century global mean change in SST differs by nearly a factor of 4 (mean ± 1 SD: 2.73 ± 0.72°C versus 0.71 ± 0.45°C), whereas global surface pH changes range from –0.33 ± 0.003 units to –0.07 ± 0.001 units). By 2100, the average global increase in mean sea level relative to preindustrial is projected to be 0.86 m for RCP8.5 and 0.60 m for RCP2.6 (28). By 2300, it will be less than 1 m for RCP2.6 and from 1 to over 3 m for RCP8.5 (10). Generally, an increase in stratification, linked to sea-surface warming and freshening, is projected; this tends to slow ocean carbon uptake and nutrient supply to the surface (29).

CO2 emissions also affect the deep ocean, although the responses are delayed by the surface-to-deep transport time and continue for centuries even after carbon emissions cease (30). The volume of ocean water that is supersaturated by more than a factor of 3 with respect to aragonite (Ωa > 3) is projected to completely vanish over the course of the century for RCP8.5 and to decrease from 2% to 1.25% of the ocean volume for RCP2.6 (Fig. 1 and Table 1). Conversely, the volume occupied by undersaturated water (Ωa < 1) that is corrosive to unprotected calcium carbonate shells and skeletons expands from 76% of the whole ocean volume in the 1990s to 91% in 2100 with RCP8.5 and to only 83% with RCP2.6. The whole ocean oxygen inventory is consistently projected to decrease (RCP8.5: –3.45 ± 0.44%; RCP2.6: –1.81 ± 0.31%) with largest changes in the subsurface mid-latitude regions. However, it remains unclear whether, and to what extent, low-oxygen regions will expand and whether the observed expansion of oxygen minimum zones over recent decades resulted from direct anthropogenic perturbation or was caused by natural variability (31, 32).

Projections of ocean warming and acidification in coastal systems follow the general trends of global and regional IPCC models but have lower confidence values because of larger contributions of processes other than CO2 uptake (3). Projected regional changes vary, with the largest sea-surface warming in the North Pacific, the tropical East Pacific, and in parts of the Arctic and the largest surface pH decrease in the Arctic (Figs. 1 and 3). By 2100, 69% of the surface ocean will warm by more than 1.5°C and acidify by more than –0.2 pH units relative to preindustrial under RCP8.5 as opposed to less than 1% under RCP2.6 (Fig. 3). The largest absolute decrease in aragonite saturation is projected for the tropical ocean, partly modulated by variability within coral reef sites (33, 34). Seasonally undersaturated conditions are already present in the northeastern Pacific and the California upwelling system (17) and in the Arctic Ocean (35) and expected for the Southern Ocean (36). pH reductions at the sea floor below 500-m depth, which includes biodiversity hot spots such as deep-sea canyons and seamounts, are projected to exceed 0.2 units (the likely bound of natural variability over the past hundreds of thousands of years) by 2100 in close to 23% of North Atlantic deep-sea canyons and 8% of seamounts under RCP8.5—including sites proposed as marine protected areas (37).

In summary, the carbon that we emit today will change the earth system irreversibly for many generations to come (10). The ocean’s content of carbon, acidity, and heat as well as sea level will continue to increase long after atmospheric CO2 is stabilized. These irreversible changes increase with increasing emissions (Fig. 2), underscoring the urgency of near-term carbon emission reduction if ocean warming and acidification are to be kept at moderate levels.

## Effects on biological processes and ecosystems

Organisms and ecosystems are changing in response to ocean warming, acidification, and deoxygenation. The inherent difficulty of distinguishing climate signals from natural variability (38), and of accounting for genetic adaptation (39), makes documenting these shifts challenging, but nevertheless broad anthropogenic impacts are evident (Figs. 2B and 3).

### Warming

Species’ range shifts, usually following a shift in isotherms or temperature extremes, are a key consequence of ocean warming (40). Recent studies strongly reiterate that many species—including various invertebrates, commercially important fish species, and marine mammals—are undergoing phenological and geographical shifts as a result of warming (41, 42). Organisms move at different rates, up to 400 km per decade, as they track temperature changes and local climate velocities according to their ecological niches (43, 44). These shifts will continue with projected ocean warming (42, 45), causing potentially permanent changes to ecosystems, including local extinctions (42), while simultaneously producing novel assemblages (46). Responses to changing temperature depend on species’ specific windows of thermal tolerance and are positively related to the degree of warming. Exceeding these limits can affect growth, body size, behavior, immune defense, feeding, and reproductive success (2), although species’ individual tolerances vary. Globally, poleward range shifts of more than 800 species of exploited marine fish and invertebrates projected under RCP8.5 are 65% faster than those under RCP2.6 by mid–21st century relative to the years 2000s (42). There is medium confidence that animals adapted to a wide range of temperatures will cope better with future conditions, whereas tropical and polar specialists are at greatest risk (2). Changes are not synchronous across trophic levels; alterations in body sizes within food webs (47) and in food web composition (48) have been reported. Recent experimental studies suggest that some species may adapt to warming projected under RCP8.5 [e.g., (49, 50)], but biogeographical shifts restrict adaptive potential and the small number of species- and population-scale studies limit the ability to generalize the importance of genetic adaptation in moderating impacts.

Reef-building corals are extremely vulnerable to warming (1, 2, 51). Warming causes mass mortality of warm-water corals through bleaching as well as through biotic diseases, resulting in declines in coral abundance and biodiversity. Coral reefs can recover from bleaching events when thermal stress is minimal and of short duration (52). However, ocean warming and acidification are expected to act synergistically to push corals and coral reefs into conditions that are unfavorable for coral reef ecosystems (53). There is limited agreement and low confidence on the potential for corals to adapt to rapid warming. Most coral species have clearly adapted to warm environments (54, 55) although the time scale of adaptation is likely to be long given the relatively lengthy generation times of corals [3 to 100 years (56)]. Recent studies have shown short-term acclimation and adaptation in some fast-growing species (57) and suggested that some genetic mechanisms may allow faster rates of change (58). It is, however, doubtful that corals will be able to adapt quickly enough to maintain populations under most emissions scenarios (56, 59, 60), especially where temperature keeps increasing over time (RCP4.5 and higher). Temperature is also an important determinant of deep-sea coral distribution, although less is known about how deep coral communities respond to thermal stress (61). The consensus is that adaptive responses of organisms will have little chance to keep current ecosystems unchanged if ocean temperature and chemistry are not stabilized, giving marine ecosystems the time needed to adapt to the new, stable environmental conditions.

### Ocean acidification

Organisms producing calcium carbonate shells and skeletons experience the strongest negative impacts from ocean acidification (62). Responses to future levels of ocean acidification expected by 2100 under RCP8.5 include reduced calcification, reduced rates of repair, and weakened calcified structures, but responses are species-specific [e.g., (63)]. Reproductive success, early life-stage survival, feeding rate, and stress-response mechanisms may also be affected (2). Most studies have investigated the effects of ocean acidification on isolated organisms; far less is known about the effects on communities and ecosystems.

Few studies measure present-day acidification effects in natural settings. However, recent field observations show a decrease in coccolith thickness over the past 12 years in the Mediterranean (64) and dissolution of live pteropod shells in the California Current system and Southern Ocean, both areas that experience significant anthropogenic acidification (65, 66). Recent investigations have also begun to report community-level responses, for example, in phytoplanktonic (67, 68), bacterial (69), seagrass (70), and algal (71) communities. Decreases in net calcification, at least partly because of ocean acidification, have also been observed in a coral reef over 1975 to 2008 (72), and conditions are already shifting some coral reefs to net erosion (73).

Most studies have investigated phenotypically plastic responses in relatively short-term, single-generation experiments and therefore did not consider the potential for transgenerational response and genetic adaptation (74). Studies published since the AR5 have expanded on the longer-term responses to ocean acidification and have found that transgenerational and evolutionary responses can partly mitigate adverse effects, for example, in phytoplankton (75), planktonic crustaceans (76), sea urchins (77), and fish (78).

### Deoxygenation

Expanding oxygen minimum zones benefits microbes and life forms adapted to hypoxia while restricting the ranges of most other species (2), with eutrophication from coastal pollution exacerbating the problem, resulting in organic matter increasing metabolic rates in deeper coastal areas (79). Moreover, higher temperatures increase species’ sensitivity to hypoxia (80), limiting the depth distribution of fish and invertebrates not adapted to hypoxic conditions (81) and leading to community-level shifts to smaller Eukarya, Bacteria, and Archaea under conditions of diminished O2 (82). Conversely, hypoxia-adapted species are likely to benefit, as illustrated by the range-expansion of a squid adapted to hypoxia (83).

### Multiple drivers

Investigations of single drivers can produce misleading inferences about organismal responses in a multivariate natural environment because interactive (additive, synergistic, or antagonistic) effects often are not predictable from single-driver studies. This is a major source of uncertainty for projections (39), but several recent studies have better characterized interactions among some drivers. Changes in temperature and pH, such as those projected under RCP8.5 for the year 2100, can have synergistic negative effects on species growth, survival, fitness, calcification, and development (8488). In some cases, hypoxic conditions can mediate negative effects of ocean acidification (89, 90); however, ocean acidification and hypoxia increase heat sensitivity and vice versa (2), and oxygen loss combined with warming is projected to contract metabolically viable habitats of marine animals on a global scale (91). Growing evidence also suggests that interactions with other environmental factors—such as irradiance, nutrient availability, geographic location, and species community composition—can strongly modulate the biological effects of warming, ocean acidification, and hypoxia (68, 9295). Few studies addressed the potential for genetic adaptation to multiple drivers, but the phytoplankton Emiliania huxleyi can adapt to simultaneous warming and acidification (49). Other direct human impacts (such as fishing) can reduce the adaptive capacity of marine species and ecosystems to CO2-related impacts. For example, fishing reduces species diversity, simplifies the trophic food web, and increases ecosystem sensitivity to climate change (96). Because relatively little is known on the interacting effects of environmental factors and the complexity of the marine food web, it is premature to make ecosystem-wide projections. However, impacts on keystone species and ecosystem engineers of three-dimensional habitats are likely to shift whole communities (97).

### Present-day impact and future risks

The observed impacts and future additional risks resulting from ocean warming and acidification vary by organism and ecosystem (Fig. 2B). Warm-water corals are already affected, as are mid-latitude seagrass, high-latitude pteropods and krill, mid-latitude bivalves, and finfish. If CO2 levels are kept to the RCP2.6 scenario, by 2100 the risk of impact increases to “high” for warm-water corals and mid-latitude bivalves. Projections with RCP8.5 indicate very high risk of impact on most marine organisms considered, except mangrove. Avoiding very high levels of risk requires limiting the increase in global surface temperature between 1990 and 2100 to below 2°C and the increase in SST below ~1.2°C. These risks of impact, based on perturbation experiments, field observations, and modeling, are consistent with the paleorecord, which indicates mass extinctions triggered by carbon perturbation events such as at the Permo-Triassic boundary [at a rate slower than the present one (98)] or severe losses of deep-sea fauna during the last glaciation, attributed to oxygen depletion (99). Evolution in response to environmental changes that occurred much slower than those projected in the coming decades did not, therefore, prevent major large-scale alterations of marine ecosystems. Levels of confidence are generally medium to very high for RCP2.6 but significantly lower for RCP8.5, except for seagrass, warm-water corals, and pteropods, for which they remain high or very high (see supplementary materials).

## Effects on ecosystem services and ocean-related human activities

Ocean warming, acidification, and deoxygenation alter earth-system-regulating processes (e.g., climate, heat distribution, weather, water flow, and waste treatment), habitat provision, and cultural services [e.g., recreation and leisure, inspiration, and cultural heritage (100)]. As a consequence, CO2-driven global change is expected to result in economic impacts for humans through the alteration of ocean-derived resources and increasing risks to public health, human development, well-being, and security (101).

### Ocean carbon uptake

Ocean uptake of anthropogenic CO2 is a key service to society that moderates climate change, although it comes at the cost of ocean acidification. CO2 uptake depends on multiple processes, many of which are sensitive to climate change [see above (102)], and the open ocean is projected to absorb a decreasing fraction of anthropogenic CO2 emissions as those emissions increase. The fraction of 21st century emissions remaining in the atmosphere consequently increases from 30% for RCP2.6 to 69% for RCP8.5 (27). The contribution of vegetated coastal ecosystems—including seagrasses, mangrove forests, and salt marshes—to contemporary carbon sequestration (103) is an order of magnitude less than that of the land biosphere and open ocean, and the coastal carbon sequestered is likely part of the natural carbon cycle rather than related to anthropogenic emissions. The projected loss of these habitats would not only reduce this relatively small uptake of CO2, but would also release carbon previously stored and thus exacerbate CO2-driven changes.

### Tourism

Decreases in the quality and abundance of coral reef cover are expected to negatively affect tourism (1, 3). Loss of coral reefs to tourism under the RCP2.6 and RCP8.5 scenarios could cost between U.S. $1.9 billion and U.S.$12 billion per year, respectively (134). Coral reef losses due to ocean warming and acidification on the Great Barrier Reef place up to A$5.7 billion and 69,000 jobs in Australia at risk (135). In addition, ocean acidification may cause an annual loss of reef ecosystem services that are valued up to U.S.$1 trillion by 2100 (136). For about a quarter of countries with reef-related tourism, mainly less-developed countries, this kind of tourism accounts for more than 15% of gross domestic product (137) and is more sustainable than extractive livelihoods.

### Human health

Ocean warming and acidification affect public health and security, although the impact pathways and associated costs are poorly understood. Hosts and parasites are likely to undergo poleward range shifts under climate change, and disease outbreaks of cholera (138) and other Vibrio infections (139) have already been linked to warmer conditions. The increased risk of pathogens and parasites in marine species and increased opportunities for pathogen transfer between hosts (140) can reduce food security (141). Increasing intensity and frequency of storm surges and sea-level rise may expand the geographical and seasonal ranges of bacteria, increasing human exposure to diseases (132). Inundation can also flood agricultural land in coastal regions, jeopardizing food security and harming human health (142).

### Present-day impact and future risks

The impacts of ocean acidification and warming have already been detected in some key ecosystem services, such as coastal protection and capture fisheries (Figs. 2C and 3). The risks of impacts increase as a function of increased temperature and decreased pH but are still moderate by 2100 for most services with the RCP2.6 scenario. However, under RCP8.5, we find that the risks of impact will become high or very high by 2100 for all seven ecosystem services considered. Fin fisheries at low latitude will be affected sooner than other services; they will face very high risk at a CO2 level corresponding to RCP2.6 in 2100. In addition, cumulative or synergistic impacts with other human-induced drivers, such as overexploitation of living resources, habitat destruction, and pollution, will likely exacerbate the risk of CO2-related impacts.

## Management options

Limiting the effects of ocean warming and acidification is critical considering the widespread risks of impacts facing natural and human systems, even under a stringent emissions scenario (RCP2.6; Fig. 2). A growing body of literature presents options for action in response to climate change and ocean acidification (143145). Drawing on Billé et al. (146), these actions can be clustered in four groups (Fig. 4): reducing the drivers of climate change and ocean acidification (mitigate), building or maintaining resilience in ecosystems (protect), adapting human societies (adapt), and repairing damage that has already occurred (repair). At present, only one of these (reducing CO2 emissions) addresses the fundamental problem; the others merely delay or decrease impacts (e.g., protecting reefs from major disturbances such as coral mining). Some actions rely on readily available technologies (e.g., sewage treatment plants to reduce exacerbating effects of coastal nutrient pollution) and socioeconomic mechanisms (e.g., coastal setback zones), whereas more engineering-intensive techniques are being developed and will require testing (e.g., removal of CO2 from the atmosphere). These options interact. For example, reducing secondary environmental stressors so as to retain ecosystem resilience works over some range of Pco2 values but is ultimately relevant only if ocean warming and acidification are drastically limited. One cannot manage coral reef resilience, for example, if there are no healthy reefs remaining (46). Importantly, some policy options are antagonistic: For example, solar radiation management could limit the increase of surface temperature but would reduce the incentive to cut greenhouse gases emissions, including CO2, thereby providing no relief from ocean acidification (147).

A positive development is that a widening range of stakeholders are testing new practices or reviving old ones, including CO2 extraction from seawater (148), assisted evolution of corals (149), coral farming (150), and customary local management (151). Such field tests provide useable information and tools for decision-makers and climate negotiators as to the costs, benefits, and timing of management options. Aquaculture, for example, has shown some potential to reduce the risk of impacts from climate change and ocean acidification through societal adaptation, such as improved monitoring and changing cultured species or farm locations (127, 152). However, the cost of adaptation measures—such as real-time monitoring of water chemistry—can be prohibitive and not within the reach of most aquaculture operations, especially those in the developing world. Ecosystem-based adaptation—or using ecosystems to reduce the vulnerability of people—appears to offer cost-efficient solutions bringing multiple co-benefits, especially for developing countries and marginalized communities (153). Stimulating ecosystem resilience by reducing the number and magnitude of local stressors and setting up marine protected areas (154) with strictly enforced no-take areas and limited pollutant inputs also stand out as tractable priorities. Moreover, some regions and local areas that are relatively less exposed to warming, hypoxia, and acidification could be climate change refugia, where more favorable environmental conditions would enable survival under CO2-driven impacts (155). Thus, identifying these climate change refugia and conserving biodiversity there contribute to building resilience to climate change (156). Nevertheless, all of these options require appropriate policy frameworks and financial commitments to cover transaction and opportunity costs, surveillance, and enforcement and monitoring and likely offer only limited protection in the face of persistent climate change and ocean acidification.

As the ocean warms and acidifies, the range of protection, adaptation, and repair options—and our confidence in those options—dwindles, while the cost of remaining options skyrockets. Lower-emissions scenarios such as RCP2.6 leave society with a greater number of effective options for safeguarding marine ecosystems and the services they provide. Therefore, actions that do not reduce carbon emissions are meaningful ocean management options only if the future climate regime entails ambitious national contributions toward the phaseout of global CO2 emissions as well as a strong funding mechanism and a relevant framework to support on-the-ground implementation of these options.

## Key messages

Maintaining ocean ecosystems and services depends in large part on the negotiation process toward a global climate agreement under the UNFCCC. In this regard, four key messages emerge from our analysis. First, the ocean strongly influences the climate system and provides important services to humans. Second, impacts on key marine and coastal organisms, ecosystems, and services from anthropogenic CO2 emissions are already detectable, and several will face high risk of impacts well before 2100, even with the stringent CO2 emissions scenario (RCP2.6). These impacts are occurring across all latitudes and have become a global concern that spans the traditional north/south divide. Third, the analysis shows that immediate and substantial reduction of CO2 emissions is required in order to prevent the massive and effectively irreversible impacts on ocean ecosystems and their services that are projected with emissions scenarios more severe than RCP2.6. Limiting emissions to below this level is necessary to meet UNFCCC's stated objectives. Management options that overlook CO2, such as solar radiation management and control of methane emission, will only minimize impacts of ocean warming and not those of ocean acidification. Fourth, as CO2 increases, the protection, adaptation, and repair options for the ocean become fewer and less effective.

Given the contrasting futures we have outlined here, the ocean provides further compelling arguments for rapid and rigorous CO2 emission reduction and eventual reduction of atmospheric CO2 content. As a result, any new global climate agreement that does not minimize the impacts on the ocean will be incomplete and inadequate.

## Supplementary Materials

www.sciencemag.org/content/349/6243/aac4722/suppl/DC1

Supplementary Text

Tables S1 and S2

References (158219)

## References and Notes

1. Acknowledgments: This is a product of “The Oceans 2015 Initiative,” an expert group supported by the Prince Albert II of Monaco Foundation, the Ocean Acidification International Coordination Centre of the International Atomic Energy Agency, the BNP Paribas Foundation, and the Monégasque Association for Ocean Acidification. This study is also a contribution to the international IMBER (Integrated Marine Biogeochemistry and Ecosystem Research) project. We are grateful for the considerable help of Y. Estrada (Technical Support Unit of IPCC WG II) and M. Khamla to finalize the illustrations and H. Flores for useful discussion. A.M. acknowledges support from the French National Research Agency (CapAdapt project, ANR-2011-JSH1-004 01). R.B. is supported by the RESCCUE project funded by the French Development Agency and the French Global Environment Facility (AFD CZZ 1647 01 F and FFEM CZZ 1667 01 H). W.W.L.C. acknowledges support from the Nippon Foundation–UBC Nereus Program and Natural Science and Engineering Research Council of Canada. U.R.S. and W.W.L.C. thank the Social Sciences and Humanities Research Council–sponsored OceanCanada Partnership for support. C.T. is supported by the UK Ocean Acidification research program. F.J. acknowledges support by the Swiss National Science Foundation and the European Commission through the European Union Framework Programme 7 project CARBOCHANGE (no. 264879). O.H.-G. is grateful for support from the University of Queensland, Australian Research Council Centre for Excellence, and his ARC Laureate Fellowship. H.-O.P. acknowledges support by the PACES and BIOACID programs. C.M.E. acknowledges support from the National Oceanic and Atmospheric Administration (NOAA) and NASA. The International Atomic Energy Agency is grateful to the government of the Principality of Monaco for the support provided to its Environment Laboratories. We acknowledge the World Climate Research Programme's Working Group on Coupled Modelling, which is responsible for the Coupled Model Intercomparison Project (CMIP), and thank the climate modeling groups for producing and making available their model output. For CMIP, the U.S. Department of Energy's Program for Climate Model Diagnosis and Intercomparison provides coordinating support and led development of software infrastructure in partnership with the Global Organization for Earth System Science Portals. The contents of this paper are solely the opinions of the authors and do not constitute a statement of policy, decision, or position on behalf of NOAA, NASA, the International Atomic Energy Agency, the U.S. government, or the Secretariat of the Pacific Community.
View Abstract