Editing IPCC:AR6/WGII/Chapter-3 (section)

=== 3.4.3 Oceanic Systems and Cross-Cutting Changes ===

<div id="h2-12-siblings" class="h2-siblings"></div>

The oceanic zone, comprising &gt;99% of the ocean’s volume, is highly exposed to climate-induced drivers because of its proximity to the atmosphere ( [[#3.2|Section 3.2]] ; [[#Pörtner--2014|Pörtner et al., 2014]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ), while its relative distance from human settlements and coastal ecosystems decreases variability and interactions, and permits many phenomena to be detected clearly and attributed to climate change. This section assesses how climate-driven changes influence oceanic biological systems over very large spatial scales and notes how impacts on the epipelagic zone affect the mesopelagic, bathypelagic and deep seafloor ecosystems.

<div id="3.4.3.1" class="h3-container"></div>

<span id="biogeography-and-species-range-shifts"></span>
==== 3.4.3.1 Biogeography and Species Range Shifts ====

<div id="h3-23-siblings" class="h3-siblings"></div>

<div id="3.4.3.1.1" class="h4-container"></div>

<span id="observed-species-range-shifts"></span>
===== 3.4.3.1.1 Observed species range shifts =====

<div id="h4-4-siblings" class="h4-siblings"></div>

Since previous assessments (Table 3.16), poleward range shifts have remained a ubiquitous response to climate change ( ''high confidence'' ), moving species from warmer regions into higher-latitude ecosystems ( [[#Fossheim--2015|Fossheim et al., 2015]] ; [[#Kumagai--2018|Kumagai et al., 2018]] ; [[#Burrows--2019|Burrows et al., 2019]] ; [[#Lenoir--2020|Lenoir et al., 2020]] ).

'''Table 3.16 |''' Summary of previous IPCC assessments of biogeography and species range shifts

{| class="wikitable"
|-
! Observations

! Projections

|-
| ''AR5 ( [[#Hoegh-Guldberg--2014|Hoegh-Guldberg et al., 2014]] ; [[#Pörtner--2014|Pörtner et al., 2014]] )''

|
|-
| The distribution and abundance of many fishes and invertebrates have shifted poleward and/or to deeper, cooler waters ( ''high confidence'' ).

On average, species’ distributions have shifted poleward by 72.0 ± 0.35 km per decade ( ''high confidence'' ).

| Spatial shifts of marine species due to projected warming will cause high-latitude invasions and high local-extinction rates in the tropics and semi-enclosed seas ( ''medium confidence'' ).

|-
|
|-
| ''SROCC ( [[#Bindoff--2019a|Bindoff et al., 2019a]] )''

|
|-
| ‘Ocean warming has contributed to observed changes in biogeography of organisms ranging from phytoplankton to marine mammals ( ''high confidence'' ).’

‘The direction of the majority of the shifts of epipelagic organisms are consistent with a response to warming ( ''high confidence'' )’ but are also shaped by oxygen concentrations and ocean currents across depth, latitudinal and longitudinal gradients ( ''high confidence'' ).

Geographic ranges have shifted since the 1950s by 51.5 ± 33.3 km per decade (mean and ''very likely'' range) and 29.0 ± 15.5 km per decade for organisms in the epipelagic and seafloor ecosystems, respectively.

| ‘Recent model projections since AR5 and SR15 continue to support global-scale range shifts of marine fishes at rates of tens to hundreds of km per decade in the 21st century, with rate of shifts being substantially higher under RCP8.5 than RCP2.6.’

|}

Thermal tolerances of epipelagic populations drive biogeographic change (Figures 3.10, 3.15), but the strength and direction of range shifts tend to be modulated by both climate-induced and non-climate drivers ( [[#Pinsky--2020b|Pinsky et al., 2020b]] ), including: (a) interactive effects of hypoxia and ocean acidification ( [[#Sampaio--2021|Sampaio et al., 2021]] ); (b) oceanic dispersal barriers ( [[#Choo--2021|Choo et al., 2021]] ), food and critical habitat availability ( [[#Alabia--2020|Alabia et al., 2020]] ; [[#Tanaka--2021|Tanaka et al., 2021]] ); (c) geographic position, including depth ( [[#Mardones--2021|Mardones et al., 2021]] ); and (d) ocean currents ( [[#Sunday--2015|Sunday et al., 2015]] ; [[#Chapman--2020|Chapman et al., 2020]] ; [[#Fuchs--2020|Fuchs et al., 2020]] ). The difference between physiological thermal tolerances ( [[#3.3.2|Section 3.3.2]] ) and local environmental conditions determines safety margins against future climate warming in ectotherms ( [[#Pinsky--2019|Pinsky et al., 2019]] ). Acclimation and evolution ( [[#3.3.4|Section 3.3.4]] ) and life-history stage ( [[#3.3.3|Section 3.3.3]] ) also alter species’ thermal tolerances. Biogeographic responses are further modulated by other interacting factors (Table 3.17).

'''Table 3.17 |''' Synthesis of selected processes conditioned by multiple environmental drivers that interact with warming to ultimately define range-shift responses

{| class="wikitable"
|-
! Factor

! Effect

! Example references

|-
| Evolution and acclimation

| Evolution of thermal tolerances and acclimation under local climatic conditions can increase resilience to future climate warming, slowing the loss of species at trailing (warm) range edges.

| [[#Palumbi--2014|Palumbi et al. (2014)]] ; [[#Miller--2020a|Miller et al. (2020a)]]

|-
| rowspan="4"| Marine heatwaves (MHWs)

| MHWs can influence the evolution of thermal tolerances by eliminating genotypes that are intolerant of elevated temperatures.

| [[#Buckley--2016|Buckley and Huey (2016)]] ; [[#Sunday--2019|Sunday et al. (2019)]]

|-
| MHWs can produce widespread die-offs of shallow-water benthic organisms triggering extensive contractions of their ranges.

| [[#Smale--2013|Smale and Wernberg (2013)]]

|-
| MHWs can facilitate range expansions by opening niches and/or enhancing recruitment of warm-affiliated species.

| [[#Leriorato--2019|Leriorato and Nakamura (2019)]] ; [[#Thomsen--2019|Thomsen et al. (2019)]] ; [[#Monaco--2021|Monaco et al. (2021)]]

|-
| Cold waves can halt or even reverse range expansions at leading edges.

| [[#Leriorato--2019|Leriorato and Nakamura (2019)]]

|-
| rowspan="3"| Ocean currents

| Ocean currents can influence range dynamics through their effect on dispersal, depending on their magnitude, direction and seasonal patterns.

| Hunt et al. (2016); [[#Kumagai--2018|Kumagai et al. (2018)]] ; [[#Fuchs--2020|Fuchs et al. (2020)]]

|-
| Where currents align with spatial gradients of warming, range expansions track thermal changes more closely. Conversely, directional mismatches result in consistently slower expansion rates and larger response lags, an effect more acute for benthic organisms relying on passive dispersion of larvae and propagules.

| García Molinos et al. (2017)

|-
| Rates of range contraction across taxa decreased (increased) under directional agreement (mismatch) with ocean currents, possibly associated with enhanced (reduced) flows of adaptive genes to warming in downstream (upstream) populations within the distributional range.

| García Molinos et al. (2017)

|-
| rowspan="2"| Climatic refugia

| Areas of locally stable climatic conditions, such as deeper waters or regions with internal tides or localised upwelling, can buffer the effects of regional warming, facilitating species persistence and conserving genetic diversity at rear-edge populations.

| [[#Smith--2014|Smith et al. (2014)]] ; [[#Assis--2016|Assis et al. (2016)]] ; [[#Lourenço--2016|Lourenço et al. (2016)]] ; [[#Wyatt--2020|Wyatt et al. (2020)]]

|-
| Distributional shifts into deeper, cooler habitats can offer an effective alternative response to latitudinal shifts, because sharper thermal gradients mean that vertical displacements, needed to compensate for the same amount of warming, are several orders of magnitude smaller than planar displacements.

| [[#Smith--2014|Smith et al. (2014)]] ; [[#Assis--2016|Assis et al. (2016)]] ; [[#Lourenço--2016|Lourenço et al. (2016)]]

|-
| rowspan="2"| Oxygen availability

| Oxygen supersaturation may extend ectotherm survival to extreme temperatures and increase thermal tolerances by compensating for the increasing metabolic demand at high temperatures.

| [[#Giomi--2019|Giomi et al. (2019)]]

|-
| Oxygen deprivation increases metabolic demand and respiration rates. Shallowing of oxygen-dead zones and subsequent hypoxic avoidance can render deep thermal refuges unsuitable for organisms.

| [[#Brown--2015|Brown and Thatje (2015)]] ; [[#Roman--2019|Roman et al. (2019)]] ; [[#Hughes--2020|Hughes et al. (2020)]]

|-
| Habitat availability and quality

| The availability and quality of habitat (underwater light conditions, adequate substrate, nutrient and food supply) set limits to the distribution of organisms and range-shift dynamics (e.g., resilience of populations to climate warming and the consolidation of range expansions).

| Krause- [[#Jensen--2019|Jensen et al. (2019)]] ; [[#Tamir--2019|Tamir et al. (2019)]]

|-
| rowspan="2"| Biotic interactions, including food availability

| Species interactions can confer resilience to warming by retarding habitat degradation and buffering the impacts of warming on organisms.

| [[#Falkenberg--2015|Falkenberg et al. (2015)]] ; [[#Giomi--2019|Giomi et al. (2019)]]

|-
| Changes in biotic interactions (e.g., altered predation rates, food availability, competition or trophic mismatches) induced by climate warming can modify range-shift dynamics.

| [[#Selden--2018|Selden et al. (2018)]] ; [[#Westerbom--2018|Westerbom et al. (2018)]] ; Figueira et al. (2019); Pinsky et al. (2020b); [[#Monaco--2021|Monaco et al. (2021)]]

|}

<div id="_idContainer057" class="Figure"></div>

[[File:4e2baa6bf62f0f7f0fb192daff36a272 IPCC_AR6_WGII_Figure_3_015.png]]

'''Figure 3.15 |''' '''Range-shift dynamics in marine ectotherms in response to climate warming.''' As the ocean warms, conditions at the edge of the species’ distribution may become warmer than the maximum thermal tolerance of the species (Figure 3.9), causing local populations to undergo a gradual decline in performance, a decreasing population size and ultimately their extirpation, resulting in a range contraction. Conversely, at the cool extreme of the distribution, habitats beyond the current range of the species will become thermally suitable in the future (i.e., within the species’ thermal tolerance range) and, providing the species can disperse to those locations, allow for the colonisation and consolidation of new populations and subsequent range expansion. These are processes conditioned by multiple drivers that interact with warming to ultimately define range-shift responses, some of which are described in Table 3.17. Note that physiological thermal tolerances relate to body temperatures of the organism rather than ambient temperatures.

A global meta-analysis of range shifts ( [[#Lenoir--2020|Lenoir et al., 2020]] ) that included data from 951 species (over half of which exhibited median range shifts consistent with climate change) estimates that marine species are moving poleward at a rate of 59.2 km per decade ( ''very likely'' range: 43.7–74.7 km per decade), closely matching the local climate velocity ( ''high confidence'' ). In some cases, warming-related distribution shifts were followed by density-dependent use of these areas, influencing associated fisheries ( [[#Baudron--2020|Baudron et al., 2020]] ), and in others, warming influenced competitive interactions: in the Arctic-Boreal Barents Sea, warming-induced increases in cod ( ''Gadus morhua'' ) abundance reduces haddock ( ''Melanogrammus aeglefinus'' ) abundance ( [[#Durant--2020|Durant et al., 2020]] ).

Biogeographic shifts lead to novel communities and biotic interactions ( ''high confidence'' ) ( [[#Zarco-Perello--2017|Zarco-Perello et al., 2017]] ; [[#Pecuchet--2020b|Pecuchet et al., 2020b]] ), with concomitant changes in ecosystem functioning and servicing ( ''high confidence'' ) ( [[#Vergés--2019|Vergés et al., 2019]] ; [[#Nagelkerken--2020|Nagelkerken et al., 2020]] ; [[#Peleg--2020|Peleg et al., 2020]] ). For instance, temperature-driven changes in distribution and abundance of copepods, the dominant zooplankton, were observed between 1960 and 2014 in the North Atlantic. These changes subsequently affect biogenic carbon cycling through alteration of microbial remineralisation and carbon sequestration in deep water ( ''medium confidence'' ) ( [[#3.4.3|Section 3.4.3.6]] ; [[#Pitois--2006|Pitois and Fox, 2006]] ; [[#Brun--2019|Brun et al., 2019]] ).

<div id="3.4.3.1.2" class="h4-container"></div>

<span id="observed-vertical-redistributions"></span>
===== 3.4.3.1.2 Observed vertical redistributions =====

<div id="h4-5-siblings" class="h4-siblings"></div>

Epipelagic isotherms have recently (1980–2015) deepened at an average of 6.6 ± 18.8 m per decade ( [[#Pinsky--2019|Pinsky et al., 2019]] ), but there is ''low agreement'' on whether species move deeper in pursuit of thermal refuge. Prior studies suggested range shifts to depth ( [[#Dulvy--2008|Dulvy et al., 2008]] ; [[#Pinsky--2013|Pinsky et al., 2013]] ; [[#Yemane--2014|Yemane et al., 2014]] ), but increasing evidence suggests that fish and planktonic communities across large parts of the North Atlantic, sub-Arctic and northeast Pacific Ocean redistribute horizontally with horizontal climate velocity, except where vertical temperature gradients are particularly steep. There is ''low confidence'' for temperature-driven depth shifts in the epipelagic zone ( [[#Burrows--2019|Burrows et al., 2019]] ; [[#Campana--2020|Campana et al., 2020]] ; [[#Caves--2021|Caves and Johnsen, 2021]] ). At the same time, decreasing oxygen concentrations and the vertical expansion of OMZs have already decreased suitable habitat of pelagic fishes, including tuna and billfishes, by ~15% primarily due to vertical compression of environmental niches ( [[#Stramma--2012|Stramma et al., 2012]] ; [[#Deutsch--2015|Deutsch et al., 2015]] ).

<div id="3.4.3.1.3" class="h4-container"></div>

<span id="projected-changes-in-species-range-shifts"></span>
===== 3.4.3.1.3 Projected changes in species range shifts =====

<div id="h4-6-siblings" class="h4-siblings"></div>

Continued changes in the biogeography of marine predators and prey are anticipated under future climate change, with climate velocity in the epipelagic zone during 2050–2100 under RCP8.5 projected to be sevenfold faster than that during 1955–2005 ( ''medium confidence'' ) (Figure 3.4; [[#Brito-Morales--2020|Brito-Morales et al., 2020]] ). This has substantial ecological implications, as projections suggest near elimination of overlaps between the distributions of certain predator–prey pairs in the northeast Atlantic Ocean when their current joint distributions (1989–2014) are compared with those projected (2037–2062) under RCP8.5 ( [[#Sadykova--2020|Sadykova et al., 2020]] ).

Deepening of epipelagic isotherms is projected to accelerate over 2006–2100 to rates of 8.5 m per decade under RCP4.5 and 32 m per decade under RCP8.5 ( [[#Jorda--2020|Jorda et al., 2020]] ). Although vertical redistribution of thermal niches is three to four orders of magnitude slower than horizontal displacement, maximum depth limits imposed by the seafloor and photic layer (both of which are projected to be reached in this century) will ''likely'' vertically compress suitable habitat for most marine organisms ( ''medium confidence'' ) ( [[#Dueri--2014|Dueri et al., 2014]] ; [[#Jorda--2020|Jorda et al., 2020]] ).

Projections from coupled biogeochemical and ecosystem models suggest a general decline in mesopelagic biomass ( [[#Lefort--2015|Lefort et al., 2015]] ), although this may vary among ocean basins. The volume of OMZs have been expanding at many locations ( ''high confidence'' ), and the oxygen content of the subsurface ocean is projected to decline to historically unprecedented conditions over the 21st century ( ''medium confidence'' ) ( [[#3.2.3|Section 3.2.3.2]] ; WGI AR6 [[IPCC:Wg2:Chapter:Chapter-5#5.3|Section 5.3.3.2]] ; [[#Canadell--2021|Canadell et al., 2021]] ) at a rate of 10–15 µM per decade in OMZs ( [[#3.2.3|Section 3.2.3.2]] ; [[#Breitburg--2018|Breitburg et al., 2018]] ). Oxygen availability and the effects of ocean acidification (Sections 3.3, 3.4.2) on zooplankton might become a dominant constraint in the upper ocean’s metabolic index, which is projected to decrease globally by 20% by 2100 ( [[#Deutsch--2015|Deutsch et al., 2015]] ; [[#Steinberg--2017|Steinberg and Landry, 2017]] ). In addition, extremely rapid acceleration of climate velocities projected in the mesopelagic under all emissions scenarios suggest that species in this ocean stratum will be even more exposed to future warming than species in the epipelagic (Figure 3.4; [[#Brito-Morales--2020|Brito-Morales et al., 2020]] ). But projections also suggest that warming-related increases in trophic efficiency lead to a 17% increase in the biomass of the deep-scattering layer (zooplankton and fish in the mesopelagic) by 2100 (low ''confidence'' ) ( [[#Bindoff--2019a|Bindoff et al., 2019a]] ). Observational studies appear to show that mesopelagic fishes adapted to warm water increased in abundance and distribution in the California Current associated with warming and the expansion of OMZ ( [[#Koslow--2019|Koslow et al., 2019]] ), suggesting that some mesopelagic fish stocks might be resilient to a changing climate ( ''medium confidence'' ).

<div id="3.4.3.2" class="h3-container"></div>

<span id="phenological-shifts-and-trophic-mismatches"></span>
==== 3.4.3.2 Phenological Shifts and Trophic Mismatches ====

<div id="h3-24-siblings" class="h3-siblings"></div>

<div id="3.4.3.2.1 " class="h4-container"></div>

<span id="observed-changes"></span>
===== 3.4.3.2.1 Observed changes =====

<div id="h4-7-siblings" class="h4-siblings"></div>

SROCC reported ''high confidence'' in phenological shifts towards earlier onset of biological events (Table 3.18), with phenological shifts among epipelagic species attributed to ocean warming ( ''high confidence'' ).

'''Table 3.18 |''' Summary of previous IPCC assessments of phenological shifts and trophic mismatches

{| class="wikitable"
|-
! Observations

! Projections

|-
| ''AR5 WGII ( [[#Hoegh-Guldberg--2014|Hoegh-Guldberg et al., 2014]] ; [[#Larsen--2014|Larsen et al., 2014]] )''

|
|-
| ‘Changes to sea temperature have altered the phenology, or timing of key life-history events such as plankton blooms, and migratory patterns, and spawning in fish and invertebrates, over recent decades ( ''medium confidence'' ). There is ''medium to high agreement'' that these changes pose significant uncertainties and risks to fisheries, aquaculture and other coastal activities.’

The highly productive high-latitude spring bloom systems in the northeast Atlantic are responding to warming ( ''medium evidence, high agreement'' ), with the greatest changes being observed since the late 1970s in the phenology, distribution and abundance of plankton assemblages, and the reorganisation of fish assemblages, with a range of consequences for fisheries ( ''high confidence'' ).

‘Observed changes in the phenology of plankton groups in the North Sea over the past 50 years are driven by climate forcing, in particular regional warming ( ''high confidence'' ).’

‘On average, spring events in the ocean have advanced by 4.4 ± 0.7 days per decade (mean ± SE).’

‘Shifts in the timing and magnitude of seasonal biomass production could disrupt matched phenologies in the food webs, leading to decreased survival of dependent species ( ''medium confidence'' ). If the timing of primary and secondary production is no longer matched to the timing of spawning or egg release, survival could be impacted, with cascading implications to higher trophic levels. This impact would be exacerbated if shifts in timing occur rapidly ( ''medium confidence'' ).’

‘There is ''medium to high confidence'' that climate-induced disruptions in the synchrony between timing of spawning and hatching of some fish and shellfish and the seasonal increases in prey availability can result in increased larval or juvenile mortality or changes in the condition factor of fish and shellfish species in the Arctic marine ecosystems.’

| Projections of phenological shifts and trophic mismatches were not assessed in this report.

|-
|
|-
| ''SROCC ( [[#Bindoff--2019a|Bindoff et al., 2019a]] )''

|
|-
| ‘Phenology of marine ectotherms in the epipelagic systems related to ocean warming ( ''high confidence'' ) and the timing of biological events has shifted earlier ( ''high confidence'' ).’

‘Timing of spring phenology of marine organisms is shifting to earlier in the year under warming, at an average rate of 4.4 ± 1.1 days per decade, although it is variable among taxonomic groups and among ocean regions.’

| Projections of phenological shifts and trophic mismatches were not assessed in this report.

|-
|
|-
| ''WGI AR6 [[IPCC:Wg2:Chapter:Chapter-2|Chapter 2]] ( [[#Gulev--2021|Gulev et al., 2021]] )''

|
|-
| ‘Phenological metrics for many species of marine organisms have changed in the last half century ( ''high confidence'' ), though many regions and many species of marine organisms remain under-sampled or even unsampled. The changes vary with location and with species ( ''high confidence'' ). There is a strong dependence of survival in higher trophic-level organisms (fish, exploited invertebrates, birds) on the availability of food at various stages in their life cycle, which in turn depends on the phenologies of both ( ''high confidence'' ). There is a gap in our understanding of how the varied responses of marine organisms to climate change, from a phenological perspective, might threaten the stability and integrity of entire ecosystems.’

| Projections of phenological shifts and trophic mismatches were not assessed in this report.

|}

'''Table 3.19 |''' Assessment of phenological shifts by taxon based on time series from field observations spanning at least 19 years published over the past 25 years

{| class="wikitable"
|-
! Taxon

! Rate of consistency of observations with climate change

! Estimated mean rate of change in seasonal timing

! Confidence

! Notes

|-
| Phytoplankton

| 78.41%

( ''n'' = 85)

| −7.5 d per decade

( ''n'' = 83)

| ''Very high confidence''

| ''Evidence most robust'' for changes in timing of blooms in the North Atlantic (e.g., [[#Chivers--2020|Chivers et al., 2020]] ) and Baltic (e.g., [[#Scharfe--2019|Scharfe and Wiltshire, 2019]] ; [[#Wasmund--2019|Wasmund et al., 2019]] ), with ''limited evidence'' from the Southern Hemisphere.

|-
| Holozooplankton

| 79.74%

( ''n'' = 77)

| −4.27 d per decade

( ''n'' = 58)

| ''Very high confidence''

| ''Evidence most robust'' in the northeast Atlantic (e.g., [[#Chevillot--2017|Chevillot et al., 2017]] ), but sparse elsewhere.

|-
| Meroplankton (taxa that are only temporarily in the plankton)

| 81.06%

( ''n'' = 72)

| −4.34 d per decade

( ''n'' = 64)

| ''Very high confidence''

| Includes earlier peak abundance of fish larvae in upwelling systems (e.g., [[#Asch--2015|Asch, 2015]] ).

|-
| Benthic invertebrates

| 72.34%

( ''n'' = 5)

| −8.5 d per decade

( ''n'' = 5)

| ''Low confidence'' ( ''limited evidence, medium agreement'' )

| ''Evidence is limited'' , uncertainty levels are high. Rate of consistency of responses with climate change is not significantly different from random chance.

|-
| Plants

| 100%

( ''n'' = 1)

| No estimate available

| ''Very low confidence''

| Just a single study for seagrasses, and only for consistency ( [[#Diaz-Almela--2007|Diaz-Almela et al., 2007]] ).

|-
| Fish

| 65.48%

( ''n'' = 109)

| −3.02 d per decade

( ''n'' = 43)

| ''Very high confidence''

| Includes earlier appearance of migratory fish in estuaries (e.g., [[#Chevillot--2017|Chevillot et al., 2017]] ), earlier spawning migrations for anadromous fish such as salmon (e.g., [[#Rubenstein--2019|Rubenstein et al., 2019]] ), earlier migrations for sole (e.g., [[#Fincham--2013|Fincham et al., 2013]] ) and tuna (e.g., [[#Dufour--2010|Dufour et al., 2010]] ), and earlier spawning of key commercial demersal (bottom-dwelling) species such as cod (e.g., [[#McQueen--2017|McQueen and Marshall, 2017]] ).

|-
| Marine reptiles

| 100.0%

( ''n'' = 4)

| −2.89 d per decade

( ''n'' = 4)

| ''Low confidence'' ( ''limited evidence, low agreement'' )

| ''Evidence is limited'' , uncertainty levels are high. Mean phenological shift is not significantly different from zero.

|-
| Seabirds

| 42.36%

( ''n'' = 56)

| +0.77 d per decade

( ''n'' = 51)

| ''Very low confidence'' ( ''limited evidence, low agreement'' )

| Neither the rate of consistency with climate change nor the phenological shift differ significantly from null expectations (50% consistency and no shift). Many seabirds are breeding earlier ( [[#Byrd--2008|Byrd et al., 2008]] ; [[#Sydeman--2009|Sydeman et al., 2009]] ), while breeding among others in temperate and polar regions has been delayed, which has been linked to later sea ice breakup or limited prey resources ( [[#Barbraud--2006|Barbraud and Weimerskirch, 2006]] ; [[#Wanless--2009|Wanless et al., 2009]] ; [[#Chambers--2014|Chambers et al., 2014]] ). Although the response of lifecycle events for many seabird species is variable in direction, there has usually been a more complex driver associated with climate that has been considered to be responsible ( [[#Sydeman--2015|Sydeman et al., 2015]] ). For many species, seasonal timing is moving earlier, especially in the Arctic (e.g., [[#Byrd--2008|Byrd et al., 2008]] ; [[#Descamps--2019|Descamps et al., 2019]] ), but for many species in the Southern Ocean, it is not ( [[#Barbraud--2006|Barbraud and Weimerskirch, 2006]] ; [[#Chambers--2014|Chambers et al., 2014]] ). This could be because of a much slower rate of warming in most of the Southern Ocean than in the Arctic.

|-
| Marine mammals

| 100.0%

( ''n'' = 4)

| −0.34 d per decade ( ''n'' = 4)

| ''Very low confidence'' ( ''limited evidence, low agreement'' )

| All studies of phenological changes for marine mammals have focused on whales (e.g., [[#Ramp--2015|Ramp et al., 2015]] ; [[#Hauser--2017|Hauser et al., 2017]] ; [[#Loseto--2018|Loseto et al., 2018]] ) or polar bears (e.g., [[#Cherry--2013|Cherry et al., 2013]] ; [[#Atwood--2016|Atwood et al., 2016]] ; [[#Escajeda--2018|Escajeda et al., 2018]] ) and have related timing to aspects of sea ice dynamics, highlighting the complexity of such processes. Mean phenological shift is not significantly different from zero at the global scale.

|}

Since SROCC, field data have continued to show that the phenology of biological events in the ocean is ''very likely'' ( ''high to very high confidence'' ) advancing in response to climate change, with 71.9% of published observations consistent with these anticipated effects (Figure 3.16a,b; Table 3.19), although most reports (95.6%) were from the Northern Hemisphere (Figure 3.16a). Biological events that are shifting earlier in response to climate change include phytoplankton blooms ( [[#Scharfe--2019|Scharfe and Wiltshire, 2019]] ; [[#Chivers--2020|Chivers et al., 2020]] ) such as: (a) those of HAB species ( [[#Forsblom--2019|Forsblom et al., 2019]] ; [[#Bucci--2020|Bucci et al., 2020]] ); (b) peaks in zooplankton abundance ( [[#Chevillot--2017|Chevillot et al., 2017]] ; [[#Forsblom--2019|Forsblom et al., 2019]] ); (c) the migration ( [[#Otero--2014|Otero et al., 2014]] ; [[#Kovach--2015|Kovach et al., 2015]] ; [[#Chust--2019|Chust et al., 2019]] ) and spawning of commercial fish ( [[#McQueen--2017|McQueen and Marshall, 2017]] ; [[#Kanamori--2019|Kanamori et al., 2019]] ), including crabs and squid ( [[#Henderson--2017|Henderson et al., 2017]] ); and (d) breeding of marine reptiles ( [[#Mazaris--2008|Mazaris et al., 2008]] ; [[#Cherkiss--2020|Cherkiss et al., 2020]] ). Moreover, different trophic levels within epipelagic food webs are responding at different rates ( ''very high confidence'' ) (Table 3.19; Figure 3.16b,c), with greater and more consistent responses by lower trophic levels (phytoplankton, holozooplankton and meroplankton) but less consistent, weaker and more varied responses by higher trophic levels. There were too few independent time series to make robust estimates for benthic invertebrates, plants, marine reptiles and mammals. This differential response across trophic levels could lead to trophic mismatches ( [[#Neuheimer--2018|Neuheimer et al., 2018]] ), where predators and their prey respond asynchronously to climate change ( [[#Edwards--2004|Edwards and Richardson, 2004]] ; [[#Rogers--2019|Rogers and Dougherty, 2019]] ; [[#Rubenstein--2019|Rubenstein et al., 2019]] ; [[#Émond--2020|Émond et al., 2020]] ), with potential population-level consequences, including declines in fish recruitment ( [[#Burthe--2012|Burthe et al., 2012]] ; [[#Chevillot--2017|Chevillot et al., 2017]] ; [[#McQueen--2017|McQueen and Marshall, 2017]] ; [[#Asch--2019|Asch et al., 2019]] ; [[#Durant--2019|Durant et al., 2019]] ; [[#Régnier--2019|Régnier et al., 2019]] ). Available evidence also suggests that feeding relationships could modulate species responses to climate change, as seen in breeding of surface-feeding and deeper-diving seabirds ( [[#Descamps--2019|Descamps et al., 2019]] ). These differential responses could determine ‘winners’ and ‘losers’ under future climate change ( [[#Lindén--2018|Lindén, 2018]] ).

<div id="_idContainer062" class="Figure"></div>

[[File:d1f25a14a1580fe6eb4b371e588bddc3 IPCC_AR6_WGII_Figure_3_016.png]]

'''Figure 3.16 |''' '''Observed responses to climate change based on a systematic Web of Science review of marine phenology studies exceeding 19 years in length to update the assessment in WGII AR5 Chapter 30 (Hoegh-Guldberg et al.''' ''', 2014).''' Error bars indicate 95% confidence limits (i.e., the ''extremely likely'' range).

'''(a)''' Global data shows changes in seasonal cycles of biota that are attributed (at least partly) to climate change (blue, ''n'' = 297 observations), and changes that are inconsistent with climate change (white, ''n'' = 116 observations). Each circle represents the centre of a study area.

'''(b)''' The proportion of phenological observations (showing means and ''extremely likely'' ranges) that are attributed to climate change (i.e., generally showing earlier timing) by taxonomic group.

'''(c)''' Observed shifts in timing (days per decade, showing means and ''extremely likely'' ranges), by taxonomic group, that are attributed to climate change. Negative shifts are earlier, positive shifts are later. (Details and additional plots are presented in 3.SM.3.3, Figure 3.SM.3 and Table 3.SM.1.)

<div id="3.4.3.2.2" class="h4-container"></div>

<span id="projected-changes"></span>
===== 3.4.3.2.2 Projected changes =====

<div id="h4-8-siblings" class="h4-siblings"></div>

The CMIP6 ESM ensembles project that, by 2100, 18.8 ± 19.0% (mean ± ''very likely'' range) and 38.9 ± 9.4% of the ocean surface will ''very likely'' undergo a change of 20 d or more (advance or delay) in the start of the phytoplankton growth period under SSP1-2.6 and SSP5-8.5, respectively (Figure 3.17a,b) ( ''low confidence'' due to the dependence with the projected changes in phytoplankton biomass, the trends of which are reported with ''low confidence'' ) ( [[#3.4.3.4|Section 3.4.3.4]] ; SROCC [[IPCC:Wg2:Chapter:Chapter-5#5.2|Section 5.2.3]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ). Phytoplankton growth is projected to begin later in the Northern Hemisphere subtropics, and earlier at high latitudes in some regions around the Antarctic Peninsula, and over large areas in the Northern Hemisphere ( ''low to medium confidence'' as there are improved constraints from historical variability in this region and consistency with CMIP5-based-studies results) ( [[#Henson--2018b|Henson et al., 2018b]] ; [[#Asch--2019|Asch et al., 2019]] ). There is ''high agreement'' in model projections that the start of the phytoplankton growth period will ''very likely'' advance in the Arctic Ocean under a high-emission scenario for CMIP5 and CMIP6 (Figure 3.17b; [[#Henson--2018b|Henson et al., 2018b]] ; [[#Asch--2019|Asch et al., 2019]] ; [[#Tedesco--2019|Tedesco et al., 2019]] ; [[#Lannuzel--2020|Lannuzel et al., 2020]] ). The CMIP6 ensemble projections further show limited changes in phenology across most of the Southern Ocean but large regional variations in the tropics (Figure 3.17). Overall, the regional patterns are qualitatively similar under SSP1-2.6 and SSP5-8.5 but with greater magnitude and larger areas under SSP5-8.5 ( ''low confidence'' ).

<div id="_idContainer065" class="Figure"></div>

[[File:aae03dd90e97291b745b7a026d4210dd IPCC_AR6_WGII_Figure_3_017.png]]

'''Figure 3.17 |''' '''Projected phytoplankton phenology.''' (a,c) Spatial patterns and (b,d) density distributions of projected change in phytoplankton phenology by 2100 under Shared Socioeconomic Pathway (SSP)1-2.6 and SSP5-8.5, respectively. Difference in the start of the phytoplankton growth period is calculated as 2090–2099 minus 1996–2013. Negative (positive) values indicate earlier (later) start of the phytoplankton growth period by 2100. The ensemble projections of global changes in phytoplankton phenology include, under SSP1-2.6 and SSP5-8.5, respectively, a total of five Coupled Model Intercomparison Project 6 Earth system models containing coupled ocean biogeochemical models that cover a wide range of complexity ( [[#Kwiatkowski--2019|Kwiatkowski et al., 2019]] ). (The phenology calculations are based on [[#Racault--2017|Racault et al., 2017]] , using updated data.)

At latitudes &gt;40°N, temperature-linked phenology of fish reproduction with high geographic fidelity to spawning grounds (geographic spawners) is projected to change at double the rate of that for phytoplankton, which will ''likely'' cause phenological mismatches resulting in increased risk of starvation for fish larvae ( ''medium to high confidence'' ) (WGI AR6 [[IPCC:Wg2:Chapter:Chapter-2#2.3|Section 2.3.4.2.3]] ; [[#Asch--2019|Asch et al., 2019]] ; [[#Durant--2019|Durant et al., 2019]] ; [[#Régnier--2019|Régnier et al., 2019]] ; [[#Gulev--2021|Gulev et al., 2021]] ; [[#Laurel--2021|Laurel et al., 2021]] ). Furthermore, under RCP8.5, trophic mismatch events exceeding ±30 days ( [[#Asch--2019|Asch et al., 2019]] ) leading to fish-recruitment failure are expected to increase tenfold for geographic spawners across much of the North Atlantic, North Pacific and Arctic Ocean basins ( ''low confidence'' ) ( [[#Neuheimer--2018|Neuheimer et al., 2018]] ). In contrast, temporal mismatches between fish that relocate spawning grounds in response to environmental variations (environmental spawners) and phytoplankton blooms are projected to remain shorter and less varied, suggesting that across ocean basins, range shifts by environmental spawners may increase their resilience. Nevertheless, this compensation mechanism might fail at locations where phytoplankton bloom phenology is not controlled by temperature-driven water-column stratification, leading to a possible sixfold local increase in extreme mismatches under climate change ( [[#Asch--2019|Asch et al., 2019]] ).

<div id="3.4.3.3" class="h3-container"></div>

<span id="changes-in-community-composition-and-biodiversity"></span>
==== 3.4.3.3 Changes in Community Composition and Biodiversity ====

<div id="h3-25-siblings" class="h3-siblings"></div>

<div id="3.4.3.3.1" class="h4-container"></div>

<span id="evidence-of-natural-adaptive-capacity-based-on-species-responses-to-past-climate-variability"></span>
===== 3.4.3.3.1 Evidence of natural adaptive capacity based on species’ responses to past climate variability =====

<div id="h4-9-siblings" class="h4-siblings"></div>

Responses to abrupt climate change in the geological past suggest that adaptive capacity is limited for marine animals (Cross-Chapter Box PALEO in Chapter 1). Temperatures during the last Interglacial (~125 ka), which were warmer than today, led to poleward range shifts of reef corals ( ''medium confidence'' ) ( [[#Kiessling--2012|Kiessling et al., 2012]] ; [[#Jones--2019a|Jones et al., 2019a]] ). Temperature has also driven marine range shifts over multi-million-year time scales ( ''medium confidence'' ) ( [[#Gibbs--2016|Gibbs et al., 2016]] ; [[#Reddin--2018|Reddin et al., 2018]] ). Warming climates, even with low ocean-warming rates, inevitably decreased tropical marine biodiversity compared with middle latitudes ( ''high confidence'' ) ( [[#Mannion--2014|Mannion et al., 2014]] ; [[#Crame--2020|Crame, 2020]] ; [[#Yasuhara--2020|Yasuhara et al., 2020]] ; [[#Raja--2021|Raja and Kiessling, 2021]] ).

The paleorecord confirms that marine biodiversity has been vulnerable to climate warming both globally and regionally ( ''very high confidence'' ) (Cross-Chapter Box PALEO in Chapter 1; [[#Stanley--2016|Stanley, 2016]] ). In extreme cases of warming (e.g., &gt;5.2°C), marine mass extinctions occurred in the geological past, and there may be a relationship between warming magnitude and extinction toll ( ''medium confidence'' ) ( [[#Song--2021b|Song et al., 2021b]] ). A combination of warming and spreading anoxia caused marine extinctions in ancient episodes of rapid climate warming ( ''high confidence'' ) ( [[#Bond--2017|Bond and Grasby, 2017]] ; [[#Benton--2018|Benton, 2018]] ; [[#Penn--2018|Penn et al., 2018]] ; [[#Them%20II--2018|Them II et al., 2018]] ; [[#Chen--2019|Chen and Xu, 2019]] ). The role of ocean acidification in ancient extinctions is yet to be resolved ( ''low confidence'' ) ( [[#Clapham--2011|Clapham and Payne, 2011]] ; [[#Clarkson--2015|Clarkson et al., 2015]] ; [[#Jurikova--2020|Jurikova et al., 2020]] ; [[#Müller--2020|Müller et al., 2020]] ).

<div id="3.4.3.3.2" class="h4-container"></div>

<span id="observed-and-projected-changes-in-contemporary-community-structure-and-biodiversity"></span>
===== 3.4.3.3.2 Observed and projected changes in contemporary community structure and biodiversity =====

<div id="h4-10-siblings" class="h4-siblings"></div>

Ocean temperature is a major driver of species richness in the global ocean at evolutionary time scales ( [[#Tittensor--2010|Tittensor et al., 2010]] ; [[#Chaudhary--2021|Chaudhary et al., 2021]] ). This, together with temperature-driven range and phenology shifts evident across taxa and ocean ecosystems (Sections 3.4.3.1, 3.4.3.2), suggests that recent ocean warming ( [[#3.2.2.1|Section 3.2.2.1]] ) should alter biodiversity at regional to global scales. Since previous assessments (Table 3.20), the most common evidence supporting these expected changes is replacement of cold-adapted species by warm-adapted species within an ecosystem as waters warm ( [[#Worm--2021|Worm and Lotze, 2021]] ). Known as tropicalisation ( [[#3.4.2.3|Section 3.4.2.3]] ), this phenomenon has been attributed to ocean warming ( ''medium to high confidence'' ) in communities as diverse as kelp, invertebrates, plankton and fish ( [[#Burrows--2019|Burrows et al., 2019]] ; [[#Flanagan--2019|Flanagan et al., 2019]] ; [[#Ajani--2020|Ajani et al., 2020]] ; [[#Villarino--2020|Villarino et al., 2020]] ; [[#Punzón--2021|Punzón et al., 2021]] ; [[#Smith--2021|Smith et al., 2021]] ).

'''Table 3.20 |''' Summary of previous IPCC assessments of community composition and biodiversity

{| class="wikitable"
|-
! Observations

! Projections

|-
| ''AR5 ( [[#Hoegh-Guldberg--2014|Hoegh-Guldberg et al., 2014]] ; [[#Pörtner--2014|Pörtner et al., 2014]] )''

|
|-
| The paleoecological record shows that global climate changes comparable in magnitudes to those projected for the 21st century under all scenarios resulted in large-scale biome shifts and changes in community composition, and that for rates projected under RCP6 and 8.5 those changes were associated with species extinctions in some groups ( ''high confidence'' ).

Loss of corals due to bleaching has a potentially critical influence on the maintenance of marine biodiversity in the tropics ( ''high confidence'' ).

| Spatial shifts of marine species due to projected warming will cause high-latitude invasions and high local-extinction rates in the tropics and semi-enclosed seas ( ''medium confidence'' ).

Species richness and fisheries catch potential are projected to increase, on average, at mid and high latitudes ( ''high confidence'' ) and decrease at tropical latitudes ( ''medium confidence'' ).

‘Shifts in the geographical distributions of marine species [...] cause changes in community composition and interactions [...]. Thereby, climate change will reassemble communities and affect biodiversity, with differences over time and between biomes and latitudes ( ''high confidence'' ).’

‘Models are currently useful for developing scenarios of directional changes in net primary productivity, species distributions, community structure, and trophic dynamics of marine ecosystems, as well as their implications for ecosystem goods and services under climate change. However, specific quantitative projections by these models remain imprecise ( ''low confidence'' ).’

|-
|
|-
| ''SROCC ( [[#Bindoff--2019a|Bindoff et al., 2019a]] )''

|
|-
| ‘Ocean warming has contributed to observed changes in biogeography of organisms ranging from phytoplankton to marine mammals ( ''high confidence'' ), consequently changing community composition ( ''high confidence'' ), and in some cases altering interactions between organisms and ecosystem function ( ''medium confidence'' ).’

| Poleward range shifts are projected to decrease species richness in tropical oceans, counterbalanced by increases in mid- to high-latitude regions, leading to global-scale species turnover ( ''medium confidence'' on trends, ''low confidence'' on magnitude because of model uncertainties and the limited number of published model simulations). ‘The projected intensity of species turnover is lower under low-emission scenarios ( ''high confidence'' ).’

‘Projections from multiple fish species distribution models show hotspots of decrease in species richness in the Indo-Pacific region, and semi-enclosed seas such as the Red Sea and Persian Gulf ( ''medium evidence, high agreement'' ). In addition, geographic barriers, such as land, [bounding the] poleward species range edge in semi-enclosed seas or low-oxygen water in deeper waters are projected to limit range shifts, resulting in a larger relative decrease in species richness ( ''medium confidence'' ).’

‘The large variation in sensitivity between different zooplankton taxa to future conditions of warming and ocean acidification suggests elevated risk on community structure and inter-specific interactions of zooplankton in the 21st century ( ''medium confidence'' ).’

|}

At local to regional scales, tropicalisation often increases species richness where warm-water species extend their ranges to overlap with existing communities and decreases species richness where warming waters extirpate species ( ''medium to high confidence'' ) ( [[#Friedland--2020a|Friedland et al., 2020a]] ; [[#Chaudhary--2021|Chaudhary et al., 2021]] ; [[#Worm--2021|Worm and Lotze, 2021]] ). Latitudinal estimates from catalogued observations show declining species richness in equatorial waters over the past 50 years, with concomitant increases in species richness at mid-latitudes; the pattern is especially prominent in free-swimming pelagic species (Figure 3.18; [[#Chaudhary--2021|Chaudhary et al., 2021]] ). Similar patterns among marine animals have been described previously for historical warming events ( [[#Song--2020b|Song et al., 2020b]] ). Tropicalisation is associated with increased representation of herbivorous species ( [[#Vergés--2016|Vergés et al., 2016]] ; [[#Zarco-Perello--2020|Zarco-Perello et al., 2020]] ; [[#Smith--2021|Smith et al., 2021]] ), although observations and theory suggest that dietary generalism can also favour range-shifting species ( [[#Monaco--2020|Monaco et al., 2020]] ; [[#Wallingford--2020|Wallingford et al., 2020]] ).

<div id="_idContainer068" class="Figure"></div>

[[File:775d59d4d5b1b9873ed3788209234bbd IPCC_AR6_WGII_Figure_3_018.png]]

'''Figure 3.18 |''' '''Changes in the latitudinal distribution of marine species richness.'''

'''(a)''' Observed species richness for three historical periods. The observed latitudinal patterns in species richness are for a suite of taxonomic groups based on 48,661 marine species ( [[#Chaudhary--2021|Chaudhary et al., 2021]] ).

'''(b)''' Projected changes in species richness under RCP4.5 and RCP8.5 are calculated as differences by grid cell by 2100 relative to 2006. Latitudinal global median (5° moving average). (Based on Figure 1b,c in [[#García%20Molinos--2016|García Molinos et al., 2016]] .) The projected latitudinal patterns in changes in species richness under climate change are based on a numerical model that includes species-specific information across a suite of taxonomic groups, based on 12,796 marine species ( [[#García%20Molinos--2016|García Molinos et al., 2016]] ).

At the community level, the magnitude and shape of projected future biodiversity changes differ depending on which groups are considered ( ''medium confidence'' ) ( [[#Chaudhary--2021|Chaudhary et al., 2021]] ). Molecular-based richness measures indicate that the most dramatic increases in diversity relative to current conditions are expected for photosynthetic eukaryotes and copepods in the Arctic Ocean ( [[#Ibarbalz--2019|Ibarbalz et al., 2019]] ). However, component eukaryotic taxa, for example diatoms ( [[#Busseni--2020|Busseni et al., 2020]] ), are projected to lose diversity by 2100 under RCP8.5. Ecosystem models project a decline in nutrient supply that drives the disappearance of less-competitive and larger phytoplankton types, leading to extinction of up to 30% of diatom types, particularly in the Northern Hemisphere, by 2100 under RCP8.5 ( [[#Henson--2021|Henson et al., 2021]] ). Models further suggest that high latitudes are ''likely'' to encounter entirely novel phytoplankton communities by 2100 under RCP8.5 (100% change in community composition; [[#Dutkiewicz--2019|Dutkiewicz et al., 2019]] ; [[#Reygondeau--2020|Reygondeau et al., 2020]] ). At the polar edges, the increased richness is projected to coincide with high species turnover and increasing dominance of smaller phytoplankton types ( [[#Henson--2021|Henson et al., 2021]] ). These imply pronounced changes to both the oceans’ ecological and biogeochemical function, as regions dominated by small phytoplankton typically support less-productive food webs ( [[#3.4.3.4|Section 3.4.3.4]] ; [[#Stock--2017|Stock et al., 2017]] ; [[#Armengol--2019|Armengol et al., 2019]] ) and sequester less particulate organic carbon (POC) in the deep ocean ( [[#3.4.3.5|Section 3.4.3.5]] ; [[#Mouw--2016|Mouw et al., 2016]] ; [[#Cram--2018|Cram et al., 2018]] ) than areas dominated by larger size classes ( ''high confidence'' ).

The profound climatic and environmental changes projected for the Arctic region by 2100 (Cross-Chapter Paper 6) are also anticipated to alter the composition of apex assemblages like marine mammals (see Box 3.2; [[#Albouy--2020|Albouy et al., 2020]] ). Under both RCP2.6 and 8.5 scenarios the most vulnerable marine mammal species will be the North Pacific right whale ( ''Eubalaena japonica'' , listed as an endangered species; [[#IUCN--2020|IUCN, 2020]] ) and the grey whale ( ''Eschrichtius robustus'' , which has critically endangered subpopulations; [[#IUCN--2020|IUCN, 2020]] ). The extinction of the most-vulnerable species will disproportionately eliminate unique and important evolutionary lineages as well as functional diversity, with consequent impacts throughout the entire marine ecosystem ( [[#3.3.4|Section 3.3.4]] ). More generally, future warming and acidification simulated in mesocosm experiments support projections of a substantial increase in biomass and productivity of primary producers and secondary consumers, but a decrease by &gt;40% of primary consumers ( [[#Nagelkerken--2020|Nagelkerken et al., 2020]] ). On longer time scales, alteration of energy flow through marine food webs may lead to ecological tipping points ( [[#Wernberg--2016|Wernberg et al., 2016]] ; [[#Harley--2017|Harley et al., 2017]] ) after which the food web collapses into shorter, bottom-heavy trophic pyramids ( ''medium confidence'' ).

Global projections anticipate a ''likely'' future reorganisation of marine life of variable magnitude, contingent on emission scenario ( [[#Beaugrand--2015|Beaugrand et al., 2015]] ; [[#Jones--2015|Jones and Cheung, 2015]] ; [[#Barton--2016|Barton et al., 2016]] ; [[#García%20Molinos--2016|García Molinos et al., 2016]] ; [[#Nagelkerken--2020|Nagelkerken et al., 2020]] ; [[#Henson--2021|Henson et al., 2021]] ). Marine organism redistributions projected under RCP4.5 and RCP8.5 include extirpations and range contractions in the tropics, strongly decreasing tropical biodiversity, and range expansions at higher latitudes, associated with increased diversity and homogenisation of marine communities (Figure 3.18b). Under continuing climate change, the projected loss of biodiversity may ultimately threaten marine ecosystem stability ( ''medium confidence'' ) ( [[#Albouy--2020|Albouy et al., 2020]] ; [[#Nagelkerken--2020|Nagelkerken et al., 2020]] ; [[#Henson--2021|Henson et al., 2021]] ), altering both the functioning and structure of marine ecosystems and thus affecting service provisioning ( ''medium confidence'' ) ( [[#3.5|Section 3.5]] ; [[#Ibarbalz--2019|Ibarbalz et al., 2019]] ; [[#Righetti--2019|Righetti et al., 2019]] ).

However, biodiversity observations remain sparse, and statistical and modelling tools can provide conflicting diversity information (e.g., [[#Righetti--2019|Righetti et al., 2019]] ; [[#Dutkiewicz--2020|Dutkiewicz et al., 2020]] ) because correlative approaches assume that the modern-day relationship between marine species distribution and environmental conditions remains the same into the future, whereas mechanistic models permit marine species to respond dynamically to changing environmental forcing. Moreover, existing global projections of future biodiversity disproportionately focus on the effects sea surface temperature ( [[#Thomas--2012|Thomas et al., 2012]] ), typically overlooking other factors such as ocean acidification, deoxygenation and nutrient availability ( [[#3.2.3|Section 3.2.3]] ), and often failing to account for natural adaptation (e.g., [[#3.3.4|Section 3.3.4]] ; see Box 3.1; [[#Barton--2016|Barton et al., 2016]] ; [[#Henson--2021|Henson et al., 2021]] ).

<div id="3.4.3.3.3" class="h4-container"></div>

<span id="abrupt-ecosystem-shifts-and-extreme-events"></span>
===== 3.4.3.3.3 Abrupt ecosystem shifts and extreme events =====

<div id="h4-11-siblings" class="h4-siblings"></div>

Climate-change-driven changes in ocean characteristics and the frequency and intensity of extreme events ( [[#3.2|Section 3.2]] ) increase the risk of persistent, rapid and abrupt ecosystem change ( ''very high confidence'' ), often referred to as ecosystem collapses or regime shifts (AR6 WGI Chapter 9; [[#Collins--2019a|Collins et al., 2019a]] ; [[#Canadell--2021|Canadell and Jackson, 2021]] ; [[#Ma--2021|Ma et al., 2021]] ). Such abrupt changes include altering ecosystem structure, function and biodiversity outside the range of natural fluctuations ( [[#Collins--2019a|Collins et al., 2019a]] ; [[#Canadell--2021|Canadell and Jackson, 2021]] ). They can involve mass-mortality events and ‘tipping points’ or ‘critical transitions’, where strong positive feedbacks within an ecosystem lead to self-sustaining change (Figure 3.19a; [[#Scheffer--2012|Scheffer et al., 2012]] ; [[#Möllmann--2015|Möllmann et al., 2015]] ; [[#Biggs--2018|Biggs et al., 2018]] ). Abrupt ecosystem shifts have been observed in both large open-ocean ecosystems and coastal ecosystems ( [[#3.4.2|Section 3.4.2]] ), with dramatic social consequences through significant loss of diverse ecosystem services ( ''high confidence'' ) ( [[#3.5|Section 3.5]] ; [[#Biggs--2018|Biggs et al., 2018]] ; [[#Pinsky--2018|Pinsky et al., 2018]] ; [[#Beaugrand--2019|Beaugrand et al., 2019]] ; [[#Collins--2019a|Collins et al., 2019a]] ; [[#Filbee-Dexter--2020b|Filbee-Dexter et al., 2020b]] ; [[#Huntington--2020|Huntington et al., 2020]] ; [[#Trisos--2020|Trisos et al., 2020]] ; [[#Turner--2020b|Turner et al., 2020b]] ; [[#Canadell--2021|Canadell and Jackson, 2021]] ; [[#Ma--2021|Ma et al., 2021]] ; [[#Ruthrof--2021|Ruthrof et al., 2021]] ). A summary of previous assessments of abrupt ecosystem shifts and extreme events is provided in Table 3.21.

'''Table 3.21 |''' Summary of previous IPCC assessments of observed and projected abrupt ecosystem shifts and extreme events

{| class="wikitable"
|-
! Observations

! Projections

|-
| ''AR5 ( [[#Wong--2014|Wong et al., 2014]] )''

|
|-
| Observations of abrupt ecosystem shifts and extreme events were not assessed in this report.

| ‘Warming and acidification will lead to coral bleaching, mortality, and decreased constructional ability ( ''high confidence'' ), making coral reefs the most vulnerable marine ecosystem with little scope for adaptation. Temperate seagrass and kelp ecosystems will decline with the increased frequency of heatwaves and sea temperature extremes as well as through the impact of invasive subtropical species ( ''high confidence'' ).’

|-
|
|-
| ''SROCC ( [[#Collins--2019a|Collins et al., 2019a]] )''

|
|-
| ‘Marine heatwaves (MHWs), periods of extremely high ocean temperatures, have negatively impacted marine organisms and ecosystems in all ocean basins over the last two decades, including critical foundation species such as corals, seagrasses and kelps ( ''very high confidence'' ).’

| ‘Marine heatwaves are projected to further increase in frequency, duration, spatial extent and intensity (maximum temperature) ( ''very high confidence'' ). Climate models project increases in the frequency of marine heatwaves by 2081–2100, relative to 1850–1900, by approximately 50 times under RCP8.5 and 20 times under RCP2.6 ( ''medium confidence'' ).’

‘Extreme El Niño and La Niña events are projected to ''likely'' increase in frequency in the 21st century and to ''likely'' intensify existing hazards, with drier or wetter responses in several regions across the globe. Extreme El Niño events are projected to occur about twice as often under both RCP2.6 and RCP8.5 in the 21st century when compared to the 20th century ( ''medium confidence'' ).’

‘Limiting global warming would reduce the risk of impacts of MHWs, but critical thresholds for some ecosystems (e.g., kelp forests, coral reefs) will be reached at relatively low levels of future global warming ( ''high confidence'' ).’

|}

<div id="_idContainer077" class="Figure"></div>

[[File:8606e2823dc86adf18dafb7ba38045b4 IPCC_AR6_WGII_Figure_3_019.png]]

'''Figure 3.19 |''' '''Observed ecological regime shifts and their drivers in the oceans.'''

'''(a)''' A conceptual representation of ecosystem resilience and regime shifts. Shift from Regime 1 to Regime 2 can be triggered by either a large shock (i.e., an abrupt environmental transition) or gradual internal or external change that erodes the dominant balancing feedbacks, reducing ecosystem resilience (indicated by the shallower dotted line, relative to the deeper ‘valley’ reflecting higher resilience). (Based on [[#Biggs--2018|Biggs et al., 2018]] ).

'''(b)''' The sum of the magnitude and extent of the abrupt community shifts that has been estimated at each geographic cell in the global ocean during 1960–2014, calculated as the ratio of the amplitude of the change in a particular year to the average magnitude of the change over the entire time series (thus, is dimensionless). (Based on [[#Beaugrand--2019|Beaugrand et al., 2019]] ).

Abrupt ecosystem shifts are associated with large-scale patterns of climate variability ( [[#Alheit--2019|Alheit et al., 2019]] ; [[#Beaugrand--2019|Beaugrand et al., 2019]] ; [[#Lehodey--2020|Lehodey et al., 2020]] ), some of which are projected to intensify with climate change ( ''medium confidence'' ) (WGI AR6 Chapter 1; [[#Wang--2017a|Wang et al., 2017a]] ; [[#Collins--2019a|Collins et al., 2019a]] ; [[#Chen--2021|Chen et al., 2021]] ). Over the past 60 years, abrupt ecosystem shifts have generally followed El Niño/Southern Oscillation events of any strength, but some periods had geographically limited ecological shifts (~0.25% of the global ocean in 1984–1987) and others more extensive shifts (14% of the global ocean in 2012–2015) ( ''medium confidence'' ) (Figure 3.19b; [[#Beaugrand--2019|Beaugrand et al., 2019]] ). Typically, interacting drivers, such as eutrophication and overharvest, reduce ecosystem resilience to climate extremes (e.g., MHWs, cyclones) or gradual warming, and hence promote ecosystem shifts ( ''high confidence'' ) (Figure 3.19a; [[#Rocha--2015|Rocha et al., 2015]] ; [[#Biggs--2018|Biggs et al., 2018]] ; [[#Babcock--2019|Babcock et al., 2019]] ; [[#Turner--2020b|Turner et al., 2020b]] ; [[#Bergstrom--2021|Bergstrom et al., 2021]] ; [[#Canadell--2021|Canadell and Jackson, 2021]] ; [[#Tait--2021|Tait et al., 2021]] ). Also, shifts in different ecosystems may be connected through common drivers or through cascading effects ( ''medium confidence'' ) ( [[#Rocha--2018a|Rocha et al., 2018a]] ).

Recent MHWs ( [[#3.2.2.1|Section 3.2.2.1]] ) have caused major ecosystem shifts and mass mortality in oceanic and coastal ecosystems, including corals, kelp forests and seagrass meadows (Sections 3.4.2.1, 3.4.2.3, 3.4.2.5, 3.4.2.6, 3.4.2.10; Cross-Chapter Box MOVING SPECIES in Chapter 5; Cross-Chapter Box EXTREMES in Chapter 2), with dramatic declines in species foundational for habitat formation or trophic flow, biodiversity declines, and biogeographic shifts in fish stocks ( ''very high confidence'' ) (Table 3.15; Cross-Chapter Box MOVING SPECIES in Chapter 5; [[#Canadell--2021|Canadell and Jackson, 2021]] ). Three major bleaching episodes on Australia’s Great Barrier Reef in 5 years corresponded with extreme temperatures in 2016, 2017 and 2020 ( [[#Pratchett--2021|Pratchett et al., 2021]] ). Between 1981 and 2017, MHWs have increased more than 20-fold due to anthropogenic climate change ( [[#3.2.2.1|Section 3.2.2.1]] ; WGI AR6 Chapter 9; [[#Laufkötter--2020|Laufkötter et al., 2020]] ; [[#Fox-Kemper--2021|Fox-Kemper et al., 2021]] ), increasing the risk of abrupt ecosystem shifts ( ''high confidence'' ) (Figure 3.19a; Cross-Chapter Box EXTREMES in Chapter 2; [[#van%20der%20Bolt--2018|van der Bolt et al., 2018]] ; [[#Garrabou--2021|Garrabou et al., 2021]] ; [[#Wernberg--2021|Wernberg, 2021]] ).

Ecosystems can recover from abrupt shifts (e.g., [[#Babcock--2019|Babcock et al., 2019]] ; [[#Christie--2019|Christie et al., 2019]] ; [[#Medrano--2020|Medrano et al., 2020]] ). However, where climate change is a dominant driver, ecosystem collapses increasingly cause permanent transitions ( ''high confidence'' ), although the extents of such transitions depend on the emission scenario ( [[#Trisos--2020|Trisos et al., 2020]] ; [[#Garrabou--2021|Garrabou et al., 2021]] ; [[#Klein--2021|Klein et al., 2021]] ; [[#Pratchett--2021|Pratchett et al., 2021]] ; [[#Wernberg--2021|Wernberg, 2021]] ). Over the coming decades, MHWs are projected to ''very likely'' become more frequent under all emission scenarios ( [[#3.2|Section 3.2]] ; WGI AR6 Chapter 9; [[#Fox-Kemper--2021|Fox-Kemper et al., 2021]] ), with intensities and rates too high for recovery of degraded foundational species, habitats or biodiversity ( ''medium confidence'' ) ( [[#Babcock--2019|Babcock et al., 2019]] ; [[#Garrabou--2021|Garrabou et al., 2021]] ; [[#Klein--2021|Klein et al., 2021]] ; [[#Serrano--2021|Serrano et al., 2021]] ; [[#Wernberg--2021|Wernberg, 2021]] ). Emission pathways that result in temperature overshoot above 1.5 o C will increase the risks of abrupt and irreversible shifts in coral reefs and other vulnerable ecosystems ( [[#3.4.4|Section 3.4.4]] ).

<div id="3.4.3.3.4" class="h4-container"></div>

<span id="time-of-emergence-species-exposure-to-altered-environments"></span>
===== 3.4.3.3.4 Time of emergence: species exposure to altered environments =====

<div id="h4-12-siblings" class="h4-siblings"></div>

Since SROCC, more studies have assessed the time of emergence for climate-induced drivers ( [[#3.2.3|Section 3.2.3]] ) and the ecosystem attributes through which the impacts manifest. However, as in previous assessments (Table 3.22), the time of emergence for a given driver or ecosystem attribute depends on the reference period, the definition of the signal emergence threshold and the spatio-temporal scales considered (see Box 5.1 in SROCC; [[#Kirtman--2013|Kirtman et al., 2013]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ).

'''Table 3.22 |''' Conclusions from previous IPCC assessments about projected time of emergence on coastal, ocean and deep-sea systems

{| class="wikitable"
|-
! Oceanic systems and chapter subsection

! Projections

|-
| Coastal ( [[#3.4.2|Section 3.4.2]] )

| ‘Multiple [climate-impact drivers] will emerge [...] in the 21st century under RCP8.5, while the time of emergence will be later and with less [climatic hazards] under RCP2.6. [Non-climate] impacts such as eutrophication add to, and in some cases, exacerbate these large-scale slow climate drivers beyond biological thresholds at local scale (e.g., deoxygenation)’ ( [[IPCC:Wg2:Chapter:Chapter-5#5.3|Section 5.3.7]] in SROCC; [[#Bindoff--2019a|Bindoff et al., 2019a]] ).

|-
| Epipelagic ( [[#3.4.3|Section 3.4.3]] )

| ‘Observed range shifts in response to climate change in some regions such as the north Atlantic are strongly influenced by warming due to both multi-decadal [climate change and] variability, suggesting that there is a longer time of emergence of range shifts from natural variability and a need for longer biological time series for robust attribution’ ( [[IPCC:Wg2:Chapter:Chapter-5#5.2|Section 5.2.3.1.1]] in SROCC; [[#Bindoff--2019a|Bindoff et al., 2019a]] ).

|-
| Open ocean ( [[#3.4.3|Section 3.4.3]] )

| ‘[The timing] for five primary drivers of marine ecosystem change (surface warming and acidification, oxygen loss, nitrate concentration and net primary production change) are all prior to 2100 for over 60% of the ocean area under RCP8.5 and over 30% under RCP2.6 ( ''very likely'' )’ (Figure 1 in Box 5.1 in SROCC, Box 5.1 in SROCC, Executive Summary in SROCC Chapter 5; [[#Bindoff--2019a|Bindoff et al., 2019a]] ).

‘Anthropogenic signals are expected to remain detectable over large parts of the ocean, even for the RCP2.6 scenario for pH and SST, but are ''likely'' [to be less conspicuous] for nutrients and NPP [net primary production] in the 21st century. For example, for the open ocean, the anthropogenic pH signal in Earth System Models’ (ESM) historical simulations is ''very likely'' to have emerged for three-quarters of the ocean prior to 1950, and it is ''very likely'' over 95% of the ocean has already been affected, with little discernible difference between scenarios’ (Executive Summary in SROCC Chapter 5, Box 5.1 in SROCC; [[#Bindoff--2019a|Bindoff et al., 2019a]] ). ‘The climate signal of oxygen loss will ''very likely'' emerge by 2050 with a ''very likely'' range of 59–80% by 2031–2050 and rising with a ''very likely'' range of 79–91% of the ocean area by 2081–2100 (RCP8.5 emissions scenario). The emergence of oxygen loss is smaller in area for the RCP2.6 scenario in the 21st century and by 2090 the [area where emergence is evident is declining].’ It has also been shown that signatures of altered oxygen solubility or utilisation may emerge earlier than for oxygen levels (Executive Summary in SROCC Chapter 5, Box 5.1 in SROCC; [[#Bindoff--2019a|Bindoff et al., 2019a]] ).

|-
| Deep sea (Box 3.3)

| ‘Emergence of risk is expected to occur later at around the mid-21st century under RCP8.5 for abyssal plain and chemosynthetic ecosystems (vents and seeps)’ ( [[IPCC:Wg2:Chapter:Chapter-5#5.2|Section 5.2.5]] in SROCC; [[#Bindoff--2019a|Bindoff et al., 2019a]] ). ‘All deep seafloor ecosystems are expected to be subject to at least moderate risk under RCP8.5 by the end of the 21st century, with cold water corals undergoing a transition from moderate to high risk below 3°C’ (SM5.2 in SROCC; [[#Bindoff--2019b|Bindoff et al., 2019b]] ).

|}

Anthropogenically driven changes in chlorophyll- ''a'' concentrations across an ensemble of 30 ESMs are expected to exceed natural variability under RCP8.5 by 2100 in ~65–80% of the global oceans, when the natural variability is calculated using the ensemble’s standard deviation ( [[#Schlunegger--2020|Schlunegger et al., 2020]] ); however, if two standard deviations are used, then significant trends in chlorophyll- ''a'' concentration are expected under RCP8.5 across ~31% of the global oceans by 2100 ( [[#Dutkiewicz--2019|Dutkiewicz et al., 2019]] ). In contrast, the anthropogenic signal in phytoplankton community structure, which has a lower natural variability, will emerge under RCP8.5 across 63% of the ocean by 2100 when two standard deviations are used ( ''limited evidence'' ) ( [[#Dutkiewicz--2019|Dutkiewicz et al., 2019]] ).

The time of emergence of climate impacts on ecosystems will be modulated jointly by species-specific adaptation potential ( [[#3.3.4|Section 3.3.4]] ; [[#Jones--2018|Jones and Cheung, 2018]] ; [[#Collins--2020|Collins et al., 2020]] ; [[#Gamliel--2020|Gamliel et al., 2020]] ; [[#Miller--2020a|Miller et al., 2020a]] ), speed of range shifts and spatial reorganisation ( ''high confidence'' ) (Sections 3.3, 3.4.2, 3.4.3). These ecosystem responses complicate projections of the time of emergence of environmental properties that impact biogeochemical cycling ( [[#Schlunegger--2019|Schlunegger et al., 2019]] ; [[#Schlunegger--2020|Schlunegger et al., 2020]] ; [[#Wrightson--2020|Wrightson and Tagliabue, 2020]] ), ecosystem structure and biodiversity (Figure 3.20a,c; [[#Dutkiewicz--2019|Dutkiewicz et al., 2019]] ; [[#Trisos--2020|Trisos et al., 2020]] ), and higher trophic levels, including fisheries targets ( [[#Cheung--2020|Cheung and Frölicher, 2020]] ). Better accounting for multiple interacting factors in ESMs (see Box 3.1) will provide insight into how marine ecosystems will respond to future climate ( ''high confidence'' ).

The time of emergence of ecosystem responses supports planning for specific time-bound actions to reduce risks to ecosystems ( [[#3.6.3.2|Section 3.6.3.2.1]] ; [[#Bruno--2018|Bruno et al., 2018]] ; [[#Trisos--2020|Trisos et al., 2020]] ). Although under RCP8.5, climate refugia from SST after 2050 are primarily in the Southern Ocean in tropical waters, these refugia are mainly from deoxygenation ( [[#Bruno--2018|Bruno et al., 2018]] ). Marine assemblages in these places will be exposed to unprecedented temperatures after 2050, peaking in 2075 (Figure 3.20a,b; [[#Trisos--2020|Trisos et al., 2020]] ). In contrast, changes in phytoplankton community structure will emerge earlier, primarily in the Pacific Ocean subtropics and through much of the North Atlantic Ocean (Figure 3.20c,d; [[#Dutkiewicz--2019|Dutkiewicz et al., 2019]] ). Under RCP8.5, changes in phytoplankton community structure and, to a lesser extent, exposure of marine species to unprecedented temperatures, will emerge earlier in marine protected areas (MPAs), covering ~7.7% of the global oceans ( [[#3.6.2.3|Section 3.6.2.3.2.1]] ; UNEP-WCMC and [[#IUCN--2020|IUCN, 2020]] ; [[#UNEP-WCMC%20and%20IUCN--2021|UNEP-WCMC and IUCN, 2021]] ) as compared with non-MPAs (Figure 3.20b,d). Such assessment can support planning for future MPA placement and extent. Because MPAs can serve as refugia from non-climate drivers (Sections 3.6.2.3, 3.6.3.2.1), they facilitate opportunities for adaptation among marine species and communities in coastal oceans ( [[#3.4.2|Section 3.4.2]] ).

<div id="_idContainer083" class="Figure"></div>

[[File:8f6c3fd05dd924e5d2f20a2442625694 IPCC_AR6_WGII_Figure_3_020.png]]

'''Figure 3.20 |''' '''Time of exposure to altered environments.'''

'''(a)''' Simulated spatial variation in the time of exposure of marine species to unprecedented temperatures under RCP8.5. Time of exposure is quantified as the median year after which local species are projected to encounter temperatures warmer than the historical maximum within their full geographic range for a period of at least 5 years. This estimate is based on 22 Coupled Model Intercomparison Project 5 (CMIP5) models, and is drawn from data presented by Trisos et al. (2020). Only regions that have times of emergence by 2100 are shown.

'''(b)''' The distribution in the time of exposure to unprecedented temperatures within marine assemblages ( [[#Trisos--2020|Trisos et al., 2020]] ) under RCP8.5 in marine protected areas (in turquoise) and in non-marine protected areas (in purple). Values were calculated after regridding to equal-area 0.5° hexagons.

'''(c)''' Time of emergence for phytoplankton community-structure changes (based on a proxy–ecosystem-model reflectance at 500 nm) under RCP8.5. Only regions with statistically significant ( ''p'' &lt; 0.05) trends that are presently largely ice free and have times of emergence by 2100 are shown. (Based on the results of one numerical model from [[#Dutkiewicz--2019|Dutkiewicz et al., 2019]] ).

'''(d)''' The distribution in the time of emergence for changes in phytoplankton community structure (same proxy as in panel c) ( [[#Dutkiewicz--2019|Dutkiewicz et al., 2019]] ) under RCP8.5 in marine protected areas (in turquoise) and in non-marine protected areas (in purple). Values were calculated after regridding to equal-area 0.5° hexagons.

<div id="3.4.3.4" class="h3-container"></div>

<span id="biomass"></span>
==== 3.4.3.4 Biomass ====

<div id="h3-26-siblings" class="h3-siblings"></div>

<div id="3.4.3.4.1" class="h4-container"></div>

<span id="observed-changes-1"></span>
===== 3.4.3.4.1 Observed changes =====

<div id="h4-13-siblings" class="h4-siblings"></div>

Observed changes in biomass in the global ocean, beyond those for phytoplankton (Table 3.23), have not routinely been attributed to climate-induced drivers, but rather to the compound effects of multiple drivers, especially fishing ( [[#Christensen--2014|Christensen et al., 2014]] ; [[#Palomares--2020|Palomares et al., 2020]] ). We therefore do not assess observed changes in ocean biomass here.

'''Table 3.23 |''' Summary of previous IPCC assessments of changes in open ocean and deep-sea biomass

{| class="wikitable"
|-
! Measure

! Observations

! Projections

|-
| ''AR5 WGII ( [[#Hoegh-Guldberg--2014|Hoegh-Guldberg et al., 2014]] ; [[#Pörtner--2014|Pörtner et al., 2014]] )''

|
|-
| Chlorophyll- ''a'' /phytoplankton biomass

| ‘Phytoplankton biomass: the approximately 15-year archived time series of satellite-chlorophyll (as a proxy of phytoplankton biomass) is too short to reveal trends over time and their causes’ (WGII AR5 [[IPCC:Wg2:Chapter:Chapter-6#6.1.2|Section 6.1.2]] ; [[#Pörtner--2014|Pörtner et al., 2014]] ).

‘Chlorophyll concentrations measured by satellites have decreased in the subtropical gyres of the North Pacific, Indian and North Atlantic oceans by 9, 12 and 11%, respectively, over and above the inherent seasonal and interannual variability from 1998 to 2010 ( ''high confidence'' ; ''p'' ≤ 0.05). Significant warming over this period has resulted in increased water-column stratification, reduced mixed-layer depth and possibly decreases in nutrient availability and ecosystem productivity ( ''limited evidence, medium agreement'' ). The short time frame of these studies against well-established patterns of long-term variability leads to the conclusion that these changes are about as ''likely'' as not due to climate change’ (WGII AR5 Chapter 30; [[#Hoegh-Guldberg--2014|Hoegh-Guldberg et al., 2014]] ).

| ‘Owing to contradictory observations there is currently uncertainty about the future trends of major upwelling systems and how their drivers (enhanced productivity, acidification and hypoxia) will shape ecosystem characteristics ( ''low confidence'' )’ (WGII AR5 [[IPCC:Wg2:Chapter:Chapter-6|Chapter 6]] Executive Summary; [[#Pörtner--2014|Pörtner et al., 2014]] ).

|-
| Animal biomass

| Observed changes in animal biomass were not assessed in this report.

| ‘The climate-change-induced intensification of ocean upwelling in some eastern boundary systems, as observed in the last decades, may lead to regional cooling, rather than warming, of surface waters and cause enhanced productivity ( ''medium confidence'' ), but also enhanced hypoxia, acidification and associated biomass reduction in fish and invertebrate stocks’ (WGII AR5 [[IPCC:Wg2:Chapter:Chapter-6|Chapter 6]] Executive Summary; [[#Pörtner--2014|Pörtner et al., 2014]] ).

|-
| ''SROCC ( [[#Bindoff--2019a|Bindoff et al., 2019a]] )''

|
|-
| Chlorophyll- ''a'' /phytoplankton biomass

| ‘[Changes reported] in overall open-ocean chlorophyll levels (a proxy of phytoplankton biomass) of less than ±1% yr –1 for individual time periods. Regionally, trends of ±4% between 2002 and 2015 for different regions are found when different satellite products are merged, with increases at high latitudes and moderate decreases at low latitudes’ (SROCC [[IPCC:Wg2:Chapter:Chapter-5#5.2.2|Section 5.2.2.6]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ).

| Projected changes in chlorophyll- ''a'' /phytoplankton biomass were not assessed in this report.

|-
| Animal biomass

| Observed changes in open-ocean and deep-sea biomass were not assessed in this report.

| ‘There is ''high agreement'' in model projections that global zooplankton biomass will ''very likely'' reduce in the 21st century, with projected decline under RCP8.5 almost doubled that of RCP2.6 ( ''very likely'' ). However, the strong dependence of the projected declines on phytoplankton production ( ''low confidence'' ) and simplification in representation of the zooplankton communities and food web render their projections having ''low confidence'' .’

The global biomass of marine animals, including those that contribute to fisheries, is projected to decrease by 4.3 ± 2.0% (95% confidence interval) and 15.0 ± 5.9% under RCP2.6 and RCP8.5, respectively, by 2080–2099 relative to 1986–2005, while the decrease is around 4.9% by 2031–2050 across all RCP2.6 and RCP8.5 ( ''very likely'' ). Regionally, total animal biomass decreases largely in tropical and mid-latitude oceans ( ''very likely'' ).

‘Projected decrease in upper-ocean export of organic carbon to the deep seafloor is expected to result in a loss of animal biomass on the deep seafloor by 5.2–17.6% by 2090–2100 compared to the present (2006–2015) under RCP8.5 with regional variations ( ''medium confidence'' ). Some increases are projected in the polar regions, due to enhanced stratification in the surface ocean, reduced primary production and shifts towards small phytoplankton ( ''medium confidence'' ). The projected impacts on biomass in the abyssal seafloor are larger under RCP8.5 than RCP4.5 ( ''very likely'' ).’

|-
| ''WGI AR6 [[IPCC:Wg2:Chapter:Chapter-2|Chapter 2]] ( [[#Gulev--2021|Gulev et al., 2021]] )''

|
|-
| Chlorophyll- ''a'' /phytoplankton biomass

| The multi-sensor time series of chlorophyll- ''a'' concentration has been updated to cover two decades (1998–2018).

‘Global trends in chlorophyll- ''a'' for the last two decades are insignificant over large areas of the global oceans, but some regions exhibit significant trends, with positive trends in parts of the Arctic and the Antarctic waters (&gt;3% yr –1 ), and both negative and positive trends (within ±3% yr –1 ) in parts of the tropics, subtropics and temperate waters.’

‘In the last two decades, the concentration of phytoplankton at the base of the marine food web, as indexed by chlorophyll concentration, has shown weak and variable trends in low and mid-latitudes and an increase in high latitudes ( ''medium confidence'' ).’

| Projected changes in open-ocean and deep-sea biomass were not assessed in this report.

|}

<div id="3.4.3.4.2" class="h4-container"></div>

<span id="projected-changes-1"></span>
===== 3.4.3.4.2 Projected changes =====

<div id="h4-14-siblings" class="h4-siblings"></div>

Based on an ensemble of CMIP5 ESMs, SROCC projected declines in global zooplankton biomass by 2100 dependent on emission scenario ( ''low confidence'' ) (Table 3.23). The new CMIP6 ESM ensemble projects a decline in global zooplankton biomass by −3.9 ± 8.2% ( ''very likely range'' ) and −9.0 ± 8.9% in the period 2081–2100 relative to 1995–2014 under SSP1-2.6 and SSP5-8.5, respectively (Figure 3.21d; [[#Kwiatkowski--2020|Kwiatkowski et al., 2020]] ), thus reinforcing the SROCC assessment albeit with greater inter-model uncertainties.

<div id="_idContainer086" class="Figure"></div>

[[File:17b6a979e794ad7f282345cd4bdd0579 IPCC_AR6_WGII_Figure_3_021.png]]

'''Figure 3.21 |''' '''Projected change in marine biomass.''' Simulated global biomass changes of (a,b,c) surface phytoplankton, (d,e,f) zooplankton, (g,h,i) animals and (j,k,l) seafloor benthos. In (a,d,g,j), the multi-model mean (solid lines) and ''very likely range'' (envelope) over 2000–2100 relative to 1995–2014, for SSP1-2.6 and SSP5-8.5. Spatial patterns of simulated change by 2090–2099 are calculated relative to 1995–2014 for (b,e,h,k) SSP1-2.6 and (c,f,i,l) SSP5-8.5. Confidence intervals can be affected by the number of models available for the Coupled Model Intercomparison Project 6 (CMIP6) scenarios and for different variables. Only one model was available for panel (j), so no confidence interval is calculated. For panels (a–f), the ensemble projections of global changes in phytoplankton and zooplankton biomasses updated based on Kwiatkowski et al. (2019) include, under SSP1-2.6 and SSP5-8.5, respectively, a total of nine and ten CMIP6 Earth system models (ESMs). For panels (b,c,e,f), unhatched areas represent regions where at least 80% of models agree on the sign of biomass anomaly. For panels (g,h,i), the ensemble projections of global changes in total animal biomass updated based on [[#Tittensor--2021|Tittensor et al. (2021)]] include six to nine published global fisheries and marine ecosystem models from the Fisheries and Marine Ecosystem Model Intercomparison Project ( [[#Tittensor--2018|Tittensor et al., 2018]] ; [[#Tittensor--2021|Tittensor et al., 2021]] ), forced with standardised outputs from two CMIP6 ESMs. For panels (j,k,l), globally integrated changes in total seafloor biomass have been updated based on [[#Yool--2017|Yool et al. (2017)]] with one benthic model ( [[#Kelly-Gerreyn--2014|Kelly-Gerreyn et al., 2014]] ) forced with the CMIP6 ESM UKESM-1.

Using an ensemble of global-scale marine ecosystem and fisheries models (Fish-MIP) ( [[#Tittensor--2018|Tittensor et al., 2018]] ) with the CMIP5 ESM ensemble, SROCC concludes that projected ocean warming and decreased phytoplankton production and biomass will reduce global marine animal biomass during the 21st century ( ''medium confidence'' ). The simulated declines (with ''very likely range'' ) are −3.5 ± 4.8% and −14.0 ± 14.6% under RCP2.6 and RCP8.5, respectively, by 2080–2099 relative to 1995–2014 (SROCC [[IPCC:Wg2:Chapter:Chapter-5#5.2|Section 5.2.3]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ; [[#Lotze--2019|Lotze et al., 2019]] ) [[#footnote-001|6]] . Updated Fish-MIP simulations with CMIP6 (Figure 3.21g,h,i) confirm the projected decline in total marine animal biomass in the 21st century ( [[#Tittensor--2021|Tittensor et al., 2021]] ). The simulated declines (with ''very likely range'' ) are −5.7 ± 4.1% and −15.5 ± 8.5% under SSP1-2.6 and SSP5-8.5, respectively, by 2080–2099 relative to 1995–2014 (Figure 3.21g), showing greater declines and lower inter-model uncertainties ( [[#Tittensor--2021|Tittensor et al., 2021]] ). These declines result from combined warming and decreased primary production (with ''low confidence'' in future changes in primary production; [[#3.4.3.5|Section 3.4.3.5]] ) and are amplified at each trophic level within all ESM and marine ecosystem model projections across all scenarios ( ''medium confidence'' ) ( [[#Kwiatkowski--2019|Kwiatkowski et al., 2019]] ; [[#Lotze--2019|Lotze et al., 2019]] ; [[#Tittensor--2021|Tittensor et al., 2021]] ). However, there is ''limited evidence'' about how underlying food-web mechanisms amplify the climate signal from primary producers to higher trophic levels, and several putative mechanisms have been proposed ( [[#3.4.4|Section 3.4.4.2.2]] ; [[#Chust--2014a|Chust et al., 2014a]] ; [[#Stock--2014|Stock et al., 2014]] ; [[#Kwiatkowski--2019|Kwiatkowski et al., 2019]] ; [[#Lotze--2019|Lotze et al., 2019]] ; [[#Heneghan--2021|Heneghan et al., 2021]] ). As assessed in SROCC, the biomass projections contain considerable regional variation with declines in tropical to temperate regions and strong increases in total animal biomass are projected in polar regions under high-emission scenarios, with climate-change effects that are spatially similar but less pronounced under lower-emission scenarios (Figure 3.21b,c,e,f,h,i; [[#Tai--2019|Tai et al., 2019]] ; [[#Tittensor--2021|Tittensor et al., 2021]] ).

SROCC assessed that reduced food supply to the deep sea will drive a reduction in abyssal seafloor biota by 2100 for RCP8.5 (Table 3.23). Simulations from one size-resolved benthic biomass model coupled to an ocean-biogeochemistry model forced with the CMIP5 ESM HadGEM2-ES ( [[#Yool--2017|Yool et al., 2017]] ) project a decline in the globally integrated total seafloor biomass of −1.1 and −17.6% by 2100 under RCP2.6 and RCP8.5, respectively ( ''limited evidence, high agreement'' ). In waters shallower than 100 m, total benthic biomass is projected to increase by 3.2% on average by 2100 under RCP8.5, primarily driven by warming-increased growth rates ( [[#Yool--2013|Yool et al., 2013]] ), while at depths &gt;2000 m (representing 83% of the ocean seafloor), declines of −32% arise from climate-driven decreases in surface primary production and POC flux to the seafloor ( [[#Yool--2013|Yool et al., 2013]] ; [[#Kelly-Gerreyn--2014|Kelly-Gerreyn et al., 2014]] ; [[#Yool--2015|Yool et al., 2015]] ; [[#Yool--2017|Yool et al., 2017]] ). These patterns are qualitatively similar under RCP2.6, except in the Pacific and Indian Ocean basins, where some increased total seafloor biomass is projected ( [[#Yool--2013|Yool et al., 2013]] ). Updated simulations with the same benthic biomass model ( [[#Kelly-Gerreyn--2014|Kelly-Gerreyn et al., 2014]] ) forced with the CMIP6 ESM UKESM-1 project declines in total seafloor biomass of −9.8 and −13.0% by 2081–2100 relative to 1995–2014 for SSP1-2.6 and SSP5-8.5, respectively (Figure 3.21j,k,l). These projected changes in benthic biomass are based on ''limited evidence'' . Development of ensemble projections forced with a range of ESMs and a benthic model that considers the ecological roles of temperature ( [[#Hunt--2006|Hunt and Roy, 2006]] ; [[#Reuman--2014|Reuman et al., 2014]] ), oxygen ( [[#Mosch--2012|Mosch et al., 2012]] ) and ocean acidification ( [[#Andersson--2011|Andersson et al., 2011]] ) will provide opportunities to better quantify uncertainty in projected declines in total seafloor biomass under climate change.

Overall, ocean warming and decreased phytoplankton production and biomass will drive a global decline in biomass for zooplankton ( ''low confidence'' ), marine animals ( ''medium confidence'' ) and seafloor benthos ( ''low confidence'' ), with regional differences in the direction and magnitude of changes ( ''high confidence'' ). There is increasing evidence that responses will amplify throughout the food web and at ocean depths, with relatively modest changes in surface primary producers translating into substantial changes at higher trophic levels and for deep-water benthic communities ( ''medium confidence'' ).

<div id="3.4.3.5" class="h3-container"></div>

<span id="changes-in-primary-production-and-biological-carbon-export-flux"></span>
==== 3.4.3.5 Changes in Primary Production and Biological Carbon Export Flux ====

<div id="h3-27-siblings" class="h3-siblings"></div>

<div id="3.4.3.5.1" class="h4-container"></div>

<span id="observed-changes-in-primary-production"></span>
===== 3.4.3.5.1 Observed changes in primary production =====

<div id="h4-15-siblings" class="h4-siblings"></div>

Analyses of satellite-derived primary production over the past two decades (1998–2018) showed generally weak and negative trends (up to −3.0%) at low and mid latitudes ( [[#Kulk--2020|Kulk et al., 2020]] ). In contrast, positive trends occurred in large areas of the South Atlantic and South Pacific Oceans, as well as in polar and coastal (upwelling) regions (up to +4.5%; Cross-Chapter Paper 6; [[#Kulk--2020|Kulk et al., 2020]] ). Data-assimilating ocean biogeochemical models estimate a global decline in primary production of 2.1% per decade in the period 1998–2015, driven by the shoaling mixed layer and decreasing nitrate concentrations ( [[#Gregg--2019|Gregg and Rousseaux, 2019]] ). This is consistent with previous assessments that identified ocean warming and increased stratification as the main drivers ( ''high confidence'' ) affecting the regional variability in primary production Bindoff et al. (2019). However, as noted in SROCC and WGI AR6 [[IPCC:Wg2:Chapter:Chapter-2|Chapter 2]] (Table 3.24; [[#Gulev--2021|Gulev et al., 2021]] ), observed interannual changes in primary production on global and regional scales are nonlinear and largely influenced by natural temporal variability, providing ''low confidence'' in the trends.

'''Table 3.24 |''' Summary of previous IPCC assessments of ocean primary production and carbon export flux

{| class="wikitable"
|-
! Process

! Observed impacts

! Projected impacts

|-
| ''SROCC ( [[#Bindoff--2019a|Bindoff et al., 2019a]] )''

|
|-
| Open-ocean

primary production

| ‘Past open-ocean productivity trends, including those determined by satellites, [are appraised with ''low confidence'' ] due to newly identified region-specific drivers of microbial growth and the lack of corroborating ''in situ'' time series datasets.’

| ‘Net primary productivity (NPP) is ''very likely'' to decline by 4–11% by 2081–2100, relative to 2006–2015, across CMIP5 models for RCP8.5, but there is ''low confidence'' for this estimate due to the ''medium agreement'' among models and the ''limited evidence'' from observations. The tropical ocean NPP will ''very likely'' to decline by 7–16% for RCP8.5, with ''medium confidence'' as there are improved constraints from historical variability in this region.’

|-
| Open-ocean carbon export

| ‘Analyses of long-term trends in primary production and particle export production, as well as model simulations, reveal that increasing temperatures, leading to enhanced stratification and nutrient limitation, will have the greatest influence on decreasing the flux of particulate organic carbon (POC) to the deep ocean. However, different lines of evidence (including observation, modelling and experimental studies) provide ''low confidence'' on the mechanistic understanding of how climate drivers affect different components of the biological pump in the epipelagic ocean, as well as changes in the efficiency and magnitude of carbon export in the deep ocean.’

| ‘The projected changes in export production can be larger than global primary production because they are affected by both, the NPP changes, but also how shifts in food-web structure modulates the ‘transfer efficiency’ of particulate organic material, which then affects the sinking speed and lability of exported particles through the ocean interior to the sea floor.’

‘As export production is a much better understood net integral of changing net nutrient supply and can be constrained by interior ocean nutrient and oxygen levels, there is ''medium confidence'' in projections for global [export production] changes [based on CMIP5 model runs].’

|-
| ''WGI AR6 Chapters 2, 5 ( [[#Canadell--2021|Canadell et al., 2021]] ; [[#Gulev--2021|Gulev et al., 2021]] )''

|
|-
| Open-ocean

primary production

| Global ocean marine primary production is estimated to be 47 ± 7.8 PgC yr –1 with ''low confidence'' because of the small number of recent studies and the insufficient length of the time series analysed. A small decrease in productivity is evident globally for the period 1998–2015, but regional changes are larger and of opposing signs ( ''low confidence'' ) (WGI AR6 [[IPCC:Wg2:Chapter:Chapter-2#2.3|Section 2.3.4.2.2]] ; [[#Gulev--2021|Gulev et al., 2021]] ).

| ‘In CMIP5 models run under RCP8.5, [POC] export flux is projected to decline by 1–12% by 2100 ( [[#Taucher--2011|Taucher and Oschlies, 2011]] ; [[#Laufkötter--2015|Laufkötter et al., 2015]] ). Similar values are predicted in 18 CMIP6 models, with declines of 2.5–21.5% (median: −14%) [...] between 1900 and 2100 under the SSP5-8.5 scenario. The mechanisms driving these changes vary widely between models due to differences in parameterisation of particle formation, remineralisation and plankton community structure’ (WGI AR6 [[IPCC:Wg2:Chapter:Chapter-5#5.4.4.2|Section 5.4.4.2]] ; [[#Canadell--2021|Canadell et al., 2021]] ).

|}

<div id="3.4.3.5.2" class="h4-container"></div>

<span id="projected-changes-in-primary-production"></span>
===== 3.4.3.5.2 Projected changes in primary production =====

<div id="h4-16-siblings" class="h4-siblings"></div>

Across 10 CMIP5 and 13 CMIP6 ESM ensembles, global mean NPP is projected to decline by 2080–2099 relative to 2006–2015, under all RCPs and SSPs ( [[#Kwiatkowski--2020|Kwiatkowski et al., 2020]] ). However, under comparable radiative forcing, the CMIP6 multi-model mean projections of primary production declines (mean ± SD: −0.56 ± 4.12% under SSP1-2.6, and −3.00 ± 9.10% under SSP5-8.5) are less than those of previous CMIP5 models (3.42 ± 2.47% under RCP2.6, and 8.54 ± 5.88% under RCP8.5) (WGI AR6 [[IPCC:Wg2:Chapter:Chapter-5#5.4.4.2|Section 5.4.4.2]] ; [[#Kwiatkowski--2020|Kwiatkowski et al., 2020]] ; [[#Canadell--2021|Canadell et al., 2021]] ). The inter-model uncertainty associated with CMIP6 NPP projections is larger than in CMIP5, and it is consistently larger than the scenario uncertainty. For each SSP across the CMIP6 ensemble, individual models project both increases and decreases in global primary production, reflecting a diverse suite of bottom-up and top-down ecological processes, which are variously parameterised across models ( [[#Laufkötter--2015|Laufkötter et al., 2015]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ). Furthermore, accurate simulation of many of the biogeochemical tracers upon which NPP depends (e.g., the distribution of iron; [[#Tagliabue--2016|Tagliabue et al., 2016]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ) remains a significant and ongoing challenge to ESMs ( ''high confidence'' ) ( [[#Séférian--2020|Séférian et al., 2020]] ).

Regionally, multi-model mean changes in primary production show generally similar patterns of large declines in the North Atlantic and the western equatorial Pacific, while in the high latitudes, primary production consistently increases in CMIP5 and CMIP6 by 2100 (Cross-Chapter Paper 6; [[#Kwiatkowski--2020|Kwiatkowski et al., 2020]] ). In the Indian Ocean and subtropical North Pacific, which were regions of consistent NPP decline in CMIP5 projections ( [[#Bopp--2013|Bopp et al., 2013]] ), the regional declines are reduced in magnitude, less spatially extensive and are typically less robust in CMIP6. Further assessment of simultaneous changes in processes such as nutrient advection, nitrogen fixation, the microbial loop and top-down grazing pressure (WGI AR6 [[IPCC:Wg2:Chapter:Chapter-5#5.4.4.2|Section 5.4.4.2]] ; [[#Laufkötter--2015|Laufkötter et al., 2015]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ; [[#Canadell--2021|Canadell et al., 2021]] ) are required to fully understand the regional primary production response in CMIP6 ( [[#Kwiatkowski--2020|Kwiatkowski et al., 2020]] ). Given the regional variations in the estimates of primary production changes and the uncertainty in the representation of the dominant drivers, there remains ''low confidence'' in the projected global decline in NPP.

<div id="3.4.3.5.3" class="h4-container"></div>

<span id="observed-processes-driving-changes-in-global-export-flux"></span>
===== 3.4.3.5.3 Observed processes driving changes in global export flux =====

<div id="h4-17-siblings" class="h4-siblings"></div>

The SROCC ''medium confidence'' assessment that warming, stratification, declines in productivity and changes in plankton community in the epipelagic zone result in reduced export of primary production to deeper layers (Table 3.24) is supported by subsequent literature ( [[#Bach--2019|Bach et al., 2019]] ; [[#Leung--2021|Leung et al., 2021]] ). Particulate organic carbon export efficiency is constrained by altered mixing and nutrient availability ( [[#Boyd--2019|Boyd et al., 2019]] ; [[#Lundgreen--2019|Lundgreen et al., 2019]] ), particle fragmentation ( [[#Briggs--2020|Briggs et al., 2020]] ) as well as viral, microbial and planktonic community structure ( [[#Fu--2016|Fu et al., 2016]] ; [[#Guidi--2016|Guidi et al., 2016]] ; [[#Flombaum--2020|Flombaum et al., 2020]] ; [[#Kaneko--2021|Kaneko et al., 2021]] ) and metabolic rates (Cavan et al., 2019). These processes are strongly interlinked, and their net effect on primary production export from the upper ocean remains difficult to quantify observationally ( [[#Boyd--2019|Boyd et al., 2019]] ). Since SROCC, there is increasing evidence that ocean deoxygenation can alter zooplankton community structure ( [[#Wishner--2018|Wishner et al., 2018]] ), zooplankton respiration rates ( [[#Cass--2014|Cass and Daly, 2014]] ; [[#Cavan--2017|Cavan et al., 2017]] ) and patterns of diel vertical migration ( [[#Aumont--2018|Aumont et al., 2018]] ), which may focus remineralisation of organic carbon at the upper margins of OMZs ( [[#3.4.3.4|Section 3.4.3.4]] on depth shifts due to OMZ; [[#Bianchi--2013|Bianchi et al., 2013]] ; [[#Archibald--2019|Archibald et al., 2019]] ).

Data on export flux from the upper ocean are limited either in coverage and consistency (ship-board sampling) or duration (sediment traps), and are subject to considerable spatial variability (as shown in satellite observations ( [[#Boyd--2019|Boyd et al., 2019]] ). As a result, trends are weak, inconsistent and often not statistically significant ( [[#Lomas--2010|Lomas et al., 2010]] ; [[#Cael--2017|Cael et al., 2017]] ; [[#Muller-Karger--2019|Muller-Karger et al., 2019]] ; [[#Xie--2019|Xie et al., 2019]] ). Deep-ocean fluxes are similarly equivocal ( [[#Smith--2018|Smith et al., 2018]] ; [[#Fischer--2019|Fischer et al., 2019]] ; [[#Fischer--2020|Fischer et al., 2020]] ). In coming years, an increasing number of Argo floats equipped with bio-optical sensors should help improve estimates of particle flux spatio-temporal variability (e.g., [[#Dall’Olmo--2016|Dall’Olmo et al., 2016]] ).

Projected changes

SROCC and WGI AR6 reported global declines in POC export flux, from −8.9 to −15.8% by 2100 relative to 2000 under RCP8.5 in CMIP5 models, and −2.5 to −21.5% (median value: −14%) between 1900 and 2100 under SSP5-8.5 in CMIP6 models (Table 3.24; WGI AR6 5.4.4.2; [[#Bindoff--2019a|Bindoff et al., 2019a]] ; [[#Canadell--2021|Canadell et al., 2021]] ). In CMIP5 model runs, the decrease in the sinking flux of organic matter from the upper ocean into the ocean interior was strongly related to the changes in stratification that reduce net nutrient supply ( [[#Fu--2016|Fu et al., 2016]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ), especially in tropical regions, and the projections for global export production changes are reported with ''medium confidence'' . Increasing model complexity with more widespread representation of ocean biogeochemical processes between CMIP5 and CMIP6, and inclusion of more than one or two classes of phyto- and zooplankton, will provide opportunities to improve assessments of how climate-induced drivers affect different components of biological carbon pump in the epipelagic ocean, as well as changes in the efficiency and magnitude of carbon export in the deep ocean ( ''high confidence'' ) (see Box 3.3; [[#Le%20Quéré--2016|Le Quéré et al., 2016]] ; [[#Séférian--2020|Séférian et al., 2020]] ; [[#Wright--2021|Wright et al., 2021]] ).

<div id="box-3.2" class="h2-container box-container"></div>

'''Box 3.2 | Marine Birds and Mammals'''
<div id="h2-28-siblings" class="h2-siblings"></div>

Marine birds (seabirds and shorebirds) and mammals include charismatic species and species that are economically, culturally and ecologically important ( [[#Sydeman--2015|Sydeman et al., 2015]] ; [[#Albouy--2020|Albouy et al., 2020]] ; [[#Pimiento--2020|Pimiento et al., 2020]] ). Their long generation times and slow population growth suggests limited evolutionary resilience to rapid climate change ( [[#3.3.4|Section 3.3.4]] ; [[#Sydeman--2015|Sydeman et al., 2015]] ; [[#Miller--2018|Miller et al., 2018]] ). According to the Red List Species Assessments of the International Union for Conservation of Nature ( [[#IUCN--2020|IUCN, 2020]] ), the greatest current hazards to these groups include human use of biological resources and areas, invasive species and pollution (see Figure Box 3.2.1; [[#Dias--2019|Dias et al., 2019]] ; [[#Lusseau--2021|Lusseau et al., 2021]] ). Impacts of climate change and severe weather are ranked among the five most-important hazards influencing 131 and 45 bird and mammal species, respectively (see Figure Box 3.2.1 for selection of species), including 24 bird and 7 mammal species that are currently listed as endangered, critically endangered or threatened. Furthermore, according to these IUCN assessments, climate change and severe weather are expected to impact an additional 122 and 18 marine bird and mammal species over the next 50–100 years, respectively (see Figure Box 3.2.1; [[#Dias--2019|Dias et al., 2019]] ).

[[File:abd6e7cedbae3d299905be104b2f8363 IPCC_AR6_WGII_Figure_3_Box_3_2_1.png]]

'''Figure Box 3.2.1 |''' '''Hazard assessment for marine birds and mammals.''' Number of (a) marine birds and (b) mammals currently impacted by different hazards (blue), and numbers of additional species expected to be exposed to these threats over the next 50–100 years (red), as assessed in the International Union for Conservation of Nature Red List ( [[#IUCN--2020|IUCN, 2020]] ). Seabird species include species in the key orders ''Sphenisciformes'' , ''Pelecaniformes, Suliformes, Anseriformes, Procellariiformes'' and ''Charadriiformes'' categorised as inhabitants of marine ecosystems ( ''n'' = 483 species, assessed in the period 2016–2019). Marine mammal species include the species reviewed by [[#Lusseau--2021|Lusseau et al. (2021)]] ( ''n'' = 136 species, assessed in the period 2008–2019).

Marine birds and mammals are vulnerable to climate-induced loss of breeding and foraging habitats such as sea ice ( [[#3.4.2.1|Section 3.4.2.1]] 2), sandy beaches ( [[#3.4.2.6|Section 3.4.2.6]] ), salt marshes ( [[#3.4.2.5|Section 3.4.2.5]] ) and seagrass beds ( ''high confidence'' ) ( [[#3.4.2.5|Section 3.4.2.5]] ; [[#Sydeman--2015|Sydeman et al., 2015]] ; [[#Bindoff--2019a|Bindoff et al., 2019a]] ; [[#Ropert-Coudert--2019|Ropert-Coudert et al., 2019]] ; [[#Von%20Holle--2019|Von Holle et al., 2019]] ; [[#Albouy--2020|Albouy et al., 2020]] ; [[#Amano--2020|Amano et al., 2020]] ; [[#Bestley--2020|Bestley et al., 2020]] ; [[#Grose--2020|Grose et al., 2020]] ). With warming, shorebird population abundances decline in the tropics, ''likely'' due to heat stress and habitat loss, and increase at higher latitudes ( [[#Amano--2020|Amano et al., 2020]] ). Marine mammals dependent on sea ice habitat are particularly vulnerable to warming ( ''medium confidence'' ) ( [[#Albouy--2020|Albouy et al., 2020]] ; [[#Bestley--2020|Bestley et al., 2020]] ; [[#Lefort--2020|Lefort et al., 2020]] ), yet vulnerability can differ between populations. Ongoing sea ice loss is decreasing some polar bear populations while others remain stable, ''likely'' related to past harvesting history, regional differences in sea ice phenology and ecosystem productivity ( [[#Hamilton--2019|Hamilton and Derocher, 2019]] ; [[#Molnár--2020|Molnár et al., 2020]] ). Nevertheless, even under an intermediate emission scenario RCP4.5, increasing ice-free periods will ''likely'' reduce both recruitment and adult survival across most polar bear populations over the next four decades, threatening their existence ( ''medium confidence'' ) (see Figure Box 3.2.2; [[#Molnár--2020|Molnár et al., 2020]] ).

Climate change is affecting marine food-web dynamics ( ''high confidence'' ) (Sections 3.4.2, 3.4.3), and the vulnerability and adaptive capacity of marine birds and mammals to such changes is linked to the species’ breeding and feeding ecology. Higher-vulnerability species include central-place foragers (confined to, for example, breeding colonies fixed in space), diet and habitat specialists, and species with restricted distributions such as marine mammal populations in SES ( ''medium confidence'' ) ( [[#McMahon--2019|McMahon et al., 2019]] ; [[#Ropert-Coudert--2019|Ropert-Coudert et al., 2019]] ; [[#Albouy--2020|Albouy et al., 2020]] ; [[#Grose--2020|Grose et al., 2020]] ; [[#Sydeman--2021|Sydeman et al., 2021]] ). Surface-feeding and piscivorous marine birds appear to be more vulnerable to food-web changes than diving seabirds and planktivorous seabirds ( ''medium confidence'' ) ( [[#Sydeman--2021|Sydeman et al., 2021]] ). During the 2014–2015 Pacific heatwave, around 1 million piscivorous common murres died along a 1500 km coastal stretch in the Pacific USA due to reduced prey availability ( [[#Jones--2018b|Jones et al., 2018b]] ; [[#Piatt--2020|Piatt et al., 2020]] ). Marine birds are vulnerable to phenological shifts in food-web dynamics, as they have limited phenotypic plasticity of reproductive timing, with potentially little scope for evolutionary adaptation ( ''medium confidence'' ) ( [[#Keogan--2018|Keogan et al., 2018]] ), although changes in reproduction timing are observed in several species ( [[#3.4.4|Section 3.4.4.1]] ; [[#Sydeman--2015|Sydeman et al., 2015]] ; [[#Descamps--2019|Descamps et al., 2019]] ; [[#Sauve--2019|Sauve et al., 2019]] ). There is ''limited evidence'' of marine mammals’ capacity to adapt to shifting phenologies, but observed responses include changes in the onset of migrations, moulting and breeding ( [[#3.4.4|Section 3.4.4.1]] ; [[#Ramp--2015|Ramp et al., 2015]] ; [[#Hauser--2017|Hauser et al., 2017]] ; [[#Beltran--2019|Beltran et al., 2019]] ; [[#Bowen--2020|Bowen et al., 2020]] ; [[#Szesciorka--2020|Szesciorka et al., 2020]] ).

[[File:b63992502eb0157e86141d72a4d2831a IPCC_AR6_WGII_Figure_3_Box_3_2_2.png]]

'''Figure Box 3.2.2 |''' '''Modelled risk timelines for demographic impacts on circumpolar polar bear subpopulations, and associated confidence assessments, due to extended fasting periods with loss of sea ice.''' Years of first impact on cub recruitment (yellow), adult male survival (blue) and adult female survival (red) are shown for the (left) RCP4.5 and (right) RCP8.5. (Data from [[#Molnár--2020|Molnár et al., 2020]] ).

Increased emergence of infectious disease in mammals and birds is expected with ocean warming, due to new transmission pathways from changing species distributions, higher species densities caused by habitat loss and increased vulnerability due to environmental stress on individuals ( ''limited evidence'' ) ( [[#Sydeman--2015|Sydeman et al., 2015]] ; [[#VanWormer--2019|VanWormer et al., 2019]] ; [[#Sanderson--2020|Sanderson and Alexander, 2020]] ). Marine birds and mammals are ''likely'' to suffer from increased mortalities due to increasing frequencies of HABs, and of extreme weather, at sea, on sea ice, and in terrestrial breeding habitats ( [[#Broadwater--2018|Broadwater et al., 2018]] ; [[#Gibble--2018|Gibble and Hoover, 2018]] ; [[#Ropert-Coudert--2019|Ropert-Coudert et al., 2019]] ; [[#Grose--2020|Grose et al., 2020]] ). Also, climate-change driven distributional shifts have strengthened interactions with other anthropogenic impacts, through, for example, increasing risks of ship strikes and bycatch ( ''medium confidence'' ) (e.g., [[#Hauser--2018|Hauser et al., 2018]] ; [[#Krüger--2018|Krüger et al., 2018]] ; [[#Record--2019|Record et al., 2019]] ; [[#Santora--2020|Santora et al., 2020]] ).

<div id="FAQ 3.3" class="h2-container"></div>

<span id="faq-3.3-are-we-approaching-so-called-tipping-points-in-the-ocean-and-what-can-we-do-about-it"></span>