Editing IPCC:AR6/SROCC/Chapter-5 (section)

== Box 5.1 Time of Emergence and Exposure to Climate Hazards ==

<div id="section-5-2-2-6changing-ocean-primary-and-export-production-block-1"></div>

The concept of time of emergence (ToE) is defined as the time at which the ''signal'' of climate change in a given variable emerges from a measure of the background variability or ''noise'' (SROCC Glossary). In associating a calendar date with the detection, attribution and projection of climate trends, the concept of a ToE has proved useful for policy and planning particularly through informing important climatic thresholds and the uncertainties associated with past and future climate change (Hawkins and Sutton, 2012). However, there is not a single agreed metric and the ToE for a given variable thus depends on choices regarding the space and time scale, the threshold at which emergence is defined and the reference period (IPCC 5th Asseessment Report (AR5) Working Group I (WGI) Section 11.3.2.1). Recently, the ToE concept has been expanded to consider variables related to climatic hazards to marine organisms and ecosystems such as pH, carbonate ion concentrations, aragonite and calcite saturation states, nutrient levels and marine primary productivity (Box 5.1, Figure 1) (Ilyina et al., 2009; Friedrich et al., 2012; Keller et al., 2014b; Lovenduski et al., 2015; Rodgers et al., 2015). ToE assessments for the ocean typically quantify the internal variability using the standard deviation of the detrended data over a given time period (Keller et al., 2014b; Rodgers et al., 2015; Henson et al., 2016; Henson et al., 2017), the scenario and model uncertainty associated with different climate scenarios and across available ESMs (Frölicher et al., 2016), and in some cases the autocorrelation of noise (Weatherhead et al., 1998). As more components of ‘noise’ are accounted for, the ToE lengthens and the ToE is also affected by whether a control simulation or historical variability is used to determine the noise (Hameau et al., 2019).

This assessment considers the ToE of hazards exposed to by marine organisms and ecosystems. These biological components of the ocean respond to climate hazards that emerge locally, rather than to the global and basin-scale averages reported in WGI AR5 (Stocker et al., 2013). Overall, ESMs show that there is an ordered emergence of the climate variables, with pH emerging rapidly across the entire open ocean, followed by sea surface temperature (SST), interior oxygen, upper ocean nutrient levels and finally NPP under both Representative Concentration Pathway (RCP)2.6 and RCP8.5 relative to the 1861–1900 reference period (Box 5.1, Figure 1). Anthropogenic signals remain detectable for over large parts of the ocean even for the RCP2.6 scenario for pH and SST, but are ''likely'' lowered for nutrients and NPP 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 discernable difference between scenarios. The climate signal of oxygen loss will ''very likely'' emerge from the historical climate 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 under RCP2.6 scenario in the 21st century and by 2090 the emerged area is declining (Henson et al., 2017) (Box 5.1 Figure 1). It has also been shown that changes to oxygen solubility or utilisation may emerge earlier than bulk oxygen levels (Hameau et al., 2019).

It must be noted that variability will be greater in the coastal ocean than for the open ocean, which will be important for both hazard exposure for coastal species and the detection of trends. For example, although signals of anthropogenic influences have already emerged from internal variability in the late 20th century for global and basin-scale averaged ocean surface and sub-surface temperature ( ''very likely'' ) (AR5 WGI Summary for Policymakers), their ToE and level of confidence vary greatly at local scales and in coastal seas (Frölicher et al., 2016). Pelagic organisms with small range size may thus be more (or less) at risk to warming with earlier (or later) ToE at the scale of the area that they inhabit. From an observational standpoint, analyses that account for autocorrelation of noise suggest time series of around a decade are sufficient to detect a trend in pH or SST, whereas datasets spanning 30 years or longer are typically needed for detection of emergence at local scales for oxygen, nitrate and primary productivity (Henson et al., 2016).

<div id="section-5-2-2-6changing-ocean-primary-and-export-production-block-2"></div>

<span id="box-5.1-figure-1"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Box 5.1, Figure 1'''

<span id="box-5.1-figure-1-time-of-emergence-of-key-ocean-condition-variables-sea-surface-temperature-sst-surface-ph-100600-m-oxygen-o2-0100-m-nitrate-no3-and-0100-m-integrated-net-primary-production-npp.-the-year-of-emergence-represents-the-year-when-the-mean-change-relative-to-the-reference-period-of-18611900-is-above"></span>
<!-- IMG CAPTION -->
'''Box 5.1, Figure 1 | Time of emergence of key ocean condition variables: sea surface temperature (SST), surface pH, 100–600 m oxygen (O2), 0–100 m nitrate (NO3), and 0–100 m integrated net primary production (NPP). The year of emergence represents the year when the mean change relative to the reference period of 1861–1900 is above […]'''
<!-- IMG FILE -->
[[File:1c0f09dc88ff1a0d0d4b76b3c11351be IPCC-SROCC-CBox_5_1.jpg]]

Box 5.1, Figure 1 | Time of emergence of key ocean condition variables: sea surface temperature (SST), surface pH, 100–600 m oxygen (O2), 0–100 m nitrate (NO3), and 0–100 m integrated net primary production (NPP). The year of emergence represents the year when the mean change relative to the reference period of 1861–1900 is above the standard deviation of each variable over the historical period (Frölicher et al. 2016) and is expressed here in terms of the rate at which different climate signals emerge as a proportion of total ocean area for the Representative Concentration Pathway (RCP)8.5 scenario. The final area (and standard deviation) by 2100 under the RCP2.6 scenario is indicated by vertical lines at 2100.

<!-- END IMG -->
<div id="section-5-2-2-6changing-ocean-primary-and-export-production-block-3"></div>

The rapidity of change and its geographic scope, encompassed in the ToE, can be linked to concepts of exposure to hazard and vulnerability of biota. As organisms have evolved to be adaptable to natural variations in the environmental conditions of their habitats, changes to their habitat conditions larger than that typically experienced or specific biological thresholds such as upper temperature or oxygen tolerance may become hazardous (Mora et al., 2013). This would then move from the statistical nature of the ‘detection and attribution’ nature of the ToE discussed above towards timescales of impacts on organisms useful for ecosystem projections. In doing so, it will be important to think about the differences in habitat suitability between different organisms, including their specific thresholds for specific drivers, for example, temperature, oxygen or calcium carbonate stability. Further, thresholds vary depending on habitat, for example, warming thresholds for coral bleaching (Pendleton et al., 2016) may differ from the temperature and oxygen thresholds for fishes such as Atlantic cod and tunas (Deutsch et al., 2015). Moreover, species with fast generation times relative to the ToE of key habitat conditions (e.g., phytoplankton) may evolve more quickly to environmental change and be less vulnerable to climate change than longer-lived, slower generation time species (e.g., large sharks) (Jones and Cheung, 2018). However, evidence on evolutionary adaptation to expected climate change is limited, thus while shorter generation time may facilitate adaptation to environmental change, it does not necessarily result in successful adaptation of organisms (Section 5.2.3.1).

Earlier ToE and their subsequent biological impacts on organisms and ecosystems increase the urgency of policy responses through both climate mitigation and adaptation (Sections 5.5). However, the rapid emergence of hazards at the local scale in the near-term (already past or in this decade) such as warming and ocean acidification and the resulting impacts on some of the more sensitivity or less adaptive biodiversity and ecosystem services may post challenges for international and regional policies as their often require multiple decades to designate and implement (Box 5.6). In contrast, scope for adaptation for national and local ocean governance can be more responsive to rapid changes (Sections 5.5.2, 5.5.3). This highlights the opportunities for multi-level adaptation that allows for reducing climate risks that are expected to emergence of stressors and impacts at different time frame (Mackenzie et al., 2014).

<span id="impacts-on-pelagic-ecosystems"></span>
=== 5.2.3 Impacts on Pelagic Ecosystems ===

<div id="section-5-2-3impacts-on-pelagic-ecosystems-block-1"></div>

Marine pelagic ecosystems (the water column extending from the surface ocean down to the deep sea floor) face increasing climate related hazards from the changing environmental conditions (see Section 5.2.2). WGII AR5 (Pörtner et al., 2014) concluded, as also confirmed in Section 5.2.2, that long time series of more than three or four decades in length are necessary for determining biological trends in the ocean. However, long-term biological observations of pelagic ecosystems are rare and biased toward mid to high-latitude systems in the Northern Hemisphere (Edwards et al., 2013; Poloczanska et al., 2013; Poloczanska et al., 2016). This assessment, therefore, combines multiple lines of evidence ranging from experiments, field observations to model simulations to detect and attribute drivers of biological changes in the past, project future climate impacts and risks of pelagic ecosystems. In this section the pelagic ecosystem is subdivided into the surface, epipelagic ocean (&lt;200 m, the uppermost part of the ocean that receives enough sunlight to allow photosynthesis) (Section 5.2.3.1) and the deep pelagic ocean, comprising the twilight, mesopelagic zone (200–1000 m) and the dark, bathypelagic zone (&gt;1000 m deep) (Section 5.2.3.2). Although the WGII AR5 Chapter 30 defined the deep sea as below 1000 m (Hoegh-Guldberg et al., 2014), the absence of photosynthetically useful light and ensuing critical ecological, biogeochemical transformations, and altered human interactions that occur on much of the sea floor below 200 m have led both pelagic and benthic biologists to include the ocean waters and seafloor below 200 m within the definition of the deep sea (Herring and Dixon, 1998; Gage, 2003).

<div id="section-5-2-3-1the-epipelagic-ocean"></div>

<span id="the-epipelagic-ocean"></span>
==== 5.2.3.1 The Epipelagic Ocean ====

<div id="section-5-2-3-1the-epipelagic-ocean-block-1"></div>

This section synthesises new evidence since AR5 to assess observed changes in relation to the effects of and the interactions between multiple climate and non-climate hazards, and to project future risks of impacts from these hazards on the epipelagic organisms, communities and food web interactions, and their scope and limitation to adapt.

<div id="section-5-2-3-1the-epipelagic-ocean-block-2"></div>

<span id="detection-and-attribution-of-biological-changes-in-the-epipelagic-ocean-temperature-driven-shifts-in-distribution-and-phenology"></span>
===== 5.2.3.1.1 Detection and attribution of biological changes in the epipelagic ocean Temperature-driven shifts in distribution and phenology =====

WGII AR5 concluded that the vulnerability of most organisms to warming is set by their physiology, which defines their limited temperature ranges and thermal sensitivity (Pörtner et al., 2014 <sup>[[#fn:r385|385]]</sup> ). Although different hypotheses have been proposed since AR5 to explain the mechanism linking temperature sensitivity of marine organisms and their physiological tolerances (Schulte, 2015 <sup>[[#fn:r386|386]]</sup> ; Pörtner et al., 2017 <sup>[[#fn:r387|387]]</sup> ; Somero et al., 2017 <sup>[[#fn:r388|388]]</sup> ), evidence from physiological experiments and observations from paleo- and contemporary periods continue to support the conclusion from AR5 on the impacts of temperature change beyond thermal tolerance ranges on biological functions such as metabolism, growth and reproduction (Payne et al., 2016 <sup>[[#fn:r389|389]]</sup> ; Pörtner and Gutt, 2016 <sup>[[#fn:r390|390]]</sup> ; Gunderson et al., 2017 <sup>[[#fn:r391|391]]</sup> ), contributing to changes in biogeography and community structure (Beaugrand et al., 2015 <sup>[[#fn:r392|392]]</sup> ; Stuart-Smith et al., 2015 <sup>[[#fn:r393|393]]</sup> ) ( ''high agreement, high confidence'' ). Comparison of biota across land and ocean suggests that marine species are generally inhabiting environment that is closer to their upper temperature limits, explaining the substantially higher rate of local extirpation related to warming relative to those on land (Pinsky et al., 2019 <sup>[[#fn:r394|394]]</sup> ). Hypoxia and acidification can also limit the temperature ranges of organisms and exacerbate their sensitivity to warming (Mackenzie et al., 2014 <sup>[[#fn:r395|395]]</sup> ; Rosas-Navarro et al., 2016 <sup>[[#fn:r396|396]]</sup> ; Pörtner et al., 2017 <sup>[[#fn:r397|397]]</sup> ), although interactions vary strongly between species and biological processes (Gobler and Baumann, 2016 <sup>[[#fn:r398|398]]</sup> ; Lefevre, 2016 <sup>[[#fn:r399|399]]</sup> ).

Shifts in distribution of marine species from phytoplankton to marine mammals continued to be observed since AR5 across all ocean regions (Poloczanska et al., 2016 <sup>[[#fn:r400|400]]</sup> ). Recent evidence continues to support that a large proportion of records of observed range shifts in the epipelagic ecosystem (Poloczanska et al., 2016 <sup>[[#fn:r401|401]]</sup> ) are correlated with ocean temperature, with an estimated average shift in distribution (including range centroids, northward and southward boundaries) from these records of 51.5 ± 33.3 km per decade since the 1950s (Figure 5.13). Such rate of shift is significantly faster than those records for organisms in the seafloor; the latter has an average rate of distribution shift of 29.0 ± 15.5 km per decade (44% of the records for seafloor species with range shifts that are consistent with expectation from the observed temperature changes)( ''very likely'' ) (Figure 5.13). Comparison of global seafloor-derived planktonic foraminifera from pre-industrial age with recent (from year 1978) communities show that the recent assemblages differ from their pre-industrial with increasing dominance of warmer or cooler species that are mostly consistent with temperature changes (Jonkers et al., 2019 <sup>[[#fn:r402|402]]</sup> ). Rate of observed responses also varies between and within animal groups among ocean regions, with zooplankton and fishes having faster recorded range shifts (Pinsky et al., 2013 <sup>[[#fn:r403|403]]</sup> ; Asch, 2015 <sup>[[#fn:r404|404]]</sup> ; Jones and Cheung, 2015 <sup>[[#fn:r405|405]]</sup> ; Poloczanska et al., 2016 <sup>[[#fn:r406|406]]</sup> ). For example, analysis of the Continuous Plankton Recorder (CPR) data-series from the north Atlantic in the last decades shows that the range of dinoflagellates tended to closely track the velocity of climate change (the rate of isotherm movement). In contrast, the distribution range of diatoms shifted much more slowly (Chivers et al., 2017 <sup>[[#fn:r407|407]]</sup> ) and its distribution seems to be primary influenced by multi-decadal variability rather than from secular temperature trends. The CPR surveys have also provided evidence that some calanoid copepods are expanding poleward in the Northeast Atlantic, at a rate up to 232 km per decade (Beaugrand, 2009 <sup>[[#fn:r408|408]]</sup> ; Chivers et al., 2017 <sup>[[#fn:r409|409]]</sup> ), although different calanoid species respond differently in the rate and direction of shifts (Philippart et al., 2003 <sup>[[#fn:r410|410]]</sup> ; Edwards and Richardson, 2004 <sup>[[#fn:r411|411]]</sup> ; Asch, 2015 <sup>[[#fn:r412|412]]</sup> ; Crespo et al., 2017 <sup>[[#fn:r413|413]]</sup> ). Overall, the observed changes in biogeography are consistent with expected responses to changes in ocean temperature for the majority of marine biota ( ''high confidence'' ). This is also consistent with theories and experimental evidence that scale from individual organisms’ physiological responses to community level effects ( ''high confidence'' ). Sensitivity of organisms’ biogeography varies between taxonomic groups ( ''high confidence'' ).

<span id="figure-5.13"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 5.13'''

<span id="figure-5.13-evidence-of-climate-change-responses-of-marine-organisms-to-changes-in-ocean-conditions-under-climate-change.-a-evidence-of-interactive-effects-including-synergistic-and-antagonistic-of-multiple-climatic-hazards-based-on-przeslawski-et-al.-2015-lefevre-2016-section-5.2.2-5.2.3-5.2.4-5.3.-others-mainly-include-mammals-seabirds-and-marine-reptiles.-the-lighter-coloured"></span>
<!-- IMG CAPTION -->
'''Figure 5.13 | Evidence of climate change responses of marine organisms to changes in ocean conditions under climate change. (a) evidence of interactive effects (including synergistic and antagonistic) of multiple climatic hazards (based on Przeslawski et al. (2015); Lefevre (2016); Section 5.2.2, 5.2.3, 5.2.4, 5.3). ‘Others’ mainly include mammals, seabirds and marine reptiles). The lighter-coloured […]'''
<!-- IMG FILE -->
[[File:c3af71145efc3e74cad05dfb98d228fb IPCC-SROCC-CH_5_13-1.jpg]]

Figure 5.13 | Evidence of climate change responses of marine organisms to changes in ocean conditions under climate change. (a) evidence of interactive effects (including synergistic and antagonistic) of multiple climatic hazards (based on Przeslawski et al. (2015); Lefevre (2016); Section 5.2.2, 5.2.3, 5.2.4, 5.3). ‘Others’ mainly include mammals, seabirds and marine reptiles). The lighter-coloured cell represents insufficient information to draw conclusion; (b–d) observations on changes in latitudinal range and (e–h) phenology (based on Poloczanska et al. 2013). For b–h, each bar represents one record.

The rate and direction of observed range shifts are shaped by the interaction between climatic and non-climatic factors (Poloczanska et al., 2013 <sup>[[#fn:r414|414]]</sup> ; Sydeman et al., 2015 <sup>[[#fn:r415|415]]</sup> ; Poloczanska et al., 2016 <sup>[[#fn:r416|416]]</sup> ), such as local temperature and oxygen gradients in the habitat across depth (Cheung et al., 2013 <sup>[[#fn:r417|417]]</sup> ; Deutsch et al., 2015 <sup>[[#fn:r418|418]]</sup> ), latitude and longitude (Burrows et al., 2014 <sup>[[#fn:r419|419]]</sup> ; Barton et al., 2016 <sup>[[#fn:r420|420]]</sup> ), ocean currents (Sunday et al., 2015 <sup>[[#fn:r421|421]]</sup> ; Barton et al., 2016 <sup>[[#fn:r422|422]]</sup> ; García Molinos et al., 2017 <sup>[[#fn:r423|423]]</sup> ), bathymetry in all or part of their life stages (for organisms living on or close to the seafloor) (Pinsky et al., 2013 <sup>[[#fn:r424|424]]</sup> ; Kleisner et al., 2015 <sup>[[#fn:r425|425]]</sup> ), geographical barriers (Pinsky et al., 2013 <sup>[[#fn:r426|426]]</sup> ; Burrows et al., 2014 <sup>[[#fn:r427|427]]</sup> ), availability of food and critical habitat (Sydeman et al., 2015 <sup>[[#fn:r428|428]]</sup> ), fishing and other non-climatic human impacts (Engelhard et al., 2014 <sup>[[#fn:r429|429]]</sup> ; Hoegh-Guldberg et al., 2014 <sup>[[#fn:r430|430]]</sup> ). Moreover, observed range shifts in respond to climate change in some regions such as the north Atlantic are strongly influenced by warming due to multi-decadal variability (Edwards et al., 2013 <sup>[[#fn:r431|431]]</sup> ; Harris et al., 2014 <sup>[[#fn:r432|432]]</sup> ), 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. The rate of shifts in biogeography of organism is influenced by multiple climatic and non-climatic factors ( ''high confidence'' ) that can result in non-synchronous shifts in community composition ( ''high confidence'' ). There is general under-representation of biogeographical records in low latitudes (Dornelas et al., 2018 <sup>[[#fn:r433|433]]</sup> ), rendering detection and attribution of shifts in biogeography in these regions having ''medium confidence'' . The variation in responses of marine biota to range shifts can cause spatial restructuring of the pelagic ecosystem with consequences for organisms at higher trophic levels (Chivers et al., 2017 <sup>[[#fn:r434|434]]</sup> ; Pecl et al., 2017 <sup>[[#fn:r435|435]]</sup> ) ( ''high confidence'' ). Marine ectotherms have demonstrated some capacity for physiological adjustment and evolutionary adaptation that lowers their sensitivity to warming and decrease in oxygen (Pörtner et al., 2014 <sup>[[#fn:r436|436]]</sup> ; Cavallo et al., 2015 <sup>[[#fn:r437|437]]</sup> ) ( ''low confidence'' ). However, historical responses in abundance and ranges of marine species to ocean warming suggest that adaptation not always suffices to mitigate projected impacts (WGII AR5 Chapter 6) ( ''high confidence'' ).

Marine reptiles, seabirds and mammals breathe air, instead of obtaining oxygen from water, and many of them spend some of their life cycle on land, being their abundance and distribution still affected by temperature (Pörtner et al., 2014 <sup>[[#fn:r438|438]]</sup> ). Long term population changes and shifts in distribution associated with climate change have been observed for temperate species of seabirds and marine mammals (Henderson et al., 2014 <sup>[[#fn:r439|439]]</sup> ; Hiscock and Chilvers, 2014 <sup>[[#fn:r440|440]]</sup> ; Ramp et al., 2015 <sup>[[#fn:r441|441]]</sup> ) ( ''high confidence'' ). For example, Laysan, ''Phoebastria immutabilis,'' and Wandering, ''Diomedea exulans'' , albatross have responded positively to climate change as they have been able to take advantage of the increased intensity of winds. This has allowed them to forage farther and faster, making their foraging trips shorter, increasing their foraging efficiency and breeding success (Descamps et al., 2015 <sup>[[#fn:r442|442]]</sup> ; Thorne et al., 2016 <sup>[[#fn:r443|443]]</sup> ). For reptiles, like sea turtles and snakes, temperature directly affects important life history traits including hatchling size, sex, viability and performance ''(high confidence)'' (Hays et al., 2003 <sup>[[#fn:r444|444]]</sup> ; Pike, 2014 <sup>[[#fn:r445|445]]</sup> ; Dudley et al., 2016 <sup>[[#fn:r446|446]]</sup> ; Santos et al., 2017 <sup>[[#fn:r447|447]]</sup> ). This is particularly important for marine turtles as changing temperatures will affect the hatchling sex ratio because sex is determined by nest site temperature ''(high confidence'' ) (Hatfield et al., 2012 <sup>[[#fn:r448|448]]</sup> ; Santidrián Tomillo et al., 2014 <sup>[[#fn:r449|449]]</sup> ; Patricio et al., 2017 <sup>[[#fn:r450|450]]</sup> ). Loss of breeding substrate, including mostly coastal habitats such as sandy beaches (Section 5.3.3), can reduce the available nesting or pupping habitat for land breeding marine turtles, lizards, seabirds and pinnipeds (Fish et al., 2005 <sup>[[#fn:r451|451]]</sup> ; Fuentes et al., 2010 <sup>[[#fn:r452|452]]</sup> ; Funayama et al., 2013 <sup>[[#fn:r453|453]]</sup> ; Reece et al., 2013 <sup>[[#fn:r454|454]]</sup> ; Katselidis et al., 2014 <sup>[[#fn:r455|455]]</sup> ; Patino-Martinez et al., 2014 <sup>[[#fn:r456|456]]</sup> ; Pike et al., 2015 <sup>[[#fn:r|]]</sup> ; Reynolds et al., 2015; Marshall et al., 2017) ( ''high confidence'' ). Climatic hazards such as SLR contributes to the loss of these coastal habitats (see Section 5.3 and Chapter 3). Changes in ocean temperature will also indirectly impact marine mammals, seabirds and reptiles by changing the abundance and distribution of their prey (Polovina, 2005 <sup>[[#fn:r479|479]]</sup> ; Polovina et al., 2011 <sup>[[#fn:r480|480]]</sup> ; Doney et al., 2012 <sup>[[#fn:r481|481]]</sup> ; Sydeman et al., 2015 <sup>[[#fn:r482|482]]</sup> ; Briscoe et al., 2017 <sup>[[#fn:r483|483]]</sup> ; Woodworth-Jefcoats et al., 2017 <sup>[[#fn:r484|484]]</sup> ) ( ''high confidence'' ). The distributions of some of these large animals is determined by the occurrence and persistence of oceanic bridges and barriers that are related to climate driven processes (Ascani et al., 2016 <sup>[[#fn:r485|485]]</sup> ; McKeon et al., 2016 <sup>[[#fn:r486|486]]</sup> ). For example, the decline of Arctic sea ice is affecting the range and migration patterns of some species and is allowing the exchange of species previously restricted to either the Pacific or Atlantic oceans (Alter et al., 2015 <sup>[[#fn:r487|487]]</sup> ; George et al., 2015 <sup>[[#fn:r488|488]]</sup> ; Laidre et al., 2015 <sup>[[#fn:r489|489]]</sup> ; MacIntyre et al., 2015 <sup>[[#fn:r490|490]]</sup> ; McKeon et al., 2016 <sup>[[#fn:r491|491]]</sup> ; Breed et al., 2017 <sup>[[#fn:r492|492]]</sup> ; Hauser et al., 2017 <sup>[[#fn:r493|493]]</sup> ) (Chapter 3). Also, the range expansion of some of these predatory megafauna can affect species endemic to the habitat; for example, while the decrease in summer sea ice in the Arctic may favour the expansion of killer whales ( ''Orcinus orca'' ), their occurrence can result in narwhale ( ''Monodon monoceros'' ) to avoid the use of key habitats to reduce the risk of killer whales’ predation (Bost et al., 2009 <sup>[[#fn:r459|459]]</sup> ; Sydeman et al., 2015 <sup>[[#fn:r460|460]]</sup> ; Breed et al., 2017 <sup>[[#fn:r461|461]]</sup> ) (see Chapter 3; section 3.2.1.4). In addition, marine mammals, seabirds and sea turtles present habitat requirements associated with bathymetric and mesoscale features that facilitate the aggregation of their prey (Bost et al., 2015 <sup>[[#fn:r462|462]]</sup> ; Kavanaugh et al., 2015 <sup>[[#fn:r463|463]]</sup> ; Hindell et al., 2016 <sup>[[#fn:r464|464]]</sup> ; Hunt et al., 2016 <sup>[[#fn:r465|465]]</sup> ; Santora et al., 2017 <sup>[[#fn:r466|466]]</sup> ). The persistence and location of these features are linked to variations in climate (Crocker et al., 2006 <sup>[[#fn:r467|467]]</sup> ; Baez et al., 2011 <sup>[[#fn:r468|468]]</sup> ; Dugger et al., 2014 <sup>[[#fn:r469|469]]</sup> ; Abrahms et al., 2017 <sup>[[#fn:r470|470]]</sup> ; Youngflesh et al., 2017 <sup>[[#fn:r471|471]]</sup> ) and to foraging success, juvenile recruitment, breeding phenology, growth rates and population stability (Costa et al., 2010 <sup>[[#fn:r472|472]]</sup> ; Ancona and Drummond, 2013 <sup>[[#fn:r473|473]]</sup> ; Ducklow et al., 2013 <sup>[[#fn:r474|474]]</sup> ; Chambers et al., 2014 <sup>[[#fn:r475|475]]</sup> ; Descamps et al., 2015 <sup>[[#fn:r476|476]]</sup> ; Abadi et al., 2017 <sup>[[#fn:r477|477]]</sup> ; Bjorndal et al., 2017 <sup>[[#fn:r478|478]]</sup> ; Fluhr et al., 2017; Youngflesh et al., 2017) ( ''high confidence'' ). Overall, recent evidence further support that impacts of climate change on some marine reptiles, mammals and birds have been observed in recent decades ( ''high confidence'' ) and that the direction of impacts vary between species, population and geographic locations (Trivelpiece et al., 2011; Hazen et al., 2013; Clucas et al., 2014; Constable et al., 2014; George et al., 2015) ( ''high confidence'' ).

Warming has contributed also to observed changes in phenology (timing of repeated seasonal activities) of marine organisms (Gittings et al., 2018 <sup>[[#fn:r505|505]]</sup> ), although observations are biased towards the northeast Atlantic (Poloczanska et al., 2016 <sup>[[#fn:r506|506]]</sup> ; Thackeray et al., 2016). Shifts in the timing of interacting species have occurred in the last decades, eventually leading to uncoupling between prey and predators, with cascading community and ecosystem consequences (Kharouba et al., 2018; Neuheimer et al., 2018 <sup>[[#fn:r507|507]]</sup> ). 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 (Poloczanska et al., 2013 <sup>[[#fn:r508|508]]</sup> ), although it is variable among taxonomic groups and among ocean regions (Lindley and Kirby, 2010 <sup>[[#fn:r509|509]]</sup> ). This is consistent with the expectations based on the close relationship between temperature and these biological events, supporting evidence from AR5 (Bruge et al., 2016 <sup>[[#fn:r510|510]]</sup> ; Poloczanska et al., 2016 <sup>[[#fn:r511|511]]</sup> ). Thus, the growing amount of literature and new studies since AR5 WGII and SR15 further support   that phenology of marine ectotherms in the epipelagic systems are related to ocean warming ( ''high confidence'' ) and that the timing of biological events has shifted earlier ( ''high confidence'' ).

''Observed impacts of multiple climatic hazards''

WGII AR5 concludes that multiple climatic hazards from ocean acidification, hypoxia and decrease in nutrient and food supplies pose risks to marine ecosystems, and the risk can be elevated when combined with warming (Riebesell and Gattuso, 2014 <sup>[[#fn:r512|512]]</sup> ; Gattuso et al., 2015 <sup>[[#fn:r513|513]]</sup> ). In a recent meta-analysis of 632 published experiments, primary production by temperate non-calcifying plankton increases with elevated temperature and CO 2 , whereas tropical plankton decreases productivity because of acidification (Nagelkerken and Connell, 2015 <sup>[[#fn:r514|514]]</sup> ). Also, temperature increases consumption and metabolic rates of herbivores but not secondary production; the latter decreases with acidification in calcifying and non-calcifying species. These effects together create a mismatch with carnivores whose metabolic and foraging costs increase with temperature (Nagelkerken and Connell, 2015 <sup>[[#fn:r515|515]]</sup> ). Warming may also exacerbate the effects of ocean acidification on the rate of photosynthesis in phytoplankton (Lefevre, 2016 <sup>[[#fn:r516|516]]</sup> ). There is some, but limited, reports of observed impacts on calcified pelagic organisms that are attributed to secular trend in ocean acidification and warming (Harvey et al., 2013 <sup>[[#fn:r517|517]]</sup> ; Kroeker et al., 2013 <sup>[[#fn:r518|518]]</sup> ; Nagelkerken et al., 2015 <sup>[[#fn:r519|519]]</sup> ; Boyd et al., 2016 <sup>[[#fn:r520|520]]</sup> ). For example , Rivero-Calle et al. (2015) reported, using CPR archives, that stocks of coccolithophores (a group of phytoplankton that forms calcium carbonate plateles) have increased by 2% to over 20% in the north Atlantic over the last five decades, and that this increase is linked to synergistic effects of increasing anthropogenic CO 2 and rising temperatures, as supported by their statistical analysis and a number of experimental studies. Most of the available evidence supports that ocean acidification and hypoxia can act additively or synergistically between each other and with temperature across different groups of biota (Figure 5.13). Limitation of nutrient and food availability and predation pressures can further increase the sensitivity of organismal groups to climate change in specific ecosystems (Riebesell et al., 2017 <sup>[[#fn:r494|494]]</sup> ). Climate change also affects organisms indirectly through the impacts on competitiveness between organisms that favour those that are more adaptive to the changing environmental conditions (Alguero-Muniz et al., 2017 <sup>[[#fn:r495|495]]</sup> ) and changes in trophic interactions (Seebacher et al., 2014 <sup>[[#fn:r496|496]]</sup> ). Overall, direct ''in situ'' observations and laboratory experiments show that there are significant responses to the multiple stressors of warming, ocean acidification and low oxygen on phytoplankton, zooplankton and fishes and that these responses can be additive or synergistic ( ''high confidence,'' Figure 5.13).

<!-- END IMG -->
<div id="section-5-2-3-1the-epipelagic-ocean-block-3"></div>

<span id="future-changes-in-the-epipelagic-ocean"></span>
===== 5.2.3.1.2 Future changes in the epipelagic ocean =====

WGII AR5 and SR15 conclude that projected ocean warming will continue to cause poleward shifts in the distribution and biomass of pelagic species, paralleled by altered seasonal timing of their activities, species abundance, migration pattern and reduction in body size in the 21st century under scenarios of increasing greenhouse gas emission (Pörtner et al., 2014 <sup>[[#fn:r497|497]]</sup> ; Hoegh-Guldberg et al., 2018 <sup>[[#fn:r498|498]]</sup> ). Simultaneously, projected expansion of OMZ and ocean acidification could lead to shifts in community composition toward hypoxia-tolerant and non-calcified organisms, respectively. However, these projected biological changes in the ocean raise questions about how individuals, communities and food webs will respond to the multiple impacts from climatic and non-climatic stressors in the future, and the feedbacks of the effects of their ecological impacts on modifying the physical and biogeochemical conditions of the ocean (Schaum et al., 2013 <sup>[[#fn:r499|499]]</sup> ; Boyd et al., 2016 <sup>[[#fn:r500|500]]</sup> ; O’Brien et al., 2016 <sup>[[#fn:r501|501]]</sup> ; Moore, 2018 <sup>[[#fn:r502|502]]</sup> ). This section focuses on addressing these questions in order to assess the future risk of impacts of climate change on the epipelagic ecosystem.

''Future projections on phytoplankton distribution, community structure and biomass''

While analysis of outputs from CMIP5 ESMs project that global average NPP and biomass of phytoplankton community will decrease in the 21st century under RCP2.6 and RCP8.5 (see Section 5.2.2.6). However, the future risk of impacts of epipelagic ecosystem can also depend on changes in community structure of phytoplankton species. Barton et al. (2016) projected the biogeography of 87 taxa of phytoplankton (diatoms and dinoflagellates) in the north Atlantic to 2051–2100 relative to the past (1951–2000) with scenarios of changes in temperature and other ocean conditions such as salinity, density and nutrients under RCP8.5. The study found that 74% of the studied taxa exhibit a poleward shift at a median rate of 12.9 km per decade, but 90% of the taxa shift eastward at a median rate of 42.7 km per decade. Such changes may affect food webs and biogeochemical cycles, and with consequence to the productivity of living marine resources (Stock et al., 2014 <sup>[[#fn:r503|503]]</sup> ; Barton et al., 2016 <sup>[[#fn:r504|504]]</sup> ).

Outputs from CMIP5 ESMs suggest that projected warming and reduction in nutrient availability in low latitudes, as a result of increasing stratification of the ocean under climate change, will increase the dominance of small-sized phytoplankton, growing more efficiently than larger taxa at low nutrient levels (Dutkiewicz et al., 2013b <sup>[[#fn:r522|522]]</sup> ). Dominant groups in subtropical oceans, like the picoplanktonic cyanobacteria ''Synechococcus'' and ''Prochlorococcus'' , are projected to expand their range of distribution towards higher latitudes and increase their abundances by 14–29%, respectively, under a future warmer ocean (Flombaum et al., 2013 <sup>[[#fn:r523|523]]</sup> ), although synergistic effects of warming and CO 2 on photosynthetic rates could lead to a dominance of ''Synechococcus'' over ''Prochlorococcus'' (Fu et al., 2007 <sup>[[#fn:r524|524]]</sup> ) ( ''low confidence'' ). Similarly, temperature-driven range shifts towards higher latitudes are also likely for tropical diazotrophic (N 2 -fixing) cyanobacteria, although they could disappear from parts of their current tropical ranges where future warming may exceed their maximum thermal tolerance limits (Hutchins and Fu, 2017 <sup>[[#fn:r525|525]]</sup> ) ( ''low confidence'' ). Modelling experiments show that the effects of warming on phytoplankton community will be exacerbated by ocean acidification at levels expected in the 21st century for RCP8.5, leading to increasing growth rate responses of some phytoplankton groups, such as diazotrophs and ''Synechococcus'' , with predicted increases in biomass up to 10% in tropical and subtropical waters (Dutkiewicz et al., 2015 <sup>[[#fn:r526|526]]</sup> ) ( ''low confidence'' ). Furthermore, warming is projected to interact with decreasing oxygen levels and increases in iron in the nutrient-impoverished subtropical waters, favoring the dominance of the diazotrophic colonial cyanobacteria ''Trichodesmium'' (Sohm et al., 2011 <sup>[[#fn:r527|527]]</sup> ; Boyd et al., 2013 <sup>[[#fn:r528|528]]</sup> ; Ward et al., 2013 <sup>[[#fn:r529|529]]</sup> ; Hutchins and Fu, 2017 <sup>[[#fn:r530|530]]</sup> ) ( ''medium confidence'' ).

Regional differences in the changes in phytoplankton community and their impacts on epipelagic ecosystem are however complex and depends on multiple interactions of co-varying climate change stressors at regional level (Boyd and Hutchins, 2012 <sup>[[#fn:r531|531]]</sup> ). Based on global ocean model simulations, Boyd et al. (2015b) show that the interaction between warming, increased CO 2 and a decline in phosphate and silicate would benefit coccolithophores against diatoms in the northern north Atlantic, despite decreasing rates of calcification. Evidence, based on long-term experiments of acclimation or adaptation to increasing temperatures in combination with elevated CO 2 , show that individual growth and carbon fixation rates of coccolithophores at high CO 2 are modulated by temperature, light, nutrients and UV radiation, and could increase calcification while the responses are also species-specific (Lohbeck et al., 2012 <sup>[[#fn:r532|532]]</sup> ; Khanna et al., 2013 <sup>[[#fn:r533|533]]</sup> ). Calcification of planktonic foraminifera will be however negatively affected by acidification (Roy et al., 2015 <sup>[[#fn:r535|535]]</sup> ), and their populations are predicted to experience the greatest decrease in diversity and abundance in sub-polar and tropical areas, under RCP8.5 (Brussaard et al., 2013 <sup>[[#fn:r535|535]]</sup> ), however environmental controls of calcite production by foraminifera are still poorly understood ( ''low confidence'' ). Boyd et al. (2015b) analysis indicate also that diatoms would benefit from the synergistic effects of increased warming and iron supply in the northern Southern Ocean, as supported by laboratory experiments and field studies with polar diatoms (Rose et al., 2009 <sup>[[#fn:r536|536]]</sup> ) ( ''low confidence)'' . At low-latitude provinces, projected concurrent increases of CO 2 and iron, and decreases in both nitrate and phosphate supply, may favour nitrogen fixers, but with ocean regional variability, since iron is thought to limit N 2 fixation in the eastern Pacific and phosphorus in the Atlantic Ocean (Gruber, 2019 <sup>[[#fn:r537|537]]</sup> ; Wang et al., 2019 <sup>[[#fn:r538|538]]</sup> ). However, recent experimental work with the diazotrophic colonial ''Trichodesmium'' and the unicellular ''Crocosphaera'' have shown a broad range of responses from rising CO 2 , with either increases or decreases in N 2 fixation rates, and with mixed evidence on co-limiting processes (Eichner et al., 2014 <sup>[[#fn:r539|539]]</sup> ; Garcia et al., 2014 <sup>[[#fn:r540|540]]</sup> ; Gradoville et al., 2014 <sup>[[#fn:r541|541]]</sup> ; Walworth et al., 2016 <sup>[[#fn:r542|542]]</sup> ; Hong et al., 2017 <sup>[[#fn:r543|543]]</sup> ; Luo et al., 2019 <sup>[[#fn:r5|5]]</sup> 44) ( ''low confidence'' ).

Overall, the response of phytoplankton to the interactive effects of multiple drivers is complex, and presently ESMs do not resolve the full complexity of their physiological responses (Breitberg et al., 2015 <sup>[[#fn:r545|545]]</sup> ; Hutchins and Boyd, 2016 <sup>[[#fn:r546|546]]</sup> ; O’Brien et al., 2016 <sup>[[#fn:r547|547]]</sup> ), precluding a clear assessment of the effects of these regional distinctive multi-stressor patterns ( ''high confidence'' ).

''Future projections on zooplankton distribution and biomass''

An ensemble of 12 CMIP5 ESMs project average declines of 6.4 ± 0.79% (95% confident limits) and 13.6 ± 1.70%) in zooplankton biomass in the 21st century relative to 1990 – 1999 historical values under RCP2.6 and RCP8.5 (Kwiatkowski et al., 2019 <sup>[[#fn:r548|548]]</sup> ). Also, production of mesozooplankton is projected from a single ESM to decrease by 7.9% between 1951 – 2000 and 2051 – 2100 under RCP8.5 (Stock et al., 2014 <sup>[[#fn:r549|549]]</sup> ). Such projected decreases in zooplankton biomass and production are partly contributed by climate-induced reduction in phytoplankton production and trophic transfer efficiency particularly in low-latitude ecosystems (Stock et al., 2014 <sup>[[#fn:r550|550]]</sup> ) (5.2.2.6). The impacts may be larger than these projections if changes in the relative abundance of carbon, nitrogen and phosphorus are considered by the models (Kwiatkowski et al., 2019 <sup>[[#fn:r551|551]]</sup> ). The overall projected decrease in zooplankton biomass is characterised by a strong latitudinal differences, with the largest decrease in tropical regions and increase in the polar regions, particularly the Arctic Ocean (Chust et al., 2014 <sup>[[#fn:r552|552]]</sup> ; Stock et al., 2014 <sup>[[#fn:r553|553]]</sup> ; Kwiatkowski et al., 2019 <sup>[[#fn:r554|554]]</sup> ) (Chapter 3) ( ''high agreement'' ). However, the projected increase in zooplankton biomass in the polar region may be affected by the seasonality of light cycle at high latitudes that may limit the bloom season at high latitude (Sundby et al., 2016 <sup>[[#fn:r555|555]]</sup> ). The projected decrease in zooplankton abundance, particularly in tropical regions, can impact marine organisms higher in the foodweb, including fish populations that are important to fisheries (Woodworth-Jefcoats et al., 2017 <sup>[[#fn:r556|556]]</sup> ). Therefore, 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'' , 5.2.2.6) and simplification in representation of the zooplankton communities and foodweb render their projections having ''low confidence'' .

Future responses of zooplankton species and communities to climate change are however affected by interactions between multiple climatic drivers. Experiments in laboratory show that acidification could partly counteract some observed effects of increased temperature on zooplankton, although the level and direction of the biological responses vary largely between species (Mayor et al., 2015 <sup>[[#fn:r557|557]]</sup> ; Garzke et al., 2016 <sup>[[#fn:r558|558]]</sup> ), with results ranging from no effects (Weydmann et al., 2012 <sup>[[#fn:r559|559]]</sup> ; McConville et al., 2013 <sup>[[#fn:r560|560]]</sup> ; Cripps et al., 2014 <sup>[[#fn:r561|561]]</sup> ; Alguero-Muniz et al., 2016 <sup>[[#fn:r562|562]]</sup> ; Bailey et al., 2016 <sup>[[#fn:r563|563]]</sup> ), to negative effects (Lischka et al., 2011 <sup>[[#fn:r564|564]]</sup> ; Cripps et al., 2014 <sup>[[#fn:r565|565]]</sup> ; Alguero-Muniz et al., 2017 <sup>[[#fn:r566|566]]</sup> ) or positive effects (Alguero-Muniz et al., 2017 <sup>[[#fn:r567|567]]</sup> ; Taucher et al., 2017 <sup>[[#fn:r568|568]]</sup> ). These differences in response can affect trophic interactions between zooplankton species ; for example, some predatory non-calcifying zooplankton may perform better under warmer and lower pH conditions, leading to increased predation on other zooplankton species (Caron and Hutchins, 2012 <sup>[[#fn:r569|569]]</sup> ; Winder et al., 2017 <sup>[[#fn:r570|570]]</sup> ). Therefore, the large variation in sensitivity between zooplankton 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'' ). Consideration of these species-specific responses may further modify the projected changes in zooplankton biomass by ESMs (Boyd et al., 2015a <sup>[[#fn:r571|571]]</sup> ).

'' Future projections on fish distribution, size and biomass''

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 (Jones and Cheung, 2015 <sup>[[#fn:r572|572]]</sup> ; Robinson et al., 2015 <sup>[[#fn:r573|573]]</sup> ; Morley et al., 2018 <sup>[[#fn:r574|574]]</sup> ). Globally, the general direction of range shifts of epipelagic fishes is poleward (Jones and Cheung, 2015 <sup>[[#fn:r575|575]]</sup> ; Robinson et al., 2015 <sup>[[#fn:r576|576]]</sup> ), while the projected directions of regional and local range shifts generally follow temperature gradients (Morley et al., 2018 <sup>[[#fn:r577|577]]</sup> ). Polewards range shifts are projected to result in decreases in species richness in tropical oceans, and increases in mid to high-latitude regions leading to global-scale species turnover (sum of species local extinction and expansion) (Ben Rais Lasram et al., 2010; Jones and Cheung, 2015 <sup>[[#fn:r578|578]]</sup> ; Cheung and Pauly, 2016 <sup>[[#fn:r579|579]]</sup> ; Molinos et al., 2016 <sup>[[#fn:r580|580]]</sup> ) ( ''medium confidence'' on trends, ''low confidence'' on magnitude because of model uncertainties and limited number of published model simulations). For example, species turnover relative to their present day richness in the tropical oceans (30 o N–30 o S) is projected to be 14–21% and 37–39% by 2031–2050 and 2081–2100 under RCP8.5 (ranges of mean projections from two sets of simulation for marine fish distributions) (Jones and Cheung, 2015 <sup>[[#fn:r581|581]]</sup> ; Molinos et al., 2016 <sup>[[#fn:r582|582]]</sup> ). In contrast, high-latitude regions (&gt;60 o N–60 o S) is projected to have higher rate of species turnover than the tropics (an average of 48% between the two data sets for region &gt;60 o N). The high species turnover in the Arctic is explained by species’ range expansion from lower-latitude and the relatively lower present day fish species richness in the Arctic. The projected intensity of species turnover is lower under lower emission scenarios (Jones and Cheung, 2015 <sup>[[#fn:r583|583]]</sup> ; Molinos et al., 2016 <sup>[[#fn:r584|584]]</sup> ) (see also Section 5.4.1) ( ''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 (Cheung et al., 2013 <sup>[[#fn:r585|585]]</sup> ; Burrows et al., 2014 <sup>[[#fn:r586|586]]</sup> ; García Molinos et al., 2015 <sup>[[#fn:r587|587]]</sup> ; Jones and Cheung, 2015 <sup>[[#fn:r588|588]]</sup> ; Wabnitz et al., 2018 <sup>[[#fn:r589|589]]</sup> ) ( ''medium evidence'' , ''high agreement'' ). In addition, geographic barriers such as land boundaries in the poleward species range edge in semi-enclosed seas or lower oxygen water in deeper waters are projected to limit range shifts, resulting in larger relative decrease in species richness ( ''medium confidence'' ) (Cheung et al., 2013 <sup>[[#fn:r590|590]]</sup> ; Burrows et al., 2014 <sup>[[#fn:r591|591]]</sup> ; García Molinos et al., 2015 <sup>[[#fn:r592|592]]</sup> ; Jones and Cheung, 2015 <sup>[[#fn:r593|593]]</sup> ; Rutterford et al., 2015 <sup>[[#fn:r594|594]]</sup> ).

Warming and decrease in oxygen content is projected to impact growth of fishes, leading to reduction in body size and contraction of suitable environmental conditions (Deutsch et al., 2015 <sup>[[#fn:r595|595]]</sup> ; Pauly and Cheung, 2017 <sup>[[#fn:r596|596]]</sup> ), with the intensity of impacts being directly related to the level of climate change. The projected reduction in abundance of larger-bodied fishes could reduce predation and exacerbate the increase in dominance of smaller-bodied fishes in the epipelagic ecosystem (Lefort et al., 2015 <sup>[[#fn:r597|597]]</sup> ). Fishes exposed to ocean acidification level expected under RCP8.5 showed impairments of sensory ability and alteration of behaviour including olfaction, hearing, vision, homing and predator avoidance (Kroeker et al., 2013 <sup>[[#fn:r598|598]]</sup> ; Heuer and Grosell, 2014 <sup>[[#fn:r599|599]]</sup> ; Nagelkerken et al., 2015 <sup>[[#fn:r600|600]]</sup> ). The combined effects of warming, ocean deoxygenation and acidification in the 21st century are projected to exacerbate the impacts on the body size, growth, reproduction and mortality of fishes, and consequently increases their risk of population decline ( ''medium evidence, high agreement, high confidence'' ).

<span id="table-5.3"></span>
<!-- START TABLE -->
'''Table 5.3'''

Projected changes in total animal biomass by the mid- and end- of the 21st century under Representative Concentration Pathway (RCP)2.6 and RCP8.5. Total animal biomass is based on 10 sets of projections for each RCP under the Fisheries and Marine Ecosystems Impact Model Intercomparison Project (FISMIP) (Lotze et al. 2018 <sup>[[#fn:r601|601]]</sup> ). The very likely ranges of the projections (95% confidence intervals) are provided. Reference period is 1986–2005.

<!-- TABLE -->
{| class="wikitable"
|-
|
| colspan="4"| '''Total animal biomass (%)'''
|-
|
| colspan="2"| '''RCP2.6'''
| colspan="2"| '''RCP8.5'''
|-
| '''Region'''
| '''2031''' – '''2050'''
| '''2081''' – '''2100'''
| '''2031''' – '''2050'''
| '''2081''' – '''2100'''
|-
| '''&gt;60''' '''o''' '''N'''
| 8.4 ± 9.3
| 8.5 ± 13.7
| 7 ± 9.2
| –1.1 ± 20.2
|-
| '''30''' '''o''' '''N''' – '''50''' '''o''' '''N'''
| –8.1 ± 4
| –4.5 ± 3.6
| –10.1 ± 4.7
| –21.3 ± 9.4
|-
| '''30''' '''o''' '''N''' – '''30''' '''o''' '''S'''
| –7.2 ± 2.7
| –7.3 ± 3.1
| –9 ± 3.6
| –23.2 ± 9.5
|-
| '''30''' '''o''' '''S''' – '''50''' '''o''' '''S'''
| –3.3 ± 2.1
| –3.5 ± 2.5
| –4.2 ± 2.9
| –9 ± 9.8
|-
| '''&lt;60''' '''o''' '''S'''
| 1.7 ± 4.5
| –0.9 ± 2.9
| 0.7 ± 3.9
| 12.4 ±11.9
|}
<!-- END TABLE -->

An ensemble of global-scale marine ecosystem and fisheries models that are part of the Fisheries and Marine Ecosystems Impact Models Intercomparison Project (FISHMIP) undertook coordinated simulation experiments and projected future changes in marine animals (mainly invertebrate and fish) globally under climate change (Lotze et al., 2018 <sup>[[#fn:r602|602]]</sup> ). These models represent marine biota and ecosystems differently, ranging from population-based to functional traits- and size-based structure and their responses are driven primarily by temperature and NPP, although oxygen, salinity and ocean advection are considered in a subset of models and play a secondary role in affecting the projected changes in biomass (Blanchard et al., 2012 <sup>[[#fn:r603|603]]</sup> ; Fernandes et al., 2013 <sup>[[#fn:r604|604]]</sup> ; Carozza et al., 2016 <sup>[[#fn:r605|605]]</sup> ; Cheung et al., 2016a <sup>[[#fn:r606|606]]</sup> ). Overall, potential total marine animal biomass is projected to decrease by 4.3 ± 2.0% (95% confident intervals) 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'' ) (Figure 5.14). Accounting for the removal of biomass by fishing exacerbates the decrease in biomass for large-bodied animals which are particularly sensitive to fishing ( ''likely'' for the direction of changes). Regionally, total animal biomass decreases largely in tropical and mid-latitude oceans ( ''very'' ''likely'' ) (Table 5.3, Figure 5.14) (Bryndum-Buchholz et al., 2019 <sup>[[#fn:r607|607]]</sup> ). The high uncertainty and the ''low confidence'' in the projection in the Arctic Ocean (Chapter 3) is because of the large variations in simulation results for this region between the ESMs and between the FISHMIP models, as well as the insufficient understanding of the oceanographic changes and their biological implications in the Arctic Ocean. In the Southern Ocean, the decrease in consumer biomass is mainly in the southern Indian Ocean while other parts of the Southern Ocean are projected to have an increase in animal biomass by 2100 under RCP8.5, reflecting mainly the projected pattern of changes in NPP from the ESMs (see Section 5.2.2.6).

<span id="figure-5.14"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 5.14'''

<span id="figure-5.14-projected-changes-in-total-animal-biomass-including-fishes-and-invertebrates-based-on-outputs-from-10-sets-of-projections-for-each-representative-concentration-pathway-rcp-from-the-fisheries-and-marine-ecosystems-impact-model-intercomparison-project-fismip-www.isimip.orggettingstartedmarine--ecosystems-fisheries-lotze-et-al.-2018-a-b-multi-model-mean-change-in-un-fished-total-marine-animal"></span>
<!-- IMG CAPTION -->
'''Figure 5.14 | Projected changes in total animal biomass (including fishes and invertebrates) based on outputs from 10 sets of projections for each Representative Concentration Pathway (RCP) from the Fisheries and Marine Ecosystems Impact Model Intercomparison Project (FISMIP, www.isimip.org/gettingstarted/marine- ecosystems-fisheries) (Lotze et al. 2018); (a, b) multi-model mean change (%) in un-fished total marine animal […]'''
<!-- IMG FILE -->
[[File:6b80fa762b7c8eed204f59ee42320efe IPCC-SROCC-CH_5_14-1.jpg]]

Figure 5.14 | Projected changes in total animal biomass (including fishes and invertebrates) based on outputs from 10 sets of projections for each Representative Concentration Pathway (RCP) from the Fisheries and Marine Ecosystems Impact Model Intercomparison Project (FISMIP, www.isimip.org/gettingstarted/marine- ecosystems-fisheries) (Lotze et al. 2018); (a, b) multi-model mean change (%) in un-fished total marine animal biomass in 2085–2099 relative to 1986–2005 under RCP2.6 and RCP8.5, respectively. Dotted area represents 8 out of 10 sets of model projections agree in the direction of change (c) projected change in global total animal biomass from 1970 to 2099 under RCP2.6 (red) and RCP8.5 (blue). Variability among different ecosystem and Earth-system model combinations (n=10) expressed as the very likely range (95% confidence interval).

''Future projections on epipelagic components of the biological pump''

A wide range of studies, from laboratory experiments, mesocosm enclosures, synthesis of observations to modeling experiments, provide insights into how the multi-faceted components of the ‘biological pump’ (the physical and biologically mediated processes responsible for transporting organic carbon from the upper ocean to depth) are projected to be altered in the coming decades. A synthesis of the individual components reported to both influence the performance of the biological pump, and which are sensitive to changing ocean conditions, is presented in Table 5.4. The table lists the putative controlling of each environmental factor, such as warming, that influences the biological pump, and the reported modification (where available) of each individual factor by changing ocean conditions for both the epipelagic ocean and the deep ocean. 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 (Bopp et al., 2013 <sup>[[#fn:r608|608]]</sup> ; Boyd et al., 2015a <sup>[[#fn:r609|609]]</sup> ; Fu et al., 2016 <sup>[[#fn:r610|610]]</sup> ; Laufkötter et al., 2016 <sup>[[#fn:r611|611]]</sup> ). However, different lines of evidence (including observation, modeling and experimental studies) provide ''low confidence'' on the mechanistic understanding of how climatic 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 (see section below and Table 5.4); this renders the projection of future contribution of the biological carbon pump to the export of POC to the deep ocean having ''low confidence'' .

<!-- END IMG -->
<span id="table-5.4"></span>
<!-- START TABLE -->
'''Table 5.4'''

Projected future changes to the ocean biological pump (adapted from Boyd et al. (2015a)). Environmental controls on individual factors that influence downward POC flux are based on published reports from experiments (denoted by '''E''' ), modelling simulations ( '''M''' ) and observations ( '''O''' ). In some cases, due to the paucity, and regional specificity, of published reports it has been indicated the sign of the projected change on export (in italics), as opposed to the magnitude. NPP: Net Primary Production; POC: Particulate Organic Carbon; DOC: Dissolved Organic Carbon; TEP: Transparent Exopolymer Particles; OA: Ocean Acidification. Climate change denotes multiple controls such as nutrients, temperature and irradiance, as parameterised in coupled ocean atmosphere models. *denotes observation for low latitudes only. '''**''' represents major uncertainty over environmental modulation of this component of the biological pump. ***denotes joint influence of temperature and acidification.

<!-- TABLE -->
{| class="wikitable"
|-
| '''Pump component'''
| '''Oceanic driver'''
| '''Projected change (by year 2100)'''
| '''Confidence'''
| '''References &amp; Lines of evidence'''
|-
| '''Epipelagic Ocean'''
|
|-
| Phytoplankton growth
| Temperature (warming)
| ~10% Faster (nutrient-replete) no change (nutrient-deplete)
| High
| (Boyd et al., 2013 <sup>[[#fn:r613|613]]</sup> ) '''E''' ; (Maranon et al., 2014 <sup>[[#fn:r614|614]]</sup> ) '''O*'''
|-
| NPP
| Climate change (temperature, nutrients, CO 2 )
| 10 – 20% decrease (low latitudes);
10 – 20% increase (high latitudes)

| Medium
| (Bopp et al., 2013) '''M'''
|-
| Partitioning of NPP (POC, TEP, DOC)
| OA
| ~20% increase in TEP production
| Medium
| (Engel et al., 2014 <sup>[[#fn:r615|615]]</sup> ) '''E''' ; (Riebesell et al., 2007 <sup>[[#fn:r616|616]]</sup> ) '''E''' ; (Seebah et al., 2014 <sup>[[#fn:r617|617]]</sup> ) '''E'''
|-
| Food web retention of NPP
| OA
| Enhanced transfer of organic matter to higher trophic levels, reduced N and P sedimentation by 10%
| Low
| (Boxhammer et al., 2018) '''E'''
|-
| Floristic shifts
| Climate change (warming, salinity, OA, iron)
| Shift to smaller or larger cells
( ''less export vs more export; inconclusive'' )

| Low
| (Moràn et al., 2010 <sup>[[#fn:r618|618]]</sup> ) '''O''' ; (Li et al., 2009 <sup>[[#fn:r619|619]]</sup> ) '''O;''' (Dutkiewicz et al., 2013a <sup>[[#fn:r620|620]]</sup> ) '''M;''' (Tréguer et al., 2018 <sup>[[#fn:r621|621]]</sup> ) '''O;''' (Sett et al., 2014 <sup>[[#fn:r622|622]]</sup> ) '''E'''
|-
| Differential susceptibility
| Temperature (warming)
| Growth-rate of grazers more temperature dependent than prey
( ''less export'' )

| Low
| (Rose and Caron, 2007 <sup>[[#fn:r626|626]]</sup> ) '''O'''
|-
| Bacterial hydrolytic effects
| Warming, OA
| Increase under warming and low pH (variable response in different plankton communities)
| Low
| (Burrell et al., 2017 <sup>[[#fn:r623|623]]</sup> ) '''E'''
|-
| Grazer physiological responses
| Warming
| Copepods had faster respiration and ingestion rates, but higher mortality
( ''inconclusive'' )

| Low
| (Isla et al., 2008 <sup>[[#fn:r627|627]]</sup> ) '''E'''
|-
| Faunistic shifts
| Temperate and subpolar zooplankton species shifts
| Temperature
( ''inconclusive'' )

| Low
| (Edwards et al., 2013 <sup>[[#fn:r628|628]]</sup> ) '''O'''
|-
| Food web amplification
| Warming
| Zooplankton negatively amplify the climate change signal that propagates up from phytoplankton in tropical regions, and positively amplify in polar regions
| Low
| (Chust et al., 2014) '''M''' ; (Stock et al., 2014) '''M'''
|-
| '''Deep Ocean'''
|
|-
| Bacterial hydrolytic enzyme activity
| Temperature
| 20% increase (resource-replete) to no change (resource-deplete)
| Low
| (Wohlers-Zöllner et al., 2011 <sup>[[#fn:r642|642]]</sup> ) '''E''' ; (Endres et al., 2014 <sup>[[#fn:r643|643]]</sup> ) '''E''' ; (Bendtsen et al., 2015 <sup>[[#fn:r644|644]]</sup> ) '''E''' ; (Piontek et al., 2015) '''E***'''
|-
| Particle sinking rates (viscosity)
| Warming
| 5% faster sinking/°C warming
| Low
| (Taucher et al., 2014) '''M'''
|-
| Mesozooplankton community composition
| Temperature '''**'''
| Shifts which increase/decrease
particle transformations

( ''less/more export, respectively'' )

| Low
| (Burd and Jackson, 2002 <sup>[[#fn:r646|646]]</sup> ) M ; (Ikeda et al., 2001 <sup>[[#fn:r647|647]]</sup> ) '''O'''
|-
| Vertical migrators
| Climate change (irradiance, temperature)
| ( ''more export'' )
| Low
| (Almén et al., 2014) '''O''' ; (Berge et al., 2014) '''O'''
|-
| Deoxygenation
| Climate change
| ( ''more export'' )
| Low
| (Rykaczewski and Dunne, 2010 <sup>[[#fn:r648|648]]</sup> ) '''M''' ; (Cocco et al., 2013 <sup>[[#fn:r649|649]]</sup> ) '''O;''' (Hofmann and Schellnhuber, 2009 <sup>[[#fn:r650|650]]</sup> ) '''M'''
|}
<!-- END TABLE -->

<div id="section-5-2-3-2the-deep-pelagic-ocean"></div>

<span id="the-deep-pelagic-ocean"></span>
==== 5.2.3.2 The Deep Pelagic Ocean ====

<div id="section-5-2-3-2the-deep-pelagic-ocean-block-1"></div>

<span id="detection-and-attribution-of-biological-changes-in-the-deep-ocean"></span>
===== 5.2.3.2.1 Detection and attribution of biological changes in the deep ocean =====

The pelagic realm of the deep ocean represents a key site for remineralisation of organic matter and long-term biological carbon storage and burial in the biosphere (Arístegui et al., 2009), but the observed effects of climate change on deep sea organisms, communities and biological processes are largely unknown ( ''high confidence'' ). Observational and model-based methods provide ''limited evidence'' that the transfer efficiency of organic carbon to the sea floor is partly controlled by temperature and oxygen in the mesopelagic zone, affecting microbial metabolism and zooplankton community structure, with highest efficiencies for high-latitude and OMZ) (see Section 5.2.2.4 for more detail on OMZs), while below 1000 m organic carbon transfer is controlled by particle sinking speed (Boyd et al., 2015a; Marsay et al., 2015; DeVries and Weber, 2017). However, there are contrasting results and ''low confidence'' on whether transfer efficiencies are highest at low or high latitudes (Boyd et al., 2015a; Marsay et al., 2015; Guidi et al., 2016; DeVries and Weber, 2017; Sweetman et al., 2017). There is also ''low confidence'' on the effects of increasing temperatures on POC remineralisation to CO 2 versus POC solubilisation to dissolved organic carbon (DOC) by microbial communities and its storage as refractory DOC (i.e., with life times of &gt;16,000 years) (Legendre et al., 2015).

<div id="section-5-2-3-2the-deep-pelagic-ocean-block-2"></div>

<span id="future-changes-in-the-deep-ocean"></span>
===== 5.2.3.2.2 Future changes in the deep ocean =====

The global magnitude of the biological pump and how this will be affected by climate change is also uncertain. Model-based studies agree in projecting a global decline in particle gravitational flux to the deep sea floor, but with regional variability in both the total particle export flux and transfer efficiency (DeVries and Weber, 2017; Sweetman et al., 2017) (see Sections 5.2.2 and 5.2.4). However, recent evidence suggest that other physical and biological processes may contribute nearly as much as the gravitational flux to the carbon transport from the surface to the deep ocean (Boyd et al., 2019), with ''low confidence'' on the future rate of change in magnitude and direction of these processes. In particular, the ‘active flux’ of organic carbon due to vertical migration of zooplankton and fishes has been reported to account from 10 to 40% of the gravitational sinking flux (Bianchi et al., 2013; Davison et al., 2013; Hudson et al., 2014; Jónasdóttir et al., 2015; Aumont et al., 2018; Gorgues et al., 2019). Predictions based on model studies suggest that mesopelagic zooplankton and fish communities living at deep scattering layers (DSLs) will increase their biomass by 2100, enhancing their trophic efficiency, because of deep-ocean warming (Section 5.2.2.1; Figures 5.2 and 5.3) and shallowing of DSL (Proud et al., 2017) ( ''low confidence'' ). Expansion of OMZs (see Section 5.2.2.4) will also widen the DSL and increase the exposure of mesopelagic organisms to shallower depths (Gilly et al., 2013; Netburn and Anthony Koslow, 2015). In the California Current, the abundance of mesopelagic fishes is closely tied to variations in the OMZ, whose dynamic is linked to the Pacific Decadal Oscillation and ENSO cycles (Koslow et al., 2015). Some large predators, like the Humboldt squid, could indirectly benefit from expanding OMZs due to the aggregation of their primary food source, myctophid fishes (Stewart et al., 2014). However, many non-adapted fish and invertebrates (like diurnal vertical migrators) will have their depth distributions compressed, affecting the carbon transport and trophic efficiency of food webs in the mesopelagic (Stramma et al., 2011; Brown and Thatje, 2014; Rogers, 2015) ( ''low confidence'' ). In OMZ waters, where zooplankton is almost absent, like in the Eastern Tropical North Pacific, the microbial remineralisation efficiency of sinking particles would be reduced, eventually increasing the transfer efficiency of organic matter to the deep ocean and thus biological carbon storage (Cavan et al., 2017) ( ''low confidence;'' Table 1). However, increases in ocean temperature may also lead to shallower remineralisation of POC in warm tropical regions, counteracting the storage of carbon in the dark ocean (Marsay et al., 2015). Overall, the direct impacts of climate change on the biological pump are not well understood for the deep pelagic organisms and ecosystems (Pörtner et al., 2014), and there is ''low confidence'' on the effect of climate change drivers on biological processes in the deep ocean (Table 5.1).

<span id="impacts-on-deep-seafloor-systems"></span>
=== 5.2.4 Impacts on Deep Seafloor Systems ===

<div id="section-5-2-4-1changes-on-the-deep-seafloor"></div>

<span id="changes-on-the-deep-seafloor"></span>
==== 5.2.4.1 Changes on the Deep Seafloor ====

<div id="section-5-2-4-1changes-on-the-deep-seafloor-block-1"></div>

The deep seafloor is assessed here as the vast area of the ocean bottom &gt;200 m deep, beyond most continental shelves (Levin and Sibuet, 2012; Boyd et al., 2019) (Figure 5.15). Below 200 m changes in light, food supply and the physical environment lead to altered benthic (seafloor) animal taxonomic composition, morphologies, lifestyles and body sizes collectively understood to represent the deep sea (Tyler, 2003).

<div id="section-5-2-4-1changes-on-the-deep-seafloor-block-2"></div>

Attachment field is blank. Please add an attachment or remove this block.

<div id="section-5-2-4-1changes-on-the-deep-seafloor-block-3"></div>

Most deep seafloor ecosystems globally are experiencing rising temperatures, declining oxygen levels, and elevated CO 2 , leading to lower pH and carbonate undersaturation (WGII AR5 30.5.7; Section 5.2.2.3). Small changes in exposure to these hazards by deep seafloor ecosystem have been confirmed by observation over the past 50 years. However, analysis using direct seafloor observations of these hazards over the past 15-29 years suggest that the environmental conditions are highly variable over time because of the strong and variable influences by ocean conditions from the sea surface (Frigstad et al., 2015; Thomsen et al., 2017). Such high environmental variability makes it difficult to attribute observed trends to anthropogenic drivers using existing datasets (Smith et al., 2013; Hartman et al., 2015; Soltwedel et al., 2016; Thomsen et al., 2017) ( ''high confidence'' ). Projections from global ESMs suggest large changes for temperature by 2100 and beyond under RCP8.5 (relative to present day variation) (Mora et al., 2013; Sweetman et al., 2017; FAO, 2019). The magnitude of the projected changes is lower under RCP2.6, and in some cases the direction of projected change to 2100 varies regionally under either scenario (FAO, 2019) ( ''high confidence'' ).

<div id="section-5-2-4-2open-ocean-seafloor-abyssal-plains-3000-6000-m"></div>

<span id="open-ocean-seafloor---abyssal-plains-3000-6000-m"></span>
==== 5.2.4.2 Open Ocean Seafloor - Abyssal Plains (3000-6000 m) ====

<div id="section-5-2-4-2open-ocean-seafloor-abyssal-plains-3000-6000-m-block-1"></div>

Abyssal communities (3000–6000 m) cover over 50% of the ocean’s surface and are considered to be extremely food limited (Gage and Tyler, 1992; Smith et al., 2018 <sup>[[#fn:r684|684]]</sup> ). There is a strong positive relationship between surface primary production, export flux, and organic matter supply to the abyssal seafloor (Smith et al., 2008 <sup>[[#fn:r685|685]]</sup> ), with pulses of surface production reflected as carbon input on the deep seafloor in days to months (Thomsen et al., 2017 <sup>[[#fn:r686|686]]</sup> ). Both vertical and horizontal transport contribute organic matter to the sea floor (Frischknecht et al., 2018 <sup>[[#fn:r687|687]]</sup> ). Food supply to the seafloor regulates faunal biomass, explaining the strong positive relationships documented between surface production and seafloor faunal biomass in the Pacific Ocean (Smith et al., 2013 <sup>[[#fn:r688|688]]</sup> ), Gulf of Mexico (Wei et al., 2011 <sup>[[#fn:r689|689]]</sup> ) and north Atlantic Ocean (Hartman et al., 2015 <sup>[[#fn:r690|690]]</sup> ). Extended time series and broad spatial coverage reveal strong positive relationship between annual POC flux and abyssal sediment community oxygen consumption (Rowe et al., 2008 <sup>[[#fn:r691|691]]</sup> ; Smith et al., 2016a <sup>[[#fn:r692|692]]</sup> ). Observed reduction in in POC flux at the abyssal seafloor enhances the relative importance of the microbial loop and reduces the importance of benthic invertebrates in carbon transfer (Dunlop et al., 2016 <sup>[[#fn:r693|693]]</sup> ) (single study, ''limited evidence'' ). However, changes in the overlying mesopelagic and bathypelagic communities (see Section 5.2.3.2) will also affect food flux to the deep seafloor, as nekton and zooplankton transfer energy to depth through diel (daily day-night) vertical migrations, ontogenetic (life staged-based) migrations and falls of dead carcasses (Gage, 2003 <sup>[[#fn:r694|694]]</sup> ). Therefore, climate change impacts on organic carbon export from the epipelagic (Section 5.2.3.1) and deeper pelagic systems (Section 5.2.3.2) can affect the energy available to support the abyssal seafloor ecosystems ( ''medium confidence'' ). However, because observations on historical changes in POC flux in abyssal seafloor ecosystems are limited to a few locations, long-term records show high variability, and mechanistic understanding of factors affecting the biological carbon pump is incomplete, there is ''limited evidence'' that the abyssal seafloor ecosystem has already been affected by changes in POC flux as a result of climate change. The metabolic rate of deep seafloor ectotherms, and consequently their demand for food, increases with temperature. Thus, observed warming in deep sea ecosystems (Hoegh-Guldberg et al., 2014 <sup>[[#fn:r695|695]]</sup> ) (Section 5.2.2.2.1) is expected to increase the sensitivity of deep seafloor biota to decrease in food supplies associated with a change in POC flux ( ''high confidence'' ). However, there is ''limited evidence'' of observed changes in abyssal biota. Small deep sea biota demonstrate increased efficiency (effective use of food energy for growth and metabolism with minimal loss) at low food inputs (due to small size and dominance by prokaryotic taxa) (Gambi et al., 2017 <sup>[[#fn:r696|696]]</sup> ). Adaptation to low food availability in abyssal ecosystems may confer higher capacity to adjust to reduced food availability than for shallow biota ( ''limited evidence).'' Overall, the risk of impacts of climate change on abyssal ecosystems through reduction in food supplies from declining POC flux in the present day is low with ''low confidence'' .

The globally integrated export flux of carbon is projected to decrease in the open ocean in the 21st century under RCP2.6 (by 1.6–4.9%) and RCP8.5 (by 8.9–15.8%) relative to 2000 ( ''medium confidence'' ) (Section 5.2.2.6). This change in export flux of carbon is projected to yield declines in POC flux at the abyssal seafloor (representing food supply to benthos) of up to –27% in the Atlantic and up to –31 to –40% in the Pacific and Indian Oceans, with some increases in polar regions (Sweetman et al. 2017 <sup>[[#fn:r697|697]]</sup> ). In some models, additional dissolution of calcium carbonate due to ocean acidification further lowers POC flux, causing the projected export production declines to be up to 38% at the northeast Atlantic seafloor (Jones et al., 2014 <sup>[[#fn:r699|699]]</sup> ). Lower POC fluxes to the abyss reduce food supply and have been projected to cause a size-shift towards smaller organisms (Jones et al., 2014), resulting in rising respiration rates, lower biomass production efficiency, and lesser energy transfer to higher trophic levels (Brown et al., 2004 <sup>[[#fn:r700|700]]</sup> ) ( ''medium confidence'' ). Changes are projected to be largest for macrofauna and lesser and similar for megafauna and meiofauna (Jones et al., 2014) ( ''limited evidence'' , ''low confidence'' ). Projections using outputs from seven CMIP5 models suggest that 97.8 ± 0.6% (95% CI) of the abyssal seafloor area will experience a biomass decline by 2091–2100 relative to 2006–2015 under RCP8.5. The projected decreases in overall POC flux to the abyssal seafloor are projected to cause a 5.2–17.6% reduction in seafloor biomass in 2090–2100, relative to 2006–2015 under RCP8.5 (Jones et al., 2014 <sup>[[#fn:r701|701]]</sup> ). The projected impacts on abyssal seafloor biomass are significantly larger under RCP8.5 than RCP4.5 (Jones et al., 2014 <sup>[[#fn:r703|703]]</sup> ). However, existing estimates are based on total POC flux changes and do not account for changes in the type or quality of the sinking material, to which macrofaunal and meiofaunal invertebrates are highly sensitive (Smith et al., 2008 <sup>[[#fn:r704|704]]</sup> ; Smith et al., 2009 <sup>[[#fn:r705|705]]</sup> ; Tittensor et al., 2011 <sup>[[#fn:r706|706]]</sup> ). The projections also do not account for direct faunal responses to changes in temperature, oxygen or the carbonate system, all of which will influence benthic responses to changing food availability (AR5 Chapter 30.5.7), reducing to ''medium confidence'' the risk assessment that is based on these projections (Figure 5.16).

Regionally, while reductions in POC flux are projected at low and mid latitudes in the Pacific, Indian and Atlantic Oceans, increases are projected at high latitudes associated in part with reduction in sea ice cover (Yool et al., 2013 <sup>[[#fn:r707|707]]</sup> ; Rogers, 2015 <sup>[[#fn:r708|708]]</sup> ; Sweetman et al., 2017 <sup>[[#fn:r709|709]]</sup> ; Yool et al. 2017 <sup>[[#fn:r710|710]]</sup> ; FAO 2019 <sup>[[#fn:r711|711]]</sup> ) (see Chapter 3) ( ''medium confidence'' ). Notably, Arctic and Southern Ocean POC fluxes at the abyssal seafloor are projected to increase by up to 38% and 21%, respectively by 2100 under RCP8.5 (Sweetman et al., 2017 <sup>[[#fn:r712|712]]</sup> ). While an increase in food supply may yield higher benthic biomass at high latitudes, warmer temperatures and reduced pH projected for the polar regions (Chapter 3) would elevate faunal metabolic demands, likely diminishing the benefit of elevated food supply to an unknown extent (Sweetman et al., 2017 <sup>[[#fn:r713|713]]</sup> ). Overall, given the limited food availability for fauna in the abyssal plains and the projected warming (Section 5.2.2.2.2) that increases the demand for food to support the elevated metabolic rates, the projected decrease in influx of organic matter and seafloor biomass will result in high risks of impacts to abyssal ecosystems by the end of the 21st century under RCP8.5 ( ''medium confidence'' ) (Figure 5.16). The risk of impacts is projected to be substantially lower under RCP4.5 or RCP2.6 ( ''high confidence'' ). The impacts on abyssal seafloor ecosystems affect functions that are important to support ecosystem services (see Section 5.4.1). For example, smaller-sized organisms exhibit reduced bioturbation intensity and depth of mixing causing reduced carbon sequestration (Smith et al., 2008 <sup>[[#fn:r714|714]]</sup> ) (Figure 5.15).

<div id="section-5-2-4-2open-ocean-seafloor-abyssal-plains-3000-6000-m-block-2"></div>

<span id="figure-5.15"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 5.15'''

<span id="figure-5.15-a-conceptual-diagram-illustrating-how-climate-drivers-are-projected-to-modify-deep-sea-ecosystems-as-discussed-in-section-5.2.4."></span>
<!-- IMG CAPTION -->
'''Figure 5.15 | A conceptual diagram illustrating how climate drivers are projected to modify deep sea ecosystems as discussed in Section 5.2.4.'''
<!-- IMG FILE -->
[[File:31366bf5606fb7d2d2844e775476da50 IPCC-SROCC-CH_5_15-3000x2465.jpg]]

Figure 5.15 | A conceptual diagram illustrating how climate drivers are projected to modify deep sea ecosystems as discussed in Section 5.2.4.

<!-- END IMG -->
<div id="section-5-2-4-3bathyal-ecosystems-200-3000-m"></div>

<span id="bathyal-ecosystems-2003000-m"></span>
==== 5.2.4.3 Bathyal Ecosystems (200–3000 m) ====

<div id="section-5-2-4-3bathyal-ecosystems-200-3000-m-block-1"></div>

Bathyal ecosystems consist of numerous geomorphic features with steep topography (Figure 5.15). These include continental slopes covering 5.2% of the seafloor, over 9400 steep-sided canyons, and &gt;9000 conical seamounts (submarine volcanos which are mainly inactive), as well as guyots and ridges which together cover ~6% of the seafloor (Harris et al., 2014) ''.'' Seamounts and canyons support high animal densities and biomass including cold water coral, sponge and bryozoan reefs, exhibit high secondary production supported by locally enhanced primary production and intensified water flow, function as diversity hotspots and serve as stepping stones for larval dispersal (Rowden et al., 2010). Canyons transport particulate organic matter, migrating plankton and coarse material from the shelf, and are sites where intensified mixing and advection of water masses occurs (De Leo et al., 2010; Levin and Sibuet, 2012; Fernandez-Arcaya et al., 2017).  Slopes, canyons and seamounts exhibit strong vertical temperature, oxygen and pH gradients generating sharp ecological zonation (Levin and Sibuet, 2012), thus changes in exposures are expected to alter the distributions of their communities (Figure 5.15, 5.16 '') (medium confidence).  ''

In some regions, observational records document changing conditions in bathyal ecosystems (Levin, 2018; Section 5.2.2.4). In the Northeast Pacific continental slopes associated with the California Current ecosystem, observations over the past 25 years show high variability but an overall trend of decreasing ocean oxygen and pH levels with oxygen declines of up to 40% and pH declines of 0.08 units in California . (Goericke et al., 2015) (high agreement, robust evidence, high confidence). Large oxygen declines are linked to past warming events on continental margins, over multiple time scales from 1–100 ky (Dickson et al., 2012; Moffitt et al., 2015). Studies across modern oxygen gradients on slopes reveal that suboxic (5–10 µMol kg -1 O 2 ) values lead to loss of biodiversity of fish (Gallo and Levin, 2016), invertebrates (Levin, 2003; Gallo and Levin, 2016; Sperling et al., 2016), and protozoans (Bernhard and Reimers, 1991; Gooday et al., 2000; Moffitt et al., 2014) (high confidence). Shoaling oxyclines on continental slopes have altered depth distributions of multiple co-occurring echinoid species over the past 25 years (Sato et al., 2017) and can reduce the growth rate, and change the skeletal structure and biochemical composition of a common sea urchin (Sato et al., 2018). In central Pacific oceanic canyons, fish abundance and diversity are reduced at 4 to 5 times higher oxygen concentrations than on continental slopes (&lt;31 µMol kg -1 O 2 ) (De Leo et al., 2012). Low oxygen on continental slopes causes reductions in faunal body size and bioturbation (Diaz and Rosenberg, 1995; Levin, 2003; Middelburg and Levin, 2009; Sturdivant et al., 2012), simplification of trophic structure reducing energy flow to upper trophic levels (Sperling et al., 2013), shifts in carbon processing pathways from metazoans to protozoans (Woulds et al., 2009), and reduced colonisation potential (Levin et al., 2013). These changes are expected to lead to altered ecosystem structure and function, with lower carbon burial (Smith et al., 2000; Levin and Dayton, 2009) (medium confidence). Both carbon sequestration and nitrogen recycling are highly sensitive to small changes in oxygenation within the suboxic zone (Deutsch et al., 2011).

Bathyal species adapted to OMZs where CO 2 levels are characteristically high, appear less vulnerable to the negative impacts of ocean acidification (Taylor et al., 2014). Benthic foraminifera, which are often the numerically dominant deep sea taxon, show no significant effect of short-term exposure to ocean acidification on survival of multiple species (Dissard et al., 2010; Haynert et al., 2011; Keul et al., 2013; McIntyre-Wressnig et al., 2014; Wit et al., 2016) and in fact hypoxia in combination with elevated pCO 2 favors survival of some foraminifera (Wit et al., 2016). However, lower pH exacerbates shallow foraminiferal sensitivity to warming (Webster et al., 2016). Limited evidence suggests that combined declines in pH and oxygen may lead to increase in some agglutinating taxa and a decrease in carbonate-producing foraminifera, including those using carbonate cement (van Dijk et al., 2017). Exposure to acidification (0.4 unit pH decrease) reduces fecundity and embryo development rate in a bathyal polychaete. Where both oxygen and CO 2 stress occur together on bathyal slopes, oxygen can be the primary driver of change (Taylor et al. 2014; Sato et al. 2018). Nematodes are sensitive to changes in temperature (Danovaro et al., 2001; Danovaro et al., 2004; Yodnarasri et al., 2008) and elevated CO 2 (Barry et al., 2004; Fleeger et al., 2006; Fleeger et al., 2010). There is low agreement about the direction of meiofaunal responses among studies, reflecting opposing responses in different regions. However, there is high agreement that meiofauna are sensitive to change in environment and food supply ( ''medium confidence'' ). Additional research is needed across all taxa on how hypoxia and pH interact (Gobler and Baumann, 2016).

Continental slopes, seamounts and canyons (200–2500 m) are projected to experience significant warming, pH decline, oxygen loss and decline in POC flux by 2081–2100 (compared to 1951–2000) under RCP8.5 (Table 5.5). In contrast, the average changes are projected to be 30–50% less under RCP2.6 (Table 5.5) by 2081-2100. Most ocean regions at bathyal depths (200–2500 m) except the Southern and Arctic Oceans are predicted to experience on average declining export POC flux under RCP8.5 by 2081–2100 (Yool et al., 2017; FAO, 2019) with the largest declines of 0.7–8.1 mg C m -2 d -1 in the Northeast Atlantic (FAO, 2019). There is a strong macroecological relationship between depth, export POC flux, biomass and zonation of macrobenthos on continental slopes (Wei et al., 2011), such that lower POC fluxes will alter seafloor community biomass and structure ( ''medium confidence'' ) (See also Section 5.2.4.1). This is modified on the local scale by near-bottom currents, which alter sediment grain size, food availability, and larval dispersal (Wei et al., 2011).

Declines in faunal biomass (6.1 ± 1.6% 95% C.I) are predicted for 96.6 ± 1.2% of seamounts under RCP8.5 by 2091–2100 relative to 2006–2015, driven by a projected 13.8 ± 3.3% drop in POC flux (Jones et al., 2014). The majority (85%) of mapped canyons are projected to experience comparable benthic biomass declines (Jones et al., 2014). By 2100 under RCP8.5, pH reductions exceeding -0.2 pH units are projected in ~ 23% of north Atlantic deep sea canyons and 8% of seamounts (Gehlen et al., 2014), with potential negative consequences for their cold water coral habitats (See Box 5.2).

Mean temperature (warming) signals are projected to emerge from background variability before 2040 in canyons of the Antarctic, northwest Atlantic, and South Pacific (FAO, 2019). Enhanced stratification and change in the intensity and frequency of downwelling processes under atmospheric forcing (including storms and density-driven cascading events would alter organic matter transported through canyons (Allen and Durrieu de Madron, 2009) ( ''low confidence'' ). Changes in the quantity and quality of transferred particulate organic matter, as well as physical disturbance during extreme events cause a complex combination of positive and negative impacts at different depths along the canyon floor (Canals et al., 2006; Pusceddu et al., 2010). Canyons and slopes are recognised as hosting many methane seeps and other chemosynthetic habitats (e.g., whale and wood falls) supported by massive transport of terrestrial organic matter (Pruski et al., 2017); their climate vulnerabilities are discussed below.

Seamounts have been proposed to serve as refugia for cold water corals facing shoaling aragonite saturation horizons (Tittensor et al., 2011), but could become too warm for deep-water corals in some regions (e.g., projections off Australia) (Thresher et al., 2015) ( ''one study, low confidence'' ). Seamounts are major spawning grounds for fishes; reproduction on seamounts may be disrupted by warming (Henry et al., 2016) ( ''one study, low confidence'' ). In the north Atlantic, models suggest seamounts are an important source of cold water coral larvae that maintain resilience under shifting NAO conditions (Fox et al., 2016), thus loss of suitable seamount habitat may have far-reaching consequences (Gehlen et al., 2014) ( ''low confidence'' ) (also see Box 5.2).

<div id="section-5-2-4-3bathyal-ecosystems-200-3000-m-block-2"></div>

<span id="table-5.5"></span>
<!-- START TABLE -->
'''Table 5.5'''

Projected climate changes from the present to 2081–2100 given as mean (min, max) at the deep seafloor for continental slopes, canyons, seamounts and cold water corals mapped from 200–2500 m under RCP8.5 and RCP2.6 Projections are based on three 3D, fully coupled earth system models (ESMs) (as part of CMIP5): the Geophysical Fluid Dynamics Laboratory’s ESM 2G (GFDL-ESM-2G); the Institut Pierre Simon Laplace’s CM6-MR (IPSL-CM5A-MR); and (iii) the Max Planck Institute’s ESM-MR (MPI-ESM-MR). Export flux at 100 m was converted to export POC flux at the seafloor (epc) using the Martin curve following the equation: epc = epc100 (depth/ export depth)-0.858. Projections were made onto the (i) slope from a global ocean basin mask from World Ocean Atlas 2013 V2 (NOAA, 2013), (ii) global distribution of submarine canyons with canyon heads shallower than 1500 m (Harris and Whiteway, 2011); (iii) global distribution of seamounts with summits between 200–2500 m (Kim et al. 2011); and (iv) global occurrence of cold water corals between 200–2500 m (Freiwald et al. 2017).

<!-- TABLE -->
{| class="wikitable"
|-
| ''' '''
| '''  Temperature '''
'''        (''' '''o''' '''C)'''

| '''      pH '''
| '''DO '''
'''(''' ''µ'' '''Mol kg''' '''-1''' ''')'''

| '''POC flux'''
'''(mgC m''' '''-2''' '''d''' '''-1''' ''')'''

|-
| ''' '''
| '''RCP2.6'''
| ''' RCP2.6'''
|-
| Continental slopes
| +0.30 (–0.44, + 2.30)
| –0.06 (–0.19, –0.02)
| –3.1 (–49.3, +61.7)
| –0.39 (–16.0, +3.9)
|-
| Canyons
| +0.31 (–0.27, +1.76)
| -0.05 (-0.13, +0.01)
| –3.5 (–44.7, +29.3)
| –0.33 (–10.53, +3.53)
|-
| Seamounts
| +0.13 (+0.01, +0.67)
| -0.02 (-0.11, +0.005)
| –3.46 (–18.9, +4.1)
| –0.15 (–2.20, +1.33)
|-
| Cold water corals
| +4.3 (–0.29, +1.85)
| -0.07 (-0.13, 0.0)
| –3.5 (–25.6, +24.7)
| –0.7 (–10.5, +3.4)
|-
|
|-
|
| '''RCP8.5'''
|-
| Continental slopes
| +0.75 (–8.4, +4.4)
| –0.14 (–0.44, –0.02)
| –10.2 (–67.8, +53.8)
|  –0.66 (–33.33, +10.3)
|-
| Canyons
| +0.19 (–0.03, +1.14)
| -0.11 (-0.35, +0.02)
| –0.8 (–28.8, +10.1)
| –0.80 (–28.76, +10.07)
|-
| Seamounts
| +0.66 (–0.75, +3.19)
| -0.03 (-0.19, +0.001)
| –0.50 (–7.2, +3.0)
| –0.50 (–7.18, +2.98)
|-
| Cold water corals
| +0.96 (–0.42, +3.84)
| -0.15 (-0.39, +0.001)
| –10.6 (–59.2, +11.1)
| –1.69 (–20.1, +4.6)
|}
<!-- END TABLE -->

<div id="section-5-2-4-4chemosynthetic-ecosystems"></div>

<span id="chemosynthetic-ecosystems"></span>
==== 5.2.4.4 Chemosynthetic Ecosystems ====

<div id="section-5-2-4-4chemosynthetic-ecosystems-block-1"></div>

Despite having nutrition derived largely from chemosynthetic sources fueled by fluids from the earth’s interior, hydrothermal vent and methane seep ecosystems are linked to surface ocean environments and water-column processes in many ways that can expose them to aspects of climate change ( ''medium confidence'' ). The reliance of vent and seep mussels on surface-derived photosynthetic production to supplement chemosynthetic food sources (Riou et al., 2010 <sup>[[#fn:r778|778]]</sup> ; Riekenberg et al., 2016 <sup>[[#fn:r779|779]]</sup> ; Demopoulos et al., 2019 <sup>[[#fn:r780|780]]</sup> ), and in some cases as a cue for synchronised gametogenesis (sperm and egg production) (Dixon et al., 2006 <sup>[[#fn:r781|781]]</sup> ; Tyler et al., 2007 <sup>[[#fn:r782|782]]</sup> ) can make them vulnerable to changing amounts or timing of POC flux to the deep seabed in most areas except high latitudes, or to changes in timing of surface production (see Section 5.2.2.5) ( ''limited evidence'' ) Most of the large, habitat-forming (foundation) species at vents and seeps such as mussels, tubeworms, and clams require oxygen to serve as electron acceptor for aerobic hydrogen-, sulfide- and methane oxidation (Dubilier et al., 2008 <sup>[[#fn:r783|783]]</sup> ) and appear unable to grow under dysoxic conditions (&lt;5–10 µmol kg –1 O 2 ) (Sweetman et al., 2017 <sup>[[#fn:r784|784]]</sup> ) ( ''medium confidence'' ). The distributions of these taxa at seeps could be constrained by climate-driven expansion of midwater oxygen minima (Stramma et al., 2008 <sup>[[#fn:r785|785]]</sup> ; Schmidtko et al., 2017 <sup>[[#fn:r786|786]]</sup> ), which is occurring at water depths where seep ecosystems typically occur on continental margins (200–1000 m). Rising bottom temperatures or shifting of warm currents on continental margins could increase dissociation of buried gas hydrates on margins (Phrampus and Hornbach, 2012 <sup>[[#fn:r787|787]]</sup> ) ( ''low confidence'' ) potentially intensifying anaerobic methane oxidation (which produces hydrogen sulfide) (Boetius and Wenzhoefer, 2013 <sup>[[#fn:r788|788]]</sup> ) and expanding cover of methane seep communities ( ''limited'' evidence). Larvae of vent species such as bathymodiolin mussels, alvinocarid shrimp, and some limpets that develop in or near surface waters (Herring and Dixon, 1998 <sup>[[#fn:r789|789]]</sup> ; Arellano et al., 2014 <sup>[[#fn:r790|790]]</sup> ), are likely to be exposed to warming waters, decreasing pH and carbonate saturation states, and in some places, reduced phytoplankton availability (Section 5.2.2), causing reduced calcification and growth rates (as in shallow water mussel larvae, Frieder et al. (2014)) ( ''limited evidence, low confidence'' ). Larvae originating at vents or seeps beneath upwelling regions may also be impaired by effects of hypoxia associated with expanding OMZ (Stramma et al., 2008 <sup>[[#fn:r791|791]]</sup> ) during migration to the surface ( ''limited evidence)'' . Warming and its effects on climate cycles have the potential to alter patterns of larval transport and population connectivity through changes in circulation (Fox et al., 2016 <sup>[[#fn:r792|792]]</sup> ) or surface generated mesoscale eddies (Adams et al., 2011 <sup>[[#fn:r793|793]]</sup> ) ( ''limited evidence'' ; ''low confidence'' ). Climate-induced changes in the distribution and cover of vent and seep foundation species may involve alteration of attachment substrate, food and refuge for the many habitat-endemic species that rely on them (Cordes et al., 2010 <sup>[[#fn:r794|794]]</sup> ) and for the surrounding deep sea ecosystems which interact through transport of nutrients and microbes, movement of vagrant predators and scavengers, and plankton interactions (Levin et al., 2016 <sup>[[#fn:r795|795]]</sup> ) ( ''limited evidence'' ; ''low confidence'' ). There is, however, insufficient analysis of faunal symbiont and nutritional requirements, life histories, larval transport and cross-system interaction to quantify the extent of the consequences described above under future climate conditions.

<div id="section-5-2-4-4chemosynthetic-ecosystems-block-2" class="box"></div>

<span id="box-5.2-cold-water-corals-and-sponges"></span>