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

== 5.3 Changing Coastal Ecosystems and Biodiversity ==

<div id="article-5-3changing-coastal-ecosystems-and-biodiversity-block-1"></div>

The world’s shelf seas and coastal waters (hereafter ‘coastal seas’) extend from the coastline to the 200 m water depth contour. They encompass diverse ecosystems, including estuaries, sandy beaches, kelp forests, mangroves and coral reefs. Although they occupy a small part of the global ocean (7.6%), coastal seas provide up to 30% of global marine primary production and about 50% of the organic carbon supplied to the deep ocean (Chen, 2003; Bauer et al., 2013) (Sections 5.2.4.1 and 5.4.1.1). Coastal seas include several frontal and upwelling areas (Box 5.3) that support high fisheries yields (Scales et al., 2014), and productive coastal ecosystems, such as wetlands (McLeod et al., 2011). Mangrove forests, seagrass meadows and kelp forests form important habitats supporting high biodiversity while offering opportunities for climate change mitigation and adaptation (Section 5.5.1.2) (Duarte et al., 2013), with mangrove forests providing physical protection against extreme events such as storms and floods (Kelleway et al., 2017a) (Sections 5.4.1.2 and 4.3.3.5.4). The regional characteristics and habitat heterogeneity of many coastal seas support endemic fauna and flora (e.g., seagrass meadows in the Mediterranean), which makes them particularly vulnerable to climate change impacts with high risk of diversity loss and alterations in ecosystem structure and functioning (Rilov, 2016; Chefaoui et al., 2018).

Near-shore coastal ecosystems are classified by their geomorphological structure (e.g., estuaries, sandy beaches and rocky shores) or foundation species (e.g., salt marshes, seagrass meadows, mangrove forests, coral reefs and kelp forests). All these coastal ecosystems are threatened to a varying degree by SLR (SLR), warming, acidification, deoxygenation and extreme weather events (Sections 5.3.1 to 5.3.7). Unlike the open ocean where detection and attribution of climate driven-physical and chemical changes are robust (Section 5.2.2), coastal ecosystems display regional complexity that can render the conclusive detection and attribution of climate effects uncertain. The hydrological complexity of coastal ecosystems that affects their biota is driven by the interactions between the land (e.g., river and groundwater discharges), the sea (e.g., circulation, tides) (Section 5.2.2.2.3) and seabed structures and substrates (Sharples et al., 2017; Chen et al., 2018; Laurent et al., 2018; Zahid et al., 2018).

Additionally, the high density of human populations on coastal land causes most of the adjacent marine ecosystems to be impacted by local anthropogenic disturbances such as eutrophication, coastline modifications, pollution and overfishing (Levin et al., 2015; Diop and Scheren, 2016; Maavara et al., 2017; Dunn et al., 2018) (Section 4.3.2.2, Cross-Chapter Box 9). Climate driven impacts interact with such human disturbances and pose a serious risk to ecosystems structure and functioning (Gattuso et al., 2015). Projections of the ecological impacts of climate change in coastal ecosystems must therefore deal with many emerging complexities such as the differentiation between the long-term climate trends (e.g., progressive ocean acidification) and the short-term natural fluctuations (Boyd et al., 2018), ranging from the seasons to interannual climate oscillations like El Niño. The ‘time of emergence’ for specific climate drivers to exceed background variability varies between ecosystems and is strongly sensitive to projected emission scenarios (Hammond et al., 2017; Reusch et al., 2018) (Box 5.1).

This section summarises our updated understanding of ecological and functional changes that coastal ecosystems are experiencing due to multiple climate and non-climatic human drivers, and their synergies. Additional experimental and long-term observational evidence since AR5 WGII (Wong et al., 2014a) and SR15 (Hoegh-Guldberg et al., 2018) improves the attribution of impacts on all the types of coastal ecosystems assessed here to climate trends (Sections 5.3.1 to 5.3.6). Moreover, the emergent impacts detected in the present strengthen the projection of risk of each ecosystem under future emission scenarios by 2100, depending on their exposure to different climate hazards (Section 5.3.7).

<span id="estuaries"></span>
=== 5.3.1 Estuaries ===

<div id="section-5-3-1estuaries-block-1"></div>

Estuarine ecosystems are defined by the river-sea interface that provides high habitat heterogeneity and supports high biodiversity across freshwater and subtidal zones (Basset et al., 2013 <sup>[[#fn:r858|858]]</sup> ). AR5 WGII (Wong et al., 2014a <sup>[[#fn:r859|859]]</sup> ) and SR15 (Hoegh-Guldberg et al., 2018 <sup>[[#fn:r860|860]]</sup> ) concluded that estuarine ecosystems have been impacted by SLR and human influences that drive salinisation, resulting in increased flooding, land degradation and erosion of coastal areas around estuaries.

Observations since AR5 provide further evidence that SLR increases seawater intrusions and raises salinity in estuaries. Salinisation of estuaries can be exacerbated by droughts and modifications of drainage area by human activities (Ross et al., 2015 <sup>[[#fn:r861|861]]</sup> ; Cardoso-Mohedano et al., 2018 <sup>[[#fn:r862|862]]</sup> ; Hallett et al., 2018 <sup>[[#fn:r863|863]]</sup> ; Zahid et al., 2018 <sup>[[#fn:r864|864]]</sup> ). The changing salinity gradients in estuaries have been linked to the observed upstream expansion of brackish and marine benthic and pelagic communities, and a reduction in the diversity and richness of freshwater fauna (Robins et al., 2016 <sup>[[#fn:r865|865]]</sup> ; Raimonet and Cloern, 2017 <sup>[[#fn:r866|866]]</sup> ; Hallett et al., 2018 <sup>[[#fn:r867|867]]</sup> ; Addino et al., 2019 <sup>[[#fn:r868|868]]</sup> ) ( ''medium confidence'' ). However, because the distribution of benthic species in estuaries is strongly determined by sediment properties like grain size, the gradient of sediment types in estuaries can be a barrier to upstream shifts of brackish and marine benthic biota, leading to a reduction in species richness in mid- to upper- estuarine areas and altering food webs (Little et al., 2017 <sup>[[#fn:r869|869]]</sup> ; Hudson et al., 2018 <sup>[[#fn:r870|870]]</sup> ; Addino et al., 2019 <sup>[[#fn:r871|871]]</sup> ). Similarly, estuarine wetlands (Section 5.3.2) and tidal flats (Murray et al. 2019) have reduced their extent and productivity in response to increased salinity, inundation and wave exposure, especially in areas with limited capacity for soil accretion or inland migration due to coastal squeezing (Sections 4.3.2.3, 5.3.2) ( ''high confidence'' ). Poleward migration of tropical and sub-tropical biota between estuaries has been observed in response to warming (Hallett et al., 2018 <sup>[[#fn:r872|872]]</sup> ) ( ''medium confidence'' ), in agreement with the global trend of biogeographic shifts of marine organisms (Sections 5.2.3.1.1; 5.3.2 − 5.3.6).

Intensive human activities around estuaries and river deltas worldwide has substantially increased nutrient and organic matter inputs into such systems since the 1970s (Maavara et al., 2017 <sup>[[#fn:r873|873]]</sup> ). Increased organic matter accumulation has been shown to interact with warming, resulting in intensification of bacterial degradation and eutrophication (Maavara et al., 2017 <sup>[[#fn:r874|874]]</sup> ; Chen et al., 2018 <sup>[[#fn:r875|875]]</sup> ; Fennel and Testa, 2019 <sup>[[#fn:r876|876]]</sup> ), contributing to an increase in the frequency and extent of hypoxic zones (Breitberg et al., 2015 <sup>[[#fn:r877|877]]</sup> ; Gobler and Baumann, 2016 <sup>[[#fn:r878|878]]</sup> ). The interaction between warming, increased nutrient loading, and hypoxia has shown to be related to the increased occurrences of HABs (Anderson et al., 2015 <sup>[[#fn:r879|879]]</sup> ; Paerl et al., 2018 <sup>[[#fn:r880|880]]</sup> ) (Box 5.4) ( ''high confidence'' ), pathogenic bacteria such as ''Vibrio'' species (Baker-Austin et al., 2017 <sup>[[#fn:r881|881]]</sup> ; Kopprio et al., 2017 <sup>[[#fn:r882|882]]</sup> ) (Section 5.4.2) ( ''low confidence'' ), and mortalities of invertebrates and fish communities (Jeppesen et al., 2018 <sup>[[#fn:r883|883]]</sup> ; Warwick et al., 2018 <sup>[[#fn:r884|884]]</sup> ) ( ''medium confidence'' ).

Fluctuations in estuarine salinity, turbidity and nutrient gradients are influenced by changes in precipitation and wind-stress caused by large-scale climatic variations such as the ENSO, the NAO and the South Atlantic Meridional Overturning Circulation (SAMOC) which have shown persistent anomalies associated with climate change since the 1970s (Wang and Cai, 2013 <sup>[[#fn:r885|885]]</sup> ; Delworth and Zeng, 2016 <sup>[[#fn:r886|886]]</sup> ; García-Moreiras et al., 2018 <sup>[[#fn:r887|887]]</sup> ). Similarly, storm surges and heat waves have increased nutrients and sediment loads in estuaries (Tweedley et al., 2016 <sup>[[#fn:r888|888]]</sup> ; Arias-Ortiz et al., 2018 <sup>[[#fn:r889|889]]</sup> ; Chen et al., 2018 <sup>[[#fn:r890|890]]</sup> ). Sustained long-term observations (15 − 40 years) provide evidence that large-scale climatic variations and extreme events affect plankton phenology and composition in estuaries worldwide with regional differences in the characteristics of the responses (Thompson et al., 2015 <sup>[[#fn:r891|891]]</sup> ; Abreu et al., 2017 <sup>[[#fn:r892|892]]</sup> ; Marques et al., 2017 <sup>[[#fn:r893|893]]</sup> ; Arias-Ortiz et al., 2018 <sup>[[#fn:r894|894]]</sup> ; López-Abbate et al., 2019 <sup>[[#fn:r895|895]]</sup> ) ( ''high confidence'' ). Although these changes in ecosystem components may be attributed to climate variability (Box 5.1), they demonstrate the sensitivity of estuarine ecosystems to climate change. Also, these large-scale climate events are ''likely'' to be intensified in the 21st century (Stocker, 2014 <sup>[[#fn:r896|896]]</sup> ) (Section 6.5.1).

Salinisation in estuaries is projected to continue in response to SLR, warming and droughts under global warming greater than 1.5°C ( ''high confidence'' ), and will pose further risks to ecosystems biodiversity and functioning (Zhou et al., 2017 <sup>[[#fn:r897|897]]</sup> ; Hallett et al., 2018 <sup>[[#fn:r898|898]]</sup> ; Zahid et al., 2018 <sup>[[#fn:r899|899]]</sup> ; Elliott et al., 2019 <sup>[[#fn:r900|900]]</sup> ) (Section 4.3.3.4, Cross-Chapter Box 7) ( ''medium confidence'' ). Estuarine wetlands are resilient to modest rates of SLR due to their sediment relocation capacity, but such adaptation is not expected to keep pace with projected rates of SLR under the RCP8.5 climate scenario (Section 5.3.2) ( ''high confidence'' ). Moreover, human activities that inhibit sediment movement and deposition in coastal deltas increase the likelihood of their shrinking as a result of SLR (Brown et al., 2018b <sup>[[#fn:r901|901]]</sup> ; Schuerch et al., 2018 <sup>[[#fn:r902|902]]</sup> ) ( ''medium'' ''confidence'' ).

Oxygen-depleted dead zones in coastal areas are already a problem; they are projected to increase under the co-occurrence and intensification of climate threats and eutrophication (Breitburg et al., 2018 <sup>[[#fn:r903|903]]</sup> ; Laurent et al., 2018 <sup>[[#fn:r904|904]]</sup> ) (Section 5.2.2.4). While warming is the primary climate driver of deoxygenation in the open ocean, eutrophication is projected to increase in estuaries due to human activities and intensified precipitation increasing riverine nitrogen loads under both RCP2.6 and RCP8.5 scenarios, both mid-century (2031–2060) and later (2071–2100) (Sinha et al., 2017 <sup>[[#fn:r905|905]]</sup> ). Moreover, enhanced stratification in estuaries in response to warming is also expected to increase the risk of hypoxia through reduced vertical mixing (Du et al., 2018 <sup>[[#fn:r906|906]]</sup> ; Hallett et al., 2018 <sup>[[#fn:r907|907]]</sup> ; Warwick et al., 2018 <sup>[[#fn:r908|908]]</sup> ).The effects of warming will be more pronounced on high-latitude and temperate shallow estuaries with limited exchange with the open ocean (e.g., Río de La Plata Estuary, Baltic Sea and Chesapeake Bay) and seasonality that already leads to dead zone development when summertime temperatures reach critical values (e.g., Black Sea) (Altieri and Gedan, 2015 <sup>[[#fn:r909|909]]</sup> ) ( ''medium confidence'' ). The coastal acidification related to this expansion of hypoxic zones (Zhang and Gao, 2016 <sup>[[#fn:r910|910]]</sup> ; Cai et al., 2017 <sup>[[#fn:r911|911]]</sup> ; Laurent et al., 2017 <sup>[[#fn:r912|912]]</sup> ) imposes risk for sensitive organisms (Beck et al., 2011 <sup>[[#fn:r913|913]]</sup> ; Duarte et al., 2013 <sup>[[#fn:r914|914]]</sup> ; Feely et al., 2016 <sup>[[#fn:r915|915]]</sup> ; Carstensen et al., 2018 <sup>[[#fn:r916|916]]</sup> ).

The interaction of SLR and changes in precipitation will have a more severe impact on shallow estuaries (&lt;10 m) than on deep basin estuaries (&gt;10 m) (Hallett et al., 2018 <sup>[[#fn:r917|917]]</sup> ; Elliott et al., 2019 <sup>[[#fn:r918|918]]</sup> ) ( ''medium confidence'' ). For a projected SLR of 1 m, climate-related risks for shallow estuaries ecosystems are estimated to increase through increased tidal current amplitudes (by 5% on average), energy dissipation, vertical mixing and salinity intrusion (Prandle and Lane, 2015 <sup>[[#fn:r919|919]]</sup> ). Estuaries with high tidal exchanges and associated well-developed sediment areas are more resilient to global climate changes than estuaries with low tidal exchanges and sediment supply, since the latter are more vulnerable to SLR and changes in river flow (Brown et al., 2018b; Warwick et al., 2018 <sup>[[#fn:r920|920]]</sup> ) ( ''medium confidence'' ).

Overall, this assessment concludes that there is evidence of upstream redistribution of marine biotic communities in estuaries driven by increased sea water intrusion ( ''medium'' ''confidence'' ). Such distribution shifts are limited by physical barriers such as the availability of benthic substrates leading to reduction of suitable habitats for estuarine communities ( ''medium confidence'' ). Warming has led to poleward range shifts of biota between estuaries ( ''medium'' ''confidence'' ). Increased nutrient inputs from intensive human development in deltas increases bacterial respiration, which in turn is exacerbated by warming, leading to an expansion of suboxic and anoxic areas ( ''high confidence'' ). These changes reduce the survival of estuarine animals ( ''medium confidence'' ), and increase the occurrence of HABs and pathogenic microbes ( ''medium confidence'' ). Projected warming, SLR and tidal changes in the 21st century will continue to expand salinisation and hypoxia in estuaries ( ''medium confidence'' ). These impacts will be more pronounced under higher emission scenarios, and in temperate and high-latitude estuaries that are eutrophic, shallow and that naturally have low sediment supply.

<span id="coastal-wetlands-salt-marshes-seagrass-meadows-and-mangrove-forests"></span>
=== 5.3.2 Coastal Wetlands (Salt Marshes, Seagrass Meadows and Mangrove Forests) ===

<div id="section-5-3-2coastal-wetlands-salt-marshes-seagrass-meadows-and-mangrove-forests-block-1"></div>

Coastal vegetated wetlands include salt marshes, mangrove forests and subtidal seagrass meadows ecosystems, considered to be the main ‘blue carbon’ habitats (Sections 5.4.1 and 5.5.1.1) (McLeod et al., 2011 <sup>[[#fn:r921|921]]</sup> ). AR5 WGII and SR15 concluded that wetland salinisation is occurring at a large geographic scale ( ''high confidence'' ); that rising water temperatures has led to shifts in plant species distribution ( ''medium confidence'' )(Wong et al., 2014b <sup>[[#fn:r923|923]]</sup> ); and that SLR and storms are causing wetland erosion and habitat loss, enhanced by human disturbances ( ''high confidence'' ) (Section 4.3.3.5.1) (Wong et al., 2014b <sup>[[#fn:r922|922]]</sup> ). This section assesses new evidence since AR5 and SR15 of observed climate impacts and future risks of these vegetated wetlands in terms of their role in supporting biodiversity and key ecosystem functions. The recent literature confirms and strengthens the SR15 conclusions (Section 5.3.7 and Figure 5.16).

Nearly 50% of the pre-industrial, natural extent of global coastal wetlands have been lost since the 19th century (Li et al., 2018a <sup>[[#fn:r924|924]]</sup> ). Such a reduction in wetlands is primarily caused by non-climatic drivers such as alteration of drainage, agriculture development, coastal settlement, hydrological alterations and reductions in sediment supply (Adam, 2002 <sup>[[#fn:r925|925]]</sup> ; Wang et al., 2014 <sup>[[#fn:r926|926]]</sup> ; Kroeger et al., 2017 <sup>[[#fn:r927|927]]</sup> ; Thomas et al., 2017 <sup>[[#fn:r928|928]]</sup> ; Li et al., 2018a <sup>[[#fn:r929|929]]</sup> ). However, large-scale mortality events of mangroves from ‘natural causes’ has also occurred globally since the 1960s; ~70% of this loss has resulted from low frequency, high intensity weather events, such as tropical cyclones (45%) and climatic extremes such as droughts, SLR variations and heat waves (Sippo et al., 2018 <sup>[[#fn:r930|930]]</sup> ) ( ''high confidence'' ). In Australia, the mangrove loss due to heat waves accounted for 22% of global mangrove forests (Sippo et al., 2018 <sup>[[#fn:r931|931]]</sup> ), with negative impacts on ecosystem biodiversity and the provisioning of services (Carugati et al., 2018 <sup>[[#fn:r932|932]]</sup> ; Saintilan et al., 2018 <sup>[[#fn:r933|933]]</sup> ) (Section 5.4). In coastal areas with sufficient sediment supply across the Indo-Pacific region, inland expansion of mangroves is occurring as a result of vertical accretion and root growth, allowing them to keep pace with current SLR (Lovelock et al., 2015 <sup>[[#fn:r934|934]]</sup> ). In seagrass meadows, temperature is the main limiting range factor, and over the past decades there have been several global die-off events (Hoegh-Guldberg et al., 2018 <sup>[[#fn:r935|935]]</sup> ). The vulnerability of seagrasses to warming varies locally depending on soil accretion and herbivory (El-Hacen et al., 2018 <sup>[[#fn:r936|936]]</sup> ; Marbà et al., 2018 <sup>[[#fn:r937|937]]</sup> ; Vergés et al., 2018 <sup>[[#fn:r938|938]]</sup> ) and on the population assemblages (e.g., expansion at high latitudes) (Beca-Carretero et al., 2018 <sup>[[#fn:r939|939]]</sup> ; Duarte et al., 2018 <sup>[[#fn:r940|940]]</sup> ). The compounding effects of heat waves, hypersaline conditions and increased turbidity and nutrient levels associated with floods have been shown to cause negative changes in the composition and biomass of co-occurring seagrass species (Nowicki et al., 2017 <sup>[[#fn:r941|941]]</sup> ; Arias-Ortiz et al., 2018 <sup>[[#fn:r942|942]]</sup> ; Lin et al., 2018 <sup>[[#fn:r943|943]]</sup> ) ( ''high confidence'' ). For example, in Shark Bay, Western Australia, a marine heat wave in austral summer 2010/2011 caused widespread losses (36% of area) of seagrass meadows, with negative implications for carbon storage (Arias-Ortiz et al., 2018 <sup>[[#fn:r944|944]]</sup> ). The poleward expansion of tropical mangroves into subtropical salt marshes as a result of increase in temperature has been also observed over the past half century on five continents (Saintilan et al., 2014 <sup>[[#fn:r945|945]]</sup> ; Saintilan et al., 2018 <sup>[[#fn:r946|946]]</sup> ) ( ''high confidence'' ); for example, in the Texas Gulf Coast (Armitage et al., 2015 <sup>[[#fn:r947|947]]</sup> ). The loss of open areas with herbaceous plants (salt marshes) reduces food and habitat availability for resident and migratory animals (Kelleway et al., 2017a <sup>[[#fn:r948|948]]</sup> ; Lin et al., 2018 <sup>[[#fn:r949|949]]</sup> ) (Section 5.4.1.2).

The ability of salt marshes to increase their elevation and withstand erosion under SLR depends on the development of new soil by the external supply of mineral sediments and organic accretion by local biota (Section 5.4.1, Figure 5.19) (Bouma et al., 2016 <sup>[[#fn:r950|950]]</sup> ). In some places, critical organic accretion rates are declining due to reduced plant productivity from stress by more frequent inundation, and increased plant and microbial respiration rates as a result of warming; consequently, the elevation of marshes from soil accretion is slower than the rate of rising sea level, resulting in reduction of salt marsh area (Carey et al., 2017 <sup>[[#fn:r951|951]]</sup> ; Watson et al., 2017b <sup>[[#fn:r952|952]]</sup> ). Vegetation loss rates were significantly negatively correlated with marsh elevation, suggesting inundation due to SLR since 1970 as the main driver, enhanced by storms and increased tidal range in back barrier marshes (Watson et al., 2017b <sup>[[#fn:r953|953]]</sup> ). Plant species that are more sensitive to higher temperatures and increases in saltwater intrusion were found to be less abundant and in some cases replaced by salinity-tolerant species (Janousek et al., 2017 <sup>[[#fn:r954|954]]</sup> ; Piovan et al., 2019 <sup>[[#fn:r955|955]]</sup> ). Plant community restructuring has resulted in biodiversity loss (Pratolongo et al., 2013 <sup>[[#fn:r956|956]]</sup> ; Raposa et al., 2017 <sup>[[#fn:r957|957]]</sup> ) and reduced above- and below-ground productivity (McLeod et al., 2011 <sup>[[#fn:r958|958]]</sup> ; Watson et al., 2017b <sup>[[#fn:r959|959]]</sup> ). As a result of tidal flooding, salt marsh soils do not dry out and high levels of carbon can accumulate under anaerobic conditions. This is coupled with generally low rates of methane emission which is strongly limited in saline marshes (Poffenbarger et al., 2011 <sup>[[#fn:r960|960]]</sup> ; Martin and Moseman-Valtierra, 2015 <sup>[[#fn:r961|961]]</sup> ; Kroeger et al., 2017 <sup>[[#fn:r962|962]]</sup> ; Tong et al., 2018 <sup>[[#fn:r963|963]]</sup> ) ( ''high confidence'' ).

Non-climatic human pressures on wetland ecosystems, including overfishing (Crotty et al., 2017 <sup>[[#fn:r964|964]]</sup> ), eutrophication (Legault II et al., 2018 <sup>[[#fn:r965|965]]</sup> ), and invasive species (Zhang et al., 2016 <sup>[[#fn:r966|966]]</sup> ), interact with climate change drivers and affect wetlands composition and structure, with the impacts varying between regions and species (Tomas et al., 2015 <sup>[[#fn:r967|967]]</sup> ; O’Brien et al., 2017; Pagès et al., 2017 <sup>[[#fn:r968|968]]</sup> ; York et al., 2017 <sup>[[#fn:r969|969]]</sup> ). The intensity of herbivory on seagrasses is expected to increase with global warming, particularly in temperate areas, because of the migration of tropical herbivores into temperate seagrass meadows (Hyndes et al., 2016 <sup>[[#fn:r970|970]]</sup> ; Vergés et al., 2018 <sup>[[#fn:r971|971]]</sup> ) ( ''medium confidence'' , Section 5.2.3.1.1). Warming also reduces the fitness of seedlings by increasing necrosis and susceptibility to consumers and pathogenic pressure while reducing establishment potential and nutritional (Olsen et al., 2016b <sup>[[#fn:r972|972]]</sup> ; Hernán et al., 2017 <sup>[[#fn:r973|973]]</sup> ). Because herbivores play a key role in modulating the biomass of plant communities, their more intense activity affects the provision of services in these ecosystems (Scott et al., 2018 <sup>[[#fn:r974|974]]</sup> ) (Section 5.4).

Globally, between 20−90% of existing coastal wetland area is projected to be lost by 2100 (Blankespoor et al., 2014 <sup>[[#fn:r975|975]]</sup> ; Crosby et al., 2016 <sup>[[#fn:r976|976]]</sup> ; Spencer et al., 2016 <sup>[[#fn:r977|977]]</sup> ), depending on different SLR projections under future emission scenarios. These projected changes vary regionally and between different types of wetlands. Gaining area may be possible, at least locally, if vertical sediment accretion occurs together with lateral re-accommodation (Brown et al., 2018b <sup>[[#fn:r978|978]]</sup> ; Schuerch et al., 2018 <sup>[[#fn:r979|979]]</sup> ) (Section 4.3.3.5.1). Local losses may also be higher; for example, in New England, where regional rates of SLR have been as much as 50% greater than the global average (from 1−5.83 mm yr -1 ; 1979-2015) (Watson et al., 2017a <sup>[[#fn:r980|980]]</sup> ) and where projections suggest that 40–95% of salt marshes will be submerged by the end of this century (Valiela et al., 2018 <sup>[[#fn:r981|981]]</sup> ). In some species of seagrasses, enhanced temperature-driven flowering (Ruiz-Frau et al., 2017 <sup>[[#fn:r982|982]]</sup> ) and greater biomass production in response to elevated CO 2 (Campbell and Fourqurean, 2018 <sup>[[#fn:r983|983]]</sup> ) may increase resilience to warming. Nevertheless, severe habitat loss (70%) of endemic species such as ''Posidonia oceanica'' is projected by 2050 with the potential for functional extinction by 2100 under RCP8.5 climate scenario. For ''Cymodosea nodosa'' , the species with the highest thermal optima (Savva et al., 2018 <sup>[[#fn:r984|984]]</sup> ), warming is expected to lead to significant reduction of meadows (46% under RCP8.5) in the Mediterranean, although potentially compensated in part by future expansion into the Atlantic (Chefaoui et al., 2018 <sup>[[#fn:r985|985]]</sup> ).

The mangrove habitats of small islands, with lack of rivers, steep topography, sediment-starved areas, groundwater extraction and coastal development, are particularly vulnerable to SLR. Although mangrove ecosystems may survive the increased storm intensity and sea levels projected until 2100 under RCP2.6 (Ward et al., 2016 <sup>[[#fn:r986|986]]</sup> ), for RCP8.5 they are only resilient up to 2050 conditions (Sasmito et al., 2016 <sup>[[#fn:r987|987]]</sup> ). Negative climate impacts will be exacerbated in cases where anthropogenic barriers cause further ‘coastal squeeze’ that prevents inland movement of plants and limits relocation of sediment ( ''medium confidence'' ) (Enwright et al., 2016 <sup>[[#fn:r988|988]]</sup> ; Borchert et al., 2018 <sup>[[#fn:r989|989]]</sup> ).

In conclusion, substantial evidence supports with ''high confidence'' that warming and salinisation of wetlands caused by SLR are causing shifts in the distribution of plant species inland and poleward, such as mangrove encroachment into subtropical salt marshes ( ''high confidence'' ) or seagrass meadows contraction at low latitudes ( ''high confidence'' ). Plants with low tolerance to flooding and extreme temperatures are particularly vulnerable and may be locally extirpated ( ''medium confidence'' ). The flooded area of salt marshes can become a mudflat or be colonised by more tolerant, invasive species, whose expansion is favoured by combined effects of warming, rising CO 2 and nutrient enrichment ( ''medium confidence'' ). The loss of vegetated coastal ecosystems causes a reduction in carbon storage with positive feedbacks to the climate system ( ''high confidence'' ) (Section 5.4.1.2). SLR and warming are expected to continue to reduce the area of coastal wetlands, with a projected global loss of 20–90% by the end of the century depending on emission scenarios. High risk of total local loss is projected under the RCP8.5 emission scenario by 2100 ( ''medium confidence'' ), especially if landward migration and sediment supply is constrained by human modification of shorelines and river flows ( ''medium confidence'' ).

<span id="sandy-beaches"></span>
=== 5.3.3 Sandy Beaches ===

<div id="section-5-3-3sandy-beaches-block-1"></div>

Sandy beaches represent 31% of the world’s ice-free shoreline (Luijendijk et al., 2018 <sup>[[#fn:r990|990]]</sup> ). They provide habitat for dune vegetation, benthic fauna and sea birds, nesting areas for marine turtles (Defeo et al., 2009 <sup>[[#fn:r991|991]]</sup> ), and several key ecosystem services (Drius et al., 2019 <sup>[[#fn:r992|992]]</sup> ) (Section 5.4.1.2). Sandy beach ecosystems are physically dynamic, where sediment movement is a key driver of benthic flora and fauna zonation (Schlacher and Thompson, 2013 <sup>[[#fn:r993|993]]</sup> ; van Puijenbroek et al., 2017 <sup>[[#fn:r994|994]]</sup> ). In AR5 WGII (Wong et al., 2014b <sup>[[#fn:r995|995]]</sup> ) and SR15 (Hoegh-Guldberg et al., 2018 <sup>[[#fn:r996|996]]</sup> ), climate impacts on sandy beach ecosystems were not assessed individually but together with other coastal systems that included beaches, barriers, sand dunes, rocky coasts, aquifers and lagoons. Those assessments concluded with ''high confidence'' that SLR, storminess, wave energy and weathering regimes will continue to erode coastal shorelines and affect the soil accretion and land-based ecosystems, with highly site-specific effects ( ''high confidence'' ). Infrastructure and geological constraints reduce shoreline movement and cause coastal squeeze ( ''high confidence'' ). Assessment in Section 4.3.3.3 supports the conclusions in AR5 and SR15 regarding the erosion of sandy coastlines. This section specifically assesses the combined climate and non-climatic impacts on sandy beach biodiversity, ecosystem structure and functioning.

Worldwide, sandy beaches show vegetation transformations caused by erosion following locally severe wave events (Castelle et al., 2017 <sup>[[#fn:r997|997]]</sup> ; Delgado-Fernandez et al., 2019 <sup>[[#fn:r998|998]]</sup> ; Zinnert et al., 2019 <sup>[[#fn:r999|999]]</sup> ) (Table SM5.7). The original dense vegetation is replaced by sparser vegetation (Zinnert et al., 2019 <sup>[[#fn:r1000|1000]]</sup> ) and has a generally slow recovery (multiple years to decades) (Castelle et al., 2017 <sup>[[#fn:r1001|1001]]</sup> ). In some instances, the changes persist over decades, resulting in a regime shift in the beach morphology (Kuriyama and Yanagishima, 2018 <sup>[[#fn:r1002|1002]]</sup> ). Such changes in vegetation and beach morphology in response to local disturbances were also related to shifts in the associated fauna composition (Carcedo et al., 2017 <sup>[[#fn:r1003|1003]]</sup> ; Delgado-Fernandez et al., 2019 <sup>[[#fn:r1004|1004]]</sup> ). Direct attribution of these observed events to climate change is not available despite early evidence (since the 1970s) and an emerging literature (Section 4.3.3.1, Table SM5.7).

Sandy beaches show similar patterns of biogeographical shifts following warming, with increased dominance of species more tolerant to higher temperatures, as observed in other ocean ecosystems (Section 5.2.3.1.1, Table SM5.7). Examples of these observed shifts in abundance and distribution of benthic fauna in sandy beaches are found in the Pacific and Atlantic coasts of North and South America, and in Australia, including increased mortality of clam populations close to their upper temperature limits with low population recovery (Orlando et al., 2019 <sup>[[#fn:r1005|1005]]</sup> ), and poleward expansion of crabs since the 1980s that were related to warming (Schoeman et al., 2015 <sup>[[#fn:r1006|1006]]</sup> ) (Table SM5.7). Also, mass mortalities of beach clams have occurred during warm phases of El Niño events (Orlando et al., 2019 <sup>[[#fn:r1007|1007]]</sup> )(Table SM5.7), parasite infestations on dense populations (Vázquez et al., 2016 <sup>[[#fn:r1008|1008]]</sup> ) and high wave exposure (Turra et al., 2016 <sup>[[#fn:r1009|1009]]</sup> ).

Human disturbances have caused coastal squeeze and morphological changes in sandy beaches (Martínez et al., 2017 <sup>[[#fn:r1010|1010]]</sup> ; Rêgo et al., 2018 <sup>[[#fn:r1011|1011]]</sup> ; Delgado-Fernandez et al., 2019 <sup>[[#fn:r1012|1012]]</sup> ). Along with SLR and climate-driven intensification of waves and offshore winds, these hazards have increased erosion rates suggesting a reduced resilience due to insufficient sediment supply and accretion capacity (Castelle et al., 2017 <sup>[[#fn:r1013|1013]]</sup> ; Houser et al., 2018 <sup>[[#fn:r1014|1014]]</sup> ; Kuriyama and Yanagishima, 2018 <sup>[[#fn:r1015|1015]]</sup> ). Narrow sandy beaches such as those in south California (Vitousek et al., 2017 <sup>[[#fn:r1016|1016]]</sup> ) or central Chile (Martínez et al., 2017 <sup>[[#fn:r1017|1017]]</sup> ) are particularly vulnerable to climate hazards when combined with human disturbances and where landward retreat of beach profile and benthic organisms is constrained due to increasing urbanisation (Hubbard et al., 2014 <sup>[[#fn:r1018|1018]]</sup> ) (Section 4.3.2.3).

Notwithstanding the uncertainty in projecting future interactions of SLR with other natural and human impacts on sandy shorelines (Le Cozannet et al., 2019; Orlando et al., 2019 <sup>[[#fn:r1019|1019]]</sup> ), they are expected to continue to reduce their area and change their topography due to SLR and increased extreme climatic erosive events. This will be especially important in low-lying coastal areas with high population and building densities ( ''medium confidence'' , SM 4.2). Megafauna that use sandy beaches during vulnerable parts of their life cycles could be particularly impacted (Laloë et al., 2017 <sup>[[#fn:r1020|1020]]</sup> ). For example, the modelled incubation temperatures of green turtles have increased by 1°C since the mid-1970s, resulting in an average 20% increase in the proportion of female hatchlings over this period (Patrício et al., 2019 <sup>[[#fn:r1021|1021]]</sup> ). By 2100, global temperatures will approach lethal levels for incubation in existing nesting sites, and hatchling success is expected to drop to 32% under RCP8.5 scenario, with 93% of the hatchlings expected to be female (76% under RCP4.5). A possible microhabitat adaptation such as shadowed vegetated areas, however, could allow for continued male production throughout the 21st century (Patrício et al., 2019 <sup>[[#fn:r1022|1022]]</sup> ). In addition, a projected global mean SLR of ~1.2 m under the upper likely range of RCP8.5 by 2100 implies a loss of 59% and 67% in the present nesting area of the green turtle and the loggerhead respectively in the Mediterranean (Varela et al., 2019 <sup>[[#fn:r1023|1023]]</sup> ), and a loss of 43% in the nesting area of green turtles in West Africa (Patrício et al., 2019 <sup>[[#fn:r1024|1024]]</sup> ). Moreover, benthic crustaceans of sandy beaches, including isopods, crabs and amphipods, generally follow the temperature-body size gradient in which body size decreases towards warmer lower-latitude regions (Jaramillo et al., 2017 <sup>[[#fn:r1025|1025]]</sup> ). Assuming that the physiological underpinning of the relationship between body size and temperature can be applied to warming (see Section 5.2.2, ''medium confidence'' ), the body size of sandy beach crustaceans is expected to decrease under warming ( ''low evidence, medium agreement'' ).

Overall, changes in sandy beach morphology have been observed from climate related events, such as storm surges, intensified offshore winds, and from coastal degradation caused by humans ( ''high confidence'' ), with impacts on beach habitats (e.g., benthic megafauna) ( ''medium confidence'' ). The direct influence of contemporary SLR on shoreline behaviour is emerging, but attribution of such changes to SLR remains difficult (Section 4.3.3.1). Projected changes in mean and extreme sea levels (Section 4.2.3) and warming (Section 5.2.1) under RCP8.5 are expected to result in high risk of impacts on sandy beach ecosystems by the end of the 21st century ( ''medium confidence'' , Figure 5.16), taking account of the slow recovery rate of sandy beach vegetation, the direct loss of habitats and the high climatic sensitivity of some fauna. Under RCP2.6, the risk of impacts on sandy beaches is expected to be only slightly higher than the present day level ( ''low confidence'' , Figure 5.16). However, pervasive coastal urbanisation lowers the buffering capacity and recovery potential of sandy beach ecosystems to impacts from SLR and warming and thus is expected to limit their resilience to climate change ( ''high confidence'' ).

<span id="coral-reefs"></span>
=== 5.3.4 Coral Reefs ===

<div id="section-5-3-4coral-reefs-block-1"></div>

Human activities and warming have already led to major impacts on shallow water tropical coral reefs caused by species replacement, bleaching and decreased coral cover while warming, ocean acidification and climate hazards will put warm water corals at very high risk even if global warming can be limited to 1.5°C above pre-industrial level (Hoegh-Guldberg et al., 2018 <sup>[[#fn:r1026|1026]]</sup> ; Kubicek et al., 2019 <sup>[[#fn:r1027|1027]]</sup> ; Sully et al., 2019 <sup>[[#fn:r1028|1028]]</sup> ). While providing new evidence to support these previous assessments (Kleypas, 2019 <sup>[[#fn:r1029|1029]]</sup> ), this assessment focuses on evaluating the variations in sensitivities and responses of coral reefs and their associated biota to highlight comparative risks and resiliences.

New evidence since AR5 and SR15 confirms the impacts of ocean warming (Kao et al., 2018 <sup>[[#fn:r1030|1030]]</sup> ; Jury and Toonen, 2019 <sup>[[#fn:r1031|1031]]</sup> ) and acidification (Jiang et al., 2018 <sup>[[#fn:r1032|1032]]</sup> ; Mollica et al., 2018 <sup>[[#fn:r1033|1033]]</sup> ; Bove et al., 2019 <sup>[[#fn:r1034|1034]]</sup> ) on coral reefs ( ''high confidence'' ), enhancing reef dissolution and bioerosion ( ''high confidence'' ), affecting coral species distribution, and leading to community changes (Agostini et al., 2018 <sup>[[#fn:r1035|1035]]</sup> ) ( ''high confidence'' ). The rate of SLR (primarily noticed in small reef islands) may outpace the growth of reefs to keep up although there is ''low agreement'' in the literature (Brown et al., 2011 <sup>[[#fn:r1036|1036]]</sup> ; Perry et al., 2018 <sup>[[#fn:r1037|1037]]</sup> ) ( ''low confidence'' ). Reefs are further exposed to other increased impacts, such as enhanced storm intensity (Lavender et al., 2018 <sup>[[#fn:r1038|1038]]</sup> ), turbidity and increased runoff from the land (Kleypas, 2019 <sup>[[#fn:r1039|1039]]</sup> ) ( ''high confidence'' ). Recovery of coral reefs resulting from repeated disturbance events is slow (Hughes et al., 2019a <sup>[[#fn:r1040|1040]]</sup> ; Ingeman et al., 2019 <sup>[[#fn:r1041|1041]]</sup> ) ( ''high confidence'' ). Only few coral reef areas show some resilience to global change drivers (Fine et al., 2019 <sup>[[#fn:r1042|1042]]</sup> ) ( ''low confidence'' ).

Globally, coral reefs and their associated communities are projected to change their species composition and biodiversity as a result of future interactions of multiple climatic and non-climatic hazards (Kleypas, 2019 <sup>[[#fn:r1043|1043]]</sup> ; Kubicek et al., 2019 <sup>[[#fn:r1044|1044]]</sup> ; Rinkevich, 2019 <sup>[[#fn:r1045|1045]]</sup> ) ( ''high evidence, very high agreement, very high confidence'' ). Multiple stressors act together to increase the risk of population declines or local extinction of reef-associated species through impacts of warming and ocean acidification on physiology and behaviours (Gunderson et al., 2017 <sup>[[#fn:r1046|1046]]</sup> ) ( ''high confidence'' ). Alteration of composition of coral reef-associated biota is exacerbated by changes in habitat conditions through increased sedimentation and nutrient concentrations from human coastal activities (Fabricius, 2005 <sup>[[#fn:r1047|1047]]</sup> ) ( ''high confidence'' ). Coral ecosystems in tropical small islands are also at high risk of being affected by extreme events, including storms, with their impacts exacerbated by SLR (Duvat et al., 2017 <sup>[[#fn:r1048|1048]]</sup> ; Harborne et al., 2017 <sup>[[#fn:r1049|1049]]</sup> ) ( ''high confidence'' ). Such risks on coral reef associated communities are substantially elevated when the level of these climatic and non-climatic hazards are above thresholds that may cause phase shifts in reef communities (McCook, 1999 <sup>[[#fn:r1050|1050]]</sup> ; Hughes et al., 2010 <sup>[[#fn:r1051|1051]]</sup> ; Graham et al., 2013 <sup>[[#fn:r1052|1052]]</sup> ; Hughes et al., 2018 <sup>[[#fn:r1053|1053]]</sup> ) ( ''high confidence'' ). A phase shift is characterised by an abrupt decrease in coral abundance or cover, with concurrent increase in the dominance of non-reef building organisms, such as algae and soft corals (Kleypas, 2019 <sup>[[#fn:r1054|1054]]</sup> ). Such phase shifts have already been observed in many coral reefs worldwide (Wernberg et al., 2016 <sup>[[#fn:r1055|1055]]</sup> ; Kleypas, 2019 <sup>[[#fn:r1056|1056]]</sup> ).

Notwithstanding the conclusion that coral reefs globally are projected to greatly decline at 2°C warming relative to pre-industrial level (Cacciapaglia and van Woesik, 2018 <sup>[[#fn:r1057|1057]]</sup> ; Dietz et al., 2018 <sup>[[#fn:r1058|1058]]</sup> ; Hoegh-Guldberg et al., 2018 <sup>[[#fn:r1059|1059]]</sup> ), climate impacts can be affected by variations in the sensitivity and adaptive capacity across coral species and coral reef ecosystems. Laboratory experiments show that some warm water corals possess the cellular, physiological or molecular machineries that could help them acclimatise or adapt to the effects of global change ( ''medium confidence'' ) (DeBiasse and Kelly, 2016 <sup>[[#fn:r1060|1060]]</sup> ; Gibbin et al., 2017 <sup>[[#fn:r1061|1061]]</sup> ; Wall et al., 2017 <sup>[[#fn:r1062|1062]]</sup> ; Camp et al., 2018 <sup>[[#fn:r1063|1063]]</sup> ; Donelson et al., 2018 <sup>[[#fn:r1064|1064]]</sup> ; Drake et al., 2018 <sup>[[#fn:r1065|1065]]</sup> ; Veilleux and Donelson, 2018 <sup>[[#fn:r1066|1066]]</sup> ; Hughes et al., 2019b <sup>[[#fn:r1067|1067]]</sup> ). For example, there are species or genotypes that show less impacts by either ocean acidification or increased temperatures (Cornwall et al., 2018 <sup>[[#fn:r1068|1068]]</sup> ; Gintert et al., 2018 <sup>[[#fn:r1069|1069]]</sup> ). Some corals and their symbionts might be able to use epigenetic (heritable [https://en.wikipedia.org/wiki/Phenotype phenotype] changes that do not involve alterations in the [https://en.wikipedia.org/wiki/DNA_sequence DNA sequence] s) mechanisms to reduce their sensitivity to temperature changes in their environment and to pass such traits to their offspring (Liew et al., 2017 <sup>[[#fn:r1070|1070]]</sup> ; Torda et al., 2017 <sup>[[#fn:r1071|1071]]</sup> ; Li et al., 2018b <sup>[[#fn:r1072|1072]]</sup> ; Liew et al., 2018 <sup>[[#fn:r1073|1073]]</sup> ). The variations in sensitivity and adaptive capacity of coral species to warming and ocean acidification contribute to changes in species composition of coral reefs as they are exposed to climatic and non-climatic hazards (Ingeman et al., 2019 <sup>[[#fn:r1074|1074]]</sup> ; Kleypas, 2019 <sup>[[#fn:r1075|1075]]</sup> ; Kubicek et al., 2019 <sup>[[#fn:r1076|1076]]</sup> ) ( ''high confidence'' ). However, it has not yet been established whether coral and coral associated biota adaptation may hold beyond 1.5°C warming. The onset of coral bleaching in the last decade has occurred at higher SSTs ( ∼ 0.5°C) than in the previous decade, suggesting that coral populations that remain after preceding bleaching events may have a higher thermal threshold (Sully et al., 2019 <sup>[[#fn:r1077|1077]]</sup> ) ( ''medium confidence'' ), potentially as a result of the increased dominance of species with lower sensitivity or higher adaptive capacity (Schulz et al., 2013 <sup>[[#fn:r1078|1078]]</sup> ; McClanahan et al., 2014 <sup>[[#fn:r1079|1079]]</sup> ; Mumby and van Woesik, 2014 <sup>[[#fn:r1080|1080]]</sup> ; Pandolfi, 2015 <sup>[[#fn:r1081|1081]]</sup> ; Folkersen, 2018 <sup>[[#fn:r1082|1082]]</sup> ) ( ''medium confidence'' ).

Coral reefs in deeper or mesophotic waters (found in tropical/subtropical regions at 30–150 m depth) may serve as refuges and sources for larval supply to those reefs exposed to disturbances (e.g., bleaching, storms, floods from land, sedimentation and tourism impacts) (Bridge et al., 2013 <sup>[[#fn:r1083|1083]]</sup> ; Thomas et al., 2015 <sup>[[#fn:r1084|1084]]</sup> ; Lindfield et al., 2016 <sup>[[#fn:r1085|1085]]</sup> ; Smith et al., 2016b <sup>[[#fn:r1086|1086]]</sup> ; Bongaerts et al., 2017 <sup>[[#fn:r1087|1087]]</sup> ). Reefs exposed to local oceanographic characteristics that reduce warming, such as upwelling, may similarly provide refuges and larval sources (Tkachenko and Soong, 2017 <sup>[[#fn:r1088|1088]]</sup> ). However, recent evidence suggests that mesophotic coral reefs are at higher risk than

previously indicated (Rocha et al., 2018 <sup>[[#fn:r1089|1089]]</sup> ). Monitoring of coral reefs worldwide shows that some areas in the eastern tropical Pacific Ocean (Smith et al., 2017 <sup>[[#fn:r1090|1090]]</sup> ), the Caribbean (Chollett and Mumby, 2013 <sup>[[#fn:r1091|1091]]</sup> ), the Red Sea (Fine et al., 2013 <sup>[[#fn:r1092|1092]]</sup> ; Osman et al., 2017 <sup>[[#fn:r1093|1093]]</sup> ), the Persian Gulf (Coles and Riegl, 2013 <sup>[[#fn:r1094|1094]]</sup> ) and the Great Barrier Reef, Australia (Hughes et al., 2010 <sup>[[#fn:r1095|1095]]</sup> ; Morgan et al., 2017 <sup>[[#fn:r1096|1096]]</sup> ) have recovered more rapidly after bleaching than the larger-scale average ( ''medium confidence'' ). There are regional differences in reef vulnerability when considering scales larger than 100 km or over latitudinal gradients (van Hooidonk et al., 2013 <sup>[[#fn:r1097|1097]]</sup> ; Heron et al., 2016 <sup>[[#fn:r1097|1097]]</sup> ; Langlais et al., 2017 <sup>[[#fn:r1099|1099]]</sup> ; McClenachan et al., 2017 <sup>[[#fn:r1100|1100]]</sup> ) ( ''high confidence'' ).

Based on findings from simulation modelling, SR15 concluded that “coral reefs are projected to decline by a further 70–90% at 1.5°C (very ''high confidence'' ) with larger losses (&gt;99%) at 2°C ( ''very high confidence'' )”. The variations in exposure, sensitivity and adaptive capacity between coral populations and regions are further projected to cause large changes in the composition and structure of the remaining coral reefs, with large regional differences (van Hooidonk et al., 2016 <sup>[[#fn:r1101|1101]]</sup> ; Hoegh-Guldberg et al., 2018 <sup>[[#fn:r1102|1102]]</sup> ; Kleypas, 2019 <sup>[[#fn:r1103|1103]]</sup> ; Kubicek et al., 2019 <sup>[[#fn:r1104|1104]]</sup> ; Sully et al., 2019 <sup>[[#fn:r1105|1105]]</sup> ).

<span id="rocky-shores"></span>
=== 5.3.5 Rocky Shores ===

<div id="section-5-3-5rocky-shores-block-1"></div>

Rocky shore ecosystems span the intertidal and shallow subtidal zones of the world’s temperate coasts and are typically dominated by calcareous mussels or seaweeds (macroalgae). Other organisms that inhabit rocky shores are coralline algae (i.e., maerl beds), polychaetes, molluscs, bryozoans and sponges. Intertidal habitats are characterised by strong environmental gradients, and are exposed to marine and atmospheric climate regimes (Hawkins et al., 2016 <sup>[[#fn:r1106|1106]]</sup> ). IPCC AR5 (Wong et al., 2014a <sup>[[#fn:r1107|1107]]</sup> ) concluded that rocky shores are among the better-understood coastal ecosystems in terms of potential impacts of climate variability and change. The high sensitivity of sessile organisms (e.g., barnacles, mussels) to extreme temperature events (e.g., mass mortality and drastic biodiversity loss of mussels beds), and to acidification (widely observed in manipulative experiments) gives ''high confidence'' that rocky shore species are at high risk of changes in distribution and abundance from these two drivers. SR15 (Hoegh-Guldberg et al., 2018 <sup>[[#fn:r1108|1108]]</sup> ) concluded that rocky coasts are already experiencing large-scale changes, and critical thresholds are expected to be reached at warming of 1.5°C and above ( ''high confidence'' ).

More observational and empirical evidence since AR5 and SR15 confirms that climate change poses high risk to rocky shore ecosystems’ biodiversity, structure and functioning through warming, acidification, SLR and extreme events (Agostini et al., 2018 <sup>[[#fn:r1109|1109]]</sup> ; Duarte and Krause-Jensen, 2018 <sup>[[#fn:r1110|1110]]</sup> ; Ullah et al., 2018 <sup>[[#fn:r1111|1111]]</sup> ; Milazzo et al., 2019 <sup>[[#fn:r1112|1112]]</sup> ). Immobile intertidal organisms are especially vulnerable to warming, due to the potential for extreme heat exposure during low tide emersion and prolonged desiccation events (Hawkins et al., 2016 <sup>[[#fn:r1113|1113]]</sup> ; Zamir et al., 2018 <sup>[[#fn:r1114|1114]]</sup> ) ( ''high confidence'' ). This effect is expected to lower the upper vertical limit of intertidal communities (Hawkins et al., 2016 <sup>[[#fn:r1115|1115]]</sup> ), reducing their suitable habitat (Harley, 2011 <sup>[[#fn:r1116|1116]]</sup> ), and accompanied by temperature-induced increases in predation by consumers (Sanford, 1999 <sup>[[#fn:r1117|1117]]</sup> ). While previous studies have documented a poleward shift in species distributions of rocky intertidal and reef algae (Duarte et al., 2013 <sup>[[#fn:r1118|1118]]</sup> ; Nicastro et al., 2013 <sup>[[#fn:r1119|1119]]</sup> ) and faunal species (Barry et al., 1995 <sup>[[#fn:r1120|1120]]</sup> ; Mieszkowska et al., 2006 <sup>[[#fn:r1121|1121]]</sup> ; Lima et al., 2007 <sup>[[#fn:r1122|1122]]</sup> ), local extinctions at the equatorial or warm edge of species ranges are increasingly being attributed to climate change (Yeruham et al., 2015 <sup>[[#fn:r1123|1123]]</sup> ; Sorte et al., 2017 <sup>[[#fn:r1124|1124]]</sup> ) ( ''high confidence'' ). Extreme heat waves are expected to cause mortality among rocky shore species (Gazeau et al., 2014 <sup>[[#fn:r1125|1125]]</sup> ; Jurgens et al., 2015 <sup>[[#fn:r1126|1126]]</sup> ) and subsequent declines or losses in important species can have cascading effects on the whole intertidal community and the services it provides (Gatti et al., 2017 <sup>[[#fn:r1127|1127]]</sup> ; Sorte et al., 2017 <sup>[[#fn:r1128|1128]]</sup> ; Sunday et al., 2017 <sup>[[#fn:r1129|1129]]</sup> ). Coralline fauna adapted to narrow environmental conditions seem especially vulnerable to heat waves, with observed mass mortalities in the Adriatic Sea in response to extreme summer temperatures (Kružić et al., 2016 <sup>[[#fn:r1130|1130]]</sup> ). The loss of thermal refugia associated with continued warming could exacerbate the impacts of heat stress on rocky intertidal communities (Lima et al., 2016 <sup>[[#fn:r1131|1131]]</sup> ). Nevertheless, experimental data indicate that some coralline algae that are well adapted to highly variable transitional environments can tolerate the warming projected for 2100 under RCP8.5; for these species, ocean acidification will constitute the main hazard (Nannini et al., 2015 <sup>[[#fn:r1132|1132]]</sup> ).

Ocean acidification is expected to decrease the net calcification ( ''high confidence'' ) and abundance ( ''medium confidence'' ) of rocky intertidal and reef-associated species (Kroeker et al., 2013 <sup>[[#fn:r1133|1133]]</sup> ), and the dissolution of calcareous species has already been documented in tide-pool communities (Kwiatkowski et al., 2016 <sup>[[#fn:r1134|1134]]</sup> ; Duarte and Krause-Jensen, 2018 <sup>[[#fn:r1135|1135]]</sup> ). Recent experimental and field studies, however, have demonstrated the importance of food resources in mediating the effects of ocean acidification on vulnerable rocky shores species (Ciais et al., 2013 <sup>[[#fn:r1136|1136]]</sup> ; Ramajo et al., 2016 <sup>[[#fn:r1137|1137]]</sup> ), suggesting that species’ vulnerability to ocean acidification may be most pronounced in areas of high heat stress and low food availability ( ''medium confidence'' ) (Kroeker et al., 2017 <sup>[[#fn:r1138|1138]]</sup> ). There is increasing evidence that the interactions between multiple climate drivers will determine species vulnerability and the ecosystem impacts of climate change (Hewitt et al., 2016 <sup>[[#fn:r1139|1139]]</sup> ).

Studies on naturally acidified rocky reef ecosystems suggest ocean acidification will simplify rocky shore ecosystems, due to an overgrowth by macroalgae, a reduction in biodiversity and a reduction in the abundance of calcareous species ( ''medium confidence'' ) (Kroeker et al., 2013 <sup>[[#fn:r1140|1140]]</sup> ; Linares et al., 2015 <sup>[[#fn:r1141|1141]]</sup> ). These shifts in community structure and function have been observed in CO 2 seep communities (Hall-Spencer et al., 2008 <sup>[[#fn:r1142|1142]]</sup> ), already exposed to levels of pCO 2 expected to generally occur by the end of the century (Agostini et al., 2018 <sup>[[#fn:r1143|1143]]</sup> ). Reductions in the abundance of calcareous herbivores that can create space for rarer species by grazing the dominant algae, are expected to contribute to the overgrowth of fleshy macroalgae on rocky shores (Baggini et al., 2015 <sup>[[#fn:r1144|1144]]</sup> ). This shift towards macroalgae is associated with a simplification of the food web at lower trophic levels (Kroeker et al., 2011 <sup>[[#fn:r1145|1145]]</sup> ).

At the local scale, warming and ocean acidification are expected to change energy flows within rocky shores ecosystems ( ''medium confidence'' ). Experiments indicate that both climate drivers may boost primary productivity in some cases (Goldenberg et al., 2017 <sup>[[#fn:r1146|1146]]</sup> ); however, increased metabolic demands and greater consumption by predators under warmer temperature increase the strength of top-down control (predation mortalities of herbivores) and thus counteracts the effects of increased bottom-up productivity (Goldenberg et al., 2017 <sup>[[#fn:r1147|1147]]</sup> ; Kordas et al., 2017 <sup>[[#fn:r1148|1148]]</sup> ). Ocean acidification could also increase species energetic costs and the grazing rate of herbivores, affecting ecosystem responses to increased primary productivity (Ghedini et al., 2015 <sup>[[#fn:r1149|1149]]</sup> ). Although these increasingly complex experiments have highlighted the potential for species interactions to mediate the effects of climate change, our understanding of the effects on intact, functioning ecosystems is limited. Despite predictions for increased production and herbivory with warming and acidification, an experimental study of a more complex food web revealed an overall reduction in the energy flow to higher trophic levels and a shift towards detritus-based food webs (Ullah et al., 2018 <sup>[[#fn:r1150|1150]]</sup> ).

Overall, intertidal rocky shores ecosystems are highly sensitive to ocean warming, acidification and extreme heat exposure during low tide emersion ( ''high confidence'' ). More field and experimental evidence shows that these ecosystems are at a moderate risk at present and this level is expected to rise to very high under the RCP8.5 scenario by the end of the century (see Section 5.3.7). Benthic species will continue to relocate in the intertidal zones and experience mass mortality events due to warming ( ''high confidence'' ). Interactive effects between acidification and warming will exacerbate the negative impacts on rocky shore communities, causing a shift towards a less diverse ecosystem in terms of species richness and complexity, increasingly dominated by macroalgae ( ''high confidence'' ).

<span id="kelp-forests"></span>
=== 5.3.6 Kelp Forests ===

<div id="section-5-3-6kelp-forests-block-1"></div>

Kelp forests are three-dimensional, highly productive coastal ecosystems with a reported global NPP between 1.02‒1.96 GtC yr –1 (Krause-Jensen and Duarte, 2016 <sup>[[#fn:r1151|1151]]</sup> ). They cover about 25% of the world’s coastline (Filbee-Dexter et al., 2016 <sup>[[#fn:r1152|1152]]</sup> ), mostly temperate and polar (Steneck et al., 2003 <sup>[[#fn:r1153|1153]]</sup> ). Canopy-forming macroalgae provide habitat for many associated invertebrates and fish communities (Pessarrodona et al., 2019 <sup>[[#fn:r1154|1154]]</sup> ). This assessment synthesises new evidence since SR15 on climate risks and impacts, and their interactions with non-climatic drivers on ecosystem biodiversity, structure and functioning.

Observational and experimental evidence since SR15 (Hoegh-Guldberg et al. 2018 <sup>[[#fn:r1155|1155]]</sup> ) supports that report’s conclusions that kelp forests are already experiencing large-scale changes, and that critical thresholds occur for some forests at 1.5°C of global warming ( ''high confidence'' ). Due to their low capacity to relocate and high sensitivity to warming, kelp forests are projected to experience higher frequency of mass mortality events as the exposure to extreme temperature rises ( ''very high confidence'' ). Moreover, changes in ocean currents have facilitated the entry of tropical herbivorous fish into temperate kelp forests decreasing their distribution and abundance ( ''medium confidence'' ). More evidence from model projections in the 21st century supports this observed range contraction of kelp forests at the warm end of their distributional margins and expansion at the poleward end with the rate being faster for high emission scenarios ( ''high confidence'' ).

New global estimates show that the abundance of kelp forests has decreased at a rate of ~2% per year over the past half century (Wernberg et al., 2019 <sup>[[#fn:r1156|1156]]</sup> ), mainly due to ocean warming and marine heat waves (e.g., in western Australia a mean loss of 43% in area followed a marine heat weave in summer 2010–2011 (Wernberg et al., 2016), Section 6.4.2.1), as well as from other human stressors ( ''high confidence'' ) (Filbee-Dexter and Wernberg, 2018 <sup>[[#fn:r1157|1157]]</sup> ). At some localities, human-driven environmental changes such as coastal eutrophication and pollution is causing severe deterioration of kelp forests adding to the loss of these ecosystems from warming, storms and heat weaves (Andersen et al., 2013 <sup>[[#fn:r1158|1158]]</sup> ; Filbee-Dexter and Wernberg, 2018 <sup>[[#fn:r1159|1159]]</sup> ).

Two global datasets and one dataset covering European coastlines (Araujo et al., 2016 <sup>[[#fn:r1160|1160]]</sup> ; Krumhans et al., 2016 <sup>[[#fn:r1161|1161]]</sup> ; Poloczanska et al., 2016 <sup>[[#fn:r1162|1162]]</sup> ) identify large local and regional variations in kelp abundance over the past half century with 38% of these ecoregions showing a decline, 27% an increase and 35% no change (Krumhans et al., 2016 <sup>[[#fn:r1163|1163]]</sup> ). These data reflect the high spatio-temporal variability and resilience of kelp forests (Reed et al., 2016 <sup>[[#fn:r1164|1164]]</sup> ; Wernberg et al., 2018 <sup>[[#fn:r1165|1165]]</sup> ). For example, a 34 year dataset of kelp canopy biomass along the California coastline does not yet show a significant response to global warming because this ecosystem responds to low frequency marine climate oscillations (Bell et al., 2018c <sup>[[#fn:r1166|1166]]</sup> ). However, between 1950‒2010 regional warming caused consistent negative responses in abundance, phenology, demography and calcification of macroalgae for the northeast Atlantic and southeast Indian Ocean (Poloczanska et al., 2016 <sup>[[#fn:r1167|1167]]</sup> ). Declines in kelp forest abundance attributed to climate change and not related to sea urchin overgrazing (which is a major driver of decline and regime shift; Ling et al. (2014)) have been documented since the 1970s and evidence has increased within the last two decades (Filbee-Dexter and Wernberg, 2018 <sup>[[#fn:r1168|1168]]</sup> ). Despite a lack of data from some regions such as South America (Pérez-Matus et al., 2017 <sup>[[#fn:r1169|1169]]</sup> ), observational evidence since SR15 supports with ''very'' ''high confidence'' that warming is driving a contraction of kelp forests at low latitudes (Franco et al., 2018b <sup>[[#fn:r1170|1170]]</sup> ; Casado-Amezúa et al., 2019 <sup>[[#fn:r1171|1171]]</sup> ; Pessarrodona et al., 2019 <sup>[[#fn:r1172|1172]]</sup> ) and expansion in polar regions ( ''medium confidence'' ) (Section 3.2.3.1.2) (Bartsch et al., 2016 <sup>[[#fn:r1173|1173]]</sup> ; Paar et al., 2016 <sup>[[#fn:r1174|1174]]</sup> ).

In many areas worldwide where the distribution range of kelp has contracted due to climatic and non-climatic drivers, it has been replaced by a less diverse and less complex turf-dominated ecosystem (Filbee-Dexter and Wernberg, 2018 <sup>[[#fn:r1175|1175]]</sup> ) ( ''high confidence'' ). Kelp supports other ecosystem components by providing food, substrate for spawning and habitat that mediate trophic interactions (O’Brien et al., 2018); its degradation therefore reduces species richness, biomass production and dependent flora and fauna species (Teagle and Smale, 2018 <sup>[[#fn:r1176|1176]]</sup> ; Pessarrodona et al., 2019 <sup>[[#fn:r1177|1177]]</sup> ). In the northeast Atlantic, the warm water species ''Laminaria ochroleuca'' is expanding poleward into regions previously dominated by the cold water species ''L. hyperborea'' which is retreating at its southern edge. These two kelp species are similar in morphology, but the cold water ''L. hyperborea'' hosts sessile communities of algae and invertebrates 12 times more diverse and richer in biomass than the warm water kelp species (Teagle and Smale, 2018 <sup>[[#fn:r1178|1178]]</sup> ). Climate-driven shifts in the species composition also affect carbon cycling, because warm-temperate kelps produce larger pools of organic matter than cold-temperate species, and their detritus is degraded faster (Pessarrodona et al., 2019 <sup>[[#fn:r1179|1179]]</sup> ).

New empirical eco-physiological studies in combination with field surveys support the evidence for climate change causing kelp forest degradation and range shifts (Franco et al., 2018b <sup>[[#fn:r1180|1180]]</sup> ; Wernberg et al., 2018 <sup>[[#fn:r1181|1181]]</sup> ). For example, interactive effects of ocean warming and acidification cause kelp degradation and disease-like symptoms, with detrimental effects on photosynthetic efficiency (Qiu et al., 2019 <sup>[[#fn:r1182|1182]]</sup> ). Enhanced herbivory due to warming and the establishment of herbivorous fish species in temperate kelp forest has been observed to enhance ecosystem degradation (Vergés et al., 2016 <sup>[[#fn:r1183|1183]]</sup> ). However, invader seaweed species driven by warming can create more complex trophic interactions, reducing the consumption by herbivorous gastropods (Miranda et al., 2019 <sup>[[#fn:r1184|1184]]</sup> ). Increased physical stress by storm events also alters the kelps community, affecting the recruitment time of kelp species. The resulting dominance of younger stages favors species with a year-round spore production or an opportunistic life strategy, reducing the kelp canopy (Pereira et al., 2017 <sup>[[#fn:r1185|1185]]</sup> ).

Projections of future distribution of kelp species based on their physiological thresholds show major species-specific range shifts under different emission scenarios. For example, under RCP2.6, laminaria and other canopy-forming seaweed species in the Northwest Atlantic are projected to show northward range shifts at their southern (warm) edge of ≤40 km, with some equatorial range expansion from 2050 to 2100. That northward range shift increases to 406 km under RCP8.5 (at 13–19 km per decade, including contractions of their warmer edges) (Wilson et al., 2019 <sup>[[#fn:r1186|1186]]</sup> ). Whilst no changes in species richness are projected under RCP2.6, more than 50% richness loss is projected under RCP8.5 in some areas (Wilson et al., 2019 <sup>[[#fn:r1187|1187]]</sup> ). Overall, model projections show that worldwide range contractions of kelps can be expected to continue at the warm end of distributional margins and range expansions at their poleward end ( ''high confidence'' ) (Raybaud et al., 2013 <sup>[[#fn:r1188|1188]]</sup> ; Assis et al., 2016 <sup>[[#fn:r1189|1189]]</sup> ; Assis et al., 2018 <sup>[[#fn:r1190|1190]]</sup> ; Wilson et al., 2019 <sup>[[#fn:r1191|1191]]</sup> ).

In summary, kelp forests have experienced large-scale habitat loss and degradation of ecosystem structure and functioning over the past half century, implying a moderate to high level of risk at present conditions of global warming ( ''high confidence'' ) (Section 5.3.7). The loss of kelp forests is followed by the colonisation of turfs, which contributes to the reduction in habitat complexity, carbon storage and diversity ( ''high confidence'' ). Kelp ecosystems are expected to continue to decline in temperate regions driven by ocean warming and intensification of extreme climate events ( ''high confidence'' ). The level of risk for the ecosystem is projected to rise to very high under RCP8.5 scenario by 2100 ( ''high confidence'' )

<span id="risk-assessment-for-coastal-ecosystems"></span>
=== 5.3.7 Risk Assessment for Coastal Ecosystems ===

<div id="section-5-3-7risk-assessment-for-coastal-ecosystems-block-1"></div>

This section synthesises the assessment of climate impacts on coastal ecosystems’ biodiversity, structure and functioning and the levels of risk under contrasting future conditions of global warming. As described in Section 5.2.5, the format for Figure 5.16 matches that of Figure 19.4 of AR5 (Oppenheimer et al., 2015) and Figure 3.20 of SR15 (Hoegh-Guldberg et al., 2018), indicating the levels of additional risk as colours (white, yellow, red and purple). The elements or burning embers for coastal ecosystems (Figure 5.16) indicate how risks increase with ocean warming, acidification, deoxygenation, SLR and extreme events with a comparison between present day conditions (2000s) and future conditions by the year 2100 under low (RCP2.6) and high (RCP8.5) CO 2 emission scenarios. The transition between the levels of risk for each type of coastal ecosystem is estimated from key evidence assessed in Sections 5.3.1 to 5.3.6. The embers are based on SST  and the transition-values may have an error of ±0.3°C depending on the consensus of expert judgment. The assessed confidence in assigning the levels of risk at present day and future scenarios are ''low, medium, high'' and ''very high'' levels of confidence. A detailed account of the procedures involved in developing the ember for each type of coastal ecosystem is given in the supplementary material (SM5.3). This includes the description of climate hazards, sensitivity of key biotic and abiotic components, natural adaptive capacity, and observed impacts and projected risks. The burning embers for seagrass meadows, warm water corals and mangrove forests are in agreement with the conclusions in SR15 (Hoegh-Guldberg et al., 2018). The more recent literature assessed here strengthens the overall confidence in the assignment of transition and the level of risk for each ecosystem.

Detection and attribution studies show that climate change impacts began over the past 50 years in coastal ecosystems, indicating a transition from undetectable risk (white areas in Figure 5.16) to moderate risk below recent sea surface temperatures for some ecosystems ( ''high confidence'' ). This transition occurs at lower global levels of sea surface warming for coral reefs (0.2°C‒0.4°C) ( ''high confidence'' ), seagrass meadows (0.5°C‒0.8°C) ( ''very'' ''high confidence'' ) and kelp forests (0.6°C‒1.0°C) ( ''high confidence'' ), with coral reefs already at high risk (0.4°C‒0.6°C) for the present day ( ''very high confidence'' ). Global common responses include large-scale coral bleaching events (Section 5.3.4) and contraction of seagrass meadows (Section 5.3.2) and kelp forests (Section 5.3.6) at low-latitudes ( ''high confidence'' ), in response to warming and marine heat waves. Degraded coral reefs and kelp forests have shifted to algal and turf-dominated ecosystem at several regions worldwide, causing loss of habitat complexity and biodiversity.

The transition from undetectable to moderate risk in salt marshes (Section 5.3.2) and rocky shores (Section 5.3.5) takes place between 0.7°C‒1.2°C of global sea surface warming ( ''medium/high confidence'' ), and between 0.9°C‒1.8°C ( ''medium confidence)'' in sandy beaches (Section 5.3.3), estuaries (Section 5.3.1) and mangrove forests (Section 5.3.2) (Figure 5.16). In all these coastal ecosystems, the detection and attribution of changes in biodiversity, structure and functioning are not as robust as in coral, seagrass and kelp ecosystems that have been extensively studied over the past decades and are highly sensitive to extreme climate events. Estuaries and sandy beaches are highly dynamic in terms of hydrological and geomorphological processes, giving them more natural adaptive capacity to climate impacts. In these systems, sediment relocation, soil accretion and landward expansion of vegetation may mitigate against flooding and habitat loss in the context of SLR and extreme climate-driven erosion. Common global responses observed since 1970 include poleward expansion of mangrove forests due to warming; transformation of salt marshes into mudflats; shifts in species composition in response to flooding and salinisation; upstream migration of estuarine biota; and redistribution of macrobenthic communities in sandy beaches. Calcified organisms in intertidal rocky shores are highly sensitive to ocean warming and acidification, marine heat waves and heat exposure during low tide, with observed mass mortality events and reduced calcification.

In all coastal ecosystems, multiple climate hazards will emerge from historical variability in the 21st century under RCP8.5 (Box 5.1), while the time of emergence will be later and with less climate hazard under RCP2.6. Non-climatic human impacts such as eutrophication add to, and in some cases, exacerbate these large-scale slow climate drivers beyond biological thresholds at local scale (e.g., deoxygenation).

All coastal ecosystems will experience high to very high risk under RCP8.5 by the end of the 21st century. The ecosystems expected to be at very high risk under the high emission scenario are coral reefs (transition from high to very high risk 0.6°C‒1.2°C) ( ''very high confidence'' ), seagrasses meadows (2.2°C‒3.0°C) ( ''high confidence'' ), kelp forests (2.2°C‒2.8°C) ( ''high confidence'' ) and rocky shores (2.9°C‒3.4°C) ( ''medium confidence'' ). These ecosystems have low to moderate adaptive capacity, as they are highly sensitive to ocean warming, marine heat waves and acidification. For example, kelp forests at low-latitudes and temperate seagrass meadows with endemic species will continue to retreat with more frequent extreme temperatures, and their low dispersal ability will elevate the risk of local extinction. Biogenic shallow reefs with calcified organisms (e.g., corals, mussels, calcified algae) are particularly sensitive to ocean acidification and compound effects with rising temperatures, deoxygenation, SLR and increasing extreme events, making these ecosystems highly vulnerable (with low resilience) to future emission scenarios. Furthermore, almost all coral reefs will greatly decline from their current levels, even if global warming remains below 2°C ( ''very high confidence'' ). Any coral reefs that do survive to the end of the century will not be the same because of irreversible changes in habitat structure and functioning, including species extinctions and food web disruptions; these changes are already taking place (e.g., the Caribbean reefs). The transition to new ecosystem states driven by unpredictable pulses of disturbance and progressive climate hazards will have negative impacts on ecosystem services (Section 5.4).

The ecosystems at moderate to high risk under future emission scenarios (Figure 5.16) are mangrove forests (transition from moderate to high risk at 2.5°C‒2.7°C of global sea surface warming), estuaries and sandy beaches (2.3°C‒3.0°C) and salt marshes (transition from moderate to high risk at 1.8°C‒2.7°C and from high to very high risk at 3.0°C‒3.4°C) ( ''medium confidence'' ). Mangrove forests and salt marshes can initially cope with SLR by plant biomass accumulation, soil accretion and sediment relocation, but the evidence shows they are unlikely to withstand the SLR projected under RCP8.5. Moreover, pervasive coastal squeeze and human-driven habitat deterioration will reduce the natural capacity of these ecosystems to adapt to climate impacts ( ''high confidence'' ). Projected warming and SLR by the end of the century will continue to expand salinisation and hypoxia in estuaries with high risk of impacts for benthic and pelagic biota. These impacts will be more pronounced under RCP8.5 in more vulnerable eutrophic, shallow and microtidal estuaries in temperate and high latitudes. Erosion in sandy beach ecosystems will continue with global warming, rising sea level and more intense and frequent storm surges and marine heat waves. The risk of losing habitats for flora and fauna is expected to rise to high level under the high emission scenario by the end of the 21st century ( ''medium confidence'' , Figure 5.16). By contrast, the risk of impacts is expected to be only slightly higher than present for a low emission scenario than today ( ''medium confidence,'' Figure 5.16).

All types of ecosystems that have been assessed in the open ocean (Sections 5.2.3 and 5.2.4) and coastal areas (Sections 5.3.1 to 5.3.6) show increased risk under both the low and the high emission scenarios (RCP2.6 and RCP8.5) compared with the present level of change (Figure 5.16). In all assessed cases with all of the factors considered (climate drivers and physiological understanding), RCP2.6 has a lower level of risk than RCP8.5 ( ''very high confidence'' ).

<div id="section-5-3-7risk-assessment-for-coastal-ecosystems-block-2"></div>

<span id="figure-5.16"></span>
<!-- START IMG -->
<!-- IMG TITLE -->
'''Figure 5.16'''

<span id="figure-5.16-risk-scenarios-for-open-ocean-upper-panel-and-coastal-lower-panel-ecosystems-based-on-observed-and-projected-climate-impacts.-present-day-corresponds-to-the-2000s-whereas-the-different-greenhouse-emissions-scenarios-representative-concentration-pathway-rcp2.6-and-rcp8.5-correspond-to-year-2100.-multiple-climatic-hazards-are-considered-including-ocean-warming-deoxygenation-acidification-changes"></span>
<!-- IMG CAPTION -->
'''Figure 5.16 | Risk scenarios for open ocean (upper panel) and coastal (lower panel) ecosystems based on observed and projected climate impacts. ‘Present day’ corresponds to the 2000s, whereas the different greenhouse emissions scenarios: Representative Concentration Pathway (RCP)2.6 and RCP8.5 correspond to year 2100. Multiple climatic hazards are considered, including ocean warming, deoxygenation, acidification, changes […]'''
<!-- IMG FILE -->
[[File:03f73e37fa8636052ef117c0dd41c161 IPCC-SROCC-CH_5_16.jpg]]

Figure 5.16 | Risk scenarios for open ocean (upper panel) and coastal (lower panel) ecosystems based on observed and projected climate impacts. ‘Present day’ corresponds to the 2000s, whereas the different greenhouse emissions scenarios: Representative Concentration Pathway (RCP)2.6 and RCP8.5 correspond to year 2100. Multiple climatic hazards are considered, including ocean warming, deoxygenation, acidification, changes in nutrients, particulate organic carbon flux and sea level rise (SLR) (see sections 5.2 and 5.3). The projected changes in sea surface temperature (SST) from an ensemble of general circulation models (left panels) indicate the level of ocean changes under RCP2.6 and RCP8.5 (see Cross Chapter Box 1 Table CB1 for the projected global average changes in average air temperature, SST and other selected ocean variables). Global average impacts/risks are represented. Regional variations of risks/impacts are described in Section 5.2.5, 5.3.7, SM5.2 and SM5.5. Impact/risk levels do not consider human risk reduction strategies such as societal adaptation, or future changes in non-climatic hazards. The grey vertical bars indicate the transition between the levels of risks, with their confidence level based on expert judgment. Note: The figure depicts climate change impacts and risks on warm water corals taken from SR15, based on global models. Observed impacts on coral reefs ecosystems outlined in Section 5.3.4 and Box 5.5 reveal a more complex situation that may result in regional differences in confidence levels.

<!-- END IMG -->
<span id="changing-marine-ecosystem-services-and-human-well-being"></span>