12.1: Introduction

Sea level rise is closely linked to increasing global temperatures.Thus, even as uncertainties remain about just how much sea levelmay rise this century, it is virtually certain that sea level risethis century and beyond will pose a growing challenge to coastalcommunities, infrastructure, and ecosystems from increased (permanent)inundation, more frequent and extreme coastal flooding, erosion ofcoastal landforms, and saltwater intrusion within coastal riversand aquifers. Assessment of vulnerability to rising sea levelsrequires consideration of physical causes, historical evidence, andprojections. A risk-based perspective on sea level rise points tothe need for emphasis on how changing sea levels alter the coastalzone and interact with coastal flood risk at local scales.

This chapter reviews the physical factors driving changes in globalmean sea level (GMSL) and those causing additional regional variationsin relative sea level (RSL). It presents geological and instrumentalobservations of historical sea level changes and an assessment ofthe human contribution to sea level change. It then describes arange of scenarios for future levels and rates of sea level change,and the relationship of these scenarios to the RepresentativeConcentration Pathways (RCPs). Finally, it assesses the impact ofchanges in sea level on extreme water levels.

While outside the scope of this chapter, it is important to notethe myriad of other potential impacts associated with RSL rise,wave action, and increases in coastal flooding. These impacts includeloss of life, damage to infrastructure and the built environment,salinization of coastal aquifers, mobilization of pollutants,changing sediment budgets, coastal erosion, and ecosystem changessuch as marsh loss and threats to endangered flora and fauna.¹ While allof these impacts are inherently important, some also have thepotential to influence local rates of RSL rise and the extent ofwave-driven and coastal flooding impacts. For example, there isevidence that wave action and flooding of beaches and marshes caninduce changes in coastal geomorphology, such as sediment build up,that may iteratively modify the future flood risk profile ofcommunities and ecosystems.²

12.2: Physical Factors Contributing to Sea Level Rise

Sea level change is driven by a variety of mechanisms operating atdifferent spatial and temporal scales (see Kopp et al. 2015³ for a review).GMSL rise is primarily driven by two factors: 1) increased volumeof seawater due to thermal expansion of the ocean as it warms, and2) increased mass of water in the ocean due to melting ice frommountain glaciers and the Antarctic and Greenland ice sheets.⁴ The overallamount (mass) of ocean water, and thus sea level, is also affectedto a lesser extent by changes in global land-water storage, whichreflects changes in the impoundment of water in dams and reservoirsand river runoff from groundwater extraction, inland sea and wetlanddrainage, and global precipitation patterns, such as occur duringphases of the El Niño–Southern Oscillation (ENSO).^4,5,6,7,8

Sea level and its changes are not uniform globally for severalreasons. First, atmosphere–ocean dynamics—driven by ocean circulation,winds, and other factors—are associated with differences in theheight of the sea surface, as are differences in density arisingfrom the distribution of heat and salinity in the ocean. Changesin any of these factors will affect sea surface height. For example,a weakening of the Gulf Stream transport in the mid-to-late 2000smay have contributed to enhanced sea level rise in the oceanenvironment extending to the northeastern U.S. coast,^9,10,11 a trend that manymodels project will continue into the future.¹²

Second, the locations of land ice melting and land water reservoirchanges impart distinct regional “static-equilibrium fingerprints”on sea level, based on gravitational, rotational, and crustaldeformation effects(Figure 12.1a–d).¹³ For example, sea levelfalls near a melting ice sheet because of the reduced gravitationalattraction of the ocean toward the ice sheet; reciprocally, it risesby greater than the global average far from the melting ice sheet.

Third, the Earth’s mantle is still moving in response to the lossof the great North American (Laurentide) and European ice sheetsof the Last Glacial Maximum; the associated changes in the heightof the land, the shape of the ocean basin, and the Earth’s gravitationalfield give rise to glacial-isostatic adjustment (Figure 12.1e). Forexample, in areas once covered by the thickest parts of the greatice sheets of the Last Glacial Maximum, such as in Hudson Bay andin Scandinavia, post-glacial rebound of the land is causing RSL to fall.Along the flanks of the ice sheets, suchas along most of the east coast of the United States, subsidenceof the bulge that flanked the ice sheet is causing RSL to rise.

Finally, a variety of other factors can cause local vertical landmovement. These include natural sediment compaction, compactioncaused by local extraction of groundwater and fossil fuels, andprocesses related to plate tectonics, such as earthquakes and moregradual seismic creep (Figure 12.1f).^14,15

Compared to many climate variables, the trend signal for sea levelchange tends to be large relative to natural variability. However,at interannual timescales, changes in ocean dynamics, density, andwind can cause substantial sea level variability in some regions.For example, there has been a multidecadal suppression of sea levelrise off the Pacific coast¹⁶ and large year-to-yearvariations in sea level along the Northeast U.S. coast.¹⁷ Local rates of landheight change have also varied dramatically on decadal timescalesin some locations, such as along the western Gulf Coast, where ratesof subsurface extraction of fossil fuels and groundwater have variedover time.¹⁸

 

Figure 12.1

(a–d) Static-equilibriumfingerprints of the relative sea level (RSL) effect of land icemelt, in units of feet of RSL change per feet of global mean sealevel (GMSL) change, for mass loss from (a) Greenland, (b) WestAntarctica, (c) East Antarctica, and (d) the median projectedcombination of melting glaciers, after Kopp et al.^3,76 (e) Model projectionsof the rate of RSL rise due to glacial-isostatic adjustment (unitsof feet/century), after Kopp et al.³ (f) Tide gauge-basedestimates of the non-climatic, long term contribution to RSL rise,including the effects of glacial isostatic adjustment, tectonics,and sediment compaction (units of feet/century).⁷⁶ (Figure source: (a)–(d)Kopp et al. 2015,³ (e) adapted from Kopp et al. 2015;³ (f) adapted from Sweetet al. 2017⁷¹).

12.3: Paleo Sea Level

Geological records of temperature and sea level indicate that duringpast warm periods over the last several millions of years, GMSL washigher than it is today.^19,20 During the Last Interglacialstage, about 125,000 years ago, global average sea surface temperaturewas about 0.5° ± 0.3°C (0.9° ± 0.5°F) above the preindustrial level[that is, comparable to the average over 1995–2014, when globalmean temperature was about 0.8°C (1.4°F) above the preindustriallevels].²¹Polar temperatures were comparable to those projected for 1°–2°C(1.8°–3.6°F) of global mean warming above the preindustrial level.At this time, GMSL was about 6–9 meters (about 20–30 feet) higherthan today (Figure 12.2a).^22,23 This geological benchmarkmay indicate the probable long-term response of GMSL to the minimummagnitude of temperature change projected for the current century.

 

Figure 12.2

VIEW

(a) The relationship between peak global mean temperature, maximumglobal mean sea level (GMSL), and source(s) of meltwater for twoperiods in the past with global mean temperature comparable to orwarmer than present. Light blue shading indicates uncertainty ofGMSL maximum. Red pie charts over Greenland and Antarctica denotefraction, not location, of ice retreat. Atmospheric CO₂levels in 2100 are shown under RCP8.5. (b) GMSL rise from −500 to1900 CE, from Kopp et al.’s³² geological and tidegauge-based reconstruction (blue), from 1900 to 2010 from Hay etal.’s³³ tidegauge-based reconstruction (black), and from 1992 to 2015 from thesatellite-based reconstruction updated from Nerem et al.³⁵ (magenta). (Figuresource: (a) adapted from Dutton et al. 2015²⁰ and (b) Sweet et al.2017⁷¹).

Similarly, during the mid-Pliocene warm period, about 3 millionyears ago, global mean temperature was about 1.8°–3.6°C (3.2°–6.5°F)above the preindustrial level.²⁴ Estimates of GMSL areless well constrained than during the Last Interglacial, due to thesmaller number of local geological sea level reconstruction and thepossibility of significant vertical land motion over millions ofyears.²⁰ Somereconstructions place mid-Pliocene GMSL at about 10–30 meters (about30–100 feet) higher than today.²⁵ Sea levels this highwould require a significantly reduced Antarctic ice sheet, highlightingthe risk of significant Antarctic ice sheet loss under such levelsof warming (Figure 12.2a).

For the period since the Last Glacial Maximum, about 26,000 to19,000 years ago,²⁶ geologists can produce detailed reconstructions of sea levelsas well as rates of sea level change. To do this, they use proxiessuch as the heights of fossil coral reefs and the populations ofdifferent salinity-sensitive microfossils within salt marshsediments.²⁷During the main portion of the deglaciation, from about 17,000 to8,000 years ago, GMSL rose at an average rate of about 12 mm/year(0.5 inches/year).²⁸ However, there were periods of faster rise. For example, duringMeltwater Pulse 1a, lasting from about 14,600 to 14,300 years ago,GMSL may have risen at an average rate about 50 mm/year (2inches/year).²⁹

Since the disappearance of the last remnants of the North American(Laurentide) Ice Sheet about 7,000 years ago³⁰ to about the start ofthe 20th century, however, GMSL has been relatively stable. Duringthis period, total GMSL rise is estimated to have been about 4meters (about 13 feet), most of which occurred between 7,000 and4,000 years ago.²⁸ The Third National Climate Assessment (NCA3) noted, based on ageological data set from North Carolina,³¹ that the 20th centuryGMSL rise was much faster than at any time over the past 2,000years. Since NCA3, high-resolution sea level reconstructions havebeen developed for multiple locations, and a new global analysisof such reconstructions strengthens this finding.³² Over the last 2,000years, prior to the industrial era, GMSL exhibited small fluctuationsof about ±8 cm (3 inches), with a significant decline of about 8cm (3 inches) between the years 1000 and 1400 CE coinciding withabout 0.2°C (0.4°F) of global mean cooling.³² The rate of rise in thelast century, about 14 cm/century (5.5 inches/century), was greaterthan during any preceding century in at least 2,800 years (Figure12.2b).³²

12.4: Recent Past Trends (20th and 21st Centuries)

12.4.1 Global Tide Gauge Network and Satellite Observations

A global tide gauge network provides the century-long observationsof local RSL, whereas satellite altimetry provides broader coverageof sea surface heights outside the polar regions starting in 1993.GMSL can be estimated through statistical analyses of either dataset. GMSL trends over the 1901–1990 period vary slightly (Hay etal. 2015:³³1.2 ± 0.2 mm/year [0.05 inches/year]; Church and White 2011:³⁴ 1.5 ± 0.2mm/year [0.06 inches/year]) with differences amounting to about1 inch over 90 years. Thus, these results indicate about 11–14 cm(4–5 inches) of GMSL rise from 1901 to 1990.

Tide gauge analyses indicate that GMSL rose at a considerably fasterrate of about 3 mm/year (0.12 inches/year) since 1993,^33,34 a result supported bysatellite data indicating a trend of 3.4 ± 0.4 mm/year (0.13 ± 0.02inches/year) over 1993–2015 (update to Nerem et al. 2010³⁵). These resultsindicate an additional GMSL rise of about 7 cm (about 3 inches)since 1990 (Figure 12.2b, Figure 12.3a) and about 16–21 cm (about7–8 inches) since 1900. Satellite (altimetry and gravity) and insitu water column (Argo floats) measurements show that, since 2005,about one third of GMSL rise has been from steric changes (primarilythermal expansion) and about two thirds from the addition of massto the ocean, which represents a growing land-ice contribution(compared to steric) and a departure from the relative contributionsearlier in the 20th century (Figure 12.3a).^{4,36,37,38,39,40}

In addition to land ice, the mass-addition contribution also includesnet changes in global land-water storage. This term varied in signover the course of the last century, with human-induced changes inland-water storage being negative (perhaps as much as about −0.6mm/year [−0.02 inches/year]) during the period of heavy damconstruction in the middle of the last century, and turning positivein the 1990s as groundwater withdrawal came to dominate.⁸ On decadal timescales,precipitation variability can dominate human-induced changes inland water storage; recent satellite-gravity estimates suggest that,over 2002–2014, a human-caused land-water contribution to GMSL of0.4 mm/year (0.02 inches/year) was more than offset by −0.7 mm/year(−0.03 inches/year) due to natural variability.⁵

Comparison of results from a variety of approaches supports theconclusion that a substantial fraction of GMSL rise since 1900 isattributable to human-caused climate change.^{32,41,42,43,44,45,46,47,48} For example, based onthe long term historical relationship between temperature and rateof GMSL change, Kopp et al.³² found that GMSL risewouldextremely likely have been less than 59% of observed in theabsence of 20th century global warming, and that it isvery likelythat GMSL has been higher since 1960 than it would have been without20th century global warming (Figure 12.3b). Similarly, using avariety of models for individual components, Slangen et al.⁴¹ found that about80% of the GMSL rise they simulated for 1970–2005 and about halfof that which they simulated for 1900–2005 was attributable toanthropogenic forcing.

Over timescales of a few decades, ocean–atmosphere dynamics drivesignificant variability in sea surface height, as can be observedby satellite (Figure 12.3c) and in tide gauge records that havebeen adjusted to account for background rates of rise due to longterm factors like glacio-isostatic adjustments. For example, theU.S. Pacific Coast experienced a slower-than-global increase betweenabout 1980 and 2011, while the western tropical Pacific experienceda faster-than-global increase in the 1990s and 2000s. This patternwas associated with changes in average winds linked to the PacificDecadal Oscillation (PDO)^16,49,50 and appears to havereversed since about 2012.⁵¹ Along the Atlantic coast,the U.S. Northeast has experienced a faster-than-global increasesince the 1970s, while the U.S. Southeast has experienced aslower-than-global increase since the 1970s. This pattern appearsto be tied to changes in the Gulf Stream,^10,12,52,53 although whether thesechanges represent natural variability or a long-term trend remainsuncertain.⁵⁴

 

Figure 12.3

(a) Contributions of ocean mass changes from land ice and land waterstorage (measured by satellite gravimetry) and ocean volume changes(or steric, primarily from thermal expansion measured by in situocean profilers) and their comparison to global mean sea level(GMSL) change (measured by satellite altimetry) since 1993. (b) Anestimate of modeled GMSL rise in the absence of 20th century warming(blue), from the same model with observed warming (red), and comparedto observed GMSL change (black). Heavy/light shading indicates the17th–83rd and 5th–95th percentiles. (c) Rates of change from 1993to 2015 in sea surface height from satellite altimetry data; updatedfrom Kopp et al.³ using data updated from Church and White.³⁴ (Figure source: (a)adapted and updated from Leuliette and Nerem 2016,⁴⁰ (b) adapted from Koppet al. 2016³² and (c) adapted and updated from Kopp et al. 2015³).

12.4.2 Ice Sheet Gravity and Altimetry and Visual Observations

Since NCA3, Antarctica and Greenland have continued to lose icemass, with mounting evidence accumulating that mass loss isaccelerating. Studies using repeat gravimetry (GRACE satellites),repeat altimetry, GPS monitoring, and mass balance calculationsgenerally agree on accelerating mass loss in Antarctica.^55,56,57,58 Together, theseindicate a mass loss of roughly 100 Gt/year (gigatonnes/year) overthe last decade (a contribution to GMSL of about 0.3 mm/year [0.01inches/year]). Positive accumulation rate anomalies in EastAntarctica, especially in Dronning Maud Land,⁵⁹ have contributed to thetrend of slight growth there (e.g., Seo et al. 2015;⁵⁷ Martín-Español et al.2016⁵⁸), butthis is more than offset by mass loss elsewhere, especially in WestAntarctica along the coast facing the Amundsen Sea,^60,61 Totten Glacier in EastAntarctica,^62,63 and along the Antarctic Peninsula.^57,58,64 Floating ice shelvesaround Antarctica are losing mass at an accelerating rate.⁶⁵ Mass loss fromfloating ice shelves does not directly affect GMSL, but does allowfaster flow of ice from the ice sheet into the ocean.

Estimates of mass loss in Greenland based on mass balance frominput-output, repeat gravimetry, repeat altimetry, and aerial imageryas discussed inChapter 11: Arctic Changesreveal a recentacceleration.⁶⁶ Mass loss averaged approximately 75 Gt/year (about 0.2 mm/year[0.01 inches/year] GMSL rise) from 1900 to 1983, continuing at asimilar rate of approximately 74 Gt/year through 2003 beforeaccelerating to 186 Gt/year (0.5 mm/year [0.02 inches/year] GMSLrise) from 2003 to 2010.⁶⁷ Strong interannualvariability does exist (seeCh. 11: Arctic Changes), such as duringthe exceptional melt year from April 2012 to April 2013, whichresulted in mass loss of approximately 560 Gt (1.6 mm/year [0.06inches/year]).⁶⁸ More recently (April 2014–April 2015), annual mass losses haveresumed the accelerated rate of 186 Gt/year.^67,69 Mass loss over the lastcentury has reversed the long-term trend of slow thickening linkedto the continuing evolution of the ice sheet from the end of thelast ice age.⁷⁰

12.5: Projected Sea Level Rise

12.5.1 Scenarios of Global Mean Sea Level Rise

No single physical model is capable of accurately representing allof the major processes contributing to GMSL and regional/local RSLrise. Accordingly, the U.S. Interagency Sea Level Rise Task Force(henceforth referred to as “Interagency”)⁷¹ has revised the GMSLrise scenarios for the United States and now provides six scenariosthat can be used for assessment and risk-framing purposes (Figure12.4a; Table 12.1). The low scenario of 30 cm (about 1 foot) GMSLrise by 2100 is consistent with a continuation of the recentapproximately 3 mm/year (0.12 inches/year) rate of rise through to2100 (Table 12.2), while the five other scenarios span a range ofGMSL rise between 50 and 250 cm (1.6 and 8.2 feet) in 2100, withcorresponding rise rates between 5 mm/year (0.2 inches/year) to 44mm/year (1.7 inches/year) towards the end of this century (Table12.2). The highest scenario of 250 cm is consistent with severalliterature estimates of the maximum physically plausible level of21st century sea level rise (e.g., Pfeffer et al. 2008,⁷² updated with Sriveret al. 2012⁷³estimates of thermal expansion and Bamber and Aspinall 2013⁷⁴ estimates of Antarcticcontribution, and incorporating land water storage, as discussedin Miller et al. 2013⁷⁵) and Kopp et al. 2014⁷⁶. Itis also consistent with the high end of recent projections ofAntarctic ice sheet melt discussed below.⁷⁷ The Interagency GMSLscenario interpretations are shown in Table 12.3.

 

Figure 12.4

VIEW

(a) Global mean sea level (GMSL) rise from 1800 to 2100, based onFigure 12.2b from 1800 to 2015, the six Interagency⁷¹ GMSL scenarios (navyblue, royal blue, cyan, green, orange, and red curves), theverylikely ranges in 2100 for different RCPs (colored boxes), and linesaugmenting thevery likely ranges by the difference between themedian Antarctic contribution of Kopp et al.⁷⁶ and the various medianAntarctic projections of DeConto and Pollard.⁷⁷ (b) Relative sea level(RSL) rise (feet) in 2100 projected for the Interagency IntermediateScenario (1-meter [3.3 feet] GMSL rise by 2100) (Figure source:Sweet et al. 2017⁷¹).

The Interagency scenario approach is similar to local RSL risescenarios of Hall et al.⁷⁸ used for all coastalU.S. Department of Defense installations worldwide. The Interagencyapproach starts with a probabilistic projection framework to generatetime series and regional projections consistent with each GMSL risescenario for 2100.⁷⁶ That framework combines probabilistic estimates of contributionsto GMSL and regional RSL rise from ocean processes, cryosphericprocesses, geological processes, and anthropogenic land-waterstorage. Pooling the Kopp et al.⁷⁶ projections across evenlower, lower, and higher scenarios (RCP2.6, 4.5, and 8.5), theprobabilistic projections are filtered to identify pathways consistentwith each of these 2100 levels, with the median (and 17th and 83rdpercentiles) picked from each of the filtered subsets.

12.5.2 Probabilities of Different Sea Level Rise Scenarios

Several studies have estimated the probabilities of different amountsof GMSL rise under different pathways (e.g., Church et al. 2013;⁴ Kopp et al.2014;⁷⁶ Slangenet al. 2014;⁷⁹ Jevrejeva et al. 2014;⁸⁰ Grinsted et al. 2015;⁸¹ Kopp et al.2016;³² Mengelet al. 2016;⁸² Jackson and Jevrejeva 2016⁸³) using a variety ofmethods, including both statistical and physical models. Most ofthese studies are in general agreement that GMSL rise by 2100 isvery likely to be between about 25–80 cm (0.8–2.6 feet) under aneven lower scenario (RCP2.6), 35–95 cm (1.1–3.1 feet) under a lowerscenario (RCP4.5), and 50–130 cm (1.6–4.3 feet) under a higherscenario (RCP8.5), although some projections extend thevery likelyrange for RCP8.5 as high as 160–180 cm (5–6 feet) (Kopp et al.2014,⁷⁶sensitivity study).^80,83 Based on Kopp et al.,⁷⁶ the probability ofexceeding the amount of GMSL in 2100 under the Interagency scenariosis shown in Table 12.4.

The Antarctic projections of Kopp et al.,⁷⁶ the GMSL projections ofwhich underlie Table 12.4, are consistent with a statistical-physicalmodel of the onset of marine ice sheet instability calibrated toobservations of ongoing retreat in the Amundsen Embayment sectorof West Antarctica.⁸⁴ Ritz et al.’s⁸⁴ 95th percentile Antarctic contribution to GMSL of 30 cm by 2100is comparable to Kopp et al.’s⁷⁶ 95th percentile projectionof 33 cm under the higher scenario (RCP8.5). However, emergingscience suggests that these projections may understate the probabilityof faster-than-expected ice sheet melt, particularly for high-endwarming scenarios. While these probability estimates are consistentwith the assumption that the relationship between global temperatureand GMSL in the coming century will be similar to that observedover the last two millennia,^32,85 emerging positive feedbacks(self-amplifying cycles) in the Antarctic Ice Sheet especially^86,87 may invalidate thatassumption. Physical feedbacks that until recently were notincorporated into ice sheet models⁸⁸ could add about 0–10 cm(0–0.3 feet), 20–50 cm (0.7–1.6 feet) and 60–110 cm (2.0–3.6 feet)to central estimates of current century sea level rise under evenlower, lower, and higher scenarios (RCP2.6, RCP4.5 and RCP8.5,respectively).⁷⁷ In addition to marine ice sheet instability, examples of theseinterrelated processes include ice cliff instability and ice shelfhydrofracturing. Processes underway in Greenland may also be leadingto accelerating high-end melt risk. Much of the research has focusedon changes in surface albedo driven by the melt-associated unmaskingand concentration of impurities in snow and ice.⁶⁹ However, ice dynamicsat the bottom of the ice sheet may be important as well, throughinteractions with surface runoff or a warming ocean. As an exampleof the latter, Jakobshavn Isbræ, Kangerdlugssuaq Glacier, and theNortheast Greenland ice stream may be vulnerable to marine ice sheetinstability.⁶⁶

12.5.3 Sea Level Rise after 2100

GMSL rise will not stop in 2100, and so it is useful to considerextensions of GMSL rise projections beyond this point. By 2200, the0.3–2.5 meter (1.0–8.2 feet) range spanned by the six InteragencyGMSL scenarios in year 2100 increases to about 0.4–9.7 meters(1.3–31.8 feet), as shown in Table 12.5. These six scenarios implyaverage rates of GMSL rise over the first half of the next centuryof 1.4 mm/year (0.06 inch/year), 4.6 mm/yr (0.2 inch/year), 16mm/year (0.6 inch/year), 32 mm/year (1.3 inches/year), 46 mm/yr(1.8 inches/year) and 60 mm/year (2.4 inches/year), respectively.Excluding the possible effects of still emerging science regardingice cliffs and ice shelves, it is very likely that by 2200 GMSLwill have risen by 0.3–2.4 meters (1.0–7.9 feet) under an even lowerscenario (RCP2.6), 0.4–2.7 meters (1.3–8.9 feet) under a lowerscenario (RCP4.5), and 1.0–3.7 meters (3.3–12 feet) under the higherscenario (RCP8.5).⁷⁶

Under most projections, GMSL rise will also not stop in 2200. Theconcept of a “sea level rise commitment” refers to the long-termprojected sea level rise were the planet’s temperature to bestabilized at a given level (e.g., Levermann et al. 2013;⁸⁹ Golledge et al.2015⁹⁰). Thepaleo sea level record suggests that even 2°C (3.6°F) of globalaverage warming above the preindustrial temperature may representa commitment to several meters of rise. One modeling study suggestinga 2,000-year commitment of 2.3 m/°C (4.2 feet/°F)⁸⁹ indicates that emissionsthrough 2100 would lock in a likely 2,000-year GMSL rise commitmentof about 0.7–4.2 meters (2.3–14 feet) under an even lower scenario(RCP2.6), about 1.7–5.6 meters (5.6–19 feet) under a lower scenario(RCP4.5), and about 4.3–9.9 meters (14–33 feet) under the higherscenario (RCP8.5).⁹¹ However, as with the 21st century projections, emerging scienceregarding the sensitivity of the Antarctic Ice Sheet may increasethe estimated sea level rise over the next millennium, especiallyfor a higher scenario.⁷⁷ Large-scale climate geoengineering might reduce these commitments,^92,93 but may not be ableto avoid lock-in of significant change.^94,95,96,97 Once changes are realized,they will be effectively irreversible for many millennia, even ifhumans artificially accelerate the removal of CO₂ fromthe atmosphere.⁷⁷

The 2,000-year commitment understates the full sea level risecommitment, due to the long response time of the polar ice sheets.Paleo sea level records (Figure 12.2a) suggest that 1°C of warmingmay already represent a long-term commitment to more than 6 meters(20 feet) of GMSL rise.^20,22,23 A 10,000-year modeling study⁹⁸ suggests that 2°C warmingrepresents a 10,000-year commitment to about 25 meters (80 feet)of GMSL rise, driven primarily by a loss of about one-third of theAntarctic ice sheet and three-fifths of the Greenland ice sheet,while 21st century emissions consistent with a higher scenario(RCP8.5) represent a 10,000-year commitment to about 38 meters (125feet) of GMSL rise, including a complete loss of the Greenland icesheet over about 6,000 years.

12.5.4 Regional Projections of Sea Level Change

Because the different factors contributing to sea level change giverise to different spatial patterns, projecting future RSL changeat specific locations requires not just an estimate of GMSL changebut estimates of the different processes contributing to GMSLchange—each of which has a different associated spatial pattern—aswell as of the processes contributing exclusively to regional orlocal change. Based on the process-level projections of the InteragencyGMSL scenarios, several key regional patterns are apparent in futureU.S. RSL rise as shown for the Intermediate (1 meter [3.3 feet]GMSL rise by 2100 scenario) in Figure 12.4b.

1. RSL rise due to Antarctic Ice Sheet melt is greater than GMSLrise along all U.S. coastlines due to static-equilibrium effects.

2. RSL rise due to Greenland Ice Sheet melt is less than GMSL risealong the coastline of the continental United States due tostatic-equilibrium effects. This effect is especially strong in theNortheast.

3. RSL rise is additionally augmented in the Northeast by theeffects of glacial isostatic adjustment.

4. The Northeast is also exposed to rise due to changes in theGulf Stream and reductions in the Atlantic meridional overturningcirculation (AMOC). Were the AMOC to collapse entirely—an outcomeviewed as unlikely in the 21st century—it could result in as muchas approximately 0.5 meters (1.6 feet) of additional regional sealevel rise (seeCh. 15: Potential Surprises for further discussion).^99,100

5. The western Gulf of Mexico and parts of the U.S. Atlantic Coastsouth of New York are currently experiencing significant RSL risecaused by the withdrawal of groundwater (along the Atlantic Coast)and of both fossil fuels and groundwater (along the Gulf Coast).Continuation of these practices will further amplify RSL rise.

6. The presence of glaciers in Alaska and their proximity to thePacific Northwest reduces RSL rise in these regions, due to boththe ongoing glacial isostatic adjustment to past glacier shrinkageand to the static-equilibrium effects of projected future losses.

7. Because they are far from all glaciers and ice sheets, RSL risein Hawai‘i and other Pacific islands due to any source of meltingland ice is amplified by the static-equilibrium effects.

12.6: Extreme Water Levels

12.6.1 Observations

Coastal flooding during extreme high-water events has become deeperdue to local RSL rise and more frequent from a fixed-elevationperspective.^{78,101,102,103} Trends in annual frequencies surpassing local emergencypreparedness thresholds for minor tidal flooding (i.e., “nuisance”levels of about 30–60 cm [1–2 feet]) that begin to floodinfrastructure and trigger coastal flood “advisories” by NOAA’sNational Weather Service have increased 5- to 10-fold or more sincethe 1960s along the U.S. coastline,¹⁰⁴ as shown in Figure 12.5a.Locations experiencing such trend changes (based upon fits of flooddays per year of Sweet and Park 2014¹⁰⁵) include Atlantic Cityand Sandy Hook, NJ; Philadelphia, PA; Baltimore and Annapolis, MD;Norfolk, VA; Wilmington, NC; Charleston, SC; Savannah, GA; Mayportand Key West, FL; Port Isabel, TX, La Jolla, CA; and Honolulu, HI.In fact, over the last several decades, minor tidal flood rateshave been accelerating within several (more than 25) East and GulfCoast cities with established elevation thresholds for minor(nuisance) flood impacts, fastest where elevation thresholds arelower, local RSL rise is higher, and extreme variability less.^104,105,106

Trends in extreme water levels (for example, monthly maxima) inexcess of mean sea levels (for example, monthly means) exist, butare not commonplace.^{48,101,107,108,109} More common are regional time dependencies in high-waterprobabilities, which can co-vary on an interannual basis withclimatic and other patterns.^{101,110,111,112,113,114,115} These patterns are oftenassociated with anomalous oceanic and atmospheric conditions.^116,117 For instance, theprobability of experiencing minor tidal flooding is compoundedduring El Niño periods along portions of the West and Mid-AtlanticCoasts¹⁰⁵from a combination of higher sea levels and enhanced synoptic forcingand storm surge frequency.^{112,118,119,120}

12.6.2 Influence of Projected Sea Level Rise on Coastal Flood Frequencies

The extent and depth of minor-to-major coastal flooding duringhigh-water events will continue to increase in the future as localRSL rises.^{71,76,78,105,121,122,123,124,125} Relative to fixed elevations, the frequency of high-water eventswill increase the fastest where extreme variability is less and therate of local RSL rise is higher.^{71,76,105,121,124,126} Under the RCP-basedprobabilistic RSL projections of Kopp et al. 2014,⁷⁶ at tide gauge locationsalong the contiguous U.S. coastline, a median 8-fold increase (rangeof 1.1- to 430-fold increase) is expected by 2050 in the annualnumber of floods exceeding the elevation of the current 100-yearflood event (measured with respect to a 1991–2009 baseline sealevel).¹²⁴Under the same forcing, the frequency of minor tidal flooding (withcontemporary recurrence intervals generally <1 year¹⁰⁴) will increase even moreso in the coming decades^105,127 and eventually occur ona daily basis (Figure 12.5b). With only about 0.35 m (<14 inches)of additional local RSL rise (with respect to the year 2000), annualfrequencies of moderate level flooding—those locally with a 5-yearrecurrence interval (Figure 12.5c) and associated with a NOAA coastalflood warning of serious risk to life and property—will increase25-fold at the majority of NOAA tide gauge locations along the U.S.coastline (outside of Alaska) by or about (±5 years) 2080, 2060,2040, and 2030 under the Interagency Low, Intermediate-Low,Intermediate, and Intermediate-High GMSL scenarios, respectively.⁷¹ Figure 12.5d,which shows the decade in which the frequency of such moderate levelflooding will increase 25-fold under the Interagency IntermediateScenario, highlights that the mid- and Southeast Atlantic, westernGulf of Mexico, California, and the Island States and Territoriesare most susceptible to rapid changes in potentially damaging floodfrequencies.

 

Figure 12.5

Figure 12.5: (a) Tidal floods (days per year)exceeding NOAA thresholds for minor impacts at 28 NOAA tide gaugesthrough 2015. (b) Historical exceedances (orange), future projectionsthrough 2100 based upon the continuation of the historical trend(blue), and future projections under median RCP2.6, 4.5 and 8.5conditions, for two of the locations—Charleston, SC and San Francisco,CA. (c) Water level heights above average highest tide associatedwith a local 5-year recurrence probability, and (d) the futuredecade when the 5-year event becomes a 0.2-year (5 or more timesper year) event under the Interagency Intermediate scenario; blackdots imply that a 5-year to 0.2-year frequency change does notunfold by 2200 under the Intermediate scenario. (Figure source: (a)adapted from Sweet and Marra 2016,¹⁶⁵ (b) adapted from Sweetand Park 2014,¹⁰⁵ (c) and (d) Sweet et al. 2017⁷¹).

12.6.3 Waves and Impacts

The combination of a storm surge at high tide with additional dynamiceffects from waves^128,129 creates the most damaging coastal hydraulic conditions.¹³⁰ Simply withhigher-than-normal sea levels, wave action increases the likelihoodfor extensive coastal erosion^131,132,133 and low-island overwash.¹³⁴ Wave runupis often the largest water level component during extreme events,especially along island coastlines where storm surge is constrainedby bathymetry.^78,121,123 On an interannual basis, wave impacts are correlated across thePacific Ocean with phases of ENSO.^135,136 Over the last halfcentury, there has been an increasing trend in wave height and powerwithin the North Pacific Ocean^137,138 that is modulated by thePDO.^137,139 Resultant increases in wave run-up have been more of a factorthan RSL rise in terms of impacts along the U.S. Northwest PacificCoast over the last several decades.¹⁴⁰ In the Northwest AtlanticOcean, no long-term trends in wave power have been observed overthe last half century,¹⁴¹ though hurricane activity drives interannual variability.¹⁴² In terms offuture conditions this century, increases in mean and maximumseasonal wave heights are projected within parts of the northeastPacific, northwest Atlantic, and Gulf of Mexico.^{138,143,144,145}

12.6.4 Sea Level Rise, Changing Storm Characteristics, and Their Interdependencies

Future probabilities of extreme coastal floods will depend upon theamount of local RSL rise, changes in coastal storm characteristics,and their interdependencies. For instance, there have been morestorms producing concurrent locally extreme storm surge and rainfall(not captured in tide gauge data) along the U.S. East and GulfCoasts over the last 65 years, with flooding further compounded bylocal RSL rise.¹⁰⁸ Hemispheric-scale extratropical cyclones may experience anorthward shift this century, with some studies projecting an overalldecrease in storm number (Colle et al. 2015¹¹⁷ and references therein).The research is mixed about strong extratropical storms; studiesfind potential increases in frequency and intensity in some regions,like within the Northeast,¹⁴⁶ whereas others projectdecreases in strong extratropical storms in some regions (e.g.,Zappa et al. 2013¹⁴⁷).

For tropical cyclones, model projections for the North Atlanticmostly agree that intensities and precipitation rates will increasethis century (seeCh. 9: Extreme Storms), although some modelevidence suggests that track changes could dampen the effect in theU.S. Mid-Atlantic and Northeast.¹⁴⁸ Assuming other stormcharacteristics do not change, sea level rise will increase thefrequency and extent of extreme flooding associated with coastalstorms, such as hurricanes and nor’easters. A projected increasein the intensity of hurricanes in the North Atlantic could increasethe probability of extreme flooding along most of the U.S. Atlanticand Gulf Coast states beyond what would be projected based solelyon RSL rise.^{110,149,150,151} In addition, RSL increases are projected to cause a nonlinearincrease in storm surge heights in shallow bathymetry environments^{152,153,154,155,156} and extend wavepropagation and impacts landward.^152,153 However, there is lowconfidence in the magnitude of the increase in intensity and theassociated flood risk amplification, and it could be offset oramplified by other factors, such as changes in storm frequency ortracks (e.g., Knutson et al. 2013,¹⁵⁷ 2015¹⁵⁸).

References