sea-level rise: assessing the risk and estimating the costs

EPIC University of Chicago

30 September 2015

Bob Kopp

Earth System Science & Policy Lab, Rutgers University E-mail: [email protected]

Sea-level rise: Assessing the risks, Estimating the costs

5 Columbus Circle New York, NY 10019 Tel: +1.212.532.1157 Fax: +1.212.532.1162 www.rhg.com

PROJECT DESCRIPTION

An American Climate Risk Assessment Next Generation, in collaboration with Bloomberg Philanthropies and the Paulson Institute, has asked Rhodium Group (RHG) to convene a team of climate scientists and economists to assess the risk to the US economy of global climate change. This assessment, to conclude in late spring 2014,will combine a review of existing literature on the current and potential impacts of climate change in the United States with original research quantifying the potential economic costs of the range of possible climate futures Americans now face. The report will inform the work of a high-level and bipartisan climate risk committee co-chaired by Mayor Bloomberg, Secretary Paulson and Tom Steyer.

BACKGROUND AND CONTEXT

From Superstorm Sandy to Midwest droughts to wildfires in the Rocky Mountains, our weather is becoming more extreme and more expensive. In 2012, climate and weather disasters cost Americans more than $110 billion. Not only was it the hottest year on record, but precipitation was 2.6 inches lower than last century’s average, leading to massive crop losses across the Heartland and wildfires covering more than 9 million acres in the west. And while Sandy was the most destructive storm of the year, it wasn’t the only one with a multi-billion dollar economic price tag.

Weather is inherently variable and no single storm, heat wave or drought can be conclusively attributed to climate change. But there is mounting scientific evidence that greenhouse gas (GHG) emissions are increasing the frequency and severity of many extreme weather events. Sea level rise resulting from human-induced climate change amplified coastal flooding during Sandy. And increased atmospheric concentrations of GHG emissions are already resulting in prolonged stretches of excessively high temperatures, heavier downpours and more severe droughts.

While our understanding of climate change has improved dramatically in recent years, predicting future impacts is still a challenge. Uncertainty surrounding the level of GHG emissions going forward and the sensitivity of the climate system to those emissions makes it difficult to know exactly how much warming will occur. And we are learning more and more every day about how human and natural systems respond to potential changes in temperature, precipitation, sea level and storm patterns.

Uncertainty, of course, is not unique to climate change. The military plans for a wide range of possible conflict scenarios and public health officials prepare for pandemics of low or unknown probability. Households buy insurance to guard against myriad potential perils. And effective risk management is critical to business success and investment performance. In all these areas, decision makers consider a range of possible futures in deciding on a course of action. They work off the best information at hand and take advantage of new information as it becomes available.

Both the US National Academies of Science and the Intergovernmental Panel on Climate Change (IPCC) have suggested that this kind of “iterative risk management” is also the right way to approach climate change. Using this framework, the scientific community is preparing two major assessments of the risks to human and natural systems under a range of possible climate futures. The IPCC’s Fifth Assessment Report (AR5) will provide a global outlook, while the US government’s National Climate Assessment (NCA) will

mailto:[email protected]

2

Research team

Jerry Mitrovica,Carling Hay, Eric Morrow

Solomon Hsiang, Trevor Houser, Kate Larsen, Amir Jina, DJ Rasmussen, James Rising, Michael Delgado, Shashank Mohan, Robert Muir-Wood, Paul Wilson

20th centuryEconomic risk

and many others…

Flood risk

Common Era Sea-Level Change

Ben Horton, Klaus Bitterman, Andrew Kemp,and others

Maya Buchanan, Rachel DiSciullo,Michael Oppenheimer, Claudia Tebaldi

3

Roadmap

• What controls global and local sea-level change?

• How can we use our understanding of sea-level physics to interpret records of past and present changes, and what do they tell us for the future?

• How can we synthesize multiple lines of knowledge to assess the probability of different levels of sea-level change?

• What are the implications of sea-level change risk for flood risk?

• What are the costs, how might we manage them?

4

Factors controlling global and local sea-level change

Sources of global sea-level change (and uncertainty)

Milne et al. 2009 5

Sources of global sea-level change (and uncertainty)

Milne et al. 2009 5

1.1 ± 0.3 mm/yr 0.4 ± 0.1 mm/yr

Glaciers: 0.9 ± 0.4 mm/yrAntarctica: 0.3 ± 0.1 mm/yrGreenland: 0.3 ± 0.1 mm/yr

Contributions over 1993–2010 (IPCC AR5)

6

Total Land Ice Hazard

Lemke et al. (2007); Bamber et al. (2001); Lythe et al. (2001)Harig & Simons (2012, 2015, polarice.princeton.edu)

Non-polar glaciers and ice caps ~26 cm [10”]

Greenland & Antarctic glaciers and ice caps ~46 cm [18”]

Greenland Ice Sheet 7 m [23’]

West Antarctic Ice Sheet 5 m [16’]

East Antarctic Ice Sheet 52 m [171’]

http://polarice.princeton.edu

6

Total Land Ice Hazard

Lemke et al. (2007); Bamber et al. (2001); Lythe et al. (2001)Harig & Simons (2012, 2015, polarice.princeton.edu)

Non-polar glaciers and ice caps ~26 cm [10”]

Greenland & Antarctic glaciers and ice caps ~46 cm [18”]

Greenland Ice Sheet 7 m [23’]

West Antarctic Ice Sheet 5 m [16’]

East Antarctic Ice Sheet 52 m [171’]

2003-2013: 0.7 mm/yr [0.3”/decade]

to GMSL

2003-2014:0.3 mm/yr [0.1”/decade]

to GMSL

http://polarice.princeton.edu

Sources of regional sea-level change (and uncertainty)

472 NATURE GEOSCIENCE | VOL 2 | JULY 2009 | www.nature.com/naturegeoscience

REVIEW ARTICLE NATURE GEOSCIENCE DOI: 10.1038/NGEO544

each observing system (altimetry, Argo and GRACE), suggesting that this is the true error bound on trend estimates for these short (4 yr) time series. Variations may be partly a result of the slightly different time spans chosen and the dominant role of interannual variability over periods of only a few years, but there are also issues with each of the observing systems.

Problems with calibration of the temperature measurements were noted above, but a significant part of the imbalance arises from the incomplete temperature sampling of the ocean, particularly the Southern Ocean14, which may be insufficient before 2004 (refs 15 and 16). The development of innovative ways to reduce sampling bias17 is important. The GRACE mass estimates have a number of complications that contribute to their uncertainty. Because of the small signal over the oceans, compared with those over land, the analysis must reduce both the sampling of the nearby land signal

along the coasts18 and the presence of correlated errors in the GRACE solutions19,20. In addition, the GRACE mission is insensi-tive to geocentre motion, that is, the motion of the Earth’s centre of figure relative to the centre of mass of the whole Earth (including cryosphere, hydrosphere and atmosphere). Ignoring this contribu-tion can introduce an underestimate of up to 30% in sea-level rise caused by Greenland ice melting18. Estimates of geocentre motion derived from GRACE products or satellite laser ranging can be used in these analyses, but the accuracy of the trend in these estimates is difficult to obtain21. Finally, vertical motion of the ocean floor — due to glacial isostatic adjustment (the isostatic response of the solid Earth to past ice–ocean mass exchange) — makes a particularly large contribution to the measured gravity changes, with values used in recent analyses7,14,15 ranging from -1 to -2 mm yr-1 (water-equivalent mass change).

Sea level is measured in one of two ways: relative to the ocean floor (known as ‘relative sea level’) or relative to the Earth’s centre of mass (known as ‘absolute sea level’). Satellite altimetry is the only method that provides a measure of absolute sea level. Both relative and abso-lute sea level are affected by a wide variety of processes (panel a). Note that absolute sea level is affected indirectly by deformation of the solid Earth owing to the corresponding changes to the gravity field and volume of the global ocean basin. All of the processes depicted in a result in a spatially variable sea-level response.

Two climate-related processes that will have central roles in governing sea-level changes over the coming decades to centuries are land-ice melting (mass contribution; b) and ocean-water density change owing to temperature and salinity changes (steric contribution; c). The spatial variability associated with these processes is depicted in b and c.

It is generally assumed that when land ice melts, the associated sea-level rise is globally uniform and proportional to the volume

of ice loss. For example, it is often stated that the Greenland ice sheet holds about 7 m of global sea-level rise. In reality, the situation is more complex because of the isostatic deforma-tion of the solid Earth along with gravitational and rotational changes driven by the ice–ocean mass exchange27,85,98. Panel b shows model predictions of the change in global sea level if the Greenland (top) or West Antarctic (bottom) ice sheets were to lose mass at 1 mm yr-1 (10 cm per century) of global mean sea-level equivalent. The predicted response departs significantly from the mean with a reduced rise and even fall in areas close to the ablat-ing ice mass and an amplified rise in areas far removed from the melt source99.

Ocean temperature and salinity changes have also been regionally variable in the past, and estimates of the resulting sea-level change reflect this variability. Panel c shows the mean rates of sea-level change over the period 1950–2003 estimated from observations of ocean temperature (taken from ref. 100).

Box 1 | Processes affecting sea level.

Ocean–atmosphereinteraction

Terrestrial waterstorage

Ice melting

Vertical landmotion

Vertical landmotion

Density changes

Ocean circulation

0.0 0.2 0.4 0.6 0.8 1.0 1.1 1.2 1.3 1.4

mm yr–1mm yr–1–2.0 –1.5 –1.0 –0.5 0.0 0.5 1.0 1.5 2.0

a b

c

ngeo_544_JUL09.indd 472 18/6/09 12:47:27

Milne et al. 2009 7

modulated by static-equilbrium effects

Global Sea Level change is not the same as local sea level change

• Ocean dynamic effects• Mass redistribution effects: Gravitational, elastic and rotational• Natural and groundwater withdrawal-related sediment compaction• Long term: Isostasy and tectonics

Yin et al. (2009)

NATURE GEOSCIENCE DOI: 10.1038/NGEO462 LETTERS

20° N

40° N

60° N

20° N

40° N

60° N

20° N100° W 60° W 20° W 80° W 40° W 0°

100° W 60° W 20° W 80° W 40° W 0°

100° W 60° W 20° W 80° W 40° W 0°

40° N

60° N

20° N

40° N

60° N

20° N

40° N

60° N

20° N

40° N

60° N

0.1

¬0.1

¬0.3

¬0.5

–0.7

–0.9

¬1.1

¬1.3

0.1

¬0.1

¬0.3

¬0.5

–0.7

–0.9

¬1.1

¬1.3

0.4

0.3

0.2

0.1

0

¬0.1

¬0.2

¬0.3

¬0.4

0.4

0.3

0.2

0.1

0

¬0.1

¬0.2

¬0.3

¬0.4

0.4

0.3

0.2

0.1

0

¬0.1

¬0.2

¬0.3

¬0.4

0.4

0.3

0.2

0.1

0

¬0.1

¬0.2

¬0.3

¬0.4

a b

c d

e f

(m)

(m)

(m)

(m)

(m)

(m)

Figure 3 | Dynamic sea levels in the GFDL CM2.1. a, Observation11 (1992–2002). b, Simulation (1992–2002). c–e, Projected anomalies (2091–2100relative to 1981–2000) in the A2 (c), A1B (d) and B1 (e) scenarios. f, The dynamic sea-level change induced by an idealized 0.1 Sv freshwater input(water-hosing) into 50�–70� N of the Atlantic for 100 years (the mean of years 2091–2100 compared with the control). In the water-hosing run, radiativeforcing is kept constant at the 1990 level and the global mean SLR induced by the global ocean mass increase is removed. The AMOC weakens by 37%over 100 years.

scenario independent. The maximum dynamic SLR occurs east ofNewfoundland, with significant rises extending to the coastal regionnorth of Cape Hatteras.

The dynamic SLR is mainly a result of the cessation of the deepconvection and deep-water formation in the Labrador Sea, and theslowdown of the subpolar gyre. During 1981–2000, vigorous deepconvection occurs in the Labrador Sea, which can reach more than1,000m depth (see Supplementary Fig. S4). Owing to ocean surfacewarming and freshening, the deep convection in the Labrador Seashuts down by the end of the twenty-first century in all threescenarios. Compared with other sites, the deep convection in theLabrador Sea is very sensitive to the anomalies of the thermohalinefluxes5, which probably results from positive feedbacks operatingin this region12. The subpolar gyre weakens significantly witha northeastward shift of the barotropic (vertically independent)streamfunction pattern (see Supplementary Fig. S4). A fall of the

dynamic sea level in the subtropical gyre and aNorthAtlantic dipolepattern13,14 are also evident in Fig. 3c–e.

The dynamic SLR on the northeast coast of the United Statesis closely related to the horizontal gradient of the steric SLRand mass redistribution in the ocean (Fig. 4). In addition toglobal thermal expansion, the weakening of the formation andsouthward propagation of North Atlantic DeepWater causes a deepwarming and extra steric SLR along the route of the deep westernboundary current (Fig. 4a). From the maximum rise of about0.35m east of Newfoundland, the magnitude of this steric SLRreduces southward. In contrast, the steric SLR on the continentalshelf is small owing to the shallow water column. The sharp stericSLR gradient across the shelf break (near the zero contour linesin Fig. 4) cannot be balanced by geostrophic currents, thereforeleading to an increase in mass loading near the northeast coastof the United States (Fig. 4b). At Boston, New York City and

NATURE GEOSCIENCE | ADVANCE ONLINE PUBLICATION | www.nature.com/naturegeoscience 3

SSH, 1992-20028

The sea is higher off Bermuda than off

the northeastern U.S. by about 2 feet

because of atmosphere and

ocean dynamics that could weaken over

the century.


ME

9



ME-MI

MI

10

Not to scale!Farrell & Clark (1976), after Woodward (1888)


Kopp et al. (2015)

WAIS ~1.1x

Static-Equilibrium Fingerprints of Greenland and WAIS melting, per meter GSL rise


11


WAISb

EAISc

Median glaciersd

<-0.4 -0.2 0 0.2 0.4 0.6 0.8 1 1.2 1.4

GISa


12

Sea-level rise due to GIA (mm/y)


Kopp et al. (2015)

13

Records of past and present change

Example: Battery Tide Gauge, Battery Park, New York City

14

100 yr storm

10 yr storm

Sandy 13.9 ft

Donna 10 ftIrene9.5 ft Dec. 93

9.8 ftsea level rise

50 yr storm

Photo: New York Times (Jan. 14, 2014)

15 Calendar year

Glo

bal

mea

n s

ea lev

el (

mm

)

Rate (mm/yr)1901-1990

Rate (mm/yr)1993-2010

Accel. (mm/yr2)1901-END

KS 1.2 ± 0.2 3.0 ± 0.7 0.017 ± 0.003

GPR 1.1 ± 0.4 --- 0.009 ± 0.008

Church and White4 1.5 ± 0.2 2.9 ± 0.5 0.009 ± 0.002

Jevrejeva et al.3 1.9 3.7 0.011 ± 0.006

LETTERdoi:10.1038/nature14093

Probabilistic reanalysis of twentieth-centurysea-level riseCarling C. Hay1,2, Eric Morrow1,2, Robert E. Kopp2,3 & Jerry X. Mitrovica1

Estimating and accounting for twentieth-century global mean sea-level (GMSL) rise is critical to characterizing current and futurehuman-induced sea-level change. Several previous analyses of tidegauge records1–6—employing different methods to accommodate thespatial sparsity and temporal incompleteness of the data and to con-strain the geometry of long-term sea-level change—have concludedthat GMSL rose over the twentieth century at a mean rate of 1.6 to1.9 millimetres per year. Efforts to account for this rate by summingestimates of individual contributions from glacier and ice-sheet massloss, ocean thermal expansion, and changes in land water storage fallsignificantly short in the period before 19907. The failure to closethe budget of GMSL during this period has led to suggestions thatseveral contributions may have been systematically underestimated8.However, the extent to which the limitations of tide gauge analyseshave affected estimates of the GMSL rate of change is unclear. Herewe revisit estimates of twentieth-century GMSL rise using probabil-istic techniques9,10 and find a rate of GMSL rise from 1901 to 1990 of1.2 6 0.2 millimetres per year (90% confidence interval). Based onindividual contributions tabulated in the Fifth Assessment Report7

of the Intergovernmental Panel on Climate Change, this estimatecloses the twentieth-century sea-level budget. Our analysis, whichcombines tide gauge records with physics-based and model-derivedgeometries of the various contributing signals, also indicates thatGMSL rose at a rate of 3.0 6 0.7 millimetres per year between 1993and 2010, consistent with prior estimates from tide gauge records4.The increase in rate relative to the 1901–90 trend is accordinglylarger than previously thought; this revision may affect some pro-jections11 of future sea-level rise.

Tide gauges provide records of local sea-level changes that, in the caseof some sites, extend back to the eighteenth century12–14. However, usingthe database of tide gauge records15 to estimate historical GMSL rise(defined as the increase in ocean volume normalized by ocean area) ischallenging. Tide gauges sample the ocean sparsely and non-uniformly,with a bias towards coastal sites and the Northern Hemisphere, andwith few sites at latitudes greater than 60u (see, for example, refs 4, 9). Inaddition, tide gauge time series show significant inter-annual to decadalvariability, and they are characterized by missing data (that is, intervalswithout observations at the start, middle or end of a time series). Fromthe perspective of estimating GMSL changes, the data are contaminatedby local and regional signals due to ongoing glacial isostatic adjustment(GIA) associated with past ice ages16,17

, the spatially non-uniform pat-tern of sea-level rise associated with changes in contemporary land icesources18–21, ocean/atmosphere dynamics22, and other local factors in-cluding tectonics, sediment compaction, groundwater pumping andharbour development.

Different approaches have been used to address these complexities inefforts to estimate twentieth-century GMSL rise23. These include aver-aging rates at sites with the longest records1,2, averaging rates deter-mined from regional binning of records3, incorporating shorter recordsinto the analysis to distinguish between secular trends and decadal-scale variability3, and using altimetry records to determine dominant

sea-level geometries and then using tide gauge records to estimate thetime-varying amplitudes of these geometries4,5. In most cases, other cri-teria were applied to cull the tide gauge sites adopted in the analysis (forexample, excluding sites near tectonic activity or major urban centres).

Estimates of twentieth-century GMSL rise from these previous ana-lyses range from 1.6 to 1.9 mm yr21 (refs 1–6) and define an importantenigma. Independent model- and data-based estimates of the individualsources of GMSL, including mass flux from glaciers and ice sheets, ther-mal expansion of oceans, and changes in land water storage, are insuf-ficient to account for the GMSL rise estimated from tide gauge records8,particularly before 19907. For example, a tabulation of contributionsto GMSL rise from 1901 to 1990 in the Fifth Assessment Report (AR5;ref. 7) of the Intergovernmental Panel of Climate Change (IPCC) total0.5 6 0.4 mm yr21 (90% confidence interval, CI) less than a recent tidegauge derived rate of 1.5 6 0.2 mm yr21 (90% CI) estimated by Churchand White4 for the same period (the confidence range for this estimateis taken from AR5; refs 7 and 23). Using IPCC terminology, the lattersuggests that it is ‘extremely likely’ (probability P 5 95%) that GMSLrise from 1901 to 1990 was greater than 1.3 mm yr21, although thebottom-up sum of contributions is ‘likely’ (P . 67%) below this level.The above discrepancy has been attributed to underestimation of almostall possible sources: thermal expansion, glacier mass balance, and Green-land or Antarctic ice sheet mass balance7,8.

In this Letter, we revisit the analysis of GMSL since the start of thetwentieth century using Kalman smoothing9 (KS; see Methods). Thisstatistical technique naturally accommodates spatially sparse and tem-porally incomplete sampling of a global sea-level field, provides a rigor-ous, probabilistic framework for uncertainty propagation, and can correctfor a distribution of GIA and ocean models. We applied the approach toanalyse annual records from 622 tide gauges included in the PermanentService for Mean Sea Level (PSMSL) Revised Local Reference data-base15,24 and reconstruct the global field of sea-level change for eachyear from 1900 to 2010.

To examine the skill with which the KS reconstruction reproducesthe tide gauge observations, we compute the time series of residuals ateach tide gauge site and examine the distribution of the mean residual(that is, bias) for each site (Fig. 1a). The mean of the mean residualsacross all 622 observations is 0.3 mm, with a standard deviation of5.1 mm, indicating minimal systemic bias.

Comparing reconstructions and tide gauge observations at a selec-tion of individual sites (Fig. 1b–f) shows generally excellent agreement,although there are a small number of outliers. An example outlier is theChamplain tide gauge (Fig. 1f), which has a mean residual of 52 mm.This particular misfit (also evident at other sites in the vicinity) can beattributed to the St Lawrence being a regulated water system where flowis dominated by anthropogenic control rather than global-scale climatedynamics25. The eight sites that have mean residuals greater than 63s(15 mm) from the mean exhibit an average interannual sea-level vari-ability (estimated as the standard deviation after detrending the tidegauge observations) of 6130 mm, more than triple the mean inter-annual variability of 640 mm across all sites. Although these outliers

1Earth and Planetary Sciences, Harvard University, Cambridge, Massachusetts 02138, USA. 2Earth and Planetary Sciences, Rutgers University, Piscataway, New Jersey 08854, USA. 3Rutgers EnergyInstitute, Rutgers University, New Brunswick, New Jersey 08901, USA.

0 0 M O N T H 2 0 1 5 | V O L 0 0 0 | N A T U R E | 1

Macmillan Publishers Limited. All rights reserved©2015

Hay et al. (2015)

From 1901-1990, global mean sea level rose at ~1.2 ± 0.2 mm/yr (~0.5”/decade). From 1993-2010, it rose ~2.5x faster.

Example: Sediment cores from Newfoundland salt marshes

16Photo courtesy Ben Horton

17

Global sea level change over the Common Era

Kopp et al. (in rev.)

20th century global sea-level rise (about 6”) was extremely likely (P = 95%) faster than any of previous 26 centuries.

From 0-1900 CE, global sea level exhibited variability of about ±12 cm (±5”)

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

a ML2,1

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

b ML2,2

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

c ML1,1

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

d Gr

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

e NC

Year (CE)

Example: Fossil coral reefs from the Bahamas

http://www.mnstate.edu/leonard/G390BPHOTOS.htmlChen et al. (1991)18

http://www.mnstate.edu/leonard/G390BPHOTOS.html

The Last Interglacial (125 thousand years ago)was slightly warmer than today

Otto-Bliesner et al. (2013)CAPE (2006), NEEM (2013),

Jouzel et al. (2007)19

5

rsta.royalsocietypublishing.orgPhilTransRSocA371:20130097

......................................................

°C

1064210.01

–0.01–1–2–4–6–10

0

Figure 2. Reconstructed mean annual surface temperature (MAT) change for LIG from modern as reconstructed by Turney &Jones [7] and McKay et al. [8]. See text for description of methods.

3. DatasetsWe compare the climate model results to two recent LIG data syntheses [7,8]. Both compilationsare based on published records with quantitative estimates of mean annual surface temperaturechange. In the marine-only dataset of McKay et al. [8], seasonal anomalies are also available withthe overall pattern similar to the annual anomalies. The only available synthesis over land is forannual anomalies [7].

The Turney & Jones [7] global dataset is made up of 263 published records that spanthe LIG and have quantitative estimates of mean annual surface temperature. Three ice coretemperature estimates from δ18O are included from Greenland and four from East Antarctica.Marine records of mean annual sea surface temperature (SST) include those obtained fromforaminifera, radiolarian and diatom transfer functions and calibrations using Mg/Ca and Sr/Caratios and alkenone unsaturation indices (i.e. Uk′

37). Absolute dating of LIG proxy records isdifficult. Because of this, Turney & Jones average the temperature estimates across the isotopicplateau associated with the LIG in the marine and ice core records. The terrestrial mean annualsurface temperature estimates are based on the original interpretations from pollen, macrofossilsand Coleoptera, taking the period of maximum warmth and assuming that this warmth is broadlysynchronous with the marine and ice core plateaus. The pollen and macrofossil records areconverted to quantitative temperature estimates in the original publications using a variety ofmethods, including modern analogue, regressions and other inversions. Annual temperaturechanges are calculated from modern using the present-day mean annual temperatures (MATs)at each palaeosite location using the 1961–1990 CRU dataset [34] for terrestrial sites and ESRLdataset [35] for ocean sites.

The McKay et al. [8] global dataset is a compilation of 76 records from publishedpalaeoceanographic sites. The annual SST estimates are from Mg/Ca in foraminifera, alkenoneunsaturation ratios Uk′

37, and faunal assemblage transfer functions for radiolaria, foraminifera,diatoms and coccoliths. Only records with published age models and an average temporalresolution of 3 kyr or better for the LIG and Late Holocene were included. Because of datinguncertainties, McKay et al. average the SST estimates for a 5000-year period centred on thewarmest temperatures between 135 and 118 ka at each site. SST changes are calculated from theLate Holocene (last 5 kyr). This compilation was additionally supplemented with the 94 CLIMAPProject LIG SST change from core-top values.

The combined datasets, which include some overlap over the oceans, give a broadly consistentglobal synthesis of global MAT change (figure 2). Annual surface temperatures were warmer thanmodern at mid- and high latitudes of both hemispheres. The data indicate strong warming in

on July 19, 2015http://rsta.royalsocietypublishing.org/Downloaded from

Land Ocean Land+Ocean

Global ~1.7°C [3.1°F] ~0.8°C [1.4°F] ~1.0°C [1.8°F]

Greenland ~5-8°C [9-14°F]

Antarctica ~3-5°C [5-9°F]

Temperatures with respectto pre-Industrial Holocene

Global mean sea level reconstructed for the Last Interglacial

Kopp et al. (2009, Nature; 2013, GJI)20

Extremely likely (95% probability) peaked at least 6.4 m [21 feet] higher than today, unlikely (1-in-3 chance) to have been higher than 8.8 m [29 feet]

Glo

ba

l S

ea

Le

ve

l (m

)

a

−40

−30

−20

−10

0

10

rate

of

GS

L c

ha

ng

e (

m/k

y) b

−40

−20

0

20

40

age (ka)

NH

ice

sh

ee

ts lo

ss (

m E

SL

) c

110 115 120 125 130 135−40

−30

−20

−10

0

10

age (ka)

SH

ice

sh

ee

ts lo

ss (

m E

SL

) d

110 115 120 125 130 135−40

−30

−20

−10

0

10

115 120 125 130

Age (thousand years ago)

21

6 m [21 feet] of committed sea-level rise will cause loss of areas currently home to ~380 million people – but we don’t know whether it’ll take centuries or millennia

22

Assessing future sea-level risk

31 Jan. 2002Larsen B Ice Shelf

http://earthobservatory.nasa.gov/IOTD/view.php?id=2288 ~40 km23

It’s still difficult for physical models to capture some key dynamics of ice sheet behavior.

http://earthobservatory.nasa.gov/IOTD/view.php?id=2288

17 Feb. 2002


Larsen B Ice Shelf



23 Feb. 2002

Larsen B Ice Shelf




5 Mar. 2002

Larsen B Ice Shelf




7 Mar. 2002

Larsen B Ice Shelf




7 Mar. 2002

Larsen B Ice Shelf



Some instabilities and feedbacks under very active research:

• marine ice sheet instability• ice cliff instability• higher temperature => darker snow• greater snowfall => greater ice sheet discharge• ice sheet melt => reduced near-field sea-level• ice sheet melt => cooler surface waters => expanded sea ice => cooler

air temperatures => reduced snow fall


One alternative approach: Semi-empirical models look at past relationship between temperature, GSL

28 Kopp et al. (in rev.)

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

a ML2,1

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

b ML2,2

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

c ML1,1

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

d Gr

−500 0 500 1000 1500 2000

−200

−100

0

100

GS

L (

mm

)

e NC

-0.8-0.6-0.4-0.2

00.20.40.6

Tem

p. a

nom

aly

t°C

i

-10

-5

0

5

10

15

20

25

GSL

tcm

i

-400 -200 0 200 400 600 800 1000 1200 1400 1600 1800 2000-10

-5

0

5

10

15

20

25

GSL

tcm

i

Year CE1800 1850 1900 1950 2000 2050 2100

-101030507090

110130

GSL

tcm

i

1 2 H

bi

RC

P 2.

6

RC

P 8.

5R

CP

4.5

di

Mann et al. 2009 calibrationMarcott et al. 2013 calibration

GSL

Maximal uncertainty for all RCPs

RCP 2.6RCP 4.5RCP 8.5

ci

ai

-0.8-0.6-0.4-0.2

00.20.40.6

Tem

p. a

nom

aly

t°C

i-10

-5

0

5

10

15

20

25

GS

L tc

mi

-400 -200 0 200 400 600 800 1000 1200 1400 1600 1800 2000-10

-5

0

5

10

15

20

25

GS

L tc

mi

Year CE1800 1850 1900 1950 2000 2050 2100

-101030507090

110130

GS

L tc

mi

1 2 H

bi

RC

P 2

.6

RC

P 8

.5R

CP

4.5

di


GSL



ci

ai One alternative approach: Semi-empirical models

look at past relationship between temperature, GSL


Projected SLR by 2100 (90% probability range):

20”-48” (50–121 cm)under RCP 8.5 (high emissions)

10”-24” (25–61 cm)under RCP 2.6 (low emissions)

-0.8-0.6-0.4-0.2

00.20.40.6

Tem

p. a

nom

aly

t°C

i-10

-5

0

5

10

15

20

25

GS

L tc

mi

-400 -200 0 200 400 600 800 1000 1200 1400 1600 1800 2000-10

-5

0

5

10

15

20

25

GS

L tc

mi

Year CE1800 1850 1900 1950 2000 2050 2100

-101030507090

110130

GS

L tc

mi

1 2 H

bi

RC

P 2

.6

RC

P 8

.5R

CP

4.5

di


GSL



ci

ai One alternative approach: Semi-empirical models

look at past relationship between temperature, GSL


Projected SLR by 2100 (90% probability range):

20”-48” (50–121 cm)under RCP 8.5 (high emissions)

10”-24” (25–61 cm)under RCP 2.6 (low emissions)

But: semi-empirical models project global (not local) changes and are calibrated from a time period when thermal expansion (and, regionally, ocean dynamics) dominated sea-level change

Since we can’t yet rely on physical models, we need to synthesize multiple lines of knowledge

30 Kopp et al. (2014)

Earth’s Future

Probabilistic 21st and 22nd century sea-level projectionsat a global network of tide-gauge sitesRobert E. Kopp1, Radley M. Horton2, Christopher M. Little3, Jerry X. Mitrovica4,Michael Oppenheimer3, D. J. Rasmussen5, Benjamin H. Strauss6, and Claudia Tebaldi6,7

1Department of Earth & Planetary Sciences, Rutgers Energy Institute, and Institute of Marine & Coastal Sciences,Rutgers University, New Brunswick, New Jersey, USA, 2Center for Climate Systems Research, Columbia University,New York, New York, USA, 3Woodrow Wilson School of Policy & International Affairs and Department of Geosciences,Princeton University, Princeton, New Jersey, USA, 4Department of Earth & Planetary Sciences, Harvard University,Cambridge, Massachusetts, USA, 5Rhodium Group, Oakland, California, USA, 6Climate Central, Princeton, New Jersey,USA, 7National Center for Atmospheric Research, Boulder, Colorado, USA

Abstract Sea-level rise due to both climate change and non-climatic factors threatens coastal settle-ments, infrastructure, and ecosystems. Projections of mean global sea-level (GSL) rise provide insufficientinformation to plan adaptive responses; local decisions require local projections that accommodate dif-ferent risk tolerances and time frames and that can be linked to storm surge projections. Here we presenta global set of local sea-level (LSL) projections to inform decisions on timescales ranging from the com-ing decades through the 22nd century. We provide complete probability distributions, informed by acombination of expert community assessment, expert elicitation, and process modeling. Between theyears 2000 and 2100, we project a very likely (90% probability) GSL rise of 0.5–1.2 m under representa-tive concentration pathway (RCP) 8.5, 0.4–0.9 m under RCP 4.5, and 0.3–0.8 m under RCP 2.6. Site-to-sitedifferences in LSL projections are due to varying non-climatic background uplift or subsidence, oceano-graphic effects, and spatially variable responses of the geoid and the lithosphere to shrinking land ice. TheAntarctic ice sheet (AIS) constitutes a growing share of variance in GSL and LSL projections. In the globalaverage and at many locations, it is the dominant source of variance in late 21st century projections,though at some sites oceanographic processes contribute the largest share throughout the century. LSLrise dramatically reshapes flood risk, greatly increasing the expected number of “1-in-10” and “1-in-100”year events.

1. Introduction

Sea-level rise figures prominently among the consequences of climate change. It impacts settlementsand ecosystems both through permanent inundation of the lowest-lying areas and by increasing thefrequency and/or severity of storm surge over a much larger region. In Miami-Dade County, Florida, forexample, a uniform 90-cm sea-level rise would permanently inundate the residences of about 5% of thecounty’s population, about the same fraction currently threatened by the storm tide of a 1-in-100 yearflood event [Tebaldi et al., 2012]. A 1-in-100 year flood on top of such a sea-level rise would, assuming geo-graphically uniform flooding, expose an additional 35% of the population (Climate Central, Surging Seas,2013, retrieved from SurgingSeas.org, updated November 2013).

The future rate of mean global sea-level (GSL) rise will be controlled primarily by the thermal expansionof ocean water and by mass loss from glaciers, ice caps, and ice sheets [Church et al., 2013]. Changes inland water storage, through groundwater depletion and reservoir impoundment, may have influencedtwentieth-century sea-level change [Gregory et al., 2013] but are expected to be relatively minor contribu-tors compared to other factors in the current century [Church et al., 2013].

Local sea-level (LSL) change can differ significantly from GSL rise [Milne et al., 2009; Stammer et al., 2013],so for adaptation planning and risk management, localized assessments are critical. The spatial variabilityof LSL change arises from: (1) non-uniform changes in ocean dynamics, heat content, and salinity [Lev-ermann et al., 2005; Yin et al., 2009], (2) perturbations in the Earth’s gravitational field and crustal height(together known as static-equilibrium effects) associated with the redistribution of mass between the

RESEARCH ARTICLE10.1002/2014EF000239

Key Points:• Rates of local sea-level rise differs

from rate of global sea-level rise• Differences arise from land motion,

ocean dynamics, and Antarctic massbalance

• Local sea-level rise can dramaticallyincrease flood probabilities

Supporting Information:• EFT2_37 Supp Info.pdf• EFT2_37 Table S06.tsv• EFT2_37 Table S05.tsv• EFT2_37 Table S09.tsv• EFT2_37 Table S08.tsv• EFT2_37 Table S07.tsv• EFT2_37 Project Code.zip• README.txt

Corresponding author:R. E. Kopp, [email protected]

Citation:Kopp, R. E., R. M. Horton, C. M. Little, J. X.Mitrovica, M. Oppenheimer, D. J.Rasmussen, B. H. Strauss, and C. Tebaldi(2014), Probabilistic 21st and 22ndcentury sea-level projections at a globalnetwork of tide-gauge sites, Earth’sFuture, 2, 383–406,doi:10.1002/2014EF000239.

Received 3 FEB 2014Accepted 6 JUN 2014Accepted article online 13 JUN 2014Published online 21 AUG 2014

This is an open access article underthe terms of the Creative CommonsAttribution-NonCommercial-NoDerivsLicense, which permits use and distri-bution in any medium, provided theoriginal work is properly cited, the useis non-commercial and no modifica-tions or adaptations are made.

KOPP ET AL. © 2014 The Authors. 383

Expert assessment

Climate models

Continuation ofhistorical trends

Ice sheets

Glaciers

Non-climaticbackground

Land waterstorage

Thermal expansionand ocean dynamics

Localsea level

Physical modelof local effects

Expert elicitation

Historical demand/populationrelationship

Projected GMSL rise and sources of uncertainty

Kopp et al. 2014

KOPP ET AL.: SEA-LEVEL PROJECTIONS 18

2000 2050 2100 2150 2200

0

0.5

1

1.5

2

2.5

3

m G

SL r

ise

RCP 8.5

RCP 4.5

RCP 2.6

Figure 3. Projections of GSL rise for the three RCPs.Heavy = median; dashed = 5th–95th percentile, dotted= 0.5th–99.5th percentiles.

2020 2040 2060 2080 21000

0.01

0.02

0.03

0.04

0.05

0.06 GSL

m2

LWSGICTEGISAIS

2020 2040 2060 2080 21000

0.2

0.4

0.6

0.8

1

Frac

tion

of v

aria

nce

2020 2040 2060 2080 21000

0.02

0.04

0.06

0.08

0.1

0.12

0.14

m2

BkgdLWSGICOceanGISAIS

2020 2040 2060 2080 21000

0.2

0.4

0.6

0.8

1

Frac

tion

of v

aria

nce

a b

c d

New York

Figure 4. Sources of variance in raw (a,c) and fractionalterms (b,d), globally (a-b) and at New York City (c-d)in RCP 8.5. AIS: Antarctic ice sheet; GIS: Greenlandice sheet; TE: thermal expansion; Ocean: oceanographicprocesses; GIC: glaciers and ice caps; LWS: land waterstorage; Bkgd: local background e↵ects.

GMSL rise from 2000 to: Likely (17-83%) 1-in-20 (95%) 1-in-200 (99.5%) Max. poss. (99.9%)

2100, RCP 8.5 (high emissions)

24”-39”(62-100 cm)

47” (121 cm)

69” (176 cm)

(69”)

96” (245 cm)

2100, RCP 2.6 (low emissions)

14”-26” (37-65 cm)

32” (82 cm)

56” (141 cm)

83” (210 cm)

2020 2040 2060 2080 21000

0.01

0.02

0.03

0.04

0.05

0.06GSL variance (RCP 8.5)

m2

LWS

GIC

TE

GIS

AIS

2020 2040 2060 2080 21000

0.2

0.4

0.6

0.8

1GSL variance (RCP 8.5)

Fra

ctio

n o

f va

riance

Antarctic Ice Sheet

thermal expansion

Greenland Ice Sheet

glaciers

LWS

Broad agreement with semi-empirical estimates based on the Common Era record – which either increases confidence in projections or raises concern that expert PDFs are overly narrow

Kopp et al. 2014 and Kopp et al., in rev.

GMSL rise from 2000 to: Likely (17-83%) 1-in-20 (95%) 1-in-200 (99.5%) Max. poss. (99.9%)

2100, RCP 8.5 (high emissions)

24”-39”(62-100 cm)

47” (121 cm)

69” (176 cm)

(69”)

96” (245 cm)

semi-empirical23”-40”

(58–101 cm)48”

(123 cm)

2100, RCP 2.6 (low emissions)

14”-26” (37-65 cm)

32” (82 cm)

56” (141 cm)

83” (210 cm)

semi-empirical11”-20”

(29–51 cm)24”

(61 cm)

Local sea-level rise projections show significant spatial variability

33Kopp et al. 2014

Median projection: RCP 8.5GSL = 0.79 m

<0

0.25

0.50

0.75

1.00

1.25

>1.50

Projection (17%−83%): RCP 8.5

0.40

0.45

0.50

0.55

0.60

0.65

0.70

a

b

34

2000 2050 2100 2150 22000

0.5

1

1.5

2

2.5

3

m G

SL

ris

e

RCP 8.5

RCP 4.5

RCP 2.6

Kopp et al. (2014)

Local sea-level rise projections show significant spatial

variability

by 2100, RCP 8.5 Likely 1-in-20 1-in-200

Globally 24”-39” (62-100 cm) 47” (121 cm) 69” (176 cm)

New York, NY 26”-50” (64-128 cm) 61” (154 cm) 83” (212 cm)

35

2000 2050 2100 2150 22000

0.5

1

1.5

2

2.5

3

m G

SL

ris

e

RCP 8.5

RCP 4.5

RCP 2.6

Kopp et al. (2014)


Globally 14”-26” (37-65 cm) 32” (82 cm) 56” (141 cm)


Aggressively reducing greenhouse gas emissions can

shave about 1’-1.5’ off of projected 21st century sea-level

rise


Globally 24”-39” (62-100 cm) 47” (121 cm) 69” (176 cm)


36

Implications for coastal flood risk

37

NN

Ship Bottom, NJ

2008(Courtesy Prof. Ken Miller)

October 31, 2012

38

Climate Central (2013)http://sealevel.climatecentral.org/

Coastal flooding

Areas submerged with 9’ sea level rise plus storm surge (= Sandy today, 1-in-10 year flood w/5’ sea-

level rise)

Below 9’ in New York

• $168 billion property• 930 thousand people

(1/3 high social vulnerability)

• 405 thousand housing units

• 65 fire and EMS stations• 28 hospitals

http://sealevel.climatecentral.org/

39

Even with normal distributions, relationships between changes in mean and changes in number of extremes is near exponential

-3 -2 -1 0 1 2 3 4 50

0.1

0.2

0.3

0.4

Probability of event > 4σ:baseline: 0.003%

40


-3 -2 -1 0 1 2 3 4 50

0.1

0.2

0.3

0.4

Probability of event > 4σ:baseline: 0.003%+1σ mean: 0.13%

41


-3 -2 -1 0 1 2 3 4 50

0.1

0.2

0.3

0.4

Probability of event > 4σ:baseline: 0.003%+1σ mean: 0.13%+2σ mean: 2.3 %

42

Extreme value statisticsExpected number of floods per year at the Battery, New York City (1920-2013) based on maximum-likelihood Generalized Pareto Distribution fit to observed extremes

Expe

cted

floo

ds/y

ear

0.001

0.01

0.1

1

Storm tide height (m MHHW)

0.5 1 1.5 2 2.5 3

(1)

Supe

rsto

rm S

andy

(2)

Hur

rica

ne D

onna

(3)

12/9

2 N

or’e

aste

r

(9)

H. G

lori

a

HurricaneNor’easterHybrid

43

Expected number of flood events changes significantly with SLRExpected number of floods per year at the Battery, New York City

Expe

cted

floo

ds/y

ear

0.001

0.01

0.1

1


0.5 1 1.5 2 2.5 3

+0.5 m

~6x

Frequency increases labeled are for the new 1-in-10 year event.

44

Expected number of flood events changes significantly Expected number of floods per year at the Battery, New York City

Expe

cted

floo

ds/y

ear

0.001

0.01

0.1

1


0.5 1 1.5 2 2.5 3

+0.5 m+1.0 m

+1.5 m+2.0 m

~6x ~22x ~64x

Frequency increases labeled are for the new 1-in-10 year event.

~150x

45

Under uncertainty, expectation heavily skews toward high end – even if SLR projection is symmetric

Expe

cted

floo

ds/y

ear

0.001

0.01

0.1

1


0.5 1 1.5 2 2.5 3

+0.5 m

+1.0 m

+1.5 m

50% 0.5 m,50% 1.5 m

~45x

46

Sea-level rise allowances reflect height adjustment needed to maintain constant flood risk, given uncertainty.

Expe

cted

floo

ds/y

ear

0.001

0.01

0.1

1


0.5 1 1.5 2 2.5 3

+0.5 m

+1.0 m

+1.5 m

50% 0.5 m,50% 1.5 m

SLR Allowance

Buchanan et al. (in prep.)

46

Sea-level rise allowances reflect height adjustment needed to maintain constant flood risk, given uncertainty.

Expe

cted

floo

ds/y

ear

0.001

0.01

0.1

1


0.5 1 1.5 2 2.5 3

+0.5 m

+1.0 m

+1.5 m

50% 0.5 m,50% 1.5 m

SLR Allowance


Sea-level rise allowances depend both upon acceptable risk of flooding and upon timeframe (instantaneously in 2030, 2050, 2100, or integrated over the duration of a project life, e.g., 2020-2050, 2020-2100)

47

The relevant time frame is often longer than we might think.

Marion Power Station, Jersey City, NJBuilt by the Public Service Electric Co. in 1905

Retired as generating station in 1961;Succeeded by Hudson Generating Station

Marion Switching Station flooded in 2011during Irene and 2012 during Sandy

1911

2012

48

The relevant time frame is often longer than we might think.

Major Switching Stations Flooded During Sandy(Red = close to location on 1911 map)

1. Bayonne2. Bayway [Elizabeth]3. Deans [North Brunswick]4. Essex [Newark]5. Federal Square [Newark]6. Hudson [Jersey City]7. Jersey City8. Kearny9. Linden10. Marion [Jersey City]11. Metuchen12. Newark13. Sewaren14. South Waterfront [Jersey City]

Source: PSE&G (2013)

http://www.pjm.com/~/media/committees-groups/stakeholder-meetings/grid-2020-insights/20130417-distributed-technology-panel-3-calore.ashx

49

Another complication: deep uncertainty


This is one thoughtfully produced, self-consistent probability distribution, drawing upon the best available knowledge – but it is

not the only justifiable probability distribution.

2000 2020 2040 2060 2080 2100

Sea le

vel (

m)

0

0.5

1

1.5

2The Battery, NY

2000 2020 2040 2060 2080 2100

m

0

0.5

1

1.5Instantaneous

1.0%10.0%0.2%

2020 2040 2060 2080 2100

m

0

0.2

0.4

0.6

0.8Integrated starting 2020

2020 2030 2040 2050 2060 2070

m

0.2

0.4

0.6

0.8

1

1.2Integrated 30 years

2000 2020 2040 2060 2080 2100

m

0

1

2

3

4Instantaneous (1.0%)

β = 0.00β = 0.50β = 0.67β = 0.90β = 0.95β = 0.99β = 1.00

2020 2040 2060 2080 2100

m

0

0.5

1

1.5Integrated starting 2020 (1.0%)

50

The robust decision making under uncertainty literature provides techniques for dealing with deep uncertainty.


2000 2020 2040 2060 2080 2100

Sea le

vel (

m)

0

0.5

1

1.5

2The Battery, NY

2000 2020 2040 2060 2080 2100

m

0

0.5

1

1.5Instantaneous

1.0%10.0%0.2%

2020 2040 2060 2080 2100

m

0

0.2

0.4

0.6


2020 2030 2040 2050 2060 2070

m

0.2

0.4

0.6

0.8

1


2000 2020 2040 2060 2080 2100

m

0

1

2

3


β = 0.00β = 0.50β = 0.67β = 0.90β = 0.95β = 0.99β = 1.00

2020 2040 2060 2080 2100

m

0

0.5

1


POT thresholds are set via trial and error to accurately approximate the Poisson distribution ofactual extreme outliers – high enough such that the distribution of selected extremes converges to theGPD asymptote (improving validity), yet low enough to extract enough sample points (improvingreliability). Following (author?) (Tebaldi 2012), we use a threshold µ equal to the hourly waterlevel’s 99th percentile.

[Need call out to Figure 1 here.]

3.2 Sea-level rise projections

To estimate the time-varying probability of sea-level rise, we employ the local sea-level rise PDFs of(author?) (Kopp 2014) for Representative Concentration Pathway (RCP) 8.5. For each RCP, Koppet al. produced a 10,000 member Monte Carlo ensemble of time series, which we employ directlyin our analysis. Kopp et al’s median SLR between 2000 and 2100 under RCP 8.5 is illustrated inFigure 1b.

The (author?) (Kopp 2014) projections provide one plausible, self-consistent set of local sea-level rise PDFs, but they are not the only plausible PDFs. To accommodate imperfect confidence inthese PDFs, we adapt the Limited Degrees of Confidence (LDC) criterion used in decision-makingunder uncertainty (Froyn 2005; McInerney et al. 2012). Taking P (�, t) from Kopp et al, we definethe LDC effective probability as

˜

P (�, t) = �P (�, t) + (1� �)�(��

t,WC) (7)

Here, � 2 [0, 1] is a measure of confidence in P (�, t), �t,WC is a worst-case projection at time t, and

� is the Dirac delta function (�(0) = 1, � = 0 otherwise). For �t,WC, we adopt the 99.9th percentile

projections of (Kopp 2014), which are comparable to other estimates of physically-plausible worst-case projections available in the literature. It follows that

N

e,LDC

(z, t,�) = �N

e

(�, t) + (1� �)N(z ��

t,WC) (8)

˜

N

e,LDC

(z, t1, t2,�) =1

t2 � t1

Zt2

t1

N

e,LDC

(z, t) (9)

4 Results

Figure 2 illustrates the calculation of SLR allowances, using Boston as an example. Moderateflooding at Boston has been proportionally more common than extreme flooding, giving rise topositive shape factor (⇠ = 0.074) and causing the flexure of the logN(z) vs z curve (black line).Incrementing sea-level rise by a fixed amount (e.g., the mean [Important: Expected is notmedian.] estimate for 2100, N +E(SL2100)) shifts the return period curve to the right by about ameter less than by shifting N(z) to the instantaneous allowance in 2100 accounting for uncertainty(N

e

(2100)). [As importantly as the amount, it is no longer parallel. I would probablystart by talking about the 2050 numbers, as the effect is more subtle there.] Thehorizontal difference between historic N and N

e

(2020� 2100) (0.7 m for N0 = 1%) is the DL-SLRallowance for this 80-year period. [Note: actually a 81-year period. Not sure if we wantto change the period to 2020-2099 or 2021-2100, or whether we just want to calli t a81-year period.]

[Tonal shift need here. We don’t talk about what we provide – we want to use ourresults to make specific points about the different sorts of allowances.] We provide a rangeof SLR allowances appropriate for different decision-contexts (Figure 3 and in the supplemental

5

Limited Degree of Confidence metric

where β is a measure of confidence in the expert PDF, and ΔWC is a worst-case projection (here taken as the 99.9th percentile)

β 2100 A 2020-2100 A

1.00 57” (145 cm) 24” (62 cm)

0.90 92” (234 cm) 36” (91 cm)

0.67 104” (265 cm) 43” (110 cm)

0.00 119” (301 cm) 53” (135 cm)

50

The robust decision making under uncertainty literature provides techniques for dealing with deep uncertainty.


2000 2020 2040 2060 2080 2100

Sea le

vel (

m)

0

0.5

1

1.5

2The Battery, NY

2000 2020 2040 2060 2080 2100

m

0

0.5

1

1.5Instantaneous

1.0%10.0%0.2%

2020 2040 2060 2080 2100

m

0

0.2

0.4

0.6


2020 2030 2040 2050 2060 2070

m

0.2

0.4

0.6

0.8

1


2000 2020 2040 2060 2080 2100

m

0

1

2

3


β = 0.00β = 0.50β = 0.67β = 0.90β = 0.95β = 0.99β = 1.00

2020 2040 2060 2080 2100

m

0

0.5

1


POT thresholds are set via trial and error to accurately approximate the Poisson distribution ofactual extreme outliers – high enough such that the distribution of selected extremes converges to theGPD asymptote (improving validity), yet low enough to extract enough sample points (improvingreliability). Following (author?) (Tebaldi 2012), we use a threshold µ equal to the hourly waterlevel’s 99th percentile.

[Need call out to Figure 1 here.]

3.2 Sea-level rise projections

To estimate the time-varying probability of sea-level rise, we employ the local sea-level rise PDFs of(author?) (Kopp 2014) for Representative Concentration Pathway (RCP) 8.5. For each RCP, Koppet al. produced a 10,000 member Monte Carlo ensemble of time series, which we employ directlyin our analysis. Kopp et al’s median SLR between 2000 and 2100 under RCP 8.5 is illustrated inFigure 1b.

The (author?) (Kopp 2014) projections provide one plausible, self-consistent set of local sea-level rise PDFs, but they are not the only plausible PDFs. To accommodate imperfect confidence inthese PDFs, we adapt the Limited Degrees of Confidence (LDC) criterion used in decision-makingunder uncertainty (Froyn 2005; McInerney et al. 2012). Taking P (�, t) from Kopp et al, we definethe LDC effective probability as

˜

P (�, t) = �P (�, t) + (1� �)�(��

t,WC) (7)

Here, � 2 [0, 1] is a measure of confidence in P (�, t), �t,WC is a worst-case projection at time t, and

� is the Dirac delta function (�(0) = 1, � = 0 otherwise). For �t,WC, we adopt the 99.9th percentile

projections of (Kopp 2014), which are comparable to other estimates of physically-plausible worst-case projections available in the literature. It follows that

N

e,LDC

(z, t,�) = �N

e

(�, t) + (1� �)N(z ��

t,WC) (8)

˜

N

e,LDC

(z, t1, t2,�) =1

t2 � t1

Zt2

t1

N

e,LDC

(z, t) (9)

4 Results

Figure 2 illustrates the calculation of SLR allowances, using Boston as an example. Moderateflooding at Boston has been proportionally more common than extreme flooding, giving rise topositive shape factor (⇠ = 0.074) and causing the flexure of the logN(z) vs z curve (black line).Incrementing sea-level rise by a fixed amount (e.g., the mean [Important: Expected is notmedian.] estimate for 2100, N +E(SL2100)) shifts the return period curve to the right by about ameter less than by shifting N(z) to the instantaneous allowance in 2100 accounting for uncertainty(N

e

(2100)). [As importantly as the amount, it is no longer parallel. I would probablystart by talking about the 2050 numbers, as the effect is more subtle there.] Thehorizontal difference between historic N and N

e

(2020� 2100) (0.7 m for N0 = 1%) is the DL-SLRallowance for this 80-year period. [Note: actually a 81-year period. Not sure if we wantto change the period to 2020-2099 or 2021-2100, or whether we just want to calli t a81-year period.]

[Tonal shift need here. We don’t talk about what we provide – we want to use ourresults to make specific points about the different sorts of allowances.] We provide a rangeof SLR allowances appropriate for different decision-contexts (Figure 3 and in the supplemental

5

Limited Degree of Confidence metric

where β is a measure of confidence in the expert PDF, and ΔWC is a worst-case projection (here taken as the 99.9th percentile)

Note the pull of the tail: β = 0.67 gives an allowance closer to that of the worst case than to that for full confidence in the expert PDF

β 2100 A 2020-2100 A

1.00 57” (145 cm) 24” (62 cm)

0.90 92” (234 cm) 36” (91 cm)

0.67 104” (265 cm) 43” (110 cm)

0.00 119” (301 cm) 53” (135 cm)

51

Additional complication: potential changes in the distribution of tropical cyclones

Emanuel (2013) Kopp et al. (in prep.)

stratification of the upper ocean is not considered here. Earlierwork showed that order 10% increases in ocean mixed layerthickness have very minor effects on tropical cyclone activity (18),but future work will incorporate modeled changed in upper oceanthermal structure.When driven by National Center for Atmospheric Research/

National Centers for Environmental Prediction (NCAR/NCEP)reanalyses during the period 1980–2006, this downscaling tech-nique with a constant rate of random seeding produces resultsthat explain as much of the observed variance and trends inNorth Atlantic tropical cyclone activity as do certain globalmodels (11, 19) and the regional downscaling model of Knutsonet al. (20), which was also driven by NCAR/NCEP reanalysisdata. The technique captures well the observed spatial and sea-sonal variability of tropical cyclones around the globe, as well asthe effects of such climate phenomena as El Nino SouthernOscillation (ENSO) and the Atlantic Meridional Mode (9).When driven by reanalysis data that extend over much of the20th century, the downscaling technique captures observedtrends in storm frequency in all ocean basins except for theeastern North Pacific (21). Thus, there are objective reasons tohave some confidence in the ability of the downscaling techniqueto simulate the effects of climate and climate change on tropicalcyclone activity. An important advantage of this technique overexplicit simulation with global and regional models is that itshigh resolution of the storm core allows it to capture the fullintensity spectrum of real storms.Our downscaling technique requires a single global calibra-

tion of the rate of seeding. Here we calibrate the seeding rateused by each model so as to produce 80 events globally withmaximum 1-min winds at 10 m altitude exceeding 40 knots (kn),averaged over the historical period 1950–2005. Because some ofthe events included in our dataset have maximum winds less than40 kn, the total storm frequencies shown here may have 1950–2005 averages slightly larger than 80. It should also be noted that,in contrast to Emanuel et al. (9), we downscale each year ofmodel data separately. We ran 600 events per year globally,for each of the years in the span 1950–2100, using historicalsimulations for the period 1950–2005 and the RepresentativeConcentration Pathway 8.5 (RCP8.5) scenario for the period2006–2100. This large number of events keeps the strictly random(Poisson) interannual variability of global storm counts at lessthan 5%.

ResultsFig. 1 shows a box plot of the global frequency of downscaledtropical cyclones, averaging each simulation over 10-y blocks. Anincrease in global mean frequency during roughly the first threequarters of the 21st century is indicated, with a total increase inthe range of 10–40%. Fig. 2, displaying the change in trackdensity averaged over the five models, shows that most of theincrease in frequency is in the North Pacific, but with substantialincreases in the North Atlantic and South Indian oceans as well.One distinct advantage of our downscaling technique is that it

captures the full spectrum of storm intensity (17), in contrast withdirect global model simulations, which truncate the high-intensityevents (11) that do a disproportionate amount of total tropicalcyclone damage (22, 23). One convenient measure of tropicalcyclone intensity is the power dissipation index, an estimate of thetotal amount of kinetic energy dissipated by tropical cyclones overtheir lifetimes (24). The power dissipation index is the integralover the lifetime of the storm of its maximum surface wind cubed.Here we also accumulate global power dissipation over each 10-yblock from 1950 to 2100 and display the result in Fig. 3. Averagedover the six models, power dissipation increases by about 45%over the 21st century. Of this increase, very nearly half comesfrom the increase in the frequency of events discussed previously;the other half comes from an increase in the cube of the surfacewinds. This is reflected in a 40% increase globally in hurricanes ofSaffir–Simpson category 3 and higher.The spatial distribution of the increase in power dissipation is

illustrated in Fig. 4. Consistent with the increase in track density,most of the increase in power dissipation is in the North Pacific,but with significant increases in the western part of the NorthAtlantic and in the South Indian Ocean as well. Averaged overthe six models, the power dissipation at landfall increases byabout 55% over the 21st century, consistent with the increase inbasin-wide power dissipation. (Landfall power dissipation is de-fined at the cube of the surface winds at the last 2-h snapshot ofa tropical cyclone before landfall. Landfall is defined in terms of1/4 × 1/4 degree bathymetry.)Overall, these results project substantial increases in tropical

cyclone activity under the RCP8.5 emissions pathway, at least forthe six models used here. In the next section, these results areanalyzed and compared and contrasted with previous work.

Analysis and Comparison with Previous WorkAlthough the physics underlying the frequency of tropical cyclo-genesis are not well understood, several indices have been de-veloped that empirically relate observed tropical cyclogenesis ratesto environmental variables thought to be important in controlling

75

80

85

90

95

100

105

110

115

1955 1965 1975 1985 1995 2005 2015 2025 2035 2045 2055 2065 2075 2085 2095Year

Ann

ual F

requ

ency

Fig. 1. Global annual frequency of tropical cyclones averaged in 10-y blocksfor the period 1950–2100, using historical simulations for the period 1950–2005 and the RCP8.5 scenario for the period 2006–2100. In each box, the redline represents the median among the six models, and the bottom and topsof the boxes represent the 25th and 75th percentiles, respectively. Thewhiskers extend to the most extreme points not considered outliers, whichare represented by the red + signs. Points are considered outliers if they liemore than 1.5 times the box height above or below the box.

Change in Track Density

0o 60oE 120oE 180oW 120oW 60oW 0o

60oS

30oS

0o

30oN

60oN

−0.01

0

0.01

0.02

0.03

0.04

0.05

0.06

0.07

0.08

Fig. 2. Change in track density, measured in number of events per 4° × 4° gridbox per year, averaged over the six models. The change is the average over theperiod 2006–2100 minus the average over 1950–2005. The white regions arewhere fewer than five of the six models agree on the sign of the change.

Emanuel PNAS | July 23, 2013 | vol. 110 | no. 30 | 12221

EART

H,ATM

OSP

HERIC,

ANDPL

ANET

ARY

SCIENCE

S

Global annual frequency of TCs, RCP 8.5 Windspeed of U.S. landfalling hurricanes

0.0000#

0.1000#

0.2000#

0.3000#

0.4000#

0.5000#

0.6000#

0.7000#

0# 50# 100# 150# 200#Knots&

US&Landfall&Wind&Speeds&

2010#

2050#

2100#

210020502010

We (i.e., Maya) are working to incorporate potential changes in storms (and thus the extreme value distribution) into the allowance framework.

52

Estimating the costs ECONOMIC R SKS OF

CL MATE CHANGEAn American Prospectus

Foreword by Michael R. Bloomberg, Henry M. Paulson, and Thomas F. Steyer

TR E V O R H O US E R , S O LO M O N H S I A N G , R O B E RT KO P P, A N D K ATE L A R S E N

Contributions by Karen Fisher-Vanden, Michael Greenstone, Geoffrey Heal, Michael Oppenheimer, Nicholas Stern, and Bob Ward

www.climateprospectus.org 53

Value of current property falling below the mean high tide lineBillion dollars (assuming current property distribution)

Diamonds indicate 99th percentile (1-in-100) projection.

0

200

400

600

800

1000

1200

1400

1600

1800

8.5 4.5 2.6 8.5 4.5 2.6 8.5 4.5 2.6

2030 2050 2100

U.S. coastal property loss assessment with the RMS North Atlantic Hurricane Model, assuming current property values, land use

54www.climateprospectus.org

C-5

RMS NORTH ATLANTIC HURRICANE MODEL

Hurricane Modeling

Stochastic Module

The stochastic module consists of a set of thousands of stochastic events that represent more than 100,000 years of hurricane activity. RMS scientists have used state-of-the-art modeling technologies to develop a stochastic event set made of events that are physically realistic and span the range of all possible storms that could occur in the coming years.

At the heart of the stochastic module is a statistical track model that relies on advanced statistical techniques (Hall & Jewson 2007) to extrapolate the HURDAT catalog (Jarvinen et al., 1984) and generate a set of stochastic tracks having similar statistical characteristics to the HURDAT historical tracks. Stochastic tracks are simulated from genesis (starting point) to lysis (last point) using a semi-parametric statistical track model that is based on historical data. Simulated hurricane tracks provide the key drivers of risk, including landfall intensity, landfall frequency, and landfall correlation.

The last step is a calibration process ensuring that simulated landfall frequencies are in agreement with the historical record. Target landfall rates are computed on a set of linear coastal segments (or RMS gates) by smoothing the historical landfall rates. The RMS smoothing technique uses long coastal segments, obtained by extending each RMS gate in both directions and keeping the orientation constant. Historical storms that cross an extended gate contribute to the landfall rate at the corresponding original segment.

Importance sampling of the simulated tracks is performed to create the computationally efficient event set used for loss cost determinations. Some statistics related to the resulting stochastic event set are presented in Table 1.

Table 1: Number of Stochastic Events “Hitting” a Country or a Combination of Countries

U.S. Canada U.S. or Canada

Number of events 32,702 10,648 37,538

Each of these events has a frequency of occurrence given by its mean Poisson rate. Because event frequencies were calibrated against history (HURDAT 1900–2011), this set of Poisson rates represents the RMS baseline model and this rate set is called the “RMS historical rate set.”

Windfield (or Wind Hazard) Module

The wind hazard module calculates windfields from both landfalling hurricanes and bypassing storms. Once tracks and intensities have been simulated by the stochastic module, the windfield module simulates 10 meter, three-second gusts on a variable resolution grid. Size and shape of the time stepping windfields are generated using an analytical wind profile derived from Willoughby (2006). Peak gusts are the driver of building damage and are used as the basis of many building codes world-wide, including the U.S.

Wind Vulnerability Module

The RMS U.S. and Canada Hurricane Models wind vulnerability module relates the expected amount of physical damage to buildings and contents at a location to the modeled peak three-second gust wind speed at that location. The severity of the damage is expressed in terms of a mean damage ratio (MDR), which is defined as the ratio of the expected cost to repair the damaged property to its replacement cost. The wind vulnerability module also estimates the variability in the damage ratio and models it with a beta distribution. Time-element losses due to business interruption and additional living expenses are estimated using occupancy-dependent facility restoration functions that relate the expected duration of the loss of use to the modeled building damage.

Storm Surge Model

The RMS North Atlantic Hurricane Model storm surge system utilizes the latest technology and storm surge hazard assessment methodologies to quantify the risk of storm surge for the U.S. Atlantic and Gulf coastlines (from Texas to Maine), as well as for parts of the Caribbean. The model system uses wind and pressure fields from the stochastic event set of the RMS North Atlantic Hurricane Model as forcing (i.e., as input) for the state-of-the-art MIKE 21 hydrodynamic model system to estimate still-water (i.e. without wave effects) storm surge. This hydrodynamic model system is also used to determine wave heights and resulting hazard statistics for offshore platforms. The impact of rainfall accumulation during an event is not considered in the modeling of storm surge levels.

55

Sea-level rise and changes in storms expand the flood 1% probability flood plain, New York City and surroundings

C EC D

Current

2030

2100

2050

Historical storms RCP 8.5 storms

www.climateprospectus.org


0

10

20

30

40

50

60

70

80

8.5 4.5 2.6 8.5 4.5 2.6 8.5 4.5 2.6

2030 2050 210056

Increased average annual coastal storm damageBillion dollars per year (assuming current property distribution, abandoning property that falls below mean sea level)

2000 baseline: $27 billion

Diamonds indicate 99th percentile (1-in-100) projection.


0%

50%

100%

150%

200%

250%

AL CT DC DE FL GA LA MA MD ME MS NC NH NJ NY PA RI SC TX VA

1-in-20

1-in-20

Likely

Range

Damages do not increase evenly distributed across the country

57

Average annual coastal flood damage due to property + business interruption RCP 8.5 2050 (percent increase by state)

58

These changes in average losses are driven by changes in extreme lossesReturn period curves in dollars, New York City Assuming no changes in storm intensity

Expe

cted

num

ber

of e

vent

s

0.001

0.01

0.1

Billion USD

0 10 20 30 40 50

6 cm SLR44 cm SLR96 cm SLR154 cm SLR212 cm SLR

Kopp et al. (in prep.)

Supe

rsto

rm S

andy

~1.5x

~2x

~5x

Frequencies increases labeled are for Sandy-scale damages.

~11x


Coastal property: Effects of changes in storm intensity

59

Effect of reweighting storms based on projected frequency/intensity RCP 8.5, as downscaled by Emanuel, 2013 (billion USD)

$0

$20

$40

$60

$80

$100

$120

$140

Historical

Activity

Projected

Activity

Historical

Activity

Projected

Activity

Historical

Activity

Projected

Activity

2030 2050 2100

1-in-20

1-in-20

Likely

Range

Historical: $27 billion

60

So what do we do?

61

Illustrative adaptation experiment: Effect of hard protection to fixed height (New York state)

Protection height (m)0 1 2 3 4 5

Expe

cted

add

ition

al lo

sses

/yea

r (m

illion

$)

-600

-400

-200

0

200

400

600NY

2020-20302020-20402020-20502020-20602020-20702020-20802020-20902020-2100

Project end year (starting 2020)2020 2040 2060 2080 2100

Dam

age-

stab

ilizin

g pr

otec

tion

heig

ht (m

)

2.8

3

3.2

3.4

3.6

3.8

4

4.2

4.4

4.6

Kopp et al. (in prep.)

Curr

ent p

rote

ctio

n he

ight

62

An illustrative example of costs: Metro Vancouver, BC

• Adaptation strategy focused on reinforcing current flood defenses, building new flood defenses, and soft measures, with a 1.0 m SLR target

• Cost: $9.5 billion

• Protects ~2880 sq. km, 2.3 million people

Cost of Adaptation - Sea Dikes & Alternative StrategiesFINAL REPORT

OCTOBER 2012EB3762

Submitted By

Ministry of Forests, Lands and Natural Resource Operations

With federal funding support through

Natural Resources Canada’s Regional

Adaptation Collaboration Program

Do we rebuild unchanged (and

assume the rest of the country will

continue to subsidize

indefinitely)?

63seasideheightsnjonline.comSeaside Heights, NJ

http://www.seasideheightsnjonline.com

Do we harden?

64Photo by www.aerolin.nl

Maeslantkering, the Netherlands

http://www.panoramio.com/photo/5820701

Do we raise (and otherwise modify our

communities to be resilient to occasional flooding)?

65NorthJersey.com

Bay Head, NJ

http://NorthJersey.com

Do we retreat and raze?

66Long Island Sound Study

http://longislandsoundstudy.net/2010/07/tidal-wetlands-restored/

How do we decide?

• Technocratically, e.g., through land use policy informed by benefit-cost analysis?

• Economically, e.g., with market tools such as accurately-priced, long-term flood insurance?

• Democratically, e.g., through deliberative community processes?

67

How do we decide?

• Technocratically, e.g., through land use policy informed by benefit-cost analysis?

• Economically, e.g., with market tools such as accurately-priced, long-term flood insurance?

• Democratically, e.g., through deliberative community processes?

67

My answer: Yes, we do all three, informed by risk assessments that reflect our best available knowledge.

The one thing we can’t do is wish (or legislate) sea-level rise away.

EPIC University of Chicago

30 September 2015

Bob Kopp

Earth System Science & Policy Lab, Rutgers University E-mail: [email protected]

Sea-level rise: Assessing the risks, Estimating the costs

5 Columbus Circle New York, NY 10019 Tel: +1.212.532.1157 Fax: +1.212.532.1162 www.rhg.com

PROJECT DESCRIPTION

An American Climate Risk Assessment Next Generation, in collaboration with Bloomberg Philanthropies and the Paulson Institute, has asked Rhodium Group (RHG) to convene a team of climate scientists and economists to assess the risk to the US economy of global climate change. This assessment, to conclude in late spring 2014,will combine a review of existing literature on the current and potential impacts of climate change in the United States with original research quantifying the potential economic costs of the range of possible climate futures Americans now face. The report will inform the work of a high-level and bipartisan climate risk committee co-chaired by Mayor Bloomberg, Secretary Paulson and Tom Steyer.

BACKGROUND AND CONTEXT

From Superstorm Sandy to Midwest droughts to wildfires in the Rocky Mountains, our weather is becoming more extreme and more expensive. In 2012, climate and weather disasters cost Americans more than $110 billion. Not only was it the hottest year on record, but precipitation was 2.6 inches lower than last century’s average, leading to massive crop losses across the Heartland and wildfires covering more than 9 million acres in the west. And while Sandy was the most destructive storm of the year, it wasn’t the only one with a multi-billion dollar economic price tag.

Weather is inherently variable and no single storm, heat wave or drought can be conclusively attributed to climate change. But there is mounting scientific evidence that greenhouse gas (GHG) emissions are increasing the frequency and severity of many extreme weather events. Sea level rise resulting from human-induced climate change amplified coastal flooding during Sandy. And increased atmospheric concentrations of GHG emissions are already resulting in prolonged stretches of excessively high temperatures, heavier downpours and more severe droughts.

While our understanding of climate change has improved dramatically in recent years, predicting future impacts is still a challenge. Uncertainty surrounding the level of GHG emissions going forward and the sensitivity of the climate system to those emissions makes it difficult to know exactly how much warming will occur. And we are learning more and more every day about how human and natural systems respond to potential changes in temperature, precipitation, sea level and storm patterns.

Uncertainty, of course, is not unique to climate change. The military plans for a wide range of possible conflict scenarios and public health officials prepare for pandemics of low or unknown probability. Households buy insurance to guard against myriad potential perils. And effective risk management is critical to business success and investment performance. In all these areas, decision makers consider a range of possible futures in deciding on a course of action. They work off the best information at hand and take advantage of new information as it becomes available.

Both the US National Academies of Science and the Intergovernmental Panel on Climate Change (IPCC) have suggested that this kind of “iterative risk management” is also the right way to approach climate change. Using this framework, the scientific community is preparing two major assessments of the risks to human and natural systems under a range of possible climate futures. The IPCC’s Fifth Assessment Report (AR5) will provide a global outlook, while the US government’s National Climate Assessment (NCA) will

mailto:[email protected]

sea-level rise: assessing the risk and estimating the costs

Documents

climate system

humaninduced climate

team of climate scientists

level of ghg emissions

range of possible futures

sea level rise

potential changes

weather disasters