This content has been downloaded from IOPscience. Please scroll down to see the full text.

Download details:

IP Address: 139.222.80.222

This content was downloaded on 19/05/2016 at 10:03

Please note that terms and conditions apply.

Poorest countries experience earlier anthropogenic emergence of daily temperature extremes

View the table of contents for this issue, or go to the journal homepage for more

2016 Environ. Res. Lett. 11 055007

(http://iopscience.iop.org/1748-9326/11/5/055007)

Home Search Collections Journals About Contact us My IOPscience

Environ. Res. Lett. 11 (2016) 055007 doi:10.1088/1748-9326/11/5/055007

LETTER

Poorest countries experience earlier anthropogenic emergence ofdaily temperature extremes

Luke JHarrington1, David J Frame1,2, ErichMFischer3, EdHawkins4,Manoj Joshi5 andChrisD Jones6

1 NewZealandClimate Change Research Institute, School ofGeography, Environment and Earth Sciences, VictoriaUniversity ofWellington,Wellington 6012,NewZealand

2 National Institute ofWater andAtmospheric Research,Wellington 6021,NewZealand3 Institute for Atmospheric andClimate Science, ETHZürich,Universitätstrasse 16, CH-8092 Zürich, Switzerland4 National Centre for Atmospheric Science, Department ofMeteorology, University of Reading, Reading RG6 6BB,UK5 Climatic ResearchUnit, School of Environmental Sciences, University of East Anglia, Norwich, UK6 MetOfficeHadley Centre, FitzRoyRoad, Exeter EX1 3PB,UK

E-mail: luke.harrington@vuw.ac.nz

Keywords: emergence, extreme temperatures, cumulative emissions, population exposure, CMIP5

Supplementarymaterial for this article is available online

AbstractUnderstanding how the emergence of the anthropogenic warming signal from the noise of internalvariability translates to changes in extreme event occurrence is of crucial societal importance. Byutilising simulations of cumulative carbon dioxide (CO2) emissions and temperature changes fromeleven earth systemmodels, we demonstrate that the inherently lower internal variability found attropical latitudes results in large increases in the frequency of extreme daily temperatures (exceedancesof the 99.9th percentile derived frompre-industrial climate simulations) occurringmuch earlier thanformid-to-high latitude regions.Most of theworld’s poorest people live at low latitudes, whenconsidering 2010GDP-PPP per capita; conversely thewealthiest population quintile disproportio-nately inhabitmore variablemid-latitude climates. Consequently, the fraction of the globalpopulation in the lowest socio-economic quintile is exposed to substantiallymore frequent dailytemperature extremes aftermuch lower increases in bothmean global warming and cumulative CO2

emissions.

1. Introduction

To understand how detectable anthropogenic influ-ences on the climate system will proliferate withtime, a large body of literature has considered thequestion of when the signal of climate changeemerges from the background noise of internalvariability. This ‘time of emergence’ concept has beenconsidered in both observational records and climatemodel simulations, for a range of climate indices suchas temperature (Hawkins and Sutton 2009, Diffen-baugh and Scherer 2011, Joshi et al 2011, Mahlsteinet al 2011, Hawkins and Sutton 2012, Hawkinset al 2014), precipitation (Giorgi and Bi 2009,Maraun 2013), the hydrological cycle (Sedláček andKnutti 2014), sea level rise (Lyu et al 2014) and eventransitions between different ecosystem regimes(Mahlstein et al 2013).

There has also been intense public and scientificinterest in recent years as whether and to what extentthe severity and frequency of extreme weather eventshave increased in response to anthropogenic climatewarming (Peterson et al 2012, 2013, Herringet al 2014). However, determining how the ‘time ofemergence’ concept can be applied in the context ofextreme events continues to be a developing area ofresearch. Recent work (Fischer and Knutti 2015) hasquantified the fraction of all moderate heat extremesand precipitation extremes globally which could beattributed to anthropogenic climate change in the pre-sent-day as well as for future climate change scenarios,while the emergence of statistically significant changesto specific climate extreme indices have also beendemonstrated (King et al 2015).

A key interpretation of studies on the time ofemergence of climate change indicators suggests that,

OPEN ACCESS

RECEIVED

27March 2016

REVISED

1May 2016

ACCEPTED FOR PUBLICATION

4May 2016

PUBLISHED

17May 2016

Original content from thisworkmay be used underthe terms of the CreativeCommonsAttribution 3.0licence.

Any further distribution ofthis workmustmaintainattribution to theauthor(s) and the title ofthework, journal citationandDOI.

© 2016 IOPPublishing Ltd

on seasonal timescales (Diffenbaugh andScherer 2011, Mahlstein et al 2011, Kirtman et al 2013,Mora et al 2013), climate signals may emerge from the‘noise’ of internal climate variability more quickly forlow latitude regions than elsewhere around the world.This has been convincingly demonstrated for meanclimate indicators (Diffenbaugh and Scherer 2011,Mahlstein et al 2011), but these implications have notbeen explored when considering the exposure of dif-ferent parts of the global population to the emergenceof extremes, particularly temperature extremes whichhave been found to occur much earlier than precipita-tion-related extremes (Fischer et al 2014, Kinget al 2015).

When considering the regional-scale impactswhich may occur in response to a changing climate,the majority of results are communicated with respectto corresponding changes in global mean temperature—this comparison is useful in the context of interna-tional climate targets, which are also framed in rela-tion to global mean temperature anomalies (Knuttiet al 2016). However, progress in recent years quanti-fying the near-linear correlation between cumulativecarbon dioxide (CO2) emissions with correspondingglobal temperature anomalies (Allen et al 2009, Mat-thews et al 2009, Meinshausen et al 2009) have per-mitted a more in-depth consideration of how theregional impacts of climate change may responddirectly to the emission of long-lived greenhousegases, with potentially important policy implications(Frame et al 2014).

In this study, we examine evidence for spatial het-erogeneity in the time-evolution of extreme temper-ature exceedances between different regions of theworld aggregated according to local population andincome characteristics, using for the first time, a directcomparison with the accumulation of simulated CO2

emissions.

2.Data andmethods

Following the methodology of previous work byFischer and Knutti (2015), we employ the use of the‘probability ratio’ (PR)metric, defined as PR=P1/P0where P0 is the probability of exceeding a certainquantile during the pre-industrial control period andP1 is the likelihood of exceedance in a more recentperiod, for example the last 30 years. PR can beinterpreted in a climate modelling framework as theincreased likelihood of an extreme temperaturethreshold being exceeded, when comparing a morerecent time periodwith a representation of a climate inthe absence of human interference (Allen 2003, Stottet al 2004).

Using models from the Coupled Model Inter-comparison Project Phase 5 (CMIP5, Taylor et al2012), we concatenate ‘Historical’ simulations for the

period 1901–2005 with corresponding representativeconcentration pathways (RCPs) for the period2006–2100, and then consider time-varying PR valuesusing moving 30-year windows. In this analysis, wechoose to define P0 as the 99.9th percentile of dailytemperatures (corresponding to a 1-in-1000 daytemperature extreme) based on 200 years of pre-industrial simulations. By focussing on changes to thenumber of exceedances above a fixed, well-definedpercentile threshold over running 30-year intervals,our analysis does not require any assumptions aboutthe shape and type of the underlying statistical dis-tribution, and thus avoids recent concerns relating tothe use of parametric analysis when considering a non-stationary time-series for the analysis of climateextremes (Sardeshmukh et al 2015, Sippel et al 2015).

In order to improve confidence in temperatureprojections by reducing model uncertainty associatedwith carbon cycle feedbacks, the RCP scenarios havebeen created with prescribed concentrations ratherthan emissions (Moss et al 2010). Consequently, theonly previous studies considering the physical climateimpacts of cumulative emissions from RCP scenarioshave inferred cumulative CO2 based on a best-guesslinear scaling of global mean temperature anomalies(Seneviratne et al 2016). Here, we instead utilise esti-mated RCP emissions calculated by Jones and collea-gues (2013), whereby the time-evolution ofatmospheric carbon and corresponding simulatedexchange of carbon with land and ocean sinks for asmaller subset of earth system models (table S1) havebeen used to infer anthropogenic emissions that arecompatible with each prescribed concentration path-way. Thus, all subsequent calculations of cumulativeemissions corresponding to extreme climate impactsinclude the uncertainty in translating global warminganomalies to cumulative emissions as well as theuncertainty in linking global temperatures to PRs(figure S1).

Previous work has considered the temporal evol-ution of PR (or similarly, risk ratios) averaged eitherglobally or over large regions (figure 1(a)). Here, wechoose to focus instead on when specific PR thresh-olds are exceeded at a single grid scale. Framing theemergence of temperature extremes in this manneralso enables a consistent method of comparisonbetween individual locations, as well as between differ-ent RCP scenarios. We define the cumulative emis-sions of probability ratio emergence (CEPREX) as thesimulated accumulation of CO2 emissions (since1870) corresponding to the central year of the 30-yearperiod over which a specific PR threshold (X) is excee-ded at a specific grid cell—we note that a PR thresholdis only considered to be exceeded when all subsequentvalues remain above the same threshold for theremainder of the time series available (2100). Forexample, figures 2(a), (c) and (e) show, respectively,the CEPRE corresponding to when the PR exceeds 2,

2

Environ. Res. Lett. 11 (2016) 055007

10 and 50 for each grid point globally. Hawkins et al(2014) highlighted possible endpoint effects with thistype of analysis, but sensitivity tests (figure S8) haveshown this is negligible in this study, due to the lowvariability in PR.

To consider how these emergent changes in PRare experienced by the populated regions of the world,we overlay global population and gross domestic pro-duct data, accounting for purchasing power parity(GDP-PPP), prepared for the year 2010 by the GlobalCarbon Project (http://www.cger.nies.go.jp/gcp/population-and-gdp.html) at a spatial resolution of0.5°×0.5° (figure S5). Future changes to populationand economic growth will have clear influences onthese results—we have chosen to focus on fixed data ofthe present-day to ascertain emergent risks for presentpopulations, and consider future projections of popu-lation and economic data in the supplementary mat-erial. While internal variability in temperature isgreater over smaller spatial scales, the aggregation ofgrid cells over areas comparable in size to populatedregions of the world have previously demonstrateddiscernible shifts in the probability distributions oftemperature and precipitation extremes (Fischeret al 2013, 2014), and thus will be suitable for analysingthe emergence of high-temperature PRs.

3. Results

Figure 2 illustrates the cumulative emissions requiredfor PR thresholds of 2, 10 and 50 to be exceeded indifferent regions of the world. The maps show themedian model CEPRE for each PR threshold usingRCP8.5 simulations, while the corresponding timingof exposure of the poorest and wealthiest quantiles ofthe global population are also presented. Consideringthe spatial distribution of CEPRE for each PR thresh-old, it is clear that fewer cumulative emissions arerequired for the continual exceedance of these PRthresholds to occur for lower latitudes, compared withhigher latitudes, while oceanic regions also generallyappear to experience a more rapid time of emergencethan corresponding land surfaces nearby. Theseresults are consistent with previous research (Diffen-baugh and Scherer 2011, Fischer and Knutti 2015) andsuggests that although land and high latitude regionsexperience larger warming signals than the globalmean, the variability in daily temperatures over oceanand low latitude regions are significantly lower,thereby resulting in the earlier emergence of morefrequent high-temperature extremes.

Considering the differences over quintiles inGDP-PPP per capita, we find increasingly large differencesin CEPRE between the wealthiest and poorest

Figure 1. Schematic illustrating how the time-evolution of high-temperature probability ratios are converted to theCEPREmetric. (a)Time-evolution of PR for an individual grid point, presented at the central year of the 30-year period (dashed red line given asexample) over which each PR valuewas calculated. (b)Corresponding globalmean temperature anomalies (relative to 1861–1880)over the period 1901–2100. (c)Compatible cumulative CO2 emissions over the same period,measured as a total since 1870. The blackasterisk corresponds to the year when a PR threshold of 10 (grey dashed line) isfirst exceeded, and continues to remain so, with thecorresponding blue asterisk indicating theCEPRE10 value for that particular grid box.

3

Environ. Res. Lett. 11 (2016) 055007

populations, as the PR threshold increases. For exam-ple, comparing the panels in figure 2 for a PR of 2(figure 2(b)) shows minor differences in the evolutionof population exposure with warming between thetwo economic groupings, but considering the cumula-tive emissions required for population exposure to amuch higher PR threshold of 50 reveals a much widergap in the time of emergence between the two sub-populations, albeit with large variability betweenmodels. It is also noted that CEPRE2 is less than400 PgC for nearly 100% of both sub-populations,which corresponds to present-day estimates of global

cumulative emissions (Peters et al 2015). Interestingly,when considering the CEPRE for RCP4.5 and RCP2.6simulations instead (figures S10 and S11), the differ-ences in the timing of PR exposure for the lower lati-tude locations are negligible. Instead, it is exceedancesof the higher PR thresholds which are not reached formid-to-high latitudes by the end of the century.

It is important to note that the wealthiest quintileof the global population is dispersed over a land areaseven times greater than that of the poorest quintile—however, even considering regions of equal land area,an equivalent fraction of the poorest populations

Figure 2. Left-hand panels show the global distribution of themedianmodel estimate of the cumulative emissions of probability ratioemergence (CEPRE). This is defined at the simulated cumulative emissions of carbon dioxide (in Petagrams of carbon, with respect to1870) at the central year of the 30-year periodwhen the probability ratio for one-day temperature extremes permanently exceeds athreshold of (a) 2; (c) 10 and (e) 50 at each individual grid point.White regions indicate where themedianmodel response shows noexceedance of the respective PR threshold by the end of 2100. Corresponding right-hand panels show the cumulative fraction of thepoorest (red) andwealthiest (blue) quintiles of the global population (based on 2010GDP-PPP per capita)which are exposed to thespecified PR threshold exceedances, also as a function of cumulative carbon emissions, for elevenCMIP5models over the period1901–2100, using ‘Historical’ and ‘RCP8.5’ simulations.

4

Environ. Res. Lett. 11 (2016) 055007

continues to experience earlier emergence of all threePR thresholds than their wealthier counterparts(figure S6).

Comparing, for each individual model, the differ-ence in cumulative emissions required for 50% of thepoorest population quintile to exceed each PR thresh-old versus the equivalent to occur for 50% of thewealthiest quintile, figure 3 reveals that nearly all mod-els across all RCP scenarios show high-temperaturePRs occurring earlier for the poorest populations thanfor their wealthier socio-economic counterparts.Under an RCP8.5 scenario for example, models sug-gest that between −10 and 560 PgC of additional car-bon would be emitted between the time when 50% ofthe poorest members of society continually experiencea 50-fold increase in 1-in-1000 day hot extremes andthe timewhen exposure occurs for an equal number ofcitizens within the wealthiest population quintile,thereby emphasising the contrasting time horizonsavailable for adaptation measures. Supplementaryanalysis reveals that models which exhibit the largestdifferences in the timing of exposure between the twosub-populations are actually those with a relativelymodest transient climate response to cumulative car-bon emissions (TCRE, Gillett et al 2013), andvice versa (figure S2). Rapid increases in PRs occur intropical latitudes for all models, while correspondingPR increases in higher latitudes only occur quickly forthe high-TCRE models (figure S4), subsequentlyresulting in smaller differences in CEPRE between thetwo sub-populations.

Further interrogating the differences in the cumu-lative distribution of CEPRE between the wealthiestand poorest socio-economic quintiles, figure 4 shows

the differences in the fractions of each populationquintile which have experienced emergence of eachPR threshold, as a function of global cumulative CO2

emissions. This figure demonstrates that (1) after agiven level of cumulative emissions, up to an addi-tional 60% of the poorest members of society crosseach of these PR thresholds than corresponding weal-thy populations; and (2) these patterns of unequalpopulation exposure in response to accumulating CO2

emissions occur consistently across all three RCP sce-narios. This shows that the differences in the timing ofemergence of temperature extremes between low lati-tude and high latitude regions are insensitive to therate of temperature change over the twenty-firstcentury.

In 2013, the United Nations Framework Conven-tion on Climate Change established theWarsaw Inter-national Mechanism to address the potential loss anddamage from climate change impacts for developingcountries (James et al 2014). The policy-relevance ofour result lies in the disparity between richer andpoorer people in terms of their exposure to the timingof emergence of temperature extremes. Whilst expo-sure to higher PRs does not result in higher vulner-ability, the adaptive capacity of a region can generallybe considered to scale with (1) climate variability and(2) income—certainly across the range implied byconsidering the richest and poorest quintiles(Grambsch and Menne 2003, Hayden et al 2011).These results do therefore suggest, ceteris paribus, ear-lier and more significant relative vulnerability totemperature extremes among the world’s poor and arethus of potential importance to policymakers.

Figure 3.The difference in cumulative emissions between the central year (of the 30-year period)when 50%of the poorest populationquintile experiences the continuous exceedance of a given PR threshold, and the corresponding central year when 50%of thewealthiest population quintile exceeds that same threshold. Onlymodel values are shownwhere greater than 50% cumulativeexposure occurs for both sub-populations by 2100. Blue, black and red circles correspond to eachmodel usingRCP2.6, RCP4.5 andRCP8.5 simulations, respectively.

5

Environ. Res. Lett. 11 (2016) 055007

4.Discussion

In this study, we have for the first time simulated thedirect influence of cumulative carbon emissions on thetime of emergence of daily extreme temperatures, byconsidering the CEPRE metric at specific PR thresh-olds. In considering only the persistent emergence ofgiven PR thresholds in the context of fractionalpopulation exposure, our results are less sensitive tointernal variability, despite the smaller spatial scalesconsidered. Moreover, previous research has foundthe emergence of more frequent heat extremes aver-aged over a longer time period (such as 5-day or 30-dayanomalies) occurs even earlier when compared with1-day extremes (Fischer andKnutti 2015).

By assessing emergent PR increases as a function ofcumulative CO2 emissions, this approach serves well

to evaluate the relative changes in heat extremesbetween different regions in the world, as well asacross scenarios, and could therefore be of use to thoseworking in vulnerability, impacts and adaptation, andto integrated assessment modellers, with the caveatthat the continuing exceedance of specific PR thresh-olds could be interpreted as a proxy for heat-relateddamages (Dunne et al 2013, Burke et al 2015, Hansenand Sato 2016).

While the compatible emissions profiles used inthis study have shown to accurately replicate the origi-nal Integrated Assessment Models used in developingthe RCPs (Jones et al 2013), it has been demonstratedthat emission-driven simulations overestimate warm-ing projections when compared with concentration-driven simulations (Friedlingstein et al 2014). Wetherefore choose to avoid specifying absolute

Figure 4.Difference in the fraction of population exposure to specified PR thresholds (colours) between the poorest 20% and thewealthiest 20%of the global population, as a function of CEPRE. Positive values indicate that a greater fraction of the poorestpopulation quintile is experiencing permanent exceedances of each specified PR threshold, comparedwith thewealthiest quintile,after a given quantity of cumulative carbon emissions.

6

Environ. Res. Lett. 11 (2016) 055007

cumulative emission targets for preventing the emer-gence of specific PR thresholds. However, even withthe added uncertainty of considering the end-to-endlink between cumulative emissions and extreme tem-peratures, the key differences in fractional exposure ofemergent high-temperature PRs between the wealth-iest and poorest global population quintiles remainclear, as evidenced by the equivalent results foundwhen considering the link with rises in global meantemperatures directly (figures S12–S14).

Even if emissions are towards the low end of therange considered by the RCP scenarios, this analysisshows that the pattern of changes in frequency of dailytemperature extremes remains robust: daily temper-ature extremes emerge to more frequently affect thepoorest 20% of the global population, when comparedwith the wealthiest 20%of the global population for allRCP scenarios, for a range of PRs. This result can beexplained primarily due to the fact that the poorestpeople in the world densely populate lower latituderegions, where the low variability in temperatureenhances the pace of emergence of a given signal-to-noise ratio when compared with higher latituderegions (Hawkins and Sutton 2009, Diffenbaugh andScherer 2011,Mahlstein et al 2011, Hawkins et al 2014,Hansen and Sato 2016).

As global cumulative carbon dioxide emissionscontinue to increase, the fractional gap in populationexposure between the poorer and wealthier membersof society will only widen for exponentially higher PRthresholds (figure 4). While all populated regionsaround the globe will enter a new regime of temper-ature extremes with no observed historical precedenceif cumulative emissions continue to increase at currentrates, the impacts, in terms of frequency of heatextremes, will become significantly worse for poorernations when compared with their wealthier counter-parts.We therefore argue that, even though our resultsshow the emergence of more severe temperatureextremes will always occur for poorer populationsfirst, the potential prevention of crossing extreme PRthresholdsmeans that the poorestmembers of the glo-bal community will always be the greatest beneficiariesof action towards a low-carbon pathway.

Acknowledgments

We acknowledge the World Climate Research Pro-gramme’s Working Group on Coupled Modelling,which is responsible for CMIP, and we thank theclimate modelling groups for producing and makingavailable their model output. For CMIP the USDepartment of Energy’s Programe for Climate ModelDiagnosis and Intercomparison provides coordinatingsupport and led development of software infrastruc-ture in partnership with the Global Organization forEarth System Science Portals. We thank the GlobalCarbon Project and the Japanese National Institute for

Environmental Studies for making their griddedpopulation and GDP scenarios freely available athttp://www.cger.nies.go.jp/gcp/population-and-gdp.html. The authors wish to thank Sam Dean andSuzanne Rosier for helpful discussions. LJH and DJFacknowledge research support from Victoria Univer-sity of Wellington and the Deep South NationalScience Challenge. EHwas funded by the UKNationalCentre for Atmospheric Science and a NERCAdvanced Fellowship. CDJ was supported by the JointUK DECC/Defra Met Office Hadley Centre ClimateProgramme (GA01101) and by the European Union’sHorizon 2020 research and innovation programmeunder grant agreementNo 641816 (CRESCENDO).

References

AllenM2003 Liability for climate changeNature 421 891–2AllenMR, FrameD J,Huntingford C, Jones CD, Lowe J A,

MeinshausenMandMeinshausenN2009Warming causedby cumulative carbon emissions towards the trillionth tonneNature 458 1163–6

BurkeM,Hsiang SMandMiguel E 2015Global non-lineareffect of temperature on economic productionNature 527235–9

DiffenbaughNS and SchererM2011Observational andmodel evidence of global emergence of permanent,unprecedented heat in the 20th and 21st centuriesClim.Change 107 615–24

Dunne J P, Stouffer R J and John JG 2013Reductions in labourcapacity fromheat stress under climate warmingNat. Clim.Change 3 563–6

Fischer EM,BeyerleU andKnutti R 2013Robust spatiallyaggregated projections of climate extremesNat. Clim. Change3 1033–8

Fischer EMandKnutti R 2015Anthropogenic contribution toglobal occurrence of heavy-precipitation and high-temperature extremesNat. Clim. Change 5 560–4

Fischer EM, Sedláček J, Hawkins E andKnutti R 2014Models agreeon forced response pattern of precipitation and temperatureextremesGeophys. Res. Lett. 41 2014GL062018

FrameD J,MaceyAH andAllenMR2014Cumulative emissionsand climate policyNat. Geosci. 7 692–3

Friedlingstein P,MeinshausenM,AroraVK, Jones CD,AnavA,Liddicoat SK andKnutti R 2014Uncertainties inCMIP5climate projections due to carbon cycle feedbacks J. Clim. 27511–26

Gillett NP, AroraVK,MatthewsD andAllenMR2013Constraining the ratio of global warming to cumulative CO2

emissions usingCMIP5 simulations J. Clim. 26 6844–58Giorgi F andBi X 2009Time of emergence (TOE) ofGHG-

forced precipitation change hot-spotsGeophys. Res. Lett. 36L06709

GrambschA andMenne B 2003Adaptation and adaptive capacity inthe public health contextClim. ChangeHealth Risks ResponsesedA JMcMichael (Geneva:WorldHealthOrganization)pp 220–36

Hansen J and SatoM2016Regional climate change and nationalresponsibilities Environ. Res. Lett. 11 034009

Hawkins E et al 2014Uncertainties in the timing of unprecedentedclimatesNature 511E3–5

Hawkins E and SuttonR 2009The potential to narrowuncertaintyin regional climate predictionsBull. Am.Meteorol. Soc. 901095–107

Hawkins E and SuttonR 2012Time of emergence of climate signalsGeophys. Res. Lett. 39 L01702

HaydenMH, Brenkert-SmithH andWilhelmiOV 2011Differential adaptive capacity to extreme heat: a phoenix,arizona, case studyWeather Clim. Soc. 3 269–80

7

Environ. Res. Lett. 11 (2016) 055007

Herring SC,HoerlingMP, Peterson TC and Stott PA 2014Explaining extreme events of 2013 from a climate perspectiveBull. Am.Meteorol. Soc. 95 S1–104

James R,Otto F, ParkerH, BoydE, Cornforth R,Mitchell D andAllenM2014Characterizing loss and damage from climatechangeNat. Clim. Change 4 938–9

Jones C et al 2013Twenty-first-century compatible CO2 emissionsand airborne fraction simulated byCMIP5 earth systemmodels under four representative concentration pathwaysJ. Clim. 26 4398–413

JoshiM,Hawkins E, SuttonR, Lowe J and FrameD2011 Projectionsof when temperature changewill exceed 2 °Cabove pre-industrial levelsNat. Clim. Change 1 407–12

KingAD,DonatMG, Fischer EM,Hawkins E, Alexander LV,KarolyD J,Dittus A J, Lewis SC and Perkins S E 2015Thetiming of anthropogenic emergence in simulated climateextremesEnviron. Res. Lett. 10 094015

KirtmanB et al 2013 Near-term climate change: projections andpredictabilityClimate Change 2013: The Physical Science Basis.Contribution ofWorkingGroup I to the Fifth Assessment Reportof the Intergovernmental Panel on Climate Change edT F Stocker et al (Cambridge: CambridgeUniversity Press)pp 953–1028 (http://ebooks.cambridge.org/ref/id/CBO9781107415324A031)

Knutti R, Rogelj J, Sedláček J and Fischer EM2016A scientificcritique of the two-degree climate change targetNat. Geosci. 913–8

LyuK, ZhangX, Church J A, SlangenABA andHu J 2014Time ofemergence for regional sea-level changeNat. Clim. Change 41006–10

Mahlstein I, Daniel J S and Solomon S 2013 Pace of shifts in climateregions increases with global temperatureNat. Clim. Change3 739–43

Mahlstein I, Knutti R, Solomon S and PortmannRW2011 Earlyonset of significant local warming in low latitude countriesEnviron. Res. Lett. 6 034009

MaraunD2013Whenwill trends in Europeanmean andheavy dailyprecipitation emerge? Environ. Res. Lett. 8 014004

MatthewsHD,Gillett NP, Stott PA andZickfeld K 2009Theproportionality of global warming to cumulative carbonemissionsNature 459 829–32

MeinshausenM,MeinshausenN,HareW, Raper SCB, Frieler K,Knutti R, FrameD J andAllenMR2009Greenhouse-gasemission targets for limiting global warming to 2 °CNature458 1158–62

MoraC et al 2013The projected timing of climate departure fromrecent variabilityNature 502 183–7

Moss RH et al 2010The next generation of scenarios for climatechange research and assessmentNature 463 747–56

Peters GP, AndrewRM, Solomon S and Friedlingstein P 2015Measuring a fair and ambitious climate agreement usingcumulative emissions Environ. Res. Lett. 10 105004

PetersonTC, Stott PA andHerring S 2012 Explaining extremeevents of 2011 from a climate perspectiveBull. Am.Meteorol.Soc. 93 1041–67

PetersonTC, Stott PA andHerring S 2013 Explaining extremeevents of 2012 from a climate perspectiveBull. Am.Meteorol.Soc. 94 S1–74

Sardeshmukh PD,CompoGP and PenlandC 2015Need forcaution in interpreting extremeweather statistics J. Clim. 289166–87

Sedláček J andKnutti R 2014Half of theworld’s populationexperience robust changes in thewater cycle for a 2 °CwarmerworldEnviron. Res. Lett. 9 044008

Seneviratne S I, DonatMG, PitmanA J, Knutti R andWilby R L2016Allowable CO2 emissions based on regional and impact-related climate targetsNature 529 477–83

Sippel S, Zscheischler J, HeimannM,Otto F E L, Peters J andMahechaMD2015Quantifying changes in climatevariability and extremes: pitfalls and their overcomingGeophys. Res. Lett. 42 2015GL066307

Stott PA, StoneDA andAllenMR2004Human contribution to theEuropean heatwave of 2003Nature 432 610–4

Taylor K E, Stouffer R J andMeehl GA 2012Anoverview ofCMIP5 and the experiment designBull. Am.Meteorol. Soc. 93485–98

8

Environ. Res. Lett. 11 (2016) 055007