surface mass balance of the greenland ice sheet in …...instructions for use title surface mass...

Instructions for use

Title Surface mass balance of the Greenland ice sheet in the regional climate model HIRHAM5 Present state and futureprospects

Author(s) Mottram Ruth Boberg Fredrik Langen Peter Yang Shuting Rodehacke Christian Christensen Jens HesselbjergMadsen Marianne Sloth

Citation 低温科学 75 105-115

Issue Date 2017-03-31

DOI 1014943lowtemsci75105

Doc URL httphdlhandlenet211565106

Type bulletin (article)

File Information 12_105-115pdf

Hokkaido University Collection of Scholarly and Academic Papers HUSCAP

Surface mass balance of the Greenland icesheet in the regional climate model HIRHAM5

Present state and future prospectsRuth Mottram1 Fredrik Boberg1 Peter Langen1 Shuting Yang1

Christian Rodehacke1 Jens Hesselbjerg Christensen1 Marianne Sloth Madsen1

Received 1 February 2017 accepted 9 February 2017

Surface mass balance (SMB) is the builder of the Greenland ice sheet and the driver of icedynamics Quantifying the past present and future state of SMB is important to understand thedrivers and climatic processes that control SMB and to both initialize and run ice sheet models whichwill help clarify sea level rise and how likely changes in ice sheet extent feedback within the climatesystem

Regional climate models (RCMs) and climate reanalysis are used to quantify SMB estimatesAlthough different models have different spatial and temporal biases and may include differentprocesses giving significant uncertainty in both SMB and the ice sheet dynamic response to it allRCMs show a recent declining trend in SMB from the Greenland ice sheet driven primarily byenhanced melt rates Here we present new simulations of the Greenland ice sheet SMB at 5 kmresolution from the RCM HIRHAM5 The RCM is driven by the ERA-Interim reanalysis and theglobal climate model (GCM) EC-Earth v23 to make future projections for climate scenarios RCP85and RCP45

Future estimates of SMB are affected by biases in driving global climate models and feedbacksbetween the ice sheet surface and the global and regional climate system are neglected likely resultingin significant underestimates of melt and precipitation over the ice sheet These challenges will needto be met to better estimate the role climate change will have in modulating the surface mass balanceof the Greenland ice sheet

Keywords Greenland ice sheet surface mass balance climate climate modelling ice sheet modelling

1 Introduction

The Greenland ice sheet is famously the secondlargest land ice mass on the planet containing around285 million km3 of ice over an area of 171 million km2equivalent to 72 m sea level rise (IPCC 2013) Resultsfrom remote sensing missions such as the GRACE

gravimetry satellites indicate a recent significantnegative mass budget with around -234plusmn20 Gt of netmass (Barletta et al 2013) being lost each year includingboth the contribution from precipitation and surfacemass balance and the dynamic contribution of calvingfrom icebergs and submarine melt (Shepherd et al 2012)The surface mass balance component is by far the mostimportant component of the mass budget since itincludes both positive (accumulation by precipitation)and negative (melt and runoff) terms though around onethird to one half of the mass lost by the ice sheet is fromcalving glaciers (Enderlin et al 2014)

低温科学 75 （2017） 105-115doi 1014943lowtemsci 75 105

Corresponding authore-mailrumdmidk) Danish Meteorological Institute CopenhagenDenmark

SMB is not only important in itself but is alsoimportant for ice sheet modelling studies where SMB orthe degree day approximation of it based on temperatureand precipitation is used to drive ice sheet dynamicsWe thus use high-resolution regional climate models toclarify both the current state of the ice sheet surfacemass balance and its future prospects under climatechange scenarios with the aim of also providing SMBforcing for dynamical ice sheet modelsPrevious work on Greenland ice sheet mass balance

has used the regional climate models MAR and RACMOas well as HIRHAM5 to determine the present-day andfuture surface mass balance (Lucas-Picher et al 2012Rae et al 2012 Langen et al 2015) Other models suchas SnowModel (Mernild et al 2009) or Hanna et al ʼs(2013) dECMWF model are not strictly regional climatemodels but use either model output to drive a separatesnow scheme (SnowModel) or use statistical methods todownscale output that is then used to calculate SMBbased on temperature index methods (dECMWF)(Church et al 2013)These models have previously estimated the mean

annual SMB to be in the range 340 to 470 Gt per year (seeTable 1) though note that both forcing data time periodresolution as well as processes and ice mask all vary inthese estimates and this can have a significant effect onany one models SMB estimate when compared to theothers as shown by Vernon et al (2013) An earlierversion of HIRHAM5 showed an average annual SMB of188 Gt year-1 over the whole ice sheet for the period 1990

to 2008 (Rae et al 2012) However this version was runat a coarser resolution of 25 km and with a significantlysimplified snow model to calculate surface mass balanceOur current version of HIRHAM5 is run at 5 kmresolution and was used by Langen et al (2015) to showregional changes in Greenland Here we use a furtherupdated version of the model (Langen et al 2017) tocalculate SMB across the Greenland ice sheet and tomake future projections when forced with a globalclimate modelThe high resolution of the HIRHAM5 model in this

set of simulations means that we can assess theperformance of the model on both the narrow ablationzone and on small outlet glaciers and peripheral ice capswithout needing to statistically account for elevationchanges as performed for example by Noeumll et al (2016)Results by Langen et al (2017) show that the modelperforms well in these marginal areas even withoutelevation correction when compared with ablation andweather station measurements predominantly in theablation zone compiled by Machguth et al (2016)Similarly analysis by Schmidt et al (submitted) showsthat the model works very well over the small Icelandicice cap Vatnajoumlkull which is analogous to peripheralglaciers in Greenland At the present day up to 10 ofthe mass loss from Greenland is currently contributed bythese small glaciers and ice caps (Bolch et al 2013) and itis therefore important to be able to account for theirsurface mass balance correctly in model projectionsIn this study we use a separate dataset of shallow

106 Ruth Mottram Fredrik Boberg Peter Langen Shuting Yang Christian Rodehacke Jens Hesselbjerg Christensen Marianne Sloth Madsen

Table 1Comparison of surface mass balance components calculated for the present day using different models Note that asVernon et al (2013) point out these models also have different resolutions and different ice masks so that the numbers are not quitecomparable Therefore we include in the right most column values from that study that compare the RACMO and MAR modelsusing the same ice mask and forcing data though with differing model resolution

Model Mean annual surfacemass balance (Gt year-1)

Mean annual precipita-tion (Gt year-1)

Mean annual runoff(Gt year-1)

Mean annual surfacemass balance (Gt year-1)1960-2008 from Vernonet al (2013)

HIRHAM5(1980-2014)(this study)

360plusmn134 866plusmn70 446plusmn109 mdash

RACMO(1991-2015)(van den Broeke et al2016)

306plusmn120 712plusmn70 363plusmn102 470

MAR(1980-2015)(Fettweis et al 2016)


firn cores from around the accumulation zone inGreenland compiled by Buchardt et al (2012) to evaluatehow the model performs at higher elevations In thefuture dynamically driven changes in ice sheet altitudedue to increasing melt rates and projected increases inprecipitation will likely have important feedbacks onboth orographic precipitation and melt rates In thissense the use of high-resolution simulations offersadditional value in both identifying and quantifyingfeedbacks between ice and atmosphereThe combination of surface mass balance modelling

with estimates of total mass budget derived from bothaltimetry and flux gate calculations (Rignot et al 2011)and the satellite gravimetry mission GRACE offers thepossibility of estimating the relative importance ofsurface mass processes and calving dynamics (Shepherdet al 2012) In this paper we focus on the surface massbudget contribution both at the present day and in thefuture

2 Regional climate model HIRHAM5

In this study we use the latest version of the RCMHIRHAM5 (Langen et al 2017) for a domain coveringGreenland Iceland and parts of Arctic Canada (Figure 1)The RCM was developed at the Danish MeteorologicalInstitute (DMI) (Christensen et al 2006) from theHIRLAM7 numerical weather prediction model (Eerola2006) and the ECHAM5 global climate model (Roeckneret al 2003) The model is very similar in set-up to thatfully described in Langen et al (2015 2017) and Lucas-Picher et al (2014) run on a 005degtimes005deg rotated polargrid for a 35 year period (1980-2014) For the present-day simulations HIRHAM5 was forced on the lateralboundaries by the ERA-Interim reanalysis product (Deeet al 2011) every 6 hours On the lower boundary seasurface temperatures (SSTs) and sea ice concentrationwere statistically interpolated from the ERA-Interimdata format to the model resolution and prescribed dailyThe model runs freely within the boundaries and onlytemperature pressure relative humidity and windvelocities are used in lateral boundary forcing Thecurrent set-up of HIRHAM5 has 31 vertical levels in theatmosphere 5 vertical levels in the soil includingglaciers and snow and a 90 second dynamical time step A

32 layer snow pack model is applied offline over glacierpoints to calculate surface mass balance using moredetailed snowpack processes (Langen et al 2017 seebelow) The topography of Greenland is taken fromBamber et al (2001) and it should be noted that althoughsnow is allowed to accumulate without limits overglaciers and ice sheets the elevation of the ice sheetsurface remains fixed during the simulation such thatthere are no feedbacks between surface mass balanceand ice dynamics The ice mask used in this simulationis updated from that used by Langen et al (2015) to thatproduced by Citterio and Ahlstroslashm (2013) with additionaldata for Iceland provided by the Icelandic Met Office(Figure 1)Using this same set-up we also forced HIRHAM5 on

the lateral boundaries with fields from the EC-Earthv23 GCM (Hazeleger et al 2012) in order to producefuture climate simulations of the surface mass balance ofGreenland In this set-up EC-Earth was run at aresolution of 125 km using the historical emissions andRCP45 and 85 climate scenarios in the CMIP5 set-up(Taylor et al 2012) Both the lateral boundaries and theocean SSTs and sea ice were provided by the drivingGCM to the HIRHAM5 model for downscaling experi-ments The future projections were run as transient 20year time slices for the mid and end of the 21st centuryusing RCP45 and RCP85 scenarios as well as a

107Surface mass balance of the Greenland ice sheet

Figure 1The domain topography and ice mask used in allHIRHAM5 simulations for Greenland

simulation using historical emissions for the period 1991-2010 used as a control run to assess performance of theEC-Earth model at the present dayPrior to running the simulations the inclusion of the

full snowpack model as detailed below meant that anadequate spin-up was required in particular for the snowand ice properties to reach equilibrium with the climateFor this purpose we ran the HIRHAM5 model for oneyear and then used the atmospheric output to run thesurface scheme offline repeatedly until decadal means ofrunoff and subsurface temperatures ceased to showtransient variability (in this case after 70 years ofsimulation time)The full details of the surface mass balance

calculation in HIRHAM5 are given in Langen et al (2017)but a quick summary is given here to assist ininterpreting results Surface mass balance is calculatedfrom the sum of the precipitation and the runoff (anegative term) Runoff is calculated from the melt ofsurface snow and ice accounting for the effects ofretention and refreezing within any snowpack thatoverlies the glacier ice Rain on snow is similarlyaccounted for though rain on bare ice is assumed torunoff directly There is currently no runoff routingscheme from the model and therefore no superimposedice formation on bare glacier ice However with theintroduction of a sophisticated firn model processes nowinclude densification snow grain growth snow state-dependent hydraulic conductivity superimposed iceformation at the base of the snow pack on glacier ice andirreducible water saturation The accommodation forwater retention in excess of the irreducible saturationmeans that formation of both perennial firn aquifers andperched ice lenses occurs within the snowpack (Langenet al 2017)The HIRHAM5 surface scheme uses a standard

energy balance scheme to calculate the melt of snow andunderlying glacier ice

Qm1048637Sin(1－)1048619Lin－Lout1048619QH1048619QL

M1048637QmwL f (1)

where Qm is the total energy flux at the surface Sin is theincoming shortwave radiation ɑ is the surface albedo Linand Lout are the incoming and outgoing longwave

radiative fluxes respectively and QH and QL are thesensible and latent heat fluxes As the skin tempera-ture cannot rise above the freezing point of 27315deg K ona glacier surface the excess energy is instantaneouslyused for melt M accounting for the density of water wand the release of latent heat LfMeltwater and rain that falls over glaciers are

grouped together in the model as the liquid waterfraction Where there is a snow layer overlying theglacier surface the liquid water percolates into thedeeper layers according to a given threshold ofirreducible water saturation and if there is a sufficientcold content refreezes within the snowpack In thesub-surface scheme over glaciers there are 32 layers ofunequal thickness of snow or ice or some combination ofthe two depending on the depth of surface snow Thethickness of the layers in the subsurface scheme overglaciers or snow-covered land is calculated in metres ofwater equivalent to allow for the easy calculation of masschange with different heat diffusion and conductivityparameter values used for water ice and snow withinthe snowpack As surface melting and rainfall occurthe water percolates into the lower layers and retentionis calculated by allowing liquid water in excess of thedensity-dependent irreducible water saturation withinthe layer to percolate to the layer below Ice is formedin the snowpack layers based on the ldquocold contentrdquo ineach layer The cold content determines how muchliquid water can freeze within that layer based on theenergy required to heat the snow and ice mass to thefreezing point in each layer This is used to instantane-ously freeze as much liquid water as is available or as thecold content allows within a single time step Thismass is then transferred to the ice fraction and thetemperature of the layer is calculated taking into accountthe latent heat release to conserve energyAs the subsurface scheme extends to a depth of 60

metres water equivalent (mwe) if there is less than 60mwe snow on the glacier surface then the layers of thesubsurface scheme have the properties of ice Nopercolation of meltwater is allowed through ice layersbut energy fluxes are diffused through all the layersWe assume that the ice layers from the glacier cominginto the column at the base of the subsurface schemehave a temperature equivalent to the annual mean


temperature of that grid square locationAs Langen et al (2017) discuss the albedo of the

model is one of the largest sources of uncertainty in themass balance estimate In the simulations we presenthere we use the internally derived albedo schemerather than an externally forced albedo product based onMODIS as also described in Langen et al (2017) Thisallows us to make a direct comparison of the present-daySMB with the future projections where data such asMODIS is of course unavailable The albedo scheme inthe model has a cold (＜－5degC) snow albedo of 085linearly decreasing to 065 as temperature increasesfrom－5degC to 0degC The albedo for bare ice is fixed at04 with a linear function for thin (＜3 cm) snow betweenthe two values Broadband values lower than 04 dooccur as shown in measurements at the PROMICEautomatic weather stations (van As et al 2016) andMODIS data products (Stroeve et al 2006) Similarlyfreshly exposed ice can have much brighter values forsurface albedo but the values in the current schemewere optimized over the ice sheet by Nielsen-Englyst(2015) Langen et al (2017) show that significant biasesstill persist in albedo values over the ice sheet howeverpartly due to processes such as dust accumulation andbiological activity that are not currently includedRecent work by Stibal et al (2015) demonstrates theimportance of microorganisms and melted out dust onthe Greenland ice sheet and the development ofappropriate parameterizations is an area of active

researchThe HIRHAM5 model performance is comprehen-

sively evaluated in Langen et al (2015 2017) usingweather station data shallow firn cores and historicalSMB observations from Machguth et al (2016) Theirresults show that the model performs extremely wellover Greenland reproducing air temperatures andaccumulation rates well on average However somebiases in radiation suggest that cloud cover or cloudoptical thickness may not be fully captured in the modelSimilarly HIRHAM5 is not able to capture deep coldinversion layers over the ice sheet and evidence fromFausto et al (2016) suggests that during extreme meltevents such as were observed in 2012 the modelunderestimated the sensible heat flux by up to 75 insome locationsAnalysis of the same ERA-Interim simulation by

Schmidt et al (submitted) over Icelandic glaciers showsthat HIRHAM5 also performs well over small glaciersand ice caps particularly with regard to accumulationrates However Schmidt et al (submitted) also identifythat orographically forced precipitation is overestimatedon the upslope with a possible small dry bias downwindThis is a common problem in hydrostatic models wherethe dynamical scheme means that precipitation ishandled diagnostically (Forbes et al 2011)


Figure 2Comparison of observed SMB from shallow core data compiled by Buchardt et al(2012) with modelled values from the ERA-Interim forced simulation and the historicalemissions scenario of EC-Earth

3 Results and discussion

31 Surface mass balanceIn Figure 2 we complement the model analysis of

Langen et al (2017) by comparing observed surface massbalance from stake measurements and shallow firn cores(Buchardt et al 2012) with those calculated from thesame grid cell in HIRHAM5 Figure 2 shows that themodel is able to reproduce the pattern of SMB over theice sheet when forced with both ERA-Interim (R2

10486370898) and EC-Earth (R210486370895) though with a veryslightly more pronounced dry bias in the historicalsimulation For this comparison we took the decadalmean SMB from firn cores that cover all or part of theperiod of the ERA-Interim driven simulation Althoughthe model reproduces the observed SMB pretty well it isimportant to note that observed SMB rates aredetermined by short- and medium-term climate variabil-ity that may not be captured by climate models thusleading to the poorer fit against these point measure-ments This point is underlined in Table 2 where thereare large differences in SMB calculated on a decadaltimescale and compared across the full 1980-2014simulation period However note that in Table 1 thetotal SMB for Greenland is close to but not exactly thesame as comparable simulations using other RCMsAs we plot both model simulations against the same

core data for comparison it is interesting to see that thepattern of high and low accumulation rates across

Greenland are reflected in both simulations (see alsoFigure 5) but the magnitude differs slightly with the EC-Earth historical run having a generally lower valueThis suggests that in the accumulation zone at least bothmodels have a dry bias though that in the EC-Earthsimulation is somewhat stronger Observational datafrom the south-east of Greenland the region with highestprecipitation are too sparse to assess how well the modelreproduces it For the only core data we have fromsouth-east Greenland the model under-estimates SMB inboth simulations though to a lesser extent in the ERA-Interim simulation More observations in this regionwould be helpful in both evaluating the model andcomparing results from this simulation with othermodels

32 Surface mass balance componentsFigure 3 shows the surface mass balance of

Greenland as a whole for the period 1980 to 2014 There


Table 2SMB calculated using the ERA-Interim drivensimulation in different decades and overall The choice ofperiod can make a large difference in the estimated annualmean SMB

Period Mean annual SMB(Gt year-1)

1980-2014 360plusmn1341980-1990 375plusmn1301991-2010 346plusmn1322000-2014 277plusmn101

Figure 3Annual mean components of surface mass balance for the present day in theERA-Interim simulation

is a strong declining trend in SMB in the most recentdecade driven largely by increased melt and runoff ratesThe interannual variability is however also large andshows the importance of long time series in analyzing theSMB trends There has also been an increase in rainfallevents over Greenland though this is barely discernableover the whole ice sheet However work by Doyle et al(2015) analyzing the ERA-Interim simulation shows thatin western Greenland rainfall events occurring morefrequently at higher altitudes over the ice sheet have adistinct impact on the dynamics of the ice sheetComparison between Figures 3 and 4a shows that

the EC-Earth historical simulation underestimatessnowfall across the whole ice sheet compared to ERA-Interim however melt rates are also lower and thesetwo components compensate for each other to someextent when estimating total SMB Note that the firncore observations in Figure 2 are confined to the higheraccumulation zone where differences in melt and runoffrates between the two historical period simulations areminorFigures 4b and 4c show the end of the 21st century

under two different climate scenarios A comparisonbetween Figures 3 and 4b show that under the RCP45scenario the model expects a similar though still highermean annual SMB to that of the last decade in ERA-Interim with both higher precipitation and higher meltrates than the historical simulation It is important tonote that in Figure 5 both the pattern of precipitation

and melt appear comparable with that given by theERA-interim simulation suggesting that EC-Earth doesnot simulate significant changes in circulationThere is a documented bias in EC-Earth v23 which

is colder in the Arctic region during the historical periodthan observations suggest (Hazeleger et al 2012) Thisis partly attributable to a larger area covered by sea icethan observed and possibly also to the albedo scheme inGreenland (Helsen et al 2016) in the standard EC-Earthset-up Ongoing work to couple EC-Earth to an icesheet model (Madsen et al in preparation) may improvethis in the futureOn the other hand the sources of high melt rates

observed over the ice sheet in recent years appear to bepersistent blocking high pressure systems that maysimply reflect internal variability within the climatesystem (Fettweis et al 2013) Although recent work byHanna et al (2016) suggests that some of the recentincreases in persistent anomalies could be a climatechange signal analysis of data from a wide range ofclimate models within the CMIP5 archive does not seemto indicate this is an expected signal of climate changeThe higher SMB in the EC-Earth simulations suggeststhat the cold bias in EC-Earth observed historically in theArctic persists through the 21st century Assumingthat the ERA-Interim driven SMB is accurate itsuggests that the EC-Earth climate simulations may beunder-estimating the rate of climate change in the Arcticover this century However the much lower SMB in


Figure 4Annual mean components of surface mass balance for a) the historical emissionssimulation and the end of the century under b) RCP45 and c) RCP85 scenarios

the RCP85 simulation at the end of the 21st centuryreflects a much greater enhancement to Arctic warmingin the higher RCP85 emissions scenarioThe change in ice sheet SMB expected under

climate scenarios RCP45 and RCP85 both showincreased precipitation and a significant increase in meltcompared in melt rates compared to the historicalsimulation These increases are scaled by scenario sothat RCP45 has a lower increase in melt than RCP85Table 3 summarizes the projected change in SMB

for the mid-century and end of the century simulationsfor the two different climate scenarios The plots inFigure 5 show that while the precipitation patternremains similar through the projections the melt areawill expand particularly in western and northernGreenland and over the saddle region in the south withthe magnitude of this expansion determined by thescenario The increase in precipitation projected by themodel is largely confined to the south-east with a verysmall increase in northern Greenland this represents anintensification of existing precipitation patternsOne potential source of uncertainty in this study is

the fixed ice sheet mask during the simulations

Neither the ablation― elevation feedback which leadsto lowering of the ice sheet margin and thus increasedablation rates nor the potential migration of orographicprecipitation as the ice sheet margin retreats areincluded Similarly with high ablation rates the lowerablation zone and peripheral glaciers will be lost at somepoint but these are still in the ice mask at the end ofcentury when SMB is calculated in these simulationsThis suggests that the very low annual SMB underRCP85 may be an understimate Further analysis ofthe difference between ice sheet and peripheral glaciers


Table 3SMB from historical mid-century and end of thecentury simulations for two different climate scenariosRCP45 and RCP85 dynamically downscaled from EC-Earthusing the HIRHAM5 set-up Note that the historical emissionsscenario is used for the period 1990-2006 and RCP45 for theremaining four years

Mean annual surface mass balance (Gtyear-1)

Simulationperiod Historical RCP45 RCP85

1991-2010 492plusmn94 mdash mdash2031-2050 mdash 460plusmn79 414plusmn652081-2100 mdash 418plusmn109 193plusmn99

Figure 5SMB precipitation and the number of melt days for the ERA-Interim and historicalemissions scenario as well as the end of the 21st century RCP45 and RCP85 scenarios

is required to account for this A dynamical ice sheetmodel run coupled with the climate model is alsorequired to answer these questions fully and sensitivityexperiments planned for future work will also help todetermine whether these effects are important

4 Conclusions

In this paper we present a summary of the present-day surface mass balance in Greenland Our results andthose reported in Langen et al (2017) show that the icesheet and peripheral glaciers are well represented at thepresent day but biases from global climate model forcingunderestimate the amount of melt and precipitationcurrently in Greenland compared to reanalysis drivensimulationsIn the future downscaling global climate model

simulations suggest that an increase in melt and runoffonly partially balanced by a small increase in precipita-tion is likely to lead to increasing mass loss from the icesheet with the total magnitude determined by theforcing due to greenhouse gas emissionsThe small size of peripheral glaciers means that

many of them may disappear entirely in this centuryHowever the fixed ice mask in this simulation meansthat neither the changing distribution of ice norelevation-related feedbacks such as ablation area in-crease due to declining elevation nor precipitationmigration due to changes in orography are consideredhere These are areas for future study

Data availability

HIRHAM5 simulation output is freely available athttp prudence dmi dkdatatempRUMHIRHAMGL2 (or contact the authors directly)

Acknowledgements

Research reported in this paper was supported inpart by the ERC Synergy project ice2ice (ERC grant610055) from the European Communityʼs SeventhFramework Programme (FP72007-2013) and theRETAIN project funded by the Danish Council forIndependent research (Grant no 4002-00234) Ruth

Mottram gratefully acknowledges the invitation to theinternational workshop ldquoGreenland ice sheet mass lossand its impact on global climate changerdquo under theSIGMA (Snow Impurity and Glacial Microbe effects onabrupt warming in the Arctic) project the ArCS (ArcticChallenge for Sustainability) Project and the JointResearch Program of the Institute of Low TemperatureScience Hokkaido University We thank Ralf Greve fora thorough review which substantially improved thepaper

References

Bamber J L R L Layberry and S P Gogineni (2001) A newice thickness and bed data set for the Greenland ice sheet1 Measurement data reduction and errors J GeophysRes Atmos 106 33773-33780 doi 1010292001JD900054Barletta V R L S Soslashrensen and R Forsberg (2013) Scatterof mass changes estimates at basin scale for Greenland andAntarctica Cryosphere 7 1411-1432 doi 105194tc-7-1411-2013Bolch T L Sandberg Soslashrensen S B Simonsen N Moumllg HMachguth P Rastner and F Paul (2013) Mass loss ofGreenlandʼs glaciers and ice caps 2003-2008 revealed fromICESat laser altimetry data Geophys Res Lett 40 875-881 doi 101002grl50270Buchardt S L H B Clausen B M Vinther and D Dahl-Jensen (2012) Investigating the past and recent 18 O-accumulation relationship seen in Greenland ice coresClim Past 8 2053-2059 doi 105194cp-8-2053-2012Christensen O B M Drews J H Christensen K Dethloff KKetelsen I Hebestadt and A Rinke (2006) The HIRHAMregional climate model version 5 DMI Technical Report06-17 URL https www dmi dkfileadminRapporterTRtr06-17pdfChurch J A P U Clark A Cazenave J M Gregory SJevrejeva A Levermann M A Merrifield G A Milne R SNerem P D Nunn A J Payne W T Pfeffer D Stammerand A S Unnikrishnan (2013) Sea level change In T FStocker D Qin G -K Plattner M Tignor S K Allen JBoschung A Nauels Y Xia V Bex and P M Midgley(Eds) Climate Change 2013 The Physical Science BasisContribution of Working Group I to the Fifth AssessmentReport of the Intergovernmental Panel on Climate ChangeCambridge University Press Cambridge United Kingdomand New York NY USA pp1137-1216Citterio M and A P Ahlstroslashm (2013) Brief communicationldquoThe aerophotogrammetric map of Greenland ice massesrdquoCryosphere 7 445-449 doi 105194tc-7-445-2013Dee D P S M Uppala A J Simmons P Berrisford P Poli S


Kobayashi U Andrae M A Balmaseda G Balsamo PBauer P Bechtold A C M Beljaars L van de Berg JBidlot N Bormann C Delsol R Dragani M Fuentes A JGeer L Haimberger S B Healy H Hersbach E V HoacutelmL Isaksen P Karingllberg M Koumlhler M Matricardi A PMcNally B M Monge-Sanz J-J Morcrette B-K Park CPeubey P de Rosnay C Tavolato J -N Theacutepaut and FVitart (2011) The ERA-Interim reanalysis configurationand performance of the data assimilation system Q J RMeteorol Soc 137 553-597 doi 101002qj828Doyle S H A Hubbard R S W van de Wal J E Box D vanAs K Scharrer T W Meierbachtol P C J P Smeets J THarper E Johansson R H Mottram A B Mikkelsen FWilhelms H Patton P Christoffersen and B Hubbard(2015) Amplified melt and flow of the Greenland ice sheetdriven by late-summer cyclonic rainfall Nat Geosci 8647-653 doi 101038ngeo2482Eerola K (2006) About the performance of HIRLAM version70 HIRLAM Newsl 51 93-102Fausto R S D van As J E Box W Colgan P L Langen andR H Mottram (2016) The implication of nonradiativeenergy fluxes dominating Greenland ice sheet exceptionalablation area surface melt in 2012 Geophys Res Lett 432649-2658 doi 1010022016GL067720Fettweis X E Hanna C Lang A Belleflamme M Erpicumand H Galleacutee (2013) Important role of the mid-troposphericatmospheric circulation in the recent surface melt increaseover the Greenland ice sheet Cryosphere 7 241-248Forbes R A M Tompkins and A Ungatch (2011) A newprognostic bulk microphysics scheme for the IFSTechnical Memorandum No649 ECMWFHanna E T Cropper R Hall and J Cappelen (2016)Greenland Blocking Index 1851-2015 a regional climatechange signal Int J Climatol 36 4847-4861 doi101002joc4673Hanna E F J Navarro F Pattyn C M Domingues XFettweis E R Ivins R J Nicholls C Ritz B Smith STulaczyk P L Whitehouse and H J Zwally (2013) Ice-sheet mass balance and climate change Nature 498 51-59Hazeleger W X Wang C Severijns S Ştefănescu RBintanja A Sterl K Wyser T Semmler S Yang B vanden Hurk T van Noije E van der Linden and K van derWiel (2012) EC-Earth V22 description and validation of anew seamless earth system prediction model Clim Dyn39 2611-2629 doi 101007s00382-011-1228-5IPCC (2013) Climate Change 2013 The Physical ScienceBasis Contribution of Working Group I to the FifthAssessment Report of the Intergovernmental Panel onClimate Change (Eds T F Stocker D Qin G-K PlattnerM Tignor S K Allen J Boschung A Nauels Y Xia VBex and P M Midgley) Cambridge University PressCambridge United Kingdom and New York NY USALangen P L R H Mottram J H Christensen F Boberg C

B Rodehacke M Stendel D van As A P Ahlstroslashm JMortensen S Rysgaard D Petersen K H Svendsen GAethalgeirsdoacutettir and J Cappelen (2015) Quantifying energyand mass fluxes controlling Godtharingbsfjord freshwater inputin a 5 km simulation (1991-2012) J Climate 28 3694-3713doi 101175JCLI-D-14-002711Langen P L R S Fausto B Vandecrux R H Mottram andJ E Box (2017) Liquid water flow and retention on theGreenland ice sheet in the regional climate modelHIRHAM5 Local and large-scale impacts Front Earth Sci4 110 doi 103389feart201600110Lucas-Picher P M Wulff-Nielsen J H Christensen GAethalgeirsdottir R Mottram and S B Simonsen (2012)Very high resolution regional climate model simulationsover Greenland Identifying added value J Geophys ResAtmos 117 D02108 doi 1010292011JD016267Machguth H H H Thomsen A Weidick A P Ahlstroslashm JAbermann M L Andersen S B Andersen A A Bjoslashrk JE Box R J Braithwaite C E Boslashggild M Citterio PClement W Colgan R S Fausto K Gleie S Gubler BHasholt B Hynek N T Knudsen S H Larsen S HMernild J Oerlemans H Oerter O B Olesen C J P PSmeets K Steffen M Stober S Sugiyama D van As M Rvan den Broeke and R S W van de Wal (2016) Greenlandsurface mass-balance observations from the ice-sheetablation area and local glaciers J Glaciol 62 861-887 doi101017jog201675Mernild S H G E Liston C A Hiemstra K Steffen EHanna and J H Christensen (2009) Greenland Ice Sheetsurface mass-balance modelling and freshwater flux for2007 and in a 1995-2007 perspective Hydrol Process 232470-2484 doi 101002hyp7354Noeumll B W J van de Berg H Machguth S Lhermitte IHowat X Fettweis and M R van den Broeke (2016) Adaily 1 km resolution data set of downscaled Greenland icesheet surface mass balance (1958-2015) Cryosphere 102361-2377 doi 105194tc-10-2361-2016Nielsen-Englyst P (2015) Impact of Albedo Parametrizationson Surface Mass Balance Runoff on the Greenland IceSheet Master Thesis University of CopenhagenRae J G L G Aethalgeirsdottir T L Edwards X Fettweis JM Gregory H T Hewitt J A Lowe P Lucas-Picher R HMottram A J Payne J K Ridley S R Shannon W J vande Berg R S W van de Wal and M R van den Broeke(2012) Greenland ice sheet surface mass balance evaluatingsimulations and making projections with regional climatemodels Cryosphere 6 1275-1294 doi 105194tc-6-1275-2012Rignot E I Velicogna M R van den Broeke A Monaghanand J T M Lenaerts (2011) Acceleration of the contribu-tion of the Greenland and Antarctic ice sheets to sea levelrise Geophys Res Lett 38 L05503 doi 1010292011GL046583


Roeckner E G Baumluml L Bonaventura R Brokopf M EschM Giorgetta S Hagemann I Kirchner L Kornblueh EManzini A Rhodin U Schlese U Schulzweida and ATompkins (2003) The atmospheric general circulationmodel ECHAM 5 Part I Model description TechnicalReport No349 Max Planck InstituteSchmidt L G Aethalgeirsdoacutettir S Guethmundsson P L LangenF Paacutelsson R Mottram S Gascoin and H Bjoumlrnsson(submitted) Evaluating the surface energy balance in theHIRHAM5 regional climate model over VatnajoumlkullIceland using automatic weather station data Cryosphere(submitted MS No tc-2017-14)Shepherd A E R Ivins G A V R Barletta M J Bentley SBettadpur K H Briggs D H Bromwich R Forsberg NGalin M Horwath S Jacobs I Joughin M A King J T MLenaerts J Li S R M Ligtenberg A Luckman S BLuthcke M McMillan R Meister G Milne J Mouginot AMuir J P Nicolas J Paden A J Payne H Pritchard ERignot H Rott L S Sorensen T A Scambos B ScheuchlE J O Schrama B Smith A V Sundal J H van Angelen

W J van de Berg M R van den Broeke D G Vaughan IVelicogna J Wahr P L Whitehouse D J Wingham D YiD Young and H J Zwally (2012) A reconciled estimate ofice-sheet mass balance Science 338 1183-1189 doi101126science1228102Stibal M E Goumlzdereliler K A Cameron J E Box I TStevens J K Gokul M Schostag J D Zarsky A EdwardsT D L Irvine-Fynn and C S Jacobsen (2015) Microbialabundance in surface ice on the Greenland Ice Sheet FrontMicrobiol 6 225 doi 103389fmicb201500225Van As D R S Fausto J Cappelen R S W van der Wal R JBraithwaite H Machguth C Charalampidis J E Box AM Solgaard A P Ahlstroslashm K Haubner M Citterio and SB Andersen (2016) Placing Greenland ice sheet ablationmeasurements in a multi-decadal context Geol Surv DenGreenl Bull 35 71-74Vernon C L J L Bamber J E Box M R van den Broeke XFettweis E Hanna and P Huybrechts (2013) Surface massbalance model intercomparison for the Greenland ice sheetCryosphere 7 599-614 doi 105194tc-7-599-2013


Surface mass balance of the Greenland icesheet in the regional climate model HIRHAM5

Present state and future prospectsRuth Mottram1 Fredrik Boberg1 Peter Langen1 Shuting Yang1

Christian Rodehacke1 Jens Hesselbjerg Christensen1 Marianne Sloth Madsen1

Received 1 February 2017 accepted 9 February 2017

Surface mass balance (SMB) is the builder of the Greenland ice sheet and the driver of icedynamics Quantifying the past present and future state of SMB is important to understand thedrivers and climatic processes that control SMB and to both initialize and run ice sheet models whichwill help clarify sea level rise and how likely changes in ice sheet extent feedback within the climatesystem

Regional climate models (RCMs) and climate reanalysis are used to quantify SMB estimatesAlthough different models have different spatial and temporal biases and may include differentprocesses giving significant uncertainty in both SMB and the ice sheet dynamic response to it allRCMs show a recent declining trend in SMB from the Greenland ice sheet driven primarily byenhanced melt rates Here we present new simulations of the Greenland ice sheet SMB at 5 kmresolution from the RCM HIRHAM5 The RCM is driven by the ERA-Interim reanalysis and theglobal climate model (GCM) EC-Earth v23 to make future projections for climate scenarios RCP85and RCP45

Future estimates of SMB are affected by biases in driving global climate models and feedbacksbetween the ice sheet surface and the global and regional climate system are neglected likely resultingin significant underestimates of melt and precipitation over the ice sheet These challenges will needto be met to better estimate the role climate change will have in modulating the surface mass balanceof the Greenland ice sheet

Keywords Greenland ice sheet surface mass balance climate climate modelling ice sheet modelling

1 Introduction

The Greenland ice sheet is famously the secondlargest land ice mass on the planet containing around285 million km3 of ice over an area of 171 million km2equivalent to 72 m sea level rise (IPCC 2013) Resultsfrom remote sensing missions such as the GRACE

gravimetry satellites indicate a recent significantnegative mass budget with around -234plusmn20 Gt of netmass (Barletta et al 2013) being lost each year includingboth the contribution from precipitation and surfacemass balance and the dynamic contribution of calvingfrom icebergs and submarine melt (Shepherd et al 2012)The surface mass balance component is by far the mostimportant component of the mass budget since itincludes both positive (accumulation by precipitation)and negative (melt and runoff) terms though around onethird to one half of the mass lost by the ice sheet is fromcalving glaciers (Enderlin et al 2014)

低温科学 75 （2017） 105-115doi 1014943lowtemsci 75 105

Corresponding authore-mailrumdmidk) Danish Meteorological Institute CopenhagenDenmark































M1048637QmwL f (1)














































4 Conclusions




Data availability


Acknowledgements



References







































M1048637QmwL f (1)














































4 Conclusions




Data availability


Acknowledgements



References

















































4 Conclusions




Data availability


Acknowledgements



References









surface mass balance of the greenland ice sheet in …...instructions for use title surface mass...

Documents