challenges of modelling surface mass balance - esa...

9/15/16

1

Challengesofmodellingsurfacemassbalance

MichielvandenBroekeInstituteforMarineandAtmosphericResearch,UtrechtUniversity(IMAU)

AdvancedTrainingCourseonRemoteSensingoftheCryosphereLeeds(UK),12-16September2016

Howdowedefinesurfacemassbalance(SMB)?

Here:thesumofsurfaceand internalmassbalance(climatic massbalance orfirnmassbalance)

Ligtenberg,PhDthesis,2014

9/15/16

2

MassBalance=SurfaceMassBalance– DischargeMB=dM/dt =SMB– D

SMBischallenging:notone,butthreebalances!

Icesheetmassbalance(MB)MB=Surfacemassbalance– Discharge [Gt yr-1]

Surfacemassbalance(SMB)SMB=Precipitation – Sublimation – Runoff- Erosion [Gt yr-1]

Liquid waterbalance(LWB)Runoff=Rain+Condensation +Melt– Refreezing– Retention [Gt yr-1]

Surfaceenergybalance(SEB)M=SWnet +LWnet +H+L+Gs [Wm-2]

J.PaulGettyMuseum

9/15/16

3

Combine EO,insituobservations andSMBmodelling

Remotesensing

Climatemodel

Insituobservations

EvaluationCalibration

Spatial/temporal extension

Firnm

odel

GRACEmasstrends(2003-2012)

Courtesy:BertWouters

9/15/16

4

13yearsoficesheetmasschangesfromGRACE

Velicogna andothers,2015

SMB:snowfalldriven

Greenlandicesheet

SMB:snowfallandrunoffdriven

Antarcticicesheet

The‘simpler’case:

Antarctica

NASA/MODIS

9/15/16

5

Masschangesin2009fromGRACE

MartinHorwath

AddedvalueofSMBmodelling:acasestudy forAntarctica

Elevationchangesin2009fromEnvisat

MartinHorwath

AddedvalueofSMBmodelling:acasestudy forAntarctica

9/15/16

6

is forced using fields of temperature, specific humidity,

zonal and meridional wind components, and surface pres-sure from either GCM or re-analysis output. Relaxation of

RACMO2 prognostic variables towards external forcings is

restricted to the boundary relaxation zone (Fig. 1). Externalforcings are updated every six hours and linearly interpo-

lated in time to yield accurate values in between. Sea

surface temperatures and sea-ice extent are also prescribedfrom the forcing model. The version of RACMO2 used for

this study includes a snow model that calculates tempera-ture, density and meltwater processes (percolation, reten-

tion, refreezing and runoff) in the snow (Ettema et al.

2009), and an improved albedo scheme, where the snowalbedo depends on snow grain size (Kuipers Munneke

et al. 2011). For this study, contributions from drifting

snow processes have not been included, because themodule of Lenaerts and Van den Broeke (2012) was not yet

fully implemented when we started the simulations.

For contemporary climate studies of the AIS (1–30 years),RACMO2 has been run on grids with 27 and 5.5 km hori-

zontal resolution (Lenaerts et al. 2012a, b). However, for the

number of simulation years considered here (660 years intotal), a horizontal resolution of 55 km is considered a good

trade-off between computational expense and spatial detail;

doubling the grid resolution would multiply the computa-tional time by a factor 10. Moreover, the annual integrated

SMB of the AIS at 55 km resolution (Van de Berg et al.

2006) is similar to that at 27 km resolution (Lenaerts et al.2012a). For the scenario runs, the largest uncertainty there-

fore derives not from the model resolution but from the

chosen forcing model and scenario. Given this information,and the fact that a 27 km resolution run is ten times as

expensive as a 27 km run, we chose 55 km as final resolu-tion. The model topography, grid resolution and lateral

relaxation boundary of the domain are shown in Fig. 1.

For the period 1980–1999, a RACMO2 reference sim-ulation, forced by ERA-40 re-analysis data from the

European Centre for Medium-Range Weather Forecasts

(Uppala et al. 2005), was performed in order to check thereliability of the GCM-forced RACMO2 simulations. In

this paper, ERA-40 has been used as forcing instead of its

successor ERA-Interim (Dee and et al. 2011), since thelatter only covered the period 1989–2009 at the time the

RACMO2 simulations were started. Other RACMO2

simulations forced by re-analysis data (ERA-40 or ERA-Interim) yielded realistic SMB results over Antarctica

Fig. 1 Map of Antarcticashowing the model domain, theboundary relaxation zone(dotted area) and modeltopography in meters above sealevel

Future SMB of Antarctica

123

Regionalclimatemodel

domain

27kmresolution

Courtesy:MelchiorvanWessem

Modelled3-hourly

precipitation

Run:RACMO2.3/ANTResolution:27km

9/15/16

7

!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.!.

$+

$+

$+

#*

$+

!(

#*

!.

!(

!(

#*

!.

!(!(!(

#*

#*

$1

!(

#*

!(!(

!(

#*

!(

#*

#*#*

!(

$1$1

!(

!(

!.

!(

$1

!(#*!(!(!(#*

#*

!(

!(

#*

!(

!(#*

#*!(

!(

!(

!(!(

#*

#*

#*

#*

!(

#*#*

!($1#*

!(

$1

!(

!(

#*

#*

!(!(

!(

!(

#*

!(

#*

$1

#*

!(

#*

!(

#*

!(

$1

!(!(!(!(!(

#*

!(

#*

!(

!(

#*

!(!(

!(

#*

!(

#*

!(

!(

#*

!(#*

#*#*!($1

#*

!(

#*

!(#*!(

#*

!(

!(

!(

!($1

#*

#*

!(

!(

#*

!(

#*

!(

#*#*

!(

#*

!(

!(

#*

#*

!(

#*

!(

!(

!(

!(

!(

!(

$1

#*

"/

#*

!(

"/

!(

!(!(

!(

#* #*$1$1

#*

!(

!(!(

#*#*

!(

!(

Schwerdtfeger

Modesta Sofia-B

Kominko-Slade (WAIS)

Lindsey Island

PANDA-NorthEagleDome A

Mt Brown

WhoopWhoop

CapePoinsett

SnyderRocks

CaseySkiway

Law Dome

HauptNunatak

See Inset Map B

See Inset Map A

Dome C (FRA)

Baldrick

PANDA-South

Cape Hallett

JASE2007

D-85

Summit/A

Halvfarryggen

Hugo Is

Margaret

Skinner Saddle

Darwin GlCentral Valley

Dome C II

Bisco IslandAWS 1

Larsen Ice Shelf

Thiel Mountains

D-0

D-3D-17

DC N

Dome Fuji

Atka Bay

D-47

Halley V

Utsteinen

Thurston Island

Evans Knoll

Bear PeninsulaSabrina

Atka &

Ekstrom &

Leningradskaya

Russkaya

Molodeznaya

Druznaja-4

Paola

Cierva CoveAmsler Island

Korff Ice Rise

Troll

Nordenskiold

AWS 18 Windy

Site 8

Emma

Austin

Kathie

Filchner

Gerlache StraitNeko Harbor

Dome C

Vito

Nico

Gill

Erin

D-10

Byrd

Henry

Harry

Wilkins RunwayWest 4m / East ICAO

AGO-5

AGO-4

AGO-3

AGO-2

AGO-1

Mizuho

Lettau Elaine

Sky-Blu

Limbert

Soerasen

Elizabeth

Dome Fuji (JPN)

Concordia

Ekstrom Shelf Ice

Siple DomeJanet

Port Martin

Kohnen Base

Irene

Giulia

Cp Denison

Relay Station

Dismal Is

Butler Is

Fossil Bluff

Bonaparte Pt

Camp Maudheimvida

AWS 17AWS 14

Possession Is

Joinville Is

Union Glacier

Theresa

Marilyn

!(

!(

!(!(!(!(!(!(!(

!(

!(#*

!(

!(

!(

!(

!(

#*

!(

Arelis

Inset Map A

!.

!.

!.!.!.!.!.!.!.

!.

!.

!.

!.

!.!.

!.

#*

!(

$1

!(

!(

#*

$1

$1

$1

#*

!(

$1

#*

$1

!(

!(

$1

$1

#*

#*

#*

#*#*!($1

!(

#*!(

!($1

#*

$1

!(!(

#*$1

$1#*

#*

EmiliaLinda

Minna Bluff

Arelis

Cp Bird

Marble Pt

Butter Pt -Ferrar

Jules

Laurie II

FerrellBiesiada Crevasse

Cp Royds

Herbie Alley N

Tent Is

WindlessBight

N Crevasse

Cp Spencer

Willie Field

Pegasus N

Ford RockRoss Ice Shelf White Out

Lorne

White Is S

Bull Pass

Minna Bluff

Bratina Island

Miers ValleyGarwood Valley

Mt Fleming

Granite Harbour

Victoria Valley

Victoria Lower Gl

Evans Piedmont Gl

Cones

AlexanderTall Tower

LTER AWS (21)

Marble Pt II

White Is

Alessandra

Silvia

Lola

Zoraida

EneideRita

0 100 km

Inset Map B

0 50 km 0 500 km

Coastline: ADD v4.1, 2003; Cartography: April 2016 Sam Batzli, SSEC, University of Wisconsin-Madison; Funding: National Science Foundation ANT-0944018

Manuela

Univ. of Wisconsin (UW)

Brazil

Italy

Netherlands

!(

!(

China/Australia

!(

!(

#* Australia!(

SPAWAR$1$+ AGO

United States AWS International AWS

United Kingdom!(

Russia!(

!. Other US

South Korea!(

Automatic Weather StationsAntarctica - 2016

#*

#*#*

#*

UW / AustraliaUW / ChinaUW / FranceUW / Japan

Commercial"/

Other AWS France!(

Japan!(

Jang Bogo

Whitlock

Belgium/Netherlands!( New Zealand!(

Germany!(

Lucia

Penguin

Maria

Finland!(

China!(

China

#* UW / New Zealand#* UW / United Kingdom

Norway!(

Virginia

UniversityofWisconsin,Madison

kgm-2 y-1

Circles:≈1200 obs.

Data:Arthern andothers,2006

Observation-based accumulationmap

9/15/16

8

ModeledAntarcticsurfacemassbalance

kgm-2 y-1

Run:RACMO2.1/ANTPeriod:1979-2011Resolution:27kmCircles:≈1200obs.

Data:Lenaertsandothers,2012

FilterSMBfieldfromregionalclimate modelRACMOtosameresolution asGRACE

CompareSMBtoGRACEmassanomalies

MartinHorwath

9/15/16

9

MartinHorwath

Convertsnowfallanomaliestofirnthickness change,usingafirn model

Comparemodelled snowdepthchangetoaltimetry

MartinHorwath



AmundsenSeaembayment

DronningMaudLand

9/15/16

10

DronningMaudLand,EastAntarctica

Lenaertsandothers,2013

a) SMBfromregionalclimatemodelb) CumulativeSMBanomalyvs.GRACE

MartinHorwath



AmundsenSeaembayment

DronningMaudLand

9/15/16

11

Amundsen Seaembayment,WestAntarctica

Sutterley andothers,2014

a) SMBandDb) MB=SMB– Dc) CumulativeMB

The‘harder’case:

Greenland

NASA/MODIS

9/15/16

12

13yearsoficesheetmasschangesfromGRACE

Velicogna andothers,2015

SMB:snowfalldriven

Greenlandicesheet

SMB:snowfallandrunoffdriven

Antarcticicesheet

SMBischallenging:notone,butthreebalances!

Icesheetmassbalance(MB)MB=Surfacemassbalance– Discharge [Gt yr-1]

Surfacemassbalance(SMB)SMB=Precipitation – Sublimation – Runoff- Erosion [Gt yr-1]

Liquid waterbalance(LWB)Runoff=Rain+Condensation +Melt– Refreezing– Retention [Gt yr-1]

Surfaceenergybalance(SEB)M=SWnet +LWnet +H+L+Gs [Wm-2]

J.PaulGettyMuseum

9/15/16

13

Greenlandautomaticandmannedweatherstations

AdaptedfromGEUS

K-transect,westGreenland

S10

NASA/MODIS

9/15/16

14

Greenlandsurfaceenergybalance

Courtesy:PeterKuipers Munneke

S5:500masl

S6:1000masl

S9:1500masl

S10:1850maslStationS6(1000masl)

Regionalclimatemodel

11kmresolution

9/15/16

15

Modelevaluation:surfaceenergy

balance

Noëlandothers,2015

TCD9, 1177–1208, 2015

Impact of summersnowfall events on

GrIS SMB

B. Noël et al.

Title Page

Abstract Introduction

Conclusions References

Tables Figures

J I

J I

Back Close

Full Screen / Esc

Printer-friendly Version

Interactive Discussion

Discussion

Pa

per

|D

iscussion

Pa

per

|D

iscussion

Pa

per

|D

iscussion

Pa

per

|

(a) (b)

(c) (d)

(e) (f)

(g) (h)

Figure 5. Absolute values of observed and modelled turbulent and net shortwave/longwavefluxes (Wm�2) at station (a) S5 for 2004–2012, (c) S9 for 2004–2008, (e) S9 for 2009–2012 and(g) S10 for 2010–2012; di◆erence in modelled and observed surface albedo and surface meltenergy (Wm�2) at stations (b) S5, (d) S9, (f) S9 and (h) S10 for the same periods, respectively.

1205

TCD9, 1177–1208, 2015

Impact of summersnowfall events on

GrIS SMB

B. Noël et al.

Title Page

Abstract Introduction

Conclusions References

Tables Figures

J I

J I

Back Close

Full Screen / Esc

Printer-friendly Version

Interactive Discussion

Discu

ssio

nP

ap

er

|D

iscu

ssio

nP

ap

er

|D

iscu

ssio

nP

ap

er

|D

iscu

ssio

nP

ap

er

|

(a) (b)

(c) (d)

(e) (f)

(g) (h)

Figure 5. Absolute values of observed and modelled turbulent and net shortwave/longwavefluxes (Wm�2) at station (a) S5 for 2004–2012, (c) S9 for 2004–2008, (e) S9 for 2009–2012 and(g) S10 for 2010–2012; di◆erence in modelled and observed surface albedo and surface meltenergy (Wm�2) at stations (b) S5, (d) S9, (f) S9 and (h) S10 for the same periods, respectively.

1205

500masl 1000masl

1500masl 1850masl

Modeledevaluation:surfacemass balance

kgm-2 y-1

Run:RACMO2.3/GRPeriod:1958-2014Resolution:11kmNoëlandothers,2016

9/15/16

16

Greenlandcumulativemassbalance(1958-2014)

VandenBroekeandothers,2016

Modelled Greenlandicesheetmelt/runoff (1958-2014)

0

200

400

600

800

1000

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Mas

s flu

x (G

t yr-1

)

Year

Melt

Runoff

Melt: 12.1 +/- 3.6 Gt yr-1

Runoff: 9.6 +/- 2.4 Gt yr-1

Refreezing: 3.0 +/- 1.7 Gt yr-1

Trends 1991-2014

Refreezing

Surface sublimation

Typical uncertainties


9/15/16

17

Modelled Greenlandicesheetmassbalance(1958-2014)

-400

-200

0

200

400

600

1960 1965 1970 1975 1980 1985 1990 1995 2000 2005 2010 2015

Mas

s flu

x (G

t yr-1

)

Year

Solid ice discharge

Surface mass balance

Discharge: 6.7 +/- 0.4 Gt yr-1

SMB: -10.7 +/- 2.9 Gt yr-1

Mass balance: -17.4 +/- 3.0 Gt yr-1

Trends 1991-2014

Ice sheet mass balance

Typical uncertainties


Byrdglacier,Antarctica

9/15/16

18

Impactofmodel resolutiononnetsnowaccumulation

Lenaerts and others: Impact of model resolution on simulated wind, drifting snow and SMB 823

Fig. 2. Modeled annual mean (2009) 10m wind speed (contours) and direction (arrows) in RACMO/5.5 (left) and RACMO/27 (right). Thelocations of the four AWSs in Figure 4a–d are indicated by their respective letters.

Drifting snow measurements are sampled at the TalosDome TALDICE drilling site (72◦ S, 159◦ E, 2315ma.s.l.;http://www.taldice.org; see Fig. 1 for location). Measure-ments are obtained with FlowCaptTM driftometer sensorsproduced by ISAW Outdoor Environmental Monitoring(Chritin and others, 1999). The instrument is composedof four sensors: two of these are placed 0.2m above thesnow surface, and the other two 1m above the surface.These instruments provide inaccurate estimates of the snowtransport fluxes (Cierco and others, 2007), but do provide arealistic estimate of the occurrence of drifting snow.Modeled SMB is compared with observations described

by Agosta and others (2012). The observations originatefrom a ∼150 km long stake line that runs from the coast ofTerre Adelie to the southwest (0–1800ma.s.l.). ComparingRACMO to these observations can be regarded as a stringenttest for model performance, because the stake line covers thestrong SMB gradient between the relatively mild and windycoastal climate and drier and calmer conditions inland athigh spatial resolution (100 data points).

RESULTSWind climateFigure 2 compares annual mean 10m wind speed ofRACMO/27 with RACMO/5.5. Although RACMO/5.5 obvi-ously shows much more detail, the regional patterns aresimilar. We find four areas of strong (>14ms−1) winds:three over outlet glaciers (Byrd, Mulock and Reeves/DavidGlaciers) in the Transantarctic Mountains, with several jetsabove 10ms−1, and one in coastal Terre Adelie, witha maximum wind speed of 16m s−1 at ∼69◦ S, 143◦ E,1100ma.s.l. On the smaller scale, the RACMO/5.5 windfield shows distinct features. Maximum wind speeds arehigher and occur closer to the grounding line. Relativelynarrow (<20 km) glacial valleys, in which the katabatic windspeeds converge and accelerate, are much better resolved at

5.5 km. In our simulation the strongest winds are found in theglacial valley of Reeves Glacier, with a maximumwind speedof 20.5m s−1, and, to a lesser extent, David Glacier nearthe Italian base (Mario Zuchelli) in Terra Nova Bay (∼75◦ S,163◦ E), and Byrd Glacier (81◦ S, 158◦ E). The occurrenceof these maxima is supported by results of Bromwich andothers (1990), who showed that Reeves Glacier is the primaryroute for katabatic winds, and David Glacier is an importantsecondary outflow valley.Lenaerts and others (2012b) showed that RACMO/27

underestimates high wind speeds in regions with complextopography. Figure 3 illustrates that in these regions, windspeeds in RACMO/5.5 agree better with observations. Theroot-mean-square error (rmse) decreases from 5.7m s−1 to3.6m s−1, and the mean bias between model and obser-vations drops from −4.3m s−1 to −1.4m s−1. Nonetheless,the extreme wind speeds (>15m s−1) in Terre Adelie remainunderestimated.Figure 3 only shows long-term mean near-surface wind

speeds. Drifting snow processes, however, are usuallyconnected to short-lived wind speed maxima. To evaluatethe model results at higher temporal resolution, Figure 4shows the daily mean 10m wind speed from RACMO/5.5,RACMO/27 and from available AWS observations. Due tolimited temporal coverage and large gaps in the data, theseAWSs are not included in Figure 3. At daily resolution,RACMO/5.5 shows clearly higher maximal wind speeds thanRACMO/27 at all stations, except for Sitry (Fig. 3b), wheretopography is smooth and the model agrees very well withthe observations, even at 27 km resolution. The other threestations are known to be major confluence areas, wherethe katabatic wind accelerates due to the convex shape ofthe glacier valley (Bromwich and others, 2000). At LarsenGlacier, the modeled timing and frequency of wind speedmaxima agree very well with the observations, whereas atPriestley and David Glaciers, observed wind speed maximaremain largely underestimated, also by RACMO/5.5. Theintense local katabatic flow at these locations is likely driven

Byrd glacier basin



Byrd glacier basin


9/15/16

19



Enhanced dynamical downscaling

Courtesy:MelchiorvanWessem

Run:RACMO2.3/APPeriod:1979-2014Resolution:5.5kmCircles:≈120obs.

9/15/16

20

EastGreenland:individual glaciersnotresolved

11kmresolution 1kmresolution

Noëlandothers,2016

Enhanced dynamical downscaling

Courtesy:WillemJanvandeBerg

SMB

9/15/16

21

Enhanced statistical downscaling

RACMO2.3(11km),EastGreenland Statisticallydownscaledto1km

(BasedonGIMP,Howatandothers,2014)

SMB(mm yr-1)

Noëlandothers,2016

ModeledGreenlandsurfacemass balance

kgm-2 y-1

Run:RACMO2.3/GRPeriod:1958-2014Resolution:11kmNoëlandothers,2016

9/15/16

22

Outlook

ObservationsFocusonaccurateradiation andturbulence measurements!Newgenerationofautonomousweather stationsNewgenerationofsatellites:GRACE-2,ICESat-2

ModelsFurther improveexistingatmospheric climatemodels(clouds)Further improveexistingsnowmodels(heterogeneous percolation)Movetoglobalmodelsystems(coupledicesheetmodels)Improveprognostic albedoschemes(dust,blackcarbon, bio-albedo)

challenges of modelling surface mass balance - esa...

Documents