Bathymetric and oceanic controls on Abbot Ice Shelf thickness and stability

13 Pages·2014·3.72 MB·English

TheCryosphere,8,877–889,2014 O
p
www.the-cryosphere.net/8/877/2014/ e
n
doi:10.5194/tc-8-877-2014 The Cryosphere A
c
c
©Author(s)2014.CCAttribution3.0License. e
s
s
Bathymetric and oceanic controls on Abbot Ice Shelf thickness and
stability
J.R.Cochran,S.S.Jacobs,K.J.Tinto,andR.E.Bell
Lamont-DohertyEarthObservatoryofColumbiaUniversity,Palisades,NY10964,USA
Correspondenceto:J.R.Cochran([email protected])
Received:29October2013–PublishedinTheCryosphereDiscuss.:19November2013
Revised:14March2014–Accepted:26March2014–Published:15May2014
Abstract. Ice shelves play key roles in stabilizing Antarc- 1 Introduction
tica’s ice sheets, maintaining its high albedo and returning
freshwatertotheSouthernOcean.Improveddatasetsofice
shelf draft and underlying bathymetry are important for as- IceshelvesarefoundalongmuchoftheAntarcticcoastline
sessingocean–iceinteractionsandmodelingiceresponseto andmakeupabout11%ofthesurfaceareaofAntarcticice
climate change. The long, narrow Abbot Ice Shelf south of (Fox and Cooper, 1994). These floating extensions of the
ThurstonIslandproducesalargevolumeofmeltwater,butis ice sheets play an important stabilizing role by buttressing
closetobeinginoverallmassbalance.HereweinvertNASA andslowingtheflowofinlandiceacrossthegroundingline
OperationIceBridge(OIB)airbornegravitydataovertheAb- (DupontandAlley,2005).Forexample,glaciersfeedinginto
botregiontoobtainsub-icebathymetry,andcombineOIBel- theLarsenAandaportionoftheLarsenBacceleratedwhen
evationandicethicknessmeasurementstoestimateicedraft. thoseiceshelvescollapsed(Rottetal.,2002;DeAngelisand
Aseriesofasymmetricfault-boundedbasinsformedduring Skvarca,2003;Scambosetal.,2004),whileglaciersfeeding
rifting of Zealandia from Antarctica underlie the Abbot Ice theremnantoftheLarsenBinScarInletdidnotaccelerate
Shelfwestof94◦WandtheCosgroveIceShelftothesouth. (Rignotetal.,2004).
Sub-icewatercolumndepthsalongOIBflightlinesaresuffi- Many West Antarctic ice shelves, particularly within the
cientlydeeptoallowwarmdeepandthermoclinewatersob- Amundsen Sea Embayment, immediately to the west of the
servednearthewesternAbboticefronttocirculatethrough AbbotIceShelf,haveexperiencedrapidthinning(e.g.,Rig-
much of the ice shelf cavity. An average ice shelf draft of not, 1998; Shepherd et al., 2004; Pritchard et al., 2012).
∼200m, 15% less than the Bedmap2 compilation, coin- Much of this thinning has been attributed to bottom melt-
cideswiththesummertransitionbetweentheoceansurface ing resulting from relatively warm Circumpolar Deep Wa-
mixedlayerandupperthermocline.Thickicestreamsfeed- ter (CDW) circulating on the continental shelf and beneath
ingtheAbbotcrossrelativelystablegroundinglinesandare those ice shelves (e.g., Jacobs et al., 1996, 2011, 2013;
rapidlythinnedbythewarmestinflow.Whiletheiceshelfis Jenkins et al., 2010; Hellmer et al., 1998), and to unpin-
presentlyinequilibrium,theoverallcorrespondencebetween ning from the sub-ice bathymetry (Jenkins et al., 2010).
draftdistributionandthermoclinedepthindicatessensitivity Pritchard et al. (2012) indicated that the Abbot is not thin-
to changes in characteristics of the ocean surface and deep ning, but assigned it a −1.01myr−1 basal melt imbalance
waters. for 2003–2008. Two studies combining satellite measure-
mentsandfirnmodelingsuggesttheAbbothasrecentlybeen
relatively stable, but differ as to whether it is gaining or
losing mass. Rignot et al. (2013) reported an area-average
ice thickening of 0.16±0.6myr−1 and basal melting of
1.75±0.6myr−1,whereasDepoorteretal.(2013)estimated
<1myr−1 thinningwithbasalmeltingof2.72±0.7myr−1.
Timmermannetal.(2012)andKusaharaandHasumi(2013)
PublishedbyCopernicusPublicationsonbehalfoftheEuropeanGeosciencesUnion.
878 J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability
2 BathymetrybeneaththeAbbotIceShelf
TheAbbotIceShelf,withasurfaceareaofabout30000km2
includingicerises(Swithinbanketal.,2004;Shepherdetal.,
2004;Rignotetal.,2013),isroughlythesizeofBelgiumand
extends for 450km along the Eights Coast of West Antarc-
ticabetween103and89◦W,withanaveragewidthofabout
65km(Fig.1).Westof95◦300W,ThurstonIslandboundsthe
northernsideoftheiceshelf,whiletheeasternhalfispinned
at the ice front by three islands and extends to the Fletcher
Peninsula(Figs.1and2).Thewesternedgeoftheiceshelf
facestheAmundsenSeaEmbayment.TheBedmap2compi-
lation(Fretwelletal.,2013)indicatesthaticethicknessover
Fig.1.MODISMOAimageoftheAbbotIceShelffrom2004show- most of the ice shelf is in the range of 200–325m. The nu-
ingthelocationoffeaturesmentionedinthetext.Contoursarefrom merousislandsandicerisesevidentinFig.1suggestacom-
theBamberetal.(2009)DEM,contouredat100mintervals.Inset plexbathymetrybeneaththeAbbot.
showsitspositionontheWestAntarcticcoastline
2.1 GravityanomaliesovertheAbbotIceShelf
modeled remarkably similar basal melt rates of 2.1myr−1
and 2.01myr−1, respectively. All of these rates are lower NASA’sOIBprogramobtained10north–south,low-altitude
airborne geophysical lines across the Abbot Ice Shelf and
thanhavebeencalculatedduringthesametimeframeforthe
an east–west line along the axis of the ice shelf during the
comparably sized George VI Ice Shelf to the east and Getz
2009OIBAntarcticcampaign(Fig.2).Spacingbetweenthe
to the west (Rignot et al., 2013), a sign of regionally vary-
north–south flights varied from 30 to 54km with a mean
ing influences on the mass balance of southeastern Pacific
separation of 39.2 km. Instrumentation on these lines in-
iceshelves.
cluded a laser altimeter, a variety of ice-penetrating radars
Knowledgeofthegeometryoftheseawatercavityunder-
and a gravimeter. With no OIB lines east of Farwell Island
lying an ice shelf is important for effective modeling of the
(∼91◦W,Fig.1),wedonotincludetheeasternmostportion
sub-ice-shelfcirculationsandassessmentofshelfvulnerabil-
oftheAbbotinouranalysis.
ity to bottom melting. Altimetry measurements of ice shelf
Surface elevation along the profiles was measured with
surfaceelevationmaybeusedtodetermineicethicknesswith
theNASAAirborneTopographicMapper(ATM)laser(Kra-
the assumption of hydrostatic equilibrium and a model for
bill, 2010), which can recover elevations with an accuracy
firndensificationwithdepth(e.g.,GriggsandBamber,2011).
of about 10cm (Krabill et al., 2002). Ice thickness was
Radarmethodscanprovideicethickness(e.g.,Hollandetal.,
determined using the Multichannel Coherent Radar Depth
2009) but cannot penetrate through the underlying water to
Sounder (MCoRDS) radar operated by the University of
maptheseafloorbathymetry.Seismictechniques(e.g.,Jarvis
Kansas Center for Remote Sensing of Ice Sheets (CRESIS)
andKing,1995;JohnsonandSmith,1997)anddirectobser-
(Leuschen,2011).Thebaseoftheiceshelfwasnotimaged
vationswithautonomousunderwatervehicles(Jenkinsetal.,
in the radar data for the very westernmost portion (west of
2010) have been used to map ice cavity morphology. Both
12km)ofourline0alongtheiceshelfaxisandformostof
methodsarelocallyaccuratebutlogisticallydifficultandex-
line1acrossthewesternmosticeshelf.Thisislikelydueto
tremelytime-consumingtoobtainreasonablecoverage.
thepresenceofmarineiceneartheicefront.Weusedleast-
NASA’s Operation IceBridge (OIB) has systematically
squares regression to determine the relationship of surface
collected airborne gravity data over ice shelves, observa-
elevationtodraftinareaswherethebaseoffloatingicewas
tionsthatcanbeinvertedforseafloorbathymetrywhencom-
clearly observed, and then utilized that relationship to esti-
binedwithaltimetryandradarmeasurements(TintoandBell,
mate the draft in the areas where the base of ice is not ob-
2011; Cochran and Bell, 2012; Muto et al., 2013). Here we
served.Ina15kmlengthofline1wherethebaseoftheice
invert OIB airborne gravity data to estimate the bathymetry
wasobserved,ourestimatedvalues,at0.2kmintervalsalong
beneaththeAbbotIceShelf(Fig.1).Wecombinethegravity-
the flight line, all agree with the observed draft to within
derivedbathymetrywithicethicknessdataandnearbyocean
10m. In our final data set, mismatches in ice draft at line
temperatureprofiles(GiuliviandJacobs,1997;Jacobsetal.,
crossings over floating ice varied from 0.8 to 26.1m with a
2011) to infer the presence of warm CDW and thermocline
standard error of 5.79m. The laser and radar data were ac-
watersbeneaththeiceshelf,anddiscussprobableice–ocean
quiredonthesameflightsasthegravitydata,sothedatasets
interactions.
arecoincidentintimeandspace.
OIBgravitydata(CochranandBell,2010)wereobtained
with a Sander Geophysics AIRGrav airborne gravimeter
TheCryosphere,8,877–889,2014 www.the-cryosphere.net/8/877/2014/
J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability 879
-100˚ -95˚
-105˚ -90˚
-110˚ BGA -85˚
-70˚ -70˚
-71˚ -71˚
100
-72˚ PGA 0 1 -72˚ 80
60
2 3 4 5 6 7 8 9 10 0 2400Gal
m
-73˚ -73˚ 0
-20
-40
-60
-74˚-110˚ -105˚ -100˚ -95˚ -90˚ -85˚-74˚ -80
Fig.2.Free-airgravityanomaliesfromOperationIceBridgeflightsovertheAbbotIceShelfplottedalongflightlines.Gravityanomalies
inoffshoreareasaresatellite-derivedfree-airanomaliesfromMcAdooandLaxon(1997)contouredat10mGalintervals.Colorscalefor
airborneandsatellitegravityanomaliesisshownontheright.NumbersidentifyprofilesshowninFig.3anddiscussedinthetext.PGA,
Peacockgravityanomaly;BGA,Bellingshausengravityanomaly.
(Argyleetal.,2000;Sanderetal.,2004).Amajoradvantage Large negative free-air anomalies of −55 to −60mGal
oftheAIRGravsystemcomparedtootherairbornegravime- are found southeast of Thurston Island over the northern
tersisthatitisabletocollecthigh-qualitydataduringdraped part of the ice shelf from 96.5 to 94.5◦W and extend west-
flights(Studingeretal.,2008)suchastheOIBflightsoverthe ward to 97◦W along the southern edge of a ridge of higher
Abbot Ice Shelf. Flights during OIB surveys are flown at a gravityextendingeastwardfromShermanandCarpenteris-
nominalheightof1500ft(457m)abovetheEarth’ssurface. lands(lines0and5–7).Free-airanomaliesovertheiceshelf
Thefree-airgravityanomalieswerefilteredwitha70sfull- west of Sherman Island are generally in the range of −20
wavelength filter, resulting in a ∼4.9km half-wavelength to−35mGal,althoughreaching−55mGaljusttothenorth-
resolution for a typical flying speed of 140ms−1 (272kn). westofShermanIsland(lines0,1and2).Free-airanomalies
Differencesinthefree-airanomaliesatlinecrossingsduring between Johnson and Farwell Islands in the eastern portion
the Abbot survey vary from 0.09 to 2.61mGal with a stan- ofthesurveyareaaregenerallybetween−15and−35mGal,
darderrorof0.75mGal. reaching−45to−50mGalinplaces(lines0,9and10).
Free-air gravity anomalies measured on OIB flights over
Abbot Ice Shelf are shown in map view in Fig. 2 and as 2.2 Inversionoffree-airgravityanomalies
profiles in Fig. 3. Observed free-air anomalies range from forbathymetry
80to−61mGal.Largepositivefree-airanomaliesarefound
over Thurston Island. The maximum anomaly on line 1, at Inversion of the gravity data for bathymetry was under-
the western tip of the island, is 42mGal (Figs. 2 and 3, taken in two dimensions along individual flight lines using
line1),butanomaliesontheotherlinesovertheislandcon- GeosoftGMSys™ software.Thesoftwaredoesiterativefor-
sistently reach 55 to 72mGal (lines 2–6). A large positive wardmodelingusingthetechniqueofTalwanietal.(1959).
anomaly of 77mGal is also measured over Dustin Island at The bed was kept fixed where it is observed in the radar
the edge of the ice shelf immediately east of Thurston Is- data and the bathymetry in water-covered areas (where the
land(line7).Otherislandsoverflownduringthesurveyalso seafloor cannot be imaged with radar) varied to obtain the
showpositivegravityanomalieswithrecordedanomaliesof bestmatchtotheobservedgravity.Themodelispinnedtothe
17.9mGaloverJohnsonIsland,24.9mGaloverShermanIs- observedgravityvalueatalocationwithintheregionwhere
land and 43.1mGal over Farwell Island (lines 0, 3 and 8). thebedcanbeobserved.
An east–west band of large positive anomalies with maxi- Two major concerns for inversion of gravity for
mumamplitudesof50–80mGalalsoextendsalongtheKing bathymetry are lateral density variations arising from geo-
Peninsula on the mainland (lines 1–4). This positive grav- logic heterogeneity and the possible presence of sediments.
ityanomalyextendsoffshoreintotheiceshelfeastof98◦W Theimportanceoftheseparametersisillustratedbycompar-
withareducedamplitudeanddiesoutnear94◦E(lines5–8). isonofaninversionofOIBaerogravitydataforbathymetry
www.the-cryosphere.net/8/877/2014/ TheCryosphere,8,877–889,2014
880 J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability
W E
0 25 50 75 100 125 150 175 200
Line 0 Gal4800 1 2 3 4 5 4800 Gal
m 0 0 m
-40 -40
1000 1000
Ice (0.915 gm/cm3) eters 5000 5000eters
Water (1.03 gm/cm3) m-500 Sherman Island -500m
-1000 -1000
Rock (2.70 gm/cm3) -1500 0 25 50 75 100 125 150 175 200-1500
Dense Rock (2.85 gm/cm3) W Model Gravity Anomaly Observed Gravity Anomaly E
175 200 225 250 275 300 325 350 375
Gal4800 5 6 7 8 9 10 4800 Gal
m 0 0 m
-40 -40
1000 1000
eters5000 00 22 5000eters
m-500 Johnson Farwell -500m
-1000 Island Island -1000
-1500 -1500
175 200 225 250 275 300 325 350 375
Kilometers
Fig.3a.Airbornegeophysicsline0alongtheaxisoftheAbbotIceShelf(Fig.2)brokenintotwosegmentswith35kmofoverlap(shaded
areainthelowerpanel).Verticalredlinesshowwhereline0intersectsnorth–southlines1–10(Fig.3b).Foreachsegment,theupperpanel
showstheobservedfree-airgravityinredandtheanomalypredictedbytheinversioninblue.Lowerpanelsshowtheobservedupperand
lowericesurfacesfromlaseraltimeterandradarmeasurements,respectively,andtheuppersurfaceofthesolidEarth.Whereiceisgrounded,
theEarth’ssurfaceisdeterminedfromradarmeasurements.Whereiceisfloating,thebathymetryisdeterminedfrominversionofgravity
anomalies.
beneaththeLarsenCIceShelf(CochranandBell,2012)with sions,butwilldiscussthepossibleeffectsofsedimentslater
a later seismic reflection survey that obtained depths at 87 inthissection.
sitesontheiceshelf(Brisbourneetal.,2014).Brisbourneet Weutilizedallavailablegeologicinformationtoconstrain
al.(2014)reportthatthedifferencebetweenseismicallyand crustaldensityvariationsandlocalheterogeneity,notingthat
gravimetricallydetermineddepthsvariedfromlessthan10m rock exposures are limited on the mainland immediately to
toamaximumof320mwitharoot-mean-square(rms)mis- the south of Abbot. The only outcrops reported in this re-
match of 162m. We sampled the Cochran and Bell (2012) gionareatLepleyNunatakandintheJonesMountains.Lep-
gravimetrically determined bathymetry grid at the locations leyNunatak,locatedat73◦070S,90◦190W(Fig.1)immedi-
oftheseismicmeasurementsandfoundthesamedifference ately to the east of our survey area, exposes granite cut by
range (1.6–329m), but a standard error of 59.0m. In fact, mafic dikes (Craddock et al., 1969; Grunow et al., 1991).
84% of the mismatches were less than the rms difference The Jones Mountains, located along the southern margin of
citedbyBrisbourneetal.(2014)(Fig.S1intheSupplement). theiceshelfbetween93◦300and94◦450Watabout73◦300S
Nonetheless, there are several factors that could con- (Fig. 1) reach an elevation of 1500m to the south of our
tribute to errors in results obtained with the formal Parker– line 8. The exposed Jones Mountains rocks consist of 500–
Oldenburginversion(Oldenburg,1974)usedbyCochranand 700mofthinMiocenealkalicbasaltflowsoverlyingabase-
Bell(2012).Thattechniqueassumesgravityanomaliesarise mentofEarlyJurassicarc-relatedgranitescutbymaficand
from relief on a single seawater–rock interface with a uni- felsicdikes,overlainbymiddle–LateCretaceousintermedi-
form density contrast, in turn implying uniform geology, atetofelsiclavas(Craddocketal.,1964;Rutfordetal.,1972;
probably unrealistic over an area as extensive as the Larsen Grunowetal.,1991;Pankhurstetal.,1993).
C Ice Shelf. The Parker–Oldenburg technique also requires Outcrops are more common on Thurston Island, where
auniformgridofgravityanomalies.Inareasofunevenand rocks have been described from about 15 locations (White
somewhatsparsegravitycoveragesuchastheLarsen,inter- and Craddock, 1987; Lopatin and Orlenko, 1972; Storey et
polationbetweenmeasuredgravitylinesisnecessary.Inthe al.,1991;Pankhurstetal.,1993;Grunowetal.,1991).West-
Larseninversion,interpolated,ratherthanmeasured,gravity ern Thurston Island, from about 99 to 102◦W, is underlain
valueswereusedatmanysites. primarily by pink granite, although diorite and granodior-
The very different technique employed in this study in- ite are found west of ∼102◦W at the westernmost end of
volves iterative two-dimensional forward modeling along the island close to our line 1. East of 99◦W, the predom-
flightlines,wherethegravityanomaly,surfaceelevationand inant rocks are gabbro and diorite. Storey et al. (1991) in-
ice thickness are known. This also allows us the incorpora- terpret aeromagnetic lines across eastern Thurston Island
tionoflateralvariationsincrustaldensitybasedonthelocal as suggesting that the southern part of the island (south of
geology. We cannot include a sediment layer in the inver- about72◦100S)“maybeunderlainbyalargegabbrobodyof
TheCryosphere,8,877–889,2014 www.the-cryosphere.net/8/877/2014/
J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability 881
Line 1N S Line 6 N S
Gal4800 -50 -25 0 25 50 4800 Gal Gal4800-50 -25 0 25 50 75 4800 Gal
m0 0 m m0 0 m
-40 -40 -40 -40
meters-5500000 Thurston King -5500000 meters meters-5500000 Thurston Eights -5500000meters
-1000 Island Peninsula -1000 -1000 Island Coast -1000
Lin-e15 020 -50 -25 0 25 50 -1500 Line 7-1500-50 -25 0 25 50 75 -1500
Gal4800 -50 -25 0 25 50 75 8400 Gal Gal8400 -50 -25 0 25 50 4800 Gal
m0 0 m m0 0 m
-40 -40 -40 -40
meters-5500000 Thurston King -5500000 meters meters-5500000 Dustin Eights -5500000 meters
-1000 Island Peninsula -1000 -1000 Island Coast -1000
Line- 13500 -50 -25 0 25 50 75 -1500 Lin-1e50 08 -50 -25 0 25 50 -1500
mGal84000-50 -25 0 25 50 75 84000 mGal mGal48000 -50 -25 0 25 50 48000 mGal
-40 -40 -40 -40
meters-1-0550000000 ThIsularsntdon ShIselramndan PenKiinnsgula --15500000000meters meters-1-5500000000 JoIshlannsodn ECiogahstts -1-5500000000 meters
LinmGale 4-4840-0001500--5500 --2255 00 2255 5500 -844000075mGal-1500 -15L00inemGal 9-4480000 --5500 --2255 00 2255 5500 -150075 -4840000 mGal
meters-5500000 Thurston King -5500000meters meters-5500000 Eights -5500000meters
-1000 Island Peninsula -1000 -1000 Coast -1000
LimGalne--48401 0005500 --5500 --2255 00 2255 5500 -844-01000500mGal Line-1 150mGal00-4480000--5500 --2255 00 2255 5500 -484000075mGal-1500
meters-1-5500000000 ThIsularsntdon ECiogahstts -1-5500000000 meters meters-1-5050000000 ECiogahstts -1-5500000000meters
-1500 -50 -25 0 25 50 -1500 -1500-50 -25 Kilom0eters 25 50 -1500
Kilometers
Fig.3b.AsinFig.3a,butforthenorth–southairbornegeophysicslines,withverticalredlinesindicatingintersectionswithalong-axisline0
(Fig.3a).
highmagnetization”.Amphibolite-gradegneisswasreported Weassumethatthehigh-densitybodiesbeneathThurston
fromtheeasternendofThurstonIslandatMorganInletand Islandarerestrictedtotheislandbasedontwoobservations.
Cape Menzel (White and Craddock, 1987; Pankhurst et al., First, the steep gravity gradient at the southern edge of the
1993) close to our line 6. Several investigators report that islandonlines4–6ismuchsteeperthanobservedonlines2
Dustin Island at 95◦W, just to the east of Thurston Island and 3, where no high-density rocks are observed (Fig. 3b).
and crossed by our line 7, is made up primarily of gabbro, Second, if we continued the high-density region beneath
withdioritealsopresent(WhiteandCraddock,1987;Storey theiceshelf,asignificantdiscrepancyemergesbetweenthe
et al., 1991; Grunow et al., 1991). In summary, granite pre- north–southandeast–westlines.Assumingthedensebodyis
dominatesfrom99to102◦Wwithdenserrocks(gabbroand present beneath the ice shelf, the inversions imply ∼500m
diorite) present at the western and eastern ends of Thurston ofwateronline4betweenShermanIslandandCarpenterIs-
IslandaswellasonDustinIsland. landwhiletheinversionsoforthogonalline0imply<300m
Webeganbyassumingabasemodelwithfourbodies:air ofwateratthatlocation(Fig.3).Theorthogonallineiswell
(0gcm−3), ice (0.915gcm−3), seawater (1.03gcm−3) and controlled as it matches the gravity over the three islands
granitic rock (2.70gcm−3). On lines 1, 4, 5 and 6 across (Sherman, Johnson and Farwell) where the bed can be ob-
ThurstonIslandandline7acrossDustinIsland,weincluded servedonradaralongline0(Fig.3a).
adensercrustalbody(2.85gcm−3)wheregabbroanddiorite Modeling bathymetry under the ice shelf while satisfy-
outcrop.Theuppersurfaceofthisdensebodyisattheradar- ing the observed gravity anomalies where the bed is ob-
determined rock–ice interface and the lower surface deter- served both on Thurston Island and the mainland requires
minedbytheneedforthefinalmodeltomatchtheobserved a high-density body within the crust corresponding to the
gravitywherethebedisobservedonradarbothontheislands areaoflargepositivegravityanomaliesovertheKingPenin-
andonthemainland. sula(Figs.3and4).Wherehigh-densitybodiesareinferred
www.the-cryosphere.net/8/877/2014/ TheCryosphere,8,877–889,2014
882 J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability
beneath both coasts, we constrained their extent and depth (Prestvik etal., 1990; Prestvikand Duncan, 1991)has been
by satisfying the observed anomalies onshore on both sides recoveredfromPeterIIsland(∼68◦500S,90◦300W).
of the ice shelf where the bed is constrained by radar while In all of these denser regions, there is a trade-off be-
alsomatchingtheseafloordepthbeneaththeiceshelfonor- tweentheassumeddensityandthethicknessdeterminedfor
thogonalalong-axisline0atlinecrossings. the body. If the actual density is greater than the assumed
The zone of high-density crust beneath the King Penin- 2.85gcm−3, the dense body will be thinner, but with min-
sulaextendsoffshoreundertheiceshelfeastof98◦E.Inthis imal changes in the inferred bathymetry. We repeated the
area,theamplitudeofthefree-airanomaliesovertheiceshelf inversion of line 6 giving the southern dense body a den-
equalsorexceedsthatoftheanomaliesonshorewheretheice sity of 3.0gcm−3 rather than 2.85gcm−3 while maintain-
surfaceisatanaltitudeof500–800m.Ifnodensebodyisin- ingthesamecriterionthatthethicknessofthedensebodyis
cludedinthisoffshoreregion,theinversionpinstheseafloor adjusted so that the seafloor just touched or approached the
to the base of the ice and is still unable to match the ob- baseoftheiceatthelocationoficerisesorabruptchanges
served anomalies. We used the observation that changes in inicethickness.Thisresultedinasignificantlythinnerhigh-
ice thickness (Fig. 3, lines 5–8) and linear sets of ice rises densitybody,butdifferencesintheinferredbathymetrywere
andicerumples(Swithinbanketal.,2004)(Fig.1)underlie lessthan50macrosstheentireregion.
peaksinthegravityanomaliestoconstrainthedensebodies. Wetestedthesensitivityofresultstothechoiceofcrustal
Wesetthebaseofthedensebodyoneachprofilesothatthe density by repeating the inversion of east–west line 0 as-
seafloorjusttouchedorapproachedthebaseoftheiceatthe suming crustal densities that varied by 0.1gcm−3 around
locationsoftheicerisesandstepsinicethickness. 2.7gcm−3 while holding the bed fixed where it is visible
TheregionofhighdensitiesalongtheKingPeninsulade- on radar records on Sherman, Johnson and Farwell islands
terminedonindividualprofilesformsacoherentzonedelin- and pinning the model to observed gravity over Sherman
eatinganeast–west-trendinghigh-densitybody(Fig.4).The Island. The differences in bathymetry varied directly with
thicknessofthishigh-densitybodyinourmodelsalsovaries water depth and reached a maximum of about 60m in the
smoothly from profile to profile, increasing from 4.5km on deepbasinwestofJohnsonIsland(∼225–255kmonline0,
line1to10–14kmonlines2–6,thendecreasingrapidlyeast- Fig. 3). A higher density results in shallower depths and a
wardto3kmonline7and2.4kmonline8.Thehigh-density lowerdensityindeeperdepths.
body modeled along the southern margin of the ice shelf is Wehavenoinformationonthepresenceordistributionof
nearly in line with a WNW–ESE-trending gravity anomaly sediments and did not include a sediment layer in the mod-
located between 105 and 110◦W on the continental shelf eling. If the structure of the seafloor beneath the ice shelf
westoftheAbbot(McAdooandLaxon,1997)(Fig.2).This developed from continental rifting in the Late Cretaceous,
“Peacockgravityanomaly”(Larteretal.,2002)wasmodeled sedimentalmostcertainlyaccumulatedinthebasinspriorto
by Gohl et al. (2007) on the basis of shipboard gravity and glaciation. This sediment may still exist or may have been
helicoptermagneticsdata,asarisingfromalargemagmatic scoured away. Bedrock is exposed on the inner portions of
intrusion. most Antarctic continental margins, including the Amund-
We also needed to introduce a higher-density region cor- senSeaEmbayment(Wellneretal.,2001;LoweandAnder-
respondingtohighergravityanomaliesattheseawardendof son, 2002). If sediment is present, it will cause the gravity-
line10(Fig.3b),withoutwhichtheinferredseafloorreaches derived bathymetry to be deeper than the actual seafloor. A
the sea surface. The depth to which this high-density body simple Bouguer slab calculation implies that each 100m of
extended was set so that the average water depth along that sedimentwithdensity2.2gcm−3willleadtoaseafloordepth
portion of the profile agrees with the average depth in that overestimateof∼30mandthedepthsobtainedfromthein-
area from the IBCSO (International Bathymetric Chart of version will lie between the actual seafloor and crystalline
the Southern Ocean) grid (Arndt et al., 2013). Large pos- basementdepths.Themostprobablelocationofathicklayer
itive gravity anomalies in this area appear to be related to ofsedimentsisinthedeepbasinsoutheastofThurstonIsland
an intense high (133mGal) in the satellite altimetry gravity (Fig.3,profiles5,6and7;Fig.4),wherethegravityinver-
field (McAdoo and Laxon, 1997) centered at 72◦26.250S, sion gives depths ranging from 900 to >1200m. If 1000m
89◦52.50W (Fig. 2), which may result from concentrated ofsedimentwerepresent,thosedepthswouldbe300mshal-
magmaticactivity.Anicetongueextendsoutfromtheeast- lower.Atopthehigherstandingareas,lesssedimentislikely
ernmostAbbottothecenterofthegravityanomaly,suggest- to have accumulated and is more likely to have been re-
ing the presence of a pinning bathymetric high at that loca- moved.Onseverallines,thegravity-determinedseafloorap-
tion(Fig.2).Thisgravityanomalyliesdirectlyonlinewith proaches or just touches the base of the ice where ice rises
Peter I Island and the De Gerlache Seamounts offshore to and rumples coincide with gravity highs along the southern
the north. Early Miocene (20–23Ma) basalts (Hagen et al., iceshelf(Fig.2–4).
1998) have been dredged from the De Gerlache Seamounts
(∼65◦S, 91◦W), and Late Pleistocene (∼0.3Ma) basalt
TheCryosphere,8,877–889,2014 www.the-cryosphere.net/8/877/2014/
J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability 883
--110055˚˚ --110000˚˚ --9955˚˚
--111100˚˚ --9900˚˚
Basin
--7700˚˚ Fault --7700˚˚
Limit of
rifting
Dense
crustal
body
--7711˚˚ --7711˚˚
--7722˚˚ --7722˚˚
0
200
400
600
800
1000
1200
1400
--7733˚˚ --7733˚˚ 221207505000 meters
2500
2750
3000
3250
3500
--7744˚˚ --110055˚˚ --110000˚˚ --9955˚˚ --7744˚˚ 43070500
--111100˚˚ --9900˚˚ 4250
4500
Fig.4.Basins,faultsandthespatialextentofriftingunderAbbotandCosgroveiceshelves,frominversionofOperationIceBridgegravity
data along flight lines. Shaded areas show the extent of the high-density crustal bodies determined from the inversion. Gravity-derived
bathymetryandradar-determinedbedareshownalongflightlines.OffshoremarinebathymetryisfromtheInternationalBathymetricChart
oftheSouthernOceans(IBCSO)(Arndtetal.,2013)contouredat100mintervalsto1500and250mintervalsforgreaterdepths.Blacklines
◦ 0 ◦ 0
alongthewesternicefrontbetween72 30.6 and72 42.4 SandacrosstheoutershelftothenorthofAbbotshowthelocationofshipboard
bathymetrylinesdiscussedinthetext.StarsshowlocationofCTDstationsnbp9402-0091(red)andnbp0901-008(blue)(GiuliviandJacobs,
1997;Jacobsetal.,2011)showninFig.6.
3 Discussion lines2and3,thestructuretakestheformofaseriesofsouth-
facing half grabens forming 10–20km wide topographic
3.1 TectonicsettingoftheAbbotIceShelf basins (Figs. 3b and 4). This pattern continues south across
theKingPeninsulatotheCosgroveIceShelf,whichalsoap-
pearstobeunderlainbyahalfgraben(Fig.5).
Our inversion for the bathymetry beneath the Abbot Ice
The tectonic pattern changes east of Carpenter Island,
Shelfrevealsaseriesofeast–west-trendingbasinsofvarying
where a set of large inward-facing faults delineate a deep
depthsandwidthsunderthewesternportionoftheiceshelf
(Figs.3,4and5).Thebasinswestof∼97◦Wareasymmet- basinthatreachesdepthsofover1200monline5(Figs.3b
and 4). Shallower small, asymmetric, fault-bounded basins
ric with a steep slope, which we take to be faulted, bound-
arelocatedalongthesouthernmarginoftheiceshelf.Foot-
ingoneside.Thesebasinsformfault-boundedhalfgrabens.
wall rims of the faults bounding the northern side of these
Thisinterpretationissupportedbytheobservationthatradar
basinsformsillsassociatedwithchangesinicethicknessand
images a deep, asymmetric basin under the King Peninsula
createlinearrowsoficerisesandicerumplesontheiceshelf
that appears fault-bounded on the north and takes the form
surface (Figs. 1 and 3b line 5 at 28.5km, line 6 at 18 and
ofahalfgraben(line3,Figs.3and5).Thenorthernwallof
37km, line 7 at 25km, line 8 at 25 and 38km). Rifting in
that basin is steeper than the walls of the gravity-delineated
this area does not extend south of the Abbot Ice Shelf as it
basins,butthegravity-determinedslopesarelowerduetothe
doesfartherwest(Fig.5).
70s gravity data filter. If we filter the radar-determined bed
We interpret the bathymetry revealed beneath the Abbot
with the spatial equivalent of the gravity filter, the basin on
IceShelfbyinversionoftheOIBgravitydataasacontinental
the King Peninsula has slopes very similar to those in the
rift related to the rifting between Zealandia and Antarctica.
basinsidentifiedfromthegravityinversion.
Thisissupportedbytheriftbasinsettingandtheobservation
The distribution of interpreted faults and basins is shown
thatthelastknownigneouseventonThurstonIslandwasthe
inFig.4.TheentiresouthernmarginofThurstonIslandap-
emplacementofasuiteofeast–west-trending,coast-parallel
pears to be comprised of overlapping, closely spaced sub-
dikes(e.g.,Storeyetal.,1991;Leatetal.,1993).Thesedikes
parallel faults. In the western portion of the ice shelf, on
www.the-cryosphere.net/8/877/2014/ TheCryosphere,8,877–889,2014
884 J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability
N Line 3 S Temperature (°C) Abbot Area (%)
Gal4800-50 -25 0 25 50 75 100 125 4800 Gal 0−2.0 −1.0 0.0 1.0 2.0 3.0 4.0 0 10 20 30 0
m100-4000 Abbot Ice Shelf Cosgrove Ice Shelf -1400000m A B
meters---115550000000000 ThIsularsntdon ShIselramndan PenKiinnsgula CPeanniinstseuola ---115550000000000meters 210000 T - Tf 210000
-50 -25 0 25 50 75 100 125
s) s)
Line 6 er er
mGal-4840000-50 -25 0 25 50 75 100 125 -4840000 mGal pth (met 430000 T 430000 pth (met
1000 Abbot Ice Shelf 1000 e e
meters-5500000 -5500000meters D 500 500 D
Thurston Eights
-1000 Island Coast -1000
-1500 -1500
-50 -25 0 25 50 75 100 125 600 NBP9402-091 600
Kilometers
NBP0901-008
Fig. 5. Entire length of airborne geophysics lines 3 and 6, as in
700 700
Fig. 3, with line locations shown in Fig. 2. Red arrows show the
Fig. 6. (a) Measured temperature (T) and calculated temperature
locationoffaultsmappedinFig.4.Notethatactualfaultswillbe
steeperthanshownherebecauseofthefilterappliedtothegravity abovetheinsitufreezingpoint(T −Tf)on13March1994(CTD
091)and14January2009(CTD008)(GiuliviandJacobs,1997;Ja-
data.
cobsetal.,2011)atthelocationsshowninFig.4.(b)Depthdistri-
butionofAbbotIceShelfdraftscalculatedfromthe1kmBedmap2
were intruded prior to about 90Ma (Leat et al., 1993) into compilation of surface elevation and ice thickness (Fretwell et
al., 2013). Solid horizontal lines show the mean draft determined
120–155Maplutonicrocks(Grunowetal.,1991)andimply
fromtheBedmap2compilation;dashedlinein(a)showsthemean
regionalextensionduringthattimeinterval.
Thisriftedterrainendsabruptlynear94◦WatJohnsonIs- draftdeterminedfromOIBsurfaceelevationandicethicknessdata
(Fig.7).
land.Thedensecrustalbodyunderthesoutherniceshelfand
theKingPeninsulaalsoterminatesatJohnsonIsland(Fig.4).
Thebathymetryinferredbeneaththeiceshelfonlines9and
10(Fig.3)locatedeastof94◦WbetweenJohnsonandFar- 3.2 Ocean–iceinteractionsbeneaththeAbbotIceShelf
wellIslandsisshallower,withgentlerslopesthaninthewest,
The IBCSO bathymetry grid shows a trough on the conti-
andshowsnofeaturessuggestingtectonicactivity.
The tectonic boundary at 94◦W is directly on line with nentalshelfthatextendsnorthwardalongthewesternendof
ThurstonIslandtotheshelfbreak(Fig.4).Withaminimum
alinear,north–south-orientedbathymetricandgravitystruc-
depthof575mat∼72◦050S,thistroughissufficientlydeep
ture, the Bellingshausen gravity anomaly (BGA) of Gohl et
toallowthesouthwardpassageofwarmdeepwater(Hellmer
al.(1997),inthedeepoceannorthofthecontinentalmargin
etal.,1998;Jacobsetal.,2011).Oceantemperatureprofiles
(Figs. 2 and 4). The BGA structure developed along a Cre-
attwoCTDstationsnearthewesternicefrontoftheAbbot,
taceous ridge–trench transform separating the Phoenix and
marked by stars in Fig. 4, encountered CDW at depths be-
Pacific plates (Larter et al., 2002). Rifting nucleated along
lowabout400m(Fig.6).Temperaturesare>3◦Cabovethe
thistransform,probablyatanintra-transformspreadingcen-
melting point of ice at those depths, and the overlying ther-
ter (Bird and Naar, 1994), at about 90Ma and propagated
moclineis>1◦abovefreezingatdepthsupto∼200–250m.
westward,splittingtheChathamRiseandCampbellPlateau
While the contoured IBCSO bathymetric grid suggests
fromAntarctica(Larteretal.,2002;Eaglesetal.,2004).The
tectonicboundaryat94◦WbeneaththeAbbotIceShelfim- shoalingbathymetryeastwardfromstation91towardtheice
front, bathymetric data were lacking in that region (Arndt
plies that this rifting originally extended into the Antarctic
et al., 2013). A post-IBCSO 2012 RRS Shackleton single-
continentbeforefocusingnorthofThurstonIsland.
The Abbot to the west of 94◦W occupies a different ge- beam echo sounder line (location shown in Fig. 4) shows
about 800m of water near the ice front at the base of the
ologic and tectonic setting than most ice shelves, which are
faultjustsouthofstation91(F.Nitsche,personalcommuni-
formed when the ice streams advance across the continen-
cation,2013).Theshoalingisthusagriddingartifactinthe
talshelf,thinandbegintofloat.TheAbbotinsteadoverlies
IBCSOgridandismaskedinFig.4.
a preexisting basin with well-defined and stable grounding
The maximum water depth is 575m beneath our west-
lines defined by the tectonic structure along both the north-
ernmost line 1 across the Abbot. This bathymetric low,
ernandsouthernboundariesofmostoftheriftbasin.
constrained by inversions of orthogonal profiles, is located
<10kmNEoftheShackletonlineand∼20kmNEofCTD
station91,whichrecordedtemperaturesto655m.Bothther-
moclineandwarmdeepwatersarethereforeatdepthswhere
TheCryosphere,8,877–889,2014 www.the-cryosphere.net/8/877/2014/
J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability 885
The Abbot Ice Shelf draft, calculated from the Bedmap2
OIB Elevation BM2 sampled Elevation
compilation (Fretwell et al., 2013), averages 230.2mb.s.l.,
70 A B 70 with the base of the ice shallower than 300mb.s.l. over
84% of its area (Fig. 6). The mean Abbot Ice Shelf draft
60 60
eters)50 50 eters) fisro1m89a.7lomng,-∼tra4c0kmOIleBssaltthiamnettehreavnadlueradoabrtaimneedasufrroemmetnhtes
m m
n (40 40 n ( Bedmap2compilationsampledatthesamelocations(Figs.6
o o
ati30 30 ati and 7). However, the OIB and Bedmap2 surface elevation
v v
Ele20 20 Ele distributionsaresimilar,withinsignificantlydifferentmeans
of39.2and40.2m(Fig.7),indicativeofanomaliesinoneor
10 10
bothdraftcalculations.
0 0 No firn corrections are applied to OIB MCoRDS radar
0 10 20 30 0 10 20 30
measurements(Leuschen,2011),whichcouldleadtothick-
OIB Draft BM2 sampled Draft
nessunderestimatesof∼10m(Fretwelletal.,2013).Radar
0 0
C D results can also underestimate total thickness where bottom
100 100 returns come from meteoric-ice–marine-ice interfaces. Ma-
s) s) rineiceisprobablynotwidespreadatthebaseoftheAbbot,
er er
et 200 200 et butmightbeexpectedwhereicedraftsarelessthanadjacent
m m
h ( h ( near-freezingwintermixedlayerdepths(Fig.6).
pt 300 300 pt The Bedmap2 compilation utilized the Griggs and Bam-
e e
D D
ber (2011) determination of ice shelf thickness, based on a
400 400 firndensificationmodeland the thicknessnecessarytosup-
port the observed elevations assuming hydrostatic equilib-
5000 10 20 30 0 10 20 30500 rium.Bedmap2appliedcorrectionsof−21to−81m,based
Abbot Area (%) Abbot Area (%) onfirn-correctedradardata,tomostoftheGriggsandBam-
Fig.7.DistributionofAbbotIceShelfsurfaceelevationsanddrafts ber (2011) ice shelf thickness values in the Amundsen sec-
belowtheseasurface.(A)SurfaceiceelevationfromOIBaltimeter tor (Fretwell et al., 2013). Since appropriate data were not
dataat0.2kmintervalsalongflightlines.(B)Surfaceiceelevation available,nocorrectionswereappliedtotheAbbot(Fretwell
fromBedmap2surfaceelevationgridsampledatthesamelocations et al., 2013). Applying a 10m firn correction to the OIB
astheOIBelevationdata.(C)DraftfromOIBaltimeterandradar radar data suggests that the Bedmap2 values need to be re-
dataat0.2kmintervalsalongflightlines.(D)Draftfromthegrid
duced by 30m to bring the two into agreement. Figure 8
ofBedmap2draftssampledatthesamelocationsastheOIBdraft
showstherevisedBedmap2icedraftdistribution.Themean
data. Horizontal lines show the mean values for each set of mea-
draft is ∼200m, with 78% of the ice shelf less than 250m
surements.
(Fig.8).Theaveragedraftcorrespondstotheupperportion
of the summer thermocline and base of the surface mixed
they have access to and from the ice shelf cavity. The shal- layer(Fig.6),similartoconditionsreportedforWilkinsIce
lowest passage between the western and eastern sectors of Shelf in the Bellingshausen Sea (Padman et al., 2012). The
the cavity is about 465m (Fig. 3b, line 4), well within the nearsurface“winterwater”layerwillcoolandthickensea-
CDWdepthrangeinFig.6. sonally, raising and lowering the upper thermocline, in turn
ThedeepeasternriftbasinmayalsoopenontotheBelling- influencing basal melt rates, as in the southern Amundsen
shausenSeacontinentalshelfbetweenDustinIslandandMc- Sea(Jacobsetal.,2013;Dutrieuxetal.,2014).
NamaraIslandthroughatroughreachingacalculateddepth Regionsofthickericeareconfinedtothesouthernmargin
of785mnear−22kmonline8(Fig.3b).Thermoclinewa- of the ice shelf near the grounding lines and to the eastern
ters and CDW may also flow through this cavity opening. areaaroundFarwellIsland(Fig.8).Thisdeepericeextends
IBCSObathymetrytothenorth,verypoorlyconstrainedby morethanafewkilometersfromgroundinglinesonlyatlo-
sparseshipboardmeasurements(Grahametal.,2011;Arndt cations where fault footwall rims marked by linear sets of
etal.,2013),doesnotdefineanybathymetrictroughsextend- ice rises limit access by CDW and lower thermocline water
ing toward the edge of the continental shelf. A 1999 RV/IB (e.g., Fig. 3 line 6). Most of the ice shelf draft thus appears
Nathaniel B. Palmer bathymetric line across the outer shelf to be controlled by the thickness of the cold surface water
north of Abbot (location shown in Fig. 4) also does not re- overlayingthethermocline(Figs.6and8).
veal any distinct troughs, but its continental shelf depths of The thick continental ice that flows into the Abbot is
397–558marecomparabletotheCDWdepthsinFig.6.The largely consumed by melting within a few kilometers of
depthofoutershelfsillsistheprimarycontrolonCDWac- glacier grounding zones (Rignot et al., 2013). That implies
cesstoiceshelfcavities(e.g.,Fig.4inJacobsetal.,2013). higherlocalmeltratesthanthoseauthors’area-averageequi-
librium rate of 1.9myr−1. Faster melting of thicker ice
www.the-cryosphere.net/8/877/2014/ TheCryosphere,8,877–889,2014
886 J.R.Cochranetal.:BathymetricandoceaniccontrolsonAbbotIceShelfthicknessandstability
corrected Bedmap2 ice draft eastern Abbot rift. The presence of 1000m of sediment in
--110000˚˚ --9955˚˚ thesebasinswouldmakeactualdepths300mshallower.The
--9900˚˚
--7722˚˚ --7722˚˚ riftstructuresalongwiththelikelythinsedimentinshallower
areaspresentpathwaysforseawatertocirculateundermuch
150
700700
  oftheiceshelf.
600600

200 250 500500
  OIBradardatashowmeanicedraftof∼200m,correct-
--7733˚˚ --7733˚˚ 400400
  ing the Bedmap2 compilation by −30m, with 78% of ice
112230505000000322110505000000

  sinheslofmhaevwinagysatdoratfhteoGfleetszsItcheanSh2e5l0f,mw.iTthhoeuAt cbobnonteicstsioimnsiltaor
00

largeicedrainagebasinsandfastorbroadicestreams,butit
--7744˚˚ --110000˚˚ --9955˚˚ --9900˚˚--7744˚˚ isthinnerandoverliestectonicallyformedbasins.Thebedin

surrounding areas on Thurston Island and the Eights Coast
Fig. 8. Abbot Ice Shelf draft calculated from the 1km Bedmap2
isnearorabovesealevel,andotherreportsplacetheAbbot
compilationofsurfaceelevationandicethickness(Fretwelletal.,
close to a state of mass balance between inflow, accumula-
2013)correctedasdiscussedinthetextandcontouredat100min-
tion,minorcalvingandbasalmelting.
tervalswith150and250mcontoursadded.GreylinesshowOIB
flightlines.Colorbargivesicedraftinmeters. Sparse oceanographic measurements show a thermocline
overlying “warm” CDW near the western Abbot ice front,
with temperatures from 1 to >3◦C above the in situ melt-
would be consistent with its exposure to CDW and lower ingpointupto200–250m.TheCDWandlowerthermocline
thermoclinewatersaswellasthemixingeffectsoftidalcur- waters will rapidly melt thicker inflowing ice. Most of the
rentsinareasoflowwatercavitythicknessnearsillsformed ice shelf draft distribution occupies the same depth range
by the footwall rims (Mueller et al., 2012). Once the shelf as the thermocline, making it sensitive to changes in thick-
icehasthinnedtothedepthrangeofupperthermoclineand ness of the surface and deep waters. With an average near-
lower surface waters, its melting or freezing will depend equilibrium thickness coinciding with the transition region
on temporal variability of the upper ocean thermal struc- between the upper thermocline and cold surface water, the
tureandthedensity-drivencavitycirculation.Upwellingand ice will respond to seasonal and longer-term forcings that
downwelling responses to “local” winds, as observed near changesurfacewatercharacteristicsandCDWvolumeonthe
theWilkinsIceShelf(L.Padman,personalcommunication, continentalshelf.
2014), may be less common near the Abbot, which is typi-
callysurroundedbynearlyfastseaiceinwinter.
Withanoverallaveragedraftof∼200m,muchoftheAb- Supplementarymaterialrelatedtothisarticleis
availableonlineathttp://www.the-cryosphere.net/8/877/
botwillbesensitivetointerannualvariabilityindepthofthe
2014/tc-8-877-2014-supplement.pdf.
surfacemixedlayer,whichcanbehighintheAmundsenSea
(Jacobsetal.,2013).Surfacewaterpropertiesdependmainly
onatmosphericforcing,asdoestheiceshelfsurface,where
seasonalmeltingcancauseinstabilitybyenhancingcrevasse
propagation(Scambosetal.,2000;Banwelletal.,2013).The
Acknowledgements. WethankM.Studinger,N.Frearson,S.Elieff
currentsurfacemeltintensityontheAbbotislessthanatice
and S. O’Rourke with OIB and the RV/IB Nathaniel B. Palmer
shelves on the Antarctic Peninsula (Trusel et al., 2012), but shipboard scientific party for their efforts during the 2009 field
itsrelativelythinstate,lowlatitudeforWestAntarcticaand season. L. Padman, T. Scambos and A. Brisbourne undertook
exposure to the shifting Amundsen Sea low (Turner et al., thoughtfulreviewsthatgreatlyimprovedthepaper.Wealsothank
2013) make it more vulnerable than thicker shelf ice to in- F.NitscheforprovidingtheRRSShackletonbathymetrylinealong
creasedsummerairtemperatures. the Abbot ice front and Alex Brisbourne for making the results
of the BAS Larsen C seismic survey available to us. S. Starke
provided technical assistance. This work was support by NASA
grants NNX09AR49G, NNX10AT69G and NNX13AD25A, NSF
4 Conclusions
grant ANT-06-32282 and Lamont-Doherty Earth Observatory of
ColumbiaUniversity.LDEOcontributionno.7790.
Inversion of NASA Operation IceBridge gravity data show
the Abbot Ice Shelf west of 94◦W to be underlain by a
Editedby:D.M.Holland
series of rift basins related to rifting between Antarctica
and Chatham Rise in the Late Cretaceous. West of 99◦W,
the region of rifting extends farther south and includes the
Cosgrove Ice Shelf; east of 94◦W, the bathymetry is shal-
lower and unmodified by tectonic activity. Gravimetrically
determined depths reach 900–1200m in deep basins of the
TheCryosphere,8,877–889,2014 www.the-cryosphere.net/8/877/2014/