Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | usivities ff Discussions , and 1 ects interact with Ocean Science ff , around three decades ects of the frozen upper 3 ff − s 2 m 5 − , M. J. M. Williams 1,2 . During the maximum flow period turbulent 1 1440 1439 − s 3 − 10 × , N. J. Robinson 1 , C. L. Stewart 1 3 ects are large, there is little dynamic influence due to temperature, but ff erences in drivers and manifestations that at lower latitudes. For example, ff ect of being close to the freezing temperature has a controlling influence on ff (or ice) tongues, formed by glacier outflows into the coastal ocean, can ex- This discussion paper is/has beenPlease under refer review to for the the journal corresponding final Ocean paper Science (OS). in OS if available. National Institute for Water and Atmospheric Research (NIWA), Greta PointDepartment Wellington, of Marine Science, UniversityIndustrial of Research Otago, Ltd. Dunedin, New (IRL), Zealand Gracefield Lower Hutt, New Zealand of meters thick inculation places. and mixing Such (Jacobs cryogenic et structures al., significantly 1981; influence Legresy local et cir- al., 2004). In the case of flow cryogenic topography that is continuallysome changing. di Oceanic mixing inrotational polar waters e has the e behaviour and there aresurface the (McPhee, additional 2008). frictional boundary e tend many tens kilometers from shore (Frezzotti et al., 1997) and be many hundreds 1 Introduction Quantifying ice–ocean interaction, especially at thechallenge small-scale, is in proving to high-latitude be a earthrelatively major long system timescales sciences and complex (e.g., thermohaline Sirevaag and pressure et e al., 2010) where reached around three times thattical of the velocity ambient shear tidal as flow largeenergy amplitude as and dissipation 3 generated ver- rates reachedgreater than a local maximum background of levels.Richardson 10 This Number is which in keeping droppedare with higher to estimates that around expected of from the unity.cryotopography. parameterization, gradient possibly Associated reflecting the vertical proximity di of the A glacier tongue floating incal the flow coastal and ocean presents influencesmicrostructure a oceanic significant observations mixing obstacle at to andpulses the a transport lo- and processes. glacier vortices tongue as Here side-wall ocean well show shear as tidally-induced concomitant flow mixing. Flow speeds within the pulses Abstract Flow and mixing around a glacierC. tongue L. Stevens Ocean Sci. Discuss., 7, 1439–1467,www.ocean-sci-discuss.net/7/1439/2010/ 2010 doi:10.5194/osd-7-1439-2010 © Author(s) 2010. CC Attribution 3.0 License. T. G. Haskell 1 New Zealand 2 3 Received: 2 August 2010 – Accepted: 3Correspondence August to: 2010 C. – L. Published: Stevens ([email protected]) 11 AugustPublished 2010 by Copernicus Publications on behalf of the European Geosciences Union. 5 20 25 15 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ries ff ect of ff (Stevens et al., 2006). 1 − ect of the ocean on the glacier ff 1442 1441 body. The resulting scales of flow are not normally associated with ff ects climate processes over large space and time scales (Hellmer, ff usive-convective layering. It has also been suggested that such glacier/ice ect of the glacier tongue on local oceanography, the present data also have ff ff Available data on near-surface residual circulation are sparse but suggest a predom- The EGT thus provides an ideal site to examine the degree of variability of, and in- The Tongue (EGT, Fig. 1) in southern McMurdo Sound has been These ice–ocean interaction processes are important in a situation like southern through a hole in the 2.3 m thick first year fast-ice for a four day period (Fig. 1c) around (Robinson and Haskell, 1990). inant northward current near(Leonard the et south al., of 2006; Erebusto Robinson, Bay the 2010; (Fig. SW Mahoney 1c), of et at the al., EGT least 2010). reached during amplitudes Tidal winter of2.2 flows around some 0.15 m 5 s km Instrumentation A 300 kHz acoustic Doppler current profiler (ADCP-RDI Workhorse) was deployed (Debenham, 1965; Holdsworth, 1982;are DeLisle around et 400 al., m 1989).region (Jacobs including et Water likely depths al., moraine near deposits 1981)calving (Fig. its but 1d). of tip highly At the the variable EGT timeevents due was of to writing were in the the March known most of geology recent to 1990 of have when the occurred a 3.5 in km section 1911 broke and away. at Similar some point during the 1940s 2 Methods 2.1 Location In November 2009sidewall we of conducted the oceanographic EGTfrom measurements which the within divides Dellbridge 30 the m Islandsglacier surface tongue of and was waters the Cape 12 of km Evanswith long, the a 2.5 (Fig. km majority basal 1c). wide of slope and of Erebus 300 At m around Bay thick 0.02 the at along time the its grounding of main line, sampling axis, and and the the tip is around 50 m thick of mixing rates atumn the structure tip and of a kinematics,supercooling glacier (ii) and tongue, (iv) mixing, generalization paying (iii) of special the mechanisms attention results for to beyond enhancement the (i) EGT. of water local col- The objective here is to present new microstructure observations providing evidence relevance for flow and melting processes at the face of ice shelves (Rignot et al., 2010). In the absence offrazil/platelets nucleation (e.g., centers, Dmitrenko this etthe supercooled al., region water 2010). (Gough remains et Such liquid al.,observations properties or 2010; in are Robinson, forms 1911 readily 2010) (Deacon, 1975). including observed during in the very firstteraction of between, such small-scale mixing anddeveloped advective relating processes. to Much the worktongue has cryomechanics (see been Squire and et thefor the al., e e 1994 and papers therein). As well as providing evidence observed to influence localtion vertical of stratification di (Jacobs ettongues generate al., local 1981) sources through ofand forma- supercooled Weeks, 1992) water – (e.g., water Debenham,advected cooled 1965; to at Je depth shallower to depths the in where situ it freezing temperature is but then colder than the new insitu freezing point. McMurdo Sound where Haskell Straitern forms Ross an Sea oceanic connection and(Fig. between the 1). the cavity west- The beneath Sound thewaters acts (Robinson combined as et Ross a al., and 2010). conduitmixing McMurdo for The processes both fate Ice in of ice Shelves the these shelf region. waterswhich waters is This and dependent exchange in on warmer influences turn transport sea Ross and ice2004; Sea a growth Dinniman in et the al., region, 2007; Robinson et al., 2010). flow around a blu oceanographic flows with the closestheadlands relevant and work being islands that (e.g.,tongue concerning flow Edwards flow can around et pass under al., the 2004), obstacle. except of course with a glacier around a glacier tongue, the mixing processes are then combined with the e 5 5 20 15 10 25 15 20 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 1 ε − ects. ff equivalent 3 − s 2 40 m at which time > (units of m ε ect and the comparison with the ff in the Sound-proper (Leonard et al., 2006; 1 − 1444 1443 ) were resolved from the dual shear probe profiler ε . In addition, a SBE (Seabird Electronics, USA) temperature 3 ect the shear data through the creation of spurious vibrations, − ff s 2 m 10 − (Roget et al., 2006). T 10 L × ). Thirty eight profiles were recorded with a total profiled distance of over 1 − set of 146 degrees was included. Compass testing did not indicate any inconsisten- A diurnal cycle of concurrent turbulence and ADCP data (day of year 323.6–324.6; The microstructure data were separated into 5 m bins that were overlapped by 50%. Turbulence properties were resolved with a Rockland VMP500 microstructure loose- ff 180 degrees from this,north-NW. heading Flows W other whilst than the during flow the directly pulse were after relatively the weak ringing but was still to a the moderate The pulse was locatedat on least, the this risingshallower was phase flows seen of were as the towardscal. flow tide the The and, acceleration EGT pulse at initially then suggestingmain the at deepened a body observation to depths of wake fill point the recirculation the pulse(using entire in an lasted measured acoustic around the analogy) depth 4 verti- h. whereby beforeperiods there decaying. It spaced were was a The moderate followed little flow by overringing, accelerations some an for moved wave-like hour short “ringing” mainly apart. eastward. Directionally the The pulse, flow including preceding its the trailing pulse was directed around tidal period and so wewater refer column. to it here as a pulse, in ani.e. otherwise moderately 20–21 quiescent November 2009)urnal is the period focus during here which (Fig. 3) the – strongest coincidentally measured this pulse-like was the flow di- occurred (Fig. 2). The largely diurnal tidesand in Pyne, the 2003) region with arecally a around generate 1 14 maximum m flows day in of “spring-neap”Stevens peak-peak 15 et cm cycle elevation s al., (Fig. (Goring 2006; 2a). Robinson,(Fig. 2010), 2c). While here these we The tides sawfaster flow currents typi- flows magnitudes in appear reach each cyclical 40 cycledataset cm (Fig. s were precluded not 2c) any phase-locked but co-spectral to analysis.was careful the the examination tide. The most reveals weak The consistent South-east the short feature.able flow duration The especially at of faster in the flow the its on low direction. the tide rising/high The tide strongest was flow more was vari- quite short in duration in each lengthscale 3 Observations these displacements resolves overturning scales described by the Thorpe overturn a statically stable profile generates a displacement lengthscale. Cumulatively summing pumped so as to notso a the spatial resolution isso around that 1 m. it TheThe was profiler was turbulence at kept data the in do the ambientconstraints not water and temperature continuously have so a removing there continuous thermalgap, were fortunately temporal a inertia during coverage number a start-up due of period to e of gaps operational slower in flow. the profiling,Dissipation most rates, velocities notably and a scalarscale 2.5 h properties to were enable all calculation interpolated of onto a this depth number of derived quantities. Sorting the profile into are considered here asADCP. Energy the dissipation focus rates is(Prandke, ( on 2005; the Roget EGT etof e al., around 2006) 3 and theand data conductivity revealed sensor a pair noise were floor in mounted terms on of the VMP500. These sensors are un- o cies due to the near-verticalBase magnetic field. near Cape Tidal Armitage, elevation data 17 were km recorded to at the Scott south. tethered free-fall profiler withkey dual property shear being sensors the (Macounto turbulent and W energy Lueck, kg 2004) dissipation with rate 11 km. the The profiles penetrated to around 300 m depth, although only the upper 120 m just beneath the sea ice (i.e.The 2.5 sampling m) recorded with 2 the m firstically thick measurements resolved velocity starting down bins 4 m to every below 5 depthsinstrument this. min. of in Good 120 these m. quality waters data This wheresuitable was is we scatterers typ- have quite (Leonard had good et far penetration al., shallower for 2006; penetration this Stevens due et type to al., of lack 2006). of A magnetic declination 1 km shoreward from the tip on the north side of the EGT (Fig. 1d). The ADCP was held 5 5 25 15 20 10 25 15 20 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | . 2 ) 1 cult and This − . ffi 3 Cape − ff s C above so that (Fig. 3h). 2 (Fig. 3b). ◦ (rad s 3 1 m 8 ect was not − − 7 − ff s . During the − 2 2 ) ect the velocity 1 ∂ρ/∂z ff m − 5 − . 3 − s (rad s 2 5 − m 10 − is gravitational acceleration g ciently enough to a reached 10 ffi 2 , although during the main pulse flow it N (Fig. 3d). At the same time, the stratifi- dropped to levels around 10 1 1 − ε − s reached a maximum of 10 s 4 1446 1445 3 − ε − 10 was at detectable limits of around 10 10 × 2 × N . 2 / 1 associated with these brief accelerations was not necessarily 323.95) showed that the temperature was around 0.1 # ε 2 =  was resolved from the vertical derivative of horizontal velocity com- was observed around the topography of the ridge running o (east-west and north-south, respectively) so that S ∂z ∂v 2 ) v  1 − + 2 and  is a reference (background) density and u ∂z ∂ρ 0 (rad s ∂z ∂u ρ 0 6 g ρ  − " = 10 A single example microstructure profile (Fig. 4) from around the time of maximum The turbulent energy dissipation rate Velocity shear Backscatter amplitude variations as seen here (Fig. 3e) often provide a qualitative = 2 × as it contained thesuggested highest the dissipation influence rates of and, the of base all of the the profiles, EGT it with most strongly strong steps in properties at then, after the finalfell backscatter to minimum the (time noise floor 324.5; of midday the 21 instrument November at 2009),dissipation a little it rate above (time 10 freezing and so supercoolingfrom was other not years apparentother suggests or times this indeed of is the likely.ambient year not fluid However, there (Mahoney the evidence is et al., case much 2010). more every The substantial microstructure year variation example of (Fig. in Fig. temperature 6) 4 in was and chosen the certainly at Such high dissipation rateswater were column mostly for a confined period toso of the around that upper four any hours. 40 elevated m Thecaptured. ringing of After was the the aliased ringing measured in had the finished profiling the N where was greatest during andin just the after the upper pulse 40period m flow after (Fig. the of pulse 3f) the and whenThese ringing water lighter values column fluid bracket appeared and other3 observations in theArmitage. Sound (Stevens et al., 2009) where This was typically inreached the six order times of thiscation 5 value (Fig. i.e. 3f) persisted 3 duringfrequency squared the is strongest calculated section from of the the vertical pulse derivative flow. of The density buoyancy S Furthermore, this diurnal variationThus, was the not drop nearly in as backscatter stronglyCuriously, at may apparent have the on been very other a end days. was response of a entirely the local to data minimum the shownseen in stronger in in backscatter flow. Fig. earlier between 3e measured 20 there tidal and was periods a 40 when m period the depth. when flowponents was there This weaker. e curring 12 h after the pulse.modulation This as, might despite be the a 24(A. diurnal h Mahoney, cycle personal daylight, in communication; a scattering Leonard variation duewith et to in biological al., a signal 2010). can period Our persistco-located when sampling in temporally coincided the the with region what pulse should was be essentially biologically-induced maximum around penetration. midnight local time so it was picture of a tracerproperty to field interpret (Leonard as ettion it of al., is the 2006; an water integrala Stevens column whereby low through et the which acoustic al., value the backscattera 2006). at acoustic at high a beams degree say given It must of 80 depth m flow pass isin is attenuation twice. might the a above a 24 indicate the Hence, di h func- sample few when volume.signal reflectors the There to backscatter at were was noise that three reduced ratio periods depth su shoal. adversely or These so reductions that were the pre- deepest and depth post-pulse, of as good well as quality a data sustained would period oc- et al., 2006; Stevens etaccompanied al., by 2006; relatively Robinson, strong 2010). verticalThe These flows main horizontal body that pulse of reached flows the up pulse were the to generated post-pulse upwards 10 ringing flow cm s contained at the both sampling up position. and However, down flows. proportion of the tidal amplitude observed at locations away from the coast (Leonard 5 5 25 15 20 10 25 15 20 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | erent cult in ff ffi 50–500 m and = L is lengthscale and , 1 − L s in the upper part of the 4 T − L 10 usive layering (Jacobs et al., × ff 2 ect on properties. ff = f erent slopes in the 1–24 cpd bands, erence might be associated with the ff the tip of the EGT as supported by the ff ff 1448 1447 is shedding frequency, f ( erences associated with direction of rotation except ff erence is the di ff f L/u distribution over the entire experiment (Fig. 5), the deeper = T L St at this time was comparable to the local depth of the EGT T L ). This intermediate flow regime suggests variable flow with buildup and 1 − erence at around 20 cpd. This latter di ve expectation that was dispelled by the observations was that the local max- ı ¨ ff ectiveness of spatial mapping of currents. However, such mapping is di ff 3 m s . 0 is velocity) is in the range 0.03–0.3 (assuming The Strouhal number Not only does melting of glacier and ice shelf walls have a strong influence on the There is some evidence for the ringing response being associated with shed vor- In the case of the EGT flow, the so-called ringing was comprised of accelerations = variable stratification near the wall, including double di u release of vorticies (Sobey,Islands 1982). embayment, vorticies Thus, are on likelyclockwise shed the spectral o energy. inward Headland flood studies (Edwards tidethe et into al., e 2004) the have Dellbridge demonstrated the presence ofMorison, fast 2002). ice requiring autonomous vehicle technologyice (e.g., dynamics Hayes (Rignot et and al., 2010; Olbers and Hellmer 2010), it can also result in highly is entirely feasible asnous (i) surface the layer less fast able icedepths to boundary other support than potentially internal that waves generates at and a which (ii) more the internal homoge- obstacle waves persist exists at (e.g.u mountain lees waves). (Fig. 5). Consideration of(Fig. rotational 7) velocity do spectra not both yield nearfor any the (i) strong surface a di and deeper weakcant di emphasis on clockwise“ringing” timescale motion seen at in the lowertwo ADCP frequencies depths, data the and (Fig. only 3). (ii)with obvious When a the di comparing deeper signifi- spectra data fromergy exhibiting the finding a a shallower pathway slope. through the This internal suggests wave there spectrum is beneath greater the en- tongue. This baroclinic waves, respectively (see Albrechttimescale et in al., the 2006 region). forboundary. a description Of of course, baroclinic the waves couldticies. be associated Certainly with the a di of local seiche across to Tent Island, being too slow and too fast for barotropic and around 70 min apart.flow Presumably, accelerates this during marks the either pulse. a wave The or 70 vortex min process period as doesn’t the particularly fit any kind imum depth of thein EGT would stratification be or consistently aapparent clear is strongly in probably the consistent data, indicative velocitynot either of shear level through the either feature. a fact along, step thatthe or That relatively the across, these weak underside its local were of mainthe stratification not the glacier longitudinal will tongue glacier axis retard closer tongue vertical (Holdsworth, toobservation is flows, the 1982). position grounding so thus line Even that spreading might the the still influence depth have of a of the strong e influence at the 4 Discussion 4.1 Kinematics and water column structure One na overturn that results in theof heaviest 108 m. water in Considering the the partiallarge overturn profile being is located seen at tooverturn a be of the depth around largest 70 recorded. mwater (at The column time pulse 323.8). consisted was preceded mainly During by the ofexample. a pulse overturns strong and After in ringing the the period ringing 0–2070 the m m had upper overturns range ceased as which (324.2), then seen fell in there away the was to Fig. a 40 m 4 relatively or large less. period of such temperatures) exhibited ansurface unusual or local minimum near-wall at fluidsome 50 swept m. large to scale Potentially structure the this inwater base was the column. of Thorpe However, the as overturning the scale EGTof profile by this do vortices. measure not go are This all possible the generates so way to that the this bed, deeper is values a lower bound as there is a large-scale 50 and 100 m. The salinity (and hence density, as it follows salinity very closely at 5 5 25 15 20 10 20 25 15 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | . ρ 1– K and and and = 100, 2 gr = ) gr 1 gr Ri − 1 usivity is Ri − Ri ff usive con- usivities of s ff ff usion. How- 2 ff (rad s m 5 3 − − 10 10 = × 2 3 . N 1 (Fig. 9). The di = estimates presented here ρ ρ does have some consistent m0 K K K gr Ri 1, implying mixing due to shear is usion-induced mixing as we have ff < usivity gr . Under these conditions ff 2 Ri − s 8 which are, like the comparison made in − usivities suggest that such processes and ρ ff 10 K 1450 1449 × may not be particularly representative of the ). These modifications, based on observations 3 ρ 2 = K 2 N S and ) rose – albeit for localized moments in time and pre- ε m K and that it is likely maximum very close to a headland. Present . There are trends between the estimate of the ε 1 m0 0 − ρ K s K + 2 + 2 3 m ) ) − 4 gr gr − 10 Ri Ri m 10 5 , × 5 K × 2 2 5 + + . 3 S N . 5 (1 (1 1 orts continue to determine the functionality of vertical mixing in response to the = ff = = = 0 gr E In order to parameterize the likelihood of instability and mixing, the gradient Richard- ρ ρ m scales in forcing.conditions that The spawned the observed mixing. It is useful to consider the relationship between whereK backgroundthe measured levels dissipation andsomewhat inferred higher vertical than di the are model10 determined band. by Fer There (2006) is –use given an especially in important the the point Shih hereStevens et though. et by al., al. The (2009), 2005However, somewhat estimate the higher for Fer than (2006) thebound (see model Osborne Zaron method uses and for the Moum, estimating 2009). Osborn The model discrepancy which possibly it lies in notes the is likely small an upper K implying mixing was unlikely. balance captured in Eq. (1)fication (e.g., of Zaron the and Pacanowski Moum, andmomentum 2009). Philander and (1981) Fer formulations (2006) buoyancy for suggests ( vertical modi- in di an Arctic Fiord and so likely relevant here, areK given as is amplified. However, thestructure (Fig. depth-time 8) distribution so of that inlikely the and core this of was the borne pulse the out pulse by feature, stratification the persisted dissipationbackground more rate shear strongly observations. so was At that more times like away from intermittent as it is the ratio of two derivative properties so that noise and uncertainty a measure of shear-induced instabilitylikely whereby values to substantially be less or than unity become are unstable, can be calculated. The flow observed here is rather pation rates and concomitantscales enhance are di important.influence. Intermediate scales of regional modeling wouldson elucidate Number, this Ri as normal background levels.ancy flux Edwards scaled et with al.modeling (2004) (Dinniman noted et how local al.,is turbulent at 2007; buoy- a Reddy scalevariability. et that However, al., as is 2010) seen unlikely(Muench in that to et benthic encompasses be observations al., the instantaneously to 2009), influenced region the over even north long by of timescales this McMurdo the degree Sound continual of focus of high energy dissi- quiescent and thus conducive toration the of accumulation the of melting meltwater. processesconditions Hence, needs at further to timescales incorporate explo- far less temporal than variation a in tidal the period. ambient 4.2 Mixing rates The energy dissipation ratesumably space ( – as much as three orders of magnitude above what might be regarded vection as observed into the compare the Jacobs magnitudes study, of itno shear-induced is complementary and di beyond observations the ofpulse scope the feature would of latter sweep away the situation. layering structureever, present generated the It by work pulse double seems is di highly transient and likely so that for the the majority of the time conditions were relatively 1981). Whilst the present study was initially motivated by the search for di 5 5 25 15 20 10 25 15 20 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | then 1 − usivities possible ff C change in the freezing point and ◦ 0.075 culty in separating advection versus actual − ect. If, on the other hand, the flow were rec- ffi ff 1452 1451 ries and Weeks, 1992). Such conditions arise ff cult to quantify as they inherently depend on their ffi then placing dissipation rate estimates on a displacement 2 ) 1 − (rad s 4 − 10 × of 2 2 N A subjective series of sketches (Fig. 11) shows the likely flow evolution around the Consequently, the relative scale of separated flow to that of the tongue is likely much The EGT has quite a large aspect ratio (length/width) when compared to the Drygal- Platelet growth rates are very di fluctuating vertical flows, and theright very next to large the dissipation ice rates wall. and di EGT. The tidally-driven flow is highlyerates quite rectified specific and mixing also events variable intranslate with time this depth and description and presumably to so space. other gen- It systems. would be useful to larger with the EGT.that However, vertical vertical flows scales mayto are be both greater the even for larger stronger. thedata ice demonstrate big the Where tongues ice small and the tongues temporal indeed present scales so of ice data variability shelf are in fronts turbulence, of the is relevance strong at and the small scale. The ski (and theups). Mertz At and the Ninnisthese same Glacier giant time, Tongues prior the glacier to EGT tongues.currents their is and recent Possibly smaller waves the break by in EGT thethe absolute Dellbridge area. is scale Islands atypical by and a in also factor that the of it generally ten is low from tidal protected flows from in Sharp headlands strongly influencewards et local al., flow 2004). through Althoughthe the rectification headland EGT’s of finger-like aspect morphology tides ratio is (e.g.,than at scale the the Ed- and bed extreme glacier end we of tongues expectflow. there are The is located headland much at in work the common ofform in Edwards surface drag terms et and rather of that al. oceanic the (2004)by response majority identified the to of the headland. tidal the split diapycnal between turbulent skin buoyancy flux and occurred close possibilities exist in the internalof waves the observed EGT at (Robinson, Cape 2010). Armitage 17 km to4.4 the south Generalization so highly influential if the local water is very cold. Similar temporally variable generation tified in some way whichmechanics, is this displacement likely represents given a the non-linearity of the “flow around headland” the individual events might betime. expected to advect water around 100 m vertically ininitial this conditions and observations havegrowth. di However, Leonard et al.on (2006) near-surface observed lines 30 after cmEGT periods cylinders situation, of as water platelets short advected growing up asto from three depth beneath it days. the is tongue Consequently unlikely were there if, immediately would as returned be in any e the increases it’s propensity to freeze. Whileperiod this observed is here, unlikely in there thewhen have particular been it relatively warm previous has times beenserved even possible kinematics at as are the relevant shown same toa in time the of weak CTD supercooling year problem. data upward The of flowhas ADCP Fig. a for data 6. half-period most do timescale show of Thus, of the the 35 min presently time. and ob- a Furthermore, vertical as speed of the around post-pulse 0.05 m “ringing” s 4.3 Local supercooling A number of authors havepercooled inferred water that (Debenham, glacier/ice 1965; tongues Je when might seawater be in generators contact of withsome su- ice later at time depth it rises ismay in cooled result the to in water the column, the in the fluid situ increase freezing actually in point. being local If, freezing colder at temperature than this temperature. This substantially an diagram it is clear that therates eastward and flowing the periods generated scales the of strongestthe the dissipation tip “action at of a the distance”of glacier were direct tongue. comparable flow-obstacle with interaction. Thus, the distance the to observations downstream might be the result displacement and energy dissipation. Assuming a timescale of 8 min associated with 5 5 25 15 20 10 25 15 20 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | ect on ff 1454 1453 ¨ okmen, T., Hallberg, R., and Padman, L.: Impacts of bottom cor- ¨ olemann, J. A., Klagge, T., Tyshko, K. P., and Busche, T.: Observa- ¨ Ozg The authors wish to thank Pat Langhorne, Andy Mahoney, Alex Gough, ˚ ¨ ahlin, A., oder, D., H ´ ephane Popinet, David Plew, Brian Stait and Peter Gerring assisted with field work. erent sized probes, J. Atmos. Ocean. Tech., 21, 284–297, 2004. ff ries, M. O. and Weeks, W. F.: Structural characteristics and development of sea ice in the Haskell, T.: Thesea seasonal ice growth, arrival in of preparation, 2010. ice shelf watercesses, Springer, in 215 pp., McMurdo 2008. Sound and itsrugations e on a dense Antarctic outflow: the NW , Geophys. Res. Lett., 36, 23, crystals in the waterIAHR Proceedings column: 20th the IAHR International view Symposium from on an Ice, Helsinki, Acousticdi 2010. Doppler Current Profiler (ADCP), Frew, R. D.:the Observations winter of ofdoi:10.1029/2005JC002952, 2006. platelet 2003 ice in growth McMurdo and Sound, oceanographic conditions , during J. Geophys. Res., 111, C04012, by melting of the Erebus Glacier Tongue, J. Geophys. Res.,Western 86, Ross C7, Sea, 6547–6555, Antarct. 1981. Sci., 5(1), 63–75, 1992. current on Mertz Glacier, Antarctica, J. Glaciol., 50, 427–435, 2004. mixing due to tidal flow past a sharp point,mon, J. 634 Phys. pp., Oceanogr., 1997. 34, 1297–1312, 2004. , Antarct. Sci., 9, 61–73, 1997. salt with an autonomous underwater vehicle, J. Atmos. Ocean.Geophys. Tech., 19, Res. 759–779, Lett. 2002. 31, L10307, doi:10.1029/2004GL019506, 2004. big, A., Schr tions of supercooling and frazilRes., ice 115, formation C05015, in doi:10.1029/2009JC005798, the 2010. Laptev Sea coastal polynya, J. Geophys. Zeal. J. Mar. Fresh., 37, 241–249, 2003. Tongue, New Zealand Antarctic Record, 9, 15–30, 1989. on circulation and waterAntarctica, mass J. formation Geophys. Res., in 112, a C11013, numerical doi:10.1029/2006JC004036, circulation 2007. model of the Ross Sea, Haskell, T. G.:orientation Supercooling in and sea ice seeding:Symposium growing on near oceanographic Ice, ice Helsinki, disruption shelves, 2010. IAHR of Proceedings preferred 20th crystal IAHR International 391–419, 1975. Observation of sub-inertial internal tides33, in McMurdo L24606 Sound, doi:10.1029/2006GL027377, Antarctica, 2006. Geophys. Res. Lett., ff Muench, R., W Mahoney, A., Gough, A., Langhorne, P., Robinson,McPhee, M. N., G.: Stevens, Air-Ice-Ocean-Interaction: C., Turbulent Ocean Williams, Boundary Layer M., Exchange Pro- and Macoun, P. and Lueck, R. G.: Modelling the spatial response of the air-foil shear probe using Leonard, G. H., Purdie, C. R., Langhorne, P. J., Williams, M. J. M., and Haskell, T. G.: Ice Leonard, G. H., Purdie, C. R. Langhorne, P. J., Haskell, T. G., Williams, M. J. M., and Je Legresy, B., Wendt, A., Tabacco, I., Remy, F., and Dietrich, R.: Influence of tides and tidal Jacobs, S. S., Huppert, H. E., Holdsworth, G., and Drewry, D. J.: Thermohaline steps induced Holdsworth, G.: Dynamics of Erebus Glacier tongue. Ann. Glaciol., 3, 131–137, 1982. Frezzotti, M.: Ice front fluctuation, calving flux and mass balance of Victoria Land Hellmer, H. H.: Impact of Antarctic ice shelf basal melting on sea ice and deep ocean properties, Emery, W. J. and Thomson, R. E.: DataFer, Analysis I.: Methods Scaling turbulent in dissipation Physical in Oceanography, an Perga- Arctic fjord, Deep-Sea Res. Pt. II, 53, 77–95, 2006. Hayes, D. R. and Morison, J. H.: Determining turbulent, vertical velocity, and fluxes of heat and Edwards, K., MacCready, P., Moum, J., Pawlak, G., Klymak, J., and Perlin, A.: Form drag and Dmitrenko, I. A., Wegner, C., Kassens, H., Kirillov, S. A., Krumpen, T., Heinemann, G., Hel- Goring, D. G. and Pyne, A.: Observations of sea-level variability in Ross Sea, Antarctica, New Dinniman, M. S., Klinck, J. M., and Smith, Jr., W. O.: Influence of sea ice cover and DeLisle, G., Chinn, T., Karlen, W., and Winters, P.: Radio echosounding of Erebus Glacier Gough, A. J., Mahoney, A., Langhorne, P. J., Williams, M. J. M., Stevens, C. L., and Debenham, F.: The glacier tongues of McMurdo Sound, Geogr. J., 131, 369–71, 1965. Deacon, G. E. R.: The oceanographic observations of Scott’s last expedition, Polar Rec., 17, Albrecht, N., Vennell, R., Williams, M., Stevens, C., Langhorne, P., Leonard, G., and Haskell, T.: of Don Lewis. References Robin Robertson and MelissaGrant, Bowen St for usefulFunding discussions and associated support were with providedFund, this by Antarctica The work. New New Zealand Zealand Brett Royalsearch (Events Society Science K131/K132) administered and Marsden and Technology (Contract the C01X0701). New This Zealand paper Foundation is for dedicated Re- to the memory Acknowledgements. 5 5 30 25 20 15 30 10 25 20 15 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 1455 1456 , J. R., Ivey, G. N., and Ferziger, J. H.: Parameterization of turbulent fluxes and ff on the Erebus Ice Tongue,377–385, McMurdo 1994. Sound, Antarctica, and nearby sea ice, J.Mech., Glaciol., 125, 40, 359–373, 1982. natural waters: methodology and applications, Progr. Oceanogr., 70, 126–148,scales 2006. using homogeneous sheared525, stably 193–214, stratified 2005. turbulence simulations, J. Fluid Mech., mixed layer measurements atdoi:10.1029/2008JC005141, 2010. Maud Rise, Weddell Sea, J. Geophys. Res., 115, C02009, ocean modeling, J. Phys. Oceanogr., 39(10), 2652–2664, 2009. shelf, Ph.D. Thesis, Univ. Otago, 2010. tions of the stratified turbulentIce, boundary-layer Proc. and platelet Sixth ice International beneathAustralia, Symposium McMurdo 2006. Sound on Sea Stratified Flows, edited by: Ivey, G. N., Perth, 1990. ice–ocean interaction beneath theC03025, doi:10.1029/2008JC005255, McMurdo 2010. Ice Shelf, Antarctica, J. Geophys. Res., 115, bulence beneath sea icedoi:10.5194/os-5-435-2009, in 2009. southern McMurdo Sound, Antarctica, Ocean Sci., 5, 435–445, Greenland Glaciers, Nat. Geosci., 3, 187–191, doi:10.1038/ngeo765, 2010. els, edited by: Baumert, H.101–109, Z., 2005. Simpson, J. H., and Sundermann, J., Cambridge Univ.time, Press, and melting:30, results 733–742, from doi:10.1016/j.csr.2010.01.007, 2010. a model of oceanic chlorofluorocarbons, Cont. Shelf Res., models of tropical oceans, J. Phys. Oceanogr., 11, 1443–1451, 1981. doi:10.1029/2009GL041347, 2009. Dynam., 60, 141–153, doi:10.1007/s10236-009-0252-z, 2010. Sobey, I. J.: Oscillatory flows at intermediate Strouhal number in asymmetric channels, J. Fluid Squire, V. A., Robinson, W. H., Meylan, M., and Haskell, T. G.: Observations of flexural waves Sirevaag, A., McPhee, M. G., Morison, J. H., Shaw, W. J., and Stanton, T. P.: Wintertime Shih, L. H., Kose Zaron, E. D. and Moum, J. N.: A new look at Richardson Number mixing schemes for equatorial Roget, E., Lozovatsky, I., Sanchez, X., and Figueroa, M.: Microstructure measurements in Robinson N. J.: The ocean boundary layer beneath Antarctic Sea ice in the vicinity of an ice Stevens, C. L., Williams, M. J. M., Robinson, N. J., Albrecht, N., and Haskell, T. G.: Observa- Robinson, N. J., Williams, M. J. M., Barrett, P. J., and Pyne, A. R.: Observations of flow and Stevens, C. L., Robinson, N. J., Williams, M. J. M., and Haskell, T. G.: Observations of tur- Robinson, W. and Haskell, T. G.: Calving of Erebus Glacier tongue, Nature 346, 615–616. Rignot, E., Koppes, M., and Velicogna, I.: Rapid submarine melting of the calving faces of West Reddy, T. E., Holland, D. M., and Arrigo, K. R.: Ross ice shelf cavity circulation, residence Prandke, H.: Microstructure sensors, in: Marine Turbulence: Theories, Observations and Mod- Pacanowski, R. C. and Philander, S. G. H.: Parameterization of vertical mixing in numerical Olbers, D. and Hellmer, H.: A box model of circulation and melting in ice shelf caverns, Ocean 5 5 30 25 20 15 10 Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | the McMurdo and (b) to highlight a five day segment (b) 1458 1457 at is expanded (a) shows velocity feather vectors for 3 depth bins centered at 11, 41 and The Erebus glacier tongue station (EGT) and the complex topography (c) ADCP (grey box in a) and turbulence profiling (green box in all panels) (c) (c) A cross-section of the EGT simplified from DeLisle et al. (1989). McMurdo Sound located in the South Western Ross Sea near (e) The tidal elevation was around 1000 m east of the tip of the EGT. The grey lines highlight known bathymetric were recorded. Panel 71 m. within which our Fig. 2. of the Dellbridge Islands (DI)Armitage are (CA), marked Haskell relative Strait to (HS)(d) Erebus and Bay Scott (EB), Base Cape (SB). Evansfeatures (CE). The whilst Cape microstructure the blue field dashed campIsland line (MSC) (TI) shows a and tongue BigZealand. of multi Razorback year Island fast (BRI). ice extending Map out excerpt to Tent courtesy of Land Information New Ross Ice Shelves. Fig. 1. (a) Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | , , t t σ σ vertical . (d) T , L density as w . (f) ε horizontal velocity mag- temperature, density as (a) (a) vertical velocity (c) and overturning scale ε 323.95 showing = turbulent energy dissipation rate 1460 1459 dissipation rate (h) (b) and 2 N and 2 N backscatter amplitude. The VMP turbulence profiler resolved horizontal direction in which the flow is going, (e) (b) ADCP and turbulence data products. The ADCP resolved Single microstructure profile from time buoyancy frequency squared buoyancy frequency squared Fig. 4. Fig. 3. nitude, shear and (g) Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | + 1462 1461 salinity profile data collected with Seabird Electronics 19 (b) Temperature and distribution. T L CTD at various locationsriod within during 500 2005, m 2008 of(2009-ms). and the 2009, tip The plus of dashed asalinity the line single of in EGT example 34.65 the from from PSU. upper the theissues. The present part October–November maximum The microstructure pe- of depth other data (a) profiles fromin all is the depth penetrated the 2005 in to profiles freezing the within were temperature region. 20 limited m at of by a the operational nominal bed indicating the large variation Fig. 6. (a) Fig. 5. Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | estimate. gr Ri the (b) 1464 1463 set upwards by two decades. Reference sloped lines are shown ff Velocity magnitude as per Fig. 3a for reference and Rotational spectra of ADCP-derived velocities from depths 9–17 m and 59–67 m. The Fig. 8. (a) with dashes. Fig. 7. 95% confidence limit is shownfrom in shallower the depths top are right o (Emery and Thomson, 1997). The upper results Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | . Y estimate and including ρ and K X usivity ff showing individual estimates (hollow squares), gr 1466 1465 Ri usivity of mass (dashed line). ff The same plot format for the di (b) ) spaced bins. Dissipation rate as a function of gr Dissipation rate as a function of “horizontal displacement” Ri 15log( . Fig. 10. Fig. 9. (a) bin-log-space-averaged (solid squares) and0 bars showing standard deviationthe of Fer (2006) estimates estimate for of vertical di Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 323.7, (a) 50 m) are solid and deeper flows dotted. < 1467 324.5. The top panel has Tent Island (TI) for approximate (e) 324.3 and (d) 324.1, (c) Interpretive sketches of flow around the tip of the glacier tongue with the left hand 324.0, (b) scale and orientation although plan and elevation views are not to the same scale. The right hand panels include mixing as red lines. Panels are at approximate times Fig. 11. panels showing plan view andEGT right looking hand to panels the showing east. elevation view On of the the left north the side upper of flows the (