Article Stratigraphic Analysis of Firn Cores from an Antarctic Ice Shelf Firn Aquifer
Shelley MacDonell 1,* , Francisco Fernandoy 2 , Paula Villar 2 and Arno Hammann 1,†
1 Centro de Estudios Avanzados en Zonas Áridas (CEAZA), Raúl Bitrán 1305, La Serena 1700000, Chile; [email protected] 2 Laboratorio de Análisis Isotópico, Universidad Andrés Bello, Viña del Mar 2531015, Chile; [email protected] (F.F.); [email protected] (P.V.) * Correspondence: [email protected] † Now at: Asiaq—Greenland Survey, Qatserisut, Nuuk 3900, Greenland.
Abstract: In recent decades, several large ice shelves in the Antarctic Peninsula region have experi- enced significant ice loss, likely driven by a combination of oceanic, atmospheric and hydrological processes. All three areas need further research, however, in the case of the role of liquid water the first concern is to address the paucity of field measurements. Despite this shortage of field observations, several authors have proposed the existence of firn aquifers on Antarctic ice shelves, however little is known about their distribution, formation, extension and role in ice shelf mechanics. In this study we present the discovery of saturated firn at three drill sites on the Müller Ice Shelf (67◦140 S; 66◦520 W), which leads us to conclude that either a large contiguous or several disconnected smaller firn aquifers exist on this ice shelf. From the stratigraphic analysis of three short firn cores extracted during February 2019 we describe a new classification system to identify the structures and morphological signatures of refrozen meltwater, identify evidence of superficial meltwater percolation, and use this information to propose a conceptual model of firn aquifer development Citation: MacDonell, S.; Fernandoy, on the Müller Ice Shelf. The detailed stratigraphic analysis of the sampled cores will provide an F.; Villar, P.; Hammann, A. invaluable baseline for modelling studies. Stratigraphic Analysis of Firn Cores from an Antarctic Ice Shelf Firn Keywords: ice shelf; firn aquifer; ice core; stratigraphy; glacier hydrology; Antarctica Aquifer. Water 2021, 13, 731. https://doi.org/10.3390/w13050731
Academic Editor: Michael Kuhn 1. Introduction Liquid water plays key roles in the cryosphere, for example: the flow of land ice is Received: 5 February 2021 predominately controlled by the quantity of liquid water present (e.g., [1]); energy fluxes in Accepted: 3 March 2021 Published: 8 March 2021 snowpacks are influenced by the presence and behaviour of liquid water (e.g., [2]); and many theories of ice shelf disintegration involve the presence and action of water (e.g., [3]).
Publisher’s Note: MDPI stays neutral While the hydrological system on land-based glaciers is well defined, very few direct field with regard to jurisdictional claims in observations have been carried out on ice shelves (e.g., [4]). published maps and institutional affil- Understanding the distribution of liquid and refrozen water on ice shelves is im- iations. portant for understanding ice mechanics and ice shelf dynamics [5,6]. Liquid water can cause hydrofracturing (when increased pressure is exerted by water on the tips of vertical fractures) [3,7] and when accumulated into ponds, can cause bending stresses and fracture formation from rapid drainage events [3,8–12]. On ice sheets, the refreezing of liquid water in superficial firn [13] can cause the impedance of water percolation, therein enhancing Copyright: © 2021 by the authors. either runoff or internal ponding. Comparatively, on Antarctic ice shelves, [14] identified Licensee MDPI, Basel, Switzerland. This article is an open access article such a refrozen surface on the Larsen C ice shelf using geophysical surveys, [6] reported distributed under the terms and from remotely sensed images meltwater being exported via a supraglacial stream network conditions of the Creative Commons and [4] described the hydrological properties of a firn aquifer on the Wilkins Ice Shelf. The Attribution (CC BY) license (https:// diversity of hydrological system structures, and possible impacts makes realistic modelling creativecommons.org/licenses/by/ of ice sheet and ice shelf processes difficult [15,16]. This is further exacerbated by the dearth 4.0/). of field measurements of the mechanical and hydrological properties of ice shelves.
Water 2021, 13, 731. https://doi.org/10.3390/w13050731 https://www.mdpi.com/journal/water Water 2021, 13, x FOR PEER REVIEW 2 of 18
Water 2021, 13, 731 by the dearth of field measurements of the mechanical and hydrological properties2 of of ice 17 shelves. Several authors have proposed the existence of firn aquifers on Antarctic ice shelves (e.g., [4,17]), however little is known about their distribution, formation, extension and Several authors have proposed the existence of firn aquifers on Antarctic ice shelves role in ice shelf mechanics. In this study we present the discovery of fully saturated firn (e.g., [4,17]), however little is known about their distribution, formation, extension and on the Müller Ice Shelf, which leads us towards the conclusion of at least one large firn role in ice shelf mechanics. In this study we present the discovery of fully saturated firn onaquifer the Müller or disconnected Ice Shelf, which smaller leads firn us aquifers towards on the this conclusion ice shelf. ofWe at will least present one large results firn aquiferbased on or disconnectedthe stratigraphic smaller analysis firn aquifers of three on short this icefirn shelf. cores We extracted will present during results February based on2019. the In stratigraphic particular, this analysis study of aims three to shortdevelop firn a coresclassification extracted system during to February identify the 2019. struc- In particular,tures and morphological this study aims signatures to develop of a refrozen classification meltwater, system identify to identify evidence the structures of superficial and morphologicalmeltwater percolation, signatures and of refrozenuse this informat meltwater,ion identify to construct evidence a conceptual of superficial understanding meltwater percolation,firn aquifers and on usethe Müller this information Ice Shelf. to construct a conceptual understanding firn aquifers on the Müller Ice Shelf. 2. Materials and Methods 2.2.1. Materials Study Area and Methods 2.1. StudyThe Müller Area Ice Shelf is the northern-most ice shelf on the Antarctic Peninsula (AP) (67°14The′ S; Müller 66°52′ IceW) Shelf(Figure is the1). It northern-most is located on the ice shelfwestern on theside Antarctic of the AP Peninsula and drains (AP) the ◦ 0 ◦ 0 (67Arrowsmith14 S; 66 52Peninsula.W) (Figure It is1 fed). It by is the located Brückner on the and western Antevs side glaciers of the and AP terminates and drains in the the ArrowsmithLallemand Fjord. Peninsula. The Itice is shelf fed by is therelatively Brückner small and Antevs(40 km2 glaciers) and has and been terminates experiencing in the 2 Lallemandphases of advance Fjord. The and ice retreat shelf issince relatively at least small the Little (40 km Ice) andAge has[18], been and experiencing now covers approx- phases ofimately advance 50% and of retreatthe area since recorded at least in the 1957 Little [19] Ice. An Age important [18], and question now covers is why approximately the Müller 50%Ice Shelf of the remains area recorded extant, whereas in 1957 its [19 neighbours]. An important (e.g., Jones question Ice shelf) is why have the collapsed Müller [19]. Ice ShelfHypotheses remains related extant, to whereas ice shelf its geometry, neighbours the (e.g.,amount Jones of ice Ice discharge shelf) have relative collapsed to catch- [19]. Hypothesesment size as related well as to protection ice shelf geometry, from the theocean amount have of been ice dischargeproposed, relative but the to relationship catchment sizebetween as well the as ice protection shelf, climate from theand ocean ocean have processes been proposed,remain unclear. but the relationship between the ice shelf, climate and ocean processes remain unclear.
(a) (b)
Figure 1. Map of the study area. (a) shows the position on the Antarctic Peninsula, (b) shows the wider Müller Ice Shelf region, and the red points correspond to the ice core extraction sites.
Water 2021, 13, x FOR PEER REVIEW 3 of 18
Water 2021, 13, 731 3 of 17 Figure 1. Map of the study area. (a) shows the position on the Antarctic Peninsula, (b) shows the wider Müller Ice Shelf region, and the red points correspond to the ice core extraction sites.
2.2. Sample Collection and Handling A series of of three three short short firn firn cores cores were were colle collectedcted from from the the Müller Müller Ice Ice Shelf Shelf between between 20 and20 and 22 February 22 February 2019 2019(Figures (Figures 1 and1 2).and The2). drill The used drill had used a maximum had a maximum reach of reach 20 m, ofbut ultimately20 m, but ultimately core lengths core were lengths determined were determined by the presence by the of presence saturated of material saturated impeding material theimpeding deeper the penetration deeper penetration of the device. of the The device. core TheM1 corewas M1located was locatedat the confluence at the confluence of the ◦ 0 ◦ 0 Brücknerof the Brückner and Antevs and Antevs glaciers glaciers (67°15′15.5 (67 ″15 S; 15.5”66°53′ S;44.5 66″ 53W)44.5” and reached W) and a reached depth of a depth16.72 m. of 16.72M2 was m. M2located was locatednear the near centre the centreof the of ice the shelf ice shelf in the in thearea area fed fed by by the the Antevs Antevs Glacier ◦ 0 ◦ 0 (67(67°1515′16.816.8”″ S; S; 66°50 66 50′59.859.8”″ W), W), and and was was 18.54 18.54 m m long. long. Comparatively, Comparatively, M3 was 3.72 m long and was extracted near the eastern limit of the ice shelf also fed by the AntevsAntevs GlacierGlacier ◦ 0 ◦ 0 (67(67°1515′37.637.6”″ S; S; 66°48 66 48′22.322.3”″ W). W). The The three three cores cores were were extracte extractedd using using a 5.7 a cm 5.7 diameter cm diameter elec- tricelectric drill drill (Icedrill.ch (Icedrill.ch AG), AG), and andwere were immediately immediately weighed, weighed, measured measured and bagged and bagged for stor- for agestorage in insulated in insulated cold cold store store boxes. boxes. At the At same the same time, time, firn temperatures firn temperatures were were recorded recorded and aand visual a visual analysis analysis of relative of relative saturation, saturation, possible possible melt melt layers layers and and textual textual characteristics were recorded. The cores were flownflown by helicopter to the Betanzos research vessel and were subsequently stored on the ship at −20 ◦C until arriving at Punta Arenas, Chile. From were subsequently stored on the ship at −20 °C until arriving at Punta Arenas, Chile. From Punta Arenas the cores were flown to the Universidad Andrés Bello, Viña del Mar where Punta Arenas the cores were flown to the Universidad Andrés Bello, Viña del Mar where they were stored in a purpose-built cold facility at −20 ◦C. they were stored in a purpose-built cold facility at −20 °C.
Figure 2.2. Photograph of core handling on the MüllerMüller Ice Shelf. Two teamteam membersmembers operatedoperated the electricelectric drill, whilst the other two members bagged and recorded the extracted cores. other two members bagged and recorded the extracted cores.
2.3. Laboratory Analysis To analyse the collected cores, and to provide a platform for future work, a purpose- built cold store facility was built at the Universidad AndrAndrésés Bello.Bello. This enabled the clean processing of the collected cores. As part of the constructed space, we built an area to cut and handle cores, as well as a light box (Figure 3 3)) toto facilitatefacilitate thethe visualizationvisualization ofof iceice crystalcrystal configurationsconfigurations within the core. Firn core processing consisted in firstly re-measuring and weighing the individual sections of each core to re-calculate the associated density. Subsequently the cores were longitudinally cut in half using a Bauker circular saw, and one half was used in this study, and the other section was retained for future analyses. A continuous stratigraphic analysis was then undertaken on the core which consisted in recording a visual description of each section, including variations in grain size and morphology, ice lens morphology and size, and interface characteristics between identified units. High resolution photographs were also taken along the length of each core to facilitate post-processing. To avoid confusing
Water 2021, 13, 731 4 of 17
ice that was refrozen in transport or storage from ice that was refrozen in situ, sections Water 2021, 13, x FOR PEER REVIEW 4 of 18 that presented a solid ice layer near the border with no bubbles were discarded as being identified as refrozen post-extraction.
FigureFigure 3. Photograph of of the the constructed constructed light light box. box. Th Thee displayed displayed core core is isapproximately approximately 90 90cm cm in in length.length.
3. ResultsFirn core processing consisted in firstly re-measuring and weighing the individual 3.1.sections Field of Observations each core to re-calculate the associated density. Subsequently the cores were longitudinallyAt each of cut the in three half coreusing sites, a Bauker a mix circular of damp saw, and and dry one firn half was was encountered, used in this andstudy, at theand base the other of each section drilled was hole, retained the firn for wasfuture saturated. analyses. However, A continuous at site stratigraphic M3, the recovered analysis corewas then was saturatedundertaken to on the the point core that which the consisted drill filled in with recording liquid a water visual (Figure description4; Video of each S1). Thissection, saturated including section variations was found in grain at a size shorter and distancemorphology, (3.72 ice m) lens from morphology the surface and than size, the otherand interface two sites characteristics (16.72 m and 18.54 between m), whichidentified is likely units. due High to its resolution lower elevation photographs with respect were toalso the taken other along two drillthe length locations. of each To analyse core to thefacilitate provenance post-processing. of the water, To theavoid conductivity confusing ofice the that lower was halfrefrozen of the in M1 transport core was oranalysed storage from in the ice lab that and was found refrozen to be <22.5in situ, mS sections along thethat length presented of the a measuredsolid ice layer core near which the corresponds border with to no fresh bubbles water. were This discarded suggests as that being the liquididentified water as foundrefrozen at eachpost-extraction. site was due to snow and firn melting as opposed to saltwater intrusion. In addition, the temperature recorded in the firn along the length of each core was3. Results between −0.1 ◦C and 0.4 ◦C, and in the case of M1 and M3, the highest temperatures ◦ were3.1. Field recorded Observations at the base of each core (0.4 C), whereas the base of M2 was at the melting point (0.0–0.1 ◦C). The recorded temperatures suggest that melting is the likely source of At each of the three core sites, a mix of damp and dry firn was encountered, and at water found in the cores. the base of each drilled hole, the firn was saturated. However, at site M3, the recovered 3.2.core Framework was saturated to Describe to the Firnpoint Stratigraphy that the drill filled with liquid water (Figure 4; Video S1). This saturated section was found at a shorter distance (3.72 m) from the surface than the In order to standardise the comparison of the firn core stratigraphies, a new framework other two sites (16.72 m and 18.54 m), which is likely due to its lower elevation with re- was developed based on the analysed cores. We classified the structures into three classes spect to the other two drill locations. To analyse the provenance of the water, the conduc- each of granulometry, contacts and recrystalized structures. tivity of the lower half of the M1 core was analysed in the lab and found to be <22.5 mS 3.2.1.along Granulometrythe length of the measured core which corresponds to fresh water. This suggests that Thethe liquid granulometry water found classification at each site is was based due on to threesnow grainand firn size melting bins: fineas opposed (<1 mm), to mediumsaltwater (1–2 intrusion. mm), and In addition, large (>3 mm)the temperatur (Figure5). Thee recorded differentiation in the firn between along predominant the length of graineach core sizes was was between assessed − using0.1 °C the and light 0.4 box°C, and (Figure in the4). Finecase grainedof M1 and sections M3, the are highest easily detectedtemperatures using were the lightbox,recorded asat littlethe base light of is each able core to penetrate (0.4 °C), throughwhereas tightlythe base packed of M2 finewas crystals.at the melting Comparatively point (0.0–0.1 more °C). light The is refracted recorded through temperatures medium suggest grain sections, that melting and finally, is the muchlikely moresource light of water is able found to pass in throughthe cores. large grained sections.
Water 2021, 13, x FOR PEER REVIEW 5 of 18
Water 2021, 13, 731 5 of 17 Water 2021, 13, x FOR PEER REVIEW 5 of 18
Figure 4. Water being drained from the upper section of the drill after the extraction of the lower section of M3. In Video S1, a video of the extraction of the lower section of core is included.
3.2. Framework to Describe Firn Stratigraphy In order to standardise the comparison of the firn core stratigraphies, a new frame- work was developed based on the analysed cores. We classified the structures into three classes each of granulometry, contacts and recrystalized structures.
3.2.1. Granulometry The granulometry classification is based on three grain size bins: fine (<1 mm), me- dium (1–2 mm), and large (>3 mm) (Figure 5). The differentiation between predominant grain sizes was assessed using the light box (Figure 4). Fine grained sections are easily detected using the lightbox, as little light is able to penetrate through tightly packed fine
crystals. Comparatively more light is refracted through medium grain sections, and fi- Figure 4. Water being drained from the upper section of the drill after the extraction of the lower nally,Figure much 4. Water more being light drained is able from to pass the upper through section large of grained the drill aftersections. the extraction of the lower sectionsection of of M3. M3. InIn VideoVideo S1, S1, a a video video of of the the extraction extraction of of the the lower lower section section of of core core is is included. included.
3.2. Framework to Describe Firn Stratigraphy In order to standardise the comparison of the firn core stratigraphies, a new frame- work was developed based on the analysed cores. We classified the structures into three classes each of granulometry, contacts and recrystalized structures.
3.2.1. Granulometry The granulometry classification is based on three grain size bins: fine (<1 mm), me- dium (1–2 mm), and large (>3 mm) (Figure 5). The differentiation between predominant grain sizes was assessed using the light box (Figure 4). Fine grained sections are easily detected using the lightbox, as little light is able to penetrate through tightly packed fine crystals. Comparatively more light is refracted through medium grain sections, and fi- nally, much more light is able to pass through large grained sections. (a) (b) (c)
Figure 5. Examples of the (a) fine (core reference: M2-4), (b) medium (core reference: M2-10), and (c) large grains (core reference: M2-11) found in the ice cores. NB: The measuring tape used for scale in each image is in cm units with mm subintervals.
Fine grains were found throughout the cores, without an apparent order or organisa-
tion. However, the structure was not homogeneous, and a range of small-scale conduits were present within fine-grained sections of crystals. Medium-sized grains are also hetero- geneously distributed throughout the cores, with no apparent order. Unlike fine grains, medium-sized ones have likely undergone some degree of metamorphosis. Refrozen ice conduits were found throughout these sections. Finally, the large-scale grains corresponded to highly metamorphosed sections which contained several large refrozen conduits, with diameters from mm to cm. Overall, this grain size was less widespread throughout the cores. (a) (b) (c)
Water 2021, 13, x FOR PEER REVIEW 6 of 18
Figure 5. Examples of the (a) fine (core reference: M2-4), (b) medium (core reference: M2-10), and (c) large grains (core reference: M2-11) found in the ice cores. NB: The measuring tape used for scale in each image is in cm units with mm subintervals.
Fine grains were found throughout the cores, without an apparent order or organi- sation. However, the structure was not homogeneous, and a range of small-scale conduits were present within fine-grained sections of crystals. Medium-sized grains are also heter- ogeneously distributed throughout the cores, with no apparent order. Unlike fine grains, medium-sized ones have likely undergone some degree of metamorphosis. Refrozen ice conduits were found throughout these sections. Finally, the large-scale grains corre- sponded to highly metamorphosed sections which contained several large refrozen con- Water 2021, 13, 731 duits, with diameters from mm to cm. Overall, this grain size was less widespread6 of 17 throughout the cores.
3.2.2. Contacts between Layers and at the Boundaries of Recrystalised Structures 3.2.2.In Contacts environments between with Layers temporally and at the well-sep Boundariesarated of precipitation Recrystalised ev Structuresents, contacts (or layerIn boundaries) environments in snowpacks with temporally and firn well-separated predominantly precipitation demarcate the events, boundary contacts between (or layerindividual boundaries) deposition in snowpacks events. As and the firn precipitat predominantlyion rate demarcate in the Müller the boundary Ice Shelf betweensector is individualrelatively high deposition at approximately events. As 2.4 the m precipitation a−1 [20] and rate consistent in the Müller throughout Ice Shelf the sectoryear, the is −1 relativelycontacts between high at approximatelylayers are likely 2.4more m repres a [20entative] and consistent of post-depos throughoutitional processes the year, such the contactsas meltwater between generation, layers are percolation likely more and representative refreezing. ofExpressed post-depositional within the processes core, contacts such ascan meltwater be visualised generation, as an ice percolation layer or as and a marked refreezing. change Expressed in grain within size. It the should core, contactsbe noted canthat be within visualised the core, as anmost ice contacts layer or ascontained a marked some change icy component, in grain size. both Itshould related beto notedin situ thatrefreezing, within theas well core, as most the contactsrefreezing contained of liquid some water icy in component, the ice core both post related extraction. to in situIt is refreezing,important to as note well that as the refreezing refreezing post of extraction liquid water is most in the likely ice to core influence post extraction. the relative It ice is importantconcentration to note in a that contact refreezing space, postas opposed extraction to its is mostform. likely to influence the relative ice concentrationOn the basis in a of contact our analysis, space, as we opposed define three to its classes form. of contacts: diffuse; sinuous, and planarOn (Figure the basis 6). of In our general, analysis, the wecontacts define cover three classesthe width of contacts:of the corresponding diffuse; sinuous, ice core. and planarDiffuse (Figure contacts6). are In general, found directly the contacts above cover or below the width sections of thecharacterised corresponding by refreezing ice core. Diffusemorphologies contacts (i.e., are foundlarger directlygrain sizes, above or or ice below layers), sections and characterisedmaintain a non-aligned, by refreezing diffuse mor- phologies (i.e., larger grain sizes, or ice layers), and maintain a non-aligned, diffuse contact contact between sections, thereby suggesting possible percolation of water through the between sections, thereby suggesting possible percolation of water through the contact. contact. Conversely, sinuous contacts have a smoother texture, and are highly continuous Conversely, sinuous contacts have a smoother texture, and are highly continuous across across a core, and are generally sub-horizontal. Lastly, the planar contacts are near-hori- a core, and are generally sub-horizontal. Lastly, the planar contacts are near-horizontal zontal features that are smooth and continuous across a core. features that are smooth and continuous across a core.
(a) (b) (c)
FigureFigure 6. 6.Examples Examples ofof pairedpaired upperupper andand lowerlower contactcontact morphologies.morphologies. ((aa)) Diffuse Diffuse (core (core reference: reference: M2-13), M2-13), ( b(b)) sinuous sinuous (core (core reference:reference: M1-17),M1-17), andand ((cc)) planar planar(core (core reference: reference: M16)M16) contactscontacts foundfound ininthe the ice ice cores. cores.NB: NB: The The measuring measuringtape tape used used for for scale in each image is in cm units with mm subintervals. scale in each image is in cm units with mm subintervals.
3.2.3. Recrystalized Structures In addition to the categorization of contact form, several continuous and discontinuous morphological features were described which relate to post-depositional processes. These structures include: ice lenses; fine-scale ice lenses; fine interlayering of firn and ice; and small-scale conduits. Ice lenses are defined here as a largely impermeable mass of recrystalised water that has been generated by the percolation and refreezing of liquid water along an internal barrier. They are at least 10 mm thick and extend horizontally or near horizontally in a lateral direction. The density of ice lenses is much higher than the surrounding snow or firn, and in many cases air bubbles are trapped during the freezing process. Ice lens contacts are often planar at their base, and may be planar, sinuous or diffuse at the upper contact (Figure7). Fine-scale ice lenses are also prevalent throughout the ice cores and are defined as being continuous ice layers less than 10 mm thick. Thin lenses have fewer visible bubbles Water 2021, 13, x FOR PEER REVIEW 7 of 18
3.2.3. Recrystalized Structures Water 2021, 13, x FOR PEER REVIEW 7 of 18 In addition to the categorization of contact form, several continuous and discontinu- ous morphological features were described which relate to post-depositional processes. These structures include: ice lenses; fine-scale ice lenses; fine interlayering of firn and ice; 3.2.3. Recrystalized Structures and small-scale conduits. In addition to theIce categorization lenses are defined of contact here as form, a largely several im permeablecontinuous mass and discontinu-of recrystalised water that ous morphologicalhas featuresbeen generated were described by the percolation which relate and to refreezing post-depositional of liquid processes.water along an internal These structuresbarrier. include: They ice lenses;are at leastfine-scale 10 mm ice thick lenses; and fine extend interlayering horizontally of firn or andnear ice; horizontally in a and small-scale lateralconduits. direction. The density of ice lenses is much higher than the surrounding snow or Ice lenses arefirn, defined and in here many as cases a largely air bubbles impermeable are trapped mass duringof recrystalised the freezing water process. that Ice lens con- Water 2021, 13, 731 has been generatedtacts byare the often percolation planar at andtheir refreezing base, and ofmay liquid be planar, water alongsinuous an or internal diffuse7 of 17 at the upper barrier. They arecontact at least (Figure 10 mm 7). thick and extend horizontally or near horizontally in a lateral direction. The density of ice lenses is much higher than the surrounding snow or firn, and in many cases air bubbles are trapped during the freezing process. Ice lens con- tactsthan are thick often ones. planar These at thin their layers base, can and present may anybe planar, type of sinuous contact andor diffuse are more at permeablethe upper contactthan thicker (Figure ice 7). lenses (Figure8).
(a) (b) (c) (d) Figure 7. Examples of ice layers found in the ice cores. (a) is ice core section M2-13. In this section an ice lens overlays a fine ice lens (8 mm thick). A thin black line has been added to indicate how contacts between layers were visualised. (b) corresponds to section M1-3, which includes planar contacts at the base, and diffuse upper contacts. (c) shows section M1- (a) 4 which includes an ice lens(b) with a sinuous contact at the base,(c) and a diffuse upper contact. ((dd)) is section M2-5 which has planar upper and lower contacts. NB: The measuring tape used for scale in each image is in cm units with mm subintervals. FigureFigure 7. ExamplesExamples of of ice ice layers layers found found in in the the ice ice cores. cores. (a ()a )is is ice ice core core section section M2-13. M2-13. In In this this section section an an ice ice lens lens overlays overlays a finea fine ice ice lens lens (8 (8mm mm thick). thick). A thin A thin black black line line has has beenFine-scale been added added to ice indicate to lenses indicate how are how contalso contactsacts prevalent between between throughout layers layers were were visualised.the visualised.ice cores (b) and are defined as corresponds to section M1-3, which includes planar contacts at the base, and diffuse upper contacts. (c) shows section M1- (b) corresponds to section M1-3, which includesbeing planar continuous contacts atice the layers base, less and than diffuse 10 uppermm thick. contacts. Thin (c lenses) shows have section fewer visible bubbles 4 which includes an ice lens with a sinuous contact at the base, and a diffuse upper contact. (d) is section M2-5 which has M1-4 which includes an ice lens with a sinuousthan contact thick at theones. base, These and athin diffuse layers upper can contact. present (d any) is sectiontype of M2-5 contact which and has are more permeable planar upper and lower contacts. NB: The measuring tape used for scale in each image is in cm units with mm subintervals. planar upper and lower contacts. NB: The measuringthan thicker tape used ice lenses for scale (Figure in each 8). image is in cm units with mm subintervals. Fine-scale ice lenses are also prevalent throughout the ice cores and are defined as being continuous ice layers less than 10 mm thick. Thin lenses have fewer visible bubbles than thick ones. These thin layers can present any type of contact and are more permeable than thicker ice lenses (Figure 8).
(a) (b)
Figure 8. Examples of fine ice lenses identified in the ice cores. (a) corresponds to section M1-6. The stippled line indicates a fine scale ice lens with planar contacts. (b) is section M1-18 which contains several fine-scale ice lenses that have diffuse contacts. NB: The measuring tape used for scale in each image(a is) in cm units with mm subintervals. (b) Fine interlayering of firn and ice was observed along the length of the cores. Interlay- ering in this context is where thin, horizontal layers of firn and ice are found sandwiched together. The thickness and frequency of firn and ice layers within these interlayered sections were generally consistent (mm-scale), and contacts included all types (Figure9). It should be mentioned that in many examples of interlayering, layers were found to be connected by small conduits. Small scale conduits are found in firn interfaces and form dendritic networks both horizontally and vertically in the direction of the ice shelf base (Figure 10). Horizontal conduits display restricted, slightly bulbous channels that are occasionally accompanied by other smaller, laterally expanding conduits. In contrast, vertical small scale conduits Water 2021, 13, x FOR PEER REVIEW 8 of 18
Figure 8. Examples of fine ice lenses identified in the ice cores. (a) corresponds to section M1-6. The stippled line indicates a fine scale ice lens with planar contacts. (b) is section M1-18 which contains several fine-scale ice lenses that have diffuse contacts. NB: The measuring tape used for scale in each image is in cm units with mm subintervals.
Fine interlayering of firn and ice was observed along the length of the cores. Interlay- ering in this context is where thin, horizontal layers of firn and ice are found sandwiched Water 2021, 13, x FOR PEER REVIEWtogether. The thickness and frequency of firn and ice layers within these interlayered sec-8 of 18 tions were generally consistent (mm-scale), and contacts included all types (Figure 9). It should be mentioned that in many examples of interlayering, layers were found to be connected by small conduits. Water Figure2021, 13 ,8. 731 Examples of fine ice lenses identified in the ice cores. (a) corresponds to section M1-6. The stippled line indicates8 of 17 a fine scale ice lens with planar contacts. (b) is section M1-18 which contains several fine-scale ice lenses that have diffuse contacts. NB: The measuring tape used for scale in each image is in cm units with mm subintervals.
displayFine vertical interlayering to sub-vertical of firn and conduits ice was that observed often terminate along the in length ice lenses. of the As cores. these Interlay- features eringare controlled in this context by the is crystal where structure, thin, horizontal they are layers highly of irregular firn and formsice are that found may sandwiched or may not together.be continuous The thickness across the and diameter frequency of a core. of firn From and observations ice layers within made these in the interlayered lab, many small sec- tionsscale conduitswere generally were discontinuous, consistent (mm-scale), and somewhat and contacts disconnected. included It shouldall types be (Figure remembered 9). It shouldthat these be observationsmentioned that correspond in many toexamples one face of of interlayering, the core, and thatlayers small were scale found conduits to be connectedmay be connected by small in conduits. other planes.
(a) (b) Figure 9. An example of interlayering in section M2-10. Both (a,b) show the same section, however in (b), the fine-scale ice lenses and conduits are outlined in black. NB: The measuring tape used for scale in each image is in cm units with mm subintervals.
Small scale conduits are found in firn interfaces and form dendritic networks both horizontally and vertically in the direction of the ice shelf base (Figure 10). Horizontal conduits display restricted, slightly bulbous channels that are occasionally accompanied by other smaller, laterally expanding conduits. In contrast, vertical small scale conduits display vertical to sub-vertical conduits that often terminate in ice lenses. As these features are( acontrolled) by the crystal structure, they are highly irregular(b) forms that may or may not be continuous across the diameter of a core. From observations made in the lab, many Figure 9. An example ofof interlayeringinterlayeringsmall scale inin conduits section section M2-10. M2-10. were Both discontinuous,Both (a (,ab,)b show) show thean thed same samesomewhat section, section, howeverdisconnected. however in (inb ),( bIt the), shouldthe fine-scale fine-scale be icere- icelenses lenses and and conduits conduits are are outlined outlinedmembered in in black. black. that NB: these NB: The The observations measuring measuring tape correspondtape used used for for to scale scale one in inface each each of image imagethe core, isis inin and cmcm that unitsunits small with scale mm subintervals. subintervals. conduits may be connected in other planes. Small scale conduits are found in firn interfaces and form dendritic networks both horizontally and vertically in the direction of the ice shelf base (Figure 10). Horizontal conduits display restricted, slightly bulbous channels that are occasionally accompanied by other smaller, laterally expanding conduits. In contrast, vertical small scale conduits display vertical to sub-vertical conduits that often terminate in ice lenses. As these features are controlled by the crystal structure, they are highly irregular forms that may or may not be continuous across the diameter of a core. From observations made in the lab, many small scale conduits were discontinuous, and somewhat disconnected. It should be re- membered that these observations correspond to one face of the core, and that small scale conduits may be connected in other planes.
(a) (b)
Figure 10. Horizontal (continuous line) and vertical (dashed line) small scale conduits identified in section M2-3. (a) shows the core section without interpretation, and (b) with interpretation. NB: The measuring tape used for scale in each image is in cm units with mm subintervals.
3.3. Observed Stratigraphy The identified structures can then be analysed at a core scale in order to identify both depositional and post-depositional processes, with the goal of understanding aquifer de- velopment and its impact on ice shelf mechanics. In this section we systematically combine the granulometry results together with the description of contacts, and morphological features present. To aid core interpretation, we also include, a qualitative comparison of relative saturation(a recorded) in the field, and density measurements(b) taken in the laboratory.
3.3.1. Müller 1 Figure 11 presents the Müller 1 stratigraphy results. The core consists predominately of medium sized crystals, but there are also large sections of fine and coarse-grained crystals. Water 2021, 13, 731 9 of 17
Interlayered ice and firn layers are found along the whole length of the core. There are two sections where ice lenses dominate, between 150–300 cm depth and 730–1000 cm depth. Fine-scaled ice lenses are prevalent between 1100–1420 cm depth. Below 1610 cm depth, there was complete saturation of the core when extracted, which corresponds to a large ice lens observed in the laboratory. This section of the core was therefore not included in the core intercomparison. Along the core, an abundance of air bubbles, signalling refrozen water, was also noted in the laboratory which included water refrozen in situ and during transportation. To further identify refrozen ice that was principally formed during transportation, the relative saturation observations taken in the field were compared with the core stratigraphy. In terms of relative saturation, the first 86 cm of the core where dry, before encountering a somewhat humid layer between 86 and 115 cm depth. The rest of the core was dry until the final section of the core. The upper dry section (0–86 cm) matched with a relatively low-density section, which in the field appeared to be relatively fresh snow. However, a thin ice lens was recorded near the surface, which suggests that this section had already undergone a small amount of melt and refreezing post-deposition and pre-extraction. Between 86–115 cm, the sampled snow was damp, and a corresponding rise in density was noted as well as the occurrence of ice lenses, and the start of mixed ice and firn layers (interlayering). Below 115, the core was relatively dry until the saturated layer was encountered at 1610 cm, which suggests little melting at these depths in the days leading Water 2021, 13, x FOR PEER REVIEW 10 of 18 up to extraction, and that the structures encountered were relatively ‘old’ (i.e. correspond to when these depths were closer to the surface).
Figure 11. The stratigraphy recorded from the Müller 1 core. Granulometry Granulometry is is divided divided into into fine fine ( (greengreen),), medium medium ( (orangeorange)) and coarse (red) grain sizes and the ice severely impacted by post extraction refreezing is shown as a hashed box. Recrys- and coarse (red) grain sizes and the ice severely impacted by post extraction refreezing is shown as a hashed box. Recrystal- talized structures are divided into ice lenses (blue), fine-scale ice lenses (purple) and interlaying of firn and ice). Ice con- ized structures are divided into ice lenses (blue), fine-scale ice lenses (purple) and interlaying of firn and ice). Ice conduits duits are not presented as these small-scale features where not continuous across the core. Additionally, relative humidity asare observed not presented in the asfield, these and small-scale density as featuresmeasured where in the not labora continuoustory are acrosspresented. the core. Additionally, relative humidity as observed in the field, and density as measured in the laboratory are presented. 3.3.2. Müller 2 Only a thin (ca. 10 cm) dry upper section corresponding to fresh snow was encoun- tered in the Müller 2 core (Figure 12). Below this section, the core was humid, or very humid until 950 cm depth, and dry from there to the base of the core at 1800 cm, where the firn was saturated with liquid water. Below 110 cm, there was an abundance of interleaved firn and ice layers throughout the core, and several relatively thick ice layers. The density of the core was largely uniform over its length. The stratigraphy of Müller 2 was largely consistent with that seen in Müller 1, however the amount of liquid water observed in the field was much higher.
Water 2021, 13, 731 10 of 17
3.3.2. Müller 2 Only a thin (ca. 10 cm) dry upper section corresponding to fresh snow was encoun- tered in the Müller 2 core (Figure 12). Below this section, the core was humid, or very humid until 950 cm depth, and dry from there to the base of the core at 1800 cm, where the firn was saturated with liquid water. Below 110 cm, there was an abundance of interleaved firn and ice layers throughout the core, and several relatively thick ice layers. The density of the core was largely uniform over its length. The stratigraphy of Müller 2 was largely Water 2021, 13, x FOR PEER REVIEW 11 of 18 consistent with that seen in Müller 1, however the amount of liquid water observed in the field was much higher.
FigureFigure 12. 12. TheThe stratigraphy stratigraphy recorded recorded from from the the Müller Müller 2 2core. core. Granulometry Granulometry is is divided divided into into fine fine (green (green),), medium medium (orange (orange) ) and coarse (red) grain sizes and the ice severely impacted by post extraction refreezing is shown as a hashed box. Recrys- and coarse (red) grain sizes and the ice severely impacted by post extraction refreezing is shown as a hashed box. Recrystal- talized structures are divided into ice lenses (blue), fine-scale ice lenses (purple) and interlaying of firn and ice). Ice con- ized structures are divided into ice lenses (blue), fine-scale ice lenses (purple) and interlaying of firn and ice). Ice conduits duits are not presented as these small-scale features where not continuous across the core. Additionally, relative humidity asare observed not presented in the field, as these and small-scale density as measured features where in the notlabora continuoustory are presented. across the core. Additionally, relative humidity as observed in the field, and density as measured in the laboratory are presented.
3.3.3.3.3.3. Müller Müller 3 3 Fine-grainedFine-grained crystals crystals were were less less prevalent prevalent in in the the Müller Müller 3 3 core core than than the the other other two, two, and and nono fine-grained fine-grained material material was was observed observed below below 130 130 cm cm (Figure (Figure 13). 13). The The extracted extracted core core was was muchmuch shorter shorter than thethe otherother two, two, as as highly highly saturated saturated material material was was found found at a 295at a cm 295 depth, cm depth,limiting limiting further further perforations perforations at this at site. this The site. density The density of the of material the material gradually gradually increased in- creasedwith depth, with fromdepth, approximately from approximately 400 kg m 400−3 near kg m the−3 surface,near the to surface, 760 kg mto− 7603 at kg depth. m−3 at depth.
Water 2021, 13, 731 11 of 17 Water 2021, 13, x FOR PEER REVIEW 12 of 18
FigureFigure 13. 13. TheThe stratigraphy stratigraphy recorded recorded from from the the Müller Müller 3 3 core. core. Granulometry Granulometry is is divided divided into into fine fine ( (greengreen),), medium medium (orange (orange)) andand coarse coarse ( (redred)) grain grain sizes sizes and the ice severely impactedimpacted byby postpost extractionextraction refreezing refreezing is is shown shown as as a a hashed hashed box. box. Recrystal- Recrys- talizedized structures structures are are divided divided into into ice ice lenses lenses (blue (blue), fine-scale), fine-scale ice ice lenses lenses (purple (purple) and) and interlaying interlaying of firn of firn and and ice). ice). Ice conduitsIce con- duitsare not are presented not presented as these as these small-scale small-scale features features where where not not continuous continuous across across the the core. core. Additionally, Additionally, relative relative humidity humidity as asobserved observed in in the the field, field, and and density density as as measured measured in in the the laboratory laboratory are are presented. presented.
3.3.4.3.3.4. Stratigraphic Stratigraphic Correlations Correlations WeWe qualitatively qualitatively correlated correlated the the three three cores cores using using the the results results of of the the measured measured grain grain sizessizes (Figure (Figure 14). 14 ).The The correlation correlation shows shows that that Müller Müller 1 and 1 Müller and Müller 2 have 2 very have similar very similar strat- igraphicstratigraphic profiles, profiles, which which suggests suggests that that they they have have been been exposed exposed to to similar similar processes, processes, whereaswhereas the the Müller Müller 3 3 core core does does not not exhi exhibitbit a a similar similar grain grain size size distribution. distribution. WhileWhile the the granulometry granulometry observations observations were were similar similar between between the the three three cores cores (Figure (Figure 14), 14), thethe distribution distribution of of recrystalized recrystalized structures structures (including (including ice ice lenses, lenses, fine-scale fine-scale ice ice lenses lenses and and interlayeringinterlayering of of ice ice and and firn) firn) was was not not (Figure (Figure 15). 15). This This suggests suggests that that whilst whilst broad broad deposi- deposi- tionaltional events events and and processes processes are are relatively relatively cons consistent,istent, at at least least in in the the central central section section of of the the iceice shelf, shelf, the the post-depositional post-depositional modification modificationss are are site site specific. specific. One One communality communality between between thethe three three cores, cores, however, however, is is that that refreezing refreezing can can be be observed observed along along the the length length of of the the core core sectionssections at at all all three three locations, locations, and and melt-freez melt-freezee processes processes across across the the ice ice shelf shelf are are ubiquitous. ubiquitous.
Water 2021, 13, 731 12 of 17 Water 2021, 13, x FOR PEER REVIEW 13 of 18
Figure 14. Approximate correlation between cores based on grain size. Red indicates coarse grain size, orange medium Figure 14. Approximate correlation between cores based on grain size. Red indicates coarse grain size, orange medium grain size and green indicates fine grain size. Uncoloured areas indicate saturated firn that was too modified post extrac- grain size and green indicates fine grain size. Uncoloured areas indicate saturated firn that was too modified post extraction tion to analyse precisely. Dashed lines indicate interpreted similarities between cores. Indicated elevations were recorded usingto analyse a Garmin precisely. eTrex Dashed 30x handheld lines indicate unit, so interpreted should be similarities treated as betweenindicative cores. only. Indicated elevations were recorded using a Garmin eTrex 30x handheld unit, so should be treated as indicative only.
Water 2021, 13, 731 13 of 17 Water 2021, 13, x FOR PEER REVIEW 14 of 18
Figure 15. Indicative distribution of recrystalized structures in each core (recrystalised features are not to scale). Dark grey Figure 15. Indicative distribution of recrystalized structures in each core (recrystalised features are not to scale). Dark grey indicates ice lenses, purple indicates fine ice lenses and pink indicates interlayered firn and ice sections. indicates ice lenses, purple indicates fine ice lenses and pink indicates interlayered firn and ice sections.
4. Discussion Discussion 4.1. Framework Framework to Describe Firn Stratigraphy The classification classification strategy combined a simplified simplified representation of crystal size, layer contact morphologies morphologies and and different different refrozen refrozen st structuresructures to to respond respond to to two two principal principal moti- mo- vations.tivations. Firstly, Firstly, to to observe observe whether whether there there was was evidence evidence of of water drainage, storage and refreezing within the profiles, and secondly to produce detailed structural information
Water 2021, 13, 731 14 of 17
refreezing within the profiles, and secondly to produce detailed structural information that could be used to inform hydrological modelling, which is often restricted to using simpli- fied measures such as density to model routing. The goal was to provide a sufficient range of categories to cover the diversity of observed structures, without creating unnecessary categories to complicate interpretation and utility of the generated information. In a generalised hydrological system, the flow of water through a porous medium is determined by the porosity, permeability and hydraulic gradient of the matrix [21]. Calculating flow is dependent on understanding the baseline properties of the system, but also the homogeneity and isotropy of both the profile and individual layers. Within a snowpack or firn layer, there are four principal features that modify the way water will move through the matrix: capillary imbibition, conductive pathways, grain structure discontinuities and refrozen structures [22]. Since all of these features are subject to potential change through time due to the phase changes of water, the baseline assessment of effective porosity in a snow/firn/ice matrix is more complex than in a medium such as soil or rock. Additionally, whilst grain size is important for determining bulk flow rates, it is likely that the distribution of structures such as ice lenses (which act to restrict liquid water flow) or conduits (which likely facilitate flow) is more important than grain morphology. This means that the structural considerations (such as grain size and contacts), must also be complemented with observations of relative saturation to determine probable pathways. Therefore, to better inform future modelling efforts of drainage and aquifer devel- opment on ice shelf surfaces, we recommend as a minimum the collection of detailed measurements of grain size, contacts and recrystalized structures, together with bulk val- ues of density and relative saturation. We also recommend the evaluation of how the morphology of contacts between layers enhances or hinders water movement through the profile.
4.2. A Conceptual Firn Aquifer Model Based on Firn Core Analysis The analysis of three firn cores showed that the distribution of grain size was fairly consistent in the central region of the ice shelf (M1 and M2), and that similar recrystalization structures were observed between sites. Even though the exact distribution of recrystalized structures was not the same between cores, the types and quantity of features was similar, which suggest that the snow that arrives to the surface undergoes similar metamorphic processes, however the exact expression is site specific. Throughout the cores, several ice lenses were identified that are likely to inhibit vertical flow through the profile, although the existence of small-scale conduits will enhance water movement. This enhancement of movement is however unlikely to facilitate the movement of large volumes of water. At the central sites, the wetted fringe was observed at a similar depth (17–18 m) and given the similar structures and the confirmation that the water was of meltwater origin, it suggests that the saturated regions come from the ice shelf itself but are unlikely to be input from the overlying snowpack, and more likely to arrive laterally from upstream sources. On this ice shelf, ponds can be periodically observed in satellite data in upglacier regions which may contribute some of the observed water, but this needs to be further investigated. Additionally, whilst we cannot assess the continuity of the aquifer across the ice shelf from the firn cores, the similarities suggest that this is either an extensive aquifer, or the firn and/or energy properties enable the development of aquifers or transfer of water at a similar depth profile. The discovery of very saturated material at <4 m depth (M3) was unexpected. Further work needs to be undertaken to determine whether such shallow water masses are localised near the ice shelf margins or widespread phenomena, and under which conditions they form. These conclusions are similar to those reached by the description of a firn aquifer on Wilkins ice shelf presented by [4], which combined measurements of the density, hydraulic conductivity and specific discharge. Their identified aquifer was found at a 13.4 m depth, which is several meters shallower than the aquifer identified at M1 and M2 on the Müller Water 2021, 13, 731 15 of 17
Ice Shelf. However, they also do not describe the presence of shallow saturated firn, such as that found at M3. Similarly to the field observations on Müller Ice Shelf, [4] did not record core temperatures significantly different to 0 ◦C, which suggests the thermal conditions within each ice shelf was very similar. We recommend that a combination of geophysical and direct observational information (i.e., drilling of firn cores or boreholes) data should be gathered simultaneously. Such strategy could lead to a better high-resolution and higher spatial distribution characterization of possible sub-surface liquid water bodies.
4.3. Future Application of Firn Core Analyses in Hydrological Modelling There are several possible applications of the core descriptions included in this work. One possible use if as input for modelling water flow through, and storage within, frozen porous media. Whilst such models are outside the scope of this study, here we briefly describe the suite of parameters needed to facilitate the use of such models in a complex matrix, such as snow or firn. To model flow through a porous medium, apart from the amount of meltwater that is supplied to the section, one also needs to know: • Porosity • Permeability • Saturated fraction Absolute porosity is defined as the ratio of pore volume to bulk volume and is differ- entiated from effective porosity which considers the interconnectedness of the pores. In snowpacks, porosity is often calculated from snow density [23], and the level of intercon- nectedness is largely ignored. The bulk density of snow combined with grain classification are basic datasets used to characterise snow layers; however density by itself only gives a coarse representation of the snow microstructure [23]. Depending on the final use of the information, we recommend including grain size into the calculation, and also discretizing the profile into largely homogeneous segments, as opposed to integrating the whole profile into a single calculation. Permeability is dependent on both the volume and the connectedness of voids between crystals, and so an understanding of grain size, compaction and the presence of associated structures (such as small-scale conduits) is needed to characterise relative permeability, and eventually classify areas according to the dominance of primary/secondary permeability. From an analysis of the prevalence of different structures, such as ice lenses, interlayering and small-scale conduits, it is possible to indirectly estimate the: • depth of the permeable layer • likelihood of connections between layers • presence of conduits or spaces within layers Finally, the level of saturation within the matrix prior to the addition of ‘new’ water is necessary to model water flow through the medium. Precise water volumes in ice and firn are difficult to measure in the field and can only be approximately determined in the laboratory; however, estimates of relative saturation can be used to inform modelling strategies. We recommend the incorporation of geophysical measurements to determine possible wetting fronts and assess the continuity of saturated layers. This study presents the detailed analysis of firn cores extracted from an Antarctic ice shelf, which provides profiles of granulometry and recrystalised structures that can be used as baseline information in hydrological models. The observed structures support the idea that aquifers are filled laterally as opposed to vertically and are likely formed at similar depths across the ice shelf in central regions. To expand on these conclusions, we recommend the incorporation of geophysical measurements and advanced glacio- hydrological modelling to assess the spatial distribution, overall structure and continuity of aquifers on the ice shelf, as well as hydrological system dynamics. A better understanding of sub-surface dynamics of the ice shelves in regions like the Antarctic Peninsula is critical to assess their future stability under a warming scenario. This Water 2021, 13, 731 16 of 17
has been proven a difficult task presently due to technical and logistical difficulties and represents a key feature to generate more accurate models of future glacier dynamics [24].
Supplementary Materials: The following are available online at https://www.mdpi.com/2073-444 1/13/5/731/s1, Video S1: MacDonell_Muller_saturated_core.MP4. Author Contributions: Conceptualization, S.M., F.F. and A.H.; methodology, S.M. and F.F.; fieldwork, S.M., F.F. and A.H.; investigation, all co-authors; formal analysis, S.M., F.F. and P.V.; writing— original draft preparation, S.M.; writing—review and editing, all co-authors; project administration, S.M.; funding acquisition, S.M. All authors have read and agreed to the published version of the manuscript. Funding: This research was funded by ANID-FONDECYT 1181540 and logistical support and funding was provided by the Instituto Antártico Chileno (INACH). Institutional Review Board Statement: Not applicable. Informed Consent Statement: Not applicable. Data Availability Statement: The data presented in this study are available on request from the corresponding author. Acknowledgments: We greatly acknowledge the field support provided by Tamara Muñoz, and the III Turkish Antarctic Expedition. Conflicts of Interest: The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.
References 1. Iken, A. The Effect of the Subglacial Water-Pressure on the Sliding Velocity of a Glacier in an Idealized Numerical-Model. J. Glaciol. 1981, 27, 407–421. [CrossRef] 2. Reijmer, C.H.; van den Broeke, M.R.; Fettweis, X.; Ettema, J.; Stap, L.B. Refreezing on the Greenland ice sheet: A comparison of parameterizations. Cryosphere 2012, 6, 743–762. [CrossRef] 3. Banwell, A.F.; MacAyeal, D.R.; Sergienko, O.V. Breakup of the Larsen B Ice Shelf triggered by chain reaction drainage of supraglacial lakes. Geophys. Res. Lett. 2013, 40, 5872–5876. [CrossRef] 4. Montgomery, L.; Miège, C.; Miller, J.; Scambos, T.A.; Wallin, B.; Miller, O.; Solomon, D.K.; Forster, R.; Koenig, L. Hydrologic Properties of a Highly Permeable Firn Aquifer in the Wilkins Ice Shelf, Antarctica. Geophys. Res. Lett. 2020, 47.[CrossRef] 5. Banwell, A. Ice-shelf stability questioned. Nature 2017, 544, 306–307. [CrossRef] 6. Bell, R.E.; Chu, W.; Kingslake, J.; Das, I.; Tedesco, M.; Tinto, K.J.; Zappa, C.J.; Frezzotti, M.; Boghosian, A.; Lee, W.S. Antarctic ice shelf potentially stabilized by export of meltwater in surface river. Nature 2017, 544, 344–348. [CrossRef] 7. Scambos, T.; Fricker, H.A.; Liu, C.-C.; Bohlander, J.; Fastook, J.; Sargent, A.; Massom, R.; Wu, A.-M. Ice shelf disintegration by plate bending and hydro-fracture: Satellite observations and model results of the 2008 Wilkins ice shelf break-ups. Earth Planet. Sci. Lett. 2009, 280, 51–60. [CrossRef] 8. MacAyeal, D.R.; Sergienko, O.V. The flexural dynamics of melting ice shelves. Ann. Glaciol. 2017, 54, 1–10. [CrossRef] 9. Banwell, A.F.; Caballero, M.; Arnold, N.S.; Glasser, N.F.; Mac Cathles, L.; MacAyeal, D.R. Supraglacial lakes on the Larsen B ice shelf, Antarctica, and at Paakitsoq, West Greenland: A comparative study. Ann. Glaciol. 2014, 55, 1–8. [CrossRef] 10. Buzzard, S.C.; Feltham, D.L.; Flocco, D. A Mathematical Model of Melt Lake Development on an Ice Shelf. J. Adv. Modeling Earth Syst. 2018, 10, 262–283. [CrossRef] 11. Kingslake, J.; Ely, J.C.; Das, I.; Bell, R.E. Widespread movement of meltwater onto and across Antarctic ice shelves. Nature 2017, 544, 349–352. [CrossRef] 12. Kingslake, J.; Ng, F.; Sole, A. Modelling channelized surface drainage of supraglacial lakes. J. Glaciol. 2017, 61, 185–199. [CrossRef] 13. Forster, R.R.; Box, J.E.; van den Broeke, M.R.; Miège, C.; Burgess, E.W.; van Angelen, J.H.; Lenaerts, J.T.M.; Koenig, L.S.; Paden, J.; Lewis, C.; et al. Extensive liquid meltwater storage in firn within the Greenland ice sheet. Nat. Geosci. 2013, 7, 95–98. [CrossRef] 14. Hubbard, B.; Luckman, A.; Ashmore, D.W.; Bevan, S.; Kulessa, B.; Kuipers Munneke, P.; Philippe, M.; Jansen, D.; Booth, A.; Sevestre, H.; et al. Massive subsurface ice formed by refreezing of ice-shelf melt ponds. Nat. Commun. 2016, 7, 11897. [CrossRef] [PubMed] 15. Siegert, M. Vulnerable Antarctic ice shelves. Nat. Clim. Chang. 2016, 7, 11–12. [CrossRef] 16. Buzzard, S.; Feltham, D.; Flocco, D. Modelling the fate of surface melt on the Larsen C Ice Shelf. Cryosphere 2018, 12, 3565–3575. [CrossRef] Water 2021, 13, 731 17 of 17
17. Lenaerts, J.T.M.; Lhermitte, S.; Drews, R.; Ligtenberg, S.R.M.; Berger, S.; Helm, V.; Smeets, C.J.P.P.; van de Broeke, M.R.; van de Berg, W.J.; van Meijgaard, E.; et al. Meltwater produced by wind–albedo interaction stored in an East Antarctic ice shelf. Nat. Clim. Chang. 2016, 7, 58–62. [CrossRef] 18. Domack, E.W.; Ishman, S.E.; Stein, A.B.; McClennen, C.E.; Jull, A.J.T. Late Holocene Advance of the Muller Ice Shelf, Antarctic Peninsula—Sedimentological, Geochemical and Paleontological Evidence. Antarct. Sci. 1995, 7, 159–170. [CrossRef] 19. Cook, A.J.; Vaughan, D.G. Overview of areal changes of the ice shelves on the Antarctic Peninsula over the past 50 years. Cryosphere 2010, 4, 77–98. [CrossRef] 20. Carrasco, J.E.; Cordero, R.R. Analyzing Precipitation Changes in the Northern Tip of the Antarctic Peninsula during the 1970–2019 Period. Atmosphere 2020, 11, 1270. [CrossRef] 21. Freeze, R.A.; Cherry, J.A. Groundwater; Prentice-Hall: Englewood Cliffs, NJ, USA, 1979; p. 604. 22. Pfeffer, W.T.; Humphrey, N.F. Determination of timing and location of water movement and ice-layer formation by temperature measurements in sub-freezing snow. J. Glaciol. 1996, 42, 292–304. [CrossRef] 23. Fierz, C.; Armstrong, R.L.; Durand, Y.; Etchevers, P.; Greene, E.; McClung, D.M.; Nishimura, K.; Satvawali, P.K.; Sokratov, S.A. The International Classification for Seasonal Snow on the Ground. In IHP-VII Technical Documents in Hydrology 83; UNESCO-IHP: Paris, France, 2009. 24. Siegert, M.; Atkinson, A.; Banwell, A.; Brandon, M.; Convey, P.; Davies, B.; Downie, R.; Edwards, T.; Hubbard, B.; Marshall, G.; et al. The Antarctic Peninsula Under a 1.5 degrees C Global Warming Scenario. Front. Environ. Sci. 2019, 7, 102. [CrossRef]