advances.sciencemag.org/cgi/content/full/4/11/eaar8173/DC1

Supplementary Materials for

A large impact crater beneath Hiawatha Glacier in northwest Greenland

Kurt H. Kjær*, Nicolaj K. Larsen, Tobias Binder, Anders A. Bjørk, Olaf Eisen, Mark A. Fahnestock, Svend Funder,

Adam A. Garde, Henning Haack, Veit Helm, Michael Houmark-Nielsen, Kristian K. Kjeldsen, Shfaqat A. Khan, Horst Machguth, Iain McDonald, Mathieu Morlighem, Jérémie Mouginot, John D. Paden, Tod E. Waight,

Christian Weikusat, Eske Willerslev, Joseph A. MacGregor

*Corresponding author. Email: kurtk@snm.ku.dk

Published 14 November 2018, Sci. Adv. 4, eaar8173 (2018)

DOI: 10.1126/sciadv.aar8173

The PDF file includes:

Supplementary Text Fig. S1. Bedrock type and lineations across Inglefield Land near Hiawatha Glacier. Fig. S2. Terminus history of Hiawatha Glacier and its transition from a floating to a grounded tongue with a proglacial floodplain. Fig. S3. CI-chondrite–normalized metal patterns for glaciofluvial sediment samples compared to upper continental crust. Fig. S4. Model mixtures of crust with mass proportions of various meteorites. Fig. S5. Radar reflectivity at the six deep Greenland ice-core sites, as measured by predecessor radar systems to that used for the Hiawatha Glacier survey. Fig. S6. Relationships between surface and radar layering. Fig. S7. Supraglacial drainage of Hiawatha Glacier. Table S1. Location and description of Hiawatha glaciofluvial sediment samples. Legends for Movies S1 to S3 Legends for Data files S1 and S2 References (41–44)

Other Supplementary Material for this manuscript includes the following: (available at advances.sciencemag.org/cgi/content/full/4/11/eaar8173/DC1)

Movie S1 (.mp4 format). The 2016 AWI airborne radar survey over Hiawatha Glacier. Movie S2 (.mp4 format). Operation IceBridge radar surveys across the Greenland Ice Sheet. Movie S3 (.mp4 format). Operation IceBridge radar surveys toward Hiawatha Glacier. Data file S1 (Microsoft Excel format). EMP data for grains studied from HW21-2016 samples. Data file S2 (Microsoft Excel format). Major element, trace element, and PGE concentrations for subsamples and sub-subsamples of HW21-2016.

Supplementary Materials

Supplementary Text

Additional interpretation of PGE data

Data file S2 provides PGE ratios normalized to CI chondrite for the Hiawatha Glacier glaciofluvial

sediment samples, various mafic and ultramafic mantle melts, chromite-rich rocks from the Platreef

PGE deposit of the Northern Bushveld Complex and meteorites (ordinary and enstatite chondrites and

various types of ion meteorites). The combination of particular PGE ratios (notably (Rh/Pt)N >1.2,

(Rh/Ru)N < 0.3 and (Pd/Pt)N >2.5) observed for many HW21-2016 analyses is anomalous and unlike

most normal terrestrial rocks. Such high (Rh/Pt)N values are only found in depleted mantle, some

komatiite lavas or in rare sulfide-rich chromitite rocks where Rh is concentrated and enriched in the Rh

sulfarsenide mineral hollingworthite (e.g., the Platreef deposit of the Bushveld Complex). Only the

latter source contains sufficiently high PGE concentrations for it to be eroded and diluted to produce

the concentrations observed in HW21-2016 but this source would inevitably produce abundant grains

of detrital chromite, which are not observed in the sample. Furthermore, the unusually high (Pd/Ir)N,

(Rh/Pt)N and (Pd/Pt)N ratios and low (Ru/Rh)N ratios do not conform to any major group of chondrite

meteorite, but such features are observed in some fractionated iron meteorites (Data file S2).

Figure S3 shows Hiawatha Ni, Cr, Co, PGE and Au data are shown normalized to CI chondrite. HW12-

2016 and HW13-2016 have similar patterns to those expected for bulk upper continental crust. All sub-

samples of HW21-2016 have elevated PGE concentrations and display a prominent negative peak at Pt

on the normalized diagrams, reflecting the high (Rh/Pt)N and (Pd/Pt)N ratios noted above.

Model mixtures for Ni, Cr, Co, PGE and Au between a local crustal end member (represented by the

mean of the HW12-2016 and HW13-2016 analyses) and CI chondrite, Platreef chromitite and different

iron meteorites are shown in fig. S4. Chondrite mixtures always produce less fractionated patterns than

observed, and chondrites can be rejected as a source for the siderophile elements in HW21-2016.

Platreef chromitite fails to reproduce the magnitude of the observed Pt anomaly and introduces too

much Cr yet too little Ni and Co to match the observed concentrations in that sample. Mixtures

involving the metals in the bulk Cape York iron meteorite measured by ref. 44 do not reproduce the

fractionated pattern shape or the magnitude of the observed Pt anomaly. However, mixtures involving a

fractionated iron composition, such as the Duchesne IVA iron meteorite, reproduce both the

fractionation and the strong negative Pt anomaly and provide the best fit for the Ni, Co and PGE data.

The Cr concentrations of all iron meteorites are too low to produce the additional 25–30 ppm Cr

observed in the HW21-2016 sample, as compared to HW12-2016 and HW13-2016. This pattern

suggests that an additional Cr-rich component is currently not represented accurately in the model. The

cause could also be a Cr-rich target rock present in the target rocks of the Hiawatha impact crater, but

not in the Inglefield Land mobile belt rocks that were eroded to form the HW12-2016 and HW13-2016

samples. Alternatively, it may reflect a Cr-rich component in addition to iron in the projectile, such as a

pallasite component.

Fig. S1. Bedrock type and lineations across Inglefield Land near Hiawatha Glacier. Immediately

adjacent to Hiawatha Glacier and superimposed on the Palaeoproterozoic tectonic structure of the

foreland basement, are moderately to steeply dipping and commonly brecciated brittle planar

structures, which are tangential to the mostly subglacial rim of the Hiawatha impact crater and dips

away from it. The fabric formed by these structures likely represents the outer, overturned part of the

crater rim. These surfaces contain slickenside striations, which also plunge away from the structure.

Inset photograph shows one of these structures. Photo credit: Kurt H. Kjær.

Fig. S2. Terminus history of Hiawatha Glacier and its transition from a floating to a grounded

tongue with a proglacial floodplain. (A) Oblique panoramic photograph by Lauge Koch showing the

floating glacier tongue in 1922 (41). (B) 1959 aerial photo showing a calving ice tongue. The 1922

photographer’s viewpoint and viewing angle in (A) are estimated based on landmarks visible in the

oblique photo and comparison to the 1959 aerial photo. The blue rectangle indicates the extent of

panels (C–G). The sampling location of HW21-2016 and selected glacier extents are shown on all

panels except (A). The approximate 1922 position of the floating glacier tongue and Lauge Koch’s

estimated grounding line are indicated based on (B) and ref. 41. (C–G) The genesis of the floodplain

between 2010 and our sampling in 2016. (H) Contrast between debris-rich ice, derived from basal

erosion in the crater, and the overlying clean Holocene ice at the front of Hiawatha Glacier. Photo

credit: Svend Funder. (I) Sampling location of HW21-2016. Photo credit: Søren T. Jørgensen.

Fig. S3. CI-chondrite–normalized metal patterns for glaciofluvial sediment samples compared to

upper continental crust (42, 43). CI chondrite normalization factors from ref. 44.

Fig. S4. Model mixtures of crust with mass proportions of various meteorites. Model mixtures of

crust with mass proportions (0.00001–0.01) of: (A) CI chondrite; (B) Platreef chromitites; (C) Cape

York IIIAB iron and (D) Duchesne IVA iron. CI chondrite normalization factors from ref. 44.

Fig. S5. Radar reflectivity at the six deep Greenland ice-core sites, as measured by predecessor

radar systems to that used for the Hiawatha Glacier survey. Radar reflectivity centered on the six

deep Greenland ice-core sites, as measured by predecessor radar systems (180–210 MHz) to that used

for the 2016 AWI survey (150–520 MHz), with the ice cores’ depth–age relationships overlain. The

core sites are ordered from left to right by their proximity to Hiawatha Glacier, with the distance given

in the panel title. Radargram horizontal extent is ±5 km from the closest approach to the ice-core site.

The ice-core depth–age relationships shown are the same as those used by ref. 4. For NEEM, the age of

three prominent LGP reflections are shown. CReSIS radargram frames for each panel are, from A to F:

20110506_01_005, 20110506_01_027, 20110506_01_019, 20140405_01_050, 20140410_01_038 and

20110425_08_012, respectively.

Fig. S6. Relationships between surface and radar layering. (A) Sentinel-2 surface image collected

during the 2016 summer melt season (29 July), identifying the location of panels (C) and (D). Contains

modified Copernicus Sentinel data, processed by ESA. (B) Synthetic reproduction of a previously

published photograph (14) using Sentinel-2 imagery and ArcticDEM, illustrating the compatibility of

satellite interpretations of surface layering and those from ground-based, direct isotopic sampling.

Toward the ice margin, this image shows the surface exposures of thick and thin debris-rich brown

layers associated with stadial ice (pre-Bølling-Allerød and Younger Dryas, respectively), along with

the overlying cleaner Holocene ice (C, D) Interpretation of the age of observed surface layers on

Hiawatha Glacier. (E) overview of locations of panels (F–P). (F–P) Three-dimensional perspective

views of the southwest and northeast corners of Hiawatha Glacier. For all views: 10:1 vertical

exaggeration; 5º viewing angle above horizontal, except for panel N, which is 45º. The same Sentinel-2

image is draped over the GIMP surface-elevation model (39). The vertical cross-sections are from the

2016 AWI airborne radar survey, and the white arrows on surface and in radargram identify the bottom

of the radiostratigraphic unit that we traced and identify as Holocene ice. The bed topography below

the ice is shown in shades of brown. The previously interpreted surface layering (14) and our

interpretation of the radar layering are compatible in two critical ways: 1. The traced bottom of the

reflection-rich unit intersects the surface exposure of the beginning of the Holocene and the end of the

Younger Dryas; 2. The location of radar-observed disturbances in this interface is consistent with the

location of interfingering of the late LGP (>14.7 ka) and Bølling-Allerød surface layers. This disturbed

interface is present in the southwest of the Hiawatha impact crater but not northeast of it.

Fig. S7. Supraglacial drainage of Hiawatha Glacier. Mapped from Sentinel-2 image collected on 31

July 2016, superimposed on hillshaded ArcticDEM. Supraglacial rivers, moulins and major crevasses

that could be identified and mapped unambiguously at 10-m spatial resolution are all shown, but not

several prominent supraglacial lakes upstream.

Table S1. Location and description of Hiawatha glaciofluvial sediment samples.

Sample

ID

Latitude (ºN) Longitude

(ºW)

Sample

depth

(m)

Sediment type &

environment

HW12-

2016

78.84183 67.29250 0.3 Poorly sorted, medium sand;

glacial outwash from

deglaciation

HW13-

2016

78.88696 65.99227 0.3 Well sorted, fine sand and silt;

modern supraglacial outwash

HW21-

2016

78.83305 67.13653 0.3 Well sorted fine sand; modern

outwash (exposed between

2010–2016 CE)

Movie S1. The 2016 AWI airborne radar survey over Hiawatha Glacier. The 26 primary

radargrams that form the 2016 AWI airborne radar survey. The view is initially normal to the

radargrams that followed the approximate regional ice-flow direction, then changes briefly to show two

radargrams oblique to ice flow, then changes by a total of 90º to show the radargrams that were

approximately across-flow. Inset map shows a Landsat-8 image, the relationship between the

radargram being shown (highlighted in green) and the other radargrams (black), the manually identified

elevated rim (magenta triangles on map, vertical dashed line on radargrams) and the central uplift (dark

magenta circles on map, vertical dotted line on radargrams). The map also shows the best-fit circle to

the rim picks (magenta circle on map) and the centre of that circle (magenta X). MPEG-4 video, 1080p

resolution, 30 fps.

Movie S2. Operation IceBridge radar surveys across the Greenland Ice Sheet. NASA Operation

IceBridge MCoRDS radargrams along the Greenland “master grid”, oriented at 22.5º counter-

clockwise from the positive x-axis direction in the NSDIC Sea Ice Polar Stereographic North projection

(EPSG:3413). View is normal to the grid. Surface image is MODIS Mosaic of Greenland, and surface

and bed topography are from ref. 3. Traced and dated radiostratigraphy is from ref. 4. Undated

reflections are shown in white. MPEG-4 video, 1080p resolution, 30 fps.

Movie S3. Operation IceBridge radar surveys toward Hiawatha Glacier. Same format as Movie

S2, but showing radargrams upstream and approximately across-flow of Hiawatha Glacier. Radargram

orientation is more variable, but the viewing angle is fixed at 45º counter-clockwise from the positive

x-axis direction in EPSG:3413. MPEG-4 video, 1080p resolution, 30 fps.

Data file S1. EMP data for grains studied from HW21-2016 samples. Microsoft Excel spreadsheet.

Data file S2. Major element, trace element, and PGE concentrations for subsamples and sub-

subsamples of HW21-2016, along with comparison between those samples previously published PGE

concentrations for other geologic materials. Microsoft Excel spreadsheet.