Workshop on Greenland's diamond potential part 1

G E U S

Contents

Introduction

Old and/or thick lithosphere beneath NW and E Greenland

Bernstein, S. ..................................................................................................................11

Consequences of Phanerozoic episodes of burial and exhumation for erosional
patterns in West Greenland

Bonow, J.M. & Japsen, P. in collaboration with Green, P.F., Chalmers, J.A. & Lidmar-
Bergström, K. 13

Crust and mantle information in Greenland from earthquakes

Dahl-Jensen, T., Darbyshire, F. & Larsen, T.B. ............................................................. 19

Tracking glacially transported kimberlite erratics in Greenland

Funder, S. 23

Patterns of kimberlite emplacement the importance of robust geochronology

Heaman, L.M. ................................................................................................................ 25

Geotectonic settings of diamond-producing cratons with implications for the
diamond potential of southwest Greenland

Helmstaedt, H.H............................................................................................................. 27

Diamondiferous kimberlites from the Garnet Lake area, west Greenland: exploration
methodologies and petrochemistry

Hutchison, M.T. 33

Regional distribution and chemistry of indicator minerals from in situ rocks and
surficial deposits in the Maniitsoq and Sarfartoq regions

Jensen, S.M., Sand, K.K. & Steenfelt, A. ....................................................................... 43

The Diamond Potential of West Greenland: some Global Insights

Kjarsgaard, B.A. 53

Mapping of the lithosphere beneath the Archaean craton and Proterozoic mobile belt
in West Greenland

Larsen, L.M. & Garrit, D. 55

G E U S

Indicator mineral signatures in basal till surrounding the Lahtojoki and Seitaperä
kimberlites, eastern Finland

Lehtonen M., Marmo, J. & Nissinen, A. ......................................................................... 57

Kimberlites and ultramafic lamprophyres in West Greenland: regional constraints 69

Nielsen, T.F.D. & Jebens, M. ......................................................................................... 69

The Majuagaa calcite-kimberlite dyke

Nielsen, T.F.D., Jensen, S.M. & Secher, K. ................................................................... 73

A Greenland petrographic atlas of kimberlites and ultramafic lamprophyres

Nielsen, T.F.D., Secher, K. & Jensen, S.M. ................................................................... 77

Assessment of diamond potential using kimberlitic indicator minerals: key principles
and applications

Nowicki, T.E. & Gurney, J.J. .......................................................................................... 79

The North Atlantic alkaline rocks probes for testing continuity of subcontinental
lithospheric mantle

O'Brien, H., Peltonen P. & Lehtonen, M. ....................................................................... 85

Age, depth and composition of the W. Greenland lithospheric mantle root and
implications for diamond prospecting

Pearson, D.G., Webb, M., Nowell, G.M., Sand, K.K., Luguet, A. & Jensen, S.M........... 87

Mantle stratigraphy of the Karelian craton implications for diamond prospecting 89

Peltonen, P., O'Brien, H., Lehtonen, M. and Brügmann, G............................................ 89

Kimberlites and lamproites in continental reconstructions implications for diamond
prospecting

Pesonen, L.J., O'Brien, H., Piispa, E., Mertanen, S., Peltonen

P. ................................. 91

Magnetic signatures of circular geological features with special emphasis on
kimberlite prospecting

Plado

J. & Pesonen, L.J. ............................................................................................... 93

Application of geophysical methods to diamond exploration in Greenland
experiences and data review

101

Rasmussen, T.M.......................................................................................................... 101

Distribution of kimberlite indicator minerals in till within the Neoproterozoic
Sarfartoq-Maniitsoq province of kimberlite and ultramafic lamprophyres, southern
West Greenland

109

Steenfelt, A., Jensen, S.M. & Sand, K.K...................................................................... 109

G E U S

Introduction

This workshop on Greenland's diamond potential is a milestone in GEUS' five-year project
on `kimberlite' geology. The project has been carried out in collaboration with the Bureau of
Minerals and Petroleum (BMP) in Nuuk, Greenland. During these five years, GEUS' De-
partment of Economic Geology conducted a new collection of data from West Greenland
with relevance to distribution of rock types in the field of kimberlite, lamproite and ultramafic
lamprophyre in the wide context, aiming at all aspects of the Greenland diamond potential.
A number of reports and accounts on results from this work have been published sOFar,
including a complete collection presented on DVD of all diamond exploration data produced
by companies, universities and GEUS.

A wealth of new kimberlite occurrences and data relevant to kimberlites and their diamond
potential have appeared over the last few years, and several research groups are at pres-
ent working in collaboration with GEUS on new samples and data from the Sarfartoq and
Maniitsoq regions. GEUS and BMP wish to convene the international workshop in order to
give involved and interested parties an opportunity to exchange and discuss recent investi-
gations and data and implications for exploration for diamonds in West Greenland and
other Archaean cratonic regions in the northern hemisphere, especially those in glaciated
terrain.

The presentations in this workshop report various results of GEUS' research in this field, as
well as a number of invited accounts covering similar subjects from the North American and
Atlantic with topics relating to the general discussion on kimberlites. The topics cover such
wide issues as the lithospheric mantle under southern West Greenland, petrogenesis of
kimberlite, geotectonic and structural setting, and exploration techniques and results.

This volume includes the abstracts delivered by the participants and contributions are
printed as they are received. Only technical changes have been carried out by the
workshop secretariat in order to match the format of the GEUS report.

The abstracts are

presented in alphabetical order.

Participants presenting material are thanked for their effort and enthusiasm directed
toward this workshop. We look forward to a productive gathering with the Greenland
diamond potential in focus.

G E U S

Old and/or thick lithosphere beneath NW and E
Greenland

Bernstein, S.

Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copen-
hagen K, Denmark

Mantle xenoliths from NW Greenland (Ubekendt Ejland) and E Greenland (Wiedemann
Fjord) have Re-depletion ages> 3 Ga. Coupled with indirect evidence of crust-formation
ages from zircon dating this suggests that large portions of Greenland is underlain by Early
Archaean lithospheric mantle. Igneous activity in the Early Palaeogene in NE Greenland
requires a lithospheric thickness in excess of 200 km, demonstrating that in at least some
parts of E Greenland the lithosphere is thick enough to host diamonds. The age and thick-
ness of the Greenlandic lithospheric mantle thus imply that in SW Greenland a diamond
potential exists in large areas in E and NW Greenland in addition to the known region.

G E U S

Consequences of Phanerozoic episodes of burial
and exhumation for erosional patterns in West
Greenland

Bonow, J.M.

& Japsen, P.

in collaboration with Green, P.F.

, Chalmers, J.A.

& Lidmar-

Bergström, K.

Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copen-

hagen K, Denmark

Geotrack International, 37 Melville Road, West Brunswick, Victoria 3055, Australia

University of Stockholm, S-106 91 Stockholm, Sweden

The preserved MesozoicCenozoic sedimentary and volcanic record of West Greenland
makes this a key area for studying the uplift of passive continental margins. Large-scale
landscapes near the passive continental margins around the northern North Atlantic are
commonly erosion surfaces at high elevation. The landscapes often lack pre-Quaternary
cover rocks and the details in the landscape are characterised by glacial scouring. The
amount of erosion in the basement has been subjected to since the final dated metamor-
phic or intrusion events is therefore difficult to ascertain. We have combined apatite fission-
track analysis (AFTA) data with landform analysis to investigate the development of West
Greenland landscapes across areas with substantially different geology (Chalmers et al.
1999; Green et al. 2002; Bonow 2004, 2005; Japsen et al. 2005; Fig. 1a).

Thermal history constraints have been extracted from the AFTA data across central West

Greenland (6571 N). Data in individual samples typically define two (sometimes three)

discrete paleo-thermal episodes. Synthesis of the timing constraints for individual cooling
episodes identified in 70 samples suggests that at least six discrete episodes of cooling
since the latest Proterozoic are required to explain all the AFTA results. The analysis
shows that the basement rocks presently at outcrop have been buried below only a few

kilometres of cover rocks during the Phanerozoic because palaeo-temperatures of 120 C

or less are found for the palaeo-thermal events in this time interval at various locations
across the area.

Identification and mapping of palaeosurfaces (preserved erosion surfaces) in West Green-
land and analysis of the relationship between re-exposed relief, new relief and the Meso-
zoic-Palaeogene cover rocks, have made it possible to obtain relative age chronologies, to
identify uplift events and to estimate the spatial distribution of erosion (Fig. 1). We find that
the mountains of West Greenland are the end result of three Cenozoic phases of uplift and
erosion. The first phase that began between 36 and 30 Ma created a planation surface
during the OligoceneMiocene. This surface was offset by reactivated faults, resulting in
megablocks that were tilted and uplifted to present-day altitudes of up to 2 km in two

G E U S

phases one that began between 11 and 10 Ma and the second that began between 7 and 2
Ma (Fig. 2).

The amount of erosion since the onset of Neogene uplift has been estimated as the differ-
ence between the altitude of the planation surface and the present topography (Fig. 3). The
area has been affected by differential erosion, even though the landscape is characterised
by glacial scouring. Maximum erosion of 800-1300 m occurs along the fjords and valleys
that cut through the highlands, while large areas within plateaux situated at high altitude are
virtually unaffected. Areas dominated by gneiss in amphibolite facies are more eroded than
areas with gneiss in granulite facies.

The multi-disciplinary approach taken here illustrates that it is possible to indicate when
basement areas have been covered and when they have been exposed to erosion as well
as to estimate the amount of erosion of basement rocks since uplift events.

G E U S

Figure 1. Maps of the study area. Figure description on next page.

(A) Topography. (B) Geology with topographical profiles (Chalmers et al. 1999; Bonow
2004, 2005). Two planation surfaces cut across Precambrian basement and Paleocene
Eocene volcanic rocks. Their formation was uniform across the study area and they must
be younger than the Eocene basalts. The planation surfaces now dip in different directions
and are offset by faults that displace them. Three significant faults and breaklines (changes
in slope gradient) relative to the planation surfaces are (1) the NS Kuugannguaq-Qunnilik
(K-Q) fault on Disko and Nuussuaq, (2) an EW fault just north of Aasiaat (Aa) where or-
thogneisses are separated from supracrustal rocks to the north (Garde et al. 2004); the
fault separates the southwards-dipping planation surface on DiskOFrom the northwards-
dipping surface south of Disko Bugt, and (3) the EW `Sisimiut Line' (SL) that coincides
with the Precambrian Ikertôq thrust zone. A hilly relief has been re-exposed from below

G E U S

CretaceousPaleocene cover rocks on Nuussuaq, Disko and south of Disko Bugt. A-A'
profile for fig. 2. LS: Labrador Sea, BB: Baffin Bay.
(C) Panorama towards west-north-west overlooking Sarfartoq. The upper planation surface,
including wide and shallow valleys, can clearly be identified and appears as a line along the
horizon. In detail the planation surface contains some undulating relief as seen in the fore-
ground. The planation surface and the shallow valleys are distinct from the deeply incised
valleys, here exemplified by the Sarfartoq valley.

Figure 2. Possible development of the relief in West Greenland. (a) 600 Ma, latest Pro-
terozoic intrusion of the Sarfartoq carbonatite complex. (b) 160 Ma, Jurassic maximum
burial before JurassicCretaceous uplift. (c) 120 Ma, Cretaceous hilly relief. (d) 35 Ma, Pa-

G E U S

laeogene maximum burial before EoceneOligocene uplift. (e) 20 Ma, OligoceneMiocene
upper planation surface. (f) Present-day relief after second late Neogene uplift.

Figure 3. Map showing the absolute amount of erosion in basement rock during the last c.
10 Ma, estimated as the difference between present topography and a reconstructed
planation surface. The planation surface was formed close to a former base level (the sea)
prior to differentiated uplift by events in the Neogene (starting c. 10 Ma). The planation
surface has been dissected after uplift, first by fluvial systems and subsequently by glacial
systems. The amount of erosion is highly variable across the area, from almost none in the
SE to more than 1500 m in the high areas cut by fjords. Preservation of the planation sur-
face occurs mainly in areas of a geographically high position and in the areas in front, as
the high area has diverted eroding ice. Erosion is less within areas of gneiss in granulite
facies than in areas of gneiss in amphibolite facies. K: Kangerlussuaq.

References

Bonow, J.M. 2004: Palaeosurfaces and palaeovalleys on North Atlantic previously glaciated

passive margins- reference forms for conclusions on uplift and erosion: PhD thesis
Dissertation 30 . Stockholm University: The Department of Physical Geography and
Quaternary Geology.

Bonow, J.M. 2005: Re-exposed basement landforms in the Disko region, West Greenland

disregarded data for estimation of glacial erosion and uplift modelling. Geomorphology
(in press).

Chalmers, J.A., Pulvertaft, C., Marcussen, C. & Pedersen, A.K. 1999: New insight into the

structure of the Nuussuaq Basin, central West Greenland. Marine and Petroleum Ge-
ology 16 , 197224.

Green, P.F., Duddy, I.R. & Hegarty, K.A. 2002: Quantifying exhumation from apatite fission-

track analysis and vitrinite reflectance data: precision, accuracy and latest results from
the Atlantic margin of NW Europe. In: Doré, A.G., Cartwright, J., Stoker, M.S., Turner,

G E U S

Crust and mantle information in Greenland from
earthquakes

Dahl-Jensen, T.

, Darbyshire, F.

& Larsen, T.B.

Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copen-

hagen K, Denmark

Natural Resources Canada, Earth Science Sector, Geological Survey of Canada, Ottawa,

Ontario, Canada K1A 0E9

The GLATIS project (Greenland
Lithosphere Analysed Teleseismi-
cally on the Ice Sheet) with colla-
borators have operated a total of
16 temporary broadband seismo-
graphs for periods from 3 months
to two years distributed over much
of Greenland from late 1999 to
present. The very first results are
presented in this paper where
receiver function analysis has
been used to map the depth to
Moho in a large region where
crustal thicknesses were previ-
ously completely unknown. The
results suggest that the Protero-
zoic part of central Greenland con-
sists of two distinct blocks with
different depths to Moho. North of
the Archean core in southern
Greenland is a zone of very thick
Proterozoic crust with an average
depth to Moho close to 48 km.
Further to the north the Protero-
zoic crust thins to 37-42 km. We
suggest that the boundary be-
tween thick and thin crust forms
the boundary between the geo-

logically defined Nagssugtoqidian and Rinkian mobile belts, which thus can be viewed as
two blocks, based on the large difference in depth to Moho (over 6 km). Depth to Moho on
the Archean crust is around 40 km. Four of the stations are placed in the interior of Green-
land on the ice sheet, where we find the data quality excellent, but receiver function analy-

G E U S

ses are complicated by strong converted phases generated at the base of ice sheet, which
in some places is more than 3 km thick.

Rayleigh wave phase velocity dispersion curves were estimated for 45 two-station paths
across Greenland, using data from large teleseismic earthquakes. The individual dispersion
curves show characteristics broadly consistent with those of continental shields worldwide,
but with significant differences across the Greenland landmass. Reliable phase velocity
measurements were made over a period range of 25160 s, providing constraint on mantle
structure to a depth of ~300 km. An isotropic tomographic inversion was used to combine
the phase velocity information from the dispersion curves, in order to calculate phase ve-
locity maps for Greenland at several different periods. The greatest lateral variation in
phase velocity is observed at intermediate periods (~5080 s), where a high-velocity
anomaly is resolved beneath central-southwestern Greenland, and a low-velocity anomaly
is resolved beneath southeastern Greenland. The results of the phase velocity inversion
were used to construct localized dispersion curves for node points along two parallel north
south profiles in southern Greenland (see figure). These curves were inverted to obtain
models of shear wave velocity structure as a function of depth, again with the assumption

G E U S

of isotropic structure. A similar inversion was carried out for two twostation dispersion
curves in northern Greenland, where the resolution of the phase velocity maps is relatively
low. The models show a high-velocity `lid' structure overlying a zone of lower velocity, be-
neath which the velocity gradually increases with depth. The `lid' structure is interpreted as
the continental lithosphere.Within the lithosphere, the shearwave velocity is ~412 per cent
above global reference models, with the highest velocities beneath central-southwestern
Greenland. However, the assumption of isotropic structure means that the maximum veloc-
ity perturbation may be overestimated by a few per cent. The lithospheric thickness varies
from ~100 km close to the southeast coast of Greenland to ~180 km beneath central-
southern Greenland.

G E U S

Tracking glacially transported kimberlite erratics in
Greenland

Funder, S.

Geological Museum, University of Copenhagen, Øster Voldgade 5-7, DK-1350 Copenha-
gen K, Denmark

Formerly glaciated mountain areas are generally areas of erosion, not deposition. Recon-
naissance studies in the Angujartorfik area, West Greenland, have indicated that kimberlite
erratics were mainly dispersed during the final phase of deglaciation, and were transported
only over short distances (hundreds of metres). During this phase of deglaciation the thick-
ness of the Greenland ice sheet over the area had decreased, and gradually the continu-
ous ice margin had been transformed into local ice dispersal centres over upland areas.
Radially from these dispersal centres small local glaciers flowed into existing valleys and
fjords. The reconnaissance indicates that the erratics' complex pathways and source can
be approximated by a combination of glacial geological field work, landscape analysis from
air photos and satellite images, and dating of selected erratics by cosmogenic nuclides to
provide a chronology for glacier movements.

G E U S

Patterns of kimberlite emplacement the impor-
tance of robust geochronology

Heaman, L.M.

Department of Earth & Atmospheric Sciences, University of Alberta, Edmonton, Alberta,
Canada T6G 2E3

For the most part, North American kimberlite magmatism spans a period of time in excess
of 1 billion years from 1.1 Ga Mesoproterozoic kimberlites in the Lake Superior and James
Bay Lowlands region of Ontario to Eocene kimberlites in the Lac de Gras field, N.W.T. An
exception to this is the Archean kimberlite-like deposits in the Wawa region. Nearly 25% of
the known kimberlites in North America have robust emplacement ages and several pat-
terns of kimberlite emplacement are emerging: 1) a Mesoproterozoic kimberlite province in
central Ontario, 2) an Eocambrian/Cambrian Labrador Sea Province in northern Québec
and Labrador, 3) an eastern Jurassic Province, 4) a central Cretaceous corridor and 5) a
western mixed domain that includes two Type-3 kimberlite provinces. For some provinces
the origin of kimberlite magmatism can be linked to known mantle heat sources such as
mantle plume hotspots and upwelling asthenosphere attendant with continental rifting. For
example, the timing and location of Mesoproterozoic kimberlites in North America coincides
with and slightly precedes the timing of 1.1 Ga intracontinental rifting that culminated in the
Midcontinent Rift centered in the Lake Superior region. The eastern Jurassic kimberlites
record an age progression where magmatism youngs in a southeast direction from the
~200 Ma Rankin Inlet kimberlites to the 155-126 Ma Timiskaming kimberlites. The location
of several kimberlite fields and clusters in Ontario and Québec lie along a continental ex-
tension of the Great Meteor hotspot track and represents one of the best examples in the
world of kimberlite magmatism triggered by mantle plumes. The central Cretaceous (103-
94 Ma) corridor extends for more than 4000 km from Somerset Island in northern Canada
through the Fort à la Corne field in Saskatchewan to the kimberlites in central U.S.A. The
possible westward younging of Cretaceous to Eocene corridors of kimberlite magmatism
could reflect major changes in plate geometry during subduction of the Kula-Farallon plate.

Several periods of kimberlite/lamprøite emplacement are currently recognized in Green-
land. Based on 24 new U-Pb perovskite and Rb-Sr phlogopite ages and previous geochro-
nology, as many as five discrete events can be discerned; 1) Mesoproterozoic lamprøites
(1284-1227 Ma), 2) three episodes of Eocambrian to Cambrian magmatism at 604-602 Ma
(n=2), 585-577 Ma (n=9), and 568-556 Ma (n=9) and a younger Jurassic event at 164 Ma.
Many of the kimberlites in the Eocambrian/Cambrian Labrador Sea Province in both Labra-
dor/Quebec and Greenland were emplaced soon after the opening of the Iapetus Ocean at
about 615 Ma and may be linked to mantle upwelling associated with continental rifting.

G E U S

Geotectonic settings of diamond-producing cratons
with implications for the diamond potential of
southwest Greenland

Helmstaedt, H.H.

Department of Geological Sciences and Geological Engineering, Queen's University, King-
ston, Ontario, Canada K7L 3N6

Judging from the regional distribution of economic `primary' diamond deposits, xenoliths
studies and geochronological evidence, most diamonds are `old', and their major source
rocks are garnet harzburgites, garnet lherzolites and eclogites located deep in the lithos-
pheric roots of Archean cratons. There, the diamonds remain hidden, unless they are
picked up by `younger' kimberlites, lamproites, or other magmatic rocks originating deep
enough to sample the source rocks and intruding fast enough for the diamonds to survive
transport to the surface or near-surface emplacement site. The evolution of `primary' dia-
mond deposits is thus multi-stage. First is the formation of diamonds within ultrahigh-
pressure ultramafic source rocks, a process that must be viewed as part of the craton as-
sembly and formation of Archean cratonic roots, with the possible addition of diamonds to
these roots during the Proterozoic. Diamonds are known to have formed at various times
and by a number of processes, with individual diamonds preserving evidence for complex
sequences of growth, resorption, brittle deformation and further overgrowth. Second is the
long-term storage in the cratonic roots, during which the diamonds must survive all igneous
and tectonic processes affecting the cratonic roots. Third is the formation of an appropriate
igneous transport medium (kimberlite, lamproite, ultramafic lamprophyre) either within or
below the cratonic roots. In order to yield a melt, the previously depleted source region re-
quires a metasomatic enrichment in incompatible elements. Whereas the processes lead-
ing to the formation of `primary' diamond deposits have operated worldwide, the timing of
the individual diamond-forming events, source area enrichment, and the transport to the
surface appear to be craton specific.

A comparison of Archean cratons worldwide suggests that not all are equally well endowed
with `primary' diamond deposits. This may be a consequence of differences in abundance
and distribution of igneous transport media, differences in the chemical environment during
transport, or differences in diamond contents within the craton roots, either as a conse-
quence of root formation or preservation. It is likely that early root formation was more or
less coeval with the earliest known diamond-forming episode at ~3.4±0.2 Ga which ap-
pears to have been a world-wide metasomatic event triggered by CO

-rich, probably sub-

duction-derived fluids that produced diamonds associated with garnet harzburgite. In a
generally hotter Archean Earth, the low geothermal gradients and pressures necessary for
diamond formation are thought to have been achieved through a build-up of early conti-
nental lithosphere by rapid tectonic stacking of imbricate slabs, whereby successively un-
derplated slabs not only thickened the overlying stack, but also kept it relatively cool

G E U S

(Helmstaedt and Schulze 1989; Gurney, Helmstaedt et al. 2005). Partial melting of hy-
drated oceanic crust in such a model would have yielded successive generations of to-
nalites, leaving behind an eclogitic residue within harzburgites into which carbon would
have been preferentially partitioned (Ireland, Rudnick et al. 1994; Rapp 1995).

Taking into account the growing evidence from deep reflection seismic sections, that Ar-
chean cratons have grown by lateral tectonic accretion (Calvert, Sawyer et al. 1995; Calvert
and Ludden 1999; White, Musacchio et al. 2003), it stands to reason that areas within the
boundaries of the oldest domains of Archean cratons should have enhanced diamond po-
tential over other on-craton and craton margin areas. This agrees with the surface geologi-
cal record showing that the exposed parts of diamond-rich cratons are composite and al-
ways appear to contain tonalitic rocks older than 3 Ga. The label "archon", meaning simply
Archean craton without further qualification (Janse 1994), is thus not a sufficient descrip-
tion, synonymous with high diamond potential, as it would also include Neoarchean (2.7 to
2.5 Ga) juvenile granite-greenstone terrains which appear to have a lower diamond poten-
tial.

As all present Archean cratons are fragments of larger, earlier Archean cratons, craton
margins after fragmentation were initially extensional, trans-tensional or transform. Latest
during the Early Proterozoic, Archean cratons or their fragments became involved in plate
motions and were built into Proterozoic cratons and continents which may have broken up
again and rearranged several times before ending up in their present continental configura-
tion. During accretion, the margins of Archean building blocks became convergent, tran-
spressional or transform. During dispersal of Proterozoic or Phanerozoic continents or su-
percontinents, some Archean building blocks split again, but more often they preserved
their integrity, as break-up occurred along earlier Proterozoic sutures. Each Archean terrain
thus has its own complex post-Archean history, at every stage of which the diamond en-
dowment in its root may have been affected. To remain diamond-prospective, an Archean
craton or subprovince should not have fragmented below a certain minimum size (shortest
dimension about 400 km). Kimberlites and related rocks emplaced prior to such fragmenta-
tion may have had a chance of picking up diamonds from the earlier root, however kimber-
lites following fragmentation and destruction of the earlier root stand only a slim chance of
being diamondiferous. The minimum craton dimensions may not be obvious from the sur-
face geology where craton margins are structurally modified and where Archean regions
are tectonically buried by Proterozoic or later orogenic belts.

Diamond exploration in Greenland sOFar has focused mainly on western Greenland, in
particular on the northwestern margin of the North Atlantic craton (Larsen 1991; Jensen,
Secher et al. 2004), where ca. 600 Ma ultramafic dykes have yielded diamond indicator
minerals as well as numerous small diamonds in the Sarfartoq and Maniitsoq areas (Olsen,
Jensen et al. 1999; Jensen, Secher et al. 2004). Initially referred to as kimberlitic and lam-
prophyric (Larsen, Rex et al. 1983; Larsen and Rex 1992), these dykes were reclassified as
melnoitic (Mitchell, Scott Smith et al. 1999), but a convincing case has now been made,
that at least one of these dykes, the diamondiferous Majuagaa dyke in the Maniitsoq area,
may be classified as a calcite-kimberlite (Nielsen and Jensen 2005).

G E U S

As the oldest Archean building block of northeastern Laurentia, the composite North Atlan-
tic craton (NAC) shows all the surface geological criteria and dimensions (> 150,000 km

required of a diamond-rich craton, and it may be assumed that it developed an early dia-
mondiferous root. Along its northern and western margins, the NAC was affected by the
Paleoproterozoic Nagssugtoqidian and Torngat orogens, respectively. Tectonic fabrics in-
dicate that convergence between the Disko craton and NAC across the broad Nagssugto-
qidian orogen was mainly orthogonal, and tectonic reconstructions suggest that the north-
ern edge of the NAC forms a northward-tapering tectonic wedge between the southerly-
verging southern Nagssugtoqidian front and a postulated southward-dipping subduction
zone (van Gool, Connelly et al. 2002). Deformation in the relatively narrow Torngat orogen,
in the west (across Davis Strait in Labrador), appears to have been dominated by sinistral
transpression and probably had little effect on the NAC exposed in West Greenland. Palep-
roterozoic preservation and modification of an Archean lithospheric root in this part of the
NAC thus depends mainly on the extent of post-Archean rifting (as indicated by the ca.
2040 Ma Kangamiut dykes) and on the angle of subduction and southward extent of mantle
wedge hydration above the subducting slab during collision with the Disko craton. On this
basis, the strength of the Archean mantle-root signature of potential diamond hosts, as
indicated by the presence and composition of harzburgitic diamond indicator minerals (G-
10 garnets, chromite), should increase towards the south, away from the Nagssugtoqidian
front.

Ages of alkaline ultramafic intrusive rocks in West Greenland have been reviewed by Lar-
sen and Rex (1992). Latest Precambrian melnoite and kimberlite magmatism in the Sar-
fartoq and Maniitsoq overlaps in time with ages of numerous alkaline rocks of the North
Atlantic alkaline province (Doig 1970), generally assumed to be related to rifting connected
with the opening of the Proto-Atlantic (Iapetus) ocean. This also includes dykes of diamon-
diferous melnoites in Labrador (Torngat Mountains) (Digonnet, Goulet et al. 2000) and kim-
berlites and kimberlite-like rocks in the northern Quebec part of the Superior Province
(Wemindji, Lac Beaver, Renard) recently referred to as Eocambrian/Cambrian Labrador
Sea kimberlite province (Heaman, Kjarsgaard et al. 2004). It is interesting that the Renard
igneous bodies, in the Otish Mountains of Quebec, although transitional between kimber-
lites and melnoites (Birkett, McCandless et al. 2004), appear to contain economic quantities
of diamonds (Ashton Exploration, various press releases).

Little is known as yet about the age and isotopic compositions of the mantle sample of the
Sarfartoq and Maniitsoq intrusions, and such information is needed to distinguish between
the ages of possible diamond source rocks (harzburgites and eclogites), for assessing re-
gional variations in ages and composition of the mantle sample and for fingerprinting the
metasomatic event(s) preceding or accompanying the intrusive events. As for the overall
diamond potential of the Sarfartoq and Maniitsoq provinces, it is encouraging to note that
indicator mineral tests of the diamondiferous Majuagaa calcite-kimberlite dyke in the
Maniitsoq area, in addition to yielding a fair number of harzburgitic garnets (G-10), also
showed a relatively strong subgroup of Na

O-rich eclogitic garnets (Jensen, Secher et al.

2004). This means that a significant part of the diamond budget of this dyke may be eclogi-
tic, a possibility that could be tested by studying the carbon isotopes of the recovered dia-
monds.

G E U S

The age of a younger kimberlite generation (ca.193-220 Ma) reviewed by Larsen and Rex
(1992) approximately coincides with initial rifting related to the opening of the Atlantic
Ocean. One dyke is located within Archean rocks in the southern part of the NAC in the
Nigerlikasik (Frederikshåb = Paamiut) area, and numerous kimberlite sheets are found
north of Ivittuut, in the slightly reworked Archean part of the Paleoproterozoic Ketilidian belt,
near the southern border of the NAC. The occurrence of microdiamonds in these sheets
suggests that the limited Mesoproterozoic rifting associated with the alkaline magmatism of
the Gardar province (ca. 1350-1150 Ma), in southern Greenland, may not have eliminated
the diamond potential of this area.

References

Birkett, T. C., T. E. McCandless, et al. (2004). "Petrology of the Renard igneous bodies:

host rocks for diamond in the northern Otish Mountains region, Quebec." Lithos 76
475-490.

Calvert, A. J. and J. Ludden (1999). "Archean continental assembly in the southeastern

Superior Province of Canada." Tectonics 18 : 412-429.

Calvert, A. J., E. W. Sawyer, et al. (1995). "Archean subduction inferred from seismic im-

ages of a mantle suture in the Superior Province." Nature 375 : 670-674.

Digonnet, S., N. Goulet, et al. (2000). "Petrology of the Abloviak Aillikite dykes, New Que-

bec: evidence for a Cambrian diamondiferous alkaline province in northeastern North
America." Canadian Journal of Earth Sciences 37 : 517-533.

Doig, R. (1970). "An alkaline rock province linking Europe and North America." Canadian

Journal Earth Sciences 7 : 22-28.

Gurney, J. J., H. H. Helmstaedt, et al. (2005). "Diamonds: Crustal distribution and formation

processes in time and space and an integrated deposit model." Society of Economic
Geologists 100th Anniversary Volume : in press.

Heaman, L., B. A. Kjarsgaard, et al. (2004). "The temporal evolution of North American

kimberlites." Lithos 76 : 377-397.

Helmstaedt, H. H. and D. J. Schulze (1989). Southern African kimberlites and their mantle

sample: implications for Archean tectonics and lithosphere evolution. Kimberlites And
Related Rocks, Vol. 1. J. Ross. Perth, Geological Society of Australia, Special Publi-
cation 14 358-368.

Ireland, T. R., R. L. Rudnick, et al. (1994). "Trace elements in diamond inclusions from ec-

logites reveal link to Archean granites." Earth and Planetary Science Letters 128 : 199-
213.

Janse, A. A. (1994). Is Clifford's Rule still valid? Affirmative examples from around the

world. Fifth International Kimberlite Conference. H. O. A. Meyer and O. H. Leonardos.
Araxa, Brazil, Companhia de Pesquisa de Recursos Minerais - CPRM. 2, Diamonds:
Characterization, Genesis and Exploration: 215-235.

Jensen, S. M., K. Secher, et al. (2004). Diamond exploration data from West Greenland:

2004 update and revision, Danmarks og Grønlands Geologiske Undersogelse Rapport
2004/117: 90 pp.

Larsen, L. M. (1991). Occurrences of kimberlite, lamproite and ultramafic lamprophyre in

Greenland, Open File Series Grønlands Geologiske Undersogelse 91/2: 36 pp.

Larsen, L. M. and D. C. Rex (1992). "A review of the 2500 Ma span of alkaline-ultramafic,

potassic and carbonatitic magmatism in West Greenland." Lithos 28 : 367-402.

G E U S

Larsen, L. M., D. C. Rex, et al. (1983). "The age of carbonatites, kimberlites and lampro-

phyres from southern West Greenland: recurrent alkaline magmatism during 2500 mil-
lion years." Lithos 16 : 215-221.

Mitchell, R. H., B. H. Scott Smith, et al. (1999). Mineralogy of ultramafic dikes from the

Sarfartoq, Sisimiut and Manitsoq areas, West Greenland. Proceedings of the VIIth In-
ternational Kimberlite Conference. J. J. Gurney, J. L. Gurney, M. D. Pascoe and S. H.
Richardson. Cape Town. 2, P.H. Nixon Volume: 574-583.

Nielsen, T. F. D. and S. M. Jensen (2005). The Majuagaa calcite-kimberlite dyke, Maniit-

soq, southern West Greenland, Danmarks og Grønlands Geologiske Undersogelse
Rapport 2005/43: 59 pp.

Olsen, H. K., S. M. Jensen, et al. (1999). "Review of diamond exploration in Greenland."

North Atlantic Minerals Symposium, Dublin, Ireland, 19-22 September, 1999 Ex-
tended Abstracts : 166-168.

Rapp, R. P. (1995). "Is eclogite in the subcontinental lithosphere the residue from melting of

subducted oceanic crust? Experimental constraints and implications for the origin of
the Archean continents." 457-459.

van Gool, J. A. M., J. N. Connelly, et al. (2002). "The Nagssugtoqidian orogen of West

Greenland: tectonic evolution and regional correlations from a West Greenland per-
spective." Canadian Journal of Earth Sciences 39 : 665-686.

White, D. J., G. Musacchio, et al. (2003). "Images of a lower-crustal slab: Direct evidence

for tectonic accretion in the Archean western Superior province." Geology 31 : 997-
1000.

G E U S

Diamondiferous kimberlites from the Garnet Lake
area, west Greenland: exploration methodologies
and petrochemistry

Hutchison, M.T.

Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copen-
hagen K, Denmark

1. Introduction

Amongst the kimberlite-affinity rocks of West Greenland, of particular interest from the point
of view of diamond prospectivity is the Garnet Lake locality that lies approximately 2 km to
the north of the Sukkertoppen Ice Cap, Sarfartoq, West Greenland (Fig. 1). Reported here
are the recovery of the largest diamond sOFar found in Greenland (1.90 x 1.70 x 1.42 mm)
in addition to the largest calculated figures for metric carats of diamond per 100 tons
(ct/100ton).

Figure 1. Map of locations of Garnet Lake and Spider Lake (yellow stars) with locations of
in-situ kimberlite shown by green triangles after Jensen et al. (2004a).

Diamond recovery data and geo-
chemistry of mineral phases are pre-
sented, arising from samples of drill
core and associated float recovered
as part of the 2004 and 2005 explora-
tion program of Hudson Resources,
Inc. Comparison is made with miner-
alogy of diamond and non-diamond
bearing rocks recovered from nearby
localities principally within the same
program with a view to clarify indica-
tions of diamond prospectivity.

G E U S

The Garnet Lake site is centred around WGS84 UTM22N grid reference (469922,
7360319). Also discussed are results from the Spider Lake site (479467, 735374) and as-
sociated Spider Hollow (479087, 7358613) approx. 10 km to the east and the Silly Kimber-
lite site (470219, 7360929) approx. 700 m to the north-east of Garnet Lake.

2. Geophysical exploration

The Garnet Lake site was discovered during 2004 ground reconnaissance following up on
publicly available reports on indicator mineralogy (references in Jensen et al., 2004a) and
an airborne DIGHEM resistivity / magnetic survey conducted at 100m line spacing for Hud-
son Resources Inc. by Fugro Airborne Surveys. The airborne magnetic survey in particular
yielded a number of positive and negative semi-spherical and possible dipole anomalies
which exhibited similarities with kimberlite pipes from elsewhere (e.g. Lockhart et al., 2004).
Furthermore a number of strong linear basement features were seen to intersect each
other within the Garnet Lake area.

Figure 2. Ground based geophysical survey of the Garnet Lake area total field. UTM
coordinate system is based on WGS84 Zone 22N. Drill site locations are indicated by
black/yellow circles. Note that lake shapes and locations are approximate and lie in reality
~100 m W of the locations shown.

Further to the successful recovery of diamonds from the Garnet Lake site, reported below,
a 50 m line spacing ground-based magnetic survey was conducted around the Garnet Lake
and Silly Kimberlite sites in order to direct drilling operations during 2005. Results are pre-
sented as total field data in Fig. 2. Within the field of view, the most prolific kimberlite-
bearing sites lie at the SE corners of Garnet Lake and the Silly Kimberlite Lake at the east-
ern extent of a strong NE-SW trending magnetic low, interpreted as a basement feature.
Furthermore, small linear features are seen to extend from these locations to the SE and it
is believed that in-situ kimberlite which was subsequently drilled at these sites may have
been emplaced preferentially along intersections of basement weaknesses indicated by the
geomagnetic trends. This observation is commonly made for kimberlite fields worldwide

G E U S

(e.g. Stubley, 2004). Amongst other intersections elsewhere, six drill holes on Garnet Lake,
three at Silly Kimberlite and six at Spider Lake were successful in intersecting kimberlite
bodies, all thought to be sills, with the principal intersection at Garnet Lake having a 3.9 m
interpreted uninterrupted true thickness.

3. Diamond recovery

Three samples of drill core and three larger samples of float taken from Garnet Lake were
crushed and processed by caustic fusion for diamond separation at the SRC Geoanalytical
Labs., Saskatoon, Canada. Results are presented in Table 1. The more voluminous float
samples were found to yield the largest diamonds with the three most significant being 1.90
x 1.70 x 1.42; 1.98 x 1.34 x 0.98 and 1.56 x 1.40 x 1.16 mm. Figure 3 shows a slightly
smaller colourless octahedral stone from float sample MHG9-7. Diamonds have been re-
covered in total from 36 samples from West Greenland (references in Jensen et al.,
2004a,b) with the largest previously reported being a single 1.62 x 1.53 x 0.22 mm stone
from Pyramidefjeld (Geisler, 1974).

Figure 3. Photograph of diamond recovered from Garnet Lake float sample MHG9-7.

Calculations of ct/100t are presented as a means of rough comparison between samples
(Table 1). Due to the small sample size however, figures should not be considered to be
suitable for comparison with those quoted for producing mines. It is notable however that
core 05DS12-D yields recovery figures most comparable with float. This core was taken
from shallow depth within a few metres of the float sampling site whereas the other drill
cores were from 35-100 m north. Values are much higher than those for diamondiferous
samples previously recorded from Greenland with the exception of overlap with a 187 kg
sample from east of Sukkertoppen Ice Cap (Bizzarro and Plouffe, 1999) which yielded 55
ct/100ton. Diamond recovery values for other sites from the current program were also
comparably smaller. Aside from a number of diamond-free samples, three micro-diamonds
were recovered from 41.05 kg Spider Lake core, two microdiamonds in 57.8 kg of float
were recovered from the Silly Kimberlite and two in 64.45 kg from another site nearby.
Garnet Lake diamond recovery in comparison with other diamondiferous rocks from West
Greenland, therefore suggests that this is a site from which a useful diamond prospectivity
methodology may be constructed.

G E U S

Table 1.

Weights and numbers of diamonds recovered from Garnet Lake float and drill

core.

Sample

Sample wt.

Diamond wt.

ct/100t

05DS08-D

14.4

0.269

9.3

05DS10-D

14.15

0.098

3.5

05DS12-D

10.95

0.662

30.2

MHG9-5

29.65

6.654

112.2

MHG9-7

21.2

4.269

100.7

MHG9-13

57.05

9.696

85.0

TOTALS

147.4

178

21.648

73.4

Samples prefixed by 05DS are drill core samples; samples prefixed by MHG are float sam-
ples; Sample weight in kg; diamond weight in mg; # :- number of diamonds. Note that
ct/100t values are not statistically robust due to the small sample sizes involved.

3. Mineralogy

Compositions of kimberlite and xenolith phases have been measured using standard
EPMA techniques (Univ. Copenhagen JEOL 733) from samples of heavy mineral separates
recovered from crushed float and core and from polished thin sections taken from both float
and core. Mineral separation was conducted at the SRC Geoanalytical Labs., Saskatoon,
Canada.

3.1 Olivine

Olivines have been analysed in abundance from both mineral separates and groundmass
and xenolith-hosted grains from thin sections. Strong trends in Ni at constant Fo content
are apparent for Garnet Lake samples particularly at Fo content of 0.86, 0.90 and 0.92.
Although there is a dominance of analyses within the proposed diamond field of Fo> 0.90
and Ni> 2250 ppm (Jago, 2004), no strong differences are observed between Garnet Lake
olivines and olivines from nearby less diamondiferous localities.

3.2 Ilmenite

Ilmenites from Garnet Lake samples are almost exclusively picro-ilmenites although with
variable Cr

-content up to 6.78 wt%. The occurrence of ilmenites with MnO greater than

1 wt% is highly variable with some samples having no such grains and one thin section
having only Mn-rich ilmenites. Variability in Mg, Cr and Mn is not strikingly different be-
tween Garnet Lake and other samples in this study with perhaps the exception of Silly Kim-
berlite samples which are typically more Cr

-rich (up to 16.5 wt%) and yield no Mn-rich

examples.

3.3 Garnet

Following the classification scheme of Grütter et al. (2004), Garnet Lake samples are rich in
harzburgitic G10D and particularly eclogitic G3D and G4D diamonds in comparison with
samples from other localities. G10D and G3D-G4D garnets comprise 10% and 11% re-
spectively in comparison with for example Spider Lake with 11% (the most comparable
G10D occurrence) and no eclogitic garnets. Garnet Lake garnet compositions in terms of a

G E U S

/CaO discriminatory diagram (Fig. 4) demonstrate the proliferation of G10D and ec-

logitic garnets in core and float mineral separates. Thin sections of float and core show a
similar spread of data and also include a single G12 wehrlitic garnet.

The Na

O content of Garnet Lake garnets is unusually high (averaging 0.19 up to 0.518

wt% and up to 0.28 wt% for G4D garnets in an eclogitic garnet-bearing xenolith from Gar-
net Lake drill core). The only Greenlandic samples otherwise reported with the range of Na
approaching this trend is the Majuagaa kimberlite, Maniitsoq (Nielsen and Jensen, 2005),
however their highest Na

O content is reported as 0.18 wt%.

CaO (wt%)

(wt%)

G10D

G10 (D)

G12

G4 (D)

G3 (D)

G1 (projected)

only non-
diamondiferous G10

G11

unclassified garnet (G0) -
MgNum too low

05DS11 21 feet core

MHG09-5
Hand Sample

05DS12 26 feet core

Figure 4. Compositional variation of garnets from Garnet Lake samples expressed as
Cr

versus CaO (wt%). Yellow diamonds:- 05DS07-162 foot thin section G4D eclogitic

garnets.

3.4 Spinel

A variety of spinel compositions have been recovered from Garnet Lake, involving Ul-
vöspinel, magnetite and magnesioferrite components to varying degrees. Notably however
few chromites have been recovered, unlike at Spider Lake and Spider Hollow where chro-
mites are common. Compositions of chromites from Garnet Lake and other localities from
this study are presented as a projection onto the reduced spinel prism (Fig. 5). Individual
grains are grouped separately according to sample with different legend sizes correspond-
ing to different samples. Where trends in composition are apparent these are annotated on
the diagram. There is a significant spread in the data, partly due to the involvement of
magnetite and Mg-rich spinel, however it is apparent that examples of both T1 and T2
trends of Mitchell (1995) plus a mixed trend for Garnet Lake sample T1 occur (similar to
that described in Mitchell et al., 1999). Mitchell (1995) describes the T1 trend as being kim-
berlitic and the T2 trend as being orangeitic.

G E U S

0.0

0.1

0.2

0.3

0.4

0.5

0.6

0.7

0.8

0.9

1.0

0.4

0.5

0.6

0.7

0.8

0.9

1.0

Fe2T / (Fe2T + Mg) cations

i/(T

i+Al+C

r

cati

magnetite - periclase in vein in xenolith in MHG9-5

magnetite - periclase grains in
kimberlite and xenolith in MHG9-6b

Spider Hollow
magnetites

Garnet Lake T1?

Silly Kimberlite T1

Spider Lake T1
04CL01b

Spider Hollow
T2 MHG5-20

Spider Lake T1?
MHG5-10

Figure 5. Compositional variation of spinels projected onto the front face of the reduced
spinel prism: expressed as Ti/(Ti+Al+Cr) cations versus (Fe2T i.e. total Fe calculated as
Fe

)/(Fe2T+Mg) cations. Red circles are Garnet Lake samples, red triangles are from the

Silly Kimberlite, yellow diamonds are from Spider Hollow and flesh-coloured triangles are
from Spider Lake samples.

3.5 Mica

Individual samples from Garnet Lake, as from the other localities studied show a range in
mica compositions. Of particular use for classification are the variations in Al, Ti and Fe.
Mica compositions in terms of Al and Ti wt% oxide are presented in Figure 6. Individual
grains are typically significantly homogeneous with the exception of rims of tetra-
ferriphlogopite. The Garnet Lake samples distinguish themselves in being particularly Ti-
rich (also compared to Greenlandic micas published elsewhere, e.g. Nielsen and Jensen,
2005). Their trend towards tetra-ferriphlogopite can be considered to be orangeitic
(Mitchell, 1995). It is notable that Spider Lake and Spider Hollow compositions have a
similarity with those from the Majuagaa calcite-kimberlite despite Spider Hollow in particular
having otherwise an orangeitic character (e.g. in terms of spinel composition).

TiO2 (wt%)

Tetra-ferri Phlogopite

Biotite

Garnet Lake

Silly Kimberlite

Spider
Hollow

Spider
Lake

KIMBERLITE

ORANGEITE

Figure 6. Compositional variation of phlogopites from Garnet Lake and Spider Lake sam-
ples expressed as Al

versus TiO

(wt%). Tetra-ferriphlogopite rims on micas from three

Garnet Lake float samples are shown as small red circles.

4. Geothermobarometry

Calculations of equilibration pressure and temperature of a Garnet Lake sample was un-
dertaken using data from phases in a garnet lherzholite xenolith taken from float sample
MHG9-6 (Figure 7). A four phase assemblage calculation is considered preferable to cal-

G E U S

culations based on fewer phases and the commonly used technique of using mineral com-
positions from mineral separates. The latter method allows no confidence of mineral equili-
bration and can result in misleading conclusions.

Two calculations were carried out using the Al in Opx barometer of Brey and Köhler (1990)
and the following thermometers:
1. Ellis and Green (1979) Fe-Mg exchange in garnet-cpx
2. Brey and Köhler (1990) Na in Opx-Cpx
3. Brey and Köhler (1990) Cpx-Opx solvus

The first calculation used averaged analyses of touching grains considered most likely to
be in chemical equilibrium. The second calculation used averaged analyses for grains of
opx, cpx, olivine and garnet taken from within the unaltered centre of the xenolith.

Equilibrium conditions are calculated to lie within the range P=61.4 to 66.7 kbar and
T=1352°C to 1327°C. Comparison with the fields of Greenland till samples calculated after
data in Jensen et al. (2004a) which cuts off at ~1270°C, indicates that the depth of origin of
the xenolith from the diamond-bearing Garnet Lake locality is greater than generally ob-
served previously. Admittedly equilibration within till sample grains, as described previ-
ously, may not be assured. Data from Garnet Lake is also consistent with a similar cold
geotherm, as in for example the Kaapvaal craton (references in Nixon, 1987).

Figure 7. Transmitted light photomicrograph of garnet lherzolite xenolith sample
tsMHG06b. Field of view is approximately 2 cm. Green macrocrysts are cpx, colourless are
opx, orange are garnet. Finer grains are almost exclusively olivine.

5. Discussion and conclusions

Similarities in terms of diamond recovery and mineral compositions between Garnet Lake
core and float, in and in contrast with data from nearby samples (e.g. Fig. 4) strongly sug-
gests that the Garnet Lake float can be considered to be close to in-situ. It is reasonable

G E U S

therefore to accommodate data from float into discussion of the significance of mineralogy
of Garnet Lake samples in general.

h11

h12

t7C

t7F

h9-5

t9-2

t9-3

t9-5

t9-6a

t9-6b

Olivine

Macrocrysts

Phenocrysts

Mica
Groundmass

Macrocrysts

Spinel

(ulv,

T1?)

(ulv,

T1?)

Perovskite

Apatite

- ?c

K ?c

Carbonate

K-richterite

Mn-ilmenite

REEphos-
phates

Barite

Ni sulphide

lowCr,Ti
macro.

K (G1)

K or O affinity

O/K

Table 2. Orangeite and Kimberlite Affinities of Najaat samples Garnet Lake core and
associated float h11 and h12 :- heavy mineral separates from cores 05DS11-21 and
05DS12-26 respectively; t7F :- thin section 05DS07-155b (apahanitic); t7C :- thin sections
05DS07-155-a and c (macrocrystal); h9-5 :- heavy mineral separate from MHG9-5; t9-2.
t9-3. t9-5, t9-6a and t9-6b :- thin sections MHG9-2, MHG9-3, MHG9-5, MHG9-6a and
MHG9-6b respectively. Orangeite and Kimberlite characteristics after Mitchell (1995); K :-
kimberlite; O :- Orangeite; ülv :- ulvöspinel; macro: :- macrocrysts; * :- contains olekmin-
skite; ?c :- composition unknown; ?p :- proportion unknown; not abundant however
absence can't be stated with confidence; - :- phase present but observed characteristics do

not allow for distinction between rock types; T1 and T2 :- spinel magmatic trends T1 and T2
respectively.

Garnet Lake samples distinguish themselves from neighbouring diamond-poor kimberlitic
rocks by the following characteristics:

implied high diamond grade;

commonly visible garnet megacrysts in the matrix;

dominance of diamond-stable peridotitic garnets;

abundance of diamond-stable Na-eclogitic garnets;

deep solution to geothermobarometry calculations

The closest analogy reported elsewhere may be the Majuagaa calcite-kimberlite (Nielsen
and Jensen, 2005). Its diamond content of 125 microdiamonds in 1060 kg (Jensen et al.,
2004b) is higher than most tested Greenlandic kimberlites but is still significantly lower than

G E U S

Garnet Lake samples. Majuagaa does contain significant eclogitic garnets and concentra-
tions of Na in garnet overlap the lower end of the Garnet Lake range.

In terms of classification, there has been some significant debate as to how to describe the
Greenlandic kimberlitic rocks (e.g. Mitchell et al., 1999 and Nielsen and Jensen, 2005).
Table 2 summarises the mineral compositional data described above in association with
additional observations in the context of Mitchell's (1995) orangeite/kimberlite classification
scheme. It is apparent that the mineralogical and geochemical characteristics of most of the
samples studied would lead to a classification of either kimberlite or orangeite depending
on which is considered to be the most important criteria. Garnet Lake samples often con-
tain low Cr, Ti-macrocrysts, occasional Ni-sulphide and Mn-ilmenites are rare. These are all
characteristics of kimberlite. On the other hand tetra-ferriphlogopite rims on phlogopite are
seen in many samples, olekminskite (Sr,Ba,Ca carbonate) is reported and olivine pheno-
crysts are typically Fo-rich. These are all characteristics of orangeite. Such conflicting char-
acteristics are not confined to mineral separates but are also seen in thin section. Similar
mixed characteristics are seen in Spider Lake, Spider Hollow and Silly Kimberlite samples.
Indeed a single core section from Garnet Lake (05DS07-262) consists of a coarse grained
rock with strong orangeitic affinity abutting a fine grained perovskite-rich rock of strong kim-
berlitic affinity.

As the orangeite classification relies heavily on southern African samples, it is perhaps not
surprising that the terminology has questionable direct application to Greenlandic rocks.
However the natural question which arises is whether or not Garnet Lake, and other kim-
berlitic rocks from Greenland represent a mixing of true primary orangeite with kimberlite in
the source region, or whether some genetic spectrum of kimberlite-orangeite primary
magma is possible. Significantly more work is required tOFully address this issue. It is at
least safe to say at this stage however, that notwithstanding uncertainty on classification,
Garnet Lake samples demonstrate that Greenlandic kimberlitic rocks can be substantially
diamond-bearing.

6. Acknowledgements

Hudson Resources Inc., Canada and James Tuer are gratefully acknowledged for access
to mineral claims and supply of samples. Grant Lockhart is thanked for provision of proc-
essed geomagnetic data presented in Fig. 2. Research was supported by Trigon GeoServi-
ces Ltd., U.S.A. and the European Community's 6

Framework Program, Marie Curie EIF

Fellowship.

Disclaimer: Non-Greenlandic place names are informal. This publication reflects the
author's views and the European Community shall not be held liable for any use of the in-
formation contained herein.

References

Bizzarro, M. and Plouffe, D. (1999) Assessment report for field work on the Sisimiut (EL

13/96) and Sarfartoq (EL 12/96) exploration licenses in relation to the Dunbar Joint
Venture, south western Kalaallit Nunaat (Greenland), 29 pp. Archives of the Geologi-
cal Survey of Denmark and Greenland; GEUS Report File 21685.

G E U S

Brey, G.P. and Köhler, T. (1990) Geothermobarometry in four-phase lherzolites II. New

thermobarometers, and practical assessment of existing thermobarometers J. Petrol.
31 , 1353-1378.

Ellis, D. and Green D. (1979) An experimental study of the effect of Ca upon garnet-

clinopyroxene Fe-Mg exchange equilibria Contrib. Mineral. Petrol. 71 , 13-22.

Geisler, R.A. (1974) Investigations on the Renzy Mines Limited Frederikshåb concession,

Greenland to June 15, 1972 Archives of the Geological Survey of Denmark and
Greenland; GEUS Report File 20054 , 6 pp.

Grütter, H.S., Gurney, J.J., Menzies, A.H. and Winter, F. (2004) An updated classification

scheme for mantle-derived garnet, for use by diamond explorers Lithos 77 , 841-857.

Jago, B.C. (2004) Diamond indicator mineral evaluation of three dyke rocks from Western

Greenland Archives of the Geological Survey of Denmark and Greenland report
number pending , 20 pp. 4 appendices.

Jensen, S.M., Secher, K., Rasmussen, T.M. and Schjøth, F. (2004a) Diamond Exploration

data from West Greenland: 2004 update and revision Danmarks of Grønlands Geolo-
giske Undersøgelse Rapport 2004/117 , 90 pp. + DVD-ROM.

Jensen, S.M., Secher, K. and Rasmussen, T.M. (2004b) Diamond content of three kimber-

litic occurrences in southern West Greenland Danmarks of Grønlands Geologiske Un-
dersøgelse Rapport 2004/119 , 41 pp.

Lockhart, G., Grütter, H. and Carlson, J. (2004) Temporal, geomagnetic and related attrib-

utes of kimberlite magmatism at Ekati, Northwest Territories, Canada Lithos 77 , 665-
682.

Mitchell, R.H. (1995) Kimberlites, Orangeites and Related Rocks Plenum, New York.

Mitchell, R.H., Scott Smith, B.H. and Larsen, L.M. (1999) Mineralogy of Ultramafic Dikes

from the Sarfartoq, Sisimiut and Maniitsoq Areas, West Greenland Proceedings of the
VIIth International Kimberlite Conference, Volume I, Gurney, J.J. et al., eds. (Red Roof
Design, Cape Town), 574-583.

Nielsen, T.F.D. and Jensen, S.M. (2005) The Majuagaa calcite-kimberlite dyke, Maniitsoq,

southern West Greenland Danmarks og Grønlands Geologiske Undersøgelse Rapport
2005/43 , 59 pp.

Nixon, P.H. (1987) Mantle Xenoliths Chichester, Wiley.

Stubley, M.P. (2004) Spatial distribution of kimberlite in the Slave craton, Canada: a geo-

metrical approach Lithos 77, 683-693.

G E U S

Regional distribution and chemistry of indicator
minerals from in situ rocks and surficial deposits in
the Maniitsoq and Sarfartoq regions

Jensen, S.M., Sand, K.K. & Steenfelt, A.

Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copenhagen
K, Denmark

Introduction

A core activity of recent work on Greenland's diamond potential by the Geological Survey
of Denmark and Greenland (GEUS) has been the compilation of data submitted to Green-
land's Bureau of Minerals and Petroleum (BMP) according to the standard terms for min-
eral exploration licences. The data were amassed over a period of more than 30 years, with
the vast majority of data being from 1994 and younger. This work resulted in a digital data
package first published in 1993 (Jensen et al. 2003), and a major update and revision of
the compilation was published the following year (Jensen et al. 2004). The data, extracted
from 164 assessment reports, consist primarily of indicator mineral analyses (ca. 96 000)
from till and stream sediment sampling programmes covering the whole Archaean craton of
southern West Greenland. The numbers of mineral grains analysed and reported by explo-
ration companies and GEUS prior to 2004 are given in Table 1.

On a regional scale, the populations and chemistries of the indicator minerals outline rela-
tively well-defined areas as potentially prospective for diamonds, both in the Sarfartoq and
Maniitsoq regions. However, it has been the experience of GEUS, and presumably of sev-
eral exploration companies, that tracing diamond-favourable indicator minerals from till
back to a kimberlite or lamprophyre source was not straightforward where it has been at-
tempted, i.e., chiefly in the Sarfartoq region. There was an apparent lack of potential in situ
kimberlite source rocks to explain the relatively high numbers of indicator minerals with
favourable chemistry. Although the sampling density for till and stream sediment in the
Maniitsoq region was fairly even and high, and despite the existence of a clearly defined
linear trend with garnets and other indicator minerals with diamond-favourable chemistry,
as well as some diamondiferous kimberlites, the amount of follow-up tracking of kimberlite
has been very limited and few potential source rocks was studied in any detail here.

GEUS sampling project, 2004

In 2004 GEUS, in a collaboration project with BMP, launched new field and laboratory in-
vestigations to address the following questions:

1. Are the numbers and chemistries of indicator minerals reported by various explora-

tion companies reproducible and are they inter-comparable?

G E U S

2. Do the large numbers of indicator minerals reflect local in situ sources, or are they

derived from distant sources (possibly below the Inland Ice)?

3. Is the linear trend of diamond-favourable garnets from till samples in the Maniitsoq

region reproducible, and are the areas outside the trend really barren?

4. Is the empirical observation from available descriptions and own field reconnais-

sance that the dykes from the Maniitsoq region are closer in composition to `real
kimberlites' than those of the Sarfartoq region verifiable by detailed petrography
and mineral chemistry?

Figure 1. Kimberlite dykes and occurrences of the Maniitsoq region, several of which
were found during GEUS fieldwork in 2004.

Thus, in 2004, new till samples were collected and a number of kimberlite and lamprophyre
dykes were sampled with a view to separation and analysis of heavy minerals. Because of
the paucity of data from potential source rocks it became a key element in the new investi-
gation to separate and analyse mineral grains from a fair number of kimberlites and lam-
prophyres, and to examine the dispersal of indicator minerals in `down-ice' directions rela-
tive to the outcrops.

A total of 131 till samples and 41 rock samples were included in the study. The amount of
sample material submitted for heavy mineral separation was approximately 20 kg of <4
mesh (ca. 6.3 mm) material for till and sediment samples, and 35 kg of rock for kimberlite
and lamprophyre samples. The heavy mineral separation and picking was conducted by
Overburden Drilling Management in Nepean, Ontario, and the subsequent mounting and
microprobing of grains by GEUS. The numbers of mineral grains analysed to date from
these samples are given in Table 2. Publication of the new mineral chemistry data is

G E U S

planned for January 2006 in conjunction with the Mineral Exploration Roundup in Vancou-
ver. A few preliminary conclusions are presented as an appetiser here.

Table 1. Mineral grains microprobed prior to 2004 (Jensen et al. 2004)

Mineral group

Sample medium

Rock

Till

n = 31

n = 15295

CHR (chromite, spinel)

1058

11332

CPX (clinopyroxene)

205

15069

GAR (garnet)

1106

8875

ILM (ilmenite)

833

43318

OPX (orthopyroxene)

732

OL (olivine)

472

12333

Table 2. Mineral grains microprobed from GEUS' 2004 sampling programme, ex-

cluding in situ grains in kimberlite reported by Nielsen & Jensen (2005)

Mineral group

Sample medium

Rock

Till

n = 27*

n = 133

CHR (chromite, spinel)

841

390

CPX (clinopyroxene)

333

235

GAR (garnet)

2524

1417

ILM (ilmenite)

446

672

OPX (orthopyroxene)

* Results from 14 samples pending

The 2004 fieldwork resulted in the recognition of much larger kimberlite dyke systems than
had been previously reported from the Maniitsoq region (Figure 1), and these new localities
were included in the latest digital data compilation (Jensen et al. 2004). For example, in
one field area (Sillissannguit), where a very small kimberlite dyke had been previously re-
ported, new dykes with a combined length of more than 10 km were found, and in another
area near the coast (Timitta Tasersua E), what is believed to one of the largest kimberlite
dyke exposures known in Greenland was found in a steep gully above a stream with nu-
merous large boulders of kimberlite. Mineral chemistry data for these occurrences, along
with several occurrences in the Sarfartoq region, have only recently become available, and
it is now possible to address the provenance of the indicator minerals from till and stream
sediment.

The discovery of new dykes in 2004 was a first indication that the dispersion of indicator
minerals `down-ice' from outcropping kimberlite is very limited in the Maniitsoq region. A
regional, remarkably linear belt defined by diamond-favourable garnets in till, for example,
is interpreted to closely reflect the occurrence of kimberlite. All of the areas that have been
checked for outcropping kimberlite, based on the distribution of diamond-favourable gar-
nets in till, have been found to have kimberlite dykes or boulders a short distance N or NE
of the till sampling sites. This reflects an apparently uniform direction of ice movement of

G E U S

240° (SW) in the region. In the Sillissannguit area glacial striation on the surface of a kim-
berlite dyke has been observed. It is suggested that the source of indicator mineral grains
in till is to be sought, most likely within 12 km distance, NE of the till sampling sites.

Figure 2. Diamond-favourable garnets (G10D class of Grütter et al. 2004 plotted as top
layer) from (A) in situ rock samples, and (B) till samples.

Distribution of diamond-favourable indicator minerals

Peridotitic garnets from the 2004 till sampling occur in comparable numbers, locations and
classes as in earlier sampling (Figure 2), but a new area was added with diamond-
favourable G10D (Grütter et al. 2004) garnets near a 2 km long kimberlite dyke system at
Tasersuatsiaq in the Maniitsoq region (Figure 1, 2), where previous sampling was sparse.
Peridotitic garnets from in situ rocks appeared in the pre-2004 data to be concentrated in
the SW part of the Sarfartoq region, and almost exclusively within the undeformed Ar-
chaean craton. With the addition of data from the 2004 investigation, the in situ kimberlites
of the Maniitsoq region appears to have a comparable, if not higher, frequency of G10D
garnets (Figure 2). Another important result of the 2004 sampling is that several in situ
rocks in the N part of the Sarfartoq region also are now also seen to contain G10D and G10

G E U S

garnets, although generally in modest numbers. This possibly explains the widespread oc-
currence of these garnet types in till samples from the area.

Figure 3. Diamond-favourable eclogitic garnets (G3D class of Grütter et al. 2004 plotted
as top layer) from (A) in situ rock samples, and (B) till samples.

Eclogitic garnets of diamond-favourable class G3D (Grütter et al. 2004) occur in most of the
in situ rocks sampled in 2004, both in the Sarfartoq and Maniitsoq regions, but in the new
till samples mainly in the Maniitsoq region (Figure 3). The pre-2004 data showed the oppo-
site trend, and this may tentatively be explained by picking bias towards purple (peridotitic)
garnets by the exploration companies that operated in the Maniitsoq region and the SW
part of the Sarfartoq region; the picking of indicator minerals from samples in the rest of the
Sarfartoq region apparently recorded a more fair representation of purple and orange (ec-
logitic) garnets.

G E U S

Figure 4. Diamond-favourable chromites (chromite-inclusion-in-diamond class, Cr2O3
TiO2 diagram of Fipke 1994, plotted as top layer) from (A) in situ rock samples, and (B) till
samples.

Spinel populations (mainly chromite) include a small, but possibly significant number of
grains with high Cr

and intermediate MgO, considered to be typical of chromite inclu-

sions in diamond (Fipke 1994). These chromites also plot in the chromite-inclusion-in-
diamond fields in Cr

vs TiO

diagrams of Fipke (Fipke 1994; Fipke et al. 1995) and

(Grütter & Apter 1998). The pre-2004 data showed that chromite-inclusion-in-diamond
grains from in situ rocks were sparse, but tended to coincide with diamondiferous localities
(Figure 4). The 2004 rock sampling added several new localities with diamond-favourable
chromites (Cr

vs TiO

diagram of Fipke 1994) in the Maniitsoq region, confirmed their

overall distribution in the S part of the Sarfartoq region, and added new localities in the N
part of the Sarfartoq region. The new chromite data from till samples largely confirms the
distribution seen in the pre-2004 data.

Clinopyroxenes (chrome-diopside) from the till and stream sediment have apparent PT
compositions (Nimis & Taylor 2000) within the diamond stability field as defined by Ken-
nedy & Kennedy (1976) in widespread areas of Western Greenland, also in some areas

G E U S

where no kimberlite or lamprophyre sources are known (Jensen et al. 2004). However, on a
regional scale, the distribution of `diamond-stable' clinopyroxenes largely coincides with the
two main regions with kimberlite and lamprophyre intrusions. In the Maniitsoq region, only
sampling sites close to kimberlites appear to contain `diamond-stable' clinopyroxenes (Fig-
ure 5); they are notably absent from till samples from the areas of two major dyke systems.
This is interpreted to reflect that no samples have been collected within a few kilometres of
the dykes in the regional `down-ice' direction SW. In the Timitta Tasersua area, the `dia-
mond-stable' clinopyroxenes appear to be absent from till only about 1 km W of the large
dyke outcrops.

Figure 5. `Diamond-stable' clinopyroxenes (Kennedy & Kennedy 1976; Nimis & Taylor
2000) in till samples from of the Maniitsoq region.

Orthopyroxene is not an abundant mineral in the Greenland kimberlites and lamprophyres,
and this is reflected in low numbers of grains recovered from both rock and till samples. In
the Sarfartoq region, pre-2004 orthopyroxene analyses with diamond-favourable composi-
tions, those falling within the `diamondiferous harzburgite' and `diamondiferous lherzolite'
fields of Ramsay & Tompkins (1994), were from localities south of the Palaeoproterozoic
deformation front, where diamonds had actually been found. The 2004 sampling added
new localities with diamond-favourable orthopyroxenes in the N part of the Sarfartoq re-
gion. The impression from the previously available data that these orthopyroxenes are very
rare in the kimberlites of the Maniitsoq region was confirmed with the new data, and it is
interesting to note that no orthopyroxenes were recovered from samples of two major
dykes systems (Sillissannguit and Tasersuatsiaq).

Ilmenite is the most common of the indicator minerals analysed prior to 2004, and thus
constitutes about half of all the analyses available. A very high proportion (about 90 %) of
the ilmenites are `kimberlitic' in the recent classification scheme of Wyatt et al. (2004), i.e.,

G E U S

Mg-rich, with generally more than 6 wt% MgO. In till samples `kimberlitic' ilmenites are
ubiquitous and therefore not very diagnostic of specific kimberlite or lamprophyre sources.
However, as do the clinopyroxenes, on a regional scale, ilmenites from till with `kimberlitic'
composition outline areas that coincide well with areas of known kimberlite or lamprophyre
occurrences. The contribution of ilmenite grains to the till from crustal rocks, for example
the voluminous Kangâmiut dolerite dykes of the Sarfartoq region, appears to be minimal.
This, and the observation that the Greenland tills have higher heavy mineral content than
average tills from the Slave Craton (Overburden Drilling Management, personal communi-
cation), may help explain the relationship between fairly high numbers of indicator minerals
in till and numerous, but generally small outcrops of possible source rocks.

Conclusions

The separation, picking and analysis of heavy minerals from in situ kimberlite and lampro-
phyre occurrences has lent support to the hypothesis that the dispersion of indicator miner-
als `down-ice' from their source is limited to a few kilometres in most areas of southern
West Greenland. This seems especially convincing in the Maniitsoq region, where glacial
transport appears to have been uniformly to the SW (240°). Kimberlite float and the most
diamond-favourable indicator minerals are always found on the S side of these, generally
WSWENE striking dyke systems. The fact that the recently found kimberlite dykes in the
Maniitsoq region have remained unknown until 2004 is largely due lack of previous sys-
tematic follow-up on indicator mineral results. That most of the kimberlite localities in the
Maniitsoq region contain abundant diamond-favourable peridotitic and eclogitic garnets,
along with diamond-favourable chromites and clinopyroxenes adds weight to the sugges-
tion that those dykes not already tested for diamond content should be sampled and sub-
jected to caustic dissolution.

References

Fipke, C.E. 1994: Significance of chromite, ilmenite, G5 Mg-almandine garnet, zircon and

tourmaline in heavy mineral detection of diamond bearing lamproite. In: Meyer, H.O.A.
& Leonardos, O.H. (eds): Proceedings of the Fifth International Kimberlite Conference
2 . CPRM Special Publication 1/B Jan/94 , 366381. Rio de Janeiro: Companhia de
Pesquisa de Recursos Minerais.

Fipke, C.E., Gurney, J.J. & Moore, R.O. 1995: Diamond exploration techniques emphasis-

ing indicator mineral geochemistry and Canadian examples. Geological Survey of
Canada, Bulletin 423 , 86 pp.

Grütter, H.S. & Apter, D.B. 1998: Kimberlite- and lamproite-borne chromite phenocrysts

with `diamond-inclusion'-type chemistries. 7th International Kimberlite Conference,
Cape Town, 1317 April, 1998. Extended abstracts, 280282.

Grütter, H.S., Gurney, J.J., Menzies, A.H. & Winter, F. 2004: An updated classification

scheme for mantle-derived garnet, for use by diamond explorers. In: Mitchell, R.H. et
al. (eds): Selected Papers from the Eighth International Kimberlite Conference. Vol-
ume 2: The J. Barry Hawthorne Volume. Lithos 77 , 841857.

Jensen, S.M., Lind, M., Rasmussen, T.M., Schjøth, F. & Secher, K. 2003: Diamond explo-

ration data from West Greenland. Danmarks og Grønlands Geologiske Undersøgelse
Rapport 2003/21 , 50 pp. + 1 DVD.

G E U S

Jensen, S.M., Secher, K., Rasmussen, T.M. & Schjøth, F. 2004: Diamond exploration data

from West Greenland: 2004 update and revision. Danmarks og Grønlands Geologiske
Undersøgelse Rapport 2004/117 , 90 pp. + 1 DVD.

Kennedy, C.S. & Kennedy, G.C. 1976: The equilibrium boundary between graphite and

diamond. Journal of Geophysical Research 81 , 24672470.

Nielsen, T.F.D. & Jensen, S.M. 2005: The Majuagaa calcite-kimberlite dyke, Maniitsoq,

southern West Greenland. Danmarks og Grønlands Geologiske Undersøgelse Rap-
port 2005/43 , 59 pp. + 1 DVD.

Nimis, P. & Taylor, W.R. 2000: Single clinopyroxene thermobarometry for garnet peri-

dotites. Part I. Calibration and testing of a Cr-in-Cpx barometer and an enstatite-in-
Cpx thermometer. Contributions to Mineralogy and Petrology 139 , 541554.

Ramsay, R.R. & Tompkins, L.A. 1994: The geology, heavy mineral concentrate mineralogy,

and diamond prospectivity of the Boa Esperança and Cana Verde pipes, Corrego
D'anta, Minas Gerais, Brazil. In: Meyer, H.O.A. & Leonardos, O.H. (eds): Proceedings
of the Fifth International Kimberlite Conference 2 . CPRM Special Publication
1/B Jan/94 , 329345. Rio de Janeiro: Companhia de Pesquisa de Recursos Minerais.

Wyatt, B.A., Baumgartner, M., Anckar, E. & Grutter, H. 2004: Compositional classification

of `kimberlitic' and `non-kimberlitic' ilmenite. In: Mitchell, R.H. et al. (eds): Selected Pa-
pers from the Eighth International Kimberlite Conference. Volume 2: The J. Barry
Hawthorne Volume. Lithos 77 , 819840.

G E U S

The Diamond Potential of West Greenland: some
Global Insights

Kjarsgaard, B.A.

Geological Survey of Canada, Ottawa, Ontario, Canada K1A 0E8

Deep seated magmas capable of transporting diamond to the surface include kimberlite,
ultramafic lamprophyre, and olivine lamproite. Typically, economic diamond mines are ob-
served in areas of Archean crust, which are underlain by Archean mantle. From this per-
spective West Greenland is an interesting exploration area for diamonds based on: 1) the
existence of bonafide kimberlite at Maniitsoq, as typified by their `Trend 1' spinels, in con-
trast to ultramafic lamprophyres (UML, var. aillikite) which are now recognized to have a
diagnostic spinel population which lies intermediate between `Trend 1' and `Trend 2'; 2) the
Maniitsoq kimberlites are emplaced through Archean crust, which preliminary studies have
shown to be underlain by a thick lithospheric root (> 190 km) with a cool paleogeotherm (36
40 mW/m

), and; 3) the lithospheric mantle in the Maniitsoq area contains subcalcic har-

zburgite. However, it must be noted that: 1) large variations in mantle stratigraphy (and
hence diamond potential) can occur over very short distances ( <50 km) as seen in the
Kaapvaal craton (S. Africa) and the Slave and the Superior cratons (Canada), and; 2) that
within kimberlite fields, and specifically within individual kimberlite bodies there can be large
variations in diamond grade that are specifically related to geochemically and temporally
discrete `diamond transporting events' as observed in Canadian (and Southern African)
kimberlites. For West Greenland, it is suggested that a structural analysis for areas of Ar-
chean crust would provide insights about possible boundaries between mantle "micro-
blocks", as well as pathways/zones of weakness for intruding kimberlite or UML magmas.
Further, there is probably an intimate relationship between `diamondiferous' kimberlite
bodies and specific structures, and also that these `diamondiferous' kimberlite bodies will
also have diagnostic geochemical and age characteristics, which distinguish them from
weakly- or non-diamondiferous kimberlites.

G E U S

Mapping of the lithosphere beneath the Archaean
craton and Proterozoic mobile belt in West Green-
land

Larsen, L.M. & Garrit, D.

Geological Survey of Denmark and Greenland, Øster Voldgade 10, DK-1350 Copen-
hagen K, Denmark

Minerals in mantle xenoliths and separated garnet and oxide megacrysts in kimberlite and
aillikite dykes were investigated from three areas in West Greenland: the Maniitsoq
(Sukkertoppen) area well within the Archaean craton, the Sarfartoq area at the boundary
between the Archaean craton and the Nagssugtoqidian mobile belt, and the Sisimiut
(Holsteinsborg) area well within the mobile belt. The mobile belt was formed during a
Proterozoic continental collision event that reworked the Archaean terranes. The alleged
collision suture is located about 80 km north of Sisimiut.

The minerals were analysed for major elements by electron microprobe and for trace
elements by proton microprobe and laser-ICP-MS. The work was carried out as a PhD
project in the period 1994-2000.

No xenolith material was at the time available from the Maniitsoq area; however the kim-
berlite samples contained frequent garnet megacrysts. From the Sarfartoq area, the xeno-
lith population (208) consisted of dunite, harzburgite, lherzolite (all three with or without
garnet), wehrlite and glimmerite. Spinel-bearing types were rare, as were granulites. From
the Sisimiut area, the xenolith population (24) consisted of spinel lherzolite, dunite, granu-
lite, wehrlite, glimmerite and rare garnet lherzolite; only few dykes contained garnet megac-
rysts.

Garnet, chromite and ilmenite separates were analysed for 4 samples (543 garnets) from the
Maniitsoq area, 12 samples (1560 garnets) from the Sarfartoq area, and 3 samples (96
garnets) from the Sisimiut area. There is in general a dearth of garnets in the Sisimiut area.

Based primarily on the methods developed by W. L. Griffin and coworkers, sections through
the lithosphere in the Maniitsoq and Sarfartoq areas were constructed, whereas there was
not enough data from the Sisimiut area, primarily because the garnets there derive only from
specific intervals. The sections show that the lithosphere in the Maniitsoq and Sarfartoq areas
are of similar thickness (c. 235 km) though the base of the lithosphere at 235190 km depth
is strongly influenced by melt metasomatism. The two areas contain similar lithologies though
in different proportions. The most frequent rock type in the lithosphere is garnet lherzolite.
The amount of harzburgite at depths of diamond stability goes up to 40% in the Maniitsoq
area and is only c. 10% in the Sarfartoq area. In both areas the lithosphere is layered, most
strongly in the Sarfartoq area where the upper part is dominated by depleted harzburgites

G E U S

typical of Archaean cratons. Whereas the lithosphere in the Maniitsoq area is of Archaean
aspect throughout, the lower lithosphere in the Sarfartoq area is dominated by lherzolites and
metasomatised lherzolites similar to some Proterozoic sections. In both areas, but particularly
in the Sarfartoq area, there is extensive phlogopite metasomatism around 150 km depth. The
Sarfartoq carbonatite complex is probably genetically related to this layer.

REE patterns of garnets and clinopyroxenes show evidence of repeated depletion and
enrichment events in the mantle. No truly primitive mantle is present in West Greenland.
Some of the re-enrichment events are of ancient Archaen age, as judged from the primary
metamorphic and undisturbed texture of the xenoliths.

The lithosphere beneath the Maniitsoq and Sarfartoq areas is similar in mineralogy and
average chemical composition to Archaean lithosphere in other parts of the World. As far as
the data allow conclusions, the lithosphere beneath the Sisimiut area is less depleted than in
the other two areas, and the metasomatic processes seem to be different. This may be
connected to mantle disturbances during the Proterozoic collision event.

www.geus.dk > Arbejdsområder > Råstoffer og kortlægning > Grønland > Rapport 2005/68 - oversigt > Siden her

Workshop on Greenland's diamond potential

Danmarks og Grønlands Geologiske Undersøgelse Rapport 2005/68