Speculations on CrustMantle Kimberlite Relations_ Churchill Province

Document Sample
Speculations on CrustMantle Kimberlite Relations_ Churchill Province Powered By Docstoc
					Speculations on Crust/Mantle
   Kimberlite Relations,
    Churchill Province

       B.A. Kjarsgaard
        GSC Ottawa

          April 8, 2005
Deep-seated kimberlite magma can pick up anything en-
route to the surface (mantle, crust, diamonds).

                                                        after Mitchell, 1991

NOTE: This is an old, out of date, simplified lithospheric section model.
Kimberlites in Churchill
     WC kimberlite fields, clusters
Somerset Island (~100-95 Ma; <1 cpht); Brodeur, north Baffin

Northern Melville Peninsula/Aviat
(~600 Ma) ~83 cpht

Wales Island

Repulse Bay

Rankin Inlet (214 Ma, 196 Ma)            geochron data from Heaman et. al., 2004
Janse’s Rule (1981, 1984)
- diamond mines in Archean terranes

                             from Kirkley et al., 1991
       Diamond Formation

                                        HIGH PRESSURE


                                      = THICK OLD, COLD

Diamonds can form in the lithosphere and asthenosphere
Diamonds from three ‘sources’:
Lithospheric mantle peridotite;
Lithospheric mantle eclogite;
Ultradeep (TZ, UM)

               P-type xenolith                             E-Type xenolith
Based on the occurrence of diamond bearing mantle xenoliths and inclusions in diamonds
    Peridotitic (P-type) inclusions
      Cr-pyrope garnet                Cr-diopside

E-type garnet in diamond   Sulphide in diamond (P & E-types)

Eclogitic (E-type) inclusions
Age of Formation of Diamonds

                                                 Lac de Gras

                                       after Pearson & Shirey, 1999

OLDEST diamond formation age is ~3.6 Ga (Mesoarchean);
 YOUNGEST diamond formation age ~85 Ma (Cretaceous).
Diamond formation ages suggest:

1.) Paleoarchean crust/mantle is not

2.) Mesoarchean crust/ mantle is
the most favourable, followed by
Neoarchean crust/mantle, and then
Paleoproterozoic crust/mantle which
has an Archean component .
                 melt extraction

                                                            Only true if
                                                              the crust
                                                               and the
                                   (oceanic lithosphere)
                                                            mantle are
                                                           i.e. old crust
                                                                = old
Modified after                     after Pearson, 1999
Paleogeotherms (P-T plots)

                                           Griffin et al
               from Griffin et al., 2004
            Garnet populations
Yellow - oceanic lithosphere (subduction)

Red - continental lithosphere


Mantle stratigraphic sections

Major + trace element data with
P-T estimates.

                                                  after Scully et al, 2004

             modified after Griffin et al, 2004

                                     after Griffin et al., 2003

Note the complexity!
Now think about Churchill Province
‘Simple’ lithosphere models: WATCH OUT!
    The classic interpretation would be:
Kirkland Lake           Lake Timiskaming
(thicker lithosphere)   (thinner lithospere)
      Superior                      (Grenville)

                                    after Mitchell, 1991
BUT, in reality the lithosphere is thicker toward
the Superior craton margin (Lake Timiskaming)

         Thinner              Thicker

                                           from Griffin et al, 2004

Lithosphere reconstruction, Kirkland Lake & Lake Timiskaming
Somerset Island kimberlites (discovered
1963, recognized as kimberlite in 1973)

                    from Kjarsgaard and Levinson, 2002
Direct Re-Os dating of lithosphere

                          from Pearson, 1999
Require PGE systematics to
 understand Re-Os ages!

                   from Irvine et al., 2003
Somerset lithosphere TRD, TMA ages

                          from Irvine et al., 2003
 Somerset Lu/Hf lithosphere ages

                                   50 km depth

                                     2.8 Ga

                                 “Post Archean”

from Schmidberger et al., 2002
                                  175 km depth
                                  One possible
 BUT, majority of Somerset lithosphere
formed over ca. ~100 m.y.(2.8 - 2.7 Ga)
Deepest & shallow samples have TRD and TMA ages of ~2.8 - 2.7 Ga,
consistent with 2.8 Ga crust formation ages (Baffin, Melville Peninsula),
Somerset (gneiss with 2.77 Ga upper intercept) and basalt/komatiite
extraction at 2.74 - 2.70 Ga (PAG, MRG).The lithosphere also records
 younger “crustal perturbations” OR growth (fluid/melt additions).
       50 km depth

      2.7 - 2.8 Ga

       <2.5 Ga
                                                  from Irvine et al., 2003

     175 km depth              Revised interpretation
                  Mantle signatures
 Somerset Island (no subduction signature)
 Aviat (subduction signature)
 Rankin Inlet (no subduction

                from www.shear

Subduction signature based on the presence of E-type garnets or of ultra-
depleted P-type garnets in kimberlite (published or public domain data).
          Mantle root destruction?
   After Helmstaedt, 2003        Notable exceptions to the ‘rules’
Plumes                       North Slave, Victoria Island
(Mackenzie, Franklin events) (diamond-bearing kimberlites)

Magmatic underplating        Premier Diamond Mine (Bushveld

Plate collision              Venetia Diamond Mine (within the
                             Limpopo collision belt)

                             Potentially consistent with the ‘rules’
Magmatic arcs/pervasive      Area of Christopher Island Formation,
metasomatism                 Central Churchill Province
                                                        Premier mantle xenoliths
                                                        showing the Bushveld
                                                        related melt metasomatism

                                              Bushveld intrusion
                                              outlined in black

                 after Ernst &Baragar, 1991

                                                                        after Hoal, 2003
Victoria Island, north Slave are
diamond-bearing kimberlites.                  Premier is a Diamond Mine
Potentially, kimberlites <1.9 Ga within areas of high to moderate
concentrations of CIF minette dykes and volcanics could have low
diamond tenor due to destruction of the mantle root by intense
metasomatic processes (at ca. 1.9 Ga) . This hypothesis is testable (and
incorrect) if kimberlites are discovered within this region with
significant diamonds.

                                                              from Peterson
                                                              et al., 2002
Based on the information presented in this talk, as well as information
presented by Peterson, and Berman et al., at the WC Project Workshop
2005 it can be suggested that:
1.) there is known, considerable variations of lithospheric mantle
stratigraphy within the Churchill Province (e.g. Somerset - Melville -
2.) there should also be considerable variations of lithospheric mantle
stratigraphy elsewhere within the Churchill Province, based on recent
subdivision of the Churchill Province into subdomains, based on
analogy with the Kaapvaal and Slave cratons.
3.) The western Rae and Queen Maude blocks (Kjarsgaard, 1998) and
southern Hearne can be suggested as prospective for diamonds,
followed by the northern Rae, and then the central Hearne.
4.) The northern Hearne, and the portion of the Rae adjacent to the
Snowbird Tectonic Zone (with abundant CIF dykes and volcanics) may
not be diamond favorable due to intense mantle metasomatism.

Shared By: