Controls on the Emplacement Style of Coherent Kimberlites in the Lac de Gras Field, Canada

In the Lac de Gras (LDG) kimberlite field, Northwest Territories, Canada, coherent kimberlites (CKs) occur as tabular dykes, pipe-shaped diatremes, and irregular bodies without well-defined geometries. Combining the morphology of CK bodies with the occurrence of fragmented olivine microcrysts allows distinction of four CK types at LDG: (1) dykes with no broken olivine; (2) CK without well-defined but probable sheet geometry and no broken olivine; (3) pipe-filling CK (pfCK) with abundant broken olivine and (4) pfCK with no broken olivine. These features suggest an intrusive origin for type 1 and, probably, type 2 CK; a high-energy extrusive emplacement for CK type 3 and a low-energy intrusive or extrusive emplacement for the CK type 4. Here, we compare petrographic and whole-rock, olivine and spinel compositional data for high-energy extrusive pfCK, low-energy pfCK and intrusive CK units to understand the factors controlling their variable emplacement styles. Extrusive CK contain more abundant groundmass phlogopite and monticellite, lower carbonate/silicate mineral abundance ratios and significantly lower dolomite and pleonaste-spinel abundances compared to intrusive CK. This indicates greater CO2 loss and higher H2O/CO2 in the melt phase for the extrusive CK during emplacement. Lower incompatible element concentrations in the extrusive CKs and different chromite Ti# and olivine rim Mg# indicate derivation from distinct primitive melt compositions. The extrusive CK feature higher eNd(i) and marginally higher eHf(i) compositions than the intrusive CK, pointing to derivation from distinct sources. These findings strongly imply that distinct primary melt compositions were largely responsible for the differences in emplacement styles of CK at LDG. Low-energy pfCKs have similar olivine rim Mg#, chromite Ti# and, hence, primitive melt compositions to the high-energy extrusive CK samples. Their marginally different emplacement styles may depend on local factors, such as changing stress regimes, or slightly different volatile concentrations. Both types of pfCK might reflect the waning stages of volcanic sequences resulting from the eruption of a segregated magma column that started with pipe excavation and the explosive emplacement of gas-rich magma (volcaniclastic kimberlite), followed by the less energetic emplacement of melt-rich magma (pfCK). This hypothesis underscores different primary melt compositions for dyke vs pipe-forming (and filling) kimberlites and hence a fundamental primary melt control on the explosivity of kimberlites.