Why lower diatremes in kimberlitic and non-kimberlitic systems are non-stratified, homogenized, and contain steep internal contacts: episodic bursts and debris jets
In both kimberlitic and non-kimberlitic systems, the volcaniclastic fill of the lower diatreme zone is often described as "homogenized" or "well mixed". Although the components come from different sources, the deposits display "a crude degree of textural and lithological consistency" (Clement and Reid, 1989, "Kimberlites and related rocks", p. 632-646). Bedding is typically absent from the lower diatreme but in some pipes, columnar bodies of volcaniclastic material occur. These bodies have steep contacts with, and a different grain size, componentry, etc. than, the enclosing host. Sometimes the difference can be subtle and the contacts gradational, making recognition difficult. Good examples are documented from Arizona and Antarctica in basaltic systems and such columnar bodies are also known in kimberlites, where they are sometimes called "feeder conduits". Both the homogenized aspect of many diatremes, and the generation of steep internal contacts, have been attributed to whole-pipe fluidization by some recent workers. This process is unlikely to occur in large pipes because it would take a huge amount of gas being emitted at a sufficient rate to fluidize the whole pipe. Other recent models call for Plinian-scale eruptions. However it is clear that small episodic bursts, not sustained Plinian plumes, must explain the genesis of the hundreds of relatively thin beds in maar tephra rims (maar- diatreme volcanoes do not generate large ignimbrites or thick widespread pyroclastic fall layers). Here we examine what these episodic bursts may do to the underground part of the maar-diatreme volcano. An explosion at deep levels in the pipe will generate enough gas to mobilize newly fragmented magma and existing debris upward into a "debris jet", typically much narrower than the width of the diatreme. Debris jets propagate within the existing diatreme fill and may or may not reach the surface. Experimental studies can be used to illustrate the processes at work. With time and repetition, debris jets due to episodic bursts will not only create columnar bodies, but will also contribute to homogenizing the vent fill and destroying any bedding features in the lower diatreme. Only the last debris jets leave a record of well contrasted columnar bodies, since older ones would likely have been smeared out or destroyed by subsequent explosions. Post- depositional shaking or movement of the tephra would also have further diffused and blurred the boundaries between these older columnar bodies.
Basaltic Diatreme To Root Zone Volcanic Processes In Tuzo Kimberlite Pipe (Gahcho Kué Kimberlite Field, NWT, Canada)
Tuzo pipe is infilled by a series of coherent and fragmental kimberlite facies types typical for a diatreme to root zone transition level. Coherent or transitional coherent kimberlite facies dominate at depth, but also occur at shallow levels, either as dikes or as individual or agglutinated coherent kimberlite clasts (CKC). Several fragmental kimberlite varieties fill the central and shallow portions of the pipe. The definition, geometry and extent of the geological units are complex and are controlled by vertical elements. Specific for Tuzo is: (1) high abundance of locally derived xenoliths (granitoids and minor diabase) between and within the kimberlite phases, varying in size from sub-millimeter to several tens of meters, frequent in a belt-like domain between 120-200 m depth in the pipe; (2) the general presence of CKC, represented by round-subround, irregular to amoeboid-shaped clasts with a macrocrystic or aphanitic texture, mainly derived from fragmentation of erupting magma and less commonly from previously solidified kimberlite, as well as recycled pyroclasts. In addition, some CKC are interpreted to be intersections of a complex dike network. This diversity attests formation by various volcanic processes, extending from intrusive to explosive; (3) the presence of bedded polymict wall- rock and kimberlite breccia occurring mostly in deep levels of the pipe below 345 m depth. The gradational contact relationships of these deposits with the surrounding kimberlite rocks and their location suggest that they formed in situ. The emplacement of Tuzo pipe involved repetitive volcanic explosions alternating with periods of relative quiescence causing at least partial consolidation of some facies. The volume deficit in the diatreme-root zone after each eruption was compensated by gravitational collapse of overlying diatreme tephra and pre-fragmented wall-rock xenoliths. Highly explosive phases were alternating with weak explosions or intrusive phases, suggesting an external factor to control the explosive behaviour of the magma. The overall constant volatile content of the kimberlite does not explain the observed extreme change in emplacement behaviour. The facies architecture of fragmental facies dominated by vertical elements is similar to that in non- kimberlitic diatremes and indicates deposition from debris jets marking separate and repeated explosive volcanic events. In basaltic pipes, such jets are generated by phreatomagmatic explosions in the explosion chamber(s) of the root zone, causing abundant country rock fragmentation and further efficient mixture of the various particles. Phases of high explosivity formed the finely fragmented kimberlites containing a high percentage of wall-rock xenoliths, while the fluidal-shaped and partly welded texturally variable and wall-rock- poor transitional coherent facies suggest phases of repetitive, hot, and low-energy fragmentation forming kimberlite spatter. Peperite hosted in kimberlite tephra is also typically found in basaltic root zones. Time gaps in between volcanic eruptive periods are indicated by cognate pyroclasts and reworked wall-rock deposits emplaced by sporadic sedimentation events in subterranean cavities under the widening roof of the pipe. The presence of temporary caves in the root zone is proposed also by the occurrence of spherical CKC in deep- seated fragmental kimberlite and by spatter found in transitional coherent rocks. Evidence for caves was mostly preserved at deeper pipe levels advocating continuously recurring processes during the life span of Tuzo.
Mineralogy of Juvenile Lapilli in Fort a la Corne Pyroclastic Kimberlites
Pyroclastic-dominated kimberlites of the Fort a la Corne area (Saskatchewan, Canada) are characterized by the presence of juvenile ash and lapilli tuffs together with crystal tuffs composed of discrete crystals of olivine. Juvenile lapilli are amoeboid-to-curviplanar in shape and composed of olivine set in a fine-grained groundmass. Welding of juvenile lapilli is extremely rare. Clasts can be set in a matrix of calcite and/or serpentine. Commonly, olivine crystals protrude from the clast margins. Many, but not all, clasts contain vesicles filled with carbonates and/or serpentine. Carbonates in the vesicles include Sr-bearing calcite, dolomite and Ba-Mg carbonate. Individual kimberlite units in some instances contain several juxtaposed texturally- and mineralogically-different varieties of juvenile lapilli. In others, clasts are of similar petrographic character that differ in only in their spinel mineralogy. In many clasts early-forming microphenocrystal prisms of calcite are present. The groundmass of the ash and lapilli consists of perovskite, serpentine pseudomorphs after monticellite, diverse discrete euhedral, resorbed and atoll spinels, apatite, serpentine, and laths of quench dolomite. Groundmass mica appears to be absent. Spinel assemblages differ between different vents. At Candle Lake the overall trend of spinel compositions follows that typical of hypabyssal kimberlites with individual clasts within a given unit exhibiting segments of this trend. Compositions belong to a spinel- magnesiochromite-chromite-qandilite-magnetite solid solution series. At Smeaton (FALC 169 body), spinel compositions follow the same overall trend but are all relatively more evolved, and typically Ti-rich and Cr-poor. The majority of compositions plot on the rear face of the reduced spinel prism and belong to the spinel- qandilite-ulvospinel-magnetite series. It is concluded that spinels in Fort a la Corne kimberlites follow the "normal" evolutionary trend of spinel compositions established for hypabyssal kimberlites on a world-wide basis. Evolved spinels of very similar composition occur in juvenile lapilli in Diavik pyroclastic kimberlites. Spinel compositional data indicates that significant differentiation of the parental magma of the some Fort a la Corne kimberlites must have occurred prior to their eruption. The common presence of carbonate-filled vesicles and a quenched dolomitic groundmass in the juvenile lapilli demonstrates that significant quantities of volatiles were present in the magma at the time of eruption.
Geochemical Dissection of a Kimberlite: What Makes up a Whole Rock Analysis?
The elemental and isotopic composition of "primary" kimberlite magma and its relationship to basaltic magmas has been the subject of considerable debate for decades. We present a trace element and isotopic study of multiple mineral phases from a kimberlite and compare these to an olivine melilitite with significantly less visible crustal input. Our objective is to determine whether different mineral phases clearly record different stages of crustal contamination in the evolving kimberlite and to assess which phases are most likely to give the best information about potential source regions. We have analysed phlogopite, perovskite, olivine and calcite from a Group I kimberlite from Jos, Somerset Island. These results are compared with olivine, melilite, phlogopite, perovskite and apatite hand picked from an olivine melilitite from Saltpetre Kop (SPK), S. Africa. Melilitites are less obviously affected by crustal contamination and are generally less altered by low-T processes than kimberlites, yet they contain several minerals in common with kimberlites and therefore offer a good, simpler analogue system. Preliminary Sr isotope data from the melilitite confirm that melilite, olivine and perovskite have similar initial 87Sr/86Sr ratios to the whole rock; while the later crytsallising phases, apatite and phlogopite, show increasingly more radiogenic values. Together they define a reasonable Rb-Sr isochron age of ~77±3.4Ma close to the published emplacement age (72.5-76.8Ma, ), suggesting that crustal assimilation is relatively minor. In contrast, preliminary results from the Jos kimberlite are more complex and not consistent with closed system behavior. Variations in initial 87Sr/86Sr both within and between the different mineral phases so far analysed can be explained by a combination of factors that include progressive crustal contamination and source heterogeneity.  Duncan et al (1978), Geological Magazine, 115, 317-396
Geothermobarometry for ultramafic assemblages from the Emeishan Large Igneous Province, Southwest China and the Nikos and Zulu Kimberlites, Nunavut, Canada
To understand and contrast the origins of ultramafic assemblages from basaltic and kimberlitic rocks and their associated deposits, such as V-Ti magnetite and Ni-Cu-(PGE) sulfide deposits and diamond, applicable thermobarometers were evaluated and applied to the ultramafic assemblages from the Emeishan Large Igneous Province (ELIP), Southwest China and from the Nikos and Zulu Kimberlites of Nunavut, Canada. The ELIP is located in the Yangtze Block, Southwest China and composed of Permian Emeishan Flood basalt (EFB) and associated layered mafic-ultramafic intrusions. Some of these intrusions host V-Ti magnetite deposits; while others contain Ni-Cu-(PGE) sulfide deposits. It is not clear why some intrusions host magnetite deposits and others contain sulfide deposits. The P-T conditions for the ultramafic assemblages from the mafic-ultramafic intrusions in the ELIP were calculated in order to understand the origins and the associated mineral deposits. The ultramafic assemblages are peridotite, olivine pyroxenite, pyroxenite in the layered intrusions and the common minerals include spinel, olivine, clinopyroxene, orthopyroxene, and minor magnetite and ilmenite. Using a two pyroxene thermometer and a Ca-Mg exchange barometer between olivine and clinopyroxene, a spinel-olivine-clinopyroxene-orthopyroxene assemblage from the Xinjie intrusion yields a T-P of 905°C and 17 kbar; and a similar assemblage from the Jinbaoshan intrusion yields a T-P of 1124°C and 31 kbar. The Nikos kimberlite, near Elwin Bay on Somerset Island, is located at the northeast end of the northeast-southwest kimberlite zone; and the Zulu kimberlite is located on the neighboring Brodeur Peninsula of Baffin Island, Nunavut. The ultramafic assemblages from the Canadian Kimberlites include garnet lherzolite, garnet-spinel lherzolite, spinel lherzolite, dunite, garnet websterite, spinel websterite and garnet clinopyroxenite. The calculated P-T conditions are in the range of 760 to 1180°C and 25 to 60 kbar, follow a continental geotherm, and overlap the stability field of diamond. The ultramafic assemblages from the ELIP, Southwest China and from the Canadian Kimberlites were sampled from different depths in the lithosphere.
Seismically Anisotropic Subcontinental Mantle Lithosphere Caused by Metasomatic Wehrlite-Pyroxenite Dyke Stockworks
Recently published observations of teleseismic wave propagation indicate that wave-speed anisotropy nearly doubles within the Lac de Gras kimberlite field in NW Canada when compared to surrounding parts of the Slave craton. The implied increase in structural fabric in the lithospheric mantle cannot be explained entirely by stronger alignment of minerals, particularly olivine, or increased temperature effects uniquely beneath the kimberlite field. Forward modelling of the observed anisotropy implies the superposition of an additional fabric related to kimberlite eruption that does not coincide in its depth range precisely with regional anisotropy. Here the source of this additional fabric is newly interpreted to be metasomatic fluid conduits that form a macroscopic stockwork of carbonated, hydrated or otherwise metasomatised peridotite dykes within depleted harzburgite. The superimposed fabric is thus composed of rock types that record passage of kimberlites that formed at greater depths from <1% partial melts of carbonated lherzolite. These metasomatised peridotite conduits probably are composed of rocks such as pyroxenites or wehrlites and occupy up to 10% of the mantle where present beneath the Lac de Gras kimberlite field. Metasomatizing events can thus be more extensive in volume and extent or much older than the kimberlite erupted to the near-surface. Using teleseismic anisotropy to reliably predict the alignment of these dyke stockworks in the mantle lithosphere beneath known diamondiferous kimberlites has general application for informed target selection in diamond exploration. Reduced bulk shear wave speeds near these stockworks may also be diagnostic of their presence deep in the sub-continental lithospheric mantle and a more reliable and continuous estimate of pyroxenite-wehrlite content in the mantle that sparse xenolith samples.