In geostatistical studies, the fitting of the linear model of coregionalization (LMC) to direct and cross experimental semivariograms is usually performed with a weighted least-squares (WLS) procedure based on the number of pairs of observations at each lag. So far, no study has investigated the efficiency of other least-squares procedures, such as ordinary least squares (OLS), generalized least squares (GLS), and WLS with other weighing functions, in the context of the LMC. In this article, we compare the statistical properties of the sill estimators obtained with eight least-squares procedures for fitting the LMC: OLS, four WLS, and three GLS. The WLS procedures are based on approximations of the variance of semivariogram estimates at each distance lag. The GLS procedures use a variance–covariance matrix of semivariogram estimates that is (i) estimated using the fourth-order moments with sill estimates (GLS1), (ii) calculated using the fourth-order moments with the theoretical sills (GLS2), and (iii) based on an approximation using the correlation between semivariogram estimates in the case of spatial independence of the observations (GLS3). The current algorithm for fitting the LMC by WLS while ensuring the positive semidefiniteness of sill matrix estimates is modified to include any least-squares procedure. A Monte Carlo study is performed for 16 scenarios corresponding to different combinations of the number of variables, number of spatial structures, values of ranges, and scale dependence of the correlations among variables. Simulation results show that the mean square error is accounted for mostly by the variance of the sill estimators instead of their squared bias. Overall, the estimated GLS1 and theoretical GLS2 are the most efficient, followed by the WLS procedure that is based on the number of pairs of observations and the average distance at each lag. On that basis, GLS1 can be recommended for future studies using the LMC. 相似文献
Ar–Ar age measurements are reported for selected eclogitic clinopyroxene and garnet inclusions in Orapa diamonds and clinopyroxene inclusions in Venetia diamonds. Laser drilling of encapsulated clinopyroxene inclusions within Venetia diamonds released a maximum of 3% of the total 40Ar, indicating little diffusive transfer and storage of radiogenic 40Ar at the diamond–inclusion boundary. Apparent ages obtained during stepped heating of three diamonds are consistent with diamond crystallisation occurring just prior to the kimberlite eruption 520 Ma ago. Stepped heating of three clinopyroxene-bearing Orapa diamonds gave ages of 906–1032 Ma, significantly above the eruption age, but consistent with previously determined isotopic ages. A few higher apparent ages hint at the presence an older generation of Orapa diamonds that formed >2500 Ma ago. Orapa garnets also contain measurable K contents, and record a range of ages between 1000 and 2500 Ma. The old apparent ages and lack of significant interface 40Ar released by the laser probe, suggests that pre-eruption radiogenic 40Ar and mantle-derived 40Ar components are trapped in microinclusions within the pyroxene and garnet inclusions. 相似文献
Geological mapping and diamond exploration in northern Quebec and Labrador has revealed an undeformed ultramafic dyke swarm in the northern Torngat Mountains. The dyke rocks are dominated by an olivine-phlogopite mineralogy and contain varying amounts of primary carbonate. Their mineralogy, mineral compositional trends and the presence of typomorphic minerals (e.g. kimzeyitic garnet), indicate that these dykes comprise an ultramafic lamprophyre suite grading into carbonatite. Recognized rock varieties are aillikite, mela-aillikite and subordinate carbonatite. Carbonatite and aillikite have in common high carbonate content and a lack of clinopyroxene. In contrast, mela-aillikites are richer in mafic silicate minerals, in particular clinopyroxene and amphibole, and contain only small amounts of primary carbonate. The modal mineralogy and textures of the dyke varieties are gradational, indicating that they represent end-members in a compositional continuum.
The Torngat ultramafic lamprophyres are characterized by high but variable MgO (10–25 wt.%), CaO (5–20 wt.%), TiO2 (3–10 wt.%) and K2O (1–4 wt.%), but low SiO2 (22–37 wt.%) and Al2O3 (2–6 wt.%). Higher SiO2, Al2O3, Na2O and lower CO2 content distinguish the mela-aillikites from the aillikites. Whereas the bulk rock major and trace element concentrations of the aillikites and mela-aillikites overlap, there is no fractional crystallization relation between them. The major and trace element characteristics imply related parental magmas, with minor olivine and Cr-spinel fractionation accounting for intra-group variation.
The Torngat ultramafic lamprophyres have a Neoproterozoic age and are spatially and compositionally closely related with the Neoproterozoic ultramafic lamprophyres from central West Greenland. Ultramafic potassic-to-carbonatitic magmatism occurred in both eastern Laurentia and western Baltica during the Late Neoproterozoic. It can be inferred from the emplacement ages of the alkaline complexes and timing of Late Proterozoic processes in the North Atlantic region that this volatile-rich, deep-seated igneous activity was a distal effect of the breakup of Rodinia. This occurred during and/or after the rift-to-drift transition that led to the opening of the Iapetus Ocean. 相似文献
Sixteen kimberlite boulders were collected from three sites on the Munro and Misema River Eskers in the Kirkland Lake kimberlite field and one site on the Sharp Lake esker in the Lake Timiskaming kimberlite field. The boulders were processed for heavy-mineral concentrates from which grains of Mg-ilmenite, chromite, garnet, clinopyroxene and olivine were picked, counted and analyzed by electron microprobe. Based on relative abundances and composition of these mineral phases, the boulders could be assigned to six mineralogically different groups, five for the Kirkland Lake area and one for the Lake Timiskaming area. Their indicator mineral composition and abundances are compared to existing data for known kimberlites in both the Kirkland Lake and Lake Timiskaming areas. Six boulders from the Munro Esker form a compositionally homogeneous group (I) in which the Mg-ilmenite population is very similar to that of the A1 kimberlite, located 7–12 km N (up-ice), directly adjacent to the Munro esker in the Kirkland Lake kimberlite field. U–Pb perovskite ages of three of the group I boulders overlap with that of the A1 kimberlite. Three other boulders recovered from the same localities in the Munro Esker also show some broad similarities in Mg-ilmenite composition and age to the A1 kimberlite. However, they are sufficiently different in mineral abundances and composition from each other and from the A1 kimberlite to assign them to different groups (II–IV). Their sources could be different phases of the same kimberlite or—more likely—three different, hitherto unknown kimberlites up-ice of the sample localities along the Munro Esker in the Kirkland Lake kimberlite field. A single boulder from the Misema River esker, Kirkland Lake, has mineral compositions that do not match any of the known kimberlites from the Kirkland Lake field. This suggests another unknown kimberlite exists in the area up-ice of the Larder Lake pit along the Misema River esker. Six boulders from the Sharp Lake esker, within the Lake Timiskaming field, form a homogeneous group with distinct mineral compositions unmatched by any of the known kimberlites in the Lake Timiskaming field. U–Pb perovskite age determinations on two of these boulders support this notion. These boulders are likely derived from an unknown kimberlite source up-ice from the Seed kimberlite, 4 km NW of the Sharp Lake pit, since indicator minerals with identical compositions to those of the Sharp Lake boulders have been found in till samples collected down-ice from Seed. Based on abundance and composition of indicator minerals, most importantly Mg-ilmenite, and supported by U–Pb age dating of perovskite, we conclude that the sources of 10 of the 16 boulders must be several hitherto unknown kimberlite bodies in the Kirkland Lake and Lake Timiskaming kimberlite fields. 相似文献
Mantle xenoliths and xenocrysts were retrieved from three of the 88–86 Ma Buffalo Hills kimberlites (K6, K11, K14) for a reconnaissance study of the subcontinental lithospheric mantle (SCLM) beneath the Buffalo Head Terrane (Alberta, Canada). The xenoliths include spinel lherzolites, one garnet spinel lherzolite, garnet harzburgites, one sheared garnet lherzolite and pyroxenites. Pyroxenitic and wehrlitic garnet xenocrysts are derived primarily from the shallow mantle and lherzolitic garnet xenocrysts from the deep mantle. Harzburgite with Ca-saturated garnets is concentrated in a layer between 135–165 km depth. Garnet xenocrysts define a model conductive paleogeotherm corresponding to a heat flow of 38–39 mW/m2. The sheared garnet lherzolite lies on an inflection of this geotherm and may constrain the depth of the lithosphere–asthenosphere boundary (LAB) beneath this region to ca 180 km depth.
A loss of >20% partial melt is recorded by spinel lherzolites and up to 60% by the garnet harzburgites, which may be related to lithosphere formation. The mantle was subsequently modified during at least two metasomatic events. An older metasomatic event is evident in incompatible-element enrichments in homogeneous equilibrated garnet and clinopyroxene. Silicate melt metasomatism predominated in the deep lithosphere and led to enrichments in the HFSE with minor enrichments in LREE. Metasomatism by small-volume volatile-rich melts, such as carbonatite, appears to have been more important in the shallow lithosphere and led to enrichments in LREE with minor enrichments in HFSE. An intermediate metasomatic style, possibly a signature of volatile-rich silicate melts, is also recognised. These metasomatic styles may be related through modification of a single melt during progressive interaction with the mantle. This metasomatism is suggested to have occurred during Paleoproterozoic rifting of the Buffalo Head Terrane from the neighbouring Rae Province and may be responsible for the evolution of some samples toward unradiogenic Nd and Hf isotopic compositions.
Disturbed Re–Os isotope systematics, evident in implausible model ages, were obtained in situ for sulfides in several spinel lherzolites and suggest that many sulfides are secondary (metasomatic) or mixtures of primary and secondary sulfides. Sulfide in one peridotite has unradiogenic 187Os/188Os and gives a model age of 1.89±0.38 Ga. This age coincides with the inferred emplacement of mafic sheets in the crust and suggests that the melts parental to the intrusions interacted with the lithospheric mantle.
A younger metasomatic event is indicated by the occurrence of sulfide-rich melt patches, unequilibrated mineral compositions and overgrowths on spinel that are Ti-, Cr- and Fe-rich but Zn-poor. Subsequent cooling is recorded by fine exsolution lamellae in the pyroxenes and by arrested mineral reactions.
If the lithosphere beneath the Buffalo Head Terrane was formed in the Archaean, any unambiguous signatures of this ancient origin may have been obliterated during these multiple events. 相似文献
We perform a global statistical investigation into the problem of Coulomb stress triggering of earthquakes by using the Harvard CMT catalogue of shallow earthquakes from 1976 to 2001. We consider ‘earthquake pairs’, i.e., pairs of successive earthquakes occurring near each other with similar focal mechanisms, and address the problem of whether or not the change of the Coulomb failure stress (CFS) produced by the first earthquake of the pair ‘encourages’ the occurrence of the second one. An interesting feature is that such a Coulomb stress-triggering phenomenon has an apparent focal mechanism dependence: for thrust earthquakes, a more evident triggering effect can be observed. 相似文献
Blooms of the brown tide organismAureccoccus anophagefferens have recurred in the coastal bays in New Jersey since 1995 and in the coastal bays of Long Island since 1985. Intracellular viral-like particles (VLPs) were documented during 1999–2000 brown tide blooms in Little Egg Harbor, New Jersey, but it was not determined whether cells were infected during the termination of the bloom. The objective of this study was to determine if VLPs infected and lysed natural populations ofA. anophagefferens in coastal bays of New Jersey and New York in 2002 with the same frequency as in 1999–2000 and especially at the termination of the bloom. Our results confirmed that the highest percentage (37.5%) of VLP-infected cells occurred at the termination of the brown tide bloom in New Jersey in 2002. Intracellular VLPs were present throughout the bloom event. The percentage of visibly infected cells was higher at the beginning of the bloom than during the peak of the bloom. The intracellular VLPs in natural populations ofA. anophagefferens were consistent in size and shape (approximately 140 nm in diameter) and comparable to those in previous studies. Concentrated viral isolates, prepared from waters during brown tide blooms in New York and New Jersey in 2002, infected healthy laboratoryA. anophagefferens cultures in vitro. The viral isolates associated with the highest laboratory viral activity (lysis positive) were concentrated from water samples having the highest viral and bacteria concentrations. The intracellular viruses in these virally infected laboratory cultures ofA. anophagefferens were similar in size and shape to those found in natural populations. The successful isolation of a virus specific toA. anophagefferens from a brown tide bloom in the field, the similarity of ultrastructure of VLPs infecting both natural populations and laboratory infected cultures, and the pattern of VLP infection during bloom activity in combination with the observed high percentage of VLP-infected cells during bloom termination, supports, the hypothesis that viruses may be a major source of mortality for brown tide blooms in regional coastal bays of New Jersey and New York. 相似文献
The structure of the mid-Norwegian volcanic Vøring margin at the onset of the Maastrichtian–Paleocene extension phase reflects the cumulative effect of earlier consecutive rifting events. Lateral structural differences present on the margin at that time are a consequence of migration of the location of maximum extension in time between Norway and Greenland. The most important imprints (Moho depth, thermal structure) of these events on the lithosphere are incorporated in a numerical simulation of the final extension phase. We focus on a possible mechanism of formation of the Vøring Marginal High and address the relationship between spatial and temporal evolution of crustal thinning and thickening, uplift of the surface and strength of the lithosphere.It is found that the Vøring Basin formed the strongest part of the margin which explains why the Maastrichtian–Paleocene rift axis was not located here but instead jumped westward with respect to the earlier rift axes locations. The modeling study predicts that local crustal thickening during extension can be expected when large lateral thermal variations are present in the lithosphere at the onset of extension. Negative buoyancy induced by lateral temperature differences increases downwelling adjacent to the rifting zone; convergence of material at the particular part of the margin is mainly taken up by the lower crust. The model shows that during the final phase of extension, the crust in the Vøring Marginal High area was thickened and the surface uplifted. It is likely that this dynamic process and the effects of magmatic intrusions both acted in concert to form the Marginal High. 相似文献
