201D Seismology

61 Route 9W - PO Box 1000

Palisades

NY

10964-8000

US

Phone: 

(845) 365-8460

collier@ldeo.columbia.edu

Basaltic lavas, together with surface exposures of mantle rock, provide one of the principal sources of information about the processes by which the Earth's interior has evolved throughout geologic time. Such processes include partial melting, reactive melt transport and focusing, melt crystallization, convective stirring, crustal recycling, and diffusive homogenization. Although I am generally interested in the chemical and dynamical evolution of partially molten regions throughout Earth's crust and mantle, my dissertation research is focused on developing new inferences about the operation of some of these processes based on geochemical variability in the melts and source rocks of the relatively "simple" basaltic magma systems found at mid-ocean ridges. 

A common theme on the observational side of my work is the integration of statistical and spatial characterizations of geochemical variability. Although some aspects of this work have been less feasible in the past due to limited data availability and computing power, there are still very few fully integrated statistical-spatial characterizations of most igneous geochemical data, despite the growing abundance of data and analysis capability. Quantitative estimates of the underlying probability densities governing geochemical variability -- together with a characterization of associated spatial patterns -- can provide more informative constraints than characterizations of variability by, e.g., regionally averaged summary statistics. Since coupled geodynamic/petrological forward models can often predict specific patterns of spatial and/or statistical variability, integrated spatial-statistical characterizations of geochemical data offer an exciting opportunity for more powerful observational tests of specific geodynamic or petrologic hypotheses.

My primary research toolbox includes exploratory data analysis and computational statistics techniques, Bayesian statistical inverse theory, geochemical and petrologic modeling (the effect of melting, mixing, crystallization, etc. on major and trace elements, isotopes), and simple two-phase porous flow magma transport models, but has also included fieldwork in Oman and California and some electron microprobe work. Recent projects: (1) an up-to-date characterization of along-axis mid-ocean ridge basalt (MORB) variability using PetDB; (2) the effects of sampling method and sampling density on apparent variability in along-axis MORB; (3) predicted changes in magma chemistry caused by various phases of mid-ocean ridge basalt petrogenesis using fluid dynamical and geochemical/petrological calculations; (4) statistical inversions for the parameters governing basalt petrogenesis based on basalt major element analyses interpreted through various geothermometry and barometry techniques, as well as pMELTS/MELTS; (5) characterization of intermediate length-scale mantle heterogeneity in the Oman ophiolite.

Among the key geologic results from this research is the argument that simultaneous crystallization and chemical reaction with surrounding mantle rock (which co-author Peter Kelemen and I term "reactive crystallization") may represent a common and important process modifying the liquid line of descent of MORB magmas within the thermal boundary layer beneath mid-ocean ridges. "Fractionation-corrected" MORB compositional variability could therefore be primarily caused by sample-to-sample variations in the relative extents of reactive vs fractional crystallization, in turn reflecting variations in melt transport through the thermal boundary layer. Reactive crystallization may obscure the effects of, e.g., variations in mantle source properties or deep melt transport within the melting region, the currently favored interpretations. 

Most recently, we have detected and characterized multiple spatially coherent regions of mantle of up to 50 km length scale with distinctive statistical variability using detrital spinel samples collected within the Wadi Tayin massif of the Oman ophiolite. This ophiolite is perhaps the best analogue for "typical" mid-ocean ridge mantle exposed at the surface world wide. Our results expand the range of heterogeneity length scales that have been observationally documented in mid-ocean ridge processed residual mantle by more than an order of magnitude, and provide the basis for new inferences about lithologic and isotopic heterogeneity in the mantle.