OVERVIEW OF COLUMBIA RIVER FLOOD BASALT GEOCHEMISTRY: WHAT IS STILL TO LEARN ABOUT MANTLE SOURCES?

The Columbia River Basalt Group (CRBG) consists of 210,000 km³ of basalt and basaltic andesite lava [1]. Recent age estimates indicate that 99% of this volume erupted between ~16.7 and 15.9 Ma for an overall magma flux of 0.26 km³/y; this value may have been exceeded by x5 during the peak eruptive periods. The lavas are overwhelmingly tholeiitic basalts to basaltic andesites. The contributing source reservoirs during the main phase [2] of lava production are EM II-type mantle (the Imnaha Component, IC), depleted Pacific MORB-source mantle variably fluid-fluxed by the adjacent descending slab, ancient cratonic north American crust, and arc-like accreted terrane crust [3]. Hypotheses for the origin of the province fall into two groups: dominantly “bottom-up”, that is, driven by a deep-seated mantle plume, as suggested by the geochemical character of IC; or “top-down”, fundamentally plate-driven. The IC shares isotopic and trace element features with some Pacific Ocean island group lavas, and has high ³He/⁴He, consistent with a plume origin. However, highly magnesian melt compositions, which would provide clear evidence for derivation from deep-seated upwelling mantle with anomalously high potential temperature, are absent. In addition, all CRBG lavas, including those with the least isotopic evidence for crustal contamination, have the elevated LILE/HFSE ratios expected from subduction-related processes. The geochemical features of the main-phase CRBG may perhaps be reconciled by a compromise model, where lithospheric instabilities affected regional upper mantle which was simultaneously being invaded by a rising plume. However, several questions remain regarding the petrogenesis of the later Wanapum and Saddle Mountains basalts, particularly the role of subcontinental lithospheric mantle vs. continental crust as the origin of elevated and highly variable ⁸⁷Sr/⁸⁶Sr, and unusual REE, LILE and HFSE characteristics. These questions may be amenable to solution via systematic Sr-Nd-Pb-Os data and high-quality ICPMS trace element analyses acquired on the same samples from these formations, but there is currently a shortfall of such data.

[1] Camp et al. (2017) USGS SIR 5022; [2] Camp & Ross (2004), JGR 109, doi:10.1029/2003JB002838; [3] Wolff & Ramos (2013) GSA SP 497, 273–291.