The most well supported hypothesis for the source of the Columbia River Flood basalts is a plume with crustal contamination (e.g. Hooper, 1984; Brandon and Goles, 1995; Brandon et al, 1993; Carlson et al, 1981). Brandon and Goles (1995) take note of two signatures seen in the CRFB’s. The early signature shows Sr and Nd isotopic values between MORB and OIB, suggesting a plume source for the Picture Gorge Basalts. Later they note a negative Epsilon-Nd value and a greater Sr-87/Sr-86 ratio, suggestive of a plume plus crustal source. Brandon et al (1993) state that Picture Gorge basalts have a plume source with up to 8 to 21% crustal contamination by volume for the most evolved flows. O, Pb, Nd and Sr data combined with phase equilibrium constraints reveal crustal contaminant.
Additional evidence for crustal contamination of a mantle plume source comes from Rhenium-Osmium data and Pressure-Temperature analyses. Rhenium-Osmium studies on the CRFB are not common, presumably due to the difficulties involved in the analysis. However, the results seem to be very conclusive, while Nd and Sr data has pointed towards a variety of sources. Chesley and Ruiz (1998) conducted a study on the CRFB based on the Re/Os isotope system. They concluded that the mafic lower crust was a contaminant but the SCLM was not involved at all. The Os signature is markedly different between mantle and crust and is not affected by metamorphism or weathering. The only reservoir with significant radiogenic Os is old continental crust, and data points towards this as a plume source contaminant.
This figure above, from Chesley and Ruiz, 1998, shows the Os isotopic evolution
of various reservoirs. The data for the CRFB fall into the crustal realm, and
are far from the SCLM.
This figure (Chesley and Ruiz, 1998) shows two components mixing and how the compositions would evolve over time. The plot A shows the evolution of dikes in the SCLM over different time periods. The value of the dikes overlaps with the CRFB Os data observed for samples collected on and off craton. Plot B demonstrates mixing of a plume with various end members including SCLM and three crustal sources which have evolved over different periods of time. The SCLM contaminant is very far from matching the observed data for the CRFB.
One final piece of support for a crustal contamination source comes from Pressure/Temperature analyses of clinopyroxenes and plagioclases in the Grande Ronde member of the Clarkston Basalt Group. The pyroxenes found in the basalts crystallized at crustal pressures. Caprarelli and Reidel (2004) suggest that the plume-derived magma ponded in the lower crust prior to being erupted.
The highest calculated pressure in a pyroxene examined by the authors is 0.62
GPa, which is placed in the lower crust.