Effects of melt depletion and metasomatism on the heat generation in the continental lithospheric mantle below Jagersfontein, South Africa

Abstract

Present-day continental lithospheric mantle (CLM) heat production estimates vary considerably and likely overestimate heat generation due to the infiltration of the host magma (i.e., kimberlite), mantle metasomatism or variable heat-producing element (HPE) ratios. We present estimates of heat production in the CLM beneath Jagersfontein, from bulk rock reconstruction of 11 peridotitic xenoliths based on in-situ analyses of primary mineralogy, to avoid kimberlite contamination. Higher concentrations of Th and U are observed in the reconstructed bulk rocks at shallower depths (< 5 GPa) and decrease towards the deepest parts of the CLM (Th: 0.5–26 versus 1–5 ppb; U: 0.4–19 versus 1–3 ppb). Moreover, the reconstructed samples have a broad range of bulk K/U (~ 70-16500) and Th/U ratios (~ 0.2–3.8), outside the expected range of the modern convecting mantle. A crucial factor is garnet, as it can control the U budget, has Th/U < 1 and is present across the CLM in the garnet stability field. The differences of the CLM with the convecting mantle challenge the use of assumedly constant HPE ratios to calculate the heat production. Our estimates of present-day heat generation from reconstructed bulk data yield ~ 0.0002–0.008 µW/m³ at shallow depths, decreasing down to ~ 0.0005 µW/m³ near the lithosphere-asthenosphere boundary, lower than typical heat generation values used in most previous models. The variable heat production in the CLM derives from the metasomatism and re-fertilization near the base caused by rising asthenospheric melts, which react and fractionate as they ascend, potentially carrying most of the HPE in a fluid phase to shallower depths.