Ambient and excess mantle temperatures, olivine thermometry, and active vs. passive upwelling

Abstract

Mantle temperatures provide a key test of the mantle plume hypothesis, and olivine-liquid equilibria provide perhaps the most certain means of estimating mantle temperatures. Here, we review mantle temperature estimates and olivine thermometers, and calculate a new convective geotherm for the upper mantle. The convective geotherm is determined from estimates of sub-mid-ocean ridge (MOR) mantle potential temperatures ( T p is the T the mantle would have if it rose adiabatically without melting, and provides a reference for measuring excess temperatures at volcanic hot spots; T ex = T p hot spot − T p MOR). The Siqueiros Transform has high MgO glass compositions that have been affected only by olivine fractionation, and yields T p Siqueiros = 1441 ± 63 °C. Most mid-ocean ridge basalts (MORB) have slightly higher FeO liq than at Siqueiros; if Fo max (= 91.5) and Fe 2+–Mg exchange at Siqueiros apply globally, then upper mantle T p is closer to1466 ± 59 °C. Since our global MORB database was not filtered for hot spots besides Iceland, Siqueiros may in fact be representative of ambient mantle, so we average these estimates to obtain T p MOR = 1454 ± 81 °C; this value is used to calculate T ex. Global MORB variations in FeO liq indicate that 95% of the sub-MORB mantle has a global T range of ± 140 °C; 68% of this range (1 σ) exhibits temperature variations of ± 34 °C. Our estimate for T p MOR defines the convective mantle geotherm; this estimate is consistent with T estimates from sea floor bathymetry, and overlaps within 1 σ estimates derived from phase transitions at the 410 km and 670 km seismic discontinuities. Mantle potential temperatures at Hawaii and Samoa are identical at 1722 °C and at Iceland is 1616 °C; hence T ex is ≈ 268 °C at Hawaii and Samoa and 162 °C at Iceland. Furthermore, T p estimates at Hawaii and Samoa exceed maximum T p estimates at MORs by > 100 °C. Our T ex estimates agree with estimates based on excess topography and dynamic models of mantle flow and melt generation. Rayleigh number calculations further show that if our values for T ex extend to depths as small as 135 km, thermally driven, active upwellings will ensue. Hawaii, Samoa and Iceland thus almost assuredly result from thermally driven active upwellings, or mantle plumes. Estimates of T ex account for generalized differences in H 2O contents between ocean islands and MORs, and are robust against variations in CO 2, and major element components, and thus cannot be explained away by the presence of volatiles or more fusible source materials. However, our temperature variations at MORs do not account for H 2O variations within the MORB source region.