A revision of the Ni-in-garnet geothermometer with special regard to its pressure dependence

We have explored the effect of temperature (T) and pressure (P) on Ni partitioning between garnet and olivine in well-equilibrated mantle xenoliths from on-craton, marginal-craton and off-craton settings and in high-P–T experiments at natural Ni abundance. Contrary to previous evaluations, the xenolith and experimental data indicate that the P effect is not negligible, consistent with the significant volume change of the garnet–olivine Mg-Ni exchange reaction. The recognition of a P effect satisfactorily resolves the discrepancies observed using previous calibrations of the Ni-in-Grt geothermometer and provides a solution to the long-standing controversy as to which Ni-in-Grt geothermometer is best applied to natural chromian pyrope compositions. A recalibrated, P-dependent Ni-in-garnet geothermometer reproduces the pyroxene T estimates for the xenoliths and the T conditions of the experiments with a standard error of estimate of 44 °C. The P dependence (ca. 40 °C/GPa) is comparable to that of the garnet–olivine Fe–Mg exchange geothermometer. A small tendency to overestimate at T < 900 °C relative to two-pyroxene thermometry is observed, which is unrelated to garnet compositional parameters. A set of simplified, geotherm-referenced calibrations permit traditional use of the Ni-in-garnet geothermometer as a single-mineral method, provided the local geotherm is known or can be estimated or inferred.