Abstract

Our ability to calculate the depths and temperatures at which magmas partially crystallize can prove crucial as petrologists test hypotheses of magma transport and evolution. Yet whereas numerous magma transport arguments involve hydrous and SiO2-rich volcanic products, current clinopyroxene-liquid thermobarometers have been calibrated only from basaltic liquid compositions. To remedy this deficiency, new thermobarometers have been calibrated using new experiments that include hydrated (water-undersaturated) and SiO2-rich liquids ranging to 71.3 wt% SiO2. As with prior models, the new calibrations are based on jadeite crystallization and jadeite-diopside + hedenbergite exchange equilibria:

 
\[\mathit{P}(kbar)\ =\ {-}88.3\ +\ 2.82\ {\times}\ 10^{{-}3}\mathit{T}(K)\ ln\left[\frac{[Jd^{cpx}]}{[Na^{liq}Al^{liq}(Si^{liq})^{2}]}\right]\ +\ 2.19\ {\times}\ 10^{{-}2}\mathit{T}(\mathit{K})\ {-}\ 25.1\ ln[Ca^{liq}Si^{liq}]\ +\ 7.03[Mg{^\prime}^{liq}]\ +\ 12.41\ ln[Ca^{liq}]\]
 
\[\frac{10^{4}}{\mathit{T}(K)}\ =\ 4.60\ {-}\ 4.37\ {\times}\ 10^{{-}1}\ ln\left[\frac{[Jd^{cpx}Ca^{liq}Fm^{liq}]}{[DiHd^{cpx}Na^{liq}Al^{liq}]}\right]\ {-}\ 6.54\ {\times}\ 10^{{-}1}\ ln[Mg{^\prime}^{liq}]\ {-}\ 3.26\ {\times}\ 10^{{-}1}\ ln[Na^{liq}]\ {-}\ 6.32\ {\times}\ 10^{{-}3}[\mathit{P}(kbar)]\ {-}\ 0.92\ ln[Si^{liq}]\ +\ 2.74\ {\times}\ 10^{{-}1}\ ln[Jd^{cpx}]\]

Here, T is in Kelvins and P is in kbar. Jdcpx is the mole fraction of jadeite in clinopyroxene, where pyroxene cations are calculated on the basis of 6 O atoms, and Jd is the lesser of Na or VIAl; remaining Al is used to form CaTs. DiHdcpx is the mole fraction of diopside + hedenbergite in clinopyroxene, calculated as the fraction of Ca remaining after forming CaTs (= VIAl – Jd), CaTiAl2O6 [= (IVAl – CaTs)/2], and CaCr2SiO6 (=Cr/2). Terms such as Alliq refer to the cation fraction of AlO1.5 in the liquid, Fmliq is the sum FeOliq + MgOliq, and Mg′liq is the cation fraction ratio MgOliq/(MgOliq + FeOliq). Errors are similar to earlier models that utilize basalt compositions only. For the barometer, R2 = 0.97 and the standard error of estimate (SEE) is 1.7 kbar for the regression data; for the thermometer R2 = 0.96 and the SEE is 33 K. The models are applied to Neogene lavas from the Tibetan Plateau, and Neogene-Quternary lavas from the eastern Snake River Plain (SRP), Idaho. Crystallization depths for Tibet cluster at the middle/lower crust boundary. Magma-crust density relationships suggest that the middle crust may act as a level of neutral buoyancy. In the SRP, however, magmas appear to bypass several possible density traps. We suggest that fracture properties, such as dike size and aspect ratio, control magma transport in the SRP.

You do not currently have access to this article.