International audienceSeveral methods have been developed to assess the thermal state of the mantle below oceanic ridges, islands, and plateaus, on the basis of the petrology and geochemistry of erupted lavas. One leads to the conclusion that mantle potential temperature (i.e., TP) of ambient mantle below oceanic ridges is 1430°C, the same as Hawaii. Another has ridges with a large range in ambient mantle potential temperature (i.e., TP = 1300–1570°C), comparable in some cases to hot spots (Klein and Langmuir, 1987; Langmuir et al., 1992). A third has uniformly low temperatures for ambient mantle below ridges, ∼1300°C, with localized 250°C anomalies associated with mantle plumes. All methods involve assumptions and uncertainties that we critically evaluate. A new evaluation is made of parental magma compositions that would crystallize olivines with the maximum forsterite contents observed in lava flows. These are generally in good agreement with primary magma compositions calculated using the mass balance method of Herzberg and O'Hara (2002), and differences reflect the well-known effects of fractional crystallization. Results of primary magma compositions we obtain for mid-ocean ridge basalts and various oceanic islands and plateaus generally favor the third type of model but with ambient mantle potential temperatures in the range 1280–1400°C and thermal anomalies that can be 200–300°C above this background. Our results are consistent with the plume model

[1] PRIMELT2.XLS software is introduced for calculating primary magma composition and mantle potential temperature (T P ) from an observed lava composition. It is an upgrade over a previous version in that it includes garnet peridotite melting and it detects complexities that can lead to overestimates in T P by >100°C. These are variations in source lithology, source volatile content, source oxidation state, and clinopyroxene fractionation. Nevertheless, application of PRIMELT2.XLS to lavas from a wide range of oceanic islands reveals no evidence that volatile-enrichment and source fertility are sufficient to produce them. All are associated with thermal anomalies, and this appears to be a prerequisite for their formation. For the ocean islands considered in this work, T P maxima are typically $1450-1500°C in the Atlantic and 1500-1600°C in the Pacific, substantially greater than $1350°C for ambient mantle. Lavas from the Galápagos Islands and Hawaii record in their geochemistry high T P maxima and large ranges in both T P and melt fraction over short horizontal distances, a result that is predicted by the mantle plume model.

No abstract

The formation of basaltic crust at mid-ocean ridges and ocean islands provides a window into the compositional and thermal state of the Earth's upper mantle. But the interpretation of geochemical and crustal-thickness data in terms of magma source parameters depends on our understanding of the melting, melt-extraction and differentiation processes that intervene between the magma source and the crust. Much of the quantitative theory developed to model these processes has neglected the role of water in the mantle and in magma, despite the observed presence of water in ocean-floor basalts. Here we extend two quantitative models of ridge melting, mixing and fractionation to show that the addition of water can cause an increase in total melt production and crustal thickness while causing a decrease in mean extent of melting. This may help to resolve several enigmatic observations in the major- and trace-element chemistry of both normal and hotspot-affected ridge basalts.

The program adiabat_1ph is a simple text‐menu driver for subroutine versions of the algorithms MELTS, pMELTS, and pHMELTS [Asimow et al., 2004; Ghiorso et al., 2002; Ghiorso and Sack, 1995]. It may be used to calculate equilibrium assemblages along a thermodynamic path set by the user and can simultaneously calculate trace element distributions. The MELTS family of algorithms is suitable for multicomponent systems, which may be anhydrous, water‐undersaturated, or water‐saturated, with the options of buffering oxygen fugacity and/or water activity. A wide variety of calculations can be performed either subsolidus or with liquid(s) present; melting and crystallization may be batch, fractional, or continuous. The software is suitable for Linux, MacOS X, and Windows, and many aspects of program execution are controlled by environment variables. Perl scripts are also provided that may be used to invoke adiabat_1ph with some command line options and to produce output that may be easily imported into spreadsheet programs, such as Microsoft Excel. Benefits include a batch mode, which allows almost complete automation of the calculation process when suitable input files are written. This technical brief describes version 1.04, which is provided as . Binaries, scripts, documentation, and example files for this and future releases may be downloaded at http://www.gps.caltech.edu/~asimow/adiabat. On a networked computer, adiabat_1ph automatically checks whether a newer version is available.

[1] The major element, trace element, and isotopic composition of mid-ocean ridge basalt glasses affected by the Azores hotspot are strongly correlated with H 2 O content of the glass. Distinguishing the relative importance of source chemistry and potential temperature in ridge-hotspot interaction therefore requires a comprehensive model that accounts for the effect of H 2 O in the source on melting behavior and for the effect of H 2 O in primitive liquids on the fractionation path. We develop such a model by coupling the latest version of the MELTS algorithm to a model for partitioning of water among silicate melts and nominally anhydrous minerals. We find that much of the variation in all major oxides except TiO 2 and a significant fraction of the crustal thickness anomaly at the Azores platform are explained by the combined effects on melting and fractionation of up to 700 ppm H 2 O in the source with only a small thermal anomaly, particularly if there is a small component of buoyantly driven active flow associated with the more H 2 Orich melting regimes. An on-axis thermal anomaly of 35°C in potential temperature explains the full crustal thickness increase of 4 km approaching the Azores platform, whereas a 75°C thermal anomaly would be required in the absence of water or active flow. The polybaric hydrous melting and fractionation model allows us to solve for the TiO 2 , trace element and isotopic composition of the H 2 O-rich component in a way that self-consistently accounts for the changes in the melting and fractionation regimes resulting from enrichment, although the presence and concentration in the enriched component of elements more compatible than Dy cannot be resolved.

An upgrade of the PRIMELT algorithm for calculating primary magma composition is given together with its implementation in PRIMELT3 MEGA.xlsm software. It supersedes PRIMELT2.xls in correcting minor mistakes in melt fraction and computed Ni content of olivine, it identifies residuum mineralogy, and it provides a thorough analysis of uncertainties in mantle potential temperature and olivine liquidus temperature. The uncertainty analysis was made tractable by the computation of olivine liquidus temperatures as functions of pressure and partial melt MgO content between the liquidus and solidus. We present a computed anhydrous peridotite solidus in T-P space using relations amongst MgO, T and P along the solidus; it compares well with experiments on the solidus. Results of the application of PRIMELT3 to a wide range of basalts shows that the mantle sources of ocean islands and large igneous provinces were hotter than oceanic spreading centers, consistent with earlier studies and expectations of the mantle plume model.

[1] We present new equation-of-state (EoS) data acquired by shock loading to pressures up to 245 GPa on both low-density samples (MgSiO 3 glass) and high-density, polycrystalline aggregates (MgSiO 3 perovskite + majorite). The latter samples were synthesized using a large-volume press. Modeling indicates that these materials transform to perovskite, postperovskite, and/or melt with increasing pressure on their Hugoniots. We fit our results together with existing P-V-T data from dynamic and static compression experiments to constrain the thermal EoS for the three phases, all of which are of fundamental importance to the dynamics of the lower mantle. The EoS for perovskite and postperovskite are well described with third-order Birch-Murnaghan isentropes, offset with a Mie-Grüneisen-Debye formulation for thermal pressure. The addition of shock data helps to distinguish among discrepant static studies of perovskite, and for postperovskite, constrain a value of K 0 significantly larger than 4. For the melt, we define for the first time a single EoS that fits experimental data from ambient pressure to 230 GPa; the best fit requires a fourth-order isentrope. We also provide a new EoS for Mg 2 SiO 4 liquid, calculated in a similar manner. The Grüneisen parameters of the solid phases decrease with pressure, whereas those of the melts increase, consistent with previous shock wave experiments as well as molecular dynamics simulations. We discuss implications of our modeling for thermal expansion in the lower mantle, stabilization of ultra-low-velocity zones associated with melting at the core-mantle boundary, and crystallization of a terrestrial magma ocean.