A Consensus on Mantle Potential Temperatures? (Invited)

Recent publications may indicate a mounting consensus regarding mantle temperatures - an agreement that can be crucial for improving our understanding of mantle dynamics. To compare temperatures at various localities, McKenzie & Bickle (1988) proposed the concept of a mantle potential temperature (Tp) as a reference; Tp is the temperature the mantle would have at the surface, if it ascended along an adiabat without undergoing melting. Perhaps the most precise method to estimate Tp involves estimating the conditions of partial melting, and then correcting for the heat of fusion. The several sources of error include estimation of: a parental liquid, an equilibrium mantle olivine, the degree of partial melting (F), and the depth at which the parental melt is generated. There is also model error inherent to any thermometer. And when correcting for the heat of fusion we assume that we are correcting up to the convective adiabat, but if the parental melt was generated within the conductive lithosphere, Tp will be low. In any case, if we accept that the highest Tp estimates at Hawaii are of most interest (since magmas generated away from a plume centerline will not reflect the full heat content of a high T source), then in spite of these sources of error, recent estimates, published over a span of 10 months by three independent research groups, indicate considerable convergence. At Hawaii maximum Tp values are: 1600 deg. C by Herzberg & Asimow (2008), ca. 1630 deg. C by Lee et al. (2009; their Fig. 2B), and using two slightly different equation sets, 1687 deg. C by Putirka (2008; Geology) and 1660 deg. C by Putirka (2008; RiMG volume 69), yielding an average of 1644±38oC. Similarly, there is convergence for mean Tp at MORs (accepting that MORs are not isothermal; Klein and Langmuir, 1987): Herzberg & Asimow (2008) and Lee et al. (2009) estimate that Tp is ca. 1350 deg. C, and Putirka (2008; Geology) estimates a Tp of 1396 deg. C; these estimates average to 1365±26oC. All three groups further estimate that MORs exhibit a T range of 100 deg. C (Herzberg & Asimow, 2008; Lee et al., 2009) or 140 deg. C (Putirka et al., 2007). The agreement is especially remarkable given that each group uses different methods and/or equations. These absolute Tp values are furthermore consistent with a broad array of geophysical and experimental observations, including ocean floor bathymetry and heat flow (Stein & Stein, 1996), estimates for the peridotite solidus (Hirschmann, 2000), seismic estimates for depths of melting beneath MORs (MELT seismic Team, 1998), phase transitions at 670 km for a pyrolite mantle (Hirose, 2002), and excess bathymetry at Hawaii (Sleep, 1990). In addition, when pressures are calculated based on Si-activity (and T) for ocean islands (Putirka, 2008; RiMG), P is correlated with other indicators of partial melting depths, such as FeOt, and Na/Ti (Putirka, 2008 RiMG); these cross correlations indicate that inter-ocean island P and T estimates are real, and that thermal differences are an important control on melt composition. Finally, 3He/4He (a possible lower mantle signature; Kurz, 1993) positively correlates with both F and Tp for ocean islands (Putirka, 2008, Geology, p. e176), providing perhaps the most direct evidence in support of Morgan’s (1971) model that thermal plumes are driven by excess heat from the core-mantle boundary.