Temperatures and depths of origin of magmas fueling the Hawaiian volcanic chain
U.S. Geological Survey
The source depths and temperatures of Hawaiian magmas provide key information on the processes creating the Hawaiian Islands. Trace and some major element features of Hawaiian tholeiites point to separation from sources shallow in the garnet lherzolite stability field, at pressures near 30 kbar (± ~ 3). This corresponds to depths of ~ 90-100 km, which matches the estimated base of Cretaceous lithosphere. This coincidence may indicate that the lithosphere act as a lid on pressure-release melting processes by arresting ascent of bouyant source materials.
Source temperatures can be deduced by two methods. Melting experiments on garnet lherzolite show that the solidus at 30 kbar is very close to 1,500°C, so if melt is present at that depth, the temperature has to be above 1,500°C (but probably not much above 1,550°C to retain clinopyroxene in the source). The second method involves adjusting bulk lava or glass analyses for fractionation of olivine to arrive at compositions that would be in equilibrium with the most magnesian olivine phenocrysts found in Hawaiian tholeiites - Fo91 - which match olivine in peridotite. Olivine-liquid geothermometers can then give temperatures at different proposed pressures. Applying this method to picritic glasses, Dave Clague and coworkers infer primary magmas averaging 16.5 wt.% MgO, and olivine-saturation temperatures at 1 bar of 1,350°C. Since the olivine-saturation surface temperature increases at 5-7 °C/kbar, the source temperature would be 1,500-1,560°C at 30 kbar, in good agreement with the first method. Applying somewhat more complicated methodology and geothermometer, I estimated primary magmas with 18-19 wt.% MgO and 1 bar olivine-saturation temperatures of about 1,425°C, or 1,575-1,635°C in the source. The first method is the most direct, and a source temperature of 1,500-1,550°C at 30 kbar is the most reliable, in my opinion.
In comparison, a typical estimated source condition for mid-ocean ridge basalt would be 1,325°C at 10 kbar. Running both MORB and Hawaiian sources to 1 bar along 1 °C/kbar adiabats gives potential temperatures of 1,315°C for MORB source, and 1,470 - 1,520°C for Hawaii source, or a temperature difference of 150 - 200 °C at an equivalent pressure. It is doubtful to me that the temperature difference could be far outside this range.
Hawaiian tholeiites generally have SiO2 contents slightly higher than expected for garnet-saturated peridotite melts, and this may indicate that magma reacts to some degree with shallower lithosphere, but if so, this lithosphere must be poor or lacking in clinopyroxene and incompatible trace elements to preserve the tholeiites' garnet signatures. Alternatively the anomalous SiO2 may result from deficiencies in the experimental database.
Using the same method, Tom estimated a 70°C temperature anomaly for Icelandic tholeiites, which is similar to the estimates of other petrological methods and geophysics. The consensus from many, though not all, studies of the temperature difference between Hawaiian and MOR magmas is that Hawaiian magmas are 150 - 200 °C hotter. However, this is less than the minimum predicted to occur in the cores of even the shallowest, smallest, coolest plumes. Nevertheless, it is still a substantial temperature anomaly, and still requires explanation.