C–O–H fluid-melt-rock interaction in graphitic granulites and problems of quantifying carbon budget in the lower continental crust

Estimates on the geological carbon cycle are subject to large uncertainties that can be reduced by thorough observation of rocks. In this contribution, we focus specifically on C-O-H fluid-melt-rock interactions in graphitic metapelitic granulites and on their bearing to the carbon budget of granulitic roots of continents. We provide robust microstructural and thermometric constraints on the coexistence of anatectic silicate melts and C-O-H fluids up to near ultrahigh temperature conditions in the archetypal crustal section of Ivrea-Verbano Zone (IVZ, Italian Alps). Fluid inclusions in garnet are investigated before and after high-temperature experiments, and contain considerable proportions of CO2, CH4, N2, but lower H2O than predicted for graphitic systems at granulite facies. When comparing and contrasting the melt compositions obtained by Perple_X and rhyoliteMELTS with natural melts from IVZ, a much better match is obtained by the former, questioning the choice of rhyolite-MELTS for modelling melting equilibria of metasedimentary rocks and for quantifying carbon budget of the lower crust. Overall, data show that assuming only a limited extent of fluid-melt immiscibility in the deep crust contradicts the evidence from natural rocks and prompts to an incomplete view of actual carbon behavior and carbonic fluids. The available experimental dataset on CO2 solubility in felsic melts cannot be used to interpret the volatile budget of melt inclusions in graphitic migmatites and granulites, as most solubility experiments were conducted under carbonate-saturated (i.e. highly oxidizing) conditions which maximize CO2 content of melt, compared to graphitic (i.e. more reducing) protoliths. As a consequence, thermodynamic models still cannot account for all the complexities related with interactions among H2O-CO2-CH4 ternary fluids, H2Oand CO2-bearing anatectic melts and graphite-bearing residues in graphitic metapelites. Targeted experimental studies are therefore crucial to boost substantial computational efforts, before any precise estimates on carbon budget and fluxes in the lower anatectic crust can be made.