Ultrafast eclogite formation via melting-induced overpressure

- We present the first petrologic constraints for eclogitization via overpressure.
- Garnet zonation records timescales of <1000 yr for transient eclogitization.
- Rates of compression (5 kbar, <500 yr) are far too fast for burial by subduction.
- Overpressure via local anataxis reconciles the transiency with tectonic rates.
- Forward modeling is applied to fit the garnet profiles with a Monte-Carlo scheme.

The conventional wisdom holds that metamorphic reactions take place at pressures near-lithostatic so that the thermodynamic pressure, reflected by the mineral assemblage, is directly correlated with depth. On the other hand, recent field-based observations and geodynamic simulations suggest that heterogeneous stress and significant pressure deviations above lithostatic (overpressure) can occur in Earth's crust. Here we show that eclogite, normally interpreted to form at great depths in subduction zones and Earth's mantle, may form at much shallower depths via local overpressure generated in crustal shear zones. The eclogites studied crop out as lenses hosted by felsic paragneiss in a sheared thrust slice and represent a local pressure and temperature anomaly in the Taconic orogenic belt, southern New England. Sharply-defined chemical zones in garnet, which record ∼5 kbar pressure rise and fall accompanied by a temperature increase of 150-200 °C, demonstrate extremely short timescales of diffusion. This requires anomalously fast compression (∼500 yrs) and decompression. We use coupled phase equilibria and garnet diffusion forward modeling to fit the observed garnet profiles and test the likely P-T-t paths using a Monte Carlo-type approach, accounting for off-center sectioning of garnet. The simulation shows that a ∼5 kbar pressure increase after the temperature peak is necessary to reproduce the garnet zoning. Remarkably, this post-peak-T compression (from 9 kbar to 14 kbar) lasted only ∼500 yrs. If the compression was due to burial along a lithostatic pressure gradient, the descent speed would exceed 30 m yr−1, defying any observed or modeled subduction rates. Local overpressure in response to partial melting in a confined volume (Vrijmoed et al., 2009) caused by transient shear heating can explain the ultra-fast compression without necessitating burial to great depth.