Genesis of metapelitic migmatites in the Adirondack Mts., New York

The Adirondack Mountains, NY are ideal for studying melting and migmatites. This well-studied region grades from the upper amphibolite facies Lowlands in the north-west to the granulite facies Highlands in the east and southeast, with migmatites found across both. U-Pb zircon ages (Heumann et al., 2006) have identified at least four major events during the Grenville orogenic cycle (~1.2-1.0Ga), three of which could have produced the conformable granitic leucosomes cutting melanocratic metasediments. To determine whether these leucocratic bands were local melts of surrounding melanosomes or externally-derived intrusions, we have correlated cathodoluminescence (CL) imaging, in situ zircon U-Pb geochronology by SHRIMP and δ¹⁸O measurements by CAMECA IMS 1280 ion microprobe, and in situ metamorphic garnet δ¹⁸O measurements by ion microprobe at 9 locations. CL imaging indicates three populations of zircons in both regions: 1. relatively featureless rounded balls (metamorphic), and rhythmically zoned (igneous) cores truncated by either 2. discordantly zoned (igneous) or 3. unzoned (metamorphic) rims. In both regions of the Adirondacks, δ¹⁸O analyses in overgrowths and whole zircons are highly variable for metamorphic zircon (6.1-13.4‰, mean 10.8±1.5‰, n=95). In contrast, garnet δ¹⁸O analyses are typically constant at each locality, differing only between leucosomes and corresponding melanosomes. Typical spot-to-spot reproducibility of δ¹⁸O in homogeneous zircon standards by ion microprobe is better than ±0.2‰ (1SD) from 10µm diameter spots. Values of δ¹⁸O in leucocratic layers are unusually high for plutonic rocks, especially in the SE, while surrounding melanosomes have similarly elevated values. Leucosomes therefore cannot represent nearby magmas and indicate melting of surrounding metapelites. The ability to measure δ¹⁸O and rare earth element (REE) patterns in co-existing zircon and garnet grains also allows us to consider the relative timing of these ‘peak’ minerals. In the Adirondacks, zircon crystallisation ages are often associated with P-T estimates from garnets (e.g., Mezger et al., 1993; McLelland et al., 2001; Alcock et al., 2004), so it is important to test this assumption. Coexisting grains in six samples from both regions were analysed for their REE patterns, and distribution coefficients calculated. In all cases, these distribution coefficients did not agree with experimental and empirical values from similar rocks. Indeed, all analysed zircon grains crystallised prior to any significant garnet growth, and therefore should not be correlated with P-T estimates based on garnet equilibria. In general, P-T-t paths should not rely on zircon geochronology to constrain peak conditions unless the growth of zircon and other minerals are shown to be coeval. Based on the combined U-Pb, δ¹⁸O and REE data, the Adirondack region experienced at least three stages of heating. The first crystallised zircons under igneous conditions (now seen as relict cores) before 1200Ma. The second phase (~1200-1100Ma) drove melting and dehydration of the rocks, crystallising zircon overgrowths and whole featureless grains. Finally, Ottawan granulite metamorphism at ~1050Ma crystallised ‘peak’ mineral assemblages, including garnet, under low a_H2O conditions due to earlier dehydration.