FAST COOLING OF THE SOUTHERN ADIRONDACK HIGHLANDS: CONSTRAINTS FROM DIFFUSIVE MODELING OF BOTH RETROGRADE EXCHANGE AND NET-TRANSFER REACTIONS

Metapelites from the Peck Lake and Green Lake formations, within the southern Adirondack Highlands, reveal peak assemblages produced during continuous vapor-absent biotite melting, at conditions of around 800 °C and 6-8 kb. The prograde history of the southern Adirondacks is similar to that of the Central Adirondack Highlands, but the retrograde—or cooling—history is poorly understood for both terrains. Reaction textures and garnet zoning patterns from the southern Adirondack rocks indicate two major retrograde reactions operative during cooling—retrograde exchange (REx) and retrograde net-transfer (ReNT). ReNT operates at higher temperatures during melt crystallization, whereas REx operates throughout cooling until closure of garnet-biotite Fe-Mg exchange.

For the first time (refer to Storm & Spear, in review), the affects of these two reactions are incorporated into a program to model multi-component diffusion at the rim of a garnet undergoing net-transfer-associated resorbtion that, in combination with retrograde exchange, produces diffusional profiles of Fe/(Fe+Mg) and spessartine (ReNT only) at the garnet rim. Spessartine profiles, produced during ReNT, are more sensitive to the early cooling history. In contrast, profiles of Fe/(Fe+Mg) in garnet are continuously modified, first by ReNT and then by REx. Together, the observed profiles of Fe/(Fe+Mg) and spessartine in garnet require, at a minimum, a three part cooling history (slow-fast-slow) with some period of rapid cooling (10² to 10³ °C/Ma). These results are in stark contrast to the estimates of Mezger et al. (1990), who reported rates of 1 to 10 °C/Ma for the Central Adirondacks.

Diffusion modeling of biotite inclusions in garnet (see Spear & Parrish, 1996) is also applied to the southern Adirondack samples to gain additional perspective on the cooling history. Biotite inclusions are shielded from the matrix and are not subjected to ReNT. Since REx alone affects the composition of the biotite inclusions, the cooling history derived from this method is more sensitive to the late cooling history. If a single cooling rate is assumed, biotite inclusions predict rates of around 1 to 10 °C/Ma. However, multi-step cooling histories similar to those predicted by diffusion profiles in garnet also provide decent fits to the biotite inclusion data.