WHAT LIES BENEATH THE ICE? USING THE GEOCHEMISTRY AND GEOCHRONOLOGY OF MODERN RIVER CLASTS TO DECIPHER THE EVOLUTION OF A GLACIATED VOLCANIC ARC (WRANGELL ARC, ALASKA, USA)

We present results from a novel technique of applying geochemical and geochronologic analyses on volcanic clasts collected from rivers presently encircling/draining the ~30 Ma-Recent Wrangell Arc (WA). The WA formed via flat-slab subduction of the Yakutat microplate and is an ideal location to study the interrelations between subduction, strike-slip faults, and slab-edge magmatic processes. However, the WA covers a huge area (~10,000 km²) and much of it is covered by glacier ice, which makes sampling bedrock on a province-wide scale challenging and inefficient. Our detrital petrological technique allows large areas of land to be targeted by relatively few sampling locales and is the first of its kind to utilize combined geochronology-geochemistry on river clasts for a petrological outcome at this scale. Bulk rock major and trace element data was obtained via XRF from 236 gravel to cobble-sized clasts collected from 17 major rivers that drain the western and central WA. New ⁴⁰Ar/³⁹Ar geochronology (n = 119 clasts with geochem. data; from 10 rivers so far) reveal that the WA initiated at ~30 Ma, mirror the more spatially restricted bedrock record, and that magmatism generally migrated southward and then northwestward through time mirroring province-wide bedrock studies. Forty-seven clasts are dated <~1 Ma and only nine clasts are dated ~29-20 Ma, whereas there are no clasts from ~17-13 Ma. Clast geochemistry shows three main WA magma types: Transitional-tholeiitic, calc-alkaline, and adakite-like. Calc-alkaline and adakite-like clasts are spatially and temporally ubiquitous in the WA, indicating that subduction and likely slab melting have been the dominant tectonic processes throughout WA history. Transitional-tholeiitic clasts are not found in southwestern WA rivers and are temporally restricted to ~11 – ~6 Ma and <1 Ma. Our results are confirmed by geochemical and geochronologic datasets from bedrock sampled from localized parts of the WA. Overall, this study shows that the sampled clast lithologies largely reproduce the known bedrock record in geochemical, temporal, and spatial contexts. The methodology applied here can be used in other locations where field conditions limit access to bedrock or as an efficient sampling technique for characterizing watershed-scale bedrock areas.