The Wrangellia composite terrane is one of the largest fragments of juvenile crust added to the North American continent since Mesozoic time, and refining its accretionary history has important implications for understanding how continents grow. New U-Pb geochronology and Hf isotopes of detrital zircons from Late Jurassic-Late Cretaceous strata from the forearc of the Wrangellia composite terrane allows more insight on the tectonic and paleogeographic history of the terrane. Our stratigraphically oldest samples from the Late Jurassic Naknek Formation have a detrital zircon U-Pb signature dominated by Early and Late Jurassic grains (195-190 Ma; 153-147 Ma). Hf isotopic compositions of these grains are juvenile to intermediate (εHf(t)=4.5-14.7). Disconformably above the Naknek Formation are two poorly understood units Ks and Kc. The Ks unit is dominated by Early to Late Jurassic grains (159-154 Ma) with a few Paleozoic grains (347-340 Ma). Hf isotopic compositions of Carboniferous-Jurassic grains are juvenile to intermediate (εHf(t)=6.0-18.8). The overlying Kc unit has Late to Early Jurassic zircons (198-161 Ma), and an increase in Paleozoic ages (374-323 Ma). Hf isotopic compositions of these grains are juvenile to intermediate (εHf(t)=4.5-14.7). Samples from the Matanuska Formation have major Late Cretaceous grains (90-71 Ma), and minor Early Cretaceous (137-106 Ma), Late to Early Jurassic (200-153 Ma), Paleozoic (367-277 Ma), and Precambrian grains (2597-1037 Ma). Hf compositions have a wider range from both the Late Cretaceous grains (εHf(t)=-1.5-14.9) and Paleozoic-Precambrian grains (εHf(t)=-23.7-16.3). Our results suggest an evolving provenance from Late Jurassic to Late Cretaceous time for the Wrangellia composite terrane forearc basin. The Late Jurassic Naknek Formation samples were dominantly derived from a juvenile to intermediate Jurassic igneous sediment source. During Early Cretaceous time, there is a slight increase in the number of Paleozoic grains in the Ks and Kc unit samples. The Early Cretaceous sediments have a mostly positive Hf isotopic compositions suggesting exhumation of Jurassic and Paleozoic juvenile igneous sediment sources. By Late Cretaceous time, our data illustrates another increase in Paleozoic grain abundances, in addition to the introduction of Precambrian grains, all with widely variable Hf isotopic compositions. We interpret this to reflect a larger sediment flux from the interior of Alaska where more evolved igneous rocks of that age are found.

Detrital zircon U-Pb geochronology is a useful tool for analyzing provenance in the sedimentary record. Differentiating recycled and first cycle populations in the detrital record, however, is not a straightforward process. A second potential problem in using detrital signatures to determine provenance of sediment lies in the assumption that detrital signatures of modern rivers reflect input from each exposed unit in the catchment boundaries. To investigate each of these problems, I present U-Pb analysis of detrital zircon (DZ) from modern river sand collected from 20 watersheds, 6 detrital apatite (DA) signatures from modern river sand, and 6 DA signatures from exposed strata, all within the Talkeetna Mountains (south-central Alaska). DA rarely survives past the first cycle of erosion and deposition due to its inability to survive chemical weathering, and thus dominantly represent igneous input in detrital signatures, whereas zircon can be of igneous origin or can survive multiple cycles of erosion and deposition. By comparing the DA signatures with the DZ signatures, I present a method to better differentiate first cycle, igneous sediment contributions from recycled populations within a detrital signature. The results of these comparisons show that DA signatures provide ages of igneous input into the detrital record; these ages are also reflected in the DZ signature, thus signaling these DZ populations as igneous in origin. This study also investigates the potential for DA recycling and DA input from recycled strata. To address the second problem, I present a method using GIS software and the most recent map of Alaska to create simulated signatures that records input on a scale proportionate to the exposed surface area of each bedrock unit. In ~35% of the watersheds tested, the simulated signatures predict trends similar to the DZ signatures from the modern river sands, in 55% of the watersheds tested the simulated signatures missed one or more populations present in the DZ signature, and in 10% of watersheds tested, the simulated signature predicted trends very different from the DZ signatures. In cases where the DZ and simulated signatures do not match, I believe this represents influences of climate and relief and zircon fertility.