Date of Award

July 2016

Degree Type


Embargo Date


Degree Name

Doctor of Philosophy (PhD)


Earth Sciences


Gregory D. Hoke


Detrital, Geochronology, Multidimensional Scaling, Provenance, Tibetan Plateau, Zircon

Subject Categories

Physical Sciences and Mathematics


Detrital provenance analysis is an important tool in our understanding paleo-fluvial drainages, erosion of landscapes, paleogeographic reconstructions, and regional geologic and tectonic histories. Here, we examine the utility of multiple provenance analysis techniques through their application to paleo, modern, and synthetic detrital datasets. Zircon is a common accessory mineral found in most detrital sediments, primarily due to its refractory nature. During formation, zircon preferentially incorporates the element U and Th into its crystal lattice while excluding Pb, making it ideal for radiometric dating. Inexpensive and time-efficient U/Pb age acquisition techniques make zircon the mineral of choice for a majority of provenance studies.

The sedimentary basins between the Yangtze River and Red River have long been used to argue for a Mississippi River-scale paleo-drainage. We examine the U/Pb zircon ages of Cenozoic deposits ranging from Eocene to Pliocene age from basins surrounding the first bend of the Yangtze River and upper reaches of the Red River. We combine this data with a comprehensive suite of zircon grain-ages from contemporaneous deposits, modern fluvial sediments, and bedrock source units from previously published literature. Using the new technique developed here, of combining age spectra deconvolution and age component interpolation maps, it becomes clear that Cenozoic deposits of the Southeastern margin of the Tibetan Plateau do not share provenance with offshore sediments associated with the Paleo Red River in the Yinggehai Song-Hong Basin. This, coupled with detailed stratigraphic measurements and interpretations, as well as paleoflow measurements strongly suggests that at least since the Eocene, there was no connectivity between the Yangtze and Red Rivers.

In a modern setting, examination of provenance of fluvial sediments collected throughout a known catchment can provide insight into regional erosional patterns. The modern Yangtze River, the largest river in Eurasia, provides a perfect setting to apply detrital zircon provenance analysis. We use a previously published zircon U/Pb age distribution dataset of fifteen trunk stream samples and ten samples of the largest tributaries feeding the Yangtze. We apply a series of age-distribution analysis techniques to examine both downstream changes in provenance of trunk stream samples as well as identify the key bedrock and tributary sources of sediment to the trunk stream samples throughout the Yangtze's reach. The original work using this dataset argued that increasing anthropogenic influences, primarily agricultural, lead to a greater than expected influence of the Han, Yuan, and Xiang Rivers, whose confluence with the Yangtze occur in the middle-to lower reaches. The quantitative analysis developed here, however, shows a consistent distribution of U/Pb ages for Yangtze River trunk stream sediments is established in the upper reaches of the Yangtze after the first bend and is maintained some 3000km downstream. The signal is most likely derived from the erosion of the geologic terranes of the Songpan Ganze Terrane and the Longmenshan range, which are sourced primarily by the Yalong, Min, and Dadu rivers. These sources of sediment are consistent with known areas of greater stream power due to higher slopes, exhumation rates, and tectonic activity.

One technique that has recently been applied to detrital zircon datasets is multidimensional scaling, or MDS. MDS transforms pairwise dissimilarity measurements of sample U/Pb age distributions into Euclidian distances and then some optimal configuration, where greater distances between sample points represents greater dissimilarity between their respective age distributions. While MDS is not new, its application to detrital zircon datasets has never been rigorously tested. We examine several important issues in the application of MDS to detrital zircon research, including how intra-sample variation is represented as well as how dissimilarities are calculated; how random sampling associated with dating a limited number of zircon grain ages can and does affect the resulting MDS configuration; and how MDS differentiation is affected by samples containing either varying degrees of overlapping, shared, or unique age components. In application of MDS to both synthetic and real-world datasets, we illustrate the usefulness of the approach in the interpretation of detrital zircon age data; which suggests that thoughtful application of MDS mapping to detrital zircon data can afford significant advantages in the geologic interpretation of zircon grain ages.


Open Access

Available for download on Wednesday, September 20, 2017