Impact of rock properties on landscape form and detrital studies
Item statusRestricted Access
Embargo end date03/07/2020
Lavarini Ferreira, Chrystiann
The nature of the rocks exposed at the surface of the Earth influence both geomorphology and the release of sediment from hillslopes to rivers. Whereas this has been known for centuries, very few studies have quantitatively linked landscape form to rock type. Even fewer studies have linked rock type, abrasion and the selective release of minerals from gravel to sand, and their impact on the information contained in a river sediment sample. This thesis aims to address these two key issues in geomorphology. Detrital studies use properties of river sediment to infer processes and rates across the catchment from which it has been sourced; they have been crucial in addressing issues such as linkage between tectonics and erosion and the recycling of materials though cycles of mountain building and destruction. However, these studies rely on the assumption that a sand sample integrates catchment–wide information without bias. In a first chapter, I perform a series of experiments with an abrasion model to test if abrasion, which controls the release of minerals from gravel to sand, influences detrital age signatures and the erosion rates retrieved from them. In the study case (Marsyandi watershed, Himalaya), I use an extensive zircon age population dataset published by Amidon et al. (2005). The results demonstrate that pebble abrasion can change the zircon mixing proportions of upstream source units as well as the age distribution of mixed fluvial sands. This change is particularly significant when there is strong contrast in rock resistance within the watershed (e.g., when pebbles of sandstone have 31 % of mass loss/km, and pebbles of quartzite have 0.15 % of mass loss/km). Pebble abrasion is one of many factors that can change the mixing proportion of sands, including hillslope gravel supply, erosion rates, and mineral fertility. In our study, the abrasion model predicts age distributions that are statistically indistinguishable from those predicted by a no– abrasion model. However, the relative erosion rates estimated by the model largely differ from the results of a no–abrasion model (e.g., in one of the samples, relative erosion rates were 29% for Tethyan Series (TTS) and 71% for Formation II-III (FII-III) from the abrasion model against 42% and 58%, respectively, from the no-abrasion model), and are closer to those from other studies that suggest a strong correlation between modern erosion rates, tectonics and precipitation intensity in the Marsyandi watershed (e.g., one study found 22.5% for TTS and 77.5% for FII-III). These findings suggest that pebble abrasion must be accounted to avoid uncorrected erosion rates estimated from detrital studies. In the following chapter, I analyse the influence of rock types on the development of mountain topography on the young island of Corsica (Mediterranean Sea). I investigate the topographic state of both drainage network and regional drainage divide in response to post–Miocene tectonics combined with variations in sea–level. The results suggest that the northern section of the drainage divide is currently moving to the east, while the southern section is moving to the opposite direction (i.e., westward). These patterns reinforce that Corsica is currently in a transient topographic state. The analysis of the drainage network highlights both rock type and structural units as major controls on the modern river profiles. I found that knickpoints are preferentially located at a similar distance from the sea level (20 to 25 km), which could suggest a common base level drop. However, given the existence of a strong structural control on the island, caution must be taken when interpreting spatial patterns of knickpoints in terms of external trigger. I have not found any correlation between the knickpoint metrics (slope, length, magnitude and relief) and the thermochonometric age domains compiled from other workers. These findings suggest that the current topographic state of the drainage network and divide of Corsica is controlled by both rock strength and structural boundaries rather than by long–term (and long–scale) exhumation patterns. Finally, I combine the techniques developed in the two previous chapters to assess what can be reliably retrieved from detrital studies, using the Tavignano watershed in Corsica (Mediterranean Sea) as a template. There, I find that the majority of the sampled zircon types are linked to the upstream sub–alkaline granites (e.g., S, J, P and G zircon types). The U–Pb ages of mixed samples show relatively similar peaks in the uppermost and lowermost sampling sites, while the intermediate site has several missing peaks. The main explanation for these results is the very small amount of grains collected for typology and U–Pb dating (from 6 to 59 zircon grains). The absence of zircon constraints in many of the contributing sources is also a major biasing factor. Through mineral mixing modelling, I demonstrate how analytical issues such as too few dated grains and age peak overlapping affect source–to–sink analysis. I also demonstrate how mineral mixing proportion in poorly constrained settings can be equally explained by different natural bias (e.g., zircon fertility or erosion) without being able to disentangle them. These results reinforce the importance of choosing the proper mineral tracer and recognising the factors acting on it to assess what controls the stratigraphic record.