Recovering uplift histories from noisy landscapes
File(s)
Author(s)
Morris, Matthew
Type
Thesis or dissertation
Abstract
The shape of modern-day topography is determined at large scales by its tectonic and erosional history. However at smaller scales, other processes—climatic or lithologic variability for instance—that can be considered as geomorphic noise, can determine landscape geometries. This presents a challenge when using topography to recover tectonic information such as uplift rate histories. In this dissertation I seek to overcome this challenge by developing an inverse modelling framework to recover uplift histories from topography generated in the presence of geomorphic noise. I first demonstrate how noise added to models of landscape evolution produces complexity at local scales and variability in geomorphic properties at larger scales. By generating ensembles of models, where noise is incorporated using strategies informed by observations of Earth’s topography, I show how uncertainties in predicted landscape geometries and derived metrics can be quantified. In Chapter 3 I then expand upon the challenge of addressing complexity at local scales in the context of inverse modelling. When comparing topography using existing Euclidean approaches, small differences in noisy elevations that are initially inserted into landscape evolution simulators generate resultant landscapes with contrasting drainage planforms. These differences in local elevation hamper the recovery of tectonics from topography. Instead, I leverage techniques from Optimal Transport theory, namely the Wasserstein distance, to develop a method that ‘sees through’ local complexity to recover uplift rates from topography. In Chapter 4 I then demonstrate how such approaches can be used to recover uplift rate histories from noisy synthetic landscapes. Calibrated Wasserstein-based objective functions successfully recover rates of uplift that vary in space and time. Uncertainties in calculated values are quantified via a probabilistic inverse modelling framework. The success of this approach suggests that future attempts to invert real topography with appropriately calibrated models can yield meaningful information about tectonic
histories on geologic timescales.
histories on geologic timescales.
Version
Open Access
Date Issued
2026-02-25
Date Awarded
2026-06-01
Copyright Statement
Attribution-NonCommercial 4.0 International Licence (CC BY-NC)
License URL
Advisor
Roberts, Gareth
Richards, Fred
Lipp, Alex
Sponsor
Engineering and Physical Sciences Research Council
Grant Number
EP/W524323/1
Publisher Department
Department of Earth Science & Engineering
Publisher Institution
Imperial College London
Qualification Level
Doctoral
Qualification Name
Doctor of Philosophy (PhD)