Project Highlights:

  • A truly broad, interplanetary PhD opportunity. It will quantify dune form and pattern throughout the largest dunefields on Earth and Mars
  • High level expertise to be gained in a wide range of remote sensing and geospatial/topographic analysis techniques and climate datasets, underpinned with desert fieldwork
  • New insights into sand seas as systems through application of cutting edge techniques to build global catalogues of individual dunes

Wind-blown sand dunes cover around 10% of Earth’s land surface, and are widespread on Mars. One exciting yet under-studied way of researching major dunefields relates to investigating the patterning (i.e. order) within large dune assemblages, where fascinating self-organisation of bedforms can be seen over tens of thousands of square kilometres. In these sand bodies, dunes demonstrate a characteristic replication of bedforms, but key questions remain surrounding the nature of these arrangements. For instance, what can dune crest patterns tell us about the state of the dune system [e.g. Ewing et al., 2006]? How do dunes organise into relationships between height and spacing [Lancaster 1988; Baddock et al., 2007]?

Central to the difficulty of studying dune systems at this largest scale is their inaccessibility. A series of landmark advances for dune research came with the advent of environmental satellite coverage in the 1970s, and the first synoptic views across the largest dune regions. Today, a wealth of new high-resolution remote sensing and digital topographic data offers a fresh impetus for the understanding of dunefields and their behaviour. This PhD will harness a wide range of geospatial and climate datasets to produce novel ways of looking at, and statistically characterising dunefields. The aim is to offer new insights into the construction and self-organisation of dune systems in sand seas. Adopting two algorithms developed from other areas of geoscience (SWT and MiMIC [Hillier & Watts, 2004; Hillier, 2008]), the project will apply these to allow automatic identification and morphometric analysis of dunes (e.g. spacing, height, asymmetry) without user subjectivity. A recent pilot study demonstrates that SWT performs as well as manually mapping simple linear dunes. This opens the exciting possibility of larger scale and ultimately global quantifications of dune form. The PhD will also draw upon recent advances in understanding other bedforms (e.g. glacial) via modelling of their sizes [Hillier et al., 2013; 2016].


Improvements to the way we define dune patterns and distributions, and how these patterns link to climatic drivers on Earth, offers a path to help our growing understanding of other planetary environments e.g., Mars [Ewing et al., 2010]. A profound step forward is possible through application of new geospatial approaches, validated on Earth, to extra-terrestrial settings.

A) Linear dunes of the central Namib Sand Sea, B) An 80 x 90 km portion of the ASTER Global Digital Elevation Model (GDEM) showing linear dunes in the Rub Al Khali, Saudi Arabia, C) Example output from an (un-optimised) run of the SWT dune identification and quantification algorithm [Hillier, 2008] for a 10 km transect of the GDEM.


This PhD will utilise a diverse range of geospatial, topographic and meteorological/climate data. Analysis of dune form based on satellite imagery and digital elevation models (DEMs) will be a core activity. Given the large spatial areas this study will cover, automating the analysis of dune landscapes is an important requirement. Full verification, and optimisation of the algorithms for the analysis of dunes is crucial. A key part of verification will be achieved by a field campaign in a major sand dune desert e.g. the Namib Sand Sea. Techniques such as detailed topographic survey, including DGPS, Structure-from-Motion and possibly drone imagery will be employed for algorithm ground truthing. Simple dune form statistics and size-distributions of morphometrics will be used to achieve additional landscape understanding from high resolution satellite imagery. Dune statistics will be investigated in the context of meteorological data, including climate reanalysis, which will define the characteristic wind-fields across dune regions, as a primary influence on dune type.

Training and Skills

This project will forge high-level skills in handling, organising and analysing large geospatial and climate datasets. Training will be provided in remote sensing data collection and processing, as well as necessary GIS techniques. To gain on-the-ground experience of dune landscapes, the student will have the opportunity to conduct fieldwork in dune settings, such as Namibia. For the algorithms, training will be provided in coding if required. Numerical description of dune pattern and distribution will also lead to the development of strong skills in statistics (e.g. in R).


Year 1: The student will familiarise themselves with and use two feature-detection algorithms (i.e. SWT and MiMIC [Hillier & Watts, 2004; Hillier, 2008]), and validate them for the purpose of aeolian dunefield analysis. A dune statistics database for selected test dunefields, leading to identification of the field testing location, will be built toward evaluation of algorithm performance.

Year 2: Desert fieldwork to ground truth algorithm output followed by broadening of the global dune dataset. Expansion of the algorithm to the challenge of recognising more complex dune types and its use within more variable dunefields. Introduction of the algorithm to dunefields of Mars.

Year 3: Synthesis of a statistically robust inventory of global dune pattern and distribution, coupled with wind fields and quantifications of sediment availability, for Earth and Mars.

Partners and collaboration (including CASE)

The external international partner for this project is Dr Ryan Ewing, Texas A&M University, USA. He is a leading scholar of Earth and Martian dunes and a pioneer of the use of dune pattern parameters for exploring dunefield-scale aeolian systems. There will be opportunity to network internationally by visiting and working with Dr Ewing in Texas.


Further Details

For information about this project, please contact Dr Matthew Baddock (m.c.baddock@lboro.ac.uk)