Understanding human induced climate change though the use of numerical models. My research focuses on investigating
  • how the Antarctic sheet will respond to climate change and the consequence for sea level rise
  • the inter-continental transport of desert dust
  • the mechanisms that control the quantity of desert dust in the atmosphere on decadal time scales
  • the interaction between climate and vegetation cover

    • Ice sheet modelling

    I am currently working on the Ice2sea project . This is an EU FP7 project which aims to provide estimates of sea level rise for the next 200 years. My role in this project is to assess the contribution of the Antarctic ice sheet to sea level rise. For this purpose I am using the glimmer-cism community ice sheet model which includes longitudinal stress gradients using the scheme by Pattyn (2003). There are three parts to the project. Initialisation, validation and simulation of the future ice sheet. In the initialisation step, we obtain a set of initial conditions for ice geometry, temperature and basal properties which match the present day ice sheet. The model is then validated against remote sensing observations of elevation changes to ensure its robustness. Finally, the response of the ice sheet to two possible climate scenarios will be tested. These scenarios are the A1B, in which greenhouse gases continue to increase at current rates and the E1 scenario in which carbon dioxide remains fixed after the second half of this century. This work will contribute to the ice2sea model inter-comparison. The inter-comparison will collate the results of similar modelling studies on Antarctica and Greenland by groups within ice2sea, to provide a range of estimates for sea level rise.

    Dust cycle modelling

    Previously I worked with the GREENCYCLES Marie Curie Research Training Network. I developed a dust cycle model that included dynamic vegetation and used this to test if vegetation changes could explain decadal variability in the atmospheric dust loading. The model consisted of the LPJ dynamic vegetation model, a dust emissions scheme to represent saltation and sandblasting and the TOMCAT chemical transport model. Threshold parameters used to calculate surface emissions were tuned using Latin Hypercube sampling. The effectiveness of three sub-cloud scavenging schemes were also tested. Dust emissions which agreed best with deposition and surface concentration observations ranged from 1600 to 2400 Mtyr-1.