- Email: email@example.com
- Thesis title: Statistical methods to quantify and visualise the complex behaviour of clouds in the climate system
- Supervisors: Professor Ken Carslaw, Dr Leighton Regayre, Dr Lindsey Lee (AMRC), Dr Jill Johnson (University of Sheffield)
I am currently doing a PhD in the Institute for Climate and Atmospheric Science and I am an active member of the Atmospheric Chemistry and Aerosols research group. In 2018, I completed an integrated MPhys degree in Edinburgh at Heriot-Watt University in mathematical physics, with research projects in classical and quantum optics in Dr Jonathan Leach’s group. Modeling work piqued my interests when I took a class in mathematical biology and it combined perfectly with my passion for climate to lead me to atmospheric physics at Leeds.
My research uses statistical emulation of high–resolution cloud simulations to quantify and visualise shallow cloud responses to changing aerosol and environmental conditions, in order to better understand aerosol-cloud interaction radiative forcing. Aerosol-cloud interactions are the largest source of uncertainty in estimates of radiative forcing since the industrial period. Some individual cloud responses to changing aerosol conditions are well understood, yet when combined with each other and environmental factors the outcome can be very difficult to predict and understand. One-at-a-time testing cannot capture such complex behaviour, but computational expense limits the number of simulations that can be run. This is especially true in shallow, warm cloud systems, which form and develop on small timescales, rapidly adjusting to changes, and are therefore difficult to simulate. High-resolution cloud-resolving models are used to study cloud processes and formulate parameterisations for use in general circulation models. Using statistical emulation, aspects of the high-resolution model output can be approximated and sampled millions of times at reduced computational cost compared with the model itself. With this method, we can feasibly study the tangled responses of clouds to many changing conditions and visualise their complex behaviour.
So far, I have produced a demonstration of how statistical emulation can be applied to a two-dimensional case. That is, where two parameters, the potential temperature inversion jump and the specific humidity inversion jump, have been perturbed and the effect on the cloud produced is studied by analysing the liquid water path after eight hours of simulation. We found, in line with other work, that there are two clear behavioural regimes with a smooth transition between them. Under warm, moist conditions, the cloud produced is thick and drizzling, which cools and stabilises the subcloud layer allowing the cloud to expand below. Under cool, dry conditions, there is evidence for strong entrainment and evaporation at cloud top, leading to thinning and in some cases dissipation of the cloud. An overview of this work is published in a student conference edition of the RMetS Weather journal (https://doi.org/10.1002/wea.4001).
I am also interested in using statistical emulation when the transition between behavioural regimes is not so smooth, for example, if there is a discontinuity present. It is difficult to emulate such a case using the standard method, but it can be adapted in a variety of ways to overcome the sharp behaviour. I am currently creating a simulation of a stratocumulus-to-cumulus transition where we expect to find a sharp transition in the perturbed aerosol parameter so that we may test the adapted methods of emulation to study the key processes behind the behaviour.
- MPhys Mathematical Physics
Research groups and institutes
- Institute for Climate and Atmospheric Science
- Atmospheric Chemistry and Aerosols