PhD projects
Update May 2012:
This page lists PhD Topics that are being considered for starting in or around October 2012. Some of these projects already have guaranteed funding (for University fees and living costs) for students from particular backgrounds (as indicated below) while others are possibly available for students with their own funding. For further information on applying please see (link to http://www.met.reading. ac.uk/pg-research/pgrapplications.html ) or email phdinfo@met.reading.ac.uk
For students with their own funding, other projects can be made available, following discussions with suitable supervisors. Please contact phdinfo@met.reading.ac.uk for more information.
THE FOLLOWING PROJECTS CURRENTLY HAVE FUNDING AVAILABLE
The generation of downwind rainbands by mountains
Supervisors: Dr Suzanne Gray (Univ. Reading), Dr David Schultz (Univ. Manchester) and Dr Dan Kirshbaum (McGill University)
Full funding available to UK students and eligible non-UK EU students
Flash floods are acute hazards with long-term social-economic consequences and it is essential that they be accurately understood and predicted. Recent devastating flash flood events include the Boscastle flood of 2004 and the Cockermouth flood of 2009. Two of the three principal mechanisms behind UK flash-flooding events are convective storms and orographic rainfall (the other being frontal systems). Although convection and orography may act independently to produce extreme rainfall, they are often closely linked over the complex UK terrain. The mechanical ascent upstream, over, and downwind of steep terrain and the thermally-driven ascent due to elevated heating lead to storms in convectively unstable airflow. Because orography is fixed in space, these storms may anchor to specific terrain features and focus their rainfall over preferred areas leading to bands of rainfall (termed rainbands). In particular, quasi-stationary rainbands, locked to the terrain, are a manifestation of orographic convection that greatly increases flood risks because they focus heavy rainfall over specific regions.
In this project we will investigate the mechanisms that lead to the formation of rainbands downwind of mountains. We will use an idealised version of the operational Met Office weather forecast model to explore potential mechanisms, exploiting the control afforded by idealised simulations.
Model improvement using data assimilation.
Supervisor: Prof Peter Jan van Leeuwen
Full funding available to UK students and eligible non-UK EU students
Model improvement is typically done on an ad-hoc basis, where certain model parameters are changed and different model runs are compared to observations. However, a much more systematic way for model improvement is available using data assimilation. Data assimilation allows one to determine the source of the difference between model estimates and observations, instead of just telling they are different.
In this project we will study the efficiency of this new use of data assimilation. The basic structure of the PhD project is to explore data assimilation for model improvement in simple systems initially, like the Lorenz 1963 model, building up to more complex models, like multi-layer shallow-water models. When successful this opens an exciting new field of model improvement that is likely to become main stream in model development in atmospheric sciences and beyond.
The PhD project is part of a larger project on data assimilation for climate models involving postdocs and PhD students in the data-assimilation group here in Reading, and strong collaboration with climate modellers from e.g. NCAS.
Plant processes and anthropogenic climate change
Supervisors : Emily Black (Meteorology), Anne Verhoef (Geography and Environmental Science), Eleanor Blyth (Centre for Ecology and Hydrology) and Martin Best (Met Office)
Full funding available to UK students and eligible non-UK EU students
In the future, under anthropogenic climate change, changes in the weather will worsen the water stress that plants suffer - potentially reducing photosynthesis and growth. On the other hand, the physiological effects of higher concentrations of CO2 in the atmosphere will reduce transpiration and hence increase plants' resilience to water stress. Simulating the net impact of these competing effects is a tough challenge for our current land-surface models. As well as determining plants' response to water stress, transpiration is a crucial component of the water cycle - and hence of regional climate. Inaccurate simulation of transpiration may therefore increase the uncertainties in projections of regional climate change. This project aims to quantify and reduce these uncertainties in the Hadley Centre models. Although the results will have global implications, the project will focus on contrasting case study regions in West Africa and Europe.
Predictability of the opening of Arctic sea routes
Supervisors: Dr Ed Hawkins, Prof Keith Haines and Dr Dan Hodson
Full funding available to UK students and eligible non-UK EU students
A project studentship is available to examine the predictability of the opening of the Arctic sea routes – an important aspect of Arctic climate which is of particular importance to industry and decision makers. Arctic sea routes are important to such stakeholders for several reasons: (i) passage of ships on shorter routes between the Atlantic and Pacific, (ii) access to reserves of minerals, oil etc., and (iii) the potential for conflict over access to, and ownership of, the sea routes. Although the main focus of the studentship will be on examining the potential for predictions of the opening of Arctic sea routes on seasonal to inter-annual timescales, it will be essential to consider longer term climate projections as well. This project will form part of a larger Arctic research programme funded by NERC.
Cloud-radiation parameterisations based on 3D radiative transfer and sensor synergy
Supervisors: Dr Christine Chiu and Prof Robin Hogan
Full funding available to UK students and non-UK EU students, but they must not have resided in the UK for more than 12 months in the last 3 years
The response of clouds to climate change is one of the major sources of uncertainty in climate prediction. This project aims to improve cloud-radiation parameterisations in climate models, which currently do not treat the effect of the 3-D nature of clouds on solar and infrared fluxes. The student will put the 3D structures of natural clouds into a detailed 3D radiative transfer model, using scanning cloud radar measurements at the Jülich Observatory for Cloud Evolution (JOYCE). Once the 3D cloud effects at the field site are understood, the student will work on improving radiation calculations in current climate models, building on recent work at Reading to devise a new scheme for incorporating 3D effects into the 1D radiation scheme used by the UK Met Office. Finally, the global effect of 3D radiative transport and its implication on climate change prediction will be investigated.
THE FOLLOWING PROJECTS MAY BE AVAILABLE FOR STUDENTS WITH THEIR OWN FUNDING
Speeding up the calculation of physics in climate models
Supervisor: Bryan Lawrence
Simulations of climate can take a long time – months to years of real time – and not only is it boring to wait for the output, it's an expensive business! Future computers are not going to have faster processors - so speeding up simulations will depend on better algorithms which either scale across more processors better, or which have been coded more efficiently. The trouble with more efficient code is that it's often harder to understand and verify and even harder to evolve as the scientific understanding improves. The key question that this project will address is: What prospect is there for using "domain specific languages" in climate models to hide efficient computational code from the modellers – allowing climate models to be both faster and easier to modify. The student taking on this project will start with some simple physics code (e.g. a piece of the radiation library) and abstract out the key scientific principles before trying to provide some faster underlying code and putting it back in a climate model. There is some prior art so it wouldn't be starting from scratch, and the work is likely to be done with partners both in the UK and elsewhere.
Synchronisation in data assimilation
Supervisor: Prof Peter Jan van Leeuwen
Numerical weather prediction is unthinkable without data assimilation. Data assimilation is a tool to combine information from the numerical weather prediction model with atmospheric observations. Present-day data assimilation methods are all based on linearisations, while the data-assimilation problem is becoming more and more nonlinear due to increasing resolution and more complex observations. Operational centres like the Met Office and ECMWF are very keen on moving into the nonlinear regime. In this project we will explore an exciting new way of looking at the problem. The model and the real atmosphere are seen as two separately evolving systems, and information flows from the real atmosphere to the model through the observations. So the observations couple the two systems, and try to 'synchronise' the model to the real atmosphere. Formulated in this way, we can explore ideas from synchronisation theory, and no restrictions on nonlinearity come into play. Starting with simple atmospheric models, like the Lorentz models and a 2-D model of the atmosphere, we will explore the possibilities of this approach. Especially, huge progress is expected when this approach is combined with the extremely efficient particle filters we are developing for much more complex models of the atmosphere. The emphasis can be changed to the ocean if the student has strong interest in that area.
Optimal assimilation of retrievals from hyperspectral infra-red satellite sounders
Supervisors: Dr S. Migliorini and Dr J. Eyre (Met Office)
Satellite measurements provide an essential source of information to operational meteorological centres, which rely on these data to provide useful weather predictions to the public. A major difficulty with remotely-sensed observations is that their resolution is usually not high enough to constrain all the components of the state that is evolved by the atmospheric model. This has led operational centres to compare model forecasts and measurements in observation space, with the aim of finding better initial conditions that fit the available observations, a process called data assimilation. Although this strategy has been very successful, it has also significantly increased the complexity of the assimilation process, particularly when high-spectral-resolution (denoted as hyperspectral) sensors are considered. This proposed PhD project is focused on exploring the theoretical and practical implications of recent advances in our understanding of the satellite remote sensing inverse problem, which show that alternatives to current radiance assimilation methodologies are possible. This research is of significant practical importance, and will benefit from the collaboration of the Met Office, which we hope will sponsor the project.
What are the physical processes controlling the existence of a bistable overturning circulation and climate?
Supervisors: Dr Remi Tailleux and Dr Robin Smith
On the basis of a hierarchy of ocean models of various complexity, itis widely thought that the ocean meridional overturning circulation (MOC) could possess at least two stable equilibria: an "on" state associated with a significant amount of poleward ocean heat transport (corresponding to current climate conditions), and an "off" state associated with very weak MOC and heat transport which could be associated with significant colder climate conditions in the northern hemisphere. The question of MOC stability is associated with abrupt climate changes that are known to have occurred in the past, and is also relevant for our future climate, as the heat and freshwater cycles that help drive the modern MOC are disrupted. The existence of MOC bistability has, however, not yet been demonstrated in reality or in the most realistic complex global coupled climate models. A key issue, therefore, is to understand whether MOC bistability could occur in the real world, or whether it is an artefact seen only in some models, arising from oversimplifying the physics of ocean/climate interactions. To make progress, a physical understanding of the processes responsible for the existence of bistability is needed. The purpose of this PhD will be to explore a number of avenues to investigate the physical processes that are key for the existence of a bistable MOC and climate, and which are required to evaluate the probability of abrupt climate change in the future.
Using polarization measurements to evaluate GCMs and light scattering models of cirrus
Supervisors: Dr Christine Chiu, Dr Anthony Baran (Met Office) and Prof Keith Shine
Recent studies have revealed a shocking model-to-model discrepancy in cloud ice content among the general circulation model (GCM) simulations contributed to the IPCC report, which highlights our insufficient knowledge of ice clouds. This leads to erro rs in both weather and climate predictions, because ice clouds not only strongly couple with weather systems via synoptic motions and deep convections, but also strongly influence global climate via their radiative impact. This project aims to fill such critical knowledge gap and reduce prediction errors using polarization measurements that are severely under-utilised.
This project includes two parts. Firstly, the project involves climate model evaluations in ice clouds using global polarization measurements from PARASOL (Polarization and Anisotropy of Reflectances for Atmospheric Sciences coupled with Observations from Lidar). Secondly, a new optimal method will be developed to globally estimate cirrus cloud properties by synergizing satellite observations from PARASOL, Caliop (lidar) and MODIS (a visible and infrared imager). Specifically, the student will:
- Retrieve global distributions of ice water content for semi-transparent cirrus (through collaboration with the University of Lille, France).
- Develop a new metric that can be used to evaluate GCM prediction in cloud phase, altitude, temperature, and corresponding shortwave flux.
- Develop a new parameterization for the Met Office suite of models by exploring the linkage between cirrus ice crystal randomization, relative humidity, temperature and ice water content.
Results from this project are expected to incorporate into the Met Office suite of models, and will greatly help reduce uncertainty in cirrus predictions for the climate model community.
Simulation in the terra incognita
Supervisor: Dr Bob Plant
As computer resources continue to increase, weather forecast and climate models continue to be run at finer and finer resolutions, capturing more and more detail and helping them to become more and more accurate. Which is a good thing. Only there is st arting to become a but to all of this. As the resolution gets finer, the scales at which forecast models operate are starting to approach the scales associated with turbulent motions near to the earth's surface. The current ways of handling those motions break down and fundamentally new modelling approaches become necessary. What are those approaches? Well, we don't really know: that's why it's called the "terra incognita" or sometimes the "grey zone". The models are getting there soon, and they will have to do something, so we'd better start figuring out what.
Investigating the microphysics and dynamics of mixed-phase clouds
Supervisors: Dr Chris Westbrook and Prof Anthony Illingworth
Clouds are currently the biggest cause of uncertainty in our predictions of future climate. Recent work here at Reading has made it increasingly clear that "mixed-phase" clouds (those containing both ice particles and liquid water droplets) are particu larly important, both for development of precipitation, and also for their impact on the earth's radiation budget. Yet mixed-phase clouds remain one of the most poorly understood, and poorly simulated cloud types in numerical weather and climate models. N umerical models predict such clouds should dissipate in ~30 minutes, yet in nature we observe them persisting for hours or even days. How does this occur?
This project will attempt to answer these questions through analysis of state-of-the-art observations of the microphysical evolution of the droplets and crystals as well as the dynamical air motions and radiative processes which may contribute to their persistence. These observations will come from the dual-polarisation Doppler radars and lidars at the Chilbolton Observatory (www.stfc.ac.uk/chilbolton), in tandem with data collected in the clouds themselve s using the probes onboard the FAAM 146 aircraft (www.faam.ac.uk).