Microsoft-funded PhDs in Computational Ecology
Applications are invited for the following PhD projects. These full time, three-year positions are fully funded by Microsoft Research and come with supplementary resources for travel and other expenses, plus a laptop. All students will be located primarily at the university listed with the project. However, students will be co-supervised by a scientist from the Computational Ecology and Environmental Science Group, Microsoft Research Cambridge (UK), and so are likely to make regular visits to Cambridge. In addition, students will have the opportunity to apply for one or more paid internships at Microsoft Research Cambridge, to be carried out during the course of the PhD. Additional information about each project is given below. For further information regarding any project, please email any of the supervisors listed below. Please note that, as a result of differential tuition fees, projects at UK universities are open to EU / UK applicants only.
Unifying food web structure and dynamics with metabolic theory: a general, modular computational approach
Supervisor: Dr Daniel Reuman, Imperial College, London (d.reuman@imperial.ac.uk)
Location: Silwood Park, Imperial College, London
MSR Supervisor: Dr Rich Williams (ricw@microsoft.com)
The other half of the equation: global variation in tree mortality
Supervisor: Prof. Oliver Phillips (o.phillips@leeds.ac.uk), Prof. Simon Lewis (s.l.lewis@leeds.ac.uk), Dr Emanuel Gloor (e.gloor@leeds.ac.uk), University of Leeds
Location: School of Geography, University of Leeds
MSR supervisor: Dr Drew Purves (dpurves@microsoft.com)
A data-constrained predictive model of tropical deforestation and resultant carbon emissions (2 positions)
Supervisor: Dr Rob Ewers, Imperial College, London (r.ewers@imperial.ac.uk)
Location: Silwood Park, Imperial College, London
MSR Supervisor: Dr Drew Purves (dpurves@nmicrosoft.com)
Funding note: these 2 positions are co-funded by Microsoft Research and the Grantham Institute for Climate Change, an Institute of Imperial College.
Data-constrained modelling of plant growth
Supervisor: Dr Lindsay Turnbull, University of Zurich (lindsayt@uwinst.uzh.ch)
Location: Institute of Environmental Sciences, University of Zurich
MSR Supervisor: Dr Drew Purves (dpurves@microsoft.com)
Modelling global plant biodiversity
Supervisors: Dr David Coomes, Cambridge University (dac18@plantsci.cam.ac.uk) & Dr Neil Brummitt, Royal Botanic Gardens, Kew (n.brummitt@kew.org)
Location: Cambridge University
MSR Supervisor: Dr Matthew Smith (mattsmi@microsoft.com)
Pattern and process at forest boundaries
Supervisor: Prof. Kevin Gaston, Univ. Sheffield (k.j.gaston@sheffield.ac.uk)
MSR Supervisors: Dr Greg McInerny and Dr Drew Purves (gregmci@microsoft.com; dpurves@microsoft.com)
The form and function of dynamic sociality in a wild bird population
Supervisors: Prof. B Sheldon (ben.sheldon@zoo.ox.ac.uk), Dr T. Wilkin (teddy.wilkin@zoo.ox.ac.uk), Prof. S Roberts (stephen.roberts@some.ox.ac.uk), University of Oxford
Location: Dept. Zoology, University of Oxford
MSR Supervisor: Dr Robin Freeman (robin.freeman@live.co.uk)
Project title: Unifying food web structure and dynamics with metabolic theory: a general, modular computational approach
Supervision:
In the laboratory of Daniel C. Reuman, Imperial College London
Co-supervised by Rich Williams of Microsoft Research
Funding:
This is a 3+ year fully funded full-time position, with salary at least at the NERC guidelines with London allowance, a laptop, and funds for travel and other necessary computing equipment provided by Microsoft Research.
Project description:
Because data are difficult to obtain, most past food web studies are broadly in one of two categories: 1) empirically based analyses of the average or snapshot structure of a food web in an ecosystem; and 2) theoretical analyses of the interacting population dynamics of species in the system. Structural studies have limited or no ability to predict the effects of human disturbance on food webs, since they do not consider changes in time. But models of interacting species population dynamics, which do evolve in time, are not sufficiently connected to data to make accurate predictions. As human impacts on all food webs increase, it is crucial to combine empirical-but-static with dynamic-but-theoretical research to produce empirically grounded predictive models of the dynamics of many interacting populations. The benefits of the research are expected to include progress toward: 1) an ability to inform management decisions using predictions of how impacts will ramify through food webs; 2) the ability to predict how human development projects will affect complex ecosystems; 3) hypotheses on how climate change may affect food web structure and dynamics.
The student will take a modular, computational approach to constructing and parameterising a suite of models of whole-food-web dynamics, and comparing predictions of the models with available data. Comparisons will progress from broad and qualitative to quantitative as modelling software is developed and models are refined. Data will come from ecosystems of a variety of types.
The candidate:
The project would suit candidates with computational or mathematical skills and interest, and a strong interest or background in ecological research. Demonstrated experience and ability in all three fields is an advantage but may not be required.
Informal inquiries and questions:
Please email Dan Reuman at d.reuman@imperial.ac.uk
To apply:
Please send CV, one-page statement of research experience and interests, and contact information for two references to Dan Reuman at d.reuman@imperial.ac.uk. Short-listed candidates will be contacted to arrange interviews and to request other application materials. Deadline 14 March 2009.
Project title: The other half of the equation: global variation in tree mortality
Supervisors: Prof Oliver Phillips, Dr Simon Lewis, Dr Emanuel Gloor (all Leeds), Dr Drew Purves (Microsoft Research)
Future changes in the global carbon cycle could have major impacts on climate. Whereas we have a relatively advanced ability to model the input of carbon into forests, via growth, almost nothing is known about global variation and the climate dependency of the outputs, which are dominated by tree mortality. By collating a unique, global forest inventory dataset, and analysing this using computational statistics, this project will uncover the global-scale climate dependency of tree mortality and so help predict the carbon cycle.
The project will produce important outcomes. First, it will provide the information necessary to test contrasting hypotheses about the patterns, and causes, of tree mortality globally. Tests of these hypotheses constitute novel findings, lending themselves to being written-up as exciting, high-profile scientific publications. Second, the project will produce tree mortality models for use in global carbon cycle analyses, and global vegetation models. Third, it will produce algorithms tailored to studying global variation in rates that are also subject to local control (species, tree size, and perhaps others). These algorithms will be made available in user-friendly form, to be taken up for future analyses of the carbon cycle. The student will receive state-of-the-art training in data analysis, modelling, and forest ecology, including in the tropical field. Scientists with these interdisciplinary skills are in short supply and are increasingly in demand.
The successful candidate will have an outstanding academic record in ecology and some experience of computational analysis, and be motivated to pursue a career in scientific ecology.
Application details can be found at http://www.geog.leeds.ac.uk/postgraduate/applications.html
Funding Notes
This prestigious position is supported by Microsoft Research. It is a fully funded studentship covering UK/EU fees,an annual maintenance allowance and research expenses. It draws on expertise in tropical ecology developed by our research group, and involves collaboration with researchers worldwide (see e.g., http://www.geog.leeds.ac.uk/projects/rainfor/)
Project title: A data-constrained predictive model of tropical deforestation and resultant carbon emissions (2 positions available)
Supervisors: Dr Rob Ewers (Imperial College, UK) and Dr Drew Purves (Microsoft Research Cambridge, UK)
Tropical forests store vast amounts of carbon and are the most biodiverse ecosystems on Earth, but are subject to continuing, widespread deforestation. To assess the impacts of this deforestation for future climate change, and for the persistence of biodiversity at a global scale, we require models that can predict where and when tropical deforestation is likely to occur in coming decades. To be most useful, such models need to make predictions at relatively fine spatial scales, e.g. locations within individual countries, rather than just providing regional or continental totals. These projects will use a range of computational approaches to help build such a deforestation model. The research will have substantial practical applications, and will also necessarily test a variety of important hypotheses about the socio-economic and climatic controls on deforestation: Is it true that deforestation peaks at intermediate levels of economic development? Are the main drivers at local scales similar to the main drivers at continental scales? To what extent do national differences in policy affect deforestation rates and patterns on the ground? What is the relationship between forest carbon, forest biodiversity, and vulnerability to deforestation? The projects would suit candidates with some experience in, and interest in continuing, data analysis, computation, and modeling in Earth Sciences / Social Sciences. As such, a demonstrated ability in statistics, mathematics and / or programming would be an advantage. These are three-year, fully funded, full time positions, co-funded by Microsoft Research and the Grantham Institute for Climate Change, an Institute of Imperial College.
For more information contact Drs Ewers (r.ewers@imperial.ac.uk) or Purves (dpurves@microsoft.com). Application deadline March 14th 2009.
Project title: Data-constrained modelling of plant growth
Supervisors Dr. Lindsay Turnbull (University of Zurich) and Dr. Drew Purves (Microsoft Research, UK)
How do plants grow? How do they allocate photosynthate to different structures such as leaves, roots and flowers, in different environments, and how do they make such decisions? This project seeks to understand plant growth with the simplest possible assumptions and will use extensive data sets and modelling to approach this problem from an entirely new angle. The project will involve extensive computer modelling including both simulations and fitting models directly to data. The project may also include new data collection depending on the interests and aptitudes of the successful candidate.
We are looking for a well-motivated person to conduct this research over a 3-year period, culminating in a PhD from the University of Zurich, Switzerland. Applicants need to have some experience with computer modelling and must be able to work in a team with other students and postdocs. A background in statistics and some knowledge of plant ecology is also desirable. The project is funded by Microsoft Research, UK and the successful candidate will receive extensive help and technical support from this source. The project will be based in Zurich, where the successful candidate will be expected to live. For more information contact Drs Turnbull (lindsayt@uwinst.uzh.ch) or Purves (dpurves@microsoft.com).
Review of applications will begin March 1st and continue until the position is filled.
Project title: Modelling global plant biodiversity
Supervisors:
David Coomes, Plant Sciences Department, University of Cambridge.
Neil Brummitt, Royal Botanic Gardens, Kew.
Matthew Smith, Microsoft Research Cambridge.
The overall objectives of this project are to assess the accuracy of current methodologies for predicting the global occurrence of plant species based on presence data, and to investigate the development of potentially better alternative methodologies. The appointed candidate will develop new methods for establishing the climatic and edaphic constraints of species from presence-only data, will compare separate techniques for the identification and proposal of protected areas and will test the coverage of the global protected area network for plants and identify additional priority areas for plant conservation. This will be a rare opportunity to conduct novel scientific research that could have a real impact on conservation policy.
The candidate should either have a good degree in statistics or mathematics and experience in applications to biology or should have a good degree in ecology / geography with sufficient numerical aptitude or enthusiasm to learn the desired techniques. Experience in the use of statistical software packages such as R and GIS software such as ArcView would be a particular advantage.
Applicants should fill in the graduate application pack available at the following site:
http://www.admin.cam.ac.uk/offices/gradstud/admissions/forms/graduate_application_pack_2009-10.pdf and send it, along with a one page CV, to the Graduate Secretary (Microsoft Research Studentship), Department of Plant Sciences, University of Cambridge, CB2 3EA by 1 March 2009.
Project title: Pattern and process at forest boundaries
Supervisors: Professor Kevin Gaston (University of Sheffield), Dr Greg MacInerny (Microsoft Research, Cambridge) and Dr Drew Purves (Microsoft Research, Cambridge).
Key words: Spatial ecology, computational biology, climate change
Project description:
The distribution of vegetation biomes at global scales is highly predictable from climate and soils. This implies that the distribution will shift in response to climate change, with major implications for the global carbon cycle and global biodiversity. But our ability to predict these shifts is currently limited, in part because rather little is known about the causes of boundaries between vegetation biomes. This project will: (i) identify key metrics related to the structure and dynamics of forest boundaries that might inform ecological theory and focus the development of predictive models; (ii) employ computational data analysis techniques and computational statistics to extract these metrics from regional or global data sets; (iii) use the results to test key ecological assumptions and hypotheses about the nature and causes of forest boundaries; and (iv) make this information available for use by others in the scientific community. An explicit focus will be comparing boundaries of different types (e.g. biome vs species) and driven by different external forces (e.g. forest-savanna vs timberline, or species boundaries caused by temperature vs precipitation). This interdisciplinary project would suit a highly motivated student with an interest in developing novel ideas and computational methods, and in developing scientific research at the intersection of ecology, ecophysiology, and Earth system science. A background in ecology, with some experience in computational biology, would be a major advantage. The project will be based at the University of Sheffield, and will include an internship at Microsoft Research Cambridge. For more information please contact Prof. Gaston (k.j.gaston@sheffield.ac.uk <mailto:k.j.gaston@sheffield.ac.uk>), Dr McInerny (gregmci@microsoft.com <mailto:gregmci@microsoft.com>) or Dr Purves (dpurves@microsoft.com <mailto:dpurves@microsoft.com>).
Starting date: 1 October 2009.
How to apply: Complete an on-line application form via University of Sheffield web site at http://www.shef.ac.uk/postgraduate/research/apply/index.html. Send a full CV, via email to Mrs S Carter, s.a.carter@sheffield.ac.uk, or a hard copy to Mrs S Carter, Department of Animal and Plant Sciences, University of Sheffield, Sheffield S10 2TN.
Closing date for applications: Monday 9 March 2009.
Project title: The Form and Function of Dynamic Sociality in a Wild Bird Population
Supervisors: Prof Ben Sheldon, Dr Teddy Wilkin, Prof Steve Roberts
All organisms display social behaviour of some form, but the extent and duration of this behaviour varies tremendously between species and over life cycles within species: understanding what causes variation in social behaviour has been a major research theme in biology for decades. Until recently, very little of this work has focussed on the type of social organisation that typifies many animals, where social groups are highly dynamic, with frequent changes in their composition, and where associations between individuals vary in their strength and consistency. However, there is currently great interest in applying techniques from network analysis to animal social behaviour. This project exploits a large ongoing study of a wild bird population that has been a model system in ecology and evolutionary biology, in which thousands of individuals are marked with transponders, and a grid of recording locations generates hundreds of thousands of records each winter. The main aims of the project are to use these data to generate biological insight into social behaviour in birds, in a social network context. The first aim of this project will be (1) for the student to develop methods for identifying individual groups from the complex temporally and spatially-structured data set based on feeding associations. Having done this the next aims will be to (2) explore the stability of groups over time, and (3) quantify the structure of groups in terms of the number, individual characteristics and relatedness of their constituents. Following these steps, the project will: (4) develop social networking techniques which integrate temporal changes in the strengths of relationships between individuals, and (5) determine the consequences, in terms of foraging and breeding performance, of social structure at the individual, group and population levels. Finally, (6) the project will explore how group stability changes in response to experimental manipulations of food availability, and simulated predator abundance, in the field.
The ideal candidate for this post is a physical scientist with a first degree in engineering, computer science, physics or mathematics, a strong quantitative background, and a desire to use these skills to understand complex biological problems. The supervisors are skilled in ecology and behaviour (Sheldon and Wilkin) and information engineering (Roberts) and the project will work very much at the interface of these fields, as part of two large, research active groups (see: http://www.zoo.ox.ac.uk/egi/ and http://www.robots.ox.ac.uk/~parg/). The stipend (tax free) is £16,000 per annum. Informal inquiries to Prof Ben Sheldon (ben.sheldon@zoo.ox.ac.uk).
Computational Science
Other opportunities
