Share on Facebook Tweet on Twitter Share on LinkedIn Share by email
Drew Purves

I am head of the Computational Ecology and Environmental Science Group (CEES), a part of the Computational Science Lab at Microsoft Research Cambridge. The goal of CEES is to develop predictive models of ecological systems, by inventing and applying new models and new scientific software tools. [CEES main page][CEES handout (pdf)]


I studied ecology at Cambridge University, did a PhD in ecological modelling at the University of York (UK, working under Prof. Richard Law), and spent nearly 6 years as a postdoc in the EEB Department at Princeton University (working under Prof. Stephen Pacala), before joining MSR Cambridge in 2007. My research has led to over 30 publications in peer-reviewed journals including Science, PNAS, Proc Roy Soc B, Global Change Biology, Ecology, Ecological Monographs and Ecology Letters. I co-supervise several PhD students at European universities (see below), and since 2008 have been an affiliate lecturer at Cambridge University. I am currently the treasurer of the British Ecological Society. Since the education system is currently a hot topic in the UK, I've decided to start pointing out that I did my GCSEs and A-levels at Beauchamp College, Leicester, which was a state comprehensive at the time, and is now a 'maintained school', which I believe is exactly the same thing. Make of that what you will!


Ecology is the science of interacting organisms, often described more dryly as 'the study of how the distribution and abundance of organisms follows from their interactions with other, and with the environment'. Therefore, ecology is crucial to addressing the many global challenges that we now face, including sustainable agriculture, fisheries and wood production; conserving biodiversity and ecosystem function; and predicting and mitigating disease outbreaks.

Within this broad definition, my own research spans many questions, taxa and scales, ranging from studies of the growth and development of small plants measured over a few days, through studies of how the continental-scale geographical distributions of tree species emerge over timescales of centuries; to global-scale models of carbon, biodiversity, and ecosystem function. However, this otherwise broad research 'programme' is united by an insistence on adopting a 'joined up' approach to ecology -- marrying models, with data, via comptutational statistics, in order to provide defensible, believeable models of ecological phenomena. Developing such models both increases our understanding of nature, and provides the ability to manage it better. Here at Microsoft I have a green light to pursue this research agenda, whilst packaging up the various novel software that it requires into reuseable tools that allow other scientists to more easily adopt the same joined up approach to ecological modelling.

Some example projects that are particularly dear to me at present: CCF1.0, the world's first fully data constrained model of the terrestrial carbon cycle (with Matthew Smith); The Madingley Model, the world's first global process-based model of global ecosystem function (think, simulating how all the eating, moving, growing, dying, and reproduction of all the world's animals and plants, somehow adds up to the biosphere as we know it!: with Tim Newbold, Mike Harfoot and Derek Tittensor); understanding geographic distributions of species (with Greg McInerny and Mark Vanderwel); understanding the underlying program that guides the growth and development of plants (with Lindsay Turnbull and Camille Guilbaud); and predictive modelling of global agriculture (with Paul Palmer).

Tools: including FetchClimate and Filzbach

A principle goal of the CEES group that I lead is to invent new scientific software to enable predictive modelling of ecological systems, then package this software into reuseable tools [read more here]. I have been involved in many experimental prototype tools since joining Microsoft Research, several of which can be found on our new tools site. Two relatively mature tools that I have been deeply involved with are FetchClimate, an easy, intelligent service for providing climate information; and Filzbach, a fast, robust, flexible adaptive MCMC library for parameterizing complex models against heterogeneous data. [try FetchClimate online][try Filzbach online][new tools site]

 Purves' current PhD students (in order of appearance):

Emily Lines (University of Cambridge)
Corresponding CEES project:  
Forest Dynamics
Silvia Caldararu (University of Edinburgh)
Corresponding CEES project: 
Forest Dynamics
Nikee Groot (University of Leeds) 
Corresponding CEES project: 
Forest Dynamics
Sadia Ahmed (Imperial College London)
Corresponding CEES project: 
Predictive Modelling of Tropical Deforestation 
Isabela Rosa (Imperial College London)
Corresponding CEES project: 
Predictive Modelling of Tropical Deforestation  
Camille Guilbaurd (University of Zurich)
Corresponding CEES project: 
Data-constrained Modelling of Plant Growth


Publications (since April 2007)
--- Contact ---
Microsoft Research,
21 Station Road,
Cambridge, CB1 2FB, UK


--- Latest News ---
Our Comment piece in Nature: Time to Model All Life on Earth 
Tim & Drew's latest (in Proc Roy Soc): how bird traits affect response to land-use change
Drew Purves one of 40 scientists worldwide invited to attend the World Economic Forum 'Summer Davos' 2012
New Madingley Model website!
Latest FetchClimate story
Check out mine and Raul's SPOM demo in html5! [try online]
FetchClimate highlighted at Techfest 2012 [video][Rick Rashid's keynote][next at Microsoft]
Emily Lines has handed in her PhD thesis. Well done Emily! [Emily's paper on tree mortality]
FetchClimate is live! [try online] [read more] 
New paper with Silvia Caldararu: 'Intelligent' Amazon trees track light [read more]
Drew Purves listed in Wired 2012 smart list [read more]
Tools site updated [go there now]
Drew Purves speaks alongside John Beddington (UK chief govt scientific advisor) on measuring sustainability [youtube]