Dr Sarah Sparrow, Applications Coordinator for the climateprediction.net (CPDN) project, talks to us about having volunteers all over the globe, publishing results in near real-time and the future for citizen science at the Centre.
When did you start at the Centre and what was your first role here?
I started at the centre in May 2015 as Applications Co-ordinator for the climateprediction.net distributive computing project.
What is your background?
I have a doctorate in atmospheric physics from the University of Oxford. Following my doctoral training I worked in the IT industry on business management systems and as a post-doctoral research scientist looking at drivers of atmospheric variability. Whilst working at the Environmental Change Institute as a scientist for climateprediction.net, I was involved with identifying how the risk of an observed extreme weather event (such as drought or flooding) changes under climate change, and how large ensembles of climate models can be used to identify new model configurations that are capable of capturing recent climate.
Summarise the research you are doing / your research interests in a few sentences.
Climateprediction.net is a citizen science project where idle time on people's home computers is used to perform climate model simulations. This enables us to generate super-ensembles of climate model data that would otherwise be too expensive to run. The very large ensembles generated by climateprediction.net can be used in a variety of ways, for example: attributing human influence on extreme weather events, finding which model settings can best represent recent climate and identifying settings that can improve ensemble forecasts. My role is to build, deploy, test and release all new applications (climate models) for the project; set up any new regions; monitor the progress of distributed simulations; and work with scientists and volunteer moderators to resolve any issues that might arise.
Recently I have produced updated sulphur dioxide emissions files for a project looking at extreme events in China, where aerosol concentrations may be important. I have also been altering our setup (in collaboration with the Met Office) to allow different vegetative model configurations for a project looking at forest dieback in the north western United States. To derive most value from the simulations performed it is necessary to ensure that the data is properly curated and usable to all in the wider community. To facilitate this we have recently been working on several data publications.
Why is this important (to the scientific community / the world at large)?
Climateprediction.net uniquely enables multi-thousand member simulations of climate models (around 1000 times the size of typical large ensembles). Improving the statistics in this way allows better characterization of extreme weather events by the scientific community.
Not only does climateprediction.net engage the public in scientific research but it also addresses very real questions about the effect of climate change on extreme weather events. This is relevant to local communities, governments, public health institutions and insurance industries.
What would you like to do next, funding permitting?
I would like to update and diversify the models that we use for distributive computing (not necessarily limited to climate models) as well as build upon the existing public engagement that we have within the project.
Are you involved in any wider collaborations? Why are these important?
Not only do we have volunteers from all over the globe but also a wide network of collaborators. I am currently actively involved in a number of projects with both national collaborators (Oxford, Met Office, Reading, Edinburgh) and international collaborators (France, Germany, India, China, Japan, USA).
I also manage simulations for projects led by the scientific arm of climateprediction.net (based at the Environmental Change Institute in Oxford) with collaborators in North Africa, Mexico, Dubai, Brazil, Australia and New Zealand. These collaborations are important not only to develop and diversify the scientific questions that we address but also to promote our research in different countries and maintain volunteer numbers for the project.
What publication /paper are you most proud of and why?
Schaller et al, (2016) Human influence on climate in the 2014 southern England winter floods and their impacts, Nature Climate Change, doi:10.1038/nclimate2927.
This paper demonstrates in many ways the capabilities of climateprediction.net. It resulted from simulations where we set up, analysed and published initial results live on our website and via The Guardian newspaper as they were returned by volunteers - all whilst the event was still occurring. This gave a near real-time picture of the event and enabled us to maintain public interest in our research. Not only was the weather@home setup used, where a high resolution regional model is nested into a coarser resolution global model, but also the output was fed into hydrological and economic impact models to produce more policy relevant output.
What do you think the most important issues/challenges in your field will be in the next decade and how is the Centre placed to address them?
Currently we rely on a large number of users donating idle time on their home computers to perform simulations. As such it is important to maintain their interest in the project. With the rise in popularity of tablets and other more portable devices the number of home computers capable of running climate models is expected to decline. Therefore diversification of the project to enable efficient computation in a variety of ways (cloud computing, use of graphics cards), whilst ensuring good visualisation techniques and data management, will be essential. The Centre has existing expertise in these areas, which will help to address these issues in a timely fashion.
What do you think the Centre does best?
Using technology to enable collaborative research with diverse research communities.
See also the climateprediction.net latest findings on the Paris floods.