The new dataset has been formed using computing resources from model simulations gathered across hundreds of volunteer’s computers around the world. The outputs offer new insights into the risk of possible strong droughts in future climates and will enable us to learn how these droughts could be managed through a risk-based modelling approach.
The application of risk-based approaches to manage extreme events require large sets of climate time series, but conventional climate projections typically provide only a small number of present and future climate model simulations. Using the computing resources of volunteers’ computers around the world to run climate models within weather@home, a new set of climate projections have been developed that provide 100 time series for the recent past (1900-2006) and for two future time slices (2030s and 2080s).
"These large sets of climate time series, which sample both climate model uncertainty through sea surface temperature and natural variability from the atmosphere through many simulations, allow us to approach the analysis and management of extreme events such as droughts from a new angle", says Dr Benoit Guillod, who developed this dataset at Oxford University’s Environmental Change Institute and since then moved to ETH Zurich in Switzerland.
This effort is part of the MaRIUS project (Managing the Risks, Impacts and Uncertainties of drought and water Scarcity), which focuses on drought. Benoit Guillod emphasises that "hydrologists and impact modellers from MaRIUS, are currently running their models based on these new projections. Ultimately, we will learn more about how droughts can be managed in the UK today and in the future". In addition to the publication describing the dataset, two publications using these have investigated low-flow frequency and discuss the treatment of risk and robustness under uncertainty.
The new projections suggest similar changes as previous datasets, such as UKCP09, but with a stronger drying signal. "We do not claim that the strong drying will necessarily occur; however, together with the large number of projected time series, our dataset is ideal to investigate the robustness and preparedness of society to possible strong droughts in the future".