News

The impacts of machine learning on climate change modelling

7 August 2017

Share

Reported by Paul Henry, BBSRC funded policy intern

What are the opportunities for developments in environmental risk assessment to inform policy decisions?

There is an increasing need for environmental data to inform policy decisions. At the same time, technical advances, including in data science, are leading to the prospect of being able to provide information about environmental risks across an increasingly broad spectrum of policy-relevant issues. CSaP's Policy Workshop on innovative climate risk assessments brought together experts from academia, policy and industry.

We are becoming increasingly able to provide detailed, multi-dimensional assessments of environmental risks such as those related to climate change. For example, recent progress has been made in modelling the impacts of a changing climate on modern systems such as global supply chains. This has been done by integrating extensive datasets gathered around many of these systems with weather data and the output from climate model simulations.

The ability to model the impacts of climate change at an international, national and local level can give invaluable insight for policy makers. Dr Emily Shuckburgh of the British Antarctic Survey introduced the purpose of this workshop, to discuss which questions should be investigated by researchers to best help decision makers across government and industry adapt to the changing climate.

Dr Shuckburgh introduced a research project connecting changes to local temperature change in Egypt with energy demand and with milk production in the region. Accurate climate impact assessments can help policy makers identify and strengthen vulnerabilities in economic and social systems. New data science tools such as machine learning offer the potential to augment more traditional climate modelling techniques and lead the way to multi-dimensional climate risk assessments.

Improved knowledge and accessibility to statistical techniques can dramatically change the way that society approaches climate change. In particular, information concerning the risk and potential impacts of extreme events can help to inform where to take targeted local intervention to improve resilience.

However, to enable widespread access to such information by decision-makers, data management infrastructure needs to be overhauled and data must be curated in such a way as to ease its application and a close dialogue needs to be fostered between data providers and data users to ensure the information is generated in a form that is fit for purpose.


Thumbnail image courtesy of Global Panorama on Flickr

Banner image courtesy of NASA Goddard Space Flight Center on Flickr

  • 15 June 2017, 4pm

    Innovative climate risk assessment

    The purpose of this workshop is to explore the potential for longer term research initiatives at Cambridge to develop data science methods for more useful climate risk assessments.