8-10 January 2014
CATS hosted a workshop for early-career researchers on 'Understanding Uncertainty in Environmental Modelling'. The workshop was funded as part of NERC's Postgraduate and Professional Skills Development Programme.
Modelling and simulation are an increasingly important part of modern science, especially in highly policy-relevant disciplines such as weather, climate, and hydrology. Good practice in the use and interpretation of models is therefore vital, both for sound science and for informing evidence-based policy decisions. The workshop presented an overview of model evaluation methods, statistical inference for model output, and the use of models in risk management and decision-making, with the aim of exposing participants to methods and insights available in environmental modelling and encouraging critical evaluation of the approaches and methodologies used in their own research. The workshop was structured around several themes, with facilitated discussion time and interactive problem-solving exercises, allowing participants to explore and understand the concepts presented by the expert lecturers. The workshop was short but intense, and we hoped participants would leave with an overview of the key issues in environmental modelling, an understanding of how this affects their own research methods and where to find expert guidance and further information.
Lectures included: evaluating model performance; statistical inference from ensembles of models; why good statistics is not enough; what do decision-makers want? What can environmental modellers provide?; effective dissemination of uncertain forecasts.
Workshop co-ordinator: Dr Erica Thompson
Evaluating model performance: A weather-like example - Emma Suckling and Erica Thompson
Statistical Inference for ensembles of models - Lindsay Lee
Challenges in the Interpretation of Ensembles: Why Good Statistical Methods Aren't Enough - Dave Stainforth
Testing models as hypotheses: how can we be scientific in the face of epistemic errors? - Keith Beven
Evaluating climate-like models - Erica Thompson
Effective dissemination of uncertain forecasts - Liz Stephens