High Impact Weather and Climate was the subject of the conference held by Royal Meteorological Society and National Centre for Atmospheric Science at the University of Manchester last week. Members of the forecasting service attended, along with many colleagues from the Met Office, universities and other practitioners. The three days were split into three broad topics – Observing, Predicting and Responding, and included key note speeches and a number of workshops, as well as poster sessions. The over-riding theme was that of impacts – how do we know when they are happening, how do we forecast them, and what is the best way to communicate this to those who need to know?

Liz Bentley from RMetS introduces the conference

The observations talks and workshops centred around the seemingly simple question – how do we know what is happening? High quality observation networks are sparse, so the point was made that crowd sourcing data, whether on weather or impacts, is a potential way forward, with an acceptance that the quality of individual observations may be less than ideal. Initiatives using smartphone technology, the Met Office WOW  site, and SEPA’s Report A Flood  are examples of this. The nature of observations was also questioned; particularly at the extremes, modelled parameters may be just as valid as ‘observed’ – Met Office Chief Scientist Prof Julia Slingo made the interesting point that an observation is just one particular version of reality.

(L) Mike Cranston discusses the Flood Guidance Statement during the ‘Perfecting the Weather Warnings’ workshop. (R) Louise Parry discusses the SFFS poster with Steve Cole from CEH.

Workshops such as those focussing on the Hazard Impact Model  and the Flooding from Intense Rainfall  projects presented the latest thinking in impacts modelling – the latter drawing on the work done for the Glasgow surface water flooding model  in 2014, as illustrated in the SFFS poster.

On the response side, the challenges ahead were presented in an excellent talk by Virginia Murray centred on the Sendai Framework for Disaster Risk Reduction , the ultimate aim of which is provide access to early warnings for all people by 2030. Currently 80% of developing countries have only basic or no warnings systems in place. Sally Priest, from the Flood Hazard Research Centre , made the point that just sending people a warning doesn’t necessarily lead to them taking action – people don’t always understand the warning, trust the authority, or act rationally. Care is particularly needed when communicating probabilistic warnings, the subject of a workshop involving the forecasting service, with partners at the Met Office and Flood Forecasting Centre. We played the Game of Making Decisions Under Uncertainty (developed by Micha Werner, to be available at the link soon) in which delegates were invited to make play the role of business owners and make cost-loss decisions based on probabilistic forecasts of flooding, and also discussed the situations in which low probability high impact warnings should take precedent over higher probability low impact situations. Even within the forecasting community there are many varying opinions on this topic.

The other benefit of the conference was the opportunity to meet and discuss topics of interest with colleagues from other organisations, progress existing projects and lay the groundwork for future collaborations.