Finding reliable flood data remains a major obstacle to modelling, says Dr Justin Butler.
This summer’s flooding has highlighted the catastrophic loss potential of UK flood risk. The extent, severity and spatial distribution of flooding took the insurance and reinsurance industry largely by surprise.
What is perhaps more surprising is many insurers’ lack of awareness with regards to the risks they are facing and more importantly the difficulties we encounter when it comes to obtaining essential flood-related data sets. As any catastrophe modeller knows, rubbish data in means rubbish data out. This is especially the case in climate and catastrophe modelling, both of which are highly dependent upon reliable input data.
The inadequacy of existing flood models stems largely from two interrelated factors. Firstly there is a failure to include all sources of potential flood risk (ie fluvial – from rivers and watercourses, pluvial – direct run-off from land, and residual). Secondly, many models are limited by the lack of historical flood information which can, and should, be used to validate model predictions.
In the UK, a wealth of high quality historical and modelled flood data is held by loss adjusters, insurers and government. Improved access to UK flood data could be used to enhance the predictive power of the latest generation of flood models to the benefit of insurers, government, emergency planners as well as the general public.
“As any catastrophe modeller knows, rubbish data in means rubbish data out
Dr Justin Butler Managing Director, Ambiental
However, unlike in Australia and the US, access to this data in the UK is often limited due to excessive cost and/or restrictive data licensing arrangements. Or it is prohibited entirely for other commercial reasons.
It would be naïve to discount the competitive advantage that can accrue from retaining proprietary flood risk and related spatial data in-house. It is equally naïve to forego the wider benefits, financial and otherwise, that can result from disseminating that information more widely.
A situation could be achieved, whereby information flows between insurers, data providers and modellers, increasing knowledge and awareness of flood risk to the benefit of all parties. This could be achieved via a more pragmatic and flexible approach to data licensing and dissemination. It remains to be seen whether will we see further catastrophic events before the way environmental data is handled and distributed in the UK is changed for good.