People living in flood-prone areas could benefit from new statistical models being developed by Lancaster researchers.
The new models, developed by researchers at Lancaster University and the University of Oslo, improve on existing flooding models by more accurately capturing the relationship between the number of insurance claims and weather conditions.
They could help guide infrastructure spending for flood defences and improved drainage systems and also help insurers to set premiums and offer advice and warnings to customers.
Without accurate forecasting models it is very difficult for insurers to know what premiums they should set for their policies.
This results in insurers factoring in additional costs in their pricing to cover the risk of the unknown.
“Existing models lack flexibility to explain what causes high numbers of claims and always under-estimate the quantity of claims so there is a clear need for better models that can explain the weather conditions that cause high numbers of insurance claims,” said Dr Christian Rohrbeck, researcher at Lancaster University and lead author of the study.
The new model demonstrates significant improvements in accuracy when there are large numbers of insurance claims that result from a weather event – such as high levels or intense rainfall on saturated land, or large amounts of melting snow.
The researchers applied a clustering technique to bring together insurance claims made over a small number of days as they believe these claims are likely to be associated from one particular weather event.
The statistical model is based on 10 years of data obtained for three Norwegian cities – Oslo, Bergen and Baerum. The researchers believe their model could be applied to the UK.