How much should historical experience guide our perception of flood risk?

When purchasing a property or taking out an insurance policy, one of the key questions asked, and one of the key things for re/insurers, banks and lenders to consider, is whether the property is at risk to flood.

But how do you know if a property is at risk to flood? Is it enough to simply assess the risk based on whether it has flooded before, as is usually questioned during the process? Or is there a danger in oversimplifying the picture by relying solely on historical experience? And, if so, how can high resolution maps and models give re/insurers and others involved in the home buying process a more complete picture?

We explore all this and more in our blog.

“It’s never flooded before…”: using historical experience to assess risk

When purchasing a property or taking out an insurance policy, homeowners faithfully answer the question, “Has any part of the property ever been flooded?” in order to assess the flood risk to the property.

But, of course, this is limited to the experience of the current homeowner. And, with the average homeowner moving house every 23 years and, in some areas of the UK every 8.6 years, this experience is clearly very limited, although those who have purchased property in the last decade will be in a more informed position, with flood reports normally provided by the solicitor including return period maps indicating the probability and severity of flooding possible at that property (with many providers using JBA’s own UK flood data).

By utilising the Environment Agency (EA) and Natural Resources Wales' (NRW) Historic Flood Event database we can calculate how many properties of the 39 million total properties in the UK have flooded during the average homeowner experience (Figure 1).

Figure 1: Number of properties flooded with the EA/NRW historic flood extents since 1947 (full record), since 1999 representing people moving every 23 years and since 2013 representing those moving every 8 years. There are 39 million properties in the UK.

However, using hydraulically modelled flood maps and catastrophe models, such as JBA’s UK flood data, we can see that this historical dataset captures only a fraction of the total properties at risk in the UK.

Flood maps and catastrophe models: lengthening the record

It’s unsurprising that a historical dataset fails to capture all properties at risk, for the following reasons.

Firstly: limitations of areas included. Even when flood events are captured in the dataset, the areas included are limited and do not represent all areas at risk. Traditionally, flooded areas were defined by manual survey, which was inherently limited by scope. Increasingly, flood extents are captured by drones and aerial imagery which increases the completeness of these footprints, but still may fail to capture the full picture.

Secondly, even with a perfect representation of historical flood events, we still wouldn’t capture the true risk. Put simply, the historical record isn’t long enough. It does not contain the full range of plausible flood events – just because a flood event hasn’t happened yet, doesn’t mean that with the right combination of rainfall, preceding conditions and topography, flooding couldn’t occur in the future.

That’s why high-resolution flood hazard maps and probabilistic flood models are a vital tool in a re/insurer, bank, or lender’s arsenal to capture the true potential and probability of flooding. By using these methods, re/insurers and property organisations can calculate accurate premiums, capture potential accumulations, and inform mortgage/ lending conditions.

For example, when using JBA’s newest UK Flood Model, we can see that the number of UK properties at risk to flood rises from between 121,000 and 1.3 million in the historical record, to 11.2 million, or one in four properties (Figure 2).

Figure 2: Insights from JBA’s flood hazard maps and flood defence datasets .

What about claims data?

Insurers will often use their claims data as the de facto source of information to decide a property’s flood history. Whilst unarguably insightful, insurers have a limited market share and have only been collecting claims data for the last ~20 years. As a result, this experience has similar limitations to external historical data.

As a result, history is not a complete guide to the future. While high risk areas will continue to be at high risk unless significant mitigation measures are put in place, as these figures demonstrate, “it’s never flooded before” provides a very limited view of flood risk.

Furthermore, with the forecasted increase in frequency and intensity of rainfall events under a changing climate, we may start to see properties which have never previously been flooded affected in the future.


Sophisticated, high-resolution tools in the form of maps and models are key to understanding the long term risk of flooding.

To find out more about how JBA’s flood data can help you assess flood risk, now and in the future, get in touch with the team.