Vb weather systems, flood defences, and the changing return period question

Speaking at the International Specialty Catastrophe Modelling (ISCM) Conference in Prague this week, Tim Edwards set out three observations from recent European flood events that have direct implications for how reinsurance markets price European flood risk.

‘Vb’ systems are an increasingly material driver of European flood risk

Vb (pronounced ‘Five-B’) cyclones are defined by an atypical track that originates over the Mediterranean and Adriatic Sea. These cyclones intensify over the sea and absorbs moisture, before pushing northward through Central Europe. Hofstätter et al. (2016) found that while Vb cyclones makes up just 3-5% of overall European storms, they account for 45% of storms that generate extreme summer precipitation in south-east Germany and much of the severe flooding across Central Europe. Indeed, Vb systems are responsible for five of the largest and most damaging flood events in Europe over the last 30 years.

Tim adds that: ‘While climate projections from Messmer et al. (2020) point to a smaller number of Vb events in future decades, due to changes in circulation patterns, they are expected to generate more extreme rainfall conditions when they do occur. This is because of rising sea surface temperatures in the Mediterranean. These trends require factoring in when calculating flood risk.’

Flood defences are doing more work than the loss data alone suggests

Howden Re’s estimation of on-levelled flood losses (for pricing and exposure change) indicates a 35 to 50% reduction in flood losses from the September 2024 ‘Boris’ event relative to the July 1997 event. However, Miroslav Hrncir notes that: ‘the more recent ‘Boris’ event produced greater precipitation extremes, so the reduction is on-levelled losses between events points to a substantial benefit from post-1997 investments.

These benefits are not limited to structural flood defences alone but also reflect major advances in flood forecasting, early warning, and emergency response systems, which enabled better preparedness, and more effective crisis management.

To illustrate the role of physical infrastructure, the modernization of the Wrocław Floodway System and construction of the Racibórz Dolny Reservoir, completed in 2020, materially attenuated peak discharges on the Odra river in September 2024 relative to prior events. In parallel, improved hydrometeorological monitoring and warning systems helped translate this attenuation into reduced human and economic impacts.

In the Czech Republic victims fell from 60 to 13 from these reference events. Rescues from 26,009 to 2,467. IRS evacuations from 33,632 to 19,173. The loss distribution shifted away from housing and toward infrastructure — a sign that residential exposure was meaningfully better protected.

Calculating return period loss estimates using hazard indices

Loss return periods are important for governmental resilience planning and for (re)insurance pricing discussions. Losses from events that occur well within model parameters are less likely to result in a corrective price behaviour than an event that is not well modelled. 

Following significant loss events, it is common to quantify the local return period of the hazard (e.g., precipitation accumulation or river discharge) at the location of the loss event. However, the local hazard return period may differ from the loss return period for a portfolio of risks. Robert Graham adds that: ‘Loss return periods should quantify the severity of hazard over the entire geographic domain of the portfolio and apply weighting reflective of the exposure at each location. Changes in climate circulation patterns and the latest resilience initiatives should also be taken into consideration.’