On the anniversary of the European floods, Jane Toothill asks whether extreme flood events are on the increase and if so how should the industry respond?

This year sees the anniversary of two of the largest and most devastating flood events to affect Europe in recent times, namely the 50th anniversary of the 1953 sea surge event and the first anniversary of the river floods that affected areas of Central Europe last year. The two events differ in cause and in nature, the 1953 surge being connected to wind patterns over the North Sea, and the 2002 floods being a river event caused by intense precipitation over Central Europe. There are, however, similarities. Most notably, in the cause in both cases was ascribed to the occurrence of unusual meteorological phenomena.

In 1953, the cause of the sea surge was primarily unusual wind conditions over the North Sea. A slow-moving depression swung southwards and travelled down the length of the North Sea towards the Netherlands before moving east over Germany and Denmark. The depression was associated with strong northerly winds across the whole region. The strength of the wind was described as 'phenomenal' at the time, but it is the catastrophic flooding for which the event is remembered. The large storm surges that hit the southeast of England and the coast of the Netherlands were not caused so much by the high windspeeds, but rather by the unusual path followed by the storm, which allowed the surge wave to build up along the entire length of the North Sea, the speed at which the depression moved and the coincidence of the surge with high tide in many areas.

In 2002, the river floods were a result of the passage of two depressions associated with intense rainfall across Central Europe in close succession. These meteorological circumstances were unusual, but not exceptional. The first event gave rise to relatively localised flooding in small catchments of Central Europe and on the eastern coast of the Black Sea. Further downstream, however, flood management and defence systems were sufficient to cope with the increased water levels, and widespread flooding did not occur. The second depression was slow-moving and caused further heavy rain over a lengthy period in Austria, the Czech Republic and the southern part of Germany. In these areas, river levels remained high following the first storm, and soil saturation was at a sufficient level that many small catchments could absorb no further rain and flooded immediately. The floodwaters travelled downstream as continuing rain affected the area, and in the large rivers such as the Elbe and the Donau, widespread flooding resulted.

But just how phenomenal were these circumstances, and what is the likelihood of similar events happening again in Europe?

The answer is extremely complex, and in part lies in the hands of the various scientists tasked with investigating the possibility of climate change mechanisms and the way in which such changes might affect meteorological extremes. 'Global warming' is a subject that generates passionate views on all sides, and dissociating facts from spin so as to obtain a basis for reasoned decisions is not an easy task. For the insurer or reinsurer trying to locate a sensible line somewhere between the global warming scaremongers and the blasé 'it won't affect us' attitude of some others, significant difficulty exists in finding a practical solution based on sound scientific evidence. In any case, one must consider whether or not the circumstances causing events such as the two floods discussed here need to be 'extreme'. In both these cases, extreme floods resulted from meteorological conditions that were not extreme, but merely unusual.

For the sake of practicality, the many questions relating to the climate change issue, its causes and its effects, may be reduced to a single simple query: are extreme flood events getting larger and more frequent, and will they continue to do so?

Even this is not a simple question, and much of the available evidence appears to give conflicting evidence. Take two very simple, but nevertheless representative, examples.

One way to obtain an answer might be to examine historical records of natural hazard events and study the frequency of flooding in Europe. One potential source of such data is the EM-DAT disaster database, which records occurrences of disasters based on information from a wide range of sources. The results of a query of this database for flooding in Europe are shown in Figure 1.

Clearly, this dataset records a steady increase in flood frequencies throughout the past century. However, whether or not the dataset itself resembles reality is another question entirely, and before a simple conclusion can be drawn from datasets such as this, one must first question the completeness of the database throughout historical times and also the changing criteria used to enter events into the database. Quite simply, from the early 1900s, data relating to floods, especially smaller events, are not available, and a large part of the trend shown in this plot merely reveals improvements in data gathering and dissemination.

A more rigorous approach to studying flood frequencies and extremes, might be to examine records of water discharge rates from gauge station data in locations where there is a continuous record of data over a given timeframe. One example is given in Figure 2 for data from the Vltava River (a major tributary of the Elbe) at Prague. In this diagram, the August 2002 event can be seen to have the highest discharge of any event that occurred during the recording period. However, it is also noticeable that there appears to be no evidence for a steady increase in frequency or strength of events in recent years. Even the 1981 event, which caused significant damage in the Czech Republic, is relatively minor in comparison with tens of events that occurred over the previous 170 years. So how could even this relatively small event cause large amounts of damage?

Vulnerability to flood
The problem with considering only the frequency and intensity of hazard events is that the real issue for the insurer, namely that of future risk from flooding, is not fully addressed. There are many definitions for the term 'risk', but one of the more commonly accepted versions is:

Risk = Hazard x Vulnerability

Hence the level of risk does not depend only on the level of hazard, but also on the vulnerability to that hazard.

The vulnerability of the affected region played a significant factor in controlling the losses suffered in both events. In 1953, both flood prevention schemes and emergency procedures were poorly managed. Dissemination of warning information in the hours prior to the event was ineffective, and action taken once flooding started was disorganised and chaotic. The extent of flooding could (and should) have been more effectively limited by the dyke systems in place, but many of these were damaged or in poor repair following the war years, and commonly did not offer the protection expected. Hence, not only the hazard, but also the vulnerability of the affected area was high.

Today, flood management is an important issue and defences and emergency systems are far better able to cope with the onset of severe flooding. However, even modern defence systems were not sufficient to cope with the excesses of the 2002 floods, and the event served to show that, even now, the vulnerability of many parts of Europe to extreme flood events remains high. Figure 3 shows land use and flood extent maps of the area surrounding Dessau, at the junction between the Elbe and the Mulde rivers. The normal path of the Elbe can clearly be seen in the top map, which depicts built up areas in red. The flat land of the floodplain can also be seen in beige, and at the northern side of the river, the towns of Coswig and Wittenberg, along with large portions of Dessau, are located on the floodplain. The extent of the flooding was sufficient to inundate these areas, and, as in other parts of Germany, much of the flooded area comprised relatively modern residential buildings.

Especially in the wake of last year's flooding, there is a growing pressure on governments in Europe to act to reduce social and economic vulnerability to flood. Much of the pressure originates in the alarm calls from the scientific community relating to global warming and the likely increase in the severity and frequency of extreme events (notably rainfall), but there are further factors such as a decreasing public acceptance of risk, and increased awareness that man's activities alone can increase flood hazard. Urbanisation in catchment areas significantly decreases the ability of the ground to absorb water, and where urbanisation occurs along rivers, not only are natural 'wet' habitats that previously acted as storage areas for excess water removed, but straightening of river channels and construction of flood defences can aid the swift passage of floodwaters to downstream areas.

The situation is exacerbated by increasing wealth, population and property values within Europe. Given an expanding population, there is an ever-present need for development, and in the absence of strict planning controls, development of desirable riverside areas is common. Development of these flood-prone regions has already occurred in many areas, and as was seen to be the case in 2002 along the Elbe, there is already a significant existing vulnerability to flood hazard. Flood management must be undertaken on the scale of an entire basin since the protection of one area may simply create a worse hazard further downstream, but this kind of planning means that some existing properties may remain vulnerable.

Vulnerability of the insurer
In terms of flood risk to insurers in Europe, vulnerability may be considered on two levels: firstly, the risk of the community, at local, regional, national or international level; and secondly, the part of that risk accepted by the insurer.

Although at first glance it may appear that insurers have little control over the risk of a community, there is a growing trend to involve insurers (along with other interested parties) in flood management processes. Insurance offers a practical and economic means of dealing with the after-effects of flood events, and the global nature of the insurance business means that even very large losses may be covered. To many governments, faced with an insurance community unwilling to cover risks affected by flooding on a regular basis, there is a need to increase communication and co-operation between the various parties involved so that a solution acceptable to all might be reached. Future decision-making processes may involve the insurance community to a greater extent, and political decisions are likely to directly impact the insurer by changes to the way in which flood is covered in insurance policies throughout Europe.

Whatever changes are implemented relating to flood management, insurance structures and the relationship between the insurance industry and governments, it is inevitable that increasing levels of flood cover will be sought throughout Europe. This will bring good business to the industry, but with it raised levels of risk. Even with the best possible planning and defence systems, flood hazard can never be entirely removed, and the re/insurers' prime means of protection against their own vulnerability must remain the careful management of the risks they choose to accept. The need to adapt to changing circumstances is essential, and given the inherent uncertainties associated not only with predictions of climate change but also of the way in which man will continue to develop and protect flood-prone areas, speed of adaptation may prove to be of prime importance.

By Jane Toothill

Jane Toothill is a natural hazard analyst at EQECAT Europe.