[Dr.Zoidberg]

To get a good survey about the rising of natural disasters you only have to look to the inquiries of the reassurer, as those companies are directly affected. As an example I picked the Munich Re Group, one of the biggest reassurer of the world. It´s a little long to read, but nevertheless very interesting.

At first let´s have a look on the trend of hurricanes:

Dr. Eberhard Faust

Changing hurricane risk in the North Atlantic

What we are concerned about

Updated to the end of the hurricane season 2005

The elevated frequency of intense storms in 2004 and 2005 — no fewer than four of the ten strongest hurricanes ever recorded occurred in 2004 or 2005 — hints at a systematic change in the hazard situation and hence a shift in the loss distribution and its parameters.

After an extremely active US hurricane season in 2004 with an absolute record of four hurricane landfalls in/near Florida and the highest overall insured loss from tropical cyclones until then, 2005 has been a season with even higher losses from hurricanes (particularly Katrina, Rita, and Wilma).
Accordingly, the current situation has to be characterised by a higher average market-wide annual loss and different return periods for market-wide claims expenditure compared with the situation a few years ago. In the following analyses, we address the question of new evidence with respect to causes of changes in hurricane frequencies and intensities.

01 Ocean temperatures and cyclone intensities worldwide

A scientific study performed by the Scripps Institute (Barnett et al. (2005) Science) compares recordings of ocean temperatures and respective computer simulations and shows that anthropogenic climate change is having a strong impact on increases in recorded temperatures of the upper ocean layers since 1960 (cf. Tourre/White GRL (2005)).
Other scientific studies by US researchers (Emanuel (2005), Nature; Webster et al. (2005), Science) have shown the following. There is evidence of a warmer trend during the summer season in all tropical oceans amounting to an average of 0.5°C since 1970. The intensity of tropical cyclones, characterised by the parameters of maximum wind speed and cumulative length of time with high wind speeds, increases in correlation with sea surface temperature (Fig. 1). As a consequence of this correlation, the global number of severe tropical cyclones (4–5 on the Saffir-Simpson Scale) has increased in relation to the annual total for all ocean basins. There has been a steep increase in absolute terms too, from about 8 per year at the beginning of the 1970s to 18 per year, i.e. more than double — in the period 2000–2004. At the same time, the proportion of weaker cyclones (Category 1) has decreased, while there is no recognisable trend as far as the moderate types (Categories 2-3) are concerned (Fig. 2).

02 Climate oscillation in the North Atlantic

In addition to this shift in the intensity distribution towards the higher categories, changes may also be observed in the total frequency in some regions. The number of cyclones occurring throughout the world every year on average is 80 (margin of deviation: 20) without any distinctive trend.
A general increase in frequency is observed in the North Atlantic since 1970, that means from a comparatively cool period to the current "warm phase" in terms of sea surface temperatures (Fig. 3). Accordingly, the hurricane season of 2005 has set an absolute record in terms of the number of named tropical storms (27, old record 21) and hurricanes (15, old record 12).
If further research findings of recent years are taken into account (Goldenberg (2001), Science; Trenberth (2005), Science), the result for the North Atlantic is such that cyclone activity is determined there both by a natural climate oscillation and by a superimposed linear warming process — most probably not explainable without anthropogenic global warming.
There are alternating phases lasting for several decades with exceptionally warm or exceptionally cool sea surface temperatures, the margin of deviation being around 0.5°C. The natural climatic fluctuation is driven by the ocean's large-scale currents (thermohaline circulation, Knight et al. (2005) GRL, Willoughby/Masters (2005)). Warm phases produce a distinct increase in hurricane frequency and also more intense storms, whereas cold phases have the opposite effect. So in the current warm phase, for example, 4.1 strong hurricanes have already occurred per year on average while in the previous cold phase this figure only was 1.5 (this means an increase by 173%). Of course, a definitive value for the average annual level of activity for the whole of the current warm phase can only be given when this phase has ended. The figures correspond to the observation possible up to 2005.

03 Global warming

At the same time, the natural fluctuation between these phases seems to be intensified by a superimposed long-term warming process so that sea surface temperature and the level of hurricane activity increase from warm phase to warm phase (Fig. 4). The increase in the number of strong hurricanes per year from 2.6 to 4.1 from the previous warm phase to the current warm phase means an increase of 58%.* There are strong arguments in favour of climate change as the long-term warming agent. The current unusually high level of activity is most probably due to the warm phase prevailing since the mid-1990s, which is supposed to continue for several years and intensified by the relatively linear process of global warming.

There is a clear indication that both the natural climatic cycle and global warming influence not only overall frequency but also landfall frequency. Between the last warm phase (approx. 1926 to approx. 1970) and the current warm phase since approx. 1995, the average annual number of landfalls increased as follows (Fig. 5):

Cat. 3—5 hurricanes+67% (from 0.6 to 1.0)
Cat. 1—5 hurricanes+33% (from 1.8 to 2.4)
Trop. storms and Cat. 1—5 hurricanes+47% (from 3.4 to 5.0)

This comparison has to be seen as being primarily linked to the influence of global warming.


The change in level between the last cold phase (approx. 1971 to approx. 1994) and the current warm phase since 1995 has the following impact on the number of landfalls (Fig. 5):

Cat. 3—5 hurricanes+233% (from 0.3 to 1.0)
Cat. 1—5 hurricanes+100% (from 1.2 to 2.4)
Trop. storms and Cat. 1—5 hurricanes+100% (from 2.5 to 5.0)

This comparison has to be seen as being primarily indicative of the natural climatic oscillation.

* The records of the period before aircraft reconnaissance started in the mid-1940s are not as reliable as the records since then. This applies primarily to intensity attributions, because one has to rely on observations made by ships.

04 Different loss distribution

These strong changes, reflected in both the number of tropical cyclones and the number of landfalls, can only mean that we must expect a different loss distribution in the current warm phase since 1995 compared with the distribution in the prior period.
We should recall that we observe an increase in terms of the annual frequency of major hurricanes in the order of 170% from the foregoing cold phase (1971 to 1994) to the current warm phase since 1995. In terms of landfalls the increase is of the order of 230%.
Even if we compare the loss distribution of the current warm phase with a loss distribution based on all years since 1900, which can be called indifferent towards the natural climate cycle, we should expect a large difference. This is strongly indicated by a comparison of hurricane intensity distributions calculated for the whole period 1900 — 2005 versus the current warm period 1995 — 2005 (Fig. 6). It is plain to see that the current warm phase is marked by a higher proportion of strong hurricanes (Categories 4 and 5 on the Saffir-Simpson Scale) and a lower proportion of weaker hurricanes (Categories 1 and 2 on the Saffir-Simpson Scale). Category 4 and 5 hurricanes account for 14% and 6% respectively in the distribution since 1900 and have increased to 20% and 10% in the current warm phase distribution. On the other hand, the Category 1 and 2 hurricanes account for 37% and 23% respectively in the distribution over all years since 1900 and have decreased to 34% and 17% in the current warm phase distribution.
None of the loss models available commercially incorporate such a change in the distribution. So it is a major challenge for the insurance industry to respond to the present-day hazard distribution and — as a consequence of this — the present-day loss distribution and to take them into consideration adequately in its risk management.

05 Glossary

Anthropogenic climate change/global warming

During the period of industrialisation, greenhouse gas emissions increased steadily and led to an atmospheric CO2 concentration of 380 ppm in 2004. The pre-industrial level was 280 to 300 ppm which at least for the past 650,000 years and probably for the last millions of years has not been exceeded. There are other greenhouse gases such as methane or dinitrogen oxide, which have increased equally fast.
Greenhouse gases alter the radiation properties of the atmosphere in such a way that much more energy from the sun is trapped by the lower parts of the atmosphere. This anthropogenic global warming comes in addition to what is called the natural greenhouse effect. Even before the appearance of mankind and of the industrial age the earth's atmosphere contained greenhouse gases (in particular CO2 and others), which have warmed the earth's surface by roughly 33°C. This natural greenhouse effect must be regarded as a precondition for the development of life on the planet.

Tropical cyclone

General expression for tropical storms forming over tropical oceans. Depending on the region and strength they are called hurricanes (Atlantic and Northeast Pacific), typhoons (Northwest Pacific), or cyclones (Indian Ocean and Australia).

Saffir-Simpson intensity scale

The Saffir-Simpson Scale is a five-stage intensity scale for tropical cyclones. The scale spans the following categories:

• Cat 1: windspeed 118—153 km/h; central pressure >= 980 hPa
• Cat 2: windspeed 154—177 km/h; central pressure 965—979 hPa
• Cat 3: windspeed 178—209 km/h; central pressure 945—964 hPa
• Cat 4: windspeed 210—249 km/h; central pressure 920—944 hPa
• Cat 5: windspeed > 250 km/h; central pressure < 920 hPa

Atlantic cold phases/warm phases

The so-called cold and warm phases in the North Atlantic are part of the Atlantic Multidecadal Oscillation (AMO). The mechanism behind it is a large-scale water flow conveyer belt in the ocean with periodically enhanced or reduced activity resulting in unusually warm or unusually cool surface waters in parts of the ocean. This overturning circulation, which is driven by water temperatures and water salinities, is called the thermohaline circulation.

Natural climate oscillation

Natural climate oscillations can be differentiated by the respective time scales. They are not driven by external influences on the earth's climate system, such as changes in solar irradiance or anthropogenic greenhouse gas emissions. Examples of natural climate oscillations are the El-Nino/Southern-Oscillation events (interdecadal time scale), the North Atlantic Oscillation (quasi-decadal Oscillation) or the Atlantic Multidecadal Oscillation (multidecadal time scale).

Latest examples of abnormal storms:

Ernst Rauch

Peak meteorological values and never-ending loss records

The last two years have been dominated by extreme tropical cyclones. The belief that the exceptional year of 2004 would be followed by a period of calm in 2005 turned out to be mistaken. The time has come for a radical rethinking of how hurricane risks are evaluated.

The record-breaking year of 2004

2004 was marked by the highest regional frequencies and intensities of tropical cyclones in the North Atlantic since the recording of meteorological tracks began in 1851.
Hurricane Ivan was particularly significant for the insurance industry: its HDP (Hurricane Destruction Potential), which is the sum of the squares of the maximum wind speed in 6-hour periods for the duration of the storm, was 71,250. For the sake of comparison, the average HDP value of all tropical cyclones recorded in the Atlantic in each entire season between 1950 and 1990 was 70,600.
Hurricane Ivan set new records in terms of duration and intensity, but the latest scientific findings suggest it will not be an exception for very long. The study (Emanuel [2005], Nature) quoted in the section "Climate cycles and global warming — Effects on risk evaluation" shows that the Power Dissipation Index (PDI), which represents the accumulated wind energy of tropical cyclones in the North Atlantic for a whole year, increased sharply in correlation with the higher sea surface temperature. The PDI is calculated in a similar way to the HDP. A closer analysis of this change makes it clear that there has been an increasing trend in the strength and duration of hurricanes and thus in their destruction potential too.

2005 — An increase is possible

In this season, both hurricane activity, i.e. the number of tropical cyclones, and the observed intensities reached new peak levels. The new maximum values were far above the old records of 21 tropical storms (1933) and 12 hurricanes (1969). A total of 27 named tropical cyclones developed in the North Atlantic, 15 of which reached hurricane force with wind speeds exceeding 118 km/h.
The intensities were no less striking. The list of the ten strongest hurricanes ever recorded includes Wilma, Rita, and Katrina, all from the year 2005. On 19 October, Wilma had a central pressure of 882 hPa, the lowest ever recorded. This suggests that it also had higher wind speeds than any other hurricane in the Caribbean since recordings began in 1851.
The beginning and end of the hurricane season in 2005 were also marked by exceptional meteorological features. The hurricane year began very actively with seven tropical cyclones in June and July — two more than the previous record of five by the end of July. Hurricane Epsilon marked the end of the season in December, along with Tropical Storm Zeta, which was still active in the Atlantic even at the beginning of January 2006: two storms that did not observe the "official" end of the hurricane season on 30 November.

01 Losses caused by the hurricane series in 2004 and 2005

The four most devastating hurricanes with landfalls in the Caribbean and the United States — Charley, Frances, Ivan, and Jeanne — presented the insurance industry with a new peak loss from tropical cyclones in the Atlantic of around US$ 30bn.
The most expensive year for insurers in this region before then was 1992, when Hurricane Andrew generated insured losses of US$ 17bn. According to Munich Re's analyses, Andrew would cost the insurance industry almost US$ 30bn today, given the increase in insured values in the affected regions of Florida and Louisiana since then.
The sum total of individual losses from hurricanes in 2004 was therefore not an extraordinary figure in itself. The surprising part was that a loss of these dimensions occurred only 13 years after Hurricane Andrew, since there are commercial models that put the "return period" for an annual market hurricane loss of US$ 30bn at well over 30 years.
The high loss accumulation from a series of moderate hurricanes was also unexpected for some risk carriers. Many insurers had responded to Hurricane Andrew by concentrating their efforts on estimating the accumulation loss potential of one major event — but these estimates were to be put to the test in 2005.
The natural catastrophe year of 2005 was marked by record losses from hurricanes in the North Atlantic, with insured losses exceeding US$ 83bn. Munich Re estimates that Hurricane Katrina alone generated privately insured market losses of US$ 45bn. This figure was boosted by Rita and Wilma, each costing around US$ 10bn, and significant insured losses from other storms like Dennis, Stan, and Emily.

A phase of rethinking is necessary

Two aspects in particular marked the year 2005: a mega-loss caused by Hurricane Katrina and a succession of moderate hurricane losses. Only a year after the most expensive natural catastrophe year in original values, the optimism displayed by many a market player proved to be unfounded. 2004 was not a solitary exception.

Losses in 2004 and 2005: Insured market losses from hurricanes

• United States (mainland only) approx. US$ 95bn
• Gulf of Mexico (offshore) approx. US$ 14-15bn
• Caribbean approx. US$ 2bn
• Mexico approx. US$ 2bn
• North Atlantic (United States, Caribbean, Mexico) US$ 115bn

In all these regions, a process of fundamental rethinking is called for in the evaluation of hurricane risks. The United States mainland is particularly important in this regard, since high insured values will inevitably lead to high insured accumulation losses when the time comes.

02 Hurricane Katrina: Meteorological aspects

Hurricane Katrina developed out of a low-pressure vortex over the Bahamas on 23 August. As the eleventh tropical cyclone of the 2005 hurricane season, it crossed South Florida in the Miami area as a Category 1 hurricane (measured on the Saffir-Simpson Hurricane Scale).
In the days that followed, Katrina moved over the eastern part of the Gulf of Mexico with a rapid increase in intensity. Over those areas where the water temperature was 1-2°C above the long-term average, the hurricane already reached force 5 on 28 August. This corresponds to wind speeds of approx. 340 km/h in peak gusts.
Shortly before making landfall on 29 August in the state of Louisiana — some 30-50 km east of New Orleans — it weakened to a Category 4 hurricane. An analysis of wind speed data published by the National Hurricane Center in Miami in December 2005 adjusted its strength at landfall again, lowering it even further to Category 3. Upon landfall in Louisiana and when it moved on to the states of Mississippi and Alabama, Katrina caused massive windstorm damage and, initially on a local scale, flood damage due to torrential rain.
Just a few hours after the hurricane vortex had passed over South Louisiana, the levees were breached on Lake Pontchartrain and on an artificial drainage canal. Large parts of New Orleans were flooded. The affected areas lie below sea level in a kind of soup bowl, and there is no natural drainage.
As draining is only possible using pumps or by natural evaporation, it took several weeks to dry out the city. It was not until early December 2005 that important infrastructure installations were back in place and access to the city of New Orleans was completely restored.

But it isn´t enough to only analyze the trend of hurricanes, to get a general view about the development of natural disasters:

NatCatSERVICE information

Increasing intensity and costs of natural catastrophes – Is this a long-term trend?

2005 broke all negative records. Natural catastrophes have never been so expensive, either for the world’s economies or for the insurance industry. It was also one of the deadliest years of recent decades.

Over the past year we have continued our research into the possibility of identifying natural hazard trends with even greater accuracy and certainty. To this end, the data stored in Munich Re’s natural catastrophe database, NatCatSERVICE®, was prepared to make it more amenable to systematic analysis. We are pleased to publish the results of our work for the first time in this edition of Topics Geo. This NatCatSERVICE® information examines whether there is a discernible trend towards larger natural catastrophes, where in the world such a trend may be evident and how it may manifest itself.

Data sources, data preparation, classification

The whole process of evaluating macroeconomic losses is subject to significant uncertainty and fluctuation, as we discussed in detail in topics — Annual Review: Natural Catastrophes 2000.
We used the Munich Re natural catastrophe categories as a basis for our investigation of possible trends (Graphic: Natural catastrophes — Breakdown into seven catastrophe categories). This seven-level scale — from 0, natural event, to 6, great catastrophe — makes it possible to assign each loss event to a particular category, even if the exact extent of the overall losses are not known or cannot be determined.
Our analysis examined 16,000 natural catastrophes in the period between 1980 and 2005. Only about a quarter of all events were backed up by reliable official figures concerning the economic losses involved. Since the mid-1990s, however, there has been a distinct improvement in the reporting of overall losses (Graphic: Natural catastrophes of economic losses).
Munich Re’s experts estimated the losses from the remaining 12,000 events on the basis of claims notifications and global comparisons with similar events, considering in each case the affected national economy.

Two examples of this procedure

Example 1

• Estimate of the overall losses on the basis of known insured losses using the factor of insurance penetration, a value that is known for all markets and for all the various types of event. This method factors in the type of natural hazard, the region of a country affected (urban, rural, population density, quality of buildings), and the classes of insurance business that were affected by losses. This information is the basis for a realistic loss estimate (Graphic: Example of a loss estimate: Hurricane Ivan).

Example 2

• If insured losses are not known, as is frequently the case in developing countries, Munich Re’s loss estimate is based on the following parameters: type and duration of the natural catastrophe, region affected (urban, rural, population density, property, infrastructure, and public utilities, the number of people involved, and the death toll. On the basis of this data, an approximation technique then searches for all comparative catastrophes in the affected region for which there is detailed and reliable information on overall losses. The events are clustered and realistic values derived for individual units (e.g. average value of a residential building in a rural area). In this way, the event can be assigned to a certain category of loss.

In order to determine the extent of the loss, all events were assigned to one of seven categories of natural catastrophe. Catastrophe category 0 was disregarded for the purposes of our analysis, as it is used for natural events which have little or no economic impact. The remaining events were divided into three main categories:

• Small-scale and moderate loss events (categories 1 and 2)
• Severe and major catastrophes (categories 3 and 4)
• Devastating and great natural catastrophes (categories 5 and 6)

The Analysis

• There were hardly any noticeable differences in the percentage breakdown of the types of event across the three main categories. The exceptions to this are earthquakes and volcanic eruptions. The proportion of windstorms in the three main groups is in fact absolutely identical. Overall, weather-related natural catastrophes dominated with a share of over 85% in all catastrophe categories (Graphic: Percentage distribution of events).
• If one considers the number of events from 1980 up to the present day in their respective categories, it can be seen that the proportion of catastrophes in category 1 has diminished while there has been a significant increase in categories 2 and 3 (Graphic: Number of events per year).
• A similar breakdown by continent shows that Asia – the continent with the most towns and conurbations – clearly dominates in terms of the number of events. Asia experienced 4,500 events, 70% of which were socalled "small loss events". At the same time, however, Asia also experienced the greatest number of devastating and great natural catastrophes (225 events).
• Asia was also hardest hit in terms of the number of fatalities (800,000). Almost 90% of these fatalities were caused by events in catastrophe categories 5 and 6 (devastating and great catastrophes).

A comparison of Europe and North America (USA and Canada) shows that the two continents were affected by about the same number of natural catastrophes (Graphic: Natural catastrophes comparison between Europe and North America). However, while Europe was hit primarily by small events, North America had to contend with a greater number of severe and great natural catastrophes (categories 3–6). This trend is also reflected in the loss figures: overall losses in North America were almost three times as high as those in Europe and insured losses about four times as high. In absolute terms, more people died in Europe, but this can be largely attributed to a single event: the 2003 heatwave, which affected the whole continent. The final death toll was more than 35,000.

Dr. Eberhard Faust

Climate review 2005

Climate change continues unabated. This is clearly confirmed by the results of research in 2005, a year that is likely to go down as the second warmest year ever recorded. According to provisional calculations by the World Meteorological Organisation (WMO), the mean global temperature in 2005 deviated by +0.47°C from the average of the climate normal period 1961–1990. It is thus one of the warmest years since recordings began in 1861 and currently ranks as the second warmest year worldwide.
The WMO will publish the final figures in February 2006. Nothing provides more striking evidence of the continual warming of our planet than the fact that the nine warmest years have all occurred between 1995 and 2005. In fact, in the northern hemisphere, 2005 is likely to go down as the warmest year ever recorded, with an anomaly of +0.65°C. In September 2005, ocean ice in the north covered less than six million square kilometres for the first time since satellite observations began in the 1970s. September is the month in which it typically reaches its minimum. The sea ice cover registered at the end of that month showed a reduction of 8% in the last 25 years.
A major part in this development was played by the North Atlantic, where the surface temperature in 2005 currently ranks as the warmest annual mean figure ever registered. The exceptionally large anomalies in a belt around 50°N and record values in the Caribbean and the tropical Atlantic were particularly noticeable. One of the effects of this was the extreme drought in the Amazon region. This was due to the higher level of evaporation and precipitation formation over the warm sea surfaces, whilst in the neighbouring region of North Brazil the prevailing conditions were a subsidence of air and cloud dispersion.
A study by the Scripps Institute of Oceanography was the first to show that anthropogenic climate change is responsible for the rising temperatures in the upper layers of all the earth’s oceans. This influence far outweighs the effects of natural climate variability and external forcings like changes in solar radiation and volcanic activity.

Examples of extreme weather patterns in 2005

• Between October 2004 and June 2005, the total volume of precipitation in western France, Spain, Portugal, and the United Kingdom was only half the long-term average. As a consequence, Spain and Portugal suffered their worst drought since the 1940s, resulting in many wildfires. And that only two years after the hot and dry hundred-year summer of 2003.
• With an anomaly of +1.75°C in the first five months, 2005 was the hottest year in Australia since recordings began in 1910.
• There was hardly any rain in Brazil, leading to extreme dryness in the south (Rio Grande do Sul) and the Amazon region and producing the worst drought for 60 years.
• In contrast, July presented Mumbai with the greatest 24-hour precipitation volume ever recorded in India.

Climate change and insurance

Records of mean global temperatures go back to 1861, and for the northern hemisphere there are reliable temperature estimates for the last 1,000 years. The records show a distinctive trend. The average temperature on earth is rising — with an increase of 0.7°C since 1900 alone. The ten warmest years ever recorded have all occurred since 1995. 1998 set a new all-time record: the maximum temperature that year was higher than in any other year throughout the past millennium. The next near record followed in 2005.
A temperature increase of 0.7°C may seem moderate. However, between ice ages and warm periods, which alternate due to natural factors, there is only a difference in mean global temperatures of 6-7°C.
The extremely pronounced warming that has been observed particularly in the past three decades cannot be explained simply by natural influences. The scientists of Munich Re's Geo Risks Research Department are therefore certain that this global warming is man-made and that it will have massive repercussions.
A survey of the years 1950-2005 reveals a massive increase in major weather-related natural catastrophes during that time. Between 1994 and 2005 there were almost three times as many weather-related natural catastrophes as in the 1960s.
The trend is even more distinct with regard to losses. Economic losses increased by a factor of 5.3 in the same period, insured losses by a factor of no less than 9.6. The main causes in both cases were floods and windstorms. The majority of fatalities, more than three-quarters, were caused by "wet storms".

Last but not least the outlook from insurances point-of-view:

Dr. Eberhard Faust

The further outlook

"Everything used to be better. Even the weather." Do such statements glorify the past? Not entirely, for the climate is indeed changing, as researchers have recently confirmed. Their findings are of profound importance for both insureds and insurers, especially as regards risk management.

Global mean annual temperatures can be followed back to 1861. In 2006, 145 years later, a trend has emerged which can no longer be ascribed to chance: the nine hottest years ever recorded were all between 1995 and 2005. According to provisional calculations by the World Meteorological Organisation (WMO), the mean global temperature in 2005 deviated by +0.47 °C from the average temperature between 1961 and 1990.
By only half a degree? Then there's nothing to worry about, is there? Well, maybe there is, as a closer look at the repercussions reveals. The area of sea ice covering the northern hemisphere in late September every year has declined by roughly 8% in the last 25 years, for example. Glaciers in mountainous areas are on the decline.
2005 was a year of weather extremes in many regions. Just two years after the "hundred-year summer" of 2003, Spain and Portugal suffered their worst drought since the 1940s. Between October 2004 and June 2005, western France, Spain, Portugal, and the United Kingdom had only half as much rain as usual. Australians sweltered in a heatwave with an average temperature that was 1.75°C above the mean in the first five months of the year.

Record year 2005

2005 turned out to be the hottest year there since records began in 1910. Major floods hit the Alpine regions in August, especially in Switzerland, Austria, and Germany. They were caused by a central European weather trough, involving a low-pressure system which picks up considerable amounts of moisture over the warm water of the northern Mediterranean and deposits them over the Alpine region and the low mountain ranges of central, eastern, and southeastern Europe as it heads (north)east.
Such weather troughs were responsible for the floods on the Odra in 1997, the Vistula in 2001, and the Danube and Elbe in 2002 — despite the fact that the amount of rain falling in an average central European summer is steadily decreasing and that the probability of very hot and dry summers has considerably increased.
The insurance industry was affected above all by the losses caused by tropical cyclones in the North Atlantic. These storms develop over tropical oceans and depending on their intensity and the region involved, are called hurricanes (Atlantic and Northeast Pacific), typhoons (Northwest Pacific), or cyclones (Indian Ocean and Australia).
Worldwide, the proportion of severe tropical cyclones — Categories 4 and 5 on the Saffir-Simpson scale — is growing steadily. Since 1970, their number has risen from an average of 8 per year to 18. In 2005, 27 tropical cyclones were recorded in the North Atlantic, including 15 of hurricane force — a record number. 2004 had been a very active season too.

Changes in the last 10-15 years

In view of such increases, the question is: what has changed in the last 10 to 15 years? An important part of this is how tropical cyclones work. They are fuelled by the difference in temperature and pressure between the surrounding atmosphere and its warm centre.
The relatively low pressure in the centre is caused by the evaporation of ocean surface water — the warmer the water, the stronger the evaporation. Climate simulations using cyclone models show that a "heated" earth with higher temperatures in tropical oceans gives rise to more intense storms characterised by higher wind speeds and heavier rainfall.
Indeed, this has been confirmed by our observations over the last few decades. At the same time, the increase is "masked" by natural oscillations. Over time, the average surface temperature of the North Atlantic has fluctuated in long waves; there have been exceptionally warm and exceptionally cool phases, each lasting several decades.

Higher North Atlantic temperatures

Higher temperatures prevailed before 1900, between the mid-1920s and the late 1960s, and again since the mid1990s. This phenomenon is probably due to a natural cause known as thermohaline circulation (THC). This means, in strongly simplified terms, that warm, saline water from the tropical North Atlantic, the Caribbean, and the Gulf of Mexico is transported northwards and eastwards in the upper sea layers by the Gulf Stream and the North Atlantic Current.
Once it has discharged its heat into the atmosphere, the water, which is very dense due to its salt concentration, sinks to the depths in the Labrador Sea and off the coast of Europe between Greenland and Scotland. Then it flows back towards the south. A more active THC contributes to higher North Atlantic temperatures.
Besides increasing the intensity of storms in the North Atlantic, warm phases like the one we are currently experiencing also generate more frequent hurricanes, whereas cold phases have the opposite effect. This alternation between warm and cold phases has now been supplemented by a new effect: the overall rise in temperature.

Intensity and frequency of hurricanes increases

The cold phases are not as cold as they used to be and the warm phases are getting hotter. 2005 made history with the highest value since 1880. Between July and September 2005, positive sea surface temperature anomalies of up to 2°C were registered in some parts of the tropical North Atlantic and the Caribbean, with average readings for.
January to November 2005 reaching record levels at several points on the map. Since the intensity and frequency of hurricanes increases with sea surface temperatures, the average number per year has also risen: from 2.6 to 4.1 hurricanes (Categories 3–5) between the last warm period and this one — an increase of around 60%.
A study by the Scripps Institute in 2005 reveals that the cause of this general warming is probably climate change, which, in turn, is due to human factors.

Effects on the insurance industry

Significantly more cyclones and a growing number of severe storms are also changing the prevailing hazard situations and loss distributions — factors of particular importance to the insurance industry.
The models used until spring 2006 were mainly based on all loss events since 1900, so that present loss levels are underestimated by insurance companies. Recent analyses by Munich Re have shown that the expected annual loss value increases on the basis of a distribution which only takes into account losses occurring in warm phases.
This is the great challenge for the insurance industry. It must respond to the current hazard situation and take it into account appropriately in its risk management.
The record losses generated by Hurricane Katrina also showed that some aspects of the total insured loss are still not sufficiently factored into loss models even today. Improvements must be made particularly with regard to the following:

• Modelling the effects of storm surge and flood
• The complex interrelation of aspects relating to business interruption covers that lead to higher losses
• The limited resources available to loss adjusters, which hampers settlement when there are large numbers of individual claims (no fewer than two million claims were filed after Katrina)
• The substantial increases in the price of materials and labour for the reconstruction work and the costs of alternative accommodation for people whose buildings have become uninhabitable
• More serious damage and delayed, more expensive repairs when the same region is hit by several storms within a short time
• The interruption of business activities in an entire region when this is aggravated by people returning to their homes slowly or not at all and by inefficient disaster management

These factors should also be taken into account in the insurance industry's risk management. Losses can be avoided if insurers additionally draw attention to the consequences of climate change and supports measures to counteract it.

All treatises from this site:
http://www.munichre.com/ (Choose English on the upper right side and than go to: --->TOPICS & SOLUTIONS --->Georisks)