With the industry facing increasingly ferocious hurricane seasons, Dr Eberhard Faust details the conflicting arguments from the scientific community on what could be behind this period of heightened activity.

The elevated hurricane activity in the North Atlantic since 1995, which not only contributed to record losses from natural catastrophes in 2004 and 2005 but also broke a few meteorological records in 2005 with 15 hurricanes, 27 named storms, and the strongest hurricane ever recorded, have made an adjustment of damage models in the insurance industry unavoidable for several reasons. First, greater consideration must be given to the secondary loss-aggravating aspects of very large natural catastrophes, as observed in the case of Katrina. These include factors such as demand surge and economic disruption. Above all, however, the evident rise in Atlantic cyclone activity, reflected in a higher probability of loss or damage, must be suitably factored into the insurance industry's loss models. For this reason, and with a view to being aware of what to expect in the coming decade, it is important to understand the climatological mechanism that is responsible for this increase in activity.

Sea surface temperatures

Scientific findings have established a strong link between increases in Atlantic tropical cyclone activity and the recent increases in tropical Atlantic temperatures over the last few decades. The sea surface temperature (SST) of tropical ocean regions - one of the major factors causing the development and intensity of tropical cyclones - has risen globally by about 0.5 deg C since 1970 (Agudelo and Curry, 2004). One might question the relative importance of sea surface temperatures and the warmth of the uppermost water layers - as the primary source of energy and moisture - in comparison with other atmospheric preconditions for hurricane development (such as low vertical tropospheric wind shear or the availability of cyclonic vorticity or of moist static energy in the main development region of the tropical Atlantic - the region where hurricanes develop).

According to a paper studying the period 1970 to 2004, there are strong indications that the increasing severity of hurricanes (Saffir-Simpson categories 4 and 5) is directly linked to the long-term trend in tropical SSTs, while the other aspects of the tropical environment, though responsible for short-term variations in hurricane intensity, do not contribute significantly to the increasing global trend in hurricane intensity over time (Hoyos et al, 2006; Webster et al, 2006). Although there is no distinctive global trend in the number of cyclones occurring every year, the percentage of category 4 and 5 hurricanes has been increasing since 1970 and indeed has more than doubled since then (see figure 1). These findings are borne out by the close correlation found by Emanuel between the intensity of tropical cyclones in the North Atlantic and West Pacific - measured in terms of the annual aggregate of wind power release - and SSTs (Emanuel, 2005). The close correlation between sea surface temperatures and global tropical cyclone intensities, which has been established by these studies, has been convincingly shown to apply to Atlantic hurricane frequencies as well. In particular, tropical cyclone counts correlate very closely with August-through-October average sea surface temperatures in the main development region (Mann & Emanuel, 2006).

Over the last 12 months, there has at times been a heated debate within the research community on the issue of a climatic mechanism behind the observed changes in sea surface temperatures and hurricane activity. This debate is still in full swing. A number of researchers say that a natural climate oscillation is responsible for the fact that in 1995 the North Atlantic entered a phase of extremely warm sea surface temperatures and, as a result, higher hurricane activity. Other scientists go a step further than assuming an exclusively natural mechanism: their view recognises the phenomenon of a natural climate oscillation but holds that it is superimposed by a progressive warming process which is due to anthropogenic (or man-made) climate change, which has been intensifying the natural warming of the oceans since the mid-1990s. The strongest opposition to a purely natural driver is maintained by researchers who see anthropogenic warming of the atmosphere and the ocean as the main reason for the current level of hurricane activity.

A natural cause?

Goldenberg et al (2001) were the first to attribute the increase in Atlantic hurricane activity in the mid-1990s to a natural climatic cycle which makes the sea surface temperatures of the Atlantic basin oscillate on a multi-decadal (50-80 year) timescale (see also Enfield et al, 2001). This phenomenon is called the Atlantic Multi-decadal Oscillation (AMO) and in the 20th century had a periodicity of about 65 years (Knight et al, 2005). The associated cold and warm phases are characterised by a margin of deviation of around 0.5 deg C in sea surface temperature. Modelling studies have identified a plausible mechanism related to the ocean's large-scale currents - known as thermohaline circulation - and its intrinsic multi-decadal variability (Vellinga and Wu, 2004; Sutton and Hodson, 2005; Knight et al, 2005). The current warm AMO phase started in 1995 and is expected to last for another ten to 20 years.

Interpreting the cyclical temperature fluctuations on this basis produces the following result. Warm phases produce a distinct increase in hurricane frequency and also more intense storms, whereas cold phases have the opposite effect (Knight et al, 2005; Pielke et al, 2005). In the current warm phase, the number of major hurricanes (Saffir-Simpson categories 3 to 5) per year has already reached an average of 4.1 whilst in the previous cold phase (1970-1994) this figure was only 1.5 (meaning an increase of approximately 170%). Of course, a definite value for the average annual level of activity for the whole warm phase can only be given when this phase has ended and the 11 years of data currently available (since 1995) is insufficient compared to the 45 years of data available for the 1926-1970 warm phase. However, the 11 years of data is all we have available at present and there is no reason to believe that it is not representative of the phase as a whole.

Natural and man-made drivers?

According to a model study performed by Barnett et al (2005) human induced climate change can be demonstrated to be the most significant driver of the increase in global sea surface temperatures. This is illustrated by a comparison of temperature trends since 1960 in all ocean basins and model simulations. The findings are consistent with the observation of a long-term warming process superimposing the multi-decadal oscillation in the North Atlantic. It is most probably caused by climate change - the resulting linear warming trend since 1870 amounts to 0.036 deg C per decade for the tropical North Atlantic. Hence, sea surface temperatures and the level of hurricane activity - frequency and intensity - appear to increase from one warm phase to the next, equally from one cold phase to the next. The increase in the average number of major hurricanes per year from 2.6 to 4.1 from the previous warm phase (mid-1920s to 1970) to the current one means an increase of about 60%. However, this increase has still to be confirmed by comparative studies using climate models (model studies) which as yet do not provide any conclusive results regarding changes in frequency in a warmer climate. Also, the projected changes in intensity are on the moderate side (Knutson and Tuleya, 2004).

Mainly global warming?

As mentioned above, Emanuel found a close correlation between the intensity of tropical cyclones in the North Atlantic and West Pacific - measured in terms of the annual aggregate of wind power release - and sea surface temperatures in the tropical ocean regions (Emanuel, 2005). Using different scientific approaches, both Webster et al (2006) and Sriver and Huber (2006) found that a trend towards increasingly intense cyclones may be observed throughout the world and - in the same way as Emanuel - implicated anthropogenic climate change as the main cause. All these studies were rejected by the opposing camp (Pielke 2006; Landsea 2006; Klotzbach 2006).

In their latest study (June 2006), Mann and Emanuel go one step further and argue that the AMO's impact on sea surface temperatures in the tropical North Atlantic and hurricane activity are more or less negligible. According to their statistical analyses, 85.5% of the decadal variability in sea surface temperatures in the Atlantic main development region (August-October) may be explained by the superimposition of global mean surface temperature (August-October) and the increasing influence of man-made aerosols from 1950 onwards. Aerosols, or airborne particulates, partly reflect solar irradiance back into space and play a part in low level cloud formation (tiny aerosol particles can "seed" clouds by providing the "nuclei" around which cloud droplets are formed). Both processes result in cooling of the lowest parts of the atmosphere - especially effective in the main development region.

If this theory is correct, the phase of cooler sea surface temperatures in the second half of the 20th century appears not to be the expression of a natural oscillation but the antidromic cooling influence of industrial aerosol emissions on the progressive anthropogenic warming trend. In their argument against the AMO hypothesis the authors also cite the unresolved problem that existing measurements tend to point to a weakening of thermohaline circulation in the North Atlantic since the 1950s (Bryden et al, 2005) whereas an acute AMO warm phase should be associated with stronger thermohaline circulation.

These critical arguments certainly carry weight. At the same time, however, model studies using the Hadley Centre's climate model have identified a multidecadal oscillation very similar to the AMO and an associated mechanism (Vellinga and Wu, 2004; Knight et al, 2005). The multidecadal oscillation is said to have been generated in other models too, such as Germany's ECHAM5/MPI-OM.

Weighing up all sides of this ongoing scientific discussion, it would appear at present that anthropogenic climate change is probably involved in both the elevated sea surface temperatures and the higher level of hurricane activity in the tropical North Atlantic. Whether these changes also involve a natural climate oscillation (AMO) or - as Mann and Emanuel believe - only anthropogenic climate change and antidromic aerosol effects, is a question that still needs to be answered in the course of the scientific debate.

Implications for the industry

The extreme hurricane losses of 2004 and 2005 must be seen in connection with the phase of high activity in the North Atlantic since 1995. The average annual frequency of major hurricane landfalls has increased by about 230% from the phase of cooler tropical sea surface temperatures between 1971 and 1994 to the current warm phase. As far as the coming decade is concerned, the insurance industry would be well advised to expect the high level of hurricane activity in the North Atlantic to persist. Most loss models have already been adjusted, even if the near-term view, ie the consideration of the current climate-driven phase of high activity, is not accepted across the board as the appropriate standard operating mode for the models. There has been no succession of extreme landfall and loss years since 1995 and, fortunately, this is likely to be the case in the near future too. Nevertheless, we should assume a higher landfall and loss probability and make this a permanent precondition of risk management.

- Dr Eberhard Faust is head of climate risks for Geo Risks Research at Munich Re.