Assessing Direct Losses

Energy2green Wind And Solar Power System

Wind Energy DIY Guide

Get Instant Access

There are two methods to assess how a change in hurricane risks, including surge and wind, would translate into destruction of buildings, equipment and infrastructure. The first one is based on physical models, while the second one is purely statistical.

In most cases, physical models have been developed by consulting companies that advise the insurance industry and help them assess their level of risk. These models are based: (i) on a comprehensive dataset of the exposure, i.e. the characteristics and value of the property exposed to a hazard at a fine spatial resolution; and (ii) on vulnerability models, which relate wind speed, flooding depth and any other physical description of a disaster, to a damage ratio, which is the share of the exposure that is destroyed or damaged for a given hazard level. These models describe a hurricane by its wind field and storm surge and estimate damages to properties. The drawback of these models is the amount of data they require - this information is for instance not available for developing countries - and the fact that it is particularly difficult to create scenarios to project exposure over long timescales.

Statistical models, on the other hand, can be very simple. They use past hurricanes, and data on the resulting direct economic losses to create statistical relationships able to predict future damages (Howard et al., 1972; Nordhaus, 2006; Hallegatte, 2007a; Sachs, 2007).

In Hallegatte (2007a) each coastal county is considered separately, using the normalized damages proposed by Pielke and Landsea (1998) and refined in Pielke et al. (2007). In this data set, historical hurricane losses, due to wind and surge, are normalized to remove the influence of changes in population, wealth and price level. As a consequence, for a given hurricane, the dataset estimates the amount of losses that would result if it occurred today. One can assume that the normalized losses due to a hurricane making landfall in the coastal county i depend upon the hurricane intensity and on the vulnerability of the landfall county. This can be approximated by:

L = a • Wd where W is the maximum wind speed of the hurricane (as provided for instance by the HURDAT database), and ai is a vulnerability parameter, which describes the vulnerability of the entire U.S. to a hurricane making landfall in the coastal county i (even if all the losses are not only in the county i). It is interesting to note that losses are assumed here to depend on the cubed wind speed (d = 3), which is a proxy for the energy dissipated by the hurricane. This relationship is very conservative compared with other analyses: the statistic analysis by Sachs suggests a much larger value for d, of 6.3; Howard et al. (1972) suggest a mean value of 4.36; and Nordhaus (2006) cites values between 4 and 9. Of course, the larger this value, the larger is the sensitivity of direct losses to a change in hurricane intensity, and our analysis can be considered as conservative on this point.

In spite of data availability problems, this method provides an assessment of local vulnerability parameters, which are reproduced in Fig. 3. Most of these values are consistent with what one can expect, with large values in New York, Miami, New Orleans, Galveston and Currituck County, but an unrealistic value appears in the Lee County. As an illustration of data availability issues, negative bars show counties where no landfall occurred in the recent past and where no vulnerability information can be estimated.

Using this assessment of county vulnerability, one can estimate the mean direct losses due to a set of synthetic hurricanes. For the present climate, this method gives a value of 1578 million U.S.$ per landfall and 980 million U.S.$ per track. These

95,426

17500

12500

2500

-2500

95,426

17500

12500

2500

County index (s)

Fig. 3 Historical local vulnerability coefficients, coastal county per coastal county. The negative bars represent the counties where no data is available

County index (s)

Fig. 3 Historical local vulnerability coefficients, coastal county per coastal county. The negative bars represent the counties where no data is available values are close to historical values, which are 1833 million U.S.$ per landfall and 1005 million U.S.$ per track. Hallegatte (2007a) shows that the difference between values estimated from the model and the historical values is not statistically significant.

Using these values, and assuming that there is no change in price level, wealth and population, one can produce an estimate of how a change in hurricane intensity or frequency would translate in terms of direct losses. Here, with the Modified Climate assumption with a 10-percent increase in potential intensity, annual direct hurricane losses increase by 54 percent, from about $8 billion to about $12 billion. This cost could then be adjusted as a function of expectations about how the region will develop (how the population will change, how wealthy this population will be, where and how the population will settle). Regardless, this 54-percent increase in the mean annual normalized loss is significant, but it does not seem really threatening for this region, which is one of the richest in the world. This change, nevertheless, should cause an equivalent rise in insurance premiums that can create local issues of insurance affordability (e.g., see RMS, 2006, on the insurability of New Orleans).

Possible changes in the frequency of extreme hurricanes, however, are more worrying: among the 3000 tracks produced by the hurricane model in the present climate, only 59 cause direct losses exceeding $10 billion, and only 4 exceeding $50 billion. Among the 3000 tracks produced for a climate in which potential intensity has been increased by 10 percent, as many as 99 cause direct losses above $10 billion, and 10 cause direct losses above $50 billion. Figure 4 shows the major increase in the probability of the most intense and destructive hurricanes.

Direct losses in US$b

Present climate ■ Modified climate (+10% PI)

Direct losses in US$b

Fig. 4 Cumulative distribution of hurricane losses, including wind and surge, in the present climate and in a climate in which the potential intensity has been increased by 10 percent. For instance, this figure shows that the 0.10% probability event causes about $55 billion losses in the present climate but up to $90 billion in the modified climate

1000 100.00%

Fig. 4 Cumulative distribution of hurricane losses, including wind and surge, in the present climate and in a climate in which the potential intensity has been increased by 10 percent. For instance, this figure shows that the 0.10% probability event causes about $55 billion losses in the present climate but up to $90 billion in the modified climate

More than the change in annual losses, therefore, it is the doubling of catastrophic event frequency that seems the most problematic for society and the economy.

One important drawback of this assessment is that sea level rise is neglected, even though it is likely to cause a large increase in coastal flooding risks in the future. Nicholls et al. (2007), for instance, found that a 50 cm sea level rise would double the population exposed to the 100-yr flood event in Miami and Greater New York. A more complete analysis would have to take into account this important factor.

Was this article helpful?

0 0
Renewable Energy Eco Friendly

Renewable Energy Eco Friendly

Renewable energy is energy that is generated from sunlight, rain, tides, geothermal heat and wind. These sources are naturally and constantly replenished, which is why they are deemed as renewable.

Get My Free Ebook


Post a comment