24 August 2010

Normalized Earthquake Damages

Following an earlier post, I have received several email queries about global trends in earthquake damages, and how they compare with those related to weather, prompting this post.

The figure above, from my colleague here at the University of Colorado, geologist Roger Bilham, shows a very similar story to that observed with respect to normalizations of weather-related losses.  Specifically, while the aggregate number of deaths related to earthquakes has increased dramatically in recent decades (red bars), over the long-term, there has been no trend in deaths after normalizing for population growth (black line). (Note how one might arrive at a different conclusion using a short time period starting in 1980.)

Last February, Bilham had an excellent essay in Nature (PDF) on the lessons from the Haiti earthquake.   And his Mallet-Milne lecture, published last year (PDF), should be required reading.  In it he asks:
. . . why the knowledge of 9,000 years of city collapse in earthquakes, and a known cumulative death toll of more than 10 million people, has not led to safer construction everywhere.
In his paper he explains that large earthquake disasters are relatively concentrated in a small part of the Earth's surface:
[T]he odds of a city being damaged by an earthquake are not evenly distributed on our planet (McGuire 2004; Dilley 2005). Twelve percent of all fatal earthquakes are found along the margins of the eastern Pacific, and fully 85% of the world’s earthquake fatalities have occurred in the Alpine/Himalayan collision belt between western Europe and eastern Asia. This comparison is based on earthquakes since 1570, i.e. since the earliest historically recorded earthquakes in the Americas. Since then roughly 1,100 people have died in earthquakes each year in the western Americas and Carribean, compared to 8,900/year along the southern edge of the Eurasian plate. This concentration of most of the world’s fatal earthquakes occurs in less than 12% of Earth’s surface area—a 150◦ longitude band between London and Tokyo, between the equator and 45◦N.
Bilham explains that the world still has a long way to go to adapt to the known threat of earthquakes:
[W]e still live in a world where deaths are expected to accompany large earthquakes near cities. Within 30min of a damaging earthquake we can quantify the number of fatalities and injuries anticipated in settlements surrounding the epicentre, before news of actual deaths are known on the ground. That this is possible admits that we have a problem in our cities that needs to be fixed. The time to have undertaken this fix was in the era of construction that started in about 1950.We have a further 30 years left in this global building boom, but it is unlikely that earthquake resistance will occur where the structures are going up most rapidly. The focus of earthquake resistance efforts should clearly be in the places where fatalities have been historically the worst—in the western Americas and in the Alpine/Himalayan/Indonesian collision belt. Given the present recession of world economies, the cost of the fix is likely to prevent the expenditure of funds where it matters most, at a time when it matters most. This suggests that urban populations will continue to be killed by earthquakes in the foreseeable future, and in greater numbers than in the documented past.
The story is similar when looking at economic damage in the United States, as the figure below shows.
That figure shows normalized damages in the US 1900 to 2005 from Vranes and Pielke (2008, PDF).  Such a long-term damage record is not available globally.

References:

Bilham, R., 2009. The Seismic Future of Cities, Twelth Annual Mallet-Milne Lecture., Bull. Earthquake. Engineering, 1-49, 10.1007/s10518-009-9147-0

Bilham, R. 2010. Lessons from the Haiti earthquake, Nature 463, 878-879 (18 February 2010) doi:10.1038/463878a

Vranes, K., and R. A. Pielke, Jr. 2009. Normalized Earthquake Damage and Fatalities in the United States: 1900 - 2005. Natural Hazards Review 84-101, doi: 10.1061/ASCE1527-6988200910:384,