|
Titel |
On predicting future economic losses from tropical cyclones: Comparing damage functions for the Eastern USA |
VerfasserIn |
Tobias Geiger, Anders Levermann, Katja Frieler |
Konferenz |
EGU General Assembly 2015
|
Medientyp |
Artikel
|
Sprache |
Englisch
|
Digitales Dokument |
PDF |
Erschienen |
In: GRA - Volume 17 (2015) |
Datensatznummer |
250101537
|
Publikation (Nr.) |
EGU/EGU2015-695.pdf |
|
|
|
Zusammenfassung |
Recent years have seen an intense scientific debate of what to expect from future tropical
cyclone activity under climate change [1,2]. Besides the projection of cyclones’ genesis
points and trajectories it is the cyclone’s impact on future societies that needs to be
quantified.
In our present work, where we focus on the Eastern USA, we start out with a
comprehensive comparison of a variety of presently available and novel functional
relationships that are used to link cyclones’ physical properties with their damage caused on
the ground. These so-called damage functions make use of high quality data sets
consisting of gridded population data, exposed capital at risk, and information on
the cyclone’s extension and its translational and locally resolved maximum wind
speed.
Based on a cross-validation ansatz we train a multitude of damage functions on a large
variety of data sets in order to evaluate their performance on an equally sized test sample.
Although different damage analyses have been conducted in the literature [3,4,5,6], the
efforts have so far primarily been focused on determining fit parameters for individual data
sets. As our analysis consists of a wide range of damage functions implemented on identical
data sets, we can rigorously evaluate which (type of) damage function (for which set of
parameters) does best in reproducing damages and should therefore be used for future loss
analysis with highest certainty.
We find that the benefits of using locally resolved data input tend to be outweighed by the
large uncertainties that accompany the data. More coarse and generalized data input therefore
captures the diversity of cyclonic features better. Furthermore, our analysis shows that a
non-linear relation between wind speed and damage outperforms the linear as well as the
exponential relationship discussed in the literature.
In a second step, the damage function with the highest predictive quality is implemented
to predict potential future cyclone losses for the Eastern USA until the year 2100. The
projection is based on downscaling five different GCM model runs for the RCP8.5 scenario,
as conducted by Emanuel et al.Â[7], and accounts for population and GDP changes relying
on the newly developed Shared Socioenonomic Pathways (SSPs) [8]. We hereby
contribute valuable input to the scientific community as well as the societies at
risk.
The possibility of extending this work to different regions in order to access the future
impact of tropical cyclones on a global scale will also be discussed.
References
[1]ÂÂÂThomasÂR. Knutson, JohnÂL. McBride, Johnny Chan, Kerry Emanuel, Greg
Holland, Chris Landsea, Isaac Held, JamesÂP. Kossin, A.ÂK. Srivastava, and
Masato Sugi. Tropical cyclones and climate change. Nature Geoscience,
3(3):157–163, 2010.
[2]ÂÂÂRobert Mendelsohn, Kerry Emanuel, Shun Chonabayashi, and Laura
Bakkensen. The impact of climate change on global tropical cyclone damage.
Nature Climate Change, 2(3):205–209, 2012.
[3]ÂÂÂSilvio Schmidt, Claudia Kemfert, and Peter Höppe. The impact of
socio-economics and climate change on tropical cyclone losses in the USA.
Regional Environmental Change, 10(1):13–26, 2009.
[4]ÂÂÂWilliamÂD. Nordhaus. The Economics of Hurricanes and Implications of
Global Warming. Climate Change Economics, 01(01):1–20, 2010.
[5]ÂÂÂKerry Emanuel. Global Warming Effects on U.S. Hurricane Damage. Weather,
Climate, and Society, 3(4):261–268, 2011.
[6]ÂÂÂRichard J. Murnane and James B. Elsner. Maximum wind speeds and US
hurricane losses. Geophysical Research Letters, 39(16):707, 2012.
[7]ÂÂÂKerry Emanuel. Downscaling CMIP5 climate models shows increased tropical
cyclone activity over the 21st century. Proceedings of the National Academy of
Sciences of the United States of America, 110(30):12219–24, 2013.
[8]ÂÂÂDetlef P.Âvan Vuuren, Keywan Riahi, and Richard Moss. A proposal for a
new scenario framework to support research and assessment in different climate
research communities. Global Environmental Change, 22(1):21–35, 2012. |
|
|
|
|
|