The Impact of Climate Change on Natural Disasters - Reducing Disaster: Early Warning Systems for Climate Change - page 32

and projections of future trends (Seneviratne et al. 2012 ). In addition to the reduced
sample size that accompanies the narrowing of scale, a substantial amount of noise
is introduced by tropical cyclone track variability (e.g., Kossin and Camargo 2009 ).
Track variability is largely driven by random day-to-day variability in atmo-
spheric wind currents, but there are also linkages operating on a broad range of
time-scales in response to known modes of climate variability such as the El Niño -
Southern Oscillation (ENSO), among many others (Ho et al. 2004 ; Wu et al. 2005 ;
Camargo et al. 2007 , 2008 ; Kossin and Vimont 2007 ; Wang et al. 2007 , 2010 ;
Chand and Walsh 2009 ; Tu et al. 2009 ; Kossin et al. 2010 ; Chu et al. 2012 ). Even
relatively small changes in tropical cyclone tracks can lead to large differences in
associated impacts at any given location. For example, a group of islands can be
impacted by multiple tropical cyclones in a single season (e.g., the Philippines in
2009) and then remain largely unaffected for many subsequent years, even while the
total number of storms in the larger basin exhibits normal variability. This type of
clustering occurs randomly, but it can also occur through more systematic and per-
sistent modulation by climate variability.
Of particular relevance to longer-range disaster planning and risk mitigation
strategies aimed at specifi c intra-ocean-basin regions is how tropical cyclone tracks
may change in a warming world (Wang et al. 2011 ; Murakami and Wang 2010 ).
This needs to be considered in addition to questions about how basin-wide changes
in tropical cyclone frequency and intensity may change. For example, conditions
that lead to increased basin-wide activity can also shift tracks such that landfall
frequency may increase proportionally more or less, thus compounding or offsetting
the impacts. Presently, there has been more research toward understanding linkages
between climate change and tropical cyclone frequency and intensity than toward
understanding linkages between climate and track variability. These are both active
areas of study of great relevance to designing EWS.
Increasing trends in land-falling tropical cyclones have not yet been detected in
any of the regions that have been studied (Wang and Lee 2008 ; Chan and Xu 2009 ;
Kubota and Chan 2009 ; Lee et al. 2012 ; Weinkle et al. 2012 ). A statistically signifi -
cant decreasing trend in the number of severe tropical cyclones making landfall over
northeastern Australia since the late nineteenth century has been identifi ed by
Callaghan and Power ( 2010 ). Contrarily, a signifi cant positive trend has been identi-
fi ed in the frequency of extreme sea-level anomaly events along the United States
East and Gulf Coast in the period 1923-2008, and this trend is argued to represent
a trend in storm surge associated with land-falling hurricanes (Grinsted et al. 2012 ).
As stated above, these trends likely represent some combination of basin-wide fre-
quency changes and track shifts (e.g., Bromirski and Kossin 2008 ). The difference
between Callaghan and Power ( 2010 ), who show a long-term decreasing trend in
Australian landfall events and Grinsted et al. ( 2012 ), who suggest a long-term
increasing trend in storm surge associated with US landfall events, emphasizes the
challenge of understanding and projecting changes in tropical cyclones that are
most relevant to coastal impacts.
Human-caused increases in greenhouse gases have very likely contributed to the
observed increase in tropical ocean temperatures over the past century (Santer et al.