The Impact of Climate Change on Natural Disasters - Reducing Disaster: Early Warning Systems for Climate Change - page 33

2006 ; Kunkel et al. 2008 ). Shorter-term decadal variability in regions where tropical
cyclones form and track is generally dominated by natural variability (e.g., Ting
et al. 2009 ; Camargo et al. 2013 ; Zhang et al. 2013 ) and factors such as volcanic
eruptions (e.g., Thompson and Solomon 2009 ; Evan 2012 ), changes in natural par-
ticulates such as African dust (e.g., Evan et al. 2009 , 2011a , 2012 ), and changes in
human-caused particulate pollution (e.g., Mann and Emanuel 2006 ; Baines and
Folland 2007 ; Chang et al. 2011 ; Booth et al. 2012 ; Evan et al. 2011b ). There is
presently some debate about the effect that globally increasing greenhouse gases
has on tropical cyclones versus the effect of regional changes in particulate concen-
trations (e.g., Emanuel and Sobel 2013 ). Increases in globally well-mixed green-
house gases are argued to be less effective at making the tropical environment more
conducive to tropical cyclone formation and intensifi cation compared to the more
local effects caused by changes in particulate pollution (Vecchi and Soden 2007 ;
Ramsay and Sobel 2011 ; Camargo et al. 2013 ), but both factors need to be consid-
ered for short- and long-range planning.
While there is currently debate about the relative contributions of natural versus
human-caused changes in tropical climate on 10-40-year time-scales, there is mount-
ing evidence that human-caused particulate pollution has played a substantial role in
some of the recent marked increases in tropical cyclone activity. In the tropical North
Atlantic Ocean, the reduction of pollution aerosols since the United States Clean Air
Act and Amendments during and after the 1970s (with further contribution from the
European Commission's Air Quality Framework Directive) has been linked to tropi-
cal sea surface temperature increases and associated increases in tropical cyclone
activity. This linkage has been related to the direct effect of reduced atmospheric
dimming allowing more sunlight to reach the ocean surface (e.g., Mann and Emanuel
2006 ), and to the indirect effects of reduced cloud albedo (Baines and Folland 2007 ;
Booth et al. 2012 ; Dunstone et al. 2013 ). In the Northern Indian Ocean, black carbon
particulate pollution has been linked to changes in sea surface temperature gradients
(Chung and Ramanathan 2006 ; Meehl et al. 2008 ), which has weakened the mean
vertical wind shear in the region. Evan et al. ( 2011b ) linked the reduced wind shear
to the observed increase in the number of very intense storms in the Arabian Sea,
including fi ve very severe cyclones that have occurred since 1998, killing over 3,500
people and causing over $6.5 billion in damages (in 2011 US dollars).
As with observational analyses, confi dence is compromised when numerical pro-
jections of tropical cyclone activity are reduced from global to regional scale (IPCC
SREX Box 3-2; Seneviratne et al. 2012 ). When assessing the results of all available
model simulations, it is likely that global tropical cyclone frequency will decrease
slightly in the twenty-fi rst century, but there is little confi dence in this on regional
scales (e.g., Ying et al. 2012 ). Mean tropical cyclone intensity and rainfall rates are
projected to increase with continued warming, and the models tend to agree better
when projecting these measures of activity (Knutson et al. 2013 ). Models that are
capable of producing very strong cyclones usually project increases in the frequency
of the most intense cyclones (Emanuel et al. 2008 ; Bender et al. 2010 ; Knutson
et al. 2010 ; Yamada et al. 2010 ; Murakami et al. 2012 ; Knutson et al. 2013 ). This
measure is highly relevant to physical and societal impacts, compared with mea-
sures of overall storm frequency or mean intensity, which can be dominated by