News

3Qs: With Sandy, climate change ‘loads the dice’

10/31/12

Though it’s dif­fi­cult to tie a spe­cific storm like Hur­ri­cane Sandy to the phe­nom­enon of cli­mate change, Auroop Gan­guly, an asso­ciate pro­fessor of civil and envi­ron­mental engi­neering at North­eastern, says “It is becoming increas­ingly harder to ignore the empir­ical evi­dence” that human-​​influenced cli­mate change affects the weather.

What role does climate change play in the development of superstorms like Hurricane Sandy, which battered much of the East Coast with high winds and floodwaters?

Cli­mate change loads the dice. How­ever, the sci­en­tific com­mu­nity has usu­ally been extra cau­tious before attributing any one weather extreme event to cli­mate change. This is espe­cially true for extremes such as hur­ri­canes where the link with cli­mate change is less well understood.

The cur­rent state of cli­mate sci­ence allows the attri­bu­tion of cer­tain types of weather extremes to cli­mate change, but pri­marily in terms of average sta­tis­tics. Thus, we can say with some con­fi­dence that heat waves have been inten­si­fying, and will con­tinue to do so at con­ti­nental to global scales. There have been recent devel­op­ments in our ability to attribute pre­cip­i­ta­tion extremes to cli­mate change. We are also get­ting better at attri­bu­tions of more local­ized extremes. In fact, one British group was able to attribute flood risks for autumn of 2000 in Eng­land and Wales to human-​​induced emis­sions. How­ever, hur­ri­canes are much harder, and one extreme event remains dif­fi­cult to attribute.

That being said, it is becoming increas­ingly harder to ignore the empir­ical evi­dence. Based on cur­rent under­standing, the more intense hur­ri­canes are expected to inten­sify fur­ther on an average. What we are seeing now cer­tainly does not dis­agree with our cur­rent under­standing of the impli­ca­tions of human-​​induced global warming.

Should storms like Sandy, which arrive late in hurricane season and pack a powerful punch, be considered something of a new normal? And if so, how must those in areas likely to be impacted change how they think about these storms?

Cer­tainly there is a need to increase pre­pared­ness levels for more intense hur­ri­canes and per­haps extended hur­ri­cane sea­sons. This is a per­fect example where the cli­mate sci­ence may not yet be as pre­cise as we would like, but impor­tant pre­pared­ness deci­sions still need to be made with some urgency. This is also an example where delayed deci­sions may cost human lives, destroy crit­ical infra­struc­tures and damage economies. The impor­tance of adap­ta­tion and pre­pared­ness in this con­text cannot be overstated.

Inten­si­fying hur­ri­canes, rising sea levels and growing storm surges, in con­junc­tion with growth of pop­u­la­tion and infra­struc­tures in regions at risk, point to the need for sus­tain­able urban and coastal plan­ning, and includes the need to develop greater resilience to nat­ural haz­ards. Advances in weather and cli­mate sci­ences, as well as early-​​warning sys­tems, can help save lives and pre­serve infra­struc­tures. Com­mu­nity and infra­struc­tural resilience, as well as emer­gency pre­pared­ness, are crit­ical and may pre­vent haz­ards from becoming dis­as­ters. Mea­sures ranging from effec­tive design strate­gies to public edu­ca­tion and evac­u­a­tion plan­ning, among others, can help.

We may not be able to attribute single storms to cli­mate change or pin­point the exact loca­tions of land­falls in advance, but we can still develop new adap­ta­tion and pre­pared­ness strate­gies. As our sci­ence improves, such strate­gies may be fine-​​tuned, but mounting evi­dence sug­gests that inac­tion may turn out to be cost­lier than action.

How can scientists tell if extreme weather incidents like Sandy are one-off storms or if they are part of a larger shift in climate? 

Attributing extreme weather to cli­mate change rather than nat­ural cli­mate vari­ability requires metic­u­lous analysis of data, typ­i­cally from both obser­va­tions and model sim­u­la­tions. The latter may include global and regional cli­mate models, as well as more local­ized numer­ical models to sim­u­late weather extremes. The sta­tis­tical tech­niques are rel­a­tively well devel­oped in cer­tain sit­u­a­tions but need fur­ther devel­op­ment for other conditions. The sci­ence of attri­bu­tion has been steadily improving, but what has remained con­stant is the need for detailed and metic­u­lous shifting through data and run­ning of models to delin­eate pos­sible causality.

The type of “fin­ger­prints” that cli­mate sci­en­tists look for are, for example, co-​​occurrence of observed pat­terns and events that are sta­tis­ti­cally robust; prece­dence struc­tures sug­ges­tive of cause-​​effect rela­tions; and sig­nif­i­cant changes in pro­jec­tions from mul­tiple model runs. The metic­u­lous studies leading to delin­eation of causality or attri­bu­tions can take a while and would typ­i­cally require gath­ering of rel­a­tively large vol­umes of observed data and gen­er­a­tion of model sim­u­la­tions. While we are grad­u­ally get­ting better, there is still some way to go before we can con­fi­dently make more pre­cise attributions.