Pardeep Pall & Myles Allen on Modeling Damaging Weather Events
Fast Breaking Papers Commentary, December 2011
Top-bottom: Pardeep Pall & Myles Allen |
Article: Anthropogenic greenhouse gas contribution to flood risk in England and Wales in autumn 2000
Authors: Pall, P;Aina, T;Stone, DA;Stott,
PA;Nozawa, T;Hilberts, AGJ;Lohmann, D;Allen,
MR |
Pardeep Pall & Myles Allen talk with ScienceWatch.com and answer a few questions about this month's Fast Breaking Paper in the field of Geosciences.
Why do you think your paper is highly
cited?
It addresses an issue of considerable topical and broad interdisciplinary concern: quantifying the impact that human influence on climate has on observed instances of extreme weather. It does so in a simple yet novel way, by comparing large ensembles of simulations of the world as it was and of the world as it might have been had a particular external driver (in this case, the human-induced increase in greenhouse gas concentrations) been absent.
As it happens, we also used a novel computing approach—employing spare CPU cycles donated by the general public to generate our large ensembles—but this does not seem to have been a particular factor in the paper's impact, indicating the growing acceptance of public resource distributed computing as a research tool.
Would you summarize the significance of your paper
in layman's terms?
"Autumn 2000 as observed (top) and as simulated in one of
the wetter members of the Pall et al. ensemble."
View an animated version of this figure.
We found that greenhouse gas emissions due to human activity substantially increased the odds of floods occurring in England and Wales in what was the record-wet autumn of 2000—with a likely increase in odds of about a doubling or more. To our knowledge, this is the first time that such an assessment has been made for a flood event.
We found this out by using a detailed computer climate model from the UK Met Office to simulate autumn 2000 weather: first in a world as it was in autumn 2000, and then in a parallel autumn 2000 world that might have been had no greenhouse gasses been emitted in the twentieth century.
To really pin down the difference in extreme weather between these two parallel worlds, we repeated our simulations thousands of times using the idle CPU time on personal computers that was volunteered by members of the public across the world via the climateprediction.net project. When we fed all these thousands of weather simulations into a flood model from Risk Management Solutions Ltd., we found, as stated, that the odds of floods occurring in autumn 2000 likely increased by about double or more.
What inspired this research?
The need to provide a method for objectively answering the question "has event X become more or less likely due to human-induced climate change," as opposed to the speculation and somewhat subjective debate that usually follows in the aftermath of a damaging weather-related event.
Addressing this question is becoming increasingly important given the amount of money put on the table by the UNFCCC Adaptation Fund intended to finance adaptation activities for those parties affected by climate change: many parties affected by a bout of damaging weather may subsequently claim they deserve this funding, so we need a method that disentangles whether the odds of such weather, and the associated damage, occurring have truly been affected by human-induced climate change, or whether those parties simply had bad luck.
Do you think the trend you reported on is
continuing and, if so, where do you see your research
leading in the future?
We only investigated a single event that occurred during a specific, relatively short, time period (autumn 2000), so we aren't saying anything about past or future trends, per se. However, our result is approximately consistent with a well-known basic physical mechanism for increased atmospheric moisture under global warming, and in that general sense one might expect to see wetter extreme weather of this type under future global warming. But again we stress that the odds of other events might be dominated by other mechanisms (such as changes in long-lived weather patterns), and so become less likely or even not be affected at all.
In short, we don't know the answer until we reapply our methodology to other events. That's now happening for whole host of past weather events via the new climateprediction.net experiment: weatherathome.net.
Indeed, we should be applying our methodology regularly and systematically
to assess the change in odds of all types of events, as soon as possible
after they happen—even if they're not that unusual. Plans are now
underway with Dr. Peter Stott at the UK Met Office, along with the
International Group on Attribution of Climate-Related Events (ACE), to make
such an assessment system operational.
Dr. Pardeep Pall
Climate Dynamics Group
Department of Physics
University of Oxford
Oxford, UK
and
Institute for Atmospheric and Climate Science
ETH Zurich
Zurich, Switzerland
Dr. Myles Allen
Climate Dynamics Group
Department of Physics
University of Oxford
Oxford, UK
(Now Professor of Geosystem Science at the School of Geography and
the Environment, University of Oxford)
KEYWORDS: ANTHROPOGENIC GREENHOUSE GAS, FLOOD RISK, ENGLAND, WALES, AUTUMN 2000, CLIMATE CHANGE, PRECIPTATION EXTREMES, UK, MODEL, EMISSIONS, RAINFALL, TRENDS.