Stormy Connection to Climate Change
7 Jan, 2008 10:20 am
Severe thunderstorms produce destructive winds, hail, and sometimes tornadoes. Climate model simulations of future global warming suggest a higher frequency of the conditions conducive to severe thunderstorm formation in the United States.
In 2006, severe-thunderstorm winds, hail and tornadoes in the United States caused over $2.5B in damage to property and crops. Flash floods led to an additional $2B in losses. This significant impact to society would obviously change if severe thunderstorms became more (or less) frequent. An obvious concern—and the objective of our ongoing collaborative research at the U.S. Purdue Climate Change Research Center—is whether severe-thunderstorm frequency could be altered by anthropogenic global warming. Unfortunately for many residents of the United States, this concern appears to be warranted.
As my colleagues and I describe in our recent article published in the Proceedings of the National Academy of Sciences, climate-change scenarios suggest a higher frequency of U.S. severe thunderstorm conditions in the future. This finding was reached with the aid of climate models, sophisticated computer programs that simulate the atmosphere over tens and even hundreds of years. The projected amount of future greenhouse gases is a critical input parameter. The climate models are then used to output much of the same information about the atmosphere over a hypothetical 30-year period as weather forecast models do over the next 48 hours.
Of particular interest are the meteorological conditions, or ingredients, that foster severe thunderstorm formation. One key ingredient is warm, humid air near the ground. The climate models show that this thunderstorm energy source becomes increasingly more abundant during the latter part of this century. Strong “wind shear”—a change in wind strength with distance above the ground— is another key ingredient, and this tends to decrease in the future. But it’s the combination of the ingredients that matters, and the net effect is an atmosphere that becomes more conducive to severe thunderstorms in the future.
Incidentally, the ingredient threshold for what is considered “conducive” derives from weather-balloon observations of actual storms and their immediate environment. It would seem obvious that we would have also searched for trends in these same storm occurrences. Indeed, the United States has a historical severe thunderstorm and tornado record, but it is relatively short, and compiled from eyewitness reports. More problematic is that as population and also severe-weather awareness has increased over the years, so also has the number of storm reports. Additional layers of biases have been introduced to this database by other non-meteorological factors, further obscuring any meteorological signal. Models, on the other hand, provided us with a tool to separate out cause and effect.
This potential global warming impact varies considerably across the United States. For example, areas in the southern and eastern states currently prone to severe thunderstorm activity should see more. In simple terms, as the temperature increases, more water from the Gulf of Mexico and Atlantic Ocean is evaporated into the atmosphere and transported inland.
It is important to reiterate that our experimental approach focused on environmental-meteorological conditions rather than thunderstorms themselves. The technique of using conditions as storm proxies was developed and tested in a separate study by co-author Harold Brooks. He statistically compared the distribution of eyewitness storm reports with global “reanalysis” data, which is a blend of meteorological observations and weather forecast model output. An acknowledged weakness of this technique once applied to climate models is that there is always some uncertainty in whether a thunderstorm would be triggered in the model. This is because the climate models are too coarse to resolve individual thunderstorms: imagine how a 15-km diameter thunderstorm would go undetected between cities spaced every 50 km along a highway or motorway.
Therefore, our next step in this research is to use an interwoven mesh of progressively higher-resolution computer models, which will ensure thunderstorm detection within the model confines. This new and extremely computationally intensive approach will yield more detailed information, like the storm type and structure, which relates to a storm’s likelihood to spawn a tornado.
Related research by Anthony Del Genio and his NASA colleagues (published in Geophysical Research Letters), and also by Patrick Marsh and collaborators (published in Atmospheric Science Letters) add to what hopefully will become a rich collection of possible severe thunderstorm projections that can be considered by government officials and policy makers when estimating costs of further anthropogenic global warming.
References:
Trapp, R. J., N. S. Diffenbaugh, H. E. Brooks, M. E. Baldwin, E. D. Robinson, and J. S. Pal, 2007: Changes in severe thunderstorm environment frequency during the 21st century caused by anthropogenically enhanced global radiative forcing. Proceedings, National Academy of Sciences, 104, 19719-19723, doi: 10.1073/pnas.0705494104.
Brooks, H. E., J. W. Lee, and J. P. Craven, 2003: The spatial distribution of severe thunderstorm and tornado environments from global reanalysis data. Atmospheric Research, 67-68, 73-94, doi: 10.1016/S0169-8095(03)00045-0.
Del Genio, A. D., M-S Yao, and J. Jonas, 2007: Will moist convection be stronger in a warmer climate? Geophysical Research Letters, 34, L16703, doi: 10.1029/2007GL030525.
Marsh, P. T., H. E. Brooks, and D. J. Karoly, 2007: Assessment of the severe weather environment in North America simulated by a global climate model. Atmospheric Science Letters, 8, 100-106, doi: 10.1002/asl.159.
Dear Editor,
here some comments about the contribution ?Stormy Connection to Climate Change?.
I would like to suggest Authors to revise the article in order to better describe to the general public purposes and goals of their research work that is here described in a confused and unclear way.
In particular, although model approach have relevant importance in climatic simulation they usually need to be validated with observation experimental and satellite data (of course, where and when available) able to improve the model reliability. The Authors should mention this in the manuscript in order to give a more complete general view of the tools able to improve our confidence in the projections of future environmental changes as an answer to climate forcings.
In my opinion, the penultimate paragraph (Our experimental ? spawn tornado) is fundamental and has to be greatly improved.
[Response] Dr. Udisti, Your comments are appreciated. I have added a sentence that explicitly states our objective, which is to test whether severe thunderstorm frequency in the U.S. could be altered by anthropogenic global warming. Your are correct in questioning if our technique has been tested and compared with observations. As now stated in the article, the technique was developed and tested in a separate study by one of the co-authors using reanalysis data. A caveat concerning the application of the technique to climate models is now also stated. Unfortunately, we are limited in how much further testing can be done. Weather radar and satellite data can provide some useful information toward this end, but neither of these sensors can unambiguously determine the severity of a thunderstorm. Best regards, Jeff Trapp