Tuesday, January 24, 2017

How Are Weather Extremes Linked to Climate Change?

Extreme event attribution: the climate versus weather blame game (Rebecca Lindsey, NOAA Climate, Dec. 15, 2016)

 Today we review a paper that describes the statistical process of attributing short term weather extreme events to the longer term changes underway as a result of climate change, whether that is due to natural or man-made burning of carbon fuels. It is important to understand the meaning of return periods. While the probability of a 100 year flood in a given year is 1%, the probability of the same flood over a period of 50 years is 40%. The blaming of an event on climate change depends on how good the observations of past events are, how well climate models can simulate the specific event and how well the physical processes are known and their association with climate change. Extended heat or cold events are more attributable than short term convective storms where the cross links are not as well understood.


Key Quotes:

“What is extreme event attribution?

“What can extreme event attribution tell us?

  • whether global warming made (or will make) an event more likely than it would have been without the rise in greenhouse gases from burning fossil fuels.
  • It can tell us if the average number of years between similar events is shorter or longer than it used to be.
  • It can tell us what the risk is for a given extreme weather event and if and how much global warming has increased that risk”
“Sigmas are used to describe the range of natural variability in a given climate or weather characteristic. For most types of climate data, there are many more observations close to the average (within 1 or 2 sigmas) than there are far away. So, if a climate expert describes a heavy rain event as a “5-sigma” event, she is talking about rainfall so extreme that it was 5 standard deviations away from the average high tide water level—way out at the tail end of the range of all values that have been observed.”

“So what does a hundred-year event mean?

  • The risk that a 100-year event will happen this year, or next year, or any single year is low: 1% chance that it will happen, 99% chance that it won’t. But the chance that it will happen within a given 20-year period is 18%. Within any 50-year period, 40%. Within any 100-year period, 63%. By the time 500 years have passed, there’s less than a 1% chance that such an event won’t have happened.”

“Why are some events more difficult than others to connect to global warming?

Attribution analysis depends on ‘three pillars’ of scientific knowledge:
  • The quality of the observational record,
  • The ability of models to simulate a given type of extreme event, and
  • How well we understand the physical processes that create an event and how global warming may influence those processes.
For event types where all three of these pillars are strong, our confidence in the results is higher. If any of the three pillars is weak, it becomes harder to conduct an attribution study, and our confidence in the results is lower. In general, scientists have the highest confidence for heat events because all three of the pillars are strong”

No comments:

Post a Comment