by Judith Curry
When might we see Category 6 hurricanes? 60C surface temperatures?
The IPCC Special Report on Extremes (SREX) included the following figure:
This diagram says that the distribution of extreme events in a changed climate can have a shifted (higher) mean, increased variability, or a changed shape. The shifted shape diagram shows a skewed diagram, with the tail on the right. Such distributions have given rise to much angst about fat tails (see recent post).
What if the distribution changed in a warming climate to have a ‘snub nose’ on the right, rather than a tail?
At the recent US-UK Workshop on Climate Science Needed to Support Robust Adaptation Decisions, Greg Holland gave a presentation entitled How Extremes Adjust to Climate Variability and Change: Extreme Temperatures. A summary of the presentation is provided below, with selected ppt slides:
Before getting to extreme temperatures, Holland first looks at hurricane intensity (recall Holland was a coauthor on the 2005 paper by Webster et al. Changes in Tropical Cyclone Number, Duration, and Intensity in a Warming Environment)
It looks like the original idea of Webster et al. (2005) is holding up, with an increase in the % of category 4 and 5 storms; however the extreme has barely moved. The presentation also shows hurricane model simulations of the 2012 season, then with imposed changes to the SST, demonstrating that we are nearing a saturation level beyond which there may not be further significant increases in Cat 4-5 proportions.
Holland then considers summer (austral) temperatures at two locations in Australia – Giles and Melbourne. The shape of the distributions for the two cities are quite different, but both have a max near 47C.
Looking at the desert location (Giles), the PDF is skewed (snub nosed to the right), with increasing hot days but no extension beyond the max of 46C.
Holland then asks the question ‘How high can we go?’ Holland argues that why the distribution may be changing with greater frequencies of more intense tropical cyclones and hot days, the maximum values are increasing very slowly with climate change.
Holland makes a thermodynamic argument (see full presentation for details) whereby the overall vertical structure of the tropical atmosphere precludes big changes to the desert maximum temperatures.
With regards to Melbourne, Holland demonstrates that the temperatures are dominated by advection from other regions, so can experience increase # of hot days as well as increase in extreme value.
JC comments: This paper was to me the most surprising one of the Workshop. It has implications for economic arguments that rely on ‘fat tail’ disasters, as well as for weather/climate predictions of extreme events.
The ideas presented here are particularly relevant to my company’s (Climate Forecast Applications Network) efforts to predict heat waves, especially when models are predicting a record-breaking value, that is outside the range for the recent 20 year period for which we have hindcasts to calibrate the forecasts using observations.
Improved understanding of the historical distributions of different extreme events at different locations is important for thinking about physical constraints that might lead to snub nosed distribution rather than fat tails.