Yonatan Zunger (zunger) wrote in climatepapers,
Yonatan Zunger

Clarification about fig. 20

After talking with some people, I realized that the explanation of the critical figure 20 in the big paper was somewhere between "unclear" and "just plain wrong." This is the figure that actually summarizes predictions for the next 100 years, so it's probably worth looking at. :) Here's the right explanation.

First of all, separate the year into four seasons. If you take the daily temperature data for any spot on Earth and plot a histogram over a long period of time, you'll get a sum of two Gaussians plus a general blur; but if you select out just the data for summer (June-July-August) or winter (December-January-February) each of those is a single bell curve. (My post about Lincoln, NE shows exactly this graph: the daily temperatures in Lincoln for every day during summer in the past 75 years or so. It's a pretty little bell curve with a mean of 90F and a sigma of about 8F. This sigma represents the average day-to-day variation of temperatures.

(Now, that posting explained how a shift in the local mean temperature translates into a change in number of hot days per year. The result of that calculation was that raising the average temperature in Lincoln would increase the number of days above 95F [the temperature that kills corn] by roughly 1 week per summer for every 1C temperature change. So this is the function that turns a change in the average summertime temperature into a change in local farming conditions.)

Next, look at the graph in the top right of figure 20a. This graph shows the average year-to-year variation. Take the average temperature for all the days in summer for each individual year; if you plot a histogram of those, you get another nice bell curve. The mean is in the same place as the mean of daily temperatures (90F) but the width of this bell curve represents how much each year differs from another. From looking at this plot, that difference for Lincoln is about 0.5C, which means that each summer is much like the rest.

Now, look at the third column of figure 20b. This shows the change in average temperatures, in degrees Celsius, for every spot on the earth in the five models. For Lincoln, this shows a change of between +4 and +12C, which means that the center of the average year's day-to-day bell curve (the same plot as in my post) would move 4-12C to the right. (Which, btw, would be really bad)

The fourth column of figure 20b compares the third column with the year-to-year variations plotted in the top right graph; i.e., if that graph shows a 5 for some point on Earth, it means that the summer average temperatures for a year around 2100 in that spot would be a 5σ event by the standards of the past century -- i.e., an average temperature which on its own would have occurred once every 3.5 million years without climate change. In essence, this column tells you two things: first, that the climate changes predicted in column 3 are statistically significant anomalies (they represent average temperatures which normally would never, ever occur). Second, since the year-to-year variation of temperatures is what all the local plants, animals and people have adapted to, this plot tells you how far from regular experience the new situation will be.

So based on this, a value of 1 or 2 on a plot in the fourth column isn't too bad: 1 means "the average year will look like the sort of year that happens once every 6 years or so nowadays," and 2 means "the average year then will look like a once-in-44-years heat wave." Both can be uncomfortable (spoken just as a nasty heat wave is passing where I live), but neither is a disaster. To take Lincoln as the example again, the average year-to-year variation there is about 0.5C, so a change of 0.5 or 1C isn't a huge deal. As per my previous post, it means another week per summer of 95 degree heat, but things could be worse.

On the other hand, a value of 7 or more on this plot suggests something really bad. For Lincoln, again, this would mean a temperature change of 3.5C; that means that the average summer would have about 45 days per summer above 95F. Corn is adapted to growing in a climate whose mean varies by 0.5C from year to year, which means that the number of days above 95F is ranging between about 16 and 22; it can't survive with 45 days of that any more than it can survive in Antarctica. Similarly all the other plants living there, and animals, and so on...

Of course, other creatures are just fine with warmer temperatures. Malaria and sleeping sickness, for instance. (As is currently being demonstrated in Africa: Nairobi was built above where the malaria line used to be. Surprise!)
  • Post a new comment


    Anonymous comments are disabled in this journal

    default userpic