Yonatan Zunger’s Blog

It occurs to me that it’s not at all obvious why people worry about temperature changes of a few degrees. I thought it would be nice to have a few pictures showing why.

I went over to the University of Nebraska’s web site and downloaded the daily temperatures for Lincoln, NE for every day from 1920 to 1998. Here’s a plot of how the daily highs looked for just the summer months, June through August:

The blue area marks days over 95°F. I chose that temperature because (according to Purdue’s site) this is a day hot enough to significantly “stress” a corn crop, i.e. knock down its yield by several percent per day. (Less hot days can do damage too if there isn’t enough moisture, but 95 is trouble no matter what)

Now, the thing is: When people talk about the average temperature rising, what it means is that the center of this plot – the average daily temperature – moves however many degrees to the right. And every time you do that, more bars move into the “over 95” category. In fact, here’s a plot of how many days per summer would be over 95 as the temperature goes up:

Right now, there are about 18 days per summer over 95. If the average temperature went up five degrees Fahrenheit, you’d suddenly have 38 – out of 92 days in June, July and August.

There’s a bit more to it, too — when people talk about a “3°C global temperature rise,” what it really means is that some places (e.g. the Antarctic Ocean) get cooler, and some get a lot hotter. For example, Scenario A2, the worst-case scenario in the GISS-E paper that I’ve been talking about, talks about a 2.7°C global average temperature rise by 2100, but a 3.5 or 4°C (6 or 7°F) temperature rise for Nebraska.

It could be worse; the same model predicts 15 to 20 degrees F rise for India.

By the way, the second plot here is almost linear (for temperature changes up to ten degrees or so): the conversion is roughly that for every 1°C (local, not global) temperature change, you get an extra 7.4 days per summer above 95.

Anyway, this is the short science journalism post for the evening. Sorry if it isn’t horribly polished.

I just finished reading through the second part of the GISS-E climate modelling paper. I’ll write a summary later, either a technical one for or a non-technical one for here, but that’s going to take a while, and this is important.

Everyone, you need to read this document carefully, specifically sections 6-8, including the figures. Those sections require not much more than knowing what a standard deviation is and that a “climate forcing” means “any input to the ecosystem that can affect climate.” The first five sections, short version, say that the model has proven pretty good at predicting global-scale climate change for 1880-2003, it’s not as good at predicting regional change, and its main weaknesses are an ocean model that doesn’t understand El Niño and a sea ice model that nobody really trusts. My professional opinion is that it’s definitely good enough to rely on its numbers for global-scale analyses, but it may underestimate ice melting. (And the authors freely admit the latter)

Sections 6-8 talk about their models for the period 2003-2100, according to five models: “alternative”, “2C”, and three from the Intl Panel on Climate Change.

I know this is a bit of a technical thing to be asking people to read, but this is probably the most important thing that’s crossed my desk in years, and we need to start planning urgently. Pay particular attention to figures 19, 20 and 22.

Ladies and gentlemen, we have a problem. This outweighs anything on the political arena short of global thermonuclear war. I don’t think I could summarize what IPCC scenario A2 would look like by 2100 and have you believe me; but I would say that that scenario implies a real chance of a major population collapse, up to and including extinction and certainly impacting the viability of civilization.

The good news is, the two best scenarios (Alt and 2C) would leave us coming out fairly OK, and both of those are reachable with modern technology and not too much trouble; getting from here to there seems to require policy changes rather than anything enormously terrifying. The figures in the paper describe what emissions goals we would have to hit.

I’ll write up a more detailed summary later, and try to pull the key information out of this document.

BTW, I posted the first part of the first paper to climatepapers

Per the previous post, I just created the climatepapers community. Everyone who’s interested, sign on!

As a result of various conversations recently (and especially hearing Al Gore give his new talk a few weeks ago — preemptive plug, when his new movie An Inconvenient Truth comes out, go see it. If it’s half as interesting as his lecture it will be worth it), I’ve started to get very interested in learning about the state of the art in understanding the climate and how it’s changing.

The short answer is somewhere between “real trouble” and “REALLY BIG trouble.” But I’d like to understand things a bit more specifically than that.

To that end, I’ve started to assemble a list of papers that seem to represent the current state-of-the-art in the field, and this list is sure to grow as I read through more of them and follow reference chains. In fact, I’m planning on posting something soon with a generally readable summary of one of them.

But this got me thinking: Learning a subject is better done with many people. Would anyone be interested in forming an impromptu online journal club to learn about climate modelling, climate change, and all things related? (For those of you who haven’t participated in these before, what would be involved is everyone picking a paper, [or part of one for a really long paper] reading it thoroughly enough to write a good summary and explain everything that goes on in it, and then posting their summary and having a discussion about it. A typical rate is every week, someone else is responsible for a paper. It’s a great way to learn a new technical subject.)

The minimum background for doing this seems to be a reasonable science or engineering background; from what I’ve read of the papers so far, they don’t have a lot of obscure jargon beyond “stratosphere” and “sea ice,” just a lot of graphs, plots, and discussion of how they got them. For those without a heavy tech background, it should still be possible (and fun, and interesting) to be part of the discussion.

I’ve got a few papers in my list already, from the GISS-E group:
Possible papers