Recently, Oxford University launched a new initiative called Climate.Basics. This Internet site provides a nice explanation and simple illustrations on what is meant by ‘climate’ and how the climate system works. The Climate.Basics site is a collaboration with the *climateprediction.net* project, which is the world’s largest climate prediction experiment. We think this is quite good, but as always, let us know if you think something could be explained better

**Technical Note:** Sorry for the recent unanticipated down-time, we had to perform some necessary updates. Please let us know if you have any problems.

While you write “The Climate.Basics site is a collaboration with the project, which is the world’s largest climate prediction experiment”, I think you mean that the Climate.Basics site is a project that Oxford University is working on together with ClimatePrediction.net, the world’s largest climate prediction experiment, do you?

[

Response:Right! climateprediction.net is the big prediction experiment. Thanks for pointing this out! -rasmus]To help ClimatePrediction.net improving climate modelling you can use your computer’s idle time to calculate your own climate model at home, and after some weeks of calculating all data will be sent back for evaluation. A few weeks ago, they started an experimental sulphur cycle model which is being given to people with an up-to-date computer. Older hardware is getting enough work with the standard model version. The software this experiment is based on is BOINC (although some different and older software is still in use), developed by Berkeley University. It is the very same programme on which SETI@home is running, the most popular distributed computing project these days. A quick introduction in the project can be found here, otherwise you can get all details at the project’s web page here.

[

Response:Thanks this input! -rasmus]I don’t have the computer power or time for Climate Basics right now (maybe over the break), but I just want to say I believe what you tell us & really appreciate it. I only need the nutshell or blurb on issues, because I trust you.

GW is a really serious problem (if what the scientists are saying is true), and for contrarian arguments to hold up for me, they’d have to prove at 99% certainty that GW was not happening, or was not due to human contributions, or would have nil effects or only positive benefits (the high level of certainty for a true negative for me is due to the seriousness of the true positive). They’d need to do their own regular research (not cherry-picking or simply critiquing other research) at a rate of 2 or 3 times the regular research being done that supports GW. They’re so far from convincing me, & getting further away.

So I trust you instead. If I had a third life (my second one being spent fighting the food industry on MSG), I’d become a scientist so I could really understand all the details.

The Climate.Basics presentation does a good job of explaining why estimates involve probabilities. But it doesn’t explain why what may appear to be a relatively small change in average global temperature can have serious serious consequences.

Well I got 5/6 on their intro questions. They think I got q6 wrong (I’m assuming the Q’s are always the same): “by 2050, I could take my winter holiday in blackpool not australia”. I think this is fiction (however much GW we get, its still going to be cold there in winter) but they think it’s “nobody knows”.

They also say “the weather and climate are also chaotic” and illustrate it with a cute butterfly. Oh dear.

I’ll happily offer long odds against anyone who is prepared to bet that Blackpool will warm up enough by 2050 to be comparable to Australia today as a destination for winter holidays. Over-hyping the uncertainty to this extent is silly IMO.

[

Response:I don't know why they included this question: to many, a good winter holiday would be cold and snowy, so that people can enjoy skiing. To others, it would be a warm climate and a sandy beach. Perhaps they mean that nobody knows whether a breakdown in the THC will produce more bitter winters to the UK (where many houses has single glazing in their windows and cold stone walls), despite a general warming? Or more rain? Perhaps Australia would get too hotand dry for the northern Europeans? The question is, to my mind, ill-posed, so I guess 'I don't know' is not a bad answer... :-) -rasmus]Re #3 The Basics should explain how small change in average global temperature will have very serious consequences. I’m concerned about what the consequences of the next major El Nino will be like.

Actually, the introduction to bell-shaped curves is faulty:

throwing two dice many times, the distribution is shaped like a triangle,

not like a bell. To show convergence to a Gaussian distribution, the

experiment should use more and more dice (instead of throwing two dice,

alternative options should be to throw 3 dice at a time, then try 4, then 5…)

[

Response:Well spotted! The outcome of the dice throws should follow Pascal's triangle, if my memory serves me right. Which implies that the outcome of many dice expermeintts (throws) ought to follow a binomial frequency distribution. Clearly, the dice analogy is meant to illustrate the aspect of uncertainty, rather than the exact probability distribution function (pdf) of the climatic variables. But, thanks! This is a very good and valid point. -rasmus]Also, I have no idea where Blackpool is. Is that a city in Great Britain?

Specifying would help foreign readers.

Re: #8, Blackpool is along the coast in northwestern England.

Check an atlas to picture its location in relation to a larger city in the UK. There are many online, http://www.mapblast.com/ being one example.

The binomial distribution approaches the normal distribution as the number of trials approaches infinity. That is a consequence of the Central Limit Theorem. In fact, you don’t even need to start with the binomial distribution; any one of a large class of plausible initial distributions will on repetition approach the normal distribution. It is often joked that according to mathematicians, scientists say the normal distribution is common in nature, but according to scientists, it has been proved by mathematicians. The heart of this “paradox” is the Central Limit Theorem.

In practice, “infinity” is often reached relatively soon, in the sense that the normal distribution becomes a good approximation for a relatively small number of trials. I think the presentation illustrates this.

Point 7 is not being well understood. It’s simply a histogram of the total of the throw of two dice. It should converge on the true probability distribution, which is exactly triangular. While the mean of expermients constituting multiple trials is normally distributed, the distribution of individual trials is triangular, and that is what is plotted.

Also, while the mean is displayed, it is only displayed to two significant figures, pretty much demolishing the point of the slide control on the left. So that part of the presentation was almost as problematic as the butterfly one.

On the whole, in my opinion, while the production values of Climate Basics are higher than one typically sees in public communication of science, the quality and depth of the presentation is not.

There is plenty of material at the junior high school level out there, and plenty of material at the postgraduate level, so this site, which is clearly in the elementary class, in my opinion adds little of value. The gap has heretofore been filled only by dry undergraduate textbooks.

Realclimate constitutes substantial progress in communicating science to the public. In my opinion Climate Basics doesn’t.

mt

[

Response:Thank you for your opinion! I think that Climate.Basics target group was people with very low scientific skills, and that the product probably (?) was heavily influenced by media/communication specialists with little mathematical/scientifical skills. Hopefully the creators of Climate.Basics will take your points. -rasmus]There are a couple of points I would like to make in relation to this:

1. Although they did show results from two different scenarios in which CO2 levels differ, I fel that they should have made the point more explicitly that much of the uncertainty in what will happen to the climat in the future is due to uncertainty about what forcings will act on it. Mainly this will be due to global GHG emissions, but also due to vulcanism, etc.

2. Impacts matter. RealClimate once made a pastiche of climate sceptics by pointing to the inevitability of NH spring. However, it’s worth noting that the annual range in climate mean temperature for somewhere like the UK [which has a fairly moderate temperate climate] is from a low of 4 degrees in winter to 17 degrees in summer. So 13 degrees in total and it feels like a lot more because the difference between the hottest and warmest days of the years will be at least double that. In that context a warming of 0.5-1 K by 2050 doesn’t sound too impressive and not a big deal.

The problem at the moment is that current generation GCMs don’t very adequately simulate the current climate in terms of extreme weather impacts [droughts, floods, etc] and so they, obviously, aren’t much use at predicting changes in such things.

Consequently when people ask “So what?” when you tell them the globe is warming the response is normally fairly hand-wavy about sea-level rise of ~1m [what's the tidal range?] and undetermined changes in rainfall patterns…

Don’t get me wrong, I’d rather not see this grand experiment come to fruition to find out what *would* happen with ~5 degree global warming, but people don’t see how it is going to affect them. The weather is generally so variable anyway, that a relatively small shift in the mean is lost in the noise for most people.

Michael is completely correct. My previous comment was irrelevant because the distribution is not a binomial distribution, as was suggested in a previous comment.

As Michael says, you would have to take the mean of a “large number” of throws of two dice. That random process would be governed by a distribution close to a normal distribution. The larger the number, the closer to normal. As I noted before, in practice, “large” need not be terribly large before the approximation becomes a good guide.

One minor quibble. The distribution of values for throws of two dice is a discrete distribution since it can only take the integral values from 2 to 12. Its envelope is triangular. A true triangular curve would apply only to a continuous random variable.

That site is not so bad. I liked the part about the leaky bucket. It helped me understand the greenhouse effect better than I did before. I looked at the site with my 9th grade daughter and she was interested and engaged.

About the dice experiment: here is a suggestion. The horizontal axis could be labelled 0-9%, 10-19%, 20-29%,…, all the way to 90-99%. The left bar could be: “choose how many dice K you want to throw at any one time” Then each throw has a result between K and 6K and one could plot the number of times it falls within a certain percentage of the maximum 6K. The distribution would become closer and closer to the normal distribution as the number of dice grows.

Alternatively, another experiment which is I think much nicer for giving a visual understanding of the normal distribution and helping one understand how it can happen in the real world is based on Pascal’s triangle. See

http://mathworld.wolfram.com/GaltonBoard.html or

http://ccl.northwestern.edu/netlogo/models/GaltonBox

I found the bunny colonies a bit odd – I had no idea what either the chaotic or the non chaotic models actually were, just that bunnies and bunny graves randomly popped up. A cute idea, but I don’t think it was a good way to explain chaotic models. I prefer the bubbles in boiling water analogy – you know bubbles will form but can’t predict where or when. Also some info on possible effects of global warming would be a useful addition. Otherwise there is a feeling of “so what if today were 1 or even 5 degrees warmer?”

At the end of part 3 of What Can We Predict the interactive concludes that climate models can tell us “what will definitely not happen to the climate in the future.” I have zero expertise in the field but that seems like to bold a statement. Shouldn’t that read, “what is most likely not to happen”?

[

Response:Yes, I would argue that too. There is always some room for surprises. -rasmus]On behalf of everyone who was involved with making the resource, thank you for all the comments. We will certainly try to improve the online version at least, as much as we can.

A couple of responses

- yes, it would be great to go into a discussion of impacts etc., but I think that would require a CD-ROM in itself to do it justice

- regarding point 4 (and 11) – please could you clarify what the problem is with the weather/ butterfly

link?

- would anyone like to suggest an alternative question for the introduction – the point being to stress that we cant say exactly what will happen in any particular place?

Thanks for all the comments. Just to confrim the ‘Basics of Climate Predcition’ resource is targeted at a non-scientific general audience, including secondary school children. We appreciate the discussion it has generated and as Sylvia has said will be seeking to improve it in new editions.

I’m dismayed by the ‘wealthy countries to 2100 horizon’ approach. Surely what matters are (i) the effects on future generations over tens of thousands of years if runaway warming of several C is allowed to occur (ii) loss of biodiversity forever (iii) effects on developing world in our lifetime as well as beyond.

None of this appears to be covered by this package.