Cherry Picking

Unfortunately I'm not talking about getting hold of a nice batch of fresh fruit. Instead, it's a particularly common dishonest tactic. It's also one that is flagrantly against the principles of doing science.

What it consists of is making a statement that is true only about a specific especially well-chosen circumstance, and then pretending that you've made a general statement about the system at hand. This is offensive to me as a scientist because in science we're trying to understand the system -- all of it. The cherry pickers abandon honesty for word games.

Suppose we're trying to understand the global mean surface air temperature. There are many other things we could try to understand, but this one is fairly often looked at. After we look a bit, we notice several things. One is that the temperature varies from year to year. As we look in to this further, we see that several things happen which affect the temperatures. This includes having a more active sun (warmer), having a recent major volcano erupt (cooler), having an El Nino (warmer) or La Nina (cooler). After doing our best to subtract out all those effects, we see that there is still some variation year to year. That's the 'free variability' (the scientist's way of saying 'stuff happens'). It turns out that there are also some contributions from anthropogenic aerosols (cooling), increased greenhouse gas levels (warming), and other human activity (depends).

As we try to make our honest understanding of this complex system (are there other things that affect global mean temperature? how much?) we also have to wonder about how much data we need to collect before we can tell the difference between that free variability and a trend caused by one source or another. Remember what happened when you tried my climate change detection experiment. Even with random numbers, you got runs of several consecutive 'years' of warming or cooling. Free variability does this to you. So if you're looking for trends or other systematic things, you need to look at a long enough period that the free variability can't lead you to a mistaken conclusion. Plus, of course, you have to make that allowance for all the things that you know happen and affect the variable (global mean temperature) you're interested in but are due to processes (solar variability, El Nino, volcanoes, ...) that you're not concerned about at the moment (greenhouse gas levels).

It can be very difficult to do this even when you're trying to do it all correctly. One of the first satellite sounding temperature analyses (Spencer and Christy, 1992 or 1993, if I remember rightly) showed a large cooling trend at the same time that all other data sets showed a warming. This was very puzzling. Not long after, however, Christy (same one) and McNider (1994 or so) showed that this was because the data record started near an anomalously warm period (strong El Nino in 1982-3) and ended near an anomalously cold period (after the eruption of Mount Pinatubo). It's anomalous because we're not (in looking for signs of whether human activity affects global mean temperature) concerned with El Nino and volcanoes. Once those two obvious anomalous events were taken out, the 'cooling' trend vanished. Science being a small world, I ran in to McNider not long after he'd published that paper and we talked about it among other things.

One thing you can look for, even with no particular knowledge, is whether the author (blogger, commenter, ...) is considering other factors that can be involved. Even easier, and the cherry-pick which prompts me here, is to see how they selected the time spans they used and the data sets that are used. In the satellite example above, for instance, it was straightforward -- the authors used all the satellite period they had data for. Fair enough.

Since 1998, though, there's been an industry that is careful to not use all the data they could. Indeed they're aggressive about ignoring data. You don't need to be a specialist to know that this doesn't square with honest understanding of a complex system. People who are seriously trying to understand climate are continually complaining about wanting more data. Throwing away good data is inconceivable to them. But in that industry, they're not concerned with honest understanding. They wish to arrive at a conclusion and if they pick the right starting year (1998) and data set (CRU rather than GISS, for instance), then they can get the answer (a cooling 'trend') that they want.

Now to get that, they have to choose only one or two years, both from recent history, as the time to start their 'analysis'. If they choose any of the 100+ years before 1998 that we have a surface temperature record for, their conclusion is gone. If they use GISS rather than CRU, their conclusion is gone.

Further, even choosing that one year as the start would not be enough to preserve their conclusion if they were honest enough to examine the other things we know affect the climate system -- that was a year with a strong El Nino (warming) and high solar activity (warming). Instead they ignore this (either dishonest or simply not doing their homework) and make various declarations against anthropogenic climate change.

With a couple questions, then, a legion of authors/sites can be pitched for being unreliable:
* Are they playing the 'global cooling since 1998' game?
* More generally, would their conclusions hold up if the start year were chosen differently?
* Are they assuming that only one thing affects global mean temperatures?

If the do the first or third, they're lying or not doing their homework. If they don't address the second, they're at least not doing their homework.

I've been aware of this particular cherry-pick for some years now, and the popularity of cherry-picking among anti-scientific groups even longer. So I'll let you do your own check of how many sites or sources within 15 minutes you can find that commit this error. Depending on your reading speed, you should make 5 easily, and 20 if you're a quicker reader and have a fast connection.

Types of Sea Ice

Earlier I talked some about types of ice in the climate system and types of ice ages. With the public discussion of what will happen to the Arctic ice pack this summer, it is time for some talk about the different types of sea ice and what their significance is. The main two involved are first year ice and multiyear ice. First year ice is in its first year of life. Multiyear ice has been around for more than 1 year. (We're really not very elaborate in our naming!)

Multiyear ice is generally thicker than first year. Part of this is because it has had an extra winter (or several) to freeze more ice on and grow this way. Part of it is because as the ice floes get shoved around by the winds and currents, they crash in to each other and can pile up. It also generally has a lower salt content than first year ice does. During the summer, the salt in ice makes the melting point lower there (same reason we put salt on roads in winter, at least if it's warm enough) and the saltier (brine) parts melt out of the ice floe, leaving behind nearly totally fresh water. Being fresh water or close to it also makes for mechanically stronger ice. It also makes the ice radiate differently than first year in the microwave, so it is possible to distinguish some between multiyear ice and first year ice from satellite sensors.

First year ice, then, is the (generally) thinner, mechanically weaker, saltier ice that formed some time during the most recent winter. All three of these properties make it easier to get rid of in the summer whether by atmospheric warming (straightforward melting), ocean warming (ditto), solar heating (easier to melt the saltier ice with the sun's rays), or by having a strong weather system hit the ice with high winds (breaking it up mechanically and helping it melt faster by exposing more surface area to the air and sea).

A different feature of thicker versus thinner ice is that thinner ice is harder to make weather-type predictions for/with. See, for example, The thermodynamic predictability of sea ice, Grumbine, Robert W., Journal of Glaciology, vol.40, Issue 135, pp.277-282, 1994.

There are many local names for various stages of growth in the first year ice. They include:
Grease ice -- small ice particles which give the ocean a 'greasy' appearance
Pancake ice -- ice floes maybe a meter or two across, more or less round like a pancake and with raised edges (collecting Grease ice)
Young ice -- let the pancakes grow and get thicker.

(The links will take you to places with good pictures and further explanation.)

A different sort is the 'fast ice'. This is ice which has frozen fast to the land. Otherwise it us much like the young and first year sea ice.

There are plenty more names and labels for sea ice types, and I'm not even starting in on the bestiary of names for iceberg types.

Women's Olympic Marathon

Watching the Women's Olympic Marathon last night was a particularly interesting experience for me. I like watching the distance races because I understand them better than sprints or other events like swimming. This time was the first that an Olympic Marathon was held on a course that I knew a fair part of. Earlier this summer I was in China, including Beijing, Tian'anmen Square, the Temple of Heaven, and Forbidden City, all of which figured in the course. I don't normally get to say 'I was there' in watching Olympics.

Understanding the race, and an injury, also changed part of my viewing this time. Paula Radcliffe was running the marathon with a recent (8 weeks ago) stress fracture to her femur. I got a stress fracture in my second metatarsal (base of second toe) at the end of April. Three months later, I was cleared by my podiatrist to start back running at 80%. 2 months after stress fracturing her femur -- the biggest, heaviest, hardest to fracture bone in the leg -- she was out running a marathon at Olympic speed, not one of my leisurely 5k jogs. I was watching in astonishment that she was still in the race and in the lead pack for most of it, while every time she landed on the bad leg and took off for her next step had to have been starting at painful and gotten worse for every one of the 12,000 or so steps she did it. If you start a marathon in perfect health and well-trained, you're still going to be hurting by the end if you race it. She took on all that ordinary pain plus the fact that she didn't start the race in perfect health or training.

As the race neared the end, and had resolved to the race for second between Catherine Ndereba and Chunxiu Zhou, I had some sort-of inside information that let me pick the winner of the duel. That is, I've raced Ndereba before. At least for a very loose sense of the word 'raced'. I ran in the Beach to Beacon (Cape Elizabeth, ME) 10k in 2000. Ndereba raced there too, on her way to the Sydney Olympics. Aside from at the awards ceremony, of course, I never saw her. Her race was to win, my race was to run hard. We both succeeded, just that her success was to be close to 30 minutes and mine was around 45. Since then, though, I'd kept an ear out for her career. That included her later world record in the marathon, and the fact that among an extremely competitive group (elite marathoners are not casual folks!) she was considered extremely tough. So, come to the end of a marathon ... Chunxiu Zhou was a former 1500 m racer, which means plenty of speed. But the end of the marathon is much less about speed than sheer mental toughness. So my guess, which held up, was Ndereba.

Basic Sciences in Climatology

Last Friday I mentioned some math that you'd probably run in to when trying to study climate. You'll also encounter some of the basic sciences (as opposed to messy sciences like climate).

Extremely common as requirements are:
College Chemistry (year long sequence)
Physics (year long calculus-based sequence)
Thermodynamics (from one or another of Chemistry, Physics, or Engineering departments)
Plus, though I've never seen it required, it seems quite common for people to take it:
Astronomy. I think this is more a matter of personality than requirement.

Many people arrive in climate by way of physics, so they'll also have a year of modern physics, intermediate mechanics, intermediate electricity and magnetism, and some physics lab courses.

If they go to climate through geology, they'll have the above extremely common basic science courses, but then a batch of different messy science (geology) courses.

Climate Change Detection

You'll want some dice or a random number generator for our first efforts to think about climate change detection. Start with one six-sided die. We'll assume that it's a fair die -- that each face is just as likely as any other to turn up. If we toss it many times, the average of the numbers that shows up will be 3.5.

Weather and Climate
This is one of our distinctions between weather and climate. Weather is what we saw on any given throw and climate is that average. But 3.5 is not a number that you can get on any single throw. What's up with this? It turns out that this, too, is a reasonable thing for thinking about climate. If you look at some very small area, and only one parameter (say temperature), it's possible that you'll see the 'climatic norm' occur. But only for that small area and limited look. As soon as you look at a large scale, you find that although the weather (instantaneous state) can be generally close to climate, it is seldom close everywhere. See, for example, this sea surface temperature anomaly map(difference from climatology).

The 1 die model of climate doesn't work very well. On the map, we see that most of the area is near climatology, and that the farther away we are (hot or cold), the less area is present. A better model then is to use 5 dice (again 6 sided). In this case our 'climate' is a total of 17.5, which still doesn't happen as weather, although you can get close. The maximum is 30, and the minimum is 5. The maximum difference from climatology is 12.5. But most of the time we'll be close to climatology. Take 5 dice, throw them, add them up and record the result for a couple hundred throws. (If you're moderately quick with your addition, this takes only a few minutes; I've done it.)

Within your 200 totals, you'll find some runs of constantly increasing values, and about as many runs of constantly decreasing values. Some of the runs will be long, and some short. But the number of each is about the same whether it's increasing or decreasing. If you compute the average of the first 5 sums, then 10 sums, ... out to the whole 200, you'll see that the average wobbles around. It tends to be closer to 17.5 as time goes on (as you average more tosses) but only in a very jerky fashion.

Climate Change
Suppose I take one of your dice and tape a 6 over the side that should read 1. The average throw now totals 18.3333 (repeating), instead of 17.5. The minimum is now 6 instead of 5, but the maximum is unchanged. What will happen, though, is we'll see the high totals more often. Each of these is more or less a fair description of what we've seen in the climate of the last 120 years -- warmer minimum temperatures, higher averages, and not much change in maximum temperatures. (Not a perfect analog, but pretty good for only 5 dice.)

Repeat the exercise of tossing the dice and adding them up. How long do you have to do it before the average is obviously different from the first run? Remember that the first time the average jumped around for a while. You'll have to go on for longer than that this time.

How many throws do you need to make before you can tell that there are fewer low numbers? More high numbers?

In doing climate change detection professionally, these sorts of analyses are applied. Are there more extreme highs? Fewer extreme lows? Higher average? Just how many data points do we need to detect that in the statistics?

Final question: when did the climate change? When you have a long enough series of throws to detect it in the statistics reliably, or when you saw me tape over the 1 face with a 6?