Describing the water

There’s a line in “As Good as it Gets” where the novelist Melvin Udall complains to the artist Simon Bishop, “I’m drowning here and you’re describing the water“. Melvin wants a solution to his romantic difficulties¹, but Simon has been, well, describing the water, laying out the situation in greater and greater and more overwhelming detail. The film came out in late 1997², which is now officially a long time ago, but that line stuck in my head³.

It came back to me today⁴. I posted a thread about global temperatures in August, which now has, appended below it, a long back and fore about whether the data are reliable or not. These days, aside from a silly remark⁵, I tend to keep out of such things. From long experience, I know that it’s doubtful whether anything I say would register. It would, most likely be an excuse for further, dreary entanglements and I don’t have the energy any more.

They’re dreary for lots of reasons, but one key one is that they describe the water. The data we now use to understand how climate has changed over the past two centuries were taken for all sorts of reasons, but climate monitoring usually wasn’t one of them. When a parson bought one of those new fangled thermometers and started jotting the daily temperatures readings from the north-facing wall of the parsonage in his diary, he scarcely could have dreamed that two centuries later those measurements might still be of use in tracking not just the temperature of Little-Dribbling-on-the-Water but the whole world. When Lt Maury started gathering sea-surface temperature measurements, he wanted to map out ocean currents to speed trade and transit across the oceans. He’d never heard of a reanalysis; no one would for over a century, but the measurement he and his expanding fleet of volunteers made now form the basis for century-long reforecasts of the earth’s climate that are widely used in climate science and beyond.

Building and sustaining the networks used to monitor temperature, rainfall and other meteorological parameters is neither cheap nor easy. Doing it at a global scale is an extraordinary undertaking. Inevitably, the resulting measurements are less than perfect. Stations are sited for convenience and cost as well as for more academic concerns such as correct exposure to the elements. Even when stations are well sited, urban growth and land use can overtake them. Sometimes stations need to be relocated completely or closed. Even now, large areas of the world are not routinely measured at the surface. All these changes, problems, difficulties are only partially documented. While modern climatologists would have liked to inherit a global measurement network that was pristine and unchanged from some idyllic rural past through to today, change is, sadly, (or happily depending on your disposition), the norm.

At sea, water and air temperatures were measured in a variety of ways. In the nineteenth century and early twentieth century many ships would use buckets to gather a water sample for measurement. With steamers and other independently powered ships, water drawn in to the ship’s engine room proved a practical and safe way for monitoring water temperature, safer anyway than hanging over the bulwark⁶ in a howling gale trailing a bucketful of water and thirty metres of wet rope. Latterly, dedicated electric sensors were attached to the hulls of ships as well as to free-floating or anchored buoys⁷. Until recently, and as with land, change was the norm. Even today, there is no means for making air temperature measurements at sea that come close to meeting climate monitoring requirements. Sea surface temperatures, on the other hand, are measured by ships, buoys, argo floats and an astonishing array of satellites.

The point of listing all these things is to describe the water. Climate scientists – at least those who pay the slightest attention to what goes into the observational products they use – are intimately acquainted with it. The water, the observations bequeathed to us by contingencies of history, the difficulties of understanding a chaotic complex system etc, are what we have to work with.

One could give in at this point and sink beneath the cold surface. It would be understandable. But over many decades of patient and dedicated work, climate scientists have come up with increasingly smart ways of not just avoiding drowning, but actively thriving in the water. It’s not always been easy, and the solutions aren’t always perfect, but that’s what science does, it builds something extraordinary on the scantest of foundations.

Stations move, so we have homogenisation. Stations move without anyone officially noting it down, so we have statistical break detection. Measurement methods change, so we have bucket corrections. We wonder if the bucket models work, so we test them in the lab and wind tunnels. Thermometer enclosures change so we have parallel measurement experiments. Each thermometer, each ship, each fleet has it’s own way of measuring SST, so we modify our understanding of the uncertainties accordingly. There are large gaps between stations, so we use statistical infilling ten different ways to estimate what happens in those gaps. We suspect that urbanisation and expanding heat islands have a disproportionate effect on land stations, so we break down the network into rural and urban components and show there is no difference. And, because scientists can never leave anything alone, we have tests to show that many of these things are effective at what they do. Often, these things have been revisited time after time after time. There is a large and still growing literature that covers all of these aspects in mind numbing detail. That’s not to say that everything is understood to the same degree, but we do know a lot about what we don’t know. There are, if you like, warning signs pointing out the rip tides and other dangers that might not be immediately apparent to the casual bather.

If you’re describing the water, you’re not telling a climate scientist anything they don’t already know. Furthermore, you are probably overlooking the amenities provided by climate science and the many ways that the water has been made safe. If a climate scientist won’t engage with you, it might just be for the simple reason that an adult demanding to be saved because they find themselves out of their depth in a paddling pool is embarrassing for everyone involved (or ought to be). It might also just be because there are ample resources out there already explaining all these things. For the dedicated, there are decades of scientific papers, but even for the casual enquirer, there are pre-digested and potted summaries aplenty.

It would be somewhat remiss, of course, to watch someone who was actually (metaphorically) drowning and not do anything, but such people really are few and far between. Far more common are the ones who know a little, but believe they know it all. They know about homogenisation, but give no indication that they understand why it’s done or how effective it is. They point out the gaps in the data and that these are filled using statistics, as if the methods used to do that weren’t widely known to be effective. They will in one breath complain about adjustments, but in another complain about station siting. They know enough to say that a station next to a parking lot might be in error, but don’t grasp that a station that’s currently in a field, might have been next to a parking lot twenty years ago. What about systematic errors? they will ask, as if there weren’t dozens of papers on the topic. They can simultaneously believe that data are of insufficient quality to determine a trend due to greenhouse gases, but nonetheless supportive of their own pet theory usually involving sun spots or some astrological scheme invoking the movement of the planets. Often they already know the answers to the questions they’re peddling, but that doesn’t stop them. These paddling pool exhibitionists are hard to take seriously.

If a person has studied the issues diligently enough to be able to quote reams of sceptical talking points from memory, but hasn’t taken the care to understand the “other side”, by which I mean the vast majority of literature published by experts on the subject, then I don’t feel it is incumbent upon me to correct them. If they were sincerely interested, they would, in their hundreds of hours of personal study, have corrected themselves. Usually, they haven’t.

Occasionally, though, one finds the rare sceptic how has taken the trouble to correct themselves, someone who has moved on from simply describing the water. Such instances are rare, but interesting and how a person reacts to them is often telling.

The most famous example might be the Berkeley Earth team, who went from publicly avowed scepticism to Koch-funded belief in a surprisingly short period of time. While the main dataset is of lasting usefulness⁸, the other analyses (concerning station siting, urbanisation, attribution) were of less obvious value. Other people had already done these things better and more thoroughly, but occasionally a person has to learn a thing for themselves. Initially the project was cheered on by various sceptic concerns, but the cheers turned to boos when the conclusions came out.

There are other, limited examples. Anthony Watt’s surface station project to document the siting of all stations in the US is one. Such a database is undeniably a useful thing to have (as long as one acknowledges its limitations⁹ and can overlook poorly hand-labelled photographs). The paper based on the data showed that, actually, there was no bias in mean temperatures because the biases in the max and min temperatures cancelled out. It wasn’t the result Watts was looking for so he promised another paper that would show that station siting did have a fatal effect on trends. Despite being heavily trailed for a few years, we’re still waiting.

Another example is a correction to a reconstruction of Antarctic temperatures measured by satellite. A group of sceptics spotted what they believed to be an error in the processing. They fixed the error and reran the reconstruction. The resulting paper is valuable in so much as there is now an improved record of Antarctic temperature change and a better method for calculating it, both of which are clearly good things. If, on the other hand, your takeaway is only that climate scientists made a booboo then you’re just splashing in shallow end of the pool.

So, if a climate scientist won’t engage with you, ask yourself this:

1. Am I just describing the water?
2. Have I made full use of the available water safety features?
3. Am I just splashing around for attention?

-fin-

1. His romantic difficulties are largely caused by his extraordinary multi-levelled unpleasantness. It’s a strange film. ↩︎
2. More time separates 1997 from us than separated 1997 from the last time we were on the moon (not you and I, but homo sapiens in a more general sense along with face mites, bacteria and the other passengers we carry around with us). ↩︎
3. Unlike (it seems) almost everyone I know, I can’t quote films from memory, which makes my retention of this line so unusual. When treated to a quote by someone else, I usually have great difficulty locating it in the cinematic multiverse and generally draw a blank, so I just laugh and nod vigorously*. ↩︎
4. Usual caveats about timeliness or lack thereof apply. These posts take me anywhere from ten minutes to ten years to put together, with longer periods denoting not special care but rather a tendency to ↩︎
5. Absence of a sense of humour is often a pretty strong tell. As an Englishman, I’m inclined to believe that climate change is too important to be taken seriously***, but I’m also aware that such an attitude doesn’t translate to other cultural contexts. ↩︎
6. My deployment of nautical terminology is apt to reduce to tears anyone who has actually been to sea. If it’s not a bulwark, it might be a gunwale or possibly a scuttlebutt, baggywrinkle or gollywobbler (amazingly only one of these is made up). ↩︎
7. While it’s common to say, there aren’t any heat islands at sea, I did find a moored buoy reporting over the GTS** that had been deployed in a theme park water feature. I removed it from the processing stream. Recently, record high sea-surface temperatures were claimed for a buoy reporting from shallow waters off the coast of Florida. It turned out not to be the case – there were higher temperatures from the same buoy in previous years – but does raise the question of what ought to be included in a sea surface temperature data set. There are numerous moored buoys in rivers, lakes, estuaries, lagoons, harbours, bayous and other transitional spaces, whose temperatures are not representative of the open ocean. Variability at these sites can be very different, with effects due to tidal processes, river outflow, shallow waters and limited circulation. While an SST data set with a 1km spatial resolution might usefully employ such measurements, a data set like HadSST, which has something more like a 500km resolution, is probably better off excluding them and does. ↩︎
8. It could have been so much more useful, but the team paid too much attention to those who were performatively splashing around in the shallows. If they’d bothered to ask a sensible climate scientist, they might have partitioned their effort differently and would most likely have focused more on the oceans where the uncertainties (of all kinds) were and are significantly larger. As it is, when they came to make a global dataset, they had to use an existing ocean analysis: HadSST4. As the author of HadSST4 I am, of course, happy and flattered, but at the same time, the HadSST4 paper very clearly points out the need to “add more data and metadata to ICOADS, reprocess existing ICOADS records, improve information on observational methods; improve physical and statistical models of SST bias, maintain and extend the range of different estimates of SST bias, and expand data sources for validation and extend the use of measures of internal consistency in validation“. The earlier HadSST3 paper said something very similar: “Until multiple, independent estimates of SST biases exist, a significant contribution to the total uncertainty will remain unexplored. This remains a key weakness of historical SST analysis.” I’ve been saying it for a long time. No one listens. ↩︎
9. Principally that it is just a snapshot of a point in time. To ensure homogeneity, one needs to know the whole station history and keep it up to date. Just because a station is now well sited doesn’t mean it always was or always will be. ↩︎

* Until I started German lessons, this was pretty much my approach to speaking German, albeit with a wider repertoire of emotional responses. Some people just like to talk and I’m happy to be there for them. I’ll nod and smile and frown and shake my head depending on the verbal and emotional cues without necessarily understanding one word in ten. I’ve spent hours in the company of strangers doing exactly this. It doesn’t always work though. After ten minutes of deep “conversation” at a bus stop with an elderly gentleman who had mostly directed his comments to the passing traffic, and which on my end had mostly comprised vague, encouraging noises and the occasional nod or shake of the head, he turned to me in the way that people do when making a very important point. I switched to my most earnest face and maintained strong, but sympathetic eye contact ready for whatever he was about to lay on me. In precise but heavily accented English he said “You have no idea what I am saying do you?” We were separated by the width of a bus-shelter bench whose dimensions had been chosen to discourage sleepers. Sitting shoulder to shoulder it had felt comfortably expansive, but now, facing each other, it felt tiny and the moment had taken on an emotional intensity for which neither one of us was fully prepared and which haunts me years later.

** GTS = Global Telecommunication System, a global telecommunication system used to share observations for meteorological and climatological purposes.

*** Too important to be taken seriously is, I was disappointed to find out, a Wildeism.

Header image is Mizen Head by Jjm596 provided under a CC BY-SA 4.0 license