A week or so ago, the Berkeley Earth project published a preliminary report on their findings as relates to land surface temperature data.
I think it is a wonderful example of how science works. The group of scientists, led by Richard Muller, that conducted this work were suspicious about claims made about the increase in average global surface temperatures since pre-industrial times.
They were not disputing that the numbers in the records were not going up, but that the methods used to derive those numbers were perhaps subject to some systematic errors. Their first cause of suspicion was that cities are warmer than the countryside. This is an issue with the temperature record because many of the weather stations that feed into most climate datasets were established in the countryside, just outside of towns and cities which over the decades and centuries have grown to surround and envelop the surrounding countryside and its weather stations. As an example, here are two maps of London showing how much it has grown (look at the Thames for an idea of scale).
So this growth of cities might have made it look like temperatures are going up when actually all that’s happening is that weather stations are coming under increasing influence of their host cities.
The other thing the Berkeley Earth team thought might be going on is that the records have systematic errors. Temperature measurement equipment has improved alot over the past 150 years. As weather stations upgrade their equipment, they might find that the old kit was systematically reading too low or too high. At the times where the equipment is changed over for different (hopefully better) devices, the old data has to be tweaked to make the records with the old kit line up with what those measurements would have been if the new kit had been there all along. This isn’t cheating, it’s not a dastardly trick on the part of climate scientists, it’s just something they do to give themselves a consistent dataset across time. The problem was that these tweaks to the existing climatological datasets were not applied automatically according to carefully crafted algorithms but by human researchers who could perhaps be making the data look how they want it to look, not how it really should look.
So the Berkeley Earth guys (with some of their funding from heavyweight fossil fuel bods) got enough data and enough computing power together to make a much bigger dataset that takes account of these growing cities and inconsistent records phenomena. They wanted to know if one or both of these effects were at play in the climate records to see if there wasn’t as much climate change as we keep hearing or even any at all.
What the data they produced actually showed was remarkably consistent with what the consensus view on climate change: Getting on for 1 degree C of average global surface temperature rise on pre-industrial levels, almost certainly attributable to human activity (i.e. deforestation and especially burning long dead things).
It is to their enormous credit that the scientists who conducted this work have published and publicised it. It is a fine example of how science is supposed to work. You see what others have done, you are maybe not convinced they’ve done things right and think you have a better method that could maybe lead to a different conclusion, you go away and try your method and you accept the results of your analysis.
Now if only we could get the same kind of scientific rigour out of the guys who gave them money…