1. Let's start

 

Over the last 100 years the population of our planet has more than trebled. This has led to an increase in deforestation, a loss of natural habitats, the  growth of more and more mega-cities, and an enormous increase in pollution. If feels inevitable that we, as a species, are therefore having an adverse affect on the natural environment, and that that should include the weather. As a physicist I would argue that this must be true. But the question is, how much of an effect is it, and is it really that dangerous to us or the planet? Yet to even pose this basic question these days seems both radical and extreme.

One could argue that the irreconcilable division between climate change deniers and believers comes down to political outlook. Both camps appear to believe what they want to believe, and select the evidence accordingly. It is then perhaps no surprise that most climate change deniers are from the extreme Right who reject climate change because it threatens the laissez-faire capitalism they worship (just as they currently object to any action to prevent the spread of Covid-19 for much the same reason). The believers (such as Extinction Rebellion), on the other hand, are invariably a polyglot of displaced activists from other extreme Left activity groupings: CND, animal rights, anti-capitalists and agraro-vegetarians. Yet it is possible to be left-wing and a climate change sceptic as Michael Moore is currently showing, much to the horror of most of the left-wing commentariat.

Part of the reason for this is the way political debate in certain domains has been radicalized and polarized. Climate sceptics are a bit like British Eurosceptics. They fall into two main categories: deniers and queriers. In climate science the deniers deny everything, just as die-hard Eurosceptics reject everything that the EU stands for and see no benefits in it at all. Queriers on the other hand are the genuine sceptics. They accept some propositions as true, reject some others, and continue to question the validity of the rest. In the case of Euroscepticism they do it on the basis of economic and social evidence; in the case of climate change it is on the basis of the science. Unfortunately, when it comes to the science of climate change there is a lot to query. That is the main purpose of this blog - to query and test the science. That means putting the physics to the test, and the data analysis. It means challenging bad science from both sides of the argument. It also means debunking a lot of the fallacious claims made by climate change supporters, deniers and the media.

One reason that there is so much scepticism about global warming is due to the many false claims made by its supporters and activists. Climate change now is being used to validate everything, even when there is no obvious causal link, such as with the current Covid-19 outbreak. Then there is the high degree of confirmation bias in reporting and analysis, particularly in the area of extreme weather. Every extreme event, it is invariably stated, must be due to climate change, even when it's not. All evidence that conflicts with the accepted narrative is treated as random noise and ignored. Climate change is no longer science because it is no longer up for debate. The only debate is over how hot the planet will be in 50 years time.

And then there is the data analysis. This relies on statistics. But as Lord Rutherford allegedly once said, "If your experiment needs statistics, you ought to have done a better experiment." I have always taken that to mean that if you can't see a trend in the original data, then no amount of statistical processing should change that. It may improve your signal to noise ratio, but if your signal is much less than your noise, then you're stuffed, or at least you should be. The problem is that climate science (and climate scientists) does to appear to accept this rule of thumb. When you look at the range of techniques used (homogenization, breakpoints, principal component analysis, cross-correlations etc.), and then compare the output data with the input, you are often left scratching your head in disbelief.

So, as part of the process of discussing the data I will look at the temperature data itself. Many others have done likewise, not least those at Berkeley Earth. Their analysis over the last decade or so has forced some much-needed transparency on a subject area that was suffering significant reputational damage after the so-called "Climategate" emails. When Robert Muller used his criticism of the "Climategate" emails and the data analysis that they revealed to inspire the formation of Berkeley Earth as an independent (i.e. non-govermental) alternative to the main research institutions on climate change (NOAA, NASA-GISS and UK Met. Office-Hadley-CRU), who are also the guardians of most of the global temperature records, many climate sceptics expected a positive outcome (or should that be negative outcome?) Instead, the Berkeley Earth results were virtually identical to the existing IPCC analysis. Why? Well, because Berkeley Earth used many of the same analysis tools, just a lot more of them.

That in essence is the nub of the problem. Most of these analysis tools are buried deep in complex multi-function computer algorithms. While that may make the operation of the analysis algorithm free from user bias vis-a-vis the treatment of different records by different rules, it still does not prevent bias being incorporated within the algorithm itself. Who decides the criteria by which a breakpoint is introduced in a data series, or the timescale over which a reference of mean values is constructed?

In addition, computer programs are capable of being run multiple times with different input parameters. So the parameters can always be optimised to generate the desired output. This is true irrespective of whether you are a climate scientist or an epidemiologist. And if you have enough variables or orthogonal functions, then you can pretty much fit anything to anything.

What is missing is any analysis of the actual raw data without using breakpoints and homogenization. The argument against such an approach is that too many temperature records contain errors due to changes in location, time of measurement, change of instrument, not to mention human error. But most of these effects are small, and they should cancel in any averaging process due to regression to the mean. But the key thing is that no one approach should be relied upon. The real scientific method is to try different approaches and compare and present the different results. When you have so much uncertainty you need to keep your options open, and most of all, be objective. What climate science needs is actually less certainty and more ambiguity.