It has certainly been a while since I have blogged. What can I say? Holidays, tax season, vacations, work, putting in a home recording studio, spring fishing.....Hey, stuff gets in the way.
This post is going to be different. I am going to dust off a hat I very rarely wear in public without getting paid for it. I am going to go full tilt nerd. Yeah, very surprising I know. Whatever... We are going to talk Anthropogenic Global Warming (AGW). You will note I am sticking with the original terminology. It changed to climate change for whatever reasons, but the heart of the debate is still the original concept. CO2 is a primary driver of temperature and we are polluting the world with the stuff. For every addition 100 PPM (parts per million) of C02 in the atmosphere temperatures go up. Thus we are cooking the planet and we are all going to die. They tell you the science is settled. It is the settled or proven part I am going to address. I am not going to prove CO2 doesn't make the temperature go up. I am not going to prove it does either. I am going to demonstrate why I think the issue is far from settled science and nothing more.
Before we begin I am going to address some key issues. I have discussed and debated this subject many times and whenever you do that people who support the theory of AGW always, always, make these ad hominem statements. So let's get that out of the way up front. No I am not a climate scientist. I never claimed to be. No, I have never published a peer reviewed article. Neither did Isaac Newton and he was no hack. I am, however a degreed mathematician. I have worked in an engineering field for some 30 years. I have extensive experience in statistics and statistical analysis as well as metrology. Metrology, by the way, is the science of measuring. Now that isn't all I have done, but it is a big part of what I have done. Okay? By way of comparison, I am at least as qualified to study and discuss the subject as Bill Nye. I equal him in education and I believe I exceed him in terms of experience. If you think about it, we should all thank Bill Nye for setting the precedent of elevating someone of less than overpowering qualifications to the level of national expert and spokesperson. Golf clap, Bill, golf clap.
As to questions on my data. The temperature data comes from Berkeley Earth which is a non profit eco organization. I have pulled data from a number of sources in the past from the CET, to the English MET, NOAA, and so forth. The folks at Berkeley Earth have basically compiled these data sets. My data on global CO2 comes from the Institute for Atmospheric and Climate Science (IAC) at Eidgenössische Technische Hochschule in Zürich, Switzerland. The data represents what is the acknowledge global record of temperature and CO2 for 265 years beginning in the year 1753. The temperature data shows temperature variations from the 1950 to 1984 average in degrees Celsius.
So shall we begin?
The chart below shows temperatures in blue and CO2 levels in grey from 1753 to 2016. This is no different from any graph you have ever seen from proponents of AGW. Except most graphs I have seen do not go back past about 1978. It is true, looking at this graph you will say it all looks pretty clear to you. I can assure you it is not. So don't stop here! In the world of statistics looking right is usually only the beginning. Can you prove it is right is the real question.
The next graph is something you probably haven't seen. This is called a scatter diagram and is a graphical representation of the measure of correlation between CO2 levels and temperature. This chart in fact does indicate a moderate amount of correlation. Not strong and certainly not beyond any doubt or debate. Call it a 70% chance of correlation. That isn't precise, but it does give you an idea. It does indicate further work is warranted. The hypothesis has not been rejected, but it hasn't been proven either. We are at the point of saying the theory appears likely. I will break this down as we continue.
Prior to this point the data we have been looking at was yearly averages through time. From this point onward we are going to be looking at monthly global averages through time. There are a number of reasons to look at the data this way, but to me the biggest reason is this. Looking at an average, especially one generated over data showing a great deal of inherent variability, does not yield a clear picture of what is really happening. If I tell you last year was a lot hotter than the year before what do I mean? Was the summer really hot? Was the winter really mild? Was it a late spring? A long fall? What does it mean to say one year is hotter than another? Well, if all you look at is just the average temperature for the year it could very well mean any or all of those things and more besides. To understand why something is happening you really must understand what is happening. The charts below will, I believe, answer some of those questions.
The next chart shows the strength of correlation between CO2 levels and temperature variations over time both for annual yearly averages and annual monthly averages. What I am charting is known as a correlation coefficient. This is a statistical measurement of correlation strength ranging from 0 to 1 for positive correlation, 0 being no correlation at all and 1 being perfect correlation. A scatter diagram with a CC of 1 would be a straight line with no deviations. Generally speaking a CC of .5 is the point at which statisticians consider there to be some evidence of correlation between two factors.
You will notice the CC for the annual averages appears to be somewhat significant, but that is not true for the monthly averages. In fact, there is a pretty wide variation between months. How can that be right? Well, I will explain that. But first, lets look at some monthly data side by side so we can see what is really happening.
The graphs below depict the historical temperature records for July and December from 1753 to 2016. The first thing you will notice is neither month bears any strong resemblance to the chart of global averages we started off with. Secondarily, you will notice there is a very distinct difference between July and December. There doesn't appear to be any appreciable change to July temperatures. It started off just below 1° and ended up just below 1°. December's data shows a general steady rise in temperature with minor cyclical oscillations. .
Let's see what happens when we compare August and March temperatures through time. Again, August appears to have netted not much of a difference between 1753 and 2016. The results are pretty similar to what we saw looking at July and December.
I won't put a chart for every month up, they are all similar to the four proceeding charts. Below is a closer look at the data for January. Notice how the graph generally fluctuates above and below the linear average creating four essential symmetrical sections. Weird, huh?
So, here is the conclusion to my piece. There is no clear proof of correlation between CO2 levels and temperature. Therefore it is reasonable to reject the hypothesis of correlation between CO2 and temperature. That doesn't mean the issue is settled by the way. It is a rejection based upon a failure to prove the hypothesis because the case for it is just not strong enough to accept.
So what explains the seemingly strong correlation between CO2 and the annual averages?
A correlation coefficient of .7 is fairly strong, but is by no means definitive. It indicates you might be on the right track and the inference is reasonable to fairly likely. Further work would be required if you wish to declare the issue settled. But, that is what I did, I performed a deeper examination of the data.
What I have found indicates there is a flaw in the basic methodology of looking at global temperatures as an annual average. As such, the picture or model of reality thus created is inaccurate and flawed.
The data infers whatever effect CO2 may have on temperature as an annual global average is not uniform throughout the year. This would indicate there are additional factors involved capable of either enhancing or counteracting whatever affect CO2 might have. Therefore it is necessary to understand that effect on a seasonal basis.
This leads me to conclude there are other factors which may skew the data and create an inaccurate picture. Such as seasonal differences between the northern and southern hemispheres, location with respect to the equator, and so forth. The simple fact is more study is needed.
From the moment I first became aware global warming was a thing, from the moment I first saw a graph showing annual temperatures and CO2, I have been interested in the subject. And yes, as crazy as it sounds, I actually enjoy looking at data and doing statistics. In my career as an Engineer I did a great deal of data analysis. Back in the day we didn't have anything like the type of programs nearly everyone has on their computers today. Programs like Excel for instance. I wrote programs to do statistical analysis myself, because such programs were not common. I used these tools to analyze and resolve real world problems where the proof of the results of my work was, as they say, in the pudding. It made the difference between good product or bad product rolling down the assembly line. It made the difference between machines making good part or spitting out piles of scrap. It is what I do. In my world looking at just an average rarely ever yields a complete picture of what is really going on. Looking at the average of a target that moves in cycles as an average over several cycles is just bad statistics. Looking at the average of results produced over several different processes is just bad statistics. Yet, that is exactly what all these charts you see supporting the theory of AGW do.
I will no doubt continue looking at this as I have been doing for quite some time. Actually presenting the results is a new and novel idea to me. I will no doubt do so again. Maybe.
We shall see.