April 2008 Update on Global Temperature, Part 2
Posted by The Diatribe Guy on April 10, 2008
Before presenting the next info on the temperature trends, I need to note that I am switching gears a little bit. It is not a substantial change, but one I should explain.
Until now, I had used this GISS table for my temperature trend analysis. This table represents the global land temperature anomalies. As I delve more into the discussions on temperature trends and global warming and such, it has become apparent that most people, when discussing the global temperature anomalies, use this table instead. This table represents the global land plus surface sea temperature anomalies.
It is reasonable to ask whether or not this change had a significant impact in the analysis. It does not appear to change any major conclusions. There is still an overall positive trend in temperature for longer periods of time, and the most recent flat/cooling trend line now dates back to May 2001, as opposed to October 2001 with the land only data. As for my review of rolling trend lines and the trend of slope values, the general observations remain. I will be getting into those analyses on future posts.
Due to the change-over in tables, I need to work out a few things on my spreadsheet formulas. Over the last month I have put together the next model phase in not only looking at varying temperature and slope trends, but I have also incorporated an analysis to better forecast future anomalies based on the trends. I should be able to explain the whole methodology shortly,with some initial results.
Today’s post is fairly straightforward, and the charts simply portray what many people have already known. But since I am switching data sets, it is worth presenting the charts as a new starting point.
The first chart demonstrates the overall temperature trend since the inception of the GISS data, warts and all (I will get to that some day). I am generously, for the time being, assuming that the warts smooth themselves out for the most part, and that the trend actually represents a real trend.
I realize that a lot of people like to focus on the last 30 or 40 years as the “real” temperature trend. I see no call for that. It is obvious from the data over time that there are both shorter and longer cycles of warming and cooling. There have been interesting studies in the harmonics of the data that demonstrate this, as well. So, I actually think the most relevant information when discussing long-term warming trends is the trend of temperatures going back as far as possible. While I strongly suspect that the GISS data is less trustworthy the further back one goes, it is the estimate we have, and we will consider it a reasonable estimate.
The following chart shows the overall trend.
A couple comments on the above chart. The slope = 0.0468, which is saying that the trend in the anomaly is 0.0468 per month over time. An anomaly of 1 = .01 C. So, this says that the overal trend in temperature is a warming trend of 0.000468 degrees Celsius per month, or 0.00562 degrees Cesisus per year. From there, the math is pretty easy. The warming trend per century has been 0.562 degrees Celsius.
Is this the right trending measure? Well, many don’t think so. We can loook at the trend of the last 30 years and suggest that temperatures are increasing at a musch swifter pace. But take a second look at the chart. If you follow the trend line, as will be the case with any long-term trend, the warming that took place up until the mid-90s was necessary just to get back to the overall trend line (as it stands now, anyway). In recent years, the anomalies are above the trend line, but we have seen over the last few years a flattening, which seems to be drawing the anomalies back towards the line. It therefore seems reasonable to think that when discussing long-term trends, the 120+ year trend tells us more than the 30 year trend does. Many, in fact, will argue that the 120 year line is still insufficient due to the longer-term solar cycles that are sure to affect temperature trends on a basis of 400 years, and other cycles that will last even longer. But for now, it is perhaps reasonable to suggest that there is a long-term warming trend of about half a degree. Whether or not it is anthropogenic, I don’t know for sure. But the alarmism about 3-4 degree per century warming (or, if Ted Turner is to believed, 9 degrees by 2040), just seems a little exaggerated.
As I mentioned, there is a recent flattening of the raw anomaly trends, and depending on your starting point, it can even be perceived as a cooling trend. In some of the anomaly measures, this trend dates back to 1998. The GISS data doesn’t show that same result, however. Below is the trend line since May 2001, which is the furthest date back for which there is a negative sign in front of the slope calculation.
Now, let me be clear about this “trend” by noting the R-Squared value. You don’t have to be an expert in statistics to see that the anomalies vary quite a bit and the “fit,” while the best fit for the data, isn’t really tight. An R-Squared near the value of 1.0 is a very good fit. A value of 0.0001… not so much. Basically what this means is that you could reasonably fit a line with a positive or negative slope that would be almost as good a fit as the line above. But the point is still a valid one – there has been no warming trend in the last seven years. As I noted above, with the anomalies above the long-term trend line, it is not necessarily surprising that we would see a downturn.
Now, some are arguing that we are about ready to enter a protracted cooling period. That is a topic for another time, and it is based on solar activity, sunspot cycles, the recent inactivity in the transition from cycle-23 to cycle-24, papers by Landscheidt, etc. All I am doing right now is presenting the data as I see it.