Long Term Temperature Records Contradict GISS Temperature Record

By Mark Fife – Re-Blogged From WUWT

I have been convinced for a long time there is something wrong with the theory of global warming. My initial response was based upon two factors. The first being in my youth I was a voracious reader. I was fascinated with history, archeology, and science. My interests varied wildly through the years. At times I was interested in the ancient peoples of South America. At other times I was interested in the Viking explorations. Obviously, the greatest wealth of actual historical material comes from Europe. Cutting to the point here, it seems obvious to me we are fortunate to be living in times where the climate is exceptionally good relative to what our ancestors endured in the past as well as what we have seen in the more recent past. I am old enough to remember the 60’s and I surely do remember the 70’s.

The other factor is the idea that CO2 going from 0.028% to 0.04% of the atmosphere would wreak doom and destruction upon the Earth just sounds ludicrous. What affect would that have on the emissivity or the heat capacity of a given volume of a gas mixture? I would think less than the measurement error and bias involved in trying to measure the difference.

Because of this, and because I am a real nerd when it comes to such things, I have been studying the issue as much as I can. What I found is the record of actual measurements is so poor, the majority are next to worthless. There are very few high-quality records which span the time frame necessary to put the current climate in its proper perspective. The rest are too short, too incomplete.

I have experimented with stringing different sets of data together, but that always creates uncertainty in the results. Unless two stations are reporting simultaneously for a good length of time you simply do not know how the two records relate. If you don’t have enough history from a single station you have no idea if it is warming from a relative cold period to a relatively normal. How do you even define what a normal range is?

I have long wondered how climatologists put all the fragments of data together to create such incontrovertible charts of impending doom to within 0.1° C going back to 1880. Especially when so few records go back that far. To be sure, I have confronted numerous climatologists and people claiming to be part of the group of people working on the data and models. I get nothing but generalities to my specific questions. Do you do area weighted averages? Have you applied spatial statistics? Did you see the study on starfish? That and silence. They just stop responding.

Though few in number, there are good quality, long term temperature records. What do they have to tell us about Global Warming or Climate Change and the role of CO2?

To begin putting this all together, I will look at the Central English Temperature records. According to HadCet, the data has been adjusted to account for urban heat island affects. I assume it has also been maintained to account for differing measurement devices. In any event, I am assuming it is as correct as they can make it.


The annual averages in the CET record show what has overall been a steady increase with shorter duration fluctuations since the lowest point of what is termed the Little Ice Age, which also corresponds to the Maunder Minimum. The Maunder Minimum is thought to have ended in 1715. The Little Ice Age is considered to have ended in 1850. This average warming has been 0.27° C per century.


When looking at the warmest month of each year the overall pattern remains the same as the annual average, except warming has only been 0.16° C per century.


Now looking at the coldest month of each year the over all pattern is again the same as the annual average, except warming has been 0.38° C per century.

It would seem to me milder, shorter winters would be a good thing. Especially compared to conditions around 1700.

I was fortunate to find two long term records from the Icelandic Met office. I also had the longest record from Greenland from a previous look at the GHCN network data. Let’s see how those records compare to the CET record. The graph below shows the absolute annual temperatures.


The following graph shows all four stations as temperature anomalies from their 1897 to 2007 stations average, which is the time frame where all four stations were reporting.


All four stations agree quite well in terms of the overall pattern. There is some variation in how much cooling or warming was experienced, which I would expect.


This graph shows the average of the four stations with the maximum and minimum annual average temperatures recorded amongst the four stations per year. It also shows a rolling five-year average. 95% of the annual averages fall within ± 1.0° C of the overall average.

As an aside, I will show the correlation of these temperature records to the record of CO2. The correlation coefficient of the overall average is .52 and that of Greenland is -.18.



I will now present the same type of data for the five longest records from the USHCN. The methodology of transcribing the data from absolute to relative anomalies is the same. Each station is shown relative to its 1874 to 2014 average.



As in the prior graph, these stations all follow the overall average within ± 1.0° C 90% of the time.

Now I will look at how well the average of the CET, Iceland, and Greenland records and the average of five long term records from the US match.


As shown in the graph above the two patterns are very similar, but there are significant differences. There are times when the amount of cooling and warming between the two is obviously different. Again, I would think that is the expected result. What was not expected, at least by me, is the timing of changes is out of phase. It appears the 30’s warming arrived and ended earlier in the US than in the other three locations. It also appears the 70’s cooling period ended earlier in the US. The following graph showing rolling five-year averages of the two averages demonstrates this apparent difference. It is a shame there are no records from the US prior to 1871.


As before, this is the correlation of the US long term station average to CO2. The correlation coefficient in this case is a definitive 0.14.


The question at this point is does it make sense to combine these long-term US station records with those of Iceland, Greenland, and the CET. The answer is yes and no. The combined average will create a reasonable approximation of the temperature record where the years being recorded are the same, but you will lose the data before 1871. The US record just doesn’t go back as far. When looking at records within a region the variation between stations stays within ± 1.0° C 90% of the time for over 120 years. However, when you combine two regions that boundary now becomes ± 3.5° C.

Based upon this limited look at just two regions it does make sense to combine records within a region where the records are similar as is the case here. Had one of these records been as dissimilar as the two overall regional averages it would not. The more dissimilar such records or averages of records are the less sense it makes to combine them into an average.

I am now going to take a brief look at the results from a previous study of records from Australia, which was covered in a previous article. Australia holds the only long-term records I have seen from the GHCN or any record set contained in the Berkeley Earth source data page in the Southern Hemisphere. I am only going to show those results from rural or small urban areas where the urban heat island affect is not evident.

At this point it should be obvious combining these records with those of the US and those from Iceland, Greenland, and the CET would not yield any useful information. The pattern of change is clearly and obviously different.


The correlation of this record from Australia to CO2 is as follows. The correlation coefficient is 0.14, which would indicate there is no correlation.


let’s see how these records compare to the GISS temperature record. All records are shown as temperature anomalies from their post 1960 average.




There are obviously substantial differences between the GISS temperature record and the long-term records I have presented. Lacking a detailed explanation of how GISS combined the many disparate and discontinuous records I can only speculate as to why those differences exist.

Now I am going to look at how well the GISS temperature record correlates to CO2. This is perhaps the most telling piece of evidence which shows just how different that record is from the long-term records both individually and as regional averages. The GISS record has a correlation coefficient of .92, which indicates a near perfect correlation. I would imagine many would find that near perfection to be suspect in and of itself as it would indicate there are no other major impactors of temperature. Which seems unlikely to say the least. This is in comparison to the individual records which range from .54 to -.18, which would seem a more reasonable outcome.



We have looked at quality, long term records from three different regions. Two of these are on opposite sides of the North Atlantic, one is in the South Pacific. The two regions bordered by the North Atlantic are similar, but not identical. The record from Australia is only similar in that temperature has varied over time and has warmed in the recent past.

In all three regions there is no evidence of any strong correlation to CO2. There is ample evidence to support a conjecture of little to no influence.

There is ample evidence, widely shown in other studies, of localized influence due to development and population growth. The CET record has a correlation of temperature to CO2 of 0.54, which is the highest correlation of any individual record in this study. This area is also the most highly developed. While this does not constitute proof, it does tend to support the supposition the weak CO2 signal is enhanced by a coincidence between rising CO2 and rising development and population.

The efficacy of combining US records with those records from Greenland, Iceland, and the UK may be subject to opinion. However, there is little doubt combining records from Australia would create an extremely misleading record. Like averaging a sine curve and a cosine curve.

It appears the GISS data set does a poor job of estimating the history of temperature in all three regions. It shows a near perfect correlation to CO2 levels which is simply not reflected in any of the individual or regional records. There are probably numerous reasons for this. I would conjecture the reasons would include the influence of short-term temperature record bias, development and population growth bias, and data estimation bias. However, a major source of error could be attributed to the simple mistake of averaging regions where the records simply are too dissimilar for an average to yield useful information.

The final question, which I hinted at early on in this article, is how well these records reflect what we know of the history of people in these various regions. The regional records which I have put together appear accurate, based upon history. The cold period corresponding to the Maunder Minimum has been well documented in both Europe and in North America. The warm period of the 1930’s extending into the 1940’s is also well documented, not only in Europe and North America but also in other parts of the world at the end of the 2nd World War. The 1970s cooling which affected America and Europe has also been well documented. In Australia there are accounts of severe heat waves in the 1800’s, such as the 1896 heat wave. According to records and personal accounts Australia experienced a severe drought at the end of the 1800s into the beginning of the 1900s and another drought at the end of WWII in the 1940s. By all accounts, working conditions in the late 1800s in Australia were particularly brutal because of hot conditions for factory workers.

Based upon a purely historical perspective the GISS temperature simply does not reflect the very real, well documented history of changes in climate in all three regions. The long term regional based averages I have presented do a much better job of describing what is known to history.



Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s