Twentieth Century temperatures

When we examine the temperature record for the last hundred or so years, we see a number of discrete intervals: Temperature data, 1850-2015

  • 1880-1910: A slight cooling
  • 1910-1940  Temperatures rising at a rate of around 0.5C per 20 years.
  • 1940-1970: A slight cooling
  • 1970-1999: Temperatures rising at a rate of around 0.5C per 20 years.
  • 2000-on: A very slight cooling

This is based on the East Anglia Climate Research Unit data; other datasets are basically similar.

The important point to note here is the similarity between the two periods of warming. The rate of increase is about the same in both, and both last for a roughly similar time. In science, when we find a correlation like this, it does tend to suggest that the underlying cause may be similar. Not always, but in most cases that is so.

The IPCC claim that the 1970-1999 period of warming was a direct consequence of rising carbon dioxide levels in the atmosphere. Over this interval the rate of rise of carbon dioxide does, in fact, track the temperature rise quite faithfully, at first sight making this a plausible claim. The IPCC themselves say that the earlier period of warming cannot be explained in the same way, since carbon dioxide emissions in the early 20thC were too small to have caused a climate change equal to that of the latter part.

It's worth recalling that science depends heavily on correct experimental procedure for its conclusions as regards any cause-and-effect relationship. Here, the input is manmade CO2,and the output is a rise in global average temperature. Part of the correct procedure is the carrying-out of a control experiment. A control experiment is where the main experiment is repeated, all parameters being the same except that the main input, the presumed cause, is deliberately omitted. If the effect is seen to be still present with the cause absent, then it follows that the effect you are observing does not necessarily depend on the supposed cause. Here, the 1910-40 warming would serve as a good control experiment, enabling us to determine whether or not warming can take place with or without a causative carbon dioxide increase.

The IPCC have put forward various explanations for the earlier warming period, none of which can be verified. The question which this raises is why do we need two explanations anyway? Is one not enough? Is it not more likely that both periods of warming have the same cause? Their characteristics do look too similar to be a coincidence, after all.

Another point of interest is that the IPCC have stated, in relation to the recent pause in warming, that a period of less than 30 years cannot be taken as representative of a climate trend. Yet, since the 1910-40  warming was not a product of human CO2  emissions, it implies that  the entire climate change theory rests on the temperature increase between 1970 and 1999. Which, is not quite thirty years.

To summarise, two questions arise concerning this data:

It is claimed that CO2 caused the 1970-1999 warming, but it was seemingly possible to have a similar period of warming  from 1910-40, minus the large human contribution of recent decades. If CO2 did not cause the earlier warming, then what is there to indicate that the two periods differ in some fundamental way? Or, were they actually due to the same cause?

If the 1910-40 warming was not due to CO2, then the entire climate change campaign is based on only 29 years of correlation between temperature and CO2 data, 1970-99. The IPCC have themselves stated that this is barely enough to reliably establish a  trend.

The question of what actually caused the 1910-40 warming would thus seem to be a pivotal issue in deciding whether the anthropogenic climate change theory is valid.

Fossil fuel usage in recent pastTo this end, it would be instructive to be able to compare carbon dioxide emissions over the whole of the 20thC warming (and cooling) periods with temperature. Unfortunately, the high-precision carbon dioxide measurements at Mauna Loa only go back to mid-century. Earlier measurements were made as far back as the 1800's using chemical procedures, but their accuracy is questionable. Thus, we don't have a direct and accurate way to relate the 1910-40 period of warming to emissions.

A little research does however reveal that figures for amounts of fossil fuels used globally per year are available for quite a long way back. Since these should in principle be directly proportional to human carbon dioxide emissions, they ought to be a good substitute for actual CO2 figures. 

What we initially notice is that the modern, very rapid increase in fossil fuel usage begins around 1950. Prior to that time, the rate of increase was dramatically less, a fact which has been acknowledged by the IPCC as regards CO2 emissions.  However, the logarithmic relationship between CO2 and warming must not be forgotten. (It so often is, even by leading commentators!) In this case it will have the opposite effect of creating a 'law of diminishing returns' as it does with large increases, and will in fact make the smaller 1910-40 fossil fuel consumption more significant that if the relationship were linear.

The burning question here, is whether the logarithmic CO2/warming relationship would elevate the significance of the 1910-40 carbon dioxide footprint -from around a doubling of fossil fuel usage- to the point where it suffices to explain the warming in that era, in spite of its smaller magnitude. If it does, then we have a strong hypothesis for an anthropological cause of twentieth-century warming, based on two 30-year periods of correlating data. Hence, a pressing argument to divest ourselves of fossil fuels for the sake of avoiding further warming.

If however the relationship cannot explain the 1910-40 warming, then we are left with a situation in which the entire modern-era global warming hypothesis is based on only 29 years of data, and that a similar period of warming occurred, minus the elevated greenhouse effect. Effectively, a dataset too small to reliably establish a trend, and counter-indicated by what amounts to a failed control experiment. Which would leave us with an extremely weak hypothesis.