There are two types of temperature record, those directly recorded in real time and those which are recreated using a variety of techniques (referred to as proxy records) to try to understand changes in temperature long before recorded temperatures were available.
It is important to understand that there is no such thing as a global temperature. What is referred to when talking about climate change is an approximation of global temperatures derived from a limited number of data sets. There is no definitive long term record of global temperatures. Surface temperature records go back to about 1850 but early records are restricted to a small number of locations mainly in the northern hemisphere. Later records are affected by the Urban Heat Island (UHI) effect: temperature recording instruments, once sited in remote locations, are now surrounded by urban environments. UHI exaggerates the perceived warming. Where records are missing, they’ve been estimated. Anthony Watts, a retired meteorologist, has undertaken extensive research into surface temperature data. http://wattsupwiththat.com/reference-pages/global-temperature/
Satellite data are probably the most reliable guide we have
but only since 1979. They do however, give us a clear view month by month as to where we stand. According to satellite data peak temperatures were reached in 1998 which was a peak El Nino year (we'll be exploring oceanic oscillations later but for now El Nino/ El Nina are short term phenomena which have a marked effect on temperatures). Satellite temperature data since show no evidence of the predicted runaway warming.
If we look back at the recorded surface temperature records they show a secular rise since the end of the Little Ice Age (c.1850) but with periods of warming and cooling. From about 1910 to 1940 there was a warming trend similar to what we saw from around 1975 to 2000. In between there was a cooling phase from about 1940 and this was in spite of rising CO2 emissions in the post war industrial boom.
If we want to go back much further than 150 years to understand how today's temperatures compare with say 1000 years ago, we need to rely on anecdotal evidence and proxy temperature reconstructions.
The first IPCC Assessment Report in 1991 contained the HH Lamb graph (see below) of temperatures over the last 1,000 years which accords with our understanding of the Medieval Warm Period (MWP) and the subsequent Little Ice Age (LIA) and for which there is ample anecdotal and archeological evidence: the Vikings settled and farmed in Greenland from about 980 to 1400AD over the MWP; the settlements collapsed with the onset of the LIA; burial sites have been found in the permafrost; Pepys wrote of the Great Frost Fair of 1683 and skating on the Thames when the river and surrounding estuary froze for weeks over winter. That temperatures have been rising since end of the LIA (c.1820) is neither surprising nor alarming.
Clearly, Lamb’s graph wasn’t going to convince us of the urgent need to reduce our CO2 emissions and for the Third Assessment Report in 2001, the IPCC came up with the Hockey Stick (below) based on proxy data. This was produced in a paper (MBH98) by Michael Mann and others, claiming that the 1990s was the hottest decade of the millennium. A retired geologist and mathematician, Steve McIntyre, together with Ross McKitrick, an economics professor and statistician, set out to replicate Mann’s work but found statistical errors and evidence of manipulation of data to produce the hockey stick. Subsequently, the Wegman Committee, appointed by the US Senate to adjudicate on the hockey stick, found in favour of McIntyre and McKitrick, confirmed the flaws and repudiated the claim that the 1990s was the hottest decade of the last 1,000 years. A 2010 paper by McShane and Wyner, two statisticians, demonstrates why the original and subsequent versions of the hockey stick are flawed and not a reliable guide to global temperatures. No evidence suggests recent temperatures are unprecedented; on the contrary, there is circumstantial evidence that temperatures have been higher in the past and possibly as recently as the 1940s.
The leaked documents and emails from the University of East Anglia’s Climatic Research Unit (CRU) exposed collusion and manipulation of data to produce hockey stick graphs. Michael Mann and Phil Jones, the head of CRU, resisted requests from other scientists for data, methodologies and programs, in direct contravention of the scientific method. Former cabinet secretary, Lord Turnbull, described the three inquiries into Climategate as “ hasty and superficial.” He has called for “a full review of the science itself.”