Exploring some of the intricacies of GW [Global Warming] science can lead to surprising results that have major consequences. In a recent invited talk at the Heartland Institutes ICCC-12 [Twelfth International Conference on Climate Change], I investigated three important topics:
1. Inconsistencies in the surface temperature record.
2. Their explanation as artifacts arising from the misuse of data.
3. Thereby explaining the failure of IPCC to find credible evidence for anthropogenic global warming (AGW).
A misleading graph
In the iconic picture of the global surface temperature of the 20th century [fig. 1, top] one can discern two warming intervalsin the initial decades (1910-42) and in the final decades, 1977 to 2000.
Fig. 1: 20th century temps; topglobal; bottomUS.
Although these two trends look similar, they are really quite different: the initial warming is genuine, but the later warming is not. What a surprise! I wouldnt exactly call it "fake," but it just does not exist; I try to demonstrate this difference as an artifact of the data-gathering process, by comparing with several independent data sets covering similar time intervals.
The later warming is contradicted by every available dataset, as follows:
**The surface record for the "lower 48" [US] shows a much lower trend; [see fig. 1, bottom]; presumably there is better control over the placement of weather-stations and their thermometers;
**The trend of global sea surface temp [SST] is much less; with 1995 temp values nearly equal to those of 1942 [according to Gouretski and Kennedy, as published in Geophysical Research Letters in 2012];
**Likewise, the trend of night-time marine air-temperatures [NMAT], measured with thermometers on ship decks, according to data from J.J. Kennedy, Hadley Centre, UK;
**Atmospheric temperature trends are uniformly much lower and close to zero (during 1979-1997), whether measured with balloon-borne radiosondes or with microwave sounding units [MSU] aboard weather satellites [see fig. 8 in ref. 2];
**Compatible data on solar activity that show nothing unusual happening. Interestingly, the solar data had been assembled for a quite different purposenamely, to disprove the connection between cosmic rays and climate change [see here fig. 14 of ref. 2], assuming that the late-century warming was real. In the absence of such warming, as I argue here, this attempted critique of the cosmic-rayclimate connection collapses;
**Proxy data also show near-zero trends, whether from tree rings or ice cores, as noted about 20 years ago [see fig. 16 in ref. 1 and figs. 2 and 3 of ref. 2; plus those that may have been withheld by Michael Mann]. [If you look carefully at Manns original 1998 paper in Nature or subsequent copies, you will note that his proxy temps cease suddenly in 1979 and are replaced by temps from thermometers from CRU-EAU, the Climate Research Unit of East Anglia University. This substitution not only supplies the "blade" of Manns hockey stick but enables the claim of IPCC-AR3  that the 20th century was the warmest in the past 1,000 years, surpassing even the high temps of the Medieval Warm Period. In the Climategate e-mails, this substitution was referred to as Mikes Nature trick." I cant help wondering if Mann s original post-1979 proxy data showed warming at all; perhaps that has some bearing on why Mann has withheld these data; it could have killed the blade and spoiled the IPCC claim.]
On the other hand, the early warming [1910-40] is supported by many proxy dataincluding temps derived from tree rings, ice cores, etc; unfortunately, we could not find any temperature data of the upper troposphere. However, I bet they would have shown an amplified warming trenda hot spot.
A Digression on Hotspot [HSp] and Hockeystick [HSt]
"Hotspot" refers to an enhanced temp trend in the tropical upper troposphere [UT]; it is produced by convection of latent energy through water vapor [WV] and is the dominant agent for heating the UT. In IPCC-AR2 , BD Santer mistakenly identified the HSp as the fingerprint for GH [greenhouse] warming, which has led to much confusion in the technical literature, fostering the mistaken claim that the HSp owes its existence to tropospheric CO2. But according to textbooks, it is merely an amplification of any temp trend at the surface through the "moist" atmospheric lapse rate. It surely existed during 1910-42 but we lack data to prove it. Virtual absence of the HSp during 1979-97 [fig. 8 of ref. 2 ] implies a near-zero surface trend in that interval. This observation also disproves the AGW hypothesis of IPCC-AR2  that led to the Kyoto Protocol.
Manns construction of his hockey-stick graph [often referred to as "Mikes Nature trick" was explained earlier [see above].
This recital of data should suffice to convince alarmists and climate skeptics alike that the late 20th-century global warming does not exist. We should note, however, that both IPCC-AR4  and AR5  rely on such (non-existing) warming in trying to prove that its cause is anthropogenic.
Explaining the climate-trend artifact
Now we tackle, using newly available data, what may have caused the fictitious temperature trend in the latter decades of the 20th century.
We first look at Ocean Data as seen from fig. 2, there was a great shift in the way Sea Surface Temperatures [SSTs] were measured;
Source: J. J. Kennedy, et al. Journal of Geophysical Research, 2011.
Fig. 2: Sources of SST data: Note the drastic changes between 1980 and 2000 as global buoys increasingly replaced bucket sampling of SSTwith also important geographic changes.
Data from floating buoys increased from zero to 60% between 1980 and 2000. But such buoys are heated directly by the sun, as indicated in the cartoon of fig. 3, showing a floating buoy in the solar-heated top layer and unheated engine inlet water in lower ocean layers; this combination leads to a spurious rise in SST when the data are mixed together.
Fig. 3: Cartoon showing floating buoy in solar-heated layer and inlet for engine cooling water.
In merging them, we must note that buoy data are global, while bucket and inlet temps are perforce confined to [mostly commercial] shipping routes. Nor do we know the ocean depths that buckets sample; inlet depths depend on ship type and degree of loading. Disentangling this mess requires data details that are not available. About all we can demonstrate is a distinct diurnal variation in the buoy temps.
The Land Data have problems of their own. During the same decades, quite independently, there was a severe reduction in "superfluous" (mostly) rural stations [fig. 12 in ref. 2]unless they were located at airports. As seen from fig. 4, the number of stations decreased drastically in the 1990s.
Fig. 4: Weather stations at airports [Source: NOAA data].
[fig. 12 of ref. 2], but the number at airports declined less sharply, leading to a major rise in the fraction of reporting stations at airports [according to basic NOAA data].
This led to a huge increase, from 35% to 80%, in the fraction of airport weather stationsproducing a spurious temperature increase from all the construction of runways and buildingshard to calculate in detail. About all we can claim is a general increase in air traffic, about 5% per year worldwide [see fig. 19 in ref. 1].
We have, however, MSU data for the lower atmosphere over both ocean and land; they show little difference; so we can assume that both land data and ocean data contribute about equally to the fictitious surface trend reported for 1977 to 1997.
The absence of a warming trend removes all of IPCCs evidence for AGW. Both IPCC-AR4  and IPCC-AR5  rely on the 1979-1997 warming trend to demonstrate anthropogenic global warming [see chapters on "Attribution" in their respective final reports].
Obviously, if there is no warming trend, these demonstrations failand so do their proofs for AGW.
Ref. 1: Singer, S.F. Hot Talk, Cold Science: Global Warmings Unfinished Debate. Independent Institute, Oakland, CA, 1997 and 1999.
Ref. 2: Singer, S.F. Nature, Not Human Activity, Rules the Climate. Heartland Institute, Chicago, IL, 2008.