Northern Hemisphere Temperatures During the Past Millennium: Inferences, Uncertainties, and LimitationsThis is the seminal work (published in two journals) that put the hockey stick “on the map”. Mann’s team used the full instrumental surface temperature record back over a century in combination and historical data from various temperature proxies (most notably tree-rings, ice cores, and corals) and showed that the Northern Hemisphere warming observed during the latter half of the 20th century is without precedent at least as far back as 1400 AD.

The Evolution of Climate Over the Last MillenniumTwo more works by Mann and various other team members that expand upon and discuss the original hockey stick research.

Rutherford et al 2005 highlightsMost of the criticism of the original hockey stick work by Mann’s team centered on the proxies they used to infer historical temperature trends, and the way they were analyzed (see the comments below under Criticisms & Responses). These two studies present historical temperature analyses that were done using expanded datasets and completely different calibration methods than those used by Mann’s team. Both obtained a hockey stick curve similar to the original. As such, they provide independent confirmation of the Mann et al. hockey stick that is immune to pretty much all of the criticisms that have been leveled against it.

Note: This raises an important point. Skeptic criticism of the hockey stick has concentrated almost exclusively on the work of Mann et al. (1998 Nature; and 1999, Geophys. Res. Lett. ). As important as these papers are for the hockey stick (particularly public awareness of it) they are hardly its only theoretical or observational basis. Numerous other independent studies using different datasets and robust methods also support the fact that 20th century warming is unprecedented (for instance, see Jones & Briffa 1992, Holocene; Pollack et al. 1998, Science; Jones et al. 1998, Holocene; Briffa 2000, Quat. Sci. Rev.; Thompson et al. 2000, Science; Sowers and Bender 1995, Science; Blunier et al. 1997, Geophys. Res.Lett.; Fischer et al. 1999, Science; Petit et al. 1999, Nature; and more).

Myth vs. Fact Regarding the "Hockey Stick"These two articles from RealClimate give an excellent and readable overview of the hockey stick controversy including discussions of the basic hockey data and methods, and the flaws in the allegations made against it.

Comment on ‘‘Reconstructing Past Climate from Noisy Data’’These are excellent and well-cited summaries of the most common misconceptions about the hockey stick and paleoclimate.

Hockey sticks: Round 27The hockey stick curve resulted mainly (but by no means exclusively) from the work of Mann et al. (1999,Geophys. Res. Lett.). When first published it garnered a great deal of attention—not only because of its conclusions, but also because more than any other research to date it visually demonstrates just how unnatural the warming of the last 50 years has been, and the fact that it’s almost certainly related to human greenhouse gas emissions and land-use activity. To no one’s surprise, it enraged industry and Far-Right interests for whom greenhouse gas reductions and other mitigation efforts will prove costly. An all-out effort was launched by these interests to discredit it, and the scientists who had done the research behind it, most notably Michael Mann of the University of Virginia and his colleagues. Among other things, these efforts resulted in the publication of four papers by industry consultants on which hockey stick critics have been almost exclusively dependent. The first two, which were by Sallie Baliunas and Willie Soon (one with two co-authors) were published almost simultaneously in 2003 by Energy & Environment (an industry publication with few if any scientific peer-review standards) and Climate Research. These papers, which were almost identical in their content, made quite a splash when they first appeared. But shortly thereafter it was discovered that both were riddled with basic errors in their data and methods, even to the point that the latter one caused a scandal at Climate Research over their peer-review standards. Eventually, several of CR’s editors resigned in protest when the problems weren’t dealt with by the journal’s publisher.

The other two papers were less obviously flawed and have since become the primary anti-hockey stick weapons in the contrarian arsenal. The first of these appeared in Energy & Environment in 2003, and the second in Geophysical Research Letters in 2005. Both were authored by two industry consultants: Steven McIntyre, who works for the mining industry, and Ross McKittrick, a University of Guelph economist affiliated with several Far-Right think-tanks including the Marshall and Hudson Institutes (a third paper by McIntyre & McKittrick failed to meet peer-review standards and was rejected by the journal Nature in 2004). McIntyre & McKittrick (MM) argued that Mann et al. had not use enough temperature proxies or data in their hockey stick work and had been overly dependent on tree-ring data from North American bristlecone pine trees, which according to them was too sparse to be used. When preparing a historical temperature/time curve (or time series) it’s necessary to synchronize, or calibrate, the indirect proxy data for past centuries with modern instrumental records that are more accurate. To do this, Mann et al. (1998) made use of a statistical method known as Principle Component Analysis (or PCA) which is commonly used to sift out underlying curves from heterogeneous and noisy datasets. MM argued that by relying too heavily on bristlecone pine data Mann et al. had artificially selected for a hockey stick shaped curve in their PCA analysis (in other words, they had overfitted their data to one). To demonstrate this they created their own time/temperature calibrations (or reconstructions) using the data differently and obtained a “double-bladed” hockey stick curve that showed 15th century warming equal to or greater than 20th century warming.

These articles and papers demonstrate that MM had applied the PCA method incorrectly resulting in time/temperature reconstructions that were nothing more than math constructs. In other words, though mathematically valid, they had no genuine “connection” to the data they were derived from and were not climatologically meaningful. The first two links are to papers which formally demonstrate these points in different ways. The final three are to thorough, yet readable summaries of the MM reconstructions and their flaws.