Future prospects - Ice Ages and Interglacials: Measurements, Interpretation, and Models - page 340

variable spatial and temporal extent, Mann et al. (1998, 1999) and Mann and
Jones (2003) attempted a comprehensive analysis of the history of global average
temperatures over the past 1,000 to 2,000 years using a multi-proxy network
consisting of ''widely distributed high-quality annual-resolution proxy climate
indicators, individually collected and formerly analyzed by many paleo-climate
researchers.'' This was intended to integrate as many proxy sources as possible
into a single comprehensive view of how a single global average temperature (or
NH average temperature) varied over the past millennium or two. A number of
closely related studies were also published by other paleoclimatologists, all of the
alarmist persuasion. The final result was a reconstruction of a single NH or global
average temperature over the past one or two millennia that took on a so-called
''hockey stick'' structure: a rather flat profile for most of the past two millennia
with a significant rise in the 20th century.
The papers by Mann, Bradley, and Hughes (MBH) are compact, full of
jargon, and dicult to follow. However, this is a characteristic shared by many
papers that deal with large datasets for historic Earth temperatures. The reference
period for calibration of proxies with actual temperature data was from 1902 to
1980. The various proxies were more numerous in recent times and much less so
in the more distant past. Each of the proxy datasets had variable geographical and
temporal distribution, and the quality of the data probably varied enormously.
The task was to combine these into a uniform function that best expressed the
putative single global average temperature over a long time span. The process
used for data reduction is too complex to discuss in any detail here. However, it
should be kept in mind that the data were very sparse and contained mostly noise.
MBH used principal component analysis (PCA) to process the data. The end
result of this work was the discovery that global temperatures were basically
unchanged (with small fluctuations) for the past 2,000 years, and then there was a
sudden turn upward in the 20th century. This is often referred to as the hockey
stick result because it has the shape of a hockey stick lying on the ground with
the blade pointing upward at about 45 . According to this result, one would
conclude that there was no Medieval Warm Period (MWP) centered around
900 ad , and no Little Ice Age (LIA) in the period 1500 to 1850. This diagram
became the rallying point of alarmists who argued that warming in the
20th century was unprecedented in recent history and must be due to human
impact. Millions of copies of reports containing this result have been widely dis-
seminated. Subsequently, Stephen McIntyre and Ross McKitrick found a
fundamental flaw in the data-processing procedure used by Mann et al. (1998,
1999) and Mann and Jones (2003) which had the effect of skewing the apparent
result to produce the hockey stick configuration when, actually, the data do not
support this result. However, McIntyre and McKitrick were thwarted in their
attempts to get this information published, and ended up disseminating their
analysis mainly from their website. 9 In addition to McIntyre and McKitrick,
others reviewed the methods of MBH. The U. S. Congress asked a noted expert,
9 See the Climate Audit website at http://www.climateaudit.org