Forecasting global climate change - Climate Change: The Facts - page 127

the central forecast of the IPCC since 1990. 19 Because CO 2 levels have been increasing
exponentially since the beginning of the Industrial Revolution, the IPCC model would
seem to apply over this whole period.
We tested the validity of the IPCC model for forecasting horizons up to 100 years
using the data on global mean temperatures that the IPCC use: the U.K. Met Office Hadley
Centre's HadCRUT3 series. 20 The Hadley temperature series are derived from selected
weather stations and sea surface records that are adjusted and aggregated to provide proxy
average global temperatures. We derived rolling IPCC-model forecasts of the HadCRUT3
series starting from the year 1851, and ending in the year 1975, before the most recent
globalwarmingtrendcommenced.Theforecastingprocedurewassimple,andisconsistent
with the IPCC's published business-as-usual forecasts: we added 0.03°C to the previous
year's actual temperature to derive a one-year-ahead forecast, and then added the same
figure to the forecast for the previous year for each subsequent forecast horizon out
to 100 years. By repeating this procedure for each subsequent year, we obtained 125
one-year-ahead forecasts, 124 two-year-ahead forecasts, and so on, up to and including 26
forecasts for 100 years ahead. 21
Given that the HadCRUT3 temperature series trends broadly upwards, 22 one would
expect the IPCC-model forecasts that we generated to track the HadCRUT3 series quite
well. To determine whether the dangerous global warming hypothesis is a credible one,
however, it is necessary to test the forecasts against forecasts from alternative hypotheses,
and to do so using scientific forecasting methods.
In the 1960s and early 1970s, scientists warned of a new ice age. 23 The scientists
provided hypotheses to support their belief that this time the climate really had changed.
Some scientists still advance the cooling hypothesis. 24
Yet despite these forecasts of cooling, starting in the mid-to-late 1970s there was
actually a warming trend, and warming alarmists began to inform us that virtually all
scientists now subscribed to the dangerous man-made global warming hypothesis. The
claim of near unanimity of scientific opinion has been discredited by Legates et al., 25
however, and stands in contrast to the 31,487 U.S. scientists who have publicly signed
a statement that they consider the dangerous manmade global warming hypothesis
inconsistent with the evidence. 26
While scientists who predict warming and those who predict cooling provide reasons
for their hypotheses, their reasons have been indecisive. In any event, science does not
advancebyaskingscientiststovoteonhypotheses,butbytestingthemincompetitionwith
alternative reasonable hypotheses. 27
We tested a cooling hypothesis of 1°C cooling per century against the HadCRUT3
global temperature data. The forecast of cooling is consistent with the various alarms over