Forecasting global climate change - Climate Change: The Facts - page 126

The most recent global warming scare started around 1976, so testing the validity of
short-term forecasts against the few years since then is possible. Such a test is limited,
however, given that it is not unusual for temperatures to trend up or down, on average,
for several years. Also, policy makers and investors who consider large expenditures that
are costly to reverse are concerned with long-term trends. We therefore devised tests of
the validity of the IPCC model's short- and long-term forecasts that made extensive use of
available data.
In1999,Michaelsexplainedthatshort-termeventswereresponsibleforrecentelevated
temperatures and offered an early test of the IPCC's short-term forecasts in the form of a
bet that temperatures would go down in the next ten years. 14 No one took the bet … and
temperatures went down.
Over the past nearly two decades, atmospheric CO 2 concentrations have risen while
global temperatures have remained flat. Despite the disconfirming evidence, the IPCC
claims to have become even more confident about the man-made global warming
hypothesis and they continue to forecast dangerous warming. The IPCC's response is
typical of how people tend to react when their forecasts are wrong: by having an even
stronger belief that they will be proven correct. 15 Moreover, scientists, like other human
beings, tend to reject evidence that contradicts their beliefs. 16
By 2007, there still had been no proper validation of the IPCC's forecasts. To generate
interest in the importance of validation, one of us (Armstrong) proposed a bet to former
U.S. Vice President Al Gore that a 'no-change' forecast of global average temperature
would be more accurate than any model or forecast that Mr. Gore would support. Gore,
advised by Professor James Hansen, 17 was at the time warning that a 'tipping point'
in global temperatures was imminent. In contrast to Gore's expectation of supporting
evidence soon, Armstrong expected that a much longer period would be needed to obtain
a clear result due to natural variations. Armstrong nevertheless proposed a ten-year bet on
the assumption that a shorter term would generate more interest, despite estimating that he
had a one third chance of losing.
In order to have an objective standard against which to compare forecasts from the
alternative hypotheses, the bet uses the University of Alabama at Huntsville (UAH) lower
troposphere series. 18 As of May, 2014, the errors from the IPCC's business-as-usual
forecast of +0.03ºC per year—standing in for Mr. Gore's tipping point due to his
unwillingness to take the bet—were more than 27 per cent larger than the errors from
Armstrong's bet on the no-change forecasts.
The models that the IPCC uses for forecasting are based on the beliefs of some
scientists that exponentially increasing levels of CO 2 in the atmosphere will cause global
mean temperature to increase at a rate of at least 0.03ºC per year. That figure has been