The Global Warming Challenge

Evidence-based forecasting for climate change

Archive for January, 2013

Armstrong wins two-thirds of months at halfway point of 10-year bet

without comments

Of the first 60 months of the 120 month (10 year) Climate Bet, Scott Armstrong’s naive model forecast* of no change in global average temperatures has been closer to the actual temperature than Al Gore’s IPCC-orignated 3°C per century warming forecast for 40 months. The updated Climate Bet Graph is to the right.

Mr Gore and much of the media are concerned about global warming. They should be relieved to learn that over the last five years (2008 to 2012) temperatures were flat or down from the previous month for 62% of months. The year 2012 ended with the global mean temperature for December the same as for the base year for the bet, 2007.

We calculate from the Hadley Center’s global average annual temperature estimates from 1850 to 2012 that the next five years would have to witness a rate of annual average temperature increase greater than 78% of previous five-year sequences in order for Mr Gore to win the bet. Perhaps, like the UK Met Office, he would like to reconsider his forecast.

*To learn more about the naive model, and the performance of no-change forecasts compared to the IPCC’s “forecasts”, see these papers:

Green, K. C., Armstrong, J. S., & Soon, W. (2009). Validity of climate change forecasting for public policy decision making. International Journal of Forecasting, 25, 826–832.

Green, K. C., Soon, W., & Armstrong, J. S. (2013). Evidence-based forecasting for climate change. [Working paper - not for citation].

Written by admin

January 10th, 2013 at 5:30 am