The Doctors for Disaster Preparedness Conference in Houston (July 12 to 15 2013) featured several talks related to the dangerous manmade global warming alarm, including one from theclimatebet.com’s Professor Scott Armstrong. A video of his talk, “Evidence-Based Forecasting for Global Warming”, is available here. Willie Soon’s talk, “Five or more failed experiments in measuring global sea level change”, is available here, and Antony Watts’s talk, “Ten tests to determine whether you should be concerned about global warming” is available here.
The global average temperature anomaly was below the 2007 average again in July 2013 after a warmer June. Does the average temperature seem to go up-and-down a lot to you? Well, it turns out that the correlation between the change in the average monthly temperature and the change in the previous month is negative (-0.3) over the period relevant to the climate bet, 2007-to-date. In other words, an increase in the monthly temperature anomaly tends to be followed by a decrease the next month, and vice versa. As folks (and the no-change forecast) say, the more things change, the more they stay the same. The latest Climate Bet chart is posted to the right.
Scott Armstrong and Willie Soon both spoke on long-term climate forecasting—Armstrong on temperatures and Soon on sea levels—on 13 and 14 July 2014 at the Doctors for Disaster Preparedness Meeting in Houston, Texas. Scott Armstrong’s talk was titled “Evidence-based forecasting for global warming” and the slides are available here. Willie Soon’s talk was titled “Five or more failed experiments in measuring global sea level change” and a video recording of it is available, here.
Both scientists describe how alarming forecasts have been derived for these poorly understood situations using complex mathematical models with many variables and judgmental adjustments. These procedures violate The Golden Rule of Forecasting, which requires forecasters to be conservative, especially in the presence of great uncertainty.
We have belatedly updated the Armstrong-Gore bet graph, to the right. Those of you with keen eyes and good memories may notice some differences in the plot of the temperature series. The UAH global mean temperature anomaly series has been revised to Version 5.6. Information about the revision is available here.
With the release of the June figure, for the second time in 2013 Al Gore’s putative global mean temperature forecast was more accurate than the no-change forecast, .005°C more accurate.
The May 2013 data has been released and shows the monthly temperature anomaly was below the 2007 average that is the starting point of the Armstrong-Gore graph for the fourth month running. So far, the total error of Mr Gore’s warming forecast is 21% larger than the error of Professor Armstrong’s no-change forecast. See the updated Climate Bet graph at right for the details.
It occurred to us that the bet would have been fairer to Mr Gore and the IPCC if we had used the data that were available to Mr Gore when he released his movie “An Inconvenient Truth”, during 2006, as the base-year for The Bet. (The base year that we use for The Bet, 2007, was the most recent data available when Professor Armstrong issued his challenge to Mr Gore.) And so we re-ran The Bet using the 2005 average (the latest full year available to Mr Gore when he released his movie) as the base year. Mr Gore’s forecast in the re-run is for a 0.03ºC p.a. increase from the 2005 average and Professor Armstrong’s is simply the 2005 average.
In the event, re-running The Bet from 2008 to date using 2005 as a base results in a total error for the Gore/IPCC alarming warming forecast that is 31% larger than the error of the no-change forecast. We think Mr Gore would likely prefer to stick with the current Bet arrangement, even though it is not as fair.
After spiking in January, temperatures in April were again well below the 2007 average that is Scott Armstrong’s forecast. (See the updated chart to the right for the state of the bet.) Over the duration of the 64 months to date of the bet, temperatures have been greater than Mr Gore’s IPCC-based warming forecast for 15 months or less than 23% of the time. In contrast, temperatures have been less than Professor Armstrong’s evidence-based forecasts for 36 months or more than 56% of the time. None of the forecasts was exactly equal to the actual temperature. The results support the contention that Mr Gore and the IPCC’s dangerous warming forecasts are insufficiently conservative given the state of knowledge about climate, and that the Green, Armstrong, and Soon (2009) no-change model provides a better representation of the considerable uncertainty that exists.
The latest, March 2013, global mean temperature data from UAH is now plotted on the Climate Bet Graph at right. The temperature anomaly was the same as it was in February and was again cooler than Scott Armstrong’s no-change forecast. Overall, Al Gore’s IPCC “bet” of warming of 0.03°C per annum has been 18% less accurate than the no-change forecast. Over the 63-month life of the bet to date, on the basis of cumulative absolute error, Al Gore’s alarmist forecast has been the better bet for the 8 month period ending January 2011 only.
Bob Carter, Willie Soon, and William Briggs describe the evidence that changes in radiation from the Sun are the major source of changes in the Earth’s climate in a new article in Quadrant. The claim may seem uncontroversial, but global warming alarmists argue that human emissions of carbon dioxide have such a big effect that they dominate solar changes and are inexorably and dangerously boosting global mean temperatures. The Climate Bet is a test of these hypotheses, with Scott Armstrong “betting” on unpredictable changes in solar radiation and Al Gore on CO2-induced warming… and you know how that has been going.
As we’ve written before, trends appear to emerge in the data, then reverse, on all timescales. January 2013′s relative warmth turned out to be a one-month spike, with temperatures in February again below the 2007 global average temperature. Since the first month of Scott Armstrong’s “bet” with Al Gore, the UAH monthly temperature anomaly has been cooler than the 2007 average for 34 out of the 62 months. In other words, to date 55% of months have been cooler
A jump in global mean temperatures of 0.3°C from December 2012 to January 2013 has helped to keep the bet alive as it enters the second lustrum of its 10-year term. To date, Professor Armstrong’s naive forecast has been more accurate for 40 months out of 61 and has been nearly 13% more accurate overall.