DISCLAIMER: There are still many unanswered questions about this model. I provide this essay for the purposes of discussion, but I give no pro or con endorsement – Anthony
Guest essay by David Archibald:
Back in July, David Evans released his Notch-Delay Model which uses Total Solar Irradience (TSI) to predict climate up to 10 years in advance. Soon had previously derived a possible mechanism for the 10 year delay that he found between TSI and tropical North Atlantic sea surface temperatures. This is the lower panel of Figure 4 from his paper:
To test the hindcast match of the Notch-Delay Model, the model was stopped at December 1991 for the TSI data up to that point and at two year intervals thereafter up to December 2012 for a total of 12 prediction runs. The predictions produced were then plotted on the UAH lower troposphere anomaly record up to August 2014:
There were two big departures in the 1990s due to the Mt Pinatubo eruption of 1991 and the 1998 El Nino. Just after that el nino, the model predicted the period from 2000 to 2004 very well with a tight grouping of forecasts corresponding to the shape of the temperature profile. From 2004 to the end of the decade, the model forecasts then dispersed with average temperatures generally above what the model forecast. The run of El Niños during those years would have played a part in the divergence. The prediction from 2004 gave an early, accurate forecast of the temperature peak in 2013 as it would have incorporated the second peak of Solar Cycle 23 in 2003.
The prediction from 2006 was the first indication of a sharp temperature fall this decade with a 0.4° fall over the last three years of the forecast. This predicted temperature decline would have been due to the sharp fall in the Ap Index in 2005. In the following prediction using data to the end of December 2008, the forecast decline increased and steepened up to a 0.8° decline over four years. The subsequent model runs of 2010 and 2012 are similar with a predicted flattening out between 2018 and early next decade.
Barring major volcanic eruptions and El Niños, the Notch-Delay Model’s resolution looks like it is of the order of 0.3° or so. That resolution matches the inter-annual variation in temperature over the last couple of decades. The next couple of years will show if it can predict major swings in climate from TSI data. If so, and I expect it to be successful, it is a major advance in climate science. Thanks to Soon and others, we were aware of the lag between solar activity and climate. David Evans’ Notch-Delay Model is the first practical application of that knowledge to quantify future temperature response to changes in solar output and will assist with planning.