Friday, April 1, 2011

Trenberths missing heat has been found!

There have been a lot of discussion about "Trenberths missing heat", e.g., see "Examining Trenberth’s ‘The heat will come back to haunt us sooner or later’ statement". What it is all about is that in the global energy budget there is approximately 0.8 W/m2 missing to confirm the AGW hypothesis based on CO2.


(Picture linked from US National Science Foundation)

After careful investigation I have today, April 1, 2011, found where the heat is! The explanation is a rather simple Y2K problem showing up in Trenberths calculations. As you all know, there are 365 days every year, except for every four years which has 366 days. Except that there are exceptions to first exception. Please consult Wikipedia's page about leap year for the details.

The problem that has happened in Tranberths calculations is that from 2000 and onwards, he has forgotten to add the leap days to the leap years. Not a big deal? Actually it is.

The energy that heats the earth is approximately 1366 W/m2, see Sunlight. The increase in energy when adding the approximately 0.25 days per year is thus: 0.25/365. This result in the total change in the energy budget with 1366*0.25/365 which is approximately 0.94 W/m2. It could possibly be that the year 2000 is correctly handled in Trenberths calculation and thus the difference will be lower.

In any case, this difference is well within the estimation error done by Trenberth. Thus, the missing heat has been found to be a simple calculation error for leap years following the year 2000.

Personally I think this is a disaster. This actually shows that the AGW hypothesis has been correct all the time but the calculations have been incorrect. Until today, I have always thought it was the other way around, i.e., the hypothesis is incorrect since it could not be possible to do simple calculation mistakes with the huge amount of review each climate science paper must undergo. Not to forget all "skeptics" that would do anything to find the smallest of errors in any paper.