Atmospheric Residency & Oomph (part 3)
A quick look at the HADCRUT5 global mean annual temperature anomaly and human emissions over the last 170 years
In part 2 of this series I decided to try out temporal causal modelling using the concept of Granger casualty in order to determine the range of likely values for atmospheric residency of human emissions using data sourced from here and here, the idea being to determine the best predictors for total atmospheric CO2 using a range of human emission residencies.
Upon cranking the handle twenty times I discovered that human emissions residencies of 80 - 140 years popped out as being statistically significant predictors, with time series data for 90 and 100 years yielding the best fitting model.
In turn this enabled the estimation of the current contribution of human emissions to total atmospheric CO2, this dropping out at 49%. This all felt very plausible until we came across a fly buzzing in the ointment: human emissions residencies of 90 - 100 years necessarily means non-human emissions are declining on a planet that is warming.
Today I’d like to illustrate this paradox with a colourful slide, and here it is: