The Met Office are proposing to use a 10 year average temperature indicator to define when and if the 1.5 degree since industrial periods will be exceeded. The pre-industrial period is defined as 1850-1900 which is different from the 1961- 1990 baseline that are used for calculating station anomalies because this maximises the available station count. The offset for the preindustrial period is consequentially based on these anomaly values before 1900.
A comparison of the latest HadCRUT5 with my Spherical triangulation results shows an almost perfect agreement between the two methods. This then results in a joint pre-industrial baseline (1850-1900) of -0.4C. Thus the Paris agreement to limit total warming to 1.5C equates to limiting these measured (baseline) temperature anomalies to less that 1.1C.
It is clearly evident that the year to year variability is large, for example due to El Nino and other transient effects which makes estimating the effect due to CO2 alone difficult. However there is another way to calculate global temperatures which allows for far longer integration periods avoiding this problem. This is based on Icosahedral 3D binning as I described here. This method allows for the integration period in each cell to be extended up to a decade. These are the results that I get using the latest complete 2010s decade measurements.
Now we compare the decadal temperature values with the normal annual calculated values by spherical triangulation.
This result proves that the decadal temperatures describe very well the underlying temperature trends in the data, unaffected by annual variability. As a consequence of this we can simply extrapolate the observed linear decadal trend forward in time to determine when the Paris agreement limit of 1.5 degrees will likely be exceeded.
The conclusion is that a net warming of 1.5C since preindustrial times exceeding the Paris Agreement will most likely occur in 2032. This figure of 1.5C, if I remember well, was anyway a compromise resolution made at the end of the conference.