A few of the last few days measurements from ERA5 have GEFS coming in a bit colder than earlier averages (about -0.05C)...

but even if they are on average cooler by -0.1C (indicating a December gistemp of 1.255C for an offset of 0.1C, although a 0 C offset seems most likely as a meta prediction still), my current model still puts most of the probability towards 1.28 C

It will take a very warm end of December (5-25%) to push it towards 1.29 C which doesn't look yet likely

@ChristopherRandles

Ok... I looked at the gistemp code and it does seem that gistemp calculates the average off the unrounded monthly means and then does the rounding on the annual, so I modified some of the code to the months data from the last run's data to get full precision as this might change the results...

This is gistemp for January to November for 2024 using ghcnm.v4.0.1.20241209 (and my own ersstv5 version) in centikelvin ...

[124.25124494551324,143.43104891032047,139.0563506025243,131.27736846330365,115.68215268652695,123.98082189837416,119.95886139265997,129.95921868877284,122.96425681924441,133.05707795777147,130.45975731699357]

As you can see it was very close in terms of November being 1.30 instead of 1.31 (the last couple runs were closer to 1.30). Hopefully the following should help anyone with their J-D calculations...

Now to work backwards from this data we have to get the threshold values for December that will average and round according to this questions bins (i.e. the bin -0.5 deg C) ...

For average 1.245, GISTEMP_DEC temperature: 0.7992184031799496
For average 1.255, GISTEMP_DEC temperature: 0.9192184031799471
For average 1.265, GISTEMP_DEC temperature: 1.039218403179948
For average 1.275, GISTEMP_DEC temperature: 1.1592184031799473
For average 1.285, GISTEMP_DEC temperature: 1.2792184031799465
For average 1.295, GISTEMP_DEC temperature: 1.3992184031799475

So the new exceedance bins for December become:

[0.7992184031799496, 0.9192184031799471, 1.039218403179948, 1.1592184031799473, 1.2792184031799465, 1.3992184031799475]

and I can use the same code and copy my notebook for december now..

without worrying about converting back to the original bins for this question (they just correspond in order to the bins in the question anyway)... and following the 0 offset (purple line) metaprediction for ERA5->GISTEMP offsets, the one with the highest probability (1.16-1.28 for december) corresponds to the 1.28C bin for J-D.

In order to make it into the 1.29C bin it would require a Dec temp of 1.279 C.

This is not outside the CI but is closer to the end ...

GIS TEMP anomaly projection (December 2024) (corrected, assuming -0.086 error, (absolute_corrected_era5: 13.243)):
  1.199 C +-0.097

I think I actually have done a bit better this year than that CI suggests (despite some poor mana returns) and would guess I won't be more than +-0.05 C off so I am guessing its fair to bet the sum of both bins down to ~15% then ... but this is a guess and judgment call with half the month left so far...

Its possible that prior months shift further also with new ghcnm and ersstv5 data also in a few weeks, but I haven't done any sensitivity analysis on that as this is as much as I want to do effort-wise.

D-N shown as 1.29 and consensus seems likely less than last years 1.35 for Dec.

At https://manifold.markets/ChristopherRandles/global-average-temperature-december
JRP seems to believe in much lower values like less than 1.2C for Dec

This would seem to suggest 1.28 or lower, but here there seems more preference for 1.29+