All the remaining data sets are complete for 2016, so here's the final update on my attempt to predict which data sets would set a record.
How the Year Progressed
BEST and GISTEMP both dropped in December, whilst NOAA and HadCRUT rose.
Predicting the Margins
All data sets finished above zero, which means everyone set a new record, which shouldn't be news to anyone. The data sets split into two groups as far as the extent of the record - UAH 6.0, RSS 3.3, HadCRUT and NOAA were close to their previous records and could be described as statistically tied, UAH 5.6, RSS 4, GISS and BEST all beat the previous record by a significant margin. It's important though to remember that the previous record is different for the satellite data than for the surface data. If we were only comparing 2016 with 1998 all surface sets were much warmer (around a quarter to a third of a degree), whereas if we were only comparing to 2015 all satellite sets were much warmer (around a fifth to a quarter of a degree warmer). Here are the full figures.
| Data Set | Over Record | Over 1998 | Over 2015 |
|---|---|---|---|
| HadCRUT 4 | 0.01 | 0.24 | 0.01 |
| UAH 6.0 | 0.02 | 0.02 | 0.25 |
| RSS 3.3 | 0.02 | 0.02 | 0.19 |
| NOAA | 0.04 | 0.30 | 0.04 |
| GISTEMP | 0.13 | 0.36 | 0.13 |
| BEST | 0.15 | 0.36 | 0.15 |
| UAH 5.6 | 0.17 | 0.17 | 0.23 |
| RSS 4 (TTT) | 0.17 | 0.17 | 0.27 |
Here are the time series for some of the data sets to put this into perspective.
Those who want to ignore the obvious reality of rising temperatures will inevitably focus on those sets that failed to break a record by a significant amount (just as I'm sure they would have emphasized the lack of significance if they had failed to beat the previous record by a similar amount). But this is misleading for a number of reasons - in particular the fact that every data set set a record makes it more likely that 2016 really was the warmest year on record. For it not to be a record requires every data set to be wrong, in the same direction, in some cases by a significant amount.
I should also repeat my opinion that whilst setting a new record is interesting it is not that important. A new record tells us little about how quickly the planet is warming, and a lack of new records tells us even less. A single exceptional year, such as 1998 in the satellite data, can mean it's almost inevitable that there will be a long period before the next record breaking year. 2017 will not be a record year, and 2016 was so warm it's unlikely will see a new record for many years to come, but that will not mean the planet has stopped warming.
The Probability Graph
For what it's worth here's the final probability graph.
Revisiting the IPCC 2001 Projection
Finally, now we have another year we can look at how the long term temperature is going.
This graph is based on a previous post making fun of Christopher Monckton's speedometer
which he introduced one month last year as replacement for his Great Pause
graph,
and then promptly forgot.
I pointed out that the IPCC predictions for short term temperature rises, 2025 compared to 1990, where for twenty year averages, and looked at how rolling 20 annual averages compared with the 2001 projection. This said that 2025 was expected to be between 0.4 and 1.1 °C warmer than 1990.
We are now seeing the 20 year climatic values for 2006, 16 years into the 35 year forecast period. Both HadCRUT and the Satellites (Monckton's preferred average of UAH 6.0 and RSS 3.3) continue to rise within the IPCC's projected range. The satellite data for 2006 is 0.23 °C above the 1990 average. HadCRUT is 0.30 °C above 1990.
