NOAA Updates Their GMST Dataset to v. 5.1

Within the last week or so, NOAA issued an update to their Global Mean Surface Temperature dataset. The update to v. 5.1 now has full global coverage (mostly due to increased coverage of the Arctic) and extends back to 1850. The results are pretty close to to the v. 5 dataset, and I suspect differences in trends between the two will not be significant over time frames of 30 years or longer. Below I plotted v. 5 and v. 5.1 with the difference between the two (v.5.1 - v.5). The changes are minimal and there is virtually no trend to the difference. The additional warming in recent years is likely due to the increased Arctic coverage.

I also found the global raw data (with no adjustments or bias correction) and plotted v. 5 and v 5.1 so that you can see the effect of bias correction on global temperatures. The largest adjustments occur before 1944, and they are dominated by increasing sea surface temperatures to account for changes in the sampling methods from ships. Contrary to the claims of some popular bloggers and YouTubers, bias correction does not increase global temperature trends. In fact, by warming past temperatures, it actually decreases the overall warming trend. The raw data I found only covers 1880-2016, but I think you get how the overall affect of bias correction affects global temperature trends.

I downloaded the monthly data for v. 5.1 as well. Since the time series now goes back to 1850, I can set it to the 1850-1900 baseline, which is the baseline used by the IPCC. Two other Datasets, HadCRUT5 and BEST, also go back to 1850.  Below I plotted the monthly data through January 2023 with an 1850-1900 baseline and a twelve month running mean.

The IPCC defines its thresholds in terms of a 30-year mean anomaly with respect to the 1850-1900 baseline (which is taken to be warming above preindustrial levels). Now we can use three major global datasets to evaluate how much the globe as warmed above preindustrial levels. I thought it would be good to average NOAA, HadCRUT5, and BEST, then plot centered running 30-year means. I then projected the most recent 30-year trend from the end the last 30-year mean. The trend projection is shown below with dotted lines. From this, the average of three datasets shows that we have exceeded 1.2 C warming above preindustrial levels, and the most recent 30-year trend is 0.21 C/decade. If you compare this post to my 2022 summary from last month, NOAA's update doesn't really change what we understand about global warming much at all. But it does add some confirming evidence to results from HadCRUT5 and BEST.