Six of the best!

I have been very busy over the last week updating an old program that I use to download and visualise the monthly global temperature data from the UKMO and NASA. I haven’t devoted enough time to the subject of Global Temperatures in recent years so decided to improve it by adding that allows me to now monitor and visualise data from all six of the world’s leading global temperature datasets:-

Global Temperature Datasets

  • HadCRUTv4 (UKMO and the Climate Research Unit at the University of East Anglia)
  • GISTempv4 (NASA and the Goddard Institute for Space Studies)
  • NOAAGlobalTemp (NOAA)
  • Berkeley Earth (University of Berkeley)
  • UAH (University of Alabama)
  • JMA (Japanese Meteorological Agency)

Instead of the more usual bar chart of annual mean anomalies since 1850 or when ever, the first graph that I decided to create was one that displayed all six datasets altogether as a series of line graphs of 12 month moving averages. I did think about using a 13 month period instead of 12 to create my averages, but I decided against this. Just for good measure I added functionality to centre my moving average data as well as being to display them as trailing or leading averages too. I did toy with combining all six series into a single series but decided against it.

Shape of the curves

As you can see there is a good deal of agreement in the overall shape of most of the datasets. Unlike the UAH series from the University of Alabama which derives its data from satellite observations of the lower troposphere, the five other datasets use data from surface observations made on land or sea, that being so the UAH series is still noticeably more reactive than any of the series even after being smoothed into a 12 month moving average.

Other idiosyncrasies

It’s also noticable that the JMA series from the Japanese Meteorological Agency has the lowest anomalies of the six, and the GISTEMP series from NASA the highest. The Berkeley Earth series seems to be very closely tied to the GISTemp series, despite it being just a little colder.

Long term averages

A lot of the differences between the various datasets can be explained by the fact that they don’t all use the same long-term average period to calculate their anomalies with. Berkeley Earth, like GISTemp, use the 1951-1980 period, which might explain why they are closely tied to each other. NOAA in contrast use the 1971-2000 period, whilst the other three series use the latest 1981-2010 period for their anomaly calculations. Don’t ask me why they just don’t all use the latest 1981-2010 period for their calculations, it would make life so much easier.

Linear Trend

Here are the linear trends from the six series for the last 30 years since 1990.

  • +0.171 – HadCRUTv4
  • +0.210 – GISTempv4
  • +0.194 – NOAAGlobalTemp
  • +0.201 – BerkeleyEarth
  • +0.133 – UAH
  • +0.152 – JMA

GISTEMP4 and Berkeley Earth have the highest trend in warming at over 0.2°C per decade, which of course is a whopping 2°C per century if global temperatures continue to rise at this rate. Surprisingly HadCRUT are showing a less steep climb in anomalies of only +0.171°C per decade. Bringing up the rear are the JMA and the UAH series with even lower increases in the last 30 years. It’s easy to understand why when seeing these 30 year trends, why the UAH data is so popular with global warming deniers.


I am surprised by the current spread of temperature increases over the last 30 years in each of the six series. The shapes of the graphs are very similar but I don’t think it can all be explained away by difference in the LTA period that they use. There is no doubt that global temperatures are bouncing back strongly after the dip that followed their meteoric rise in 2015 and 2016. Each of the six datasets show no signs of this steep rise relaxing in early 2020 as far as I can see. At least now I have an application that can produce a very detailed way of visualising it.

Because of the importance of global temperatures I’m going to include and update this chart in its own static page on my blog.

Scroll to Top
%d bloggers like this: