NASA and its continued use of the 1951-1980 long term averages

I noticed this graphic of global temperatures for the last winter [DJF] that NASA posted in their Twitter feed and thought it looked overly warm for IONA when anomalies this winter had been negative particularly in the north. I generated my own anomaly chart using reanalysis data which I also download from NASA. The difference in the anomalies is that I use 1981-2010 long term average to calculate them with, whilst NASA, for some unknown reason only know to themselves, use the 1951-1980 averages. By doing this it has the effect of shrinking any cooler areas quite dramatically, in fact it would give you the distinct impression that this winter as a whole had been mild across IONA when it wasn’t. I think the guidance from the WMO about long term averages – not that anyone listens to them – is that for climatological purposes the period 1961-1990 should be used. At the moment of the seven global temperature datasets two use the period 1951-1980, one the 1971-2000, three the 1981-2010, and two use the 1991-2020. You would think that between them they could agree on the same period that they all could use. With regard to this particular chart the climate skeptic could rightly complain that NASA does this to exacerbate the warming that’s gone on since the relatively cold thirty year period between 1951 and 1980 and now, and I for one wouldn’t blame them.

The note at the bottom of the NASA graphic is curious. They say that ocean data is not used over land which you would assume is how it should be. They also add that ocean data is not used within a 100 km of a reporting land station. They may do that to help in the kriging of the observational data, but I assume this means that where prevailing winds are off the ocean, a strip of sea 100 km wide will have values that are interpolated with regard to land observations rather than the sea surface temperature. I just hope that this is not how they calculate their gridded reanalysis 2 metre temperatures.

4 thoughts on “NASA and its continued use of the 1951-1980 long term averages”

  1. A baseline has to be adopted and stuck to. If not then by 2080 last winter 20/21 will look like a beast compared to the 2051-80 average. I think NASA should stick to the baseline they are using but I can see some sense in everyone sticking to the same baseline.

    1. I agree that they should all use the same baseline and that should be the latest thirty year period (1991-2020) as both the UAH and ERA series now do.
      The WMO should be taking the lead in enforcing observational standards and collecting climate and observational data from all its member nations to produce a global and impartial temperature series of it’s own.
      Instead what we have is at least seven national Met services all producing slightly different estimates all from differing long term averages.
      Unfortunately the WMO is very similar to the UN in that it seems to have forgotten what it’s real purpose is – if it ever had one at all.

  2. “and that should be the latest thirty year period (1991-2020) as both the UAH and ERA series now do.”

    In terms of measuring how the weather differs from the current ‘normal’ it makes sense to use current means as a comparison.

    *But*, by, say, 2121 it’s highly likely we’ll have warmed by a degree to four more and using that most recent mean then (after a long period of warming) would obscure change, statistically air brush it away. Informed people of 2021 would know the UK is ridiculously warm compared to realistic averages. However lay people would just think ‘it’s a bit warmer than normal’ – no need for any concern then, everything is fine compared to ‘normal’ (2091-2120)…?

    Likewise a sea level rise could be wished away by updating MSL to keep up with rate of sea level rise. Or declines in wildlife could wished away by saying ‘this is 2121, yes there was more wildlife in 2021 and 1921 but compared to our normal of there being no wildlife things are perfectly ok and there is no need for concern about the state of the planet’.

    I don’t want problems conveniently wished away, glossed away, statically airbrushed away by the use of changing comparisons.

    And it works the other way. If the climate cools I don’t want cooling masked by updating average. I don’t know how cold winters in the 1680 were compared to, um, the 1680s but arguable a lot less cold than we would think them.

    1. Hi Peter

      The nice thing about my DIY series is that I can produce anomalies that aren’t just restricted to 30 years.
      Personally I think anomalies should be calculated for the length of the entire data series or a minimum of 100 years.
      That should get rid of any bias that may have occurred in more recent years.


Comments are closed.

Scroll to Top
%d bloggers like this: