6 thoughts on “Only 7.9% of NOAA temperature stations are accurate to less than 1C”

  1. If 70% of the stations are accurate to +/- 2% or more, how can any reading with 2% or less variation be valid? I’m not a scientist, but?

  2. In actual fact most temperature devices are pretty good if they are within a 1°-2°C accuracy

  3. “pretty good if they are within 1°-2°” is fine for normal use…but don’t use something like that to suggest we have to change the human lifestyle because of estimates of a hundredth of a degree.

  4. I recently posted a comment on fabiusmaximus explaining how researchers get incredibly small errors using imprecise data. It has to do with taking the mean value of a large umber of mean values. The Probable Error of the Mean (PEM) has a much smaller variance than that of the raw data.
    The researchers are conflating measurements (raw data) and mean values (derived data) and appropriating the variance of the latter as the uncertainty of the former.

  5. Well, I have discussed this with someone who seems convinced that all that is needed is more sensors, as the results of each cancels the errors of the others (or something like that; the logic was getting bizarre). For instance, if the thermometers have an acceptable error of 1°C, then ten thermometers will reduce this error margin to 0.1°C; 100 will reduce it to 0.01°C. This, it would seem, is how they can declare with such confidence that the world’s oceans have increased in temperature by 0.02°C!

    What raw data can give us is trends; accuracy of the instrument is not paramount, though stability is (obviously). Interestingly, “homogenisationing” (is there really such a word?) of the data seems to have reversed the trends evident in so many stations, worldwide. What a coincidence!

  6. With NOAA eliminating hundreds of stations, they can get the desired result. Especially eliminating the rural stations that do not show warmth.

Comments are closed.