We consider the Westdorpe barometer as one barometer to be tested for drift based on the #51 analysis. We expect for this barometer to have no drift. This is the resulting plot:
The red curve looks pretty stable: no drift at all. Now let's estimate the error of the red curve like in #55:
Very nice Gaussian distribution. Since we based our graph in #55 on whisker width, this KNMI barometer with whisker width of 6.3 cmH2O is situated somewhere in the middle of the comparison graph in #55.
Note that the KNMI data was aggregated to 2 observations per day (12h and 24h), and that the measurements were converted to cmH2O. Also note the metadata of the KNMI series: "Opmerking: door stationsverplaatsingen en veranderingen in waarneemmethodieken zijn deze tijdreeksen van uurwaarden mogelijk inhomogeen!"
Here we analyse the Westdorpe barometer.
We consider the Westdorpe barometer as one barometer to be tested for drift based on the #51 analysis. We expect for this barometer to have no drift. This is the resulting plot:
The red curve looks pretty stable: no drift at all. Now let's estimate the error of the red curve like in #55:
Very nice Gaussian distribution. Since we based our graph in #55 on whisker width, this KNMI barometer with whisker width of 6.3 cmH2O is situated somewhere in the middle of the comparison graph in #55.
Note that the KNMI data was aggregated to 2 observations per day (12h and 24h), and that the measurements were converted to cmH2O. Also note the metadata of the KNMI series: "Opmerking: door stationsverplaatsingen en veranderingen in waarneemmethodieken zijn deze tijdreeksen van uurwaarden mogelijk inhomogeen!"