fitting new variograms using the up-to-date dwd weather data (here for precipitation)

DWD weather data is downloaded from https://opendata.dwd.de/climate_environment/CDC/observations_germany/climate/daily/more_precip/historical/

A comparison to the precipitation station data stored in /data/stohyd/data/processed/Germany/DWD/pre/pre_1900-2015/ to the new station data showed that some stations changed:

Nr Stations| Old LUT     | New LUT     | Same        | Abandoned   | New Stations 

pre             |5394           |5616             |5380          |14                     |236         
tavg           |1059           |1213              |1055         |4                        |158         
tmax         |1059            |1213             |1055          |4                       |158         
tmin          |1059           |1213             |1055          |4                        |158

However it is unclear which data was used for calculation of the variograms used in the current drought monitor version (status: 2019-11-28).

Calculation of the variogram resulted in the following theoretical variogram parameters (using exponential variogram type)

   nugget        sill       range  Type    Easting   Northing     BIAS     RMSE        r

  0.3199E+00  0.9526E+00  0.6429E+05     2        0.0        0.0   0.0098   0.0009   0.9946

The Old variogram parameters were:

    nugget           sill          range    varTyp

0.2000E+00     0.9014E+00     0.3990E+05         2

In the graph the new variogram is plotted using new theoretical and experimental variogram data)

Interestingly, the old nugget seems to fit the data better. The range is remarkable bigger in the new data. @thober any possible explanations?

Edited Nov 29, 2019 by Friedrich Boeing