There are some reasons of this variation.
First, hfe is not very tightly controlled parameter, because wafer to wafer some process parameters show quite large deviations from nominal value that's why hfe is changed in a large region.
Second, hfe is tightly dependent on collector current and it has a maxima and two minimas.
Third, eventough BC548B is industry standard BJT, it's manufactured by different companies and that's why it has some different process parameters and variable.