Small bugfixes:
- ISIMIP debiaser: Adjusted distribution fitting to be on np.float64, added an option to manually adjust the KS goodness of fit threshold for parametric fitting and reduced the default argument to 0.15, fixing #40.
- ISIMIP debiaser: ensured that the scaling of how many values are set to the respective upper and lower bound should be symmetric in the rare cases where more values than available in cm_future need to be set.
Full Changelog: v1.3.0...v1.3.1