Currently, first the full model (including the VPD effect) is fitted, and only if this is not successful reduced versions are fitted.
If there is no large VPD in the dataset, one could omit the full model and start with model version neglecting the VPD.
Currently, first the full model (including the VPD effect) is fitted, and only if this is not successful reduced versions are fitted. If there is no large VPD in the dataset, one could omit the full model and start with model version neglecting the VPD.