All three reviewers expressed major concerns, but they also provide useful guidance on how to improve the manuscript. Given the substantial concern, it is crucial that the authors address all the comments carefully when revising their manuscript. The revised manuscript will be send out for review again.

Best regards,
Jan Seibert

Thanks for your efforts with revising the manuscript. As indicated by the one reviewer from the previous round who looked again at the manuscript, the manuscript has clearly been improved However, the reviewer also noted that no new analyses have been performed although this would have been motivated. This is unfortunate as it means the need for another round of reviews. It is really important that all concerns of the reviewers (both first and second round) are addressed.

best regards,
Jan Seibert

Thanks for the continued efforts revising this potentially very interesting manuscript. To be fully honest, I have to say that the revisions in this and especially also the previous rounds are a bit on the minimalistic side. There are several comments where more substantial changes and/or new computations would be possible/desired.

There are still multiletter variable names (Eq 5, 6, 10), please follow the author guidelines.

All reviewers state that the manuscript is hard to read. Splitting results and discussion would clearly help to make the manuscript more readable (if you do not believe me, almost all scientific writing advice books recommend this for 'normal' journal papers!)

"the efficiency of the model when compared to real discharge data might not be accurate”

The point raised by the reviewer here is that there is a clear risk of being right for the wrong reason (see J.Kirchner, 2006, WRR). Since you want to make the point that including BC leads to better simulations it is crucial that the rest of the model works for the right reasons. Otherwise the improved fit might just be because some other error is compensated.

While it is good that model uncertainties now are better discussed than in previous versions, the reviewer comments actually would have motivated more new calculations. Especially a quantification of uncertainties would be useful. (see Pappenberger and Beven, 2006)

Thanks for your efforts with revising the manuscript, I would say it clearly became (even) better. There are a few minor issues listed below, which I am confident you will easily implement these technical corrections.

Best regards,
Jan Seibert

Minor, technical issues:

P11L10, change NSE to E_NS

P11: I do not think the text between the headings 5. And 5.1. is needed, consider removing this

Table: 'Kirchner parameter' does not sound as an useful description, could you change this, perhaps just say ‘empirical coefficient’

Figures: some of the numbering along the axes is too close to the axes, please check and correct (e.g. Fig 7 left axis)

Fig 4: please consider changing the colors, red and green is not good for the color blind