Dear author,
You have convincingly responded to the issues raised by the reviewers. Please implement the suggested changes in a revised manuscript.
Best regards,
Anne Van Loon

see attachment

Dear authors,

Thanks for the revised manuscript. Most of the pending questions and reviewer comments have now been addressed. You did well in leaving out the resilience concept, the upscaling of the plot scale measurements, and parts of the introduction.

One issue that has not been dealt with in the newest version of the manuscript is the lack of discussion. Your reply to this comment in the response is not satisfactory. You have to put in a separate Discussion section, in which your results are placed in a wider context, compared with literature, and the uncertainties of the methodology discussed, so that the value of your results can be evaluated. Both reviewers indicated items that need discussion and you can use discussion paragraphs and sentences that are already spread around in the manuscript.

After the major changes you made in the content, the text also needs restructuring and the abstract needs more rewriting. I do some suggestions below, but I also suggest you use your co-authors or other colleagues (preferably native English speakers) for proofreading, both for structure and spelling.

Comments:
- The abstract needs more new results (like in conclusions) and a concluding discussion sentence. The recovery results mentioned in the abstract are not in the manuscript anymore, which I think they should be.
- “Capacity recovery”? Should this be “recovery capacity” or just “recovery”?
- Methods section: please reorder the subsections 3.1 and 3.2 in a more logical sequence and rephrase the titles, e.g. 3.2 “Catchment modelling”, 3.2.1 “Calibration & validation”.
- 3.1: You do not use Eq. 1 or the methods mentioned in this section in the rest of the manuscript.
- P.12, l.1-3: should this text be removed since you deleted this part from the analysis? Instead you can write that Ks & Kv cannot be measured and need to be determined indirectly by modelling actual ET. You should also add a sentence about why you want to know Ks and Kv, i.e. to compare the catchments.
- P.12, l.6-7: is model used to assess the impact of soil moisture on ET? I think you use it for other purposes.
- P.13, l.26: did you use observed or simulated Q in the drought analysis? Also specify that in the figure.
- 3.3.1: the recovery analysis explained in this section (l.25-27) is not followed by results in the Results section. Did you do a recovery analysis on the model simulations? Then that should be discussed with the results of the recovery analysis on the observations, mentioned in the abstract. Both should be included in the Methods and the Results sections.
- 3.3.2: the description of the recovery analysis (l.10-13) is unclear. How is the recovery period from vegetation stress quantified exactly?
- 3.3.3: How are the scenario results compared to original simulation?
- 4.2: This section should be split up in multiple sections, one on general model results, one on ET and some parts that should go in de Methods or Discussion section..
- P.16, l.21-26 & l.29-30: this part should go in the Methods section
- P.17, l.6-9: Sn is out of range, discuss this.
- P.17, l.19-30: Move this paragraph to the Discussion section. Last sentence is unclear statement.
- P.18, l.1-13: first part of this paragraph should be moved to before modelling results, second part should go to Discussion section.
- 4.3: change title, e.g. “Drought severity”?
- Remove “on the other hand” in several sentences.
- 4.5: these new results need a figure (p.19 l.28 – p.20 l.4) and/or a table (p.20 l.5-15) to visualise the results. Also the vegetation stress recovery needs to be compared with the soil moisture drought recovery.
- P.20, l.16-25: why do you need the CDD analysis? Why is it mentioned in this paragraph?
- P.21, l.1-2: what are these numbers? Total P & ET over the modelling period? Or per year?
- Table 3: does this show normal NS or NS based on log Q?
- Figs. 3-5: clarify which variables are simulated and which observed
- Fig. 6: chose clearer colour scheme and make a better legend / give more information in the caption.
- For drought recovery, you might want to look for inspiration at this recent paper in HESSD by Parry et al. (in review).

Parry, S., Wilby, R. L., Prudhomme, C., and Wood, P. J.: A systematic assessment of drought termination in the United Kingdom, Hydrol. Earth Syst. Sci. Discuss., doi:10.5194/hess-2015-476, in review, 2016.http://www.hydrol-earth-syst-sci-discuss.net/hess-2015-476/

Dear authors,

The paper can now be published if the remaining grammar and spelling mistakes are corrected, the abstract is updated to reflect the changes in the manuscript, and one last item is solved (see below).

p.18, l.24-30: “So, seemingly the drought events characteristics are similar and independent of the climate. This in line with the literature. For instance, based on a global map of Köppen-Geiger climate types (Wanders et al., 2010) and using a similarity index SI (Kim et al., 2003), Van Lanen et al.(2013) concluded that independent of the climate type, similar combinations of duration and standardized deficit volume were found in a large number of drought events. They analysed 1495 locations around the world and a data set over the period 1958 to 2001.”
And p.23, l.6-7: “Initially, we can infer that the drought events are independent of the climate.”

>> drought events characteristics are independent of the climate by definition with the method you used. Firstly because the threshold you calculated is based on a percentile of your timeseries and secondly because you standardised the deficits. Therefore this text should be removed.