Dear Dr. Malek and co-authors,

I'm happy to now have 2 reviews for your revised paper (one of the original ones and a new reviewer). I apologize for the long turnaround time this revision needed.
However, both referees are now largely content with your revisions and only referee #3 raises 2 last points.
On top of these points, I have some more minor points, all with respect to your discussion of other approaches.
I do see that referee #3's first point may be somewhat difficult to address, as deciding which model to use under which circumstances is often hypothetical and would require some testing and model (setup) evaluation. So maybe you can discuss this with some specific examples where some models may be fit for purpose by addressing relevant processes and others may not as they don't.

Specific comments:
As I happen to know the gridded crop model development to some extent, and LPJmL in particular, I noticed that you list JULES and ORCHIDEE as simplified model, which is not always true, see
* Wang et al. 2016 http://www.geosci-model-dev.net/9/857/2016/
* Williams et al. 2017 http://www.geosci-model-dev.net/10/1291/2017/

Also, the Fader et al. 2015 reference is not the best one for discussing the hydrology-crop linkage in LPJmL, maybe better try
* Rost et al. 2008 http://onlinelibrary.wiley.com/doi/10.1029/2007WR006331/abstract
* Biemans et al. 2011 http://onlinelibrary.wiley.com/doi/10.1029/2009WR008929/abstract
* Jägermeyr et al. 2015 http://www.hydrol-earth-syst-sci.net/19/3073/2015/hess-19-3073-2015.html

and the LPJmL model has been used and tested for regional studies (contrary to your claim on page 6), see
* Biemans et al. 2013 http://www.sciencedirect.com/science/article/pii/S0048969713006451
* Langerwisch et al. 2013 http://www.hydrol-earth-syst-sci.net/17/2247/2013/hess-17-2247-2013.html
and it is also no longer true that LPJmL uses country-specific efficiency data, but explicitly simulates the loss terms
* Jägermeyr et al. 2015 http://www.hydrol-earth-syst-sci.net/19/3073/2015/hess-19-3073-2015.html
* Jägermeyr et al. 2016 http://iopscience.iop.org/article/10.1088/1748-9326/11/2/025002/meta

Lastly, the sentence on page 6 reads as if LPJmL had a simpler implementation of LAI and root development than EPIC, but in fact is uses the EPIC scheme for that.

I'm confident that all of the issues raised by referee #3 and myself can be easily addressed and look forward to the final version of the manuscript.

Best regards
Christoph Müller

Dear Dr. Malek and co-authors,

thanks for the revisions.
In your response letter, you mention to have removed the sentence that states that LPJmL uses country-specific efficiency data. However, the statement is still in your revised manuscript (page 5, line 34 to page 6, line 1).
Surely just an oversight, as you have answered 2 different aspects (regional application & national efficiency assumptions) jointly but then only executed the first correction.

Please correct this.

Best regards
Christoph Müller