Suggestions for revision or reasons for rejection  The authors had obviously problems to reply to my critics in respect to the ‘ reproducibility’. So I will elaborate on this point a little bit more explicit. For reproducing the work, I would expect a detailed description of the equations (no f of something), a table with the original parameter values and the intervals set around them. I assume the authors started with an initial parameter setting and set the intervals around them, which could be done in a general way, but has to be described. If you have freely chosen the intervals, you have to motivate each of them. Furthermore, I would like to see the expected values of all parameter distributions and how they shifted from the original ones.
GMD has well established standards for the reporting of model details. Last but not least it is one of the central aims of the journal. A hint to an existing model copy that is available on request is like referring to a lab book instead giving the exact details of an analytical procedure. You might use
http://www.geoscimodeldev.net/9/4133/2016/gmd941332016.pdf as an orientation for your revision.  

Suggestions for revision or reasons for rejection  My initial concerns with this paper have all been addressed satisfactorily, with one partial exception. The authors do not include residual variance when estimating prediction uncertainty, explaining that they are interested in predicting true as opposed to observed values. This assumes that residual variance is due solely to measurement error, which is almost certainly not the case for very simplified models of complex systems. The residual error is probably, to a very large extent, due to model error and thus will also contribute to prediction error. This should be acknowledged.
Other than that, there are some typing errors that should be corrected: P4L6 "description discussion", P11L3 "that that", P13L33 "can" should be "could", P14L16 "begin" should be "being".  

Dear Dr. Smith,
with the two reviews in I have now not a much clearer picture and after my own reading I think you should
* address the remaining issue of referee #3, which should be easy enough
* expand more on the description of details in your study, as requested by referee #2
To this end, I would ask you to mainly spill out the details of equation 2, which is currently presented as Gleaf = f(some parameters and other functions)Rplant. As GMD is dedicated to supply all the page space to fully describe models, I concur with referee #2 that a more explicit description of your model is desirable. I see that the Guilbaud et al. 2014 also did not describe the model in more detail and that the equations have been at least partially described elsewhere, but still the detail of model description is insufficient here. You may want to expand the full set of equations in an appendix or in the paper, as you see fit.
Secondly, I would like to see an extra column in table 2 with the predefined parameter space used in your study and (if possible) any justification or reference for that.
I think that would make the paper of greater value to the research community.
Best regards
Christoph 
Thanks so much for this fast, adequate and comprehensive response.
I still have 2 issues I need to bug you with, which is:
* on page 18 you now state "The diffuse radiation coefficient is assumed to be a constant"  could you please add the value and units of that constant?
* in table 2 in column "Fixed" the variable fJ has "np", which I suppose should be "no", but please correct/explain as necessary.
Cheers
Christoph 