Dear Dahlin et al.,

Both reviewers find your study of sufficient interest to merit publication in Biogeosciences. Whilst I agree with the reviewer's concern that the lack of an assessment of different aspects of the seasonal cycle of LAI (e.g., bud-burst and senescence) limits the scope of the analysis, I also understand your argument that, given the global scale of the study and the complex nature of the seasonal evolution of LAI in water stressed environments, such an analysis would need to be extensive in order to be of value.

I look forward to receiving your revised manuscript.

Best wishes,
Trevor

Dear Authors,

Thank you for submitting your revised manuscript, and incorporating the suggestions of the referee's. I noticed quite a few typos and grammatical errors in the revisions (please revise the language throughout), and would suggest some reinterpretation of the results of the Latin Hypercube sampling. See details below.

Trevor


Type: Line 219 "to achieve"

Line 225: "The Latin hypercube methodology is a particularly efficient means of investigating a multi- dimensional parameter space"
The latin hypercube method is one of the most inefficient methods for searching multi-dimensional parameter space. What you mean is it is the least computationally intensive, as you can define the number of parameter sets to run. More intelligent algorithms (e.g., MCMC or Kalman filter) search efficiently, in that they can find the optimum without searching all combinations. Please revise.

Line 266: "which highlights"

Line 272: "Third: the parameter exploration method may not be effective for identifying global mimima". You should probably include some discussion regarding this.

Line 273: "eliminate"

Line 283: "as function "

Line 292: " this metric of phenological trigger. " Rephrase?

Line 397: "never did CLM-MOD predict more than two peaks in a year". Rephrase