Ensembles for Viticulture Climate Classifications of the Willamette Valley Wine Region
Round 1
Reviewer 1 Report
This work proposed a new method to select subsets of climate models for Viticulture Climate Classifications projection.
Comments
#1 My main concern is with the robustness and significance of the proposed method. The authors didn’t demonstrate the superiority of their method in relation to the other similar methods already well-established/widely used in the literature.
# 2 What is the problem with inter-model dependency in multi-model ensemble projection and uncertainty assessment? This must be discussed thoroughly.
#3 L98-100: First discuss the sensitivity of grape production to these climate indices.
#4 The authors selected the subset of better-performing models in the current climate. But the model performing better in the current climate may not necessarily work better in the future climate (different climate). How did the authors address this issue? The models’ convergence in future climate should also be considered.
Raftery, A. E., Gneiting, T., Balabdaoui, F., & Polakowski, M. (2005). Using Bayesian model averaging to calibrate forecast ensembles. Monthly weather review, 133(5), 1155-1174.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
The paper has taken a relevant topic that has gathered recent interest and applied it to a specific case study. The strategies considered are not novel, but their application in the viticulture context is new and can be useful for the specific region. As the authors acknowledge, results are very location- and variable- specific. In fact, that is the motivation for this type of work. So it is unclear how the results will translate to other contexts outside of this one geographic extent and viticulture suitability context. It would be good to elaborate on this in the conclusions/discussion section.
A few other lingering questions.
- Would the historical skill translate to future climate projections as well? If not, what are the implications? Can you comment on this aspect in the discussions?
- Yes, ensemble selection results in a difference. However, is the difference big enough to make a difference on the ultimate decision? For example, is the difference between using all models and weighting them equally versus finding the optimal ensemble weights large enough that it will impact cultivar climate classification?
Addressing these will strengthen the discussion.
Some parts of the paper are hard to parse through with wordy sentences. I had to read multiple times. It would be good to try and simplify the writing where possible.
As noted below, I was unable to find evidence for some sentences in the results section in the figures. Would be good to check and edit those aspects.
Some specific comments are noted below.
Line 50. Too many references are lumped together. Each phrase in the paragraph should be linked to the appropriate reference. There are other instances as well. Please check throughout.
Line 89: 2 power n includes an empty set. So perhaps you should minus 1 from the count?
Line 154: Is it necessary to include the term CMIP3. It creates some confusion given you ue CMIP5
Line 174 to 184: It is unclear how the dryness index is calculated. Where do the soil moisture, potential transpiration and soil evaporation data come from?
Also why is DI expressed as a function of DI in equation 6. Also GMCC is equation 7. Similarly in quotations 1 through 4 as well. This representation of X = X() is unclear. I think you are yrying to say X is function of a few variables or X=f(.. , ..) So, for example, DI = f(precip, Tmax and Tmin)
Line 187 to 190 I had to read multiple times to figure out what the sentence is trying to convey. For example rather than use DI, this work uses precip, T-max and T-min as DI is a function of these variables. This section could benefit from some rewriting for clarity.
Line 222: Do you mean sections 2.4.2 – 2.4.5 ?
Line 351 to 353: What is basis for this sentence? Can you reference figures. I don’t see this result from the figures.
Line 402: I don’t understand this result in relation to Figure 5. How is this skill weighting different from what is discussed earlier in the paragraph.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
The paper is really interesting and useful for future applications and consequently worth for publication.
There is one point that should be addressed by the authors: Is the daily LOCA CMIP5 datasets downscaling based on the 553 grid points of the Livneh dataset? If the answer is yes, are the performances of the comparison between observed and simulated data influenced by this relationship?
Additionaly, Just other two notes:
1) Materials and Methods - Study Area: Please provide information about viticultural production of the area
2) Table 2 - Wrong caption (from figure 1)
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 1 Report
#1 (Comment #4 in my previous review)
I suggest authors briefly discuss the issue that the optical subset of models in the current climate might not work better in future climates. Several recent studies account for the models' convergence in future climate in making ensemble averaging/projection. See the following papers (I referred to wrong paper in my previous review. My apologies for any inconveniences)
Giorgi, F., & Mearns, L. O. (2003). Probability of regional climate change based on the Reliability Ensemble Averaging (REA) method. Geophysical research letters, 30(12).
Aryal, Y., & Zhu, J. (2020). Multimodel ensemble projection of meteorological drought scenarios and connection with climate based on spectral analysis. International Journal of Climatology, 40(7), 3360-3379.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
