Modeling Sugar Beet Responses to Irrigation with AquaCrop for Optimizing Water Allocation
Round 1
Reviewer 1 Report
REVIEW
The manuscript “Modeling sugar beet responses to irrigation with AquaCrop for optimizing water allocation” by Garcia-Vila and collaborators presents computational work. Data about water stress conditions were used to calibrate and validate AquaCrop, a statistical model used for determination of parameters affecting sugar beet growth. Authors collected experimental data and applied to the available AquaCrop to contextualize different experimental measurements, evaluating the growth of sugar beet. Overall, the study will be useful for a broad experimental audience; however the comparison with the previously reported model is needed as well as the proof of concept that experimental data improve simulations. The manuscript is novel and explores new constraint strategies for the sugar beet; although the manuscript needs to be extensible proofread for English.
Critical issues
The introductory paragraph should be redrafted, avoiding vague sentences. For example, in line 58 the sentence “Simple models such as the one presented by [8], based on [9] approach, or PLANTGRO, a simple water balance model [5] have been proposed”. Since this manuscript presents modeling work, authors should give enough background describing previous types of models. Additionally, description of the state-of-the-art of parameters determination should be discussed for the previous AquaCrop model with special attention on beets. Authors should make clear their contribution in version 6.1 and make data available for future improvement.
Authors should add a comparison of the model predictions of AquaCrop original version and AquaCrop v6 in Figure 1 and 2 as well as in the supplementary information.
Authors should include statistical tests to determine significant differences among treatments in Fig. 5, 6, 7.
Overall, the manuscript provides new insights into the field. Author should provide the model in a general format in order to be able to check for the structure of the model.
Additionally, authors should perform a robustness analysis evaluating differences in green canopy cover, dry biomass, and root dry yield, while varying each parameter used into the model. I found some issues in the manuscript that can be addressed for a better interpretation of the goals of this manuscript including a comparison with the previous version of AquaCrop.
Minor general issues
Line 11. Replace “Simulation” by Statistical
Line 11. Please briefly introduce properties of AquaCrop.
Line 16. Delete “subsequently” and “and”
Line 16. Rephrase line 16-17.
Line 18. Replace “over a wide range of values was very good” by was accurate.
Line 32. Mention the production range by ten world's largest countries.
Line 42. Provide a worldwide parameter or define Mediterranean.
Line 45. Complete the sentence. Sugar beet is sensitive to water limitation of XXXX.
Line 52. Change “Crop responses to different water availability scenarios are complex, and” by The response of crop to water availability is complex, but
Line 53. Add citation.
Line 58. Describe each model briefly and their main differences.
Line 113. Change to: To validate the model eight commercial farms part of AIMCRA were selected.
Line 355. Combine Figure 8 and 9.
Line 369. Delete “In the view of the calibration and validation results (Figures 1-4)”
Line 369. AquaCrop v6?
Line 390. Indicate the name of the performed test, number of replicates, and p-values.
Author Response
Please see the attachment
Author Response File: 
Author Response.docx
Reviewer 2 Report
I enjoyed reading this comprehensive study on AquaCrop model.
My main concern is about calibration strategy which could improve Figure 1d.
Table 2: The parameter values seem to vary very little for different studies. I assume that the study areas are different and parameter regionalization should be important in a good model. Different climate, different soil type, aspect, hydrologic processes etc should all have effect on the best parameter value.
Line 133: Apparently a manual calibration was done. Automatic calibration using different tools Pest (Doherty 2005), multi-objective functions (one focusing on Figure 1d fit only) and global search algorithms like SCE-UA and CMAES could definitely improve the results as compared to the one at a time local (manual) calibration.
Table 1: only one experimental setup in different conditions like winter sowing date and plant density for calibration seem to be less as compared to the validation data. Remote sensing data (MODIS-AET, Soil moisture etc) from the fields could be incorporated.
Line 59-62: DSSAT model can be added to the model list?
Author Response
Please see the attachment
Author Response File: Author Response.docx
Round 2
Reviewer 1 Report
Authors addressed most of the reviewer comments, however there is still a raising point about the statistical test.
The comment previously said "Authors should include statistical tests to determine significant differences among treatments in Fig. 5, 6, 7.", since author already have the information to generate box plot, a t-test analysis results in a natural test to compare treatments. This information will provide a quantitative platform to define what is significantly different or not. Authors should cite the type of statistical test used, the number of replicates and the p-value.
Other than this point, I agree with the content of the current version of the manuscript and I recommend it for publication.
Author Response
Comment: The comment previously said "Authors should include statistical tests to determine significant differences among treatments in Fig. 5, 6, 7.", since author already have the information to generate box plot, a t-test analysis results in a natural test to compare treatments. This information will provide a quantitative platform to define what is significantly different or not. Authors should cite the type of statistical test used, the number of replicates and the p-value.
Response: In the updated manuscript version, the type of statistical test and the p-value have been indicated in lines 303, 324, 467 and 472. The number of replicates (30 years) were already indicated in the previous versions of the manuscript.
Many thanks for your comments.
Reviewer 2 Report
Thanks for the revised version of the manuscript.
Author Response
Thank you for the revision.
