Comparison of Methods for Filling Daily and Monthly Rainfall Missing Data: Statistical Models or Imputation of Satellite Retrievals?
, Klebber Teodomiro Martins Formiga 2 and Veber Afonso Figueiredo Costa 1
Round 1
Reviewer 1 Report
At the first, I would like to thank the Authors for their very hard work and interesting manuscript. I have read the presented work with pleasure. In my opinion, the manuscript has been prepared very well. In Introduction the Authors reviewed the literature. The study area is described in detail. The Methodology is proper. Obtained results are supported by discussion. I will be very happy to see published work in the “Water”.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
In this paper authors have tried to identify the suitable methods for the gap filling the rainfall over the midwestern region of Brazil. I think the work is important and useful. However, before the acceptance the authors must answer the following queries and revise their manuscript accordingly. 1. Introduce the whole map of Brazil and locate the study area within the country so that readers can easily familiar with the locations. So modify the figure 1. 2. Water year 2016-17, any reason why author choose particularly this year? 3. 6 test gauging station for daily data and 4 test gauging station for monthly data why only limited stations, could you clarify it in the manuscript? 4. The performance index in Table 3 for Monti vidiu station, why the RMSE: GPM 49.28 is not in bold. What is the criteria to be the bold for RMSE? 5. Figure 5 is repeated two times, I guess the last figure should be figure 10? 6. In discussion part it would be good if authors could add few more lines on comparisons with the previous studies and findings of their study?Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Reviewer 3 Report
In the present paper, the authors have investigated the faisability of several approaches for filling missing values of the daily and monthly rainfall data. While I find the paper intersting and in overall easy to read and to understand what the authors have tried to demonstrate for the readers, I have a comments to the authors which need to be clearly addressed and clarified.
1. Please provide a clear description of the overall objectives of your paper in comparison to the literature review, what is the novelty of your paper?
2. Section results and discussion is out of scope and unacceptable. More details, more analysis, in depth comparison, and the raisons of the difference between the approaches should be clearly highlighted.
3. Regarding the results reported in Table 2, two majors issues need to be clearly addressed by the authors. First, it is clear that all percent bias values are negative except two, which leads to conclude that the proposed algorithms for filling missing values provides underestimation bias, this statement should be clearly justified by the authors and a comparison between what is reported in the literature can help in better understanding this result. The second issue is major, and risks calling into question the obtained results in the present study, why? It is clear from the obtained results reported in Table 2 that, all CC are approximately below 0.60 (R2=0.36) and the maximum is 0.72 (R2=0.49), which are very low values. These results are in contradiction with the statement reported in the abstract [at the daily scale, the satellite retrievals, albeit biased, provided more accurate estimates, with better represen-20 stations of the temporal dynamics of the dry and wet states and of the extreme events of the observed rainfall in most testing sites]. My comment is that this statement might be jeopardized, and at least clarified and reformulated.
4. The proposed algorithms worked more accurately at monthly time scale compared to the daily scale, and this is obvious for a single series as in this study, but when a long series this is no longer obvious. The authors should understand that, the fluctuation at monthly time scale is less than at daily time scale, and this should be clearly highlighted by the authors.
5. The quality of figures should be improved.
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 3 Report
The paper was improved and the authors correctly did the necessary revision, it can now be accepted.
