Evaluation of Five Satellite-Based Precipitation Products for Extreme Rainfall Estimations over the Qinghai-Tibet Plateau

Comments: The paper is methodologically well written, has very interesting content and seems to be of great practical importance. Very clearly and precisely written. I have almost no comments in PDF of manuscript.

Response: Thank you very much for taking the time to review this manuscript. Please find the detailed responses below and the corresponding revisions/corrections highlighted/in track changes in the re-submitted files.

1.       I suggest: "...to climate change...".

Response: Thank you for your valuable suggestion. We agree with this comment. Therefore, we have made the necessary change to replace "climate" with "climate change," which is a more accurate representation. See P2, L52 in the revised manuscript.

2.       I suggest: "...source..."

Response: Thank you for pointing this out. We agree with this comment.  We have accordingly made the revision by changing "reservoir" to "source". See, P2, L72 in the revised manuscript.

Comments: The authors evaluate the performance of five satellite-based precipitation products to estimate extreme rainfall over the Qinghai-Tibet Plateau. The presentation is clear, and the methods used to evaluate the retrievals performance appear to be adequate. In addition, the contents should be of interest for remote sensing readers interested in satellite-based precipitation products. My major comment is that the discussion of the POD, FAR, and CSI should describe more clearly the accuracy of the satellite products in estimating the daily precipitation values (Figs. 8,9,10). In my opinion, the performance of the retrievals is mediocre which indicate that the retrievals do not provide good estimations of the precipitation for particular days; the retrievals seem to capture the distribution. If this is correct, this should be stated clearly in the text and in the conclusions where it appears the statement that the retrievals capture the daily precipitation well. More specific comments are provided below these lines.

Response: Thank you very much for your valuable feedback and suggestions during the review process. Your professional insights and recommendations are crucial for improving the quality of our paper. As you suggestion, the descriptions on the performance of IMERG at daily scale have been revised, the specific revisions see responses to the questions of 4, 5, and 8. 

SPECIFIC COMMENTS

1.     first line on page 2 there is a letter "f" in the text that should be removed.

Response: We have removed the letter “f”, as your suggestion, see P2, L47 in the revised manuscript.

2.       Botton of page 5, replace "required for required" with "required"

Response: We have replaced "required for required" with "required" in your suggestion, see P5, L202 in the revised manuscript.

3.       Section 3.2.1. "IMERG also meets the criteria..." What is the criteria?

Response: The criteria refers to the CC value of SEPI should be higher than 0.4. The original sentence was revised as follows, see P11, L353~357 in the revised manuscript.

“Among all ten SEPIs, three SEPIs of IMERG products meets the criterion of having a CC value higher than 0.4. TRMM3B42 shows similar performance, with three SEPIs meeting the criterion of having a CC value higher than 0.4, although its CC values are lower than IMERG. PERSIANN only has one SEPI that meets the criterion of having a CC value higher than 0.4.”

4.       Section 3.4. Figure 8 shows similar POD and FAR for precipitation events (Figs. 8a,b) and larger FAR than POD for extreme precipitation (Figs 8d,e). Fig. 9 also shows mediocre performance of the retrievals. And the detections at the seasonal level, Fig. 10, also show in general mediocre performance of the retrievals in terms of POD and FAR. However, the description in the text does not mention the limited value of the retrievals at the daily scale.

Response: Thank you for pointing this out. We agree with this comment. We lack a comprehensive description of the retrievals at the daily scale. Therefore, we have supplemented and described the limited retrieval capability of satellite products for extreme precipitation at a daily scale. The specific revision are as follows.

Add the supplementary description to figure 8, see P17, L481~483.

“However, the similar POD and FAR for precipitation events (Figs. 8a,b) and larger FAR than POD for extreme precipitation (Figs 8d,e) also demonstrate that the advantage of IMERG is not significant enough at the daily scale.”

Add the supplementary description to figure 9, see P18, L502~505.

“It was also found that the PODs, FARs, and CISs for five SPPs in the QTP ranged from 0.263 to 0.383, from 0.689 to 0.781, and from 0.16 to 0.183, respectively. Obviously, the advantage of IMERG is slightly significant compared with other SPPs at the daily scale.”

Add the supplementary description to figure10, see P18, L523~526.

“Noted also that except for winter, the POD of IMERG is slight lower than PERSIANN, while CIS of IMERG is significantly higher than other SPPs. It may be related to overestimate in PERSIANN.”

5.       Conclusions. It is stated that IMERG has exceptional performance but at the view of Figs 8, 9, and 10 this may be questionable.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made additional revisions to the corresponding conclusion section in light of these findings. The original sentence and the modified sentence are as follows, see P23, L727~731:

The original sentence:

“The accuracy assessment of extreme precipitation occurrence reveals that IMERG demonstrates a robust capability in accurately detecting both daily rainfall events and extreme rainfall events (defined as rainfall exceeding the 75th percentile of rainy days) in the QTP. On the contrary, the other four SPPs exhibit weak performance and consistently overestimate both daily rainfall events and extreme precipitation events.”

The modified sentence:

“The accuracy assessment of extreme precipitation occurrence reveals that IMERG demonstrates a slightly advantage compared with other SPPs in detecting both daily rainfall events and extreme rainfall events (defined as rainfall exceeding the 75th percentile of rainy days) in the QTP. Overall, the PODs, FARs, and CISs for five SPPs in the QTP ranged from 0.263 to 0.383, from 0.689 to 0.781, and from 0.16 to 0.183, respectively.”

6.       Conclusions. It is mentioned that the retrievals are corrected by ground observations. If the ground observations used in this evaluation were used in the creation of the retrievals the evaluation presented is of limited value since the observations used to evaluate the performance should not be used in the generation of the retrievals. Please clarify.

Response: Thank you for pointing this out. We apologize for any confusion caused by our previous statement and will provide a clarification on this matter.

In fact, some retrieval products (e.g., CMORPH, IMERG, PERSIANN, TRMM-3B42V7 in this study)have merged some ground data to ensure the accuracies of retrieval data, which is different from the original or near real-time retrieval data(e.g, TRMM-3B42RT). As shown in Figs2-4 and Table 5, the performance of the TRMM-3B42RT is worse than the merged TRMM-3B42V7.

In addition, IMRGE merged the global monthly GPCC (Global Precipitation Climatology Centre) ground dataset from WMO(World Meteorological Organization), which has more than 700 stations in China. While, our ground evaluated data are from China Meteorological Administration, which includes more than 2400 stations in China at daily scale (104 stations in QTP after our quality control based on the criteria of continuity, uniformity, and consistency in time). In other words, both the number of stations and the temporal resolution are different for our ground observation and the ground observation for calibrating the IMERG-Final products. Certainly, the number of the ground observations used in this evaluation is far more than that of IMERG, and thus the evaluation results are more reliable.

7.       Page 20, KGE can not be larger than 1.

Response: Yes, KGE can not be larger than 1. I am sorry for the misunderstanding. In fact, the phrase "KGE reduced by 0.48-3.15" in this context indicates a substantial reduction in KGE, ranging from 0.48 to 3.15 (decreasing from -0.53 to -3.68). This wording can indeed be misleading, and we acknowledge the need for revision to enhance clarity. We revised the sentence as follows, see P20, L578~584.

“In terms of precipitation-frequency-based indices, PERSIANN, TRMM3B42 and TRMM3B42RT for higher rainfall intensities (>20 mm d^-1) tended to overestimate rainfall. Additionally, their KGE values at higher rainfall intensities (-0.01, 0.18, and -3.68, respectively) are notably lower than the KGE values at lower rainfall intensities (0.47, 0.68, and -0.53, respectively). Conversely, IMERG and CMORPH had higher consistency in the frequency distributions of precipitation with the site, consistently yielding KGE values above 0.46 for both rainfall intensity levels.”

8.       Conclusions. The use of the retrievals to monitor the events does not seem possible for individual days given the large FAR in Figs. 8, 9, 11. I think this conclusion should be modified.

Response: Thank you for pointing this out. We agree with this comment. Therefore, we have made additional revisions to the corresponding conclusion section in light of these findings. The revisions see the response to question 5.

Comments 1: The manuscript of the authors from China is concerned with " Evaluation of Five Satellite-Based Precipitation Products for Extreme Rainfall Estimations over the Qinghai-Tibet Plateau”.

Thank you for the opportunity to review this useful and actual research study.

The manuscript presents an evaluation of the effectiveness of five satellite precipitation products - SPPs in identifying variations in the occurrence and distribution of intense precipitation occurrences across the chosen place during the period from 2001 to 2015.

This submitted article could be with the aim and scope of the MDPI journal Remote Sensing.

• Abstract & introduction: These two parts are focused on the paper's main aim and the new contributions of authors to the state of the art. The abstract with keywords very effectively summarizes the manuscript.

The key objective for the authors is to identify the more suitable satellite precipitation products for the long-term surveillance of intense rainfall occurrences on the Qinghai-Tibet Plateau.

• Materials & methods: Based on the existing research, the authors used the NCAR Command Language software to perform bilinear interpolating gridded precipitation data from the SPPs to the station scale for point-to-point comparison.

It is clear how all of the data were obtained. This section gives readers enough information so that they can use the study for other areas. 

• Results & discussion: The data are well-controlled and robust and results are well-presented with relevant and current tables, figures, and references. A reference dataset was generated using rainfall measurements collected from 104 rainfall stations distributed across the QTP. Ten standard extreme precipitation indices were the evaluation's main focus, encompassing parameters such as precipitation duration, amount, frequency, and intensity.

Results section were obtained and the methods used to analyze the data are progressive and scientific sound. With some adaptation, this described methodology could be used more generally.

I hope that the authors will continue with such interesting and inspiring research.

I have no comments on improving the manuscript before publishing.

Response: Thank you for reviewing our manuscript. We appreciate your positive feedback and valuable comments on our research study. We are glad to hear that you found our manuscript useful and relevant. We sincerely appreciate your positive feedback on this article, encompassing the affirmation provided for the abstract, introduction, materials, methods, results, and discussion sections. Your support is highly valued and deeply acknowledged, and we will continue with such research.

Comments: Comments on “Evaluation of Five Satellite Precipitation Products for Extreme Rainfall Estimations over the Qinghai-Tibet Plateau” by Zhang et al.

General comments: To provide a comprehensive assessment regarding the performance of SPPs over the QTP, this study evaluate the effectiveness of 5 SPPs, including CMORPH, IMERG-Final, PERSIANN-CDR, TRMM-3B42V7, and TRMM-3B42RT, in identifying variations in the occurrence and distribution of intense precipitation occurrences across the QTP during the period from 2001 to 2015. In short, the logic is clear and the manuscript is also well written. I have only two major comments as follows:

Response: Thank you very much for your recognition of the manuscript and your valuable feedback and suggestions during the review process. Your professional insights and recommendations are crucial for improving the quality of our paper. In response to your suggestions, we explain the reasons for the best performance of IMERG-Final products in the discussion section and compare the results with previous studies in other regions.

SPECIFIC COMMENTS

1.       My biggest concern is on the reasons for the best performance of IMERG-Final. I fully agree with the authors that IMERG-Final exhibit highest skills in detecting extreme precipitation events on annual, seasonal, and daily scales. However, the authors should find the possible reasons from the difference between IMERG-Final and the other products.

Response: We have consulted the literature to find possible reasons for the differences between IMERG-Final and other products and have added to them. The additions are listed below, see P21, L625~635 in the revised manuscript:

“The reasons for the best performance of IMERG-Final may be as follows: the first reason is that the IMERG product has higher spatial and temporal resolution. The second reason is that the other products are TRMM program and its derived precipitation data, while IMERG belongs to the new generation of Global Precipitation Measurement (GPM) program, which inherits the successful experience and results of TRMM, further improves the precipitation inversion algorithm, and at the same time, utilizes the more advanced passive GPM Microwave Imager (GMI) and Dual-frequency Precipitation Radar (DPR) sensor to improve the spatial coverage of TRMM in combination with the improved TRMM correction algorithm[33,34]. IMERG-Final, as the IMERG product with monthly scale correction from global ground stations, is the closest in accuracy to the actual precipitation data.”

2.       I would like to know the performance of the 5 SPPs over the other regions. The authors should compare their results with the previous studies on this issue over the other regions, and explain whether IMERG-Final perform better in the most regions or just QTP.

Response: We compare the results of this paper with previous studies conducted in other regions, and found that some studies also proved that IMERG-Final performs better. At present, our results can only prove that IMERG-Final performs better in the QTP region, and whether it performs better in the majority of the regions needs to be explored, and we will make more efforts to explore this aspect in our future studies. The additions are listed below, see P19, L540~549 in the revised manuscript:

“Several scholars have already conducted assessments on the accuracy of different SPPs in other regions. For instance, Wang et al. [59] assessed the performance of TRMM 3B42, CMORPH, PERSIANN, and CHIRPS in the Ganjiang River Basin. Their findings indicated that TRMM 3B42 outperformed CMORPH and PERSIANN. Fang et al. [26] and Tang et al. [60] evaluated the performance of IMERG and TRMM-3B42 in mainland China, highlighting that GPM IMERG demonstrated better detection capabilities for extreme precipitation events and daily scales compared to TRMM-3B42. These findings align with the findings of this study in the QTP region. However, further research is still needed to investigate whether IMERG-Final performs better in most regions. ”

59.  Wang, Q.; Xia, J.; She, D.; Zhang, X.; Liu, J.; Zhang, Y. Assessment of four latest long-term satellite-based precipitation products in capturing the extreme precipitation and streamflow across a humid region of southern China. Atmospheric Research 2021, 257, 105554.