Estimation of Daily Potential Evapotranspiration in Real-Time from GK2A/AMI Data Using Artificial Neural Network for the Korean Peninsula
Round 1
Reviewer 1 Report
Estimation of Daily Potential Evapotranspiration in Real-Time from GK2A/AMI data using Artificial Neural Network for the 3 Korean Peninsula
In this manuscript, the authors estimate daily potential evapotranspiration (PET) in the Korean Peninsula using an Artificial Neural Network (ANN) and data from the GEO-KOMPSAT 2A satellite, which includes an Advanced Meteorological Imager (GK2A/AMI). Digital elevation, precipitation, and other meteorological data are used as input data for the ANN. The authors then compare daily PET derived from GK2A/AMI with in-situ ET (EC-ET) measured at NIFoS flux towers, as well as with MODIS data, and show that the proposed model offers more locally optimized PET estimates for the Korean Peninsula.
The approach is novel and offers the possibility to estimate real-time 1 km resolution daily ET when in-situ measurements do not allow for it. The manuscript is concise, and the methods and data used are robust. However, the text needs proofreading as a lot of grammatical and syntactic errors were detected. In the annotated manuscript I have noted some of them, but further attention is needed. In addition, the manuscript could be improved by enhancing the methodology and discussion sections. Therefore, a moderate revision is proposed.
More specifically, since the manuscript is relatively short, the authors could provide additional information on the ANN model for the readers who are not familiar with some of its aspects (i.e., batch normalization layer, exponential linear unit e.t.c.). At the very least, the authors should provide references. References are also missing from the “Data and Methods” section (mainly sub-section 2.1), where the remote sensing data are mentioned. In addition, a short review of other artificial intelligence methods used for the estimation of PET and a better justification of the choice of the specific AI method (ANN) would be useful (for example, (Adnan et al., 2017; Chia et al., 2020; Granata et al., 2020; Laqui et al., 2019; Park et al., 2018; Torres et al., 2011). It would also be helpful to include a table of abbreviations in the manuscript since you use so many of them.
In the abstract, as well as in the main body of the text, the authors mention that “The accuracy of the GK2A/AMI-derived PET, in comparison with the flux tower-measured ET, was consistent with findings in previous studies”. I guess that, by mentioning this, the authors refer to lines 377-396, where they provide possible explanations for the deviations of the results of their model from the flux tower-measured ET, most of which are due to the Penman-Monteith assumptions. However, this does not show consistency with previous studies. The authors should rephrase. To enhance the findings of their study, the authors should include a discussion section, where they would refer to other studies conducted in their study site (or similar study sites) for the estimation of PET (for example, (Birhanu et al., 2018; Lim et al., 2017)).
Other minor remarks:
Lines 19-20 and lines 150, 366, 465: “To examine the availability of the GK2A/AMI-derived PET data, we compared it with in-situ ET measured at flux towers, and with MODIS PET data.” I don’t think that “availability” is the right word. Do you mean “efficiency”?
Lines 285-289: here you introduce the permutation test in the results section. The description of the permutation test should be moved to the methods section.
Please refer to the annotated manuscript for other remarks concerning grammar and syntax.
References
Adnan, M., Ahsan, M., -, A.-R., Nazir, M., 2017. Estimating Evapotranspiration using Machine Learning Techniques. ijacsa 8. https://doi.org/10.14569/IJACSA.2017.080915
Birhanu, D., Kim, H., Jang, C., Park, S., 2018. Does the Complexity of Evapotranspiration and Hydrological Models Enhance Robustness? Sustainability 10, 2837. https://doi.org/10.3390/su10082837
Chia, M.Y., Huang, Y.F., Koo, C.H., Fung, K.F., 2020. Recent Advances in Evapotranspiration Estimation Using Artificial Intelligence Approaches with a Focus on Hybridization Techniques—A Review. Agronomy 10, 101. https://doi.org/10.3390/agronomy10010101
Granata, F., Gargano, R., de Marinis, G., 2020. Artificial intelligence based approaches to evaluate actual evapotranspiration in wetlands. Science of The Total Environment 703, 135653. https://doi.org/10.1016/j.scitotenv.2019.135653
Laqui, W., Zubieta, R., Rau, P., Mejía, A., Lavado, W., Ingol, E., 2019. Can artificial neural networks estimate potential evapotranspiration in Peruvian highlands? Model. Earth Syst. Environ. 5, 1911–1924. https://doi.org/10.1007/s40808-019-00647-2
Lim, C.-H., Kim, S.H., Choi, Y., Kafatos, M.C., Lee, W.-K., 2017. Estimation of the Virtual Water Content of Main Crops on the Korean Peninsula Using Multiple Regional Climate Models and Evapotranspiration Methods. Sustainability 9, 1172. https://doi.org/10.3390/su9071172
Park, N.Y., Sohn, E.H., Jang, J.D., 2018. Estimation of satellite-based daily evapotranspiration using deep-learning approahces 2018, H33A-07.
Torres, A.F., Walker, W.R., McKee, M., 2011. Forecasting daily potential evapotranspiration using machine learning and limited climatic data. Agricultural Water Management 98, 553–562. https://doi.org/10.1016/j.agwat.2010.10.012
Comments for author File: 
Comments.pdf
Author Response
========== Reviewer 1 ====================== Reviewer(s) Comments: Reviewer: 1 Comments to the Author In this manuscript, the authors estimate daily potential evapotranspiration (PET) in the Korean Peninsula using an Artificial Neural Network (ANN) and data from the GEO-KOMPSAT 2A satellite, which includes an Advanced Meteorological Imager (GK2A/AMI). Digital elevation, precipitation, and other meteorological data are used as input data for the ANN. The authors then compare daily PET derived from GK2A/AMI with in-situ ET (EC-ET) measured at NIFoS flux towers, as well as with MODIS data, and show that the proposed model offers more locally optimized PET estimates for the Korean Peninsula. 1. [Grammatical and syntactic errors] The approach is novel and offers the possibility to estimate real-time 1 km resolution daily ET when in-situ measurements do not allow for it. The manuscript is concise, and the methods and data used are robust. However, the text needs proofreading as a lot of grammatical and syntactic errors were detected. In the annotated manuscript I have noted some of them, but further attention is needed. In addition, the manuscript could be improved by enhancing the methodology and discussion sections. Therefore, a moderate revision is proposed. 2. [More information] More specifically, since the manuscript is relatively short, the authors could provide additional information on the ANN model for the readers who are not familiar with some of its aspects (i.e., batch normalization layer, exponential linear unit e.t.c.). At the very least, the authors should provide references. References are also missing from the “Data and Methods” section (mainly sub-section 2.1), where the remote sensing data are mentioned. In addition, a short review of other artificial intelligence methods used for the estimation of PET and a better justification of the choice of the specific AI method (ANN) would be useful (for example, (Adnan et al., 2017; Chia et al., 2020; Granata et al., 2020; Laqui et al., 2019; Park et al., 2018; Torres et al., 2011). It would also be helpful to include a table of abbreviations in the manuscript since you use so many of them. 3. [Discussions] In the abstract, as well as in the main body of the text, the authors mention that “The accuracy of the GK2A/AMI-derived PET, in comparison with the flux tower-measured ET, was consistent with findings in previous studies”. I guess that, by mentioning this, the authors refer to lines 377-396, where they provide possible explanations for the deviations of the results of their model from the flux tower-measured ET, most of which are due to the Penman-Monteith assumptions. However, this does not show consistency with previous studies. The authors should rephrase. To enhance the findings of their study, the authors should include a discussion section, where they would refer to other studies conducted in their study site (or similar study sites) for the estimation of PET (for example, (Birhanu et al., 2018; Lim et al., 2017)). 4. [Minor Revisions] Other minor remarks: Lines 19-20 and lines 150, 366, 465: “To examine the availability of the GK2A/AMI-derived PET data, we compared it with in-situ ET measured at flux towers, and with MODIS PET data.” I don’t think that “availability” is the right word. Do you mean “efficiency”? Lines 285-289: here you introduce the permutation test in the results section. The description of the permutation test should be moved to the methods section. Please refer to the annotated manuscript for other remarks concerning grammar and syntax. References Adnan, M., Ahsan, M., -, A.-R., Nazir, M., 2017. Estimating Evapotranspiration using Machine Learning Techniques. ijacsa 8. https://doi.org/10.14569/IJACSA.2017.080915 Birhanu, D., Kim, H., Jang, C., Park, S., 2018. Does the Complexity of Evapotranspiration and Hydrological Models Enhance Robustness? Sustainability 10, 2837. https://doi.org/10.3390/su10082837 Chia, M.Y., Huang, Y.F., Koo, C.H., Fung, K.F., 2020. Recent Advances in Evapotranspiration Estimation Using Artificial Intelligence Approaches with a Focus on Hybridization Techniques—A Review. Agronomy 10, 101. https://doi.org/10.3390/agronomy10010101 Granata, F., Gargano, R., de Marinis, G., 2020. Artificial intelligence based approaches to evaluate actual evapotranspiration in wetlands. Science of The Total Environment 703, 135653. https://doi.org/10.1016/j.scitotenv.2019.135653 Laqui, W., Zubieta, R., Rau, P., Mejía, A., Lavado, W., Ingol, E., 2019. Can artificial neural networks estimate potential evapotranspiration in Peruvian highlands? Model. Earth Syst. Environ. 5, 1911–1924. https://doi.org/10.1007/s40808-019-00647-2 Lim, C.-H., Kim, S.H., Choi, Y., Kafatos, M.C., Lee, W.-K., 2017. Estimation of the Virtual Water Content of Main Crops on the Korean Peninsula Using Multiple Regional Climate Models and Evapotranspiration Methods. Sustainability 9, 1172. https://doi.org/10.3390/su9071172 Park, N.Y., Sohn, E.H., Jang, J.D., 2018. Estimation of satellite-based daily evapotranspiration using deep-learning approahces 2018, H33A-07. Torres, A.F., Walker, W.R., McKee, M., 2011. Forecasting daily potential evapotranspiration using machine learning and limited climatic data. Agricultural Water Management 98, 553–562. https://doi.org/10.1016/j.agwat.2010.10.012 [Line #] (e.g. L15 means the location of 15 Line) R1. [Grammatical and syntactic errors] We revised the manuscript and figures following your suggestion. We also used an English language editing service to proofread the manuscript. R2. [More information] We have added further details for the readers who are not familiar with remote sensing data and the ANN model. [Section of Data and Methods, L134-144 & L148-156 & L263-283] In addition, we added a short review and appropriate references on other artificial intelligence methods. [Section of Introduction, L111-120] We have added abbreviations. [Section of Abbreviations, L575] R3. [Clarity of Discussions] We corrected the paragraphs on the comparison with flux tower-based ET. In addition, we added a Discussion dealing with other studies conducted in the study area. [Section of Discussion, L425-451 & L517-536] R4. [Minor Revisions] We have revised the manuscript accordingly. [Section of Data and Methods, L285-295] The authors greatly appreciate your valuable comments.
Author Response File: 
Author Response.docx
Reviewer 2 Report
The authors have done interesting work over the Korean peninsula which is needed to understand the hydrological cycle. It paper presents an ANN model that retrieves daily PET in real-time using GK2A/AMI-derived data for the mentioned study period.
However, there are some points described below that have to be considered before publication. The overall presentation in the Introduction section lacks synergy and exists in bits and pieces. Though authors have identified the research gaps the literature survey part can be more streamlined and while coming towards the problem statement.
I have a big concern in the Introduction, as the authors have missed providing detailed discussion on the important aspect of different classification of ET estimation methods. There is a vast literature on this I would like to suggest few lines following this which author should add is “The ETo estimation models available in the literature may be broadly classified as (1) fully physically-based combination models that account for mass and energy conservation principles; (2) semi-physically based models that deal with either mass or energy conservation; and (3) black-box models based on artificial neural networks, empirical relationships, and fuzzy and genetic algorithms”. I would recommend adding these recent references to add more scientific weight in their Introduction not only to the classification but also to the MODIS and standardisation processes.
Srivastava, A., Sahoo, B., Raghuwanshi, N. S., & Singh, R. (2017). Evaluation of variable-infiltration capacity model and MODIS-terra satellite-derived grid-scale evapotranspiration estimates in a River Basin with Tropical Monsoon-Type climatology. Journal of Irrigation and Drainage Engineering, 143(8), 04017028. https://doi.org/10.1061/(ASCE)IR.1943-4774.0001199
Elbeltagi, A., Kumari, N., Dharpure, J. K., Mokhtar, A., Alsafadi, K., Kumar, M., ... & Kuriqi, A. (2021). Prediction of combined terrestrial evapotranspiration index (CTEI) over large river basin based on machine learning approaches. Water, 13(4), 547.
Extensive English editing is required as there many problems with sentence restructuring, grammatical errors, punctuations. I suggest authors to consider the English editing a serious concern in this manuscript and with the help of native speaker they can improve this version of the manuscript adequately.
In the entire manuscript, the fundamental equations utilised in the ANN model has not been mentioned which should be included in the manuscript to understand the processes behind the model used.
Elaborate the standardization processes (section 2.4.3) with more clear explanation.
Authors have missed to include the drawbacks of other satellite products such as AVHRR, LANDSAT based indirect vegetation proxies. I encourage and recommend the authors to incorporate this portion related to the use of satellite-based remote sensing products in the estimation of ET. Authors may like below find study in line of their statements to add the scientific weight in their observations.
In the statistical analysis section citations are missing except for IOA.
Authors should mention in the conclusion where these models/relationship applied need to be improved or any suggestions for future research in terms of models' performance improvement and application
Author Response
========== Reviewer 2 ======================
Reviewer(s) Comments:
Reviewer: 2
Comments to the Author
The authors have done interesting work over the Korean peninsula which is needed to understand the hydrological cycle. It paper presents an ANN model that retrieves daily PET in real-time using GK2A/AMI-derived data for the mentioned study period. However, there are some points described below that have to be considered before publication. The overall presentation in the Introduction section lacks synergy and exists in bits and pieces. Though authors have identified the research gaps the literature survey part can be more streamlined and while coming towards the problem statement.
1. [Introduction]
I have a big concern in the Introduction, as the authors have missed providing detailed discussion on the important aspect of different classification of ET estimation methods. There is a vast literature on this I would like to suggest few lines following this which author should add is “The ETo estimation models available in the literature may be broadly classified as (1) fully physically-based combination models that account for mass and energy conservation principles; (2) semi-physically based models that deal with either mass or energy conservation; and (3) black-box models based on artificial neural networks, empirical relationships, and fuzzy and genetic algorithms”. I would recommend adding these recent references to add more scientific weight in their Introduction not only to the classification but also to the MODIS and standardisation processes.
Srivastava, A., Sahoo, B., Raghuwanshi, N. S., & Singh, R. (2017). Evaluation of variable-infiltration capacity model and MODIS-terra satellite-derived grid-scale evapotranspiration estimates in a River Basin with Tropical Monsoon-Type climatology. Journal of Irrigation and Drainage Engineering, 143(8), 04017028. https://doi.org/10.1061/(ASCE)IR.1943-4774.0001199
Elbeltagi, A., Kumari, N., Dharpure, J. K., Mokhtar, A., Alsafadi, K., Kumar, M., ... & Kuriqi, A. (2021). Prediction of combined terrestrial evapotranspiration index (CTEI) over large river basin based on machine learning approaches. Water, 13(4), 547.
2. [Grammatical and syntactic errors]
Extensive English editing is required as there many problems with sentence restructuring, grammatical errors, punctuations. I suggest authors to consider the English editing a serious concern in this manuscript and with the help of native speaker they can improve this version of the manuscript adequately.
3. [More information of ANN model]
In the entire manuscript, the fundamental equations utilised in the ANN model has not been mentioned which should be included in the manuscript to understand the processes behind the model used.
4. [Clarity]
Elaborate the standardization processes (section 2.4.3) with more clear explanation.
5. [Limitation of remote sensing data]
Authors have missed to include the drawbacks of other satellite products such as AVHRR, LANDSAT based indirect vegetation proxies. I encourage and recommend the authors to incorporate this portion related to the use of satellite-based remote sensing products in the estimation of ET. Authors may like below find study in line of their statements to add the scientific weight in their observations.
6. [Reference of statistical analysis]
In the statistical analysis section citations are missing except for IOA.
7. [Further research]
Authors should mention in the conclusion where these models/relationship applied need to be improved or any suggestions for future research in terms of models' performance improvement and application
[Line #]
(e.g. L15 means the location of 15 Line)
R1. [Introduction]
We have revised the Introduction following your comments and have added information on the recent studies associated with ET and drought monitoring.
[Section of Introduction, L38-49 & L55-65]
R2. [Grammatical and syntactic errors]
We revised the manuscript and figures following your suggestion. We also used an English language editing service to proofread the manuscript.
R3. [More information of ANN model]
We have revised the text and have added additional information on the ANN model.
[Section of Data and Methods, L263-283]
R4. [Clarity]
We have revised the text to improve calrity.
[Section of Data and Methods, L240-254]
R5. [Limitation of remote sensing data]
We have revised the text accordingly.
[Section of Introduction, L73-84]
R6. [Reference of statistical analysis]
We have added the citations.
[Section of Data and Methods, L298-301]
R7. [Further research]
We have revised the text accordingly.
[Section of Conclusions, L553-567]
We greatly appreciate your valuable comments.
Author Response File: Author Response.docx
Round 2
Reviewer 2 Report
The authors have addressed all previous concerns expressed by the reviewers and in the process have improved the work, confirmed the validity of their findings and gained confidence in their introduction, methods, results and conclusions. I would like to congratulate the authors for an interesting and well executed work and I recommend this manuscript for publication in Hydrology (MDPI) in its current form.
