the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Jan 2021
Evaluating Spatiotemporal Patterns and Trends of Drought in Japan Associated with Global Climatic Drivers
Abstract. Drought disasters, such as water scarcity and wildfires, are serious natural disasters in Japan that are also affected by climate change. However, as drought generally has widespread impacts and the duration of drought can vary considerably, it is difficult to assess the spatiotemporal characteristics and the climatic causes of drought. Therefore, to identify the drought homogeneous regions and understand climatic causes of regional drought over Japan, this study provides a spatiotemporal analysis for historical droughts patterns and teleconnections associated with global climatic drivers. The trends of meteorological elements, which are the basis of drought index calculation, was first assessed. Then, drought characterized by the Self-calibrating Palmer Drought Severity Index (scPDSI) was investigated. Trends and patterns of drought were identified through the trend-free pre-whitening Mann-Kendall test and distinct empirical orthogonal function. The continuous wavelet transform and cross wavelet transform together with wavelet coherence were utilized to depict the links between drought and global climatic drivers. The results are described as follows: (1) the trends of precipitation were insignificant. However, temperature and potential evapotranspiration increasing trends were detected over Japan; (2) the drought trend over Japan varied seasonally, increasing in spring and summer and decreasing in autumn and winter; (3) two major subregions of drought variability—the western Japan (W region) and most of the northernmost Japan near the Pacific (N region) were identified; (4) wildfires with large burned area were more likely to occur when the scPDSI was less than −1; and (5) the North Atlantic Index (NAOI) showed the strongest coherence connections with Distinguished Principle Components-1 among four climatic drivers. Additionally, Distinguished Principle Components-2 showed stronger coherence connections with NAOI and Arctic Oscillation Index. This study is the first to identify homogeneous regions with distinct drought characteristics over Japan and connect the drought in Japan with the global climatic drivers.
This preprint has been withdrawn.
-
Withdrawal notice
This preprint has been withdrawn.
-
Preprint
(1868 KB)
Interactive discussion
Status: closed
-
CC1: 'Comment on nhess-2020-416', Francesco Serinaldi, 06 Jan 2021
The Authors may be interested in the paper below, showing that TFPW-MK test used in this study suffers from mathmatical inconsistencies that affect its effective significance, and make results and their interpretation generally incorrect.
Serinaldi F, Kilsby CG. . Stochastic Environmental Research and Risk Assessment 2016, 30(2), 763-777.
The Authors may also be intersted in the discussion reported in
Serinaldi F, Kilsby CG, Lombardo F. . Advances in Water Resources 2018, 111, 132-155.
and references therein.
Sincerely,
Francesco Serinaldi
Citation: https://doi.org/10.5194/nhess-2020-416-CC1 -
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
Dear Francesco Serinaldi
Thank you so much for your comments. The choice of method is really important to the trend analysis results. We will reconsider TFPW-MK test and use other versions of MK test to compare results in the revised manuscript. The effective significance of the trend analysis results will also be discussed in the revised manuscript.
Best regards,
Ke Shi
Citation: https://doi.org/10.5194/nhess-2020-416-AC1 -
AC2: 'Reply on RC1', Ke Shi, 22 Jan 2021
We greatly appreciate you for your constructive comments and suggestions.
First of all, please allow us to make a brief reply. When the open preprint is over (until 17 Feb 2021), we will make a detailed reply and upload the corresponding revised manuscript.
Point 1: L49-59: While some relevant works about Japan drought have been cited, I feel the literature review appears weak. I recommend the authors to include all studies focusing on Japan drought. Discuss your results and compare them with the results from the previous studies in the appropriate places (results/discussion sections).
Response 1: Thank you so much for your comment. We will continue to look for papers about drought in Japan as much as possible to include all current papers related to drought in Japan in our introduction.
Point 2: L116: The study area of this paper is small, therefore, is the 0.5° resolution meet the purpose of this research? The author should explain that.
Response 2: Thanks for your comment. In fact, a resolution of 0.5° resolution is indeed challenging in identifying drought conditions in some specific areas. But the main purpose of this paper is to identify the drought homogeneous regions of Japan through distinct empirical orthogonal function (DEOF). And the change in resolution does not have much effect on the DEOF result. So, 0.5° resolution can meet the needs of this research.
Point 3~6:
L118: Add space before ‘The’.
L163: I wasn't able to access the details of the data following the link. It appears that there may be a typo in the link, and potentially an access issue.
L169: Check the reference.
L167, L201, L218: Please check the font size.
Response 3~6: Thanks for your comments. Sorry for these mistakes, we will correct them in the revised manuscript.
Point 7, Point 9:
L313-314: This part is interesting. The author should try to further explain the physical mechanism about the impact of spatial variation of rainfall on drought.
L330-339: Expand the explanation for physical interpretation of the DEOF1 and DEOF2. If needed, make a reference to other variables.
Response 7, 9: Thanks for your comment. We will add some physical mechanism descriptions in these parts.
Point 8, 10~12:
L321: Replace ‘Spatial and temporal’ with ‘Spatiotemporal’. It should be consistent with the rest of the document.
L348 Figure 8: Show the W and N regions in your figure.
L393 Figure 12: Replace ‘In’ with ‘in’.
L526 Figure A4: There is a strange horizontal line under the title of Figure A4(b). Please check all figures.
Response 8, 10~12: Thanks for your comments. Sorry for these mistakes, we will correct them and check all the figures in the revised manuscript.
Citation: https://doi.org/10.5194/nhess-2020-416-AC2
-
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
-
RC1: 'Comment on nhess-2020-416', Anonymous Referee #1, 11 Jan 2021
This paper tried to explore the spatiotemporal characteristics of droughts over Japan and also linked the droughts with global circulation indices. Exploring the occurrences, development and underlying mechanisms of drought is the key to understanding them. From this point of view, this paper will be of great importance to Japan drought research.
Therefore, it is an interesting topic and suitable for Natural Hazards and Earth System Sciences. However, there are some issues that should be addressed by the author before publication. Especially lack of some physical mechanism discussion. Specific comments are as follows:
1. Introduction
L49-59: While some relevant works about Japan drought have been cited, I feel the literature review appears weak. I recommend the authors to include all studies focusing on Japan drought. Discuss your results and compare them with the results from the previous studies in the appropriate places (results/discussion sections).
2. Materials and Methodology
L116: The study area of this paper is small, therefore, is the 0.5° resolution meet the purpose of this research? The author should explain that.
L118: Add space before ‘The’.
L163: I wasn't able to access the details of the data following the link. It appears that there may be a typo in the link, and potentially an access issue.
L169: Check the reference.
L167, L201, L218: Please check the font size.
3. Results and Discussion
L313-314: This part is interesting. The author should try to further explain the physical mechanism about the impact of spatial variation of rainfall on drought.
L321: Replace ‘Spatial and temporal’ with ‘Spatiotemporal’. It should be consistent with the rest of the document.
L330-339: Expand the explanation for physical interpretation of the DEOF1 and DEOF2. If needed, make a reference to other variables.
L348 Figure 8: Show the W and N regions in your figure.
L393 Figure 12: Replace ‘In’ with ‘in’
Appendix
L526 Figure A4: There is a strange horizontal line under the title of Figure A4(b). Please check all figures.
Citation: https://doi.org/10.5194/nhess-2020-416-RC1 -
AC2: 'Reply on RC1', Ke Shi, 22 Jan 2021
We greatly appreciate you for your constructive comments and suggestions.
First of all, please allow us to make a brief reply. When the open preprint is over (until 17 Feb 2021), we will make a detailed reply and upload the corresponding revised manuscript.
Point 1: L49-59: While some relevant works about Japan drought have been cited, I feel the literature review appears weak. I recommend the authors to include all studies focusing on Japan drought. Discuss your results and compare them with the results from the previous studies in the appropriate places (results/discussion sections).
Response 1: Thank you so much for your comment. We will continue to look for papers about drought in Japan as much as possible to include all current papers related to drought in Japan in our introduction.
Point 2: L116: The study area of this paper is small, therefore, is the 0.5° resolution meet the purpose of this research? The author should explain that.
Response 2: Thanks for your comment. In fact, a resolution of 0.5° resolution is indeed challenging in identifying drought conditions in some specific areas. But the main purpose of this paper is to identify the drought homogeneous regions of Japan through distinct empirical orthogonal function (DEOF). And the change in resolution does not have much effect on the DEOF result. So, 0.5° resolution can meet the needs of this research.
Point 3~6:
L118: Add space before ‘The’.
L163: I wasn't able to access the details of the data following the link. It appears that there may be a typo in the link, and potentially an access issue.
L169: Check the reference.
L167, L201, L218: Please check the font size.
Response 3~6: Thanks for your comments. Sorry for these mistakes, we will correct them in the revised manuscript.
Point 7, Point 9:
L313-314: This part is interesting. The author should try to further explain the physical mechanism about the impact of spatial variation of rainfall on drought.
L330-339: Expand the explanation for physical interpretation of the DEOF1 and DEOF2. If needed, make a reference to other variables.
Response 7, 9: Thanks for your comment. We will add some physical mechanism descriptions in these parts.
Point 8, 10~12:
L321: Replace ‘Spatial and temporal’ with ‘Spatiotemporal’. It should be consistent with the rest of the document.
L348 Figure 8: Show the W and N regions in your figure.
L393 Figure 12: Replace ‘In’ with ‘in’.
L526 Figure A4: There is a strange horizontal line under the title of Figure A4(b). Please check all figures.
Response 8, 10~12: Thanks for your comments. Sorry for these mistakes, we will correct them and check all the figures in the revised manuscript.
Citation: https://doi.org/10.5194/nhess-2020-416-AC2 -
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
Dear Francesco Serinaldi
Thank you so much for your comments. The choice of method is really important to the trend analysis results. We will reconsider TFPW-MK test and use other versions of MK test to compare results in the revised manuscript. The effective significance of the trend analysis results will also be discussed in the revised manuscript.
Best regards,
Ke Shi
Citation: https://doi.org/10.5194/nhess-2020-416-AC1
-
AC2: 'Reply on RC1', Ke Shi, 22 Jan 2021
-
RC2: 'Comment on nhess-2020-416', Anonymous Referee #2, 21 Jan 2021
Review of: Evaluating Spatiotemporal Patterns and Trends of Drought in Japan Associated with Global Climatic Drivers.
This article analyses drought variability in Japan by means of global datasets and the PDSI. I think the manuscript needs to reinterpret and rework several issues and to change several issues affecting both formal and substance aspects.
35-36. This should be qualified in some way. In which regions? You note e.g. the different view by Sheffield et al. (2012) Nature, also based on the PDSI.
43-44. which events' Why these continents and not others?
44-46: Under rainfall deficits is not expected an increase of ET (as the soil moisture is constrained), then what has an effect is the increase of the atmospheric evaporative demand (this term is absolutely better to use than potential evapotranspiration, see e.g. https://onlinelibrary.wiley.com/doi/full/10.1002/wcc.632), which increases plant stress, leaf temperature (given changes in the partition of latent and sensible heat).
60-61: also thermodynamic forcing (e.g. radiative CO2 forcing increases VPD, but also land-atmopshere feedbacks: see https://nyaspubs.onlinelibrary.wiley.com/doi/10.1111/nyas.13912)
95-103: This is better to be moved to the methods.
116-117: I think the capability of this dataset must be tested for the use of the assessment of droughts at the regional scale. Some comparison with local stations/datasets would be desired.
122-144: This is not needed. There are several references reviewing PDSI. Authors should cite the pioneer study by W. palmer of 1965. It is not necessary to show the formulation of the FAO-56 Penman-Monteith reference evapotranspiration.
145-157: This is also not needed. Authors are using directly the PDSI data generated by Gerard Van der Schrier, so they may simply refer to this dataset and do not explain how this author generate it.
159-166: Why do the authors select these circulation indices and not others. Note that the effect of ENSO can be very different as a function of the selected index given the different physical mechanisms related (https://link.springer.com/article/10.1007/s00382-016-3082-y).
180-199: Not needed to describe the work of Yue et al. 2002.
What is the prupose of using the methods described in 2.3 and 2.4? The authors should justify its use in the context of the objectives of the study.
238: In figure 1 it is not necessary to include the trend in temperature. It is not a metric included in the analysis of drought by means of the PDSI. Please, include precipitation and atmospheric demand in absolute units using the same scale.
259-260: This is confuse. Gridded data from CRU is based on interpolation of station data. With this approach the authors suggest a re-interpolation based on mann-kendall results, but this should not be necessary as the input data is already a gridded dataset.
278: Averaging PDSI is not a suitable approach given spatial differences of autocorrelation characteristics in the PDSI (https://journals.ametsoc.org/view/journals/hydr/11/4/2010jhm1224_1.xml) so spatial comparability is poor. If authors are interested to show a time series over Japan, figure 4 showing surface area affected by drought (in %) is much better.
296-308. This figure is redundant. Given strong autocorrelation of the PDSI (https://www.sciencedirect.com/science/article/pii/S0022169414009305), it is not expected a difference in drought conditions at the seasonal level considering the PDSI.
Figure 7. I do not find this figure very logical as precipitation is declining more than PDSI. PDSI is low sensitive to the atmospheric demand as there is an alpha parameter used to obtain the cafec PET that limit the role of PET, but this role is always negative so I do not find logical that under increased PET, the precipitation may show a more declining trend that the PDSI.
353-358: Same comments related to the relevance of seasonal differences. If the authors want to analyse droughts at the seasonal scale, the PDSI is not the best choice. More useful alternatives are the SPEI or the SPDI (Ma et al., 2014, Hyd Procc). These indices provide exact seasonal information as the time scales are defined a priory and they are known.3.4. This section seems to be very disconnected to the rest of the study. The authors provide a very simple analysis to connect drought index and area affected by fire. In addition, there is not any critical discussion of the obtained results.
3.5. This section would gain clarity if in addition to the power spectrum analysis the authors would include further information in order to determine possible anomalies in the selected atmospheric circulation indices during drought conditions (in average for the whole Japan and for the two regions). A simple box-plot or some maps of PDSI anomalies corresponding to high/low phases of these circulation indices would be enough. With the current information it is really difficult to distill if drought events may be affected by these circulation drivers.Section 3 although it is named as "results and discussion" shows very limited critical discussion of the obtained results, but this is necessary in any serious scientific research.
Citation: https://doi.org/10.5194/nhess-2020-416-RC2 -
AC3: 'Reply on RC2', Ke Shi, 22 Jan 2021
We greatly appreciate you for your constructive comments and suggestions.
First of all, please allow us to make a brief reply. When the open preprint is over (until 17 Feb 2021), we will make a detailed reply and upload the corresponding revised manuscript.
Point 1: 35-36. This should be qualified in some way. In which regions? You note e.g. the different view by Sheffield et al. (2012) Nature, also based on the PDSI.
Response 1: Thank you so much for your comment. We will add a more specific description and refer to Sheffield et al. (2012) in the revised manuscript.
Point 2: 43-44. which events' Why these continents and not others?
Response 2: Thank you so much for your comment. I'm sorry that we didn't give a detailed description and made this sentence weak. We will specify different drought events in different regions to explain clearly in the revised manuscript.
Point 3: 44-46: Under rainfall deficits is not expected an increase of ET (as the soil moisture is constrained), then what has an effect is the increase of the atmospheric evaporative demand (this term is absolutely better to use than potential evapotranspiration, see e.g. https://onlinelibrary.wiley.com/doi/full/10.1002/wcc.632), which increases plant stress, leaf temperature (given changes in the partition of latent and sensible heat).
Response 3: Thank you so much for your comment. We agree with your point. It is really important to use the correct terminology. So, we will rewrite this sentence and discuss in detail the type of drought targeted in this paper and the impact factors of drought.
Point 4: 60-61: also thermodynamic forcing (e.g. radiative CO2 forcing increases VPD, but also land-atmosphere feedbacks: see https://nyaspubs.onlinelibrary.wiley.com/doi/10.1111/nyas.13912)
Response 4: Thank you so much for your comment. We are grateful for your recommended paper, and we will check Diego G. Miralles(2018) carefully. We will rewrite this sentence to include more accurate physical mechanisms.
Point 5: 95-103: This is better to be moved to the methods.
Response 5: Thank you so much for your comment. We will move this part to the method in the revised manuscript.
Point 6: 116-117: I think the capability of this dataset must be tested for the use of the assessment of droughts at the regional scale. Some comparison with local stations/datasets would be desired.
Response 6: Thank you so much for your comment. We are also cautious about the capability of datasets. Therefore, in the appendix, we compare the results of the CRU dataset with the data from Japanese meteorological stations and Dynamical Regional Downscaling Using the JRA-55 Reanalysis (DSJAR-55). See Appendix and Figure A4 for details.
Point 7: 122-144: This is not needed. There are several references reviewing PDSI. Authors should cite the pioneer study by W. palmer of 1965. It is not necessary to show the formulation of the FAO-56 Penman-Monteith reference evapotranspiration.
Response 7: Thank you so much for your comment. The purpose of showing the key formula of scPDSI is to facilitate the understanding of the impact of meteorological variables such as precipitation and potential evapotranspiration on scPDSI. We agree with you that the formula for the potential evapotranspiration is really unnecessary, and we will remove this part.
Point 8: 145-157: This is also not needed. Authors are using directly the PDSI data generated by Gerard Van der Schrier, so they may simply refer to this dataset and do not explain how this author generate it.
Response 8: Thank you so much for your comment. The purpose of showing this part is mainly to reflect the advantages of scPDSI; that is, it can consider snow. We are sorry that we did not explain clearly. We will first improve the introduction, explain the advantages of scPDSI and correspond to the content of the introduction here.
Point 9: 159-166: Why do the authors select these circulation indices and not others. Note that the effect of ENSO can be very different as a function of the selected index given the different physical mechanisms related (https://link.springer.com/article/10.1007/s00382-016-3082-y).
Response 9: Thank you so much for your comment. The reason for choosing these climatic drivers is that there have been some papers discussing these climatic drivers on Japan's climate. We will further explain the reasons for the selection and add references in the revised manuscript.
Point 10: 180-199: Not needed to describe the work of Yue et al. 2002.
Response 10: Thank you so much for your comment. We will remove this part in the revised manuscript.
Point 11: What is the purpose of using the methods described in 2.3 and 2.4? The authors should justify its use in the context of the objectives of the study.
Response 11: Thank you so much for your comment. First, we will strengthen the description of the purpose of this article in the introduction. Distinct empirical orthogonal function (DEOF) is mainly to identify drought homogeneous regions in Japan. In this way, we can find specific regions from the long-term Japan drought index time series. And TFPW-MK can identify drought trend changes in homogeneous regions. The drought homogeneous regions will be helpful for drought management and prevention, and it will strengthen our understanding of the characteristics of drought in Japan.
Point 12: 238: In figure 1 it is not necessary to include the trend in temperature. It is not a metric included in the analysis of drought by means of the PDSI. Please, include precipitation and atmospheric demand in absolute units using the same scale.
Response 12: Thank you so much for your comment. The purpose of showing the temperature is that although the temperature is not used directly in the calculation of scPDSI, the temperature is used in the calculation of PET. After all, the temperature has a great influence on PET. We will modify the figure 1 using the same scale in the revised manuscript.
Point 13: 259-260: This is confuse. Gridded data from CRU is based on interpolation of station data. With this approach the authors suggest a re-interpolation based on mann-kendall results, but this should not be necessary as the input data is already a gridded dataset.
Response 13: Thank you so much for your comment. Our study area is small. The 0.5° resolution is challenging to reflect the trends of meteorological variables and drought in different regions of Japan. Interpolation is used to try to distinguish the trend differences between different regions in Japan.
Point 14: 278: Averaging PDSI is not a suitable approach given spatial differences of autocorrelation characteristics in the PDSI (https://journals.ametsoc.org/view/journals/hydr/11/4/2010jhm1224_1.xml) so spatial comparability is poor. If authors are interested to show a time series over Japan, figure 4 showing surface area affected by drought (in %) is much better.
Response 14: Thank you so much for your comment. We will remove Figure 3 in the revised manuscript.
Point 15, 17: 296-308. This figure is redundant. Given strong autocorrelation of the PDSI (https://www.sciencedirect.com/science/article/pii/S0022169414009305), it is not expected a difference in drought conditions at the seasonal level considering the PDSI.
353-358: Same comments related to the relevance of seasonal differences. If the authors want to analyse droughts at the seasonal scale, the PDSI is not the best choice. More useful alternatives are the SPEI or the SPDI (Ma et al., 2014, Hyd Procc). These indices provide exact seasonal information as the time scales are defined a priory and they are known.
Response 15, 17: Thank you so much for your comment. Analysing the seasonality of scPDSI is indeed challenging. But our purpose in these parts is to find the drought changes in different seasons. Different droughts often correspond to different events. For example, the spring drought in Japan often corresponds to wildfires. Summer drought in Japan corresponds to water shortage. We will add some explanations in these parts, explaining that these parts of the result needs to be treated with caution due to the autocorrelation of the scPDSI.
Point 16: Figure 7. I do not find this figure very logical as precipitation is declining more than PDSI. PDSI is low sensitive to the atmospheric demand as there is an alpha parameter used to obtain the cafec PET that limit the role of PET, but this role is always negative so I do not find logical that under increased PET, the precipitation may show a more declining trend that the PDSI.
Response 16: Thank you so much for your comment. We are sorry that Figure 7 (b) and (c) did not use a unified unit. In fact, precipitation has not decreased more significantly than drought. We will modify Figure 7 in the revised manuscript.
Point 18: 3.4. This section seems to be very disconnected to the rest of the study. The authors provide a very simple analysis to connect drought index and area affected by fire. In addition, there is not any critical discussion of the obtained results.
Response 18: Thank you so much for your comment. We will try to strengthen this part and add critical discussions in the revised manuscript.
Point 19: 3.5. This section would gain clarity if in addition to the power spectrum analysis the authors would include further information in order to determine possible anomalies in the selected atmospheric circulation indices during drought conditions (in average for the whole Japan and for the two regions). A simple box-plot or some maps of PDSI anomalies corresponding to high/low phases of these circulation indices would be enough. With the current information it is really difficult to distill if drought events may be affected by these circulation drivers.
Response 19: Thank you so much for your comment. The main purpose of our selection of wavelet analysis is that the time-lag effect of climatic drivers on drought. And out results also showed that the impact of climatic drivers on drought is complex, and it often takes several months to have an impact on drought. We will rewrite the introduction and explain the purpose of each method in the introduction to make the subsequent results more logical.
Point 20: Section 3 although it is named as "results and discussion" shows very limited critical discussion of the obtained results, but this is necessary in any serious scientific research.
Response 20: Thank you so much for your comment. We will add some critical discussions in the revised manuscript.
Citation: https://doi.org/10.5194/nhess-2020-416-AC3 -
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
Dear Francesco Serinaldi
Thank you so much for your comments. The choice of method is really important to the trend analysis results. We will reconsider TFPW-MK test and use other versions of MK test to compare results in the revised manuscript. The effective significance of the trend analysis results will also be discussed in the revised manuscript.
Best regards,
Ke Shi
Citation: https://doi.org/10.5194/nhess-2020-416-AC1
-
AC3: 'Reply on RC2', Ke Shi, 22 Jan 2021
Interactive discussion
Status: closed
-
CC1: 'Comment on nhess-2020-416', Francesco Serinaldi, 06 Jan 2021
The Authors may be interested in the paper below, showing that TFPW-MK test used in this study suffers from mathmatical inconsistencies that affect its effective significance, and make results and their interpretation generally incorrect.
Serinaldi F, Kilsby CG. . Stochastic Environmental Research and Risk Assessment 2016, 30(2), 763-777.
The Authors may also be intersted in the discussion reported in
Serinaldi F, Kilsby CG, Lombardo F. . Advances in Water Resources 2018, 111, 132-155.
and references therein.
Sincerely,
Francesco Serinaldi
Citation: https://doi.org/10.5194/nhess-2020-416-CC1 -
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
Dear Francesco Serinaldi
Thank you so much for your comments. The choice of method is really important to the trend analysis results. We will reconsider TFPW-MK test and use other versions of MK test to compare results in the revised manuscript. The effective significance of the trend analysis results will also be discussed in the revised manuscript.
Best regards,
Ke Shi
Citation: https://doi.org/10.5194/nhess-2020-416-AC1 -
AC2: 'Reply on RC1', Ke Shi, 22 Jan 2021
We greatly appreciate you for your constructive comments and suggestions.
First of all, please allow us to make a brief reply. When the open preprint is over (until 17 Feb 2021), we will make a detailed reply and upload the corresponding revised manuscript.
Point 1: L49-59: While some relevant works about Japan drought have been cited, I feel the literature review appears weak. I recommend the authors to include all studies focusing on Japan drought. Discuss your results and compare them with the results from the previous studies in the appropriate places (results/discussion sections).
Response 1: Thank you so much for your comment. We will continue to look for papers about drought in Japan as much as possible to include all current papers related to drought in Japan in our introduction.
Point 2: L116: The study area of this paper is small, therefore, is the 0.5° resolution meet the purpose of this research? The author should explain that.
Response 2: Thanks for your comment. In fact, a resolution of 0.5° resolution is indeed challenging in identifying drought conditions in some specific areas. But the main purpose of this paper is to identify the drought homogeneous regions of Japan through distinct empirical orthogonal function (DEOF). And the change in resolution does not have much effect on the DEOF result. So, 0.5° resolution can meet the needs of this research.
Point 3~6:
L118: Add space before ‘The’.
L163: I wasn't able to access the details of the data following the link. It appears that there may be a typo in the link, and potentially an access issue.
L169: Check the reference.
L167, L201, L218: Please check the font size.
Response 3~6: Thanks for your comments. Sorry for these mistakes, we will correct them in the revised manuscript.
Point 7, Point 9:
L313-314: This part is interesting. The author should try to further explain the physical mechanism about the impact of spatial variation of rainfall on drought.
L330-339: Expand the explanation for physical interpretation of the DEOF1 and DEOF2. If needed, make a reference to other variables.
Response 7, 9: Thanks for your comment. We will add some physical mechanism descriptions in these parts.
Point 8, 10~12:
L321: Replace ‘Spatial and temporal’ with ‘Spatiotemporal’. It should be consistent with the rest of the document.
L348 Figure 8: Show the W and N regions in your figure.
L393 Figure 12: Replace ‘In’ with ‘in’.
L526 Figure A4: There is a strange horizontal line under the title of Figure A4(b). Please check all figures.
Response 8, 10~12: Thanks for your comments. Sorry for these mistakes, we will correct them and check all the figures in the revised manuscript.
Citation: https://doi.org/10.5194/nhess-2020-416-AC2
-
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
-
RC1: 'Comment on nhess-2020-416', Anonymous Referee #1, 11 Jan 2021
This paper tried to explore the spatiotemporal characteristics of droughts over Japan and also linked the droughts with global circulation indices. Exploring the occurrences, development and underlying mechanisms of drought is the key to understanding them. From this point of view, this paper will be of great importance to Japan drought research.
Therefore, it is an interesting topic and suitable for Natural Hazards and Earth System Sciences. However, there are some issues that should be addressed by the author before publication. Especially lack of some physical mechanism discussion. Specific comments are as follows:
1. Introduction
L49-59: While some relevant works about Japan drought have been cited, I feel the literature review appears weak. I recommend the authors to include all studies focusing on Japan drought. Discuss your results and compare them with the results from the previous studies in the appropriate places (results/discussion sections).
2. Materials and Methodology
L116: The study area of this paper is small, therefore, is the 0.5° resolution meet the purpose of this research? The author should explain that.
L118: Add space before ‘The’.
L163: I wasn't able to access the details of the data following the link. It appears that there may be a typo in the link, and potentially an access issue.
L169: Check the reference.
L167, L201, L218: Please check the font size.
3. Results and Discussion
L313-314: This part is interesting. The author should try to further explain the physical mechanism about the impact of spatial variation of rainfall on drought.
L321: Replace ‘Spatial and temporal’ with ‘Spatiotemporal’. It should be consistent with the rest of the document.
L330-339: Expand the explanation for physical interpretation of the DEOF1 and DEOF2. If needed, make a reference to other variables.
L348 Figure 8: Show the W and N regions in your figure.
L393 Figure 12: Replace ‘In’ with ‘in’
Appendix
L526 Figure A4: There is a strange horizontal line under the title of Figure A4(b). Please check all figures.
Citation: https://doi.org/10.5194/nhess-2020-416-RC1 -
AC2: 'Reply on RC1', Ke Shi, 22 Jan 2021
We greatly appreciate you for your constructive comments and suggestions.
First of all, please allow us to make a brief reply. When the open preprint is over (until 17 Feb 2021), we will make a detailed reply and upload the corresponding revised manuscript.
Point 1: L49-59: While some relevant works about Japan drought have been cited, I feel the literature review appears weak. I recommend the authors to include all studies focusing on Japan drought. Discuss your results and compare them with the results from the previous studies in the appropriate places (results/discussion sections).
Response 1: Thank you so much for your comment. We will continue to look for papers about drought in Japan as much as possible to include all current papers related to drought in Japan in our introduction.
Point 2: L116: The study area of this paper is small, therefore, is the 0.5° resolution meet the purpose of this research? The author should explain that.
Response 2: Thanks for your comment. In fact, a resolution of 0.5° resolution is indeed challenging in identifying drought conditions in some specific areas. But the main purpose of this paper is to identify the drought homogeneous regions of Japan through distinct empirical orthogonal function (DEOF). And the change in resolution does not have much effect on the DEOF result. So, 0.5° resolution can meet the needs of this research.
Point 3~6:
L118: Add space before ‘The’.
L163: I wasn't able to access the details of the data following the link. It appears that there may be a typo in the link, and potentially an access issue.
L169: Check the reference.
L167, L201, L218: Please check the font size.
Response 3~6: Thanks for your comments. Sorry for these mistakes, we will correct them in the revised manuscript.
Point 7, Point 9:
L313-314: This part is interesting. The author should try to further explain the physical mechanism about the impact of spatial variation of rainfall on drought.
L330-339: Expand the explanation for physical interpretation of the DEOF1 and DEOF2. If needed, make a reference to other variables.
Response 7, 9: Thanks for your comment. We will add some physical mechanism descriptions in these parts.
Point 8, 10~12:
L321: Replace ‘Spatial and temporal’ with ‘Spatiotemporal’. It should be consistent with the rest of the document.
L348 Figure 8: Show the W and N regions in your figure.
L393 Figure 12: Replace ‘In’ with ‘in’.
L526 Figure A4: There is a strange horizontal line under the title of Figure A4(b). Please check all figures.
Response 8, 10~12: Thanks for your comments. Sorry for these mistakes, we will correct them and check all the figures in the revised manuscript.
Citation: https://doi.org/10.5194/nhess-2020-416-AC2 -
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
Dear Francesco Serinaldi
Thank you so much for your comments. The choice of method is really important to the trend analysis results. We will reconsider TFPW-MK test and use other versions of MK test to compare results in the revised manuscript. The effective significance of the trend analysis results will also be discussed in the revised manuscript.
Best regards,
Ke Shi
Citation: https://doi.org/10.5194/nhess-2020-416-AC1
-
AC2: 'Reply on RC1', Ke Shi, 22 Jan 2021
-
RC2: 'Comment on nhess-2020-416', Anonymous Referee #2, 21 Jan 2021
Review of: Evaluating Spatiotemporal Patterns and Trends of Drought in Japan Associated with Global Climatic Drivers.
This article analyses drought variability in Japan by means of global datasets and the PDSI. I think the manuscript needs to reinterpret and rework several issues and to change several issues affecting both formal and substance aspects.
35-36. This should be qualified in some way. In which regions? You note e.g. the different view by Sheffield et al. (2012) Nature, also based on the PDSI.
43-44. which events' Why these continents and not others?
44-46: Under rainfall deficits is not expected an increase of ET (as the soil moisture is constrained), then what has an effect is the increase of the atmospheric evaporative demand (this term is absolutely better to use than potential evapotranspiration, see e.g. https://onlinelibrary.wiley.com/doi/full/10.1002/wcc.632), which increases plant stress, leaf temperature (given changes in the partition of latent and sensible heat).
60-61: also thermodynamic forcing (e.g. radiative CO2 forcing increases VPD, but also land-atmopshere feedbacks: see https://nyaspubs.onlinelibrary.wiley.com/doi/10.1111/nyas.13912)
95-103: This is better to be moved to the methods.
116-117: I think the capability of this dataset must be tested for the use of the assessment of droughts at the regional scale. Some comparison with local stations/datasets would be desired.
122-144: This is not needed. There are several references reviewing PDSI. Authors should cite the pioneer study by W. palmer of 1965. It is not necessary to show the formulation of the FAO-56 Penman-Monteith reference evapotranspiration.
145-157: This is also not needed. Authors are using directly the PDSI data generated by Gerard Van der Schrier, so they may simply refer to this dataset and do not explain how this author generate it.
159-166: Why do the authors select these circulation indices and not others. Note that the effect of ENSO can be very different as a function of the selected index given the different physical mechanisms related (https://link.springer.com/article/10.1007/s00382-016-3082-y).
180-199: Not needed to describe the work of Yue et al. 2002.
What is the prupose of using the methods described in 2.3 and 2.4? The authors should justify its use in the context of the objectives of the study.
238: In figure 1 it is not necessary to include the trend in temperature. It is not a metric included in the analysis of drought by means of the PDSI. Please, include precipitation and atmospheric demand in absolute units using the same scale.
259-260: This is confuse. Gridded data from CRU is based on interpolation of station data. With this approach the authors suggest a re-interpolation based on mann-kendall results, but this should not be necessary as the input data is already a gridded dataset.
278: Averaging PDSI is not a suitable approach given spatial differences of autocorrelation characteristics in the PDSI (https://journals.ametsoc.org/view/journals/hydr/11/4/2010jhm1224_1.xml) so spatial comparability is poor. If authors are interested to show a time series over Japan, figure 4 showing surface area affected by drought (in %) is much better.
296-308. This figure is redundant. Given strong autocorrelation of the PDSI (https://www.sciencedirect.com/science/article/pii/S0022169414009305), it is not expected a difference in drought conditions at the seasonal level considering the PDSI.
Figure 7. I do not find this figure very logical as precipitation is declining more than PDSI. PDSI is low sensitive to the atmospheric demand as there is an alpha parameter used to obtain the cafec PET that limit the role of PET, but this role is always negative so I do not find logical that under increased PET, the precipitation may show a more declining trend that the PDSI.
353-358: Same comments related to the relevance of seasonal differences. If the authors want to analyse droughts at the seasonal scale, the PDSI is not the best choice. More useful alternatives are the SPEI or the SPDI (Ma et al., 2014, Hyd Procc). These indices provide exact seasonal information as the time scales are defined a priory and they are known.3.4. This section seems to be very disconnected to the rest of the study. The authors provide a very simple analysis to connect drought index and area affected by fire. In addition, there is not any critical discussion of the obtained results.
3.5. This section would gain clarity if in addition to the power spectrum analysis the authors would include further information in order to determine possible anomalies in the selected atmospheric circulation indices during drought conditions (in average for the whole Japan and for the two regions). A simple box-plot or some maps of PDSI anomalies corresponding to high/low phases of these circulation indices would be enough. With the current information it is really difficult to distill if drought events may be affected by these circulation drivers.Section 3 although it is named as "results and discussion" shows very limited critical discussion of the obtained results, but this is necessary in any serious scientific research.
Citation: https://doi.org/10.5194/nhess-2020-416-RC2 -
AC3: 'Reply on RC2', Ke Shi, 22 Jan 2021
We greatly appreciate you for your constructive comments and suggestions.
First of all, please allow us to make a brief reply. When the open preprint is over (until 17 Feb 2021), we will make a detailed reply and upload the corresponding revised manuscript.
Point 1: 35-36. This should be qualified in some way. In which regions? You note e.g. the different view by Sheffield et al. (2012) Nature, also based on the PDSI.
Response 1: Thank you so much for your comment. We will add a more specific description and refer to Sheffield et al. (2012) in the revised manuscript.
Point 2: 43-44. which events' Why these continents and not others?
Response 2: Thank you so much for your comment. I'm sorry that we didn't give a detailed description and made this sentence weak. We will specify different drought events in different regions to explain clearly in the revised manuscript.
Point 3: 44-46: Under rainfall deficits is not expected an increase of ET (as the soil moisture is constrained), then what has an effect is the increase of the atmospheric evaporative demand (this term is absolutely better to use than potential evapotranspiration, see e.g. https://onlinelibrary.wiley.com/doi/full/10.1002/wcc.632), which increases plant stress, leaf temperature (given changes in the partition of latent and sensible heat).
Response 3: Thank you so much for your comment. We agree with your point. It is really important to use the correct terminology. So, we will rewrite this sentence and discuss in detail the type of drought targeted in this paper and the impact factors of drought.
Point 4: 60-61: also thermodynamic forcing (e.g. radiative CO2 forcing increases VPD, but also land-atmosphere feedbacks: see https://nyaspubs.onlinelibrary.wiley.com/doi/10.1111/nyas.13912)
Response 4: Thank you so much for your comment. We are grateful for your recommended paper, and we will check Diego G. Miralles(2018) carefully. We will rewrite this sentence to include more accurate physical mechanisms.
Point 5: 95-103: This is better to be moved to the methods.
Response 5: Thank you so much for your comment. We will move this part to the method in the revised manuscript.
Point 6: 116-117: I think the capability of this dataset must be tested for the use of the assessment of droughts at the regional scale. Some comparison with local stations/datasets would be desired.
Response 6: Thank you so much for your comment. We are also cautious about the capability of datasets. Therefore, in the appendix, we compare the results of the CRU dataset with the data from Japanese meteorological stations and Dynamical Regional Downscaling Using the JRA-55 Reanalysis (DSJAR-55). See Appendix and Figure A4 for details.
Point 7: 122-144: This is not needed. There are several references reviewing PDSI. Authors should cite the pioneer study by W. palmer of 1965. It is not necessary to show the formulation of the FAO-56 Penman-Monteith reference evapotranspiration.
Response 7: Thank you so much for your comment. The purpose of showing the key formula of scPDSI is to facilitate the understanding of the impact of meteorological variables such as precipitation and potential evapotranspiration on scPDSI. We agree with you that the formula for the potential evapotranspiration is really unnecessary, and we will remove this part.
Point 8: 145-157: This is also not needed. Authors are using directly the PDSI data generated by Gerard Van der Schrier, so they may simply refer to this dataset and do not explain how this author generate it.
Response 8: Thank you so much for your comment. The purpose of showing this part is mainly to reflect the advantages of scPDSI; that is, it can consider snow. We are sorry that we did not explain clearly. We will first improve the introduction, explain the advantages of scPDSI and correspond to the content of the introduction here.
Point 9: 159-166: Why do the authors select these circulation indices and not others. Note that the effect of ENSO can be very different as a function of the selected index given the different physical mechanisms related (https://link.springer.com/article/10.1007/s00382-016-3082-y).
Response 9: Thank you so much for your comment. The reason for choosing these climatic drivers is that there have been some papers discussing these climatic drivers on Japan's climate. We will further explain the reasons for the selection and add references in the revised manuscript.
Point 10: 180-199: Not needed to describe the work of Yue et al. 2002.
Response 10: Thank you so much for your comment. We will remove this part in the revised manuscript.
Point 11: What is the purpose of using the methods described in 2.3 and 2.4? The authors should justify its use in the context of the objectives of the study.
Response 11: Thank you so much for your comment. First, we will strengthen the description of the purpose of this article in the introduction. Distinct empirical orthogonal function (DEOF) is mainly to identify drought homogeneous regions in Japan. In this way, we can find specific regions from the long-term Japan drought index time series. And TFPW-MK can identify drought trend changes in homogeneous regions. The drought homogeneous regions will be helpful for drought management and prevention, and it will strengthen our understanding of the characteristics of drought in Japan.
Point 12: 238: In figure 1 it is not necessary to include the trend in temperature. It is not a metric included in the analysis of drought by means of the PDSI. Please, include precipitation and atmospheric demand in absolute units using the same scale.
Response 12: Thank you so much for your comment. The purpose of showing the temperature is that although the temperature is not used directly in the calculation of scPDSI, the temperature is used in the calculation of PET. After all, the temperature has a great influence on PET. We will modify the figure 1 using the same scale in the revised manuscript.
Point 13: 259-260: This is confuse. Gridded data from CRU is based on interpolation of station data. With this approach the authors suggest a re-interpolation based on mann-kendall results, but this should not be necessary as the input data is already a gridded dataset.
Response 13: Thank you so much for your comment. Our study area is small. The 0.5° resolution is challenging to reflect the trends of meteorological variables and drought in different regions of Japan. Interpolation is used to try to distinguish the trend differences between different regions in Japan.
Point 14: 278: Averaging PDSI is not a suitable approach given spatial differences of autocorrelation characteristics in the PDSI (https://journals.ametsoc.org/view/journals/hydr/11/4/2010jhm1224_1.xml) so spatial comparability is poor. If authors are interested to show a time series over Japan, figure 4 showing surface area affected by drought (in %) is much better.
Response 14: Thank you so much for your comment. We will remove Figure 3 in the revised manuscript.
Point 15, 17: 296-308. This figure is redundant. Given strong autocorrelation of the PDSI (https://www.sciencedirect.com/science/article/pii/S0022169414009305), it is not expected a difference in drought conditions at the seasonal level considering the PDSI.
353-358: Same comments related to the relevance of seasonal differences. If the authors want to analyse droughts at the seasonal scale, the PDSI is not the best choice. More useful alternatives are the SPEI or the SPDI (Ma et al., 2014, Hyd Procc). These indices provide exact seasonal information as the time scales are defined a priory and they are known.
Response 15, 17: Thank you so much for your comment. Analysing the seasonality of scPDSI is indeed challenging. But our purpose in these parts is to find the drought changes in different seasons. Different droughts often correspond to different events. For example, the spring drought in Japan often corresponds to wildfires. Summer drought in Japan corresponds to water shortage. We will add some explanations in these parts, explaining that these parts of the result needs to be treated with caution due to the autocorrelation of the scPDSI.
Point 16: Figure 7. I do not find this figure very logical as precipitation is declining more than PDSI. PDSI is low sensitive to the atmospheric demand as there is an alpha parameter used to obtain the cafec PET that limit the role of PET, but this role is always negative so I do not find logical that under increased PET, the precipitation may show a more declining trend that the PDSI.
Response 16: Thank you so much for your comment. We are sorry that Figure 7 (b) and (c) did not use a unified unit. In fact, precipitation has not decreased more significantly than drought. We will modify Figure 7 in the revised manuscript.
Point 18: 3.4. This section seems to be very disconnected to the rest of the study. The authors provide a very simple analysis to connect drought index and area affected by fire. In addition, there is not any critical discussion of the obtained results.
Response 18: Thank you so much for your comment. We will try to strengthen this part and add critical discussions in the revised manuscript.
Point 19: 3.5. This section would gain clarity if in addition to the power spectrum analysis the authors would include further information in order to determine possible anomalies in the selected atmospheric circulation indices during drought conditions (in average for the whole Japan and for the two regions). A simple box-plot or some maps of PDSI anomalies corresponding to high/low phases of these circulation indices would be enough. With the current information it is really difficult to distill if drought events may be affected by these circulation drivers.
Response 19: Thank you so much for your comment. The main purpose of our selection of wavelet analysis is that the time-lag effect of climatic drivers on drought. And out results also showed that the impact of climatic drivers on drought is complex, and it often takes several months to have an impact on drought. We will rewrite the introduction and explain the purpose of each method in the introduction to make the subsequent results more logical.
Point 20: Section 3 although it is named as "results and discussion" shows very limited critical discussion of the obtained results, but this is necessary in any serious scientific research.
Response 20: Thank you so much for your comment. We will add some critical discussions in the revised manuscript.
Citation: https://doi.org/10.5194/nhess-2020-416-AC3 -
AC1: 'Reply on CC1', Ke Shi, 07 Jan 2021
Dear Francesco Serinaldi
Thank you so much for your comments. The choice of method is really important to the trend analysis results. We will reconsider TFPW-MK test and use other versions of MK test to compare results in the revised manuscript. The effective significance of the trend analysis results will also be discussed in the revised manuscript.
Best regards,
Ke Shi
Citation: https://doi.org/10.5194/nhess-2020-416-AC1
-
AC3: 'Reply on RC2', Ke Shi, 22 Jan 2021
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 893 | 608 | 66 | 1,567 | 47 | 83 |
- HTML: 893
- PDF: 608
- XML: 66
- Total: 1,567
- BibTeX: 47
- EndNote: 83
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1