the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 03 Nov 2025
Towards an operational European Drought Impacts Database (EDID)
Abstract. Drought impact information is essential to move from reactive management to a proactive approach. Data on drought impacts provide regional insight into vulnerability and support robust risk assessment and sustainable adaptation strategies. Drought impact data are also essential to build and validate models for advanced impact forecasting, including AI enhanced tools. While there is increasing consensus on the operational use of specific physical drought hazard indices, to date there is no generally accepted convention on drought impact data collection and use. Based on experience and content of several regional research databases, the development of a European Drought Impact Database (EDID) explicitly aims for operational application within the framework of the Copernicus European Drought Observatory. This article gives insight into the implementation of EDID, its structure and attributes, and provides an analysis of the content. Among the nine impacted systems, agriculture, public water supply and aquatic ecosystems contribute a majority of the impact records. Over the covered time period, impacts became more variable in the system they describe and recent years show some more extremely severe impacts according to a newly introduced severity score. Mapped at country scale, the impacts confirm previously identified European sectorial impact hotspots. The work and product show that regional datasets can be integrated and add valuable information to an international European database. Public accessibility now provides the opportunity for update and improvement by mobilizing the European drought community.
-
Notice on discussion status
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
-
Preprint
(5048 KB)
-
Supplement
(1566 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5048 KB) - Metadata XML
-
Supplement
(1566 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-4806', Anonymous Referee #1, 17 Nov 2025
Review of: Towards an Operational European Drought Impacts Database (EDID)
This study addresses a very important gap in the development of drought-impact databases—an essential resource for assessing drought vulnerability and accurately evaluating drought risk. The manuscript not only presents the methodology for constructing such a database at the European scale, but also clearly describes its structure and provides a preliminary sectoral analysis of drought impacts. Moreover, the database is conceived as a living resource, with the capacity to be regularly updated and to incorporate regional information, which gives the study strong potential for future enhancement.
The manuscript is well written and well structured. The introduction is well documented, offering a thorough, up-to-date state of the art and clearly establishing the relevance of creating such a database. The description of the database is particularly informative, detailing not only the categories of impacts but also the associated intensity levels. It might have been desirable to include additional information sources on drought impacts at national and regional scales, although I understand that such data may not be available for many countries. In any case, since this is a dynamic and open database that will continue to evolve, new databases and information sources can be incorporated in the future. The semi-automatic approach using AI is also valuable and will likely yield further improvements in the coming years.
A notable strength of this work is the integration of diverse data sources. Combining information from media reports, government documents, scientific literature, and other sources is a challenging task, and the authors have made a commendable effort to harmonize these data and classify them by impact type and severity.
The spatial and temporal analysis of drought impacts is particularly useful, as it highlights the interannual variability of impacts and their increasing dynamism. The spatial analysis also reflects current limitations in the database, particularly regarding the uneven geographic coverage, with southern European countries being under-represented despite frequently experiencing severe drought impacts. This is an issue to address in the future, perhaps by engaging with research teams working on drought-impact monitoring in these regions.
Finally, the discussion section is very informative and well organized. It effectively explains the relevance of the new database, acknowledges its current limitations, and outlines the planned steps for its improvement.
In summary, this is a high-quality and novel study that fills an important gap. It represents the result of several years of effort dedicated to compiling, classifying, and harmonizing drought-impact information from multiple sources. I recommend publication as is.
Citation: https://doi.org/10.5194/egusphere-2025-4806-RC1 -
AC2: 'Reply on RC1', Kerstin Stahl, 20 Dec 2025
We thank Reviewer 1 for these positive comments and look forward to the manuscript's publication. Indeed, we also strongly hope (with this paper) to motivate researchers and practitioners to engage also in data collection. It is tedious work to collect and categorize drought impacts from text reports (even when the task is facilitated by the new automatic search tools), but, all contributions will add to our knowledge on drought.
Citation: https://doi.org/10.5194/egusphere-2025-4806-AC2
-
AC2: 'Reply on RC1', Kerstin Stahl, 20 Dec 2025
-
RC2: 'Comment on egusphere-2025-4806', Anonymous Referee #2, 14 Dec 2025
This study presents the development of the European Drought Impacts Database (EDID), a significant step toward operationalizing drought impact monitoring within the Copernicus European Drought Observatory. The work addresses a critical gap in drought risk management by consolidating fragmented impact data and introducing a structured, operational framework. The paper is well-written, timely, and relevant for both scientific and operational communities.
Below, I provide general and specific comments aimed at improving clarity and completeness. Once addressed, I believe the manuscript will make a valuable contribution to NHESS.
General comments:
1) Scope of coverage: the manuscript states that EDID operates at the European scale; however, the dataset does not include all European countries and extends beyond the EU (e.g., Turkey). Please clarify the rationale for defining the geographic scope and selection criteria for included countries.2) Long-term maintenance as a challenge / missing discussion point: While the Abstract and Conclusion mention EDID’s operational integration within Copernicus, Section 4.2 does not explicitly discuss one of the main challenges: ensuring long-term maintenance and sustainability of the database under a specific institutional umbrella. Adding details on governance, funding, and community engagement plans would provide valuable insight into how EDID will remain operational and updated over time.
3) Underrepresented systems: in Section 4.1 (around L306), the manuscript notes that some systems are less frequently represented. It would strengthen the discussion to explicitly name these systems and provide possible reasons for their lower representation (e.g., data availability, reporting practices, sector-specific challenges).
4) Additional challenges: Section 4.2 could be expanded to include other challenges, such as harmonizing impact reports with demographic or socio-economic factors (e.g., population density), which are critical for interpreting vulnerability and exposure.
Specific comments:
Section 2.2 (L158): The text states that “measured and systematically observed data were found to be either weak or incomplete” and thus excluded from EDID. However, operational sources like propluvia were included. Please clarify why propluvia is not considered part of the “measured and systematically observed data” category to avoid confusion.
Abstract (L22): Consider adding a strong concluding sentence beyond “mobilizing the drought community,” emphasizing the broader significance of EDID for risk management and policy.
L28: Please specify what is meant by “risk” in contrast to “impact” (e.g., risk as potential impact) for clarity.
L30: “Impact-oriented” should start with lowercase “i.”
Figure 1 (L103): Indicate what the left and right screenshots represent for clarity.
Table 1: Typo: “Bachmaier” should be “Bachmair.”
Table 1: Please ensure all references listed in Table 1 appear in the Reference section (e.g., Bachmair et al., 2017 seems missing).
L199: Briefly explain what a Sankey diagram is for readers unfamiliar with the term.
Figure 2: Excellent synthesis of complex information. Two suggestions: 1) Consider improving readability by enlarging boxes or arranging text vertically. 2) Fig. would benefit from adding percentage values directly in the figure for “Original Source Type” and “EDID Systems” categories.
Section 4.1 (L290): Consider revising the title for clarity: “4.1 A new baseline of drought impacts across Europe: spatio-temporal patterns.”
Citation: https://doi.org/10.5194/egusphere-2025-4806-RC2 -
AC1: 'Reply on RC2', Kerstin Stahl, 20 Dec 2025
Dear Reviewer 2,
thanks a lot for the positive review. We respond to all comments below
General Comments:
1) Scope of coverage: the manuscript states that EDID operates at the European scale; however, the dataset does not include all European countries and extends beyond the EU (e.g., Turkey). Please clarify the rationale for defining the geographic scope and selection criteria for included countries.
Response: Funding came mostly from the EU, both for prior research projects and recently for the EDORA project that made possible the EDID development. Therefore, priority was given to EU countries and, when data sources allowed, other countries were included too. We agree that the inclusion of databases from Turkey or Northern Africa will be highly desirable, also considering that the physical drought indicators of the European Drought Observatory have been extended to that region. Hopefully such impact databases are being gathered in the respective countries for inclusion into EDID. This paper aims to encourage such efforts and show the way to proceed.
2) Long-term maintenance as a challenge / missing discussion point: While the Abstract and Conclusion mention EDID’s operational integration within Copernicus, Section 4.2 does not explicitly discuss one of the main challenges: ensuring long-term maintenance and sustainability of the database under a specific institutional umbrella. Adding details on governance, funding, and community engagement plans would provide valuable insight into how EDID will remain operational and updated over time.
Response: Indeed long-term maintenance is a challenge and we agree that it will be useful to elaborate on this in the discussion and conclusion: "Maintenance and updates of EDID are ensured by JRC in the framework of the Copernicus European Drought Observatory and leveraging also its role as co-chair of the EU Working Group on water scarcity and drought. JRC has already developed a long-term plan based on: the Europe Media Monitor and the use of large language model; the engagement of drought experts that can contribute via a dedicated component of EDID (about to be released) by adding new records."
3) Underrepresented systems: in Section 4.1 (around L306), the manuscript notes that some systems are less frequently represented. It would strengthen the discussion to explicitly name these systems and provide possible reasons for their lower representation (e.g., data availability, reporting practices, sector-specific challenges).
Response: We agree that we can certainly name them and elaborate further on those systems in the revised version. Indeed a bit of interpretation is already in the results section and we will harmonize this in the revision. In brief, the two energy systems and navigation are much more restricted to locations where they can occur, i.e. the infrastructure that can be affected doesn't exist everywhere (e.g. transportation routes, reservoirs). In addition, the impacts on people are less tangible and may be modified by market forces, which will likely affect the degree of awareness and hence reporting. The system ecosystems-terrestrial includes forestry (among other affected ecosystems). Drought response occurs with delay and will be moderated by resistance and resilience of the ecosystems. More research is needed on this system and here monitored vegetation health indicators can provide a useful comparison. We suggest that we will revise this discussion point in this direction and point out some opportunities and needs for further research on these systems.
4) Additional challenges: Section 4.2 could be expanded to include other challenges, such as harmonizing impact reports with demographic or socio-economic factors (e.g., population density), which are critical for interpreting vulnerability and exposure.
Response: Thanks for this important suggestion. Efforts were made to remove duplicates but indeed the consideration for cultural (are drought impacts found worth reporting?), demographic (more people more awareness?) and socioeconomic (more/better news?) is certainly worth discussing and hopefully at some point considering in analyses as well. We agree to add this aspect to the revised discussion.
Specific comments:
Section 2.2 (L158): The text states that “measured and systematically observed data were found to be either weak or incomplete” and thus excluded from EDID. However, operational sources like propluvia were included. Please clarify why propluvia is not considered part of the “measured and systematically observed data” category to avoid confusion.
Response: Thanks for pointing out that this is imprecise and indeed there is no very clear distinction. We suggest to revise the sentence to say 'More ...datasets were considered initially but... '.
Abstract (L22): Consider adding a strong concluding sentence beyond “mobilizing the drought community,” emphasizing the broader significance of EDID for risk management and policy.
Response: Good idea - we will add that in the revised manuscript.
L28: Please specify what is meant by “risk” in contrast to “impact” (e.g., risk as potential impact) for clarity.
Response: We would like to avoid a long discussion on where impacts fit into the risk concepts as there are different opinions about this and we don't want to restrict the interpretation or use of the data for any type of assessment. We suggest to revise: ..."do not directly provide data on observed impacts, nor on vulnerabilities that may contribute to drought risk."
L30: “Impact-oriented” should start with lowercase “i.”
Response: Thanks, this will be corrected.
Figure 1 (L103): Indicate what the left and right screenshots represent for clarity.
Response: Thanks, this will be corrected.
Table 1: Typo: “Bachmaier” should be “Bachmair.”
Response: Thanks, this will be corrected.
Table 1: Please ensure all references listed in Table 1 appear in the Reference section (e.g., Bachmair et al., 2017 seems missing).
Response: Sorry, of course this will be corrected.
L199: Briefly explain what a Sankey diagram is for readers unfamiliar with the term.
Response: Yes, we suggest to revise the text to explain that we visualize the impact data flow from one categorization to another (adding Sankey diagram in parentheses).
Figure 2: Excellent synthesis of complex information. Two suggestions: 1) Consider improving readability by enlarging boxes or arranging text vertically. 2) Fig. would benefit from adding percentage values directly in the figure for “Original Source Type” and “EDID Systems” categories.
Response: We will work on the readability and try if the numbers can fit in for a revised version.
Section 4.1 (L290): Consider revising the title for clarity: “4.1 A new baseline of drought impacts across Europe: spatio-temporal patterns.”
Response: Agreed, this will be changed. Thanks for the suggestion.
Citation: https://doi.org/10.5194/egusphere-2025-4806-AC1
-
AC1: 'Reply on RC2', Kerstin Stahl, 20 Dec 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-4806', Anonymous Referee #1, 17 Nov 2025
Review of: Towards an Operational European Drought Impacts Database (EDID)
This study addresses a very important gap in the development of drought-impact databases—an essential resource for assessing drought vulnerability and accurately evaluating drought risk. The manuscript not only presents the methodology for constructing such a database at the European scale, but also clearly describes its structure and provides a preliminary sectoral analysis of drought impacts. Moreover, the database is conceived as a living resource, with the capacity to be regularly updated and to incorporate regional information, which gives the study strong potential for future enhancement.
The manuscript is well written and well structured. The introduction is well documented, offering a thorough, up-to-date state of the art and clearly establishing the relevance of creating such a database. The description of the database is particularly informative, detailing not only the categories of impacts but also the associated intensity levels. It might have been desirable to include additional information sources on drought impacts at national and regional scales, although I understand that such data may not be available for many countries. In any case, since this is a dynamic and open database that will continue to evolve, new databases and information sources can be incorporated in the future. The semi-automatic approach using AI is also valuable and will likely yield further improvements in the coming years.
A notable strength of this work is the integration of diverse data sources. Combining information from media reports, government documents, scientific literature, and other sources is a challenging task, and the authors have made a commendable effort to harmonize these data and classify them by impact type and severity.
The spatial and temporal analysis of drought impacts is particularly useful, as it highlights the interannual variability of impacts and their increasing dynamism. The spatial analysis also reflects current limitations in the database, particularly regarding the uneven geographic coverage, with southern European countries being under-represented despite frequently experiencing severe drought impacts. This is an issue to address in the future, perhaps by engaging with research teams working on drought-impact monitoring in these regions.
Finally, the discussion section is very informative and well organized. It effectively explains the relevance of the new database, acknowledges its current limitations, and outlines the planned steps for its improvement.
In summary, this is a high-quality and novel study that fills an important gap. It represents the result of several years of effort dedicated to compiling, classifying, and harmonizing drought-impact information from multiple sources. I recommend publication as is.
Citation: https://doi.org/10.5194/egusphere-2025-4806-RC1 -
AC2: 'Reply on RC1', Kerstin Stahl, 20 Dec 2025
We thank Reviewer 1 for these positive comments and look forward to the manuscript's publication. Indeed, we also strongly hope (with this paper) to motivate researchers and practitioners to engage also in data collection. It is tedious work to collect and categorize drought impacts from text reports (even when the task is facilitated by the new automatic search tools), but, all contributions will add to our knowledge on drought.
Citation: https://doi.org/10.5194/egusphere-2025-4806-AC2
-
AC2: 'Reply on RC1', Kerstin Stahl, 20 Dec 2025
-
RC2: 'Comment on egusphere-2025-4806', Anonymous Referee #2, 14 Dec 2025
This study presents the development of the European Drought Impacts Database (EDID), a significant step toward operationalizing drought impact monitoring within the Copernicus European Drought Observatory. The work addresses a critical gap in drought risk management by consolidating fragmented impact data and introducing a structured, operational framework. The paper is well-written, timely, and relevant for both scientific and operational communities.
Below, I provide general and specific comments aimed at improving clarity and completeness. Once addressed, I believe the manuscript will make a valuable contribution to NHESS.
General comments:
1) Scope of coverage: the manuscript states that EDID operates at the European scale; however, the dataset does not include all European countries and extends beyond the EU (e.g., Turkey). Please clarify the rationale for defining the geographic scope and selection criteria for included countries.2) Long-term maintenance as a challenge / missing discussion point: While the Abstract and Conclusion mention EDID’s operational integration within Copernicus, Section 4.2 does not explicitly discuss one of the main challenges: ensuring long-term maintenance and sustainability of the database under a specific institutional umbrella. Adding details on governance, funding, and community engagement plans would provide valuable insight into how EDID will remain operational and updated over time.
3) Underrepresented systems: in Section 4.1 (around L306), the manuscript notes that some systems are less frequently represented. It would strengthen the discussion to explicitly name these systems and provide possible reasons for their lower representation (e.g., data availability, reporting practices, sector-specific challenges).
4) Additional challenges: Section 4.2 could be expanded to include other challenges, such as harmonizing impact reports with demographic or socio-economic factors (e.g., population density), which are critical for interpreting vulnerability and exposure.
Specific comments:
Section 2.2 (L158): The text states that “measured and systematically observed data were found to be either weak or incomplete” and thus excluded from EDID. However, operational sources like propluvia were included. Please clarify why propluvia is not considered part of the “measured and systematically observed data” category to avoid confusion.
Abstract (L22): Consider adding a strong concluding sentence beyond “mobilizing the drought community,” emphasizing the broader significance of EDID for risk management and policy.
L28: Please specify what is meant by “risk” in contrast to “impact” (e.g., risk as potential impact) for clarity.
L30: “Impact-oriented” should start with lowercase “i.”
Figure 1 (L103): Indicate what the left and right screenshots represent for clarity.
Table 1: Typo: “Bachmaier” should be “Bachmair.”
Table 1: Please ensure all references listed in Table 1 appear in the Reference section (e.g., Bachmair et al., 2017 seems missing).
L199: Briefly explain what a Sankey diagram is for readers unfamiliar with the term.
Figure 2: Excellent synthesis of complex information. Two suggestions: 1) Consider improving readability by enlarging boxes or arranging text vertically. 2) Fig. would benefit from adding percentage values directly in the figure for “Original Source Type” and “EDID Systems” categories.
Section 4.1 (L290): Consider revising the title for clarity: “4.1 A new baseline of drought impacts across Europe: spatio-temporal patterns.”
Citation: https://doi.org/10.5194/egusphere-2025-4806-RC2 -
AC1: 'Reply on RC2', Kerstin Stahl, 20 Dec 2025
Dear Reviewer 2,
thanks a lot for the positive review. We respond to all comments below
General Comments:
1) Scope of coverage: the manuscript states that EDID operates at the European scale; however, the dataset does not include all European countries and extends beyond the EU (e.g., Turkey). Please clarify the rationale for defining the geographic scope and selection criteria for included countries.
Response: Funding came mostly from the EU, both for prior research projects and recently for the EDORA project that made possible the EDID development. Therefore, priority was given to EU countries and, when data sources allowed, other countries were included too. We agree that the inclusion of databases from Turkey or Northern Africa will be highly desirable, also considering that the physical drought indicators of the European Drought Observatory have been extended to that region. Hopefully such impact databases are being gathered in the respective countries for inclusion into EDID. This paper aims to encourage such efforts and show the way to proceed.
2) Long-term maintenance as a challenge / missing discussion point: While the Abstract and Conclusion mention EDID’s operational integration within Copernicus, Section 4.2 does not explicitly discuss one of the main challenges: ensuring long-term maintenance and sustainability of the database under a specific institutional umbrella. Adding details on governance, funding, and community engagement plans would provide valuable insight into how EDID will remain operational and updated over time.
Response: Indeed long-term maintenance is a challenge and we agree that it will be useful to elaborate on this in the discussion and conclusion: "Maintenance and updates of EDID are ensured by JRC in the framework of the Copernicus European Drought Observatory and leveraging also its role as co-chair of the EU Working Group on water scarcity and drought. JRC has already developed a long-term plan based on: the Europe Media Monitor and the use of large language model; the engagement of drought experts that can contribute via a dedicated component of EDID (about to be released) by adding new records."
3) Underrepresented systems: in Section 4.1 (around L306), the manuscript notes that some systems are less frequently represented. It would strengthen the discussion to explicitly name these systems and provide possible reasons for their lower representation (e.g., data availability, reporting practices, sector-specific challenges).
Response: We agree that we can certainly name them and elaborate further on those systems in the revised version. Indeed a bit of interpretation is already in the results section and we will harmonize this in the revision. In brief, the two energy systems and navigation are much more restricted to locations where they can occur, i.e. the infrastructure that can be affected doesn't exist everywhere (e.g. transportation routes, reservoirs). In addition, the impacts on people are less tangible and may be modified by market forces, which will likely affect the degree of awareness and hence reporting. The system ecosystems-terrestrial includes forestry (among other affected ecosystems). Drought response occurs with delay and will be moderated by resistance and resilience of the ecosystems. More research is needed on this system and here monitored vegetation health indicators can provide a useful comparison. We suggest that we will revise this discussion point in this direction and point out some opportunities and needs for further research on these systems.
4) Additional challenges: Section 4.2 could be expanded to include other challenges, such as harmonizing impact reports with demographic or socio-economic factors (e.g., population density), which are critical for interpreting vulnerability and exposure.
Response: Thanks for this important suggestion. Efforts were made to remove duplicates but indeed the consideration for cultural (are drought impacts found worth reporting?), demographic (more people more awareness?) and socioeconomic (more/better news?) is certainly worth discussing and hopefully at some point considering in analyses as well. We agree to add this aspect to the revised discussion.
Specific comments:
Section 2.2 (L158): The text states that “measured and systematically observed data were found to be either weak or incomplete” and thus excluded from EDID. However, operational sources like propluvia were included. Please clarify why propluvia is not considered part of the “measured and systematically observed data” category to avoid confusion.
Response: Thanks for pointing out that this is imprecise and indeed there is no very clear distinction. We suggest to revise the sentence to say 'More ...datasets were considered initially but... '.
Abstract (L22): Consider adding a strong concluding sentence beyond “mobilizing the drought community,” emphasizing the broader significance of EDID for risk management and policy.
Response: Good idea - we will add that in the revised manuscript.
L28: Please specify what is meant by “risk” in contrast to “impact” (e.g., risk as potential impact) for clarity.
Response: We would like to avoid a long discussion on where impacts fit into the risk concepts as there are different opinions about this and we don't want to restrict the interpretation or use of the data for any type of assessment. We suggest to revise: ..."do not directly provide data on observed impacts, nor on vulnerabilities that may contribute to drought risk."
L30: “Impact-oriented” should start with lowercase “i.”
Response: Thanks, this will be corrected.
Figure 1 (L103): Indicate what the left and right screenshots represent for clarity.
Response: Thanks, this will be corrected.
Table 1: Typo: “Bachmaier” should be “Bachmair.”
Response: Thanks, this will be corrected.
Table 1: Please ensure all references listed in Table 1 appear in the Reference section (e.g., Bachmair et al., 2017 seems missing).
Response: Sorry, of course this will be corrected.
L199: Briefly explain what a Sankey diagram is for readers unfamiliar with the term.
Response: Yes, we suggest to revise the text to explain that we visualize the impact data flow from one categorization to another (adding Sankey diagram in parentheses).
Figure 2: Excellent synthesis of complex information. Two suggestions: 1) Consider improving readability by enlarging boxes or arranging text vertically. 2) Fig. would benefit from adding percentage values directly in the figure for “Original Source Type” and “EDID Systems” categories.
Response: We will work on the readability and try if the numbers can fit in for a revised version.
Section 4.1 (L290): Consider revising the title for clarity: “4.1 A new baseline of drought impacts across Europe: spatio-temporal patterns.”
Response: Agreed, this will be changed. Thanks for the suggestion.
Citation: https://doi.org/10.5194/egusphere-2025-4806-AC1
-
AC1: 'Reply on RC2', Kerstin Stahl, 20 Dec 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 394 | 191 | 30 | 615 | 58 | 26 | 23 |
- HTML: 394
- PDF: 191
- XML: 30
- Total: 615
- Supplement: 58
- BibTeX: 26
- EndNote: 23
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Kerstin Stahl
Kathrin Szillat
Veit Blauhut
Monika Hlavsova
Lauro Rossi
Dario Masante
Andrea Toreti
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(5048 KB) - Metadata XML
-
Supplement
(1566 KB) - BibTeX
- EndNote
- Final revised paper