the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 25 Oct 2024
Automated Rapid Estimation of Flood Depth using Digital Elevation Model and EOS-04 Satellite derived Flood inundation
Abstract. Rapid flood assessment is vital for effective relief, rehabilitation, and flood mitigation strategies. Developing and implementing automated, rapid methods for flood depth and inundation estimation is necessary for near real-time information dissemination. This paper presents an end-to-end automated floodwater delineation and depth estimation process using EOS-04 (RISAT 1A) Synthetic Aperture Radar (SAR) images and a Digital Elevation Model (DEM). Flood inundation is estimated using an Automated Tile-based Segmentation technique. Flood depth is estimated by the Trend Surface Analysis (TSA) method, a novel technique requiring only the inundated water layer and DEM, unlike various hydrodynamic models needing extensive data. This method is applied to most flood prone areas in Andhra Pradesh, Assam, Bihar, and Uttar Pradesh states of India. Water levels estimated at river gauge stations using TSA technique are validated with real-time field measurements and compared with Floodwater Depth Estimation Tool (FwDET) derived results. The TSA technique outperforms FwDET, showing lower RMSE values.
-
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
(1801 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1801 KB) - Metadata XML
- BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-2278', Shewandagn Lemma Tekle, 27 Nov 2024
Dear Authors,
Thank you so much for submitting this manuscript. The manuscript raised a good methodology for deriving flood extent and depth information from SAR images, which are less affected by clouds. The results are useful for preliminary flood management decisions, particularly for non-structural measures like preliminary floodplain zoning and evacuation planning. The results of this study can also benefit the calibration and validation of hydrodynamic models, especially in data-scarce regions. However, there are some points to take into account to make the manuscript more impactful. Please find the comments below for your consideration.
Comments:
- What is the novelty of this paper? Several studies have already been conducted on this topic.
- The methods are purely used to reconstruct historical floods. However, the manuscript becomes more impactful if it could be associated with storm characteristics like storm magnitude and return periods to generalize it for future applications. These will provide the probable extent of a storm event with a specific return period.
- Hydrodynamic models are indeed time- and data-intensive but capable of handling the limitations of the approach raised in this manuscript. This approach is probably effective in flood extent and depth estimation but not sure about its skill on other hydrodynamic characteristics of the flood, e.g., flood velocity. Is it possible to include this property in your analysis?
- It is encouraged to specify the limitations of your results in real-world applications and indicate what kind of flood management decisions can be made confidently. This will provide confidence to end-users.
- In section 4 of the manuscript, results are presented but the discussion part is missing, which is important to connect your results with similar previous studies.
Minor Comments
- Indicate permanent and seasonal water bodies in your study area (if any)
- Improve the quality of Figure 10, legends are not readable
- Put table titles consistently before/after the table
- Correct grammatical and punctuation errors (missing spaces between words, inappropriate use of full stops, colon, etc.)
Citation: https://doi.org/10.5194/egusphere-2024-2278-RC1 -
AC1: 'Reply on RC1', Amani chimata, 29 Nov 2024
Reply to Comments :
We would like to sincerely thank the referee for their valuable comments and constructive suggestions. The feedback provided has been instrumental in improving the clarity and depth of our manuscript.
(1) What is the novelty of this paper? Several studies have already been conducted on this topic.
Reply: Though there are similar works done, the novelty of this work lies in the development and application of a comprehensive, automated framework for floodwater delineation and depth estimation using EOS-04 (RISAT-1A) SAR data and Digital Elevation Models (DEMs). Integrating the Automated Tile-based Segmentation technique with the Trend Surface Analysis (TSA) method eliminates the need for extensive input data required by traditional hydrodynamic models. The TSA method, capable of generating accurate flood surfaces using only inundated water layers and DEMs, is particularly innovative in its ability to adapt to varying spatial trends and elevation changes in large and complex river systems and efficiently perform in highly flood-affected study areas of India. Furthermore, the incorporation of public and fine-resolution DEMs based on terrain type ensures adaptability and precision in both plain and steep areas. The study also addresses common challenges in SAR data analysis, such as hill shadows, by effectively leveraging the HAND tool to eliminate false water areas. Validation of this methodology against field-measured data and comparison with the Floodwater Depth Estimation Tool (FwDET) demonstrates its superior accuracy, with lower RMSE values, highlighting its potential as a robust, efficient, and scalable solution for real-time flood assessment.
(2) The methods are purely used to reconstruct historical floods. However, the manuscript becomes more impactful if it could be associated with storm characteristics like storm magnitude and return periods to generalize it for future applications. These will provide the probable extent of a storm event with a specific return period.
Reply: As noted in the comments, hydrological and hydrodynamic studies often face challenges due to insufficient and unsuitable data across various terrains. Storm and rainfall events can occur in different areas, impacting administrative units with data limitations. The methods in this study focus on reconstructing historical floods through floodwater delineation and depth estimation. Integrating storm characteristics, such as magnitude and return periods, would enhance the framework's ability to predict future flood scenarios. By linking flood extents and depths with diverse storm events, this methodology could provide probabilistic flood maps for specific return periods, improving decision support for disaster preparedness.
(3) Hydrodynamic models are indeed time- and data-intensive but capable of handling the limitations of the approach raised in this manuscript. This approach is probably effective in flood extent and depth estimation but not sure about its skill on other hydrodynamic characteristics of the flood, e.g., flood velocity. Is it possible to include this property in your analysis?
Reply: The presented approach is designed to efficiently estimate flood extent and depth with minimal data requirements, addressing challenges such as SAR data limitations and terrain-specific DEM resolution needs. While hydrodynamic models can simulate additional flood characteristics, such as velocity, they require extensive input data and computational resources. Incorporating flood velocity into this framework is theoretically possible but would require additional data, such as flow rates and channel properties, and potentially adapting or integrating simplified hydrodynamic modelling techniques. Future work could explore combining this methodology with complementary tools or datasets to estimate flood velocity and other dynamic characteristics, enhancing its scope. However, the primary focus of the current approach is on rapid flood assessment with practical applicability, prioritizing efficiency and accessibility over the detailed outputs of traditional hydrodynamic models.
(4) It is encouraged to specify the limitations of your results in real-world applications and indicate what kind of flood management decisions can be made confidently. This will provide confidence to end-users.
Reply: The presented approach has few limitations in real-world applications. First, the accuracy of flood depth estimation is sensitive to the resolution and alignment of the DEM with the flood layer, particularly in steep terrain, where high-resolution DEMs are essential which is taken care during the selection of satellite datasets and derived flood layers. Second, while the methodology effectively delineates flood extent and depth, it does not account for hydrodynamic characteristics such as flood velocity or temporal variations in flood behaviour, however, the information generated through the proposed approach is of great help in real time relief and rehabilitation, rescue operations in the field . Third, the approach performs best in areas with gentle slopes and may other complex terrains can be handled with slope / land use information in a contextual referencing approach.
Despite these limitations, the results are highly reliable for flood extent mapping and depth estimation in plain and moderately sloped regions, enabling decisions such as identifying flood-prone areas, prioritizing evacuation zones, and planning resource allocation for relief and rehabilitation. The rapid and automated nature of the framework makes it suitable for near real-time flood assessment, supporting emergency response efforts. The management decisions, especially during the relief and rehabilitation activities and rescue operations can be made efficiently in terms of deployments of rescue materials like boats/ type boats, and suitably skilled manpower, End-users can confidently use this tool for planning mitigation strategies, such as floodplain zoning and infrastructure protection, while recognizing its constraints in predicting dynamic flood behaviours etc
(5) In section 4 of the manuscript, results are presented but the discussion part is missing, which is important to connect your results with similar previous studies.
Reply: We would like to clarify that the discussion of the results has already been integrated into Section 4. We have included interpretations and comparisons with previous studies throughout the results section to maintain a smooth flow. However, if you feel a separate, more distinct discussion section would improve clarity, we are happy to make revisions accordingly.
Minor Comments :
(1) Indicate permanent and seasonal water bodies in your study area (if any)
Reply : In our study area, we have identified both permanent and seasonal water bodies. These features have been incorporated into our analysis, as they can influence flood behaviour and hydrological patterns. We will ensure to clearly indicate these water bodies in the revised manuscript.
(2) Improve the quality of Figure 10, legends are not readable
Reply : The figure is updated with improved legend (attached as supplement material )
(3) Put table titles consistently before/after the table
Reply : Corrected in the Manuscript and will be updated through revised manuscript.
(4) Correct grammatical and punctuation errors (missing spaces between words, inappropriate use of full stops, colon, etc.)
Reply : Corrected in the Manuscript and will be updated through revised manuscript.
-
RC2: 'Comment on egusphere-2024-2278', Anonymous Referee #2, 27 Dec 2024
The paper presents a method for estimating the depth and extent of floods from SAR imagery. Since SAR data are already widely used in the literature for flood monitoring, it would be helpful for the authors to highlight the differences between their approach and existing ones, clearly highlighting the strengths and any limitations. In addition, it is suggested that the authors describe the method in greater detail and clarity so that it can be easily understood and used by a wider audience. The following are additional specific suggestions for improving the quality of the work:
Line 16: It would be good to make explicit what is meant by the term “lower” and also include numerical performance results for clarity. This would help to better understand the effectiveness of the proposed method.
Lines 40-49: It is recommended to revise the text and punctuation in these lines in order to improve fluency and clarity. Some sentences are indeed a bit complex and could benefit from restructuring.
Lines 110-111: It is necessary to better specify what is meant by “limit within the active channel.” This should be clarified to avoid ambiguity and allow a more precise understanding of the method.
Line 117: The case studies should be described in more detail, including information such as the size of the watersheds and the physical and hydrological characteristics of each. In addition, it would be helpful to add a picture showing the watersheds in relation to the closure sections to enhance visual understanding of the context.
Line 133: It is important to explain the reason why satellite images with different spatial resolution (e.g., CRS and MRS) were used. Also, it would be helpful to clarify what the temporal resolution of acquisition of these images is, especially in relation to the five types of spatial resolution used.
Line 142: It would be helpful to know how many level measurements were extracted from the CWC site. It is suggested that these measurements be reported in a graph or table for clearer and more immediate visualization.
Lines 155-185: The authors should explain in more detail the workflow illustrated in Figure 3. In particular, it would be useful to supplement the figure with a textual description that would allow even readers who are not experts in the field to understand the methodological choices made, as well as how the process was replicated.
Line 274: It is recommended that the Landsat images used to validate the method be introduced in the paragraph devoted to the data used. This would allow for better contextualization of the data and their use in the validation process.
Line 275: It would be appropriate to run the validation on a larger number of dates and create a confusion matrix comparing water and non-water areas. This would allow for a more accurate assessment of method performance. In addition, it would be useful to calculate other performance metrics such as accuracy, precision, and recall to provide a more complete evaluation.
Discussion: The discussion section lacks a comparison with other work in the literature. Authors should highlight the strengths of their method compared to what has been proposed before, pointing out any significant innovations or improvements.
Citation: https://doi.org/10.5194/egusphere-2024-2278-RC2 - AC2: 'Reply on RC2', Amani chimata, 03 Jan 2025
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2024-2278', Shewandagn Lemma Tekle, 27 Nov 2024
Dear Authors,
Thank you so much for submitting this manuscript. The manuscript raised a good methodology for deriving flood extent and depth information from SAR images, which are less affected by clouds. The results are useful for preliminary flood management decisions, particularly for non-structural measures like preliminary floodplain zoning and evacuation planning. The results of this study can also benefit the calibration and validation of hydrodynamic models, especially in data-scarce regions. However, there are some points to take into account to make the manuscript more impactful. Please find the comments below for your consideration.
Comments:
- What is the novelty of this paper? Several studies have already been conducted on this topic.
- The methods are purely used to reconstruct historical floods. However, the manuscript becomes more impactful if it could be associated with storm characteristics like storm magnitude and return periods to generalize it for future applications. These will provide the probable extent of a storm event with a specific return period.
- Hydrodynamic models are indeed time- and data-intensive but capable of handling the limitations of the approach raised in this manuscript. This approach is probably effective in flood extent and depth estimation but not sure about its skill on other hydrodynamic characteristics of the flood, e.g., flood velocity. Is it possible to include this property in your analysis?
- It is encouraged to specify the limitations of your results in real-world applications and indicate what kind of flood management decisions can be made confidently. This will provide confidence to end-users.
- In section 4 of the manuscript, results are presented but the discussion part is missing, which is important to connect your results with similar previous studies.
Minor Comments
- Indicate permanent and seasonal water bodies in your study area (if any)
- Improve the quality of Figure 10, legends are not readable
- Put table titles consistently before/after the table
- Correct grammatical and punctuation errors (missing spaces between words, inappropriate use of full stops, colon, etc.)
Citation: https://doi.org/10.5194/egusphere-2024-2278-RC1 -
AC1: 'Reply on RC1', Amani chimata, 29 Nov 2024
Reply to Comments :
We would like to sincerely thank the referee for their valuable comments and constructive suggestions. The feedback provided has been instrumental in improving the clarity and depth of our manuscript.
(1) What is the novelty of this paper? Several studies have already been conducted on this topic.
Reply: Though there are similar works done, the novelty of this work lies in the development and application of a comprehensive, automated framework for floodwater delineation and depth estimation using EOS-04 (RISAT-1A) SAR data and Digital Elevation Models (DEMs). Integrating the Automated Tile-based Segmentation technique with the Trend Surface Analysis (TSA) method eliminates the need for extensive input data required by traditional hydrodynamic models. The TSA method, capable of generating accurate flood surfaces using only inundated water layers and DEMs, is particularly innovative in its ability to adapt to varying spatial trends and elevation changes in large and complex river systems and efficiently perform in highly flood-affected study areas of India. Furthermore, the incorporation of public and fine-resolution DEMs based on terrain type ensures adaptability and precision in both plain and steep areas. The study also addresses common challenges in SAR data analysis, such as hill shadows, by effectively leveraging the HAND tool to eliminate false water areas. Validation of this methodology against field-measured data and comparison with the Floodwater Depth Estimation Tool (FwDET) demonstrates its superior accuracy, with lower RMSE values, highlighting its potential as a robust, efficient, and scalable solution for real-time flood assessment.
(2) The methods are purely used to reconstruct historical floods. However, the manuscript becomes more impactful if it could be associated with storm characteristics like storm magnitude and return periods to generalize it for future applications. These will provide the probable extent of a storm event with a specific return period.
Reply: As noted in the comments, hydrological and hydrodynamic studies often face challenges due to insufficient and unsuitable data across various terrains. Storm and rainfall events can occur in different areas, impacting administrative units with data limitations. The methods in this study focus on reconstructing historical floods through floodwater delineation and depth estimation. Integrating storm characteristics, such as magnitude and return periods, would enhance the framework's ability to predict future flood scenarios. By linking flood extents and depths with diverse storm events, this methodology could provide probabilistic flood maps for specific return periods, improving decision support for disaster preparedness.
(3) Hydrodynamic models are indeed time- and data-intensive but capable of handling the limitations of the approach raised in this manuscript. This approach is probably effective in flood extent and depth estimation but not sure about its skill on other hydrodynamic characteristics of the flood, e.g., flood velocity. Is it possible to include this property in your analysis?
Reply: The presented approach is designed to efficiently estimate flood extent and depth with minimal data requirements, addressing challenges such as SAR data limitations and terrain-specific DEM resolution needs. While hydrodynamic models can simulate additional flood characteristics, such as velocity, they require extensive input data and computational resources. Incorporating flood velocity into this framework is theoretically possible but would require additional data, such as flow rates and channel properties, and potentially adapting or integrating simplified hydrodynamic modelling techniques. Future work could explore combining this methodology with complementary tools or datasets to estimate flood velocity and other dynamic characteristics, enhancing its scope. However, the primary focus of the current approach is on rapid flood assessment with practical applicability, prioritizing efficiency and accessibility over the detailed outputs of traditional hydrodynamic models.
(4) It is encouraged to specify the limitations of your results in real-world applications and indicate what kind of flood management decisions can be made confidently. This will provide confidence to end-users.
Reply: The presented approach has few limitations in real-world applications. First, the accuracy of flood depth estimation is sensitive to the resolution and alignment of the DEM with the flood layer, particularly in steep terrain, where high-resolution DEMs are essential which is taken care during the selection of satellite datasets and derived flood layers. Second, while the methodology effectively delineates flood extent and depth, it does not account for hydrodynamic characteristics such as flood velocity or temporal variations in flood behaviour, however, the information generated through the proposed approach is of great help in real time relief and rehabilitation, rescue operations in the field . Third, the approach performs best in areas with gentle slopes and may other complex terrains can be handled with slope / land use information in a contextual referencing approach.
Despite these limitations, the results are highly reliable for flood extent mapping and depth estimation in plain and moderately sloped regions, enabling decisions such as identifying flood-prone areas, prioritizing evacuation zones, and planning resource allocation for relief and rehabilitation. The rapid and automated nature of the framework makes it suitable for near real-time flood assessment, supporting emergency response efforts. The management decisions, especially during the relief and rehabilitation activities and rescue operations can be made efficiently in terms of deployments of rescue materials like boats/ type boats, and suitably skilled manpower, End-users can confidently use this tool for planning mitigation strategies, such as floodplain zoning and infrastructure protection, while recognizing its constraints in predicting dynamic flood behaviours etc
(5) In section 4 of the manuscript, results are presented but the discussion part is missing, which is important to connect your results with similar previous studies.
Reply: We would like to clarify that the discussion of the results has already been integrated into Section 4. We have included interpretations and comparisons with previous studies throughout the results section to maintain a smooth flow. However, if you feel a separate, more distinct discussion section would improve clarity, we are happy to make revisions accordingly.
Minor Comments :
(1) Indicate permanent and seasonal water bodies in your study area (if any)
Reply : In our study area, we have identified both permanent and seasonal water bodies. These features have been incorporated into our analysis, as they can influence flood behaviour and hydrological patterns. We will ensure to clearly indicate these water bodies in the revised manuscript.
(2) Improve the quality of Figure 10, legends are not readable
Reply : The figure is updated with improved legend (attached as supplement material )
(3) Put table titles consistently before/after the table
Reply : Corrected in the Manuscript and will be updated through revised manuscript.
(4) Correct grammatical and punctuation errors (missing spaces between words, inappropriate use of full stops, colon, etc.)
Reply : Corrected in the Manuscript and will be updated through revised manuscript.
-
RC2: 'Comment on egusphere-2024-2278', Anonymous Referee #2, 27 Dec 2024
The paper presents a method for estimating the depth and extent of floods from SAR imagery. Since SAR data are already widely used in the literature for flood monitoring, it would be helpful for the authors to highlight the differences between their approach and existing ones, clearly highlighting the strengths and any limitations. In addition, it is suggested that the authors describe the method in greater detail and clarity so that it can be easily understood and used by a wider audience. The following are additional specific suggestions for improving the quality of the work:
Line 16: It would be good to make explicit what is meant by the term “lower” and also include numerical performance results for clarity. This would help to better understand the effectiveness of the proposed method.
Lines 40-49: It is recommended to revise the text and punctuation in these lines in order to improve fluency and clarity. Some sentences are indeed a bit complex and could benefit from restructuring.
Lines 110-111: It is necessary to better specify what is meant by “limit within the active channel.” This should be clarified to avoid ambiguity and allow a more precise understanding of the method.
Line 117: The case studies should be described in more detail, including information such as the size of the watersheds and the physical and hydrological characteristics of each. In addition, it would be helpful to add a picture showing the watersheds in relation to the closure sections to enhance visual understanding of the context.
Line 133: It is important to explain the reason why satellite images with different spatial resolution (e.g., CRS and MRS) were used. Also, it would be helpful to clarify what the temporal resolution of acquisition of these images is, especially in relation to the five types of spatial resolution used.
Line 142: It would be helpful to know how many level measurements were extracted from the CWC site. It is suggested that these measurements be reported in a graph or table for clearer and more immediate visualization.
Lines 155-185: The authors should explain in more detail the workflow illustrated in Figure 3. In particular, it would be useful to supplement the figure with a textual description that would allow even readers who are not experts in the field to understand the methodological choices made, as well as how the process was replicated.
Line 274: It is recommended that the Landsat images used to validate the method be introduced in the paragraph devoted to the data used. This would allow for better contextualization of the data and their use in the validation process.
Line 275: It would be appropriate to run the validation on a larger number of dates and create a confusion matrix comparing water and non-water areas. This would allow for a more accurate assessment of method performance. In addition, it would be useful to calculate other performance metrics such as accuracy, precision, and recall to provide a more complete evaluation.
Discussion: The discussion section lacks a comparison with other work in the literature. Authors should highlight the strengths of their method compared to what has been proposed before, pointing out any significant innovations or improvements.
Citation: https://doi.org/10.5194/egusphere-2024-2278-RC2 - AC2: 'Reply on RC2', Amani chimata, 03 Jan 2025
Peer review completion
Journal article(s) based on this preprint
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 308 | 109 | 225 | 642 | 15 | 26 |
- HTML: 308
- PDF: 109
- XML: 225
- Total: 642
- BibTeX: 15
- EndNote: 26
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Lakshmi Amani Chimata
Suresh Babu Anuvala Setty Venkata
Shashi Vardhan Reddy Patlolla
Durga Rao Korada Hari Venkata
Sreenivas Kandrika
Prakash Chauhan
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(1801 KB) - Metadata XML