Volumetric Analysis of Reservoirs in Drought-Prone Areas Using Remote Sensing Products
Round 1
Reviewer 1 Report
Overall comments:
This manuscript demonstrates an application of freely available remote sensing products to quantify volumetric fluctuations in reservoirs in the western United States. The authors utilize datasets of processed satellite imagery in conjunction with satellite-derived digital elevation models to develop elevation-area relationships and then estimate volumetric fluctuations for Lake Mead and Lake Powell. Estimated volumetric fluctuations in these two reservoirs are then validated using freely available estimates for these reservoirs. The methodology is then applied to various reservoirs across California and compared with a drought index to explore the potential for drought monitoring.
Overall, this is an interesting manuscript and addresses the increasingly important issue of water scarcity in drought-prone regions such as the western United States. The methodology is technically sound and the results are clear and straightforward. However, there are many grammatical errors throughout the manuscript that need to be corrected. I have made some specific suggestions for obvious mistakes but the whole paper needs to be thoroughly edited prior to publication.
Specific comments:
Line 1: ‘number of dams increased during the 20th century’
Line 3: ‘amid’ or ‘amidst’ or ‘in the midst of’
Line 4: ‘Pertaining to’
Line 40: This is generally true but there are several databases that have lake/reservoir data for developing nations. Maybe you should change ‘restricted to’ to ‘primarily available for’
Line 85: I would say ‘high resolution’ or ‘near-continuous’ here.
Line 96: ‘Dam’
Line 106: ‘out of’
Section 2.2: Need more information about the reservoirs that were studied. How many were used in the example application? What was the average size, range of sizes, etc.?
Section 3.3: This is much coarser spatial resolution than the processed imagery and DEM data. Is there any way to downscale this data and make the results more accurate? This is probably outside the scope of this study but may be worth mentioning as a caveat.
Line 180: What was the dataset that the Median DEM values were taken from? Was it evenly spaced points along the reservoir boundary polygon or elevation values for each grid cell across the reservoir? Need to be more specific here.
Figures 7 and 8: What are the different color datapoints and vertical lines? Need to specify in the figure caption.
Line 270: Need to include units for RMSE (i.e., km3)
Line 271: What was the RMSE for VH estimates?
Line 299: ‘Validated’? Do you mean ‘quantified’?
Line 304: ‘estimates of volumetric fluctuations’
Table 1: What is the ‘GRAND ID’ column? Are these individual reservoirs from the GRanD database? If so, need to list the names of the reservoirs.
Line 361: This could also be attributed to local topography.
Line 363: So if the methodology only worked for 40 out of 158 reservoirs in this study, is it really globally applicable? I think that it could be with a different approach. For reservoirs with nonlinear area-elevation relationships, instead of using equations 3 and 4, you could create a polynomial area-elevation relationship (e.g., WA = a* WH 2 + b* WH + c) and integrate this relationship to develop a volumetric fluctuation relationship. This is an approach that has been used in several altimetry-based studies (e.g., Muala et al., 2014; Keys and Scott, 2018) and I believe it would work well here. Although it’s outside the scope of this study, it would be helpful to mention this here because I think this is an easy way to make this globally applicable.
Line 380: Again I would say ‘high resolution’ instead of ‘continuous’.
Line 390: ‘Water time series’ needs to be more specific.
I think sections 2,3,4 should all be combined into a single ‘Methods’ section.
All figures: numbers should be listed as superscripts (e.g., km2 should be km2)
Author Response
Point 1: Line 1- ‘number of dams increased during the 20th century’
Response 1: Changed accordingly
Point 2 : Line 3- ‘amid’ or ‘amidst’ or ‘in the midst of’
Response 2: Changed accordingly
Point 3: Line 4- ‘Pertaining to’
Response 3: Changed accordingly
Point 4: Line 40- This is generally true but there are several databases that have lake/reservoir data for developing nations. Maybe you should change ‘restricted to’ to ‘primarily available for’
Response 4: Agreed. Changed accordingly
Point 5: Line 85- I would say ‘high resolution’ or ‘near-continuous’ here.
Response 5: I think near continuous would be more appropriate as it highlights the long term continuous temporal availability of the GSW.
Point 6: Line 96- ‘Dam’
Response 6: Changed accordingly.
Point 7: Line 106: ‘out of’
Response 7: Changed accordingly
Point 8: Section 2.2: Need more information about the reservoirs that were studied. How many were used in the example application? What was the average size, range of sizes, etc.?
Response 8: Additional information on reservoirs in California has been added.
Point 9: Section 3.3- This is much coarser spatial resolution than the processed imagery and DEM data. Is there any way to downscale this data and make the results more accurate? This is probably outside the scope of this study but may be worth mentioning as a caveat
Response 9: We acknowledge this point but since drought indices are meteorological variables, downscaling may not always be the optimal solution because spatial extents of droughts are generally large. Geo-hydrological features such as the river basins or water tables can span across wider spatial extents where downscaling may not be useful.
Point 10: Line 180-What was the dataset that the Median DEM values were taken from? Was it evenly spaced points along the reservoir boundary polygon or elevation values for each grid cell across the reservoir? Need to be more specific here.
Response 10: They were taken separately from all the three DEM rasters at the monthly reservoir boundaries. A median was calculated for all the values along reservoir boundaries.
Point 11: Figures 7 and 8-What are the different color datapoints and vertical lines? Need to specify in the figure caption.
Response 11: The vertical lines with darker color-codes indicate the lower standard deviation of the residuals from the regression line (lower RMSE). The lighter color-codes indicate higher deviations and higher RMSE. The captions have been changed with appropriate details.
Point 12: Line 270- Need to include units for RMSE (i.e., km3)
Response 12: Units for RMSE have been included.
Point 13: Line 271- What was the RMSE for VH estimates?
Response 13: RMSE has been added to elevation derived estimates.
Point 14: Line 299- ‘Validated’? Do you mean ‘quantified’?
Response 14: Has been changed to quantified.
Point 15: Line 304- ‘estimates of volumetric fluctuations’
Response 15: Has been changed accordingly
Point 16: Table 1- What is the ‘GRAND ID’ column? Are these individual reservoirs from the GRanD database? If so, need to list the names of the reservoirs.
Response 16: GRanD IDs are now replaced by reservoir/lake names (taken from the GRanD metadata)
Point 17: Line 361- This could also be attributed to local topography.
Response 17: We agree. This additional point has been added.
Point 18: Line 363- So if the methodology only worked for 40 out of 158 reservoirs in this study, is it really globally applicable? I think that it could be with a different approach. For reservoirs with nonlinear area-elevation relationships, instead of using equations 3 and 4, you could create a polynomial area-elevation relationship (e.g., WA = a* WH 2 + b* WH + c) and integrate this relationship to develop a volumetric fluctuation relationship. This is an approach that has been used in several altimetry-based studies (e.g., Muala et al., 2014; Keys and Scott, 2018) and I believe it would work well here. Although it’s outside the scope of this study, it would be helpful to mention this here because I think this is an easy way to make this globally applicable.
Response 18: We acknowledge this. We have mentioned them in the discussion section with corresponding citations.
Point 19: Line 380-Again I would say ‘high resolution’ instead of ‘continuous’.
Response 19: Has been changed accordingly
Point 20: Line 390- ‘Water time series’ needs to be more specific.
Response 20: We were referring to the same GSW-monthly water history product. That has been specified accordingly.
Point 21: I think sections 2,3,4 should all be combined into a single ‘Methods’ section.
Response 21: We have combined sections 3 (Data) and 4 (Methods) together into one 'Materials and Methods' section. However, we think 'Study Area' section should remain as it is. It serves as a good introduction to the geographical locations used in the study and is a useful transition into methodological part.
Point 22: All figures-numbers should be listed as superscripts (e.g., km2 should be km2)
Response 22: All figures have been changed accordingly.
Reviewer 2 Report
Review for the article titled “Volumetric analysis of reservoirs in drought prone areas using open source remote sensing products” by Bhagwat et al. The authors utilized the global surface water dataset in combination with 3-DEMs data to estimate the volumetric areas for Lake Mead, Lake Powell, and reservoirs in California. The methodology relies on an area-elevation hypsometry relationship to calculate the volumetric variation in reservoir storage. Overall, the topic is quite interesting, and the paper is well written. Generating volumetric time series of lake levels is curcial to assess water availability in natural water reservoirs, very good research topic. Still, a few minor comments and the will be ready for publication. Title; use drought-prone. Also, “open source remote sensing”, seems more technical (software/code) related not data, I would prefer the use “remote sensing products” directly. Abstract; L3 and L4: not true, monitoring studies and related programs are there, please rewrite this statement. L5 “Open source”, not quite a good choice of word, change or eliminate. Likewise, elsewhere in the article. Keywords; use more general/non-technical words instead.
Author Response
Point 1: Use drought-prone
Response 1: The title has been changed accordingly.
Point 2: “open source remote sensing”, seems more technical (software/code) related not data, I would prefer the use “remote sensing products” directly
Response 2: Has been changed accordingly.
Point 3: L3 and L4: not true, monitoring studies and related programs are there, please rewrite this statement
Response 3: The statement is removed and replaced.
Point 4: L5 “Open source”, not quite a good choice of word, change or eliminate. Likewise, elsewhere in the article
Response 4: By open source, we meant that these products are freely accessible and therefore cost-effective. The data is not acquired through any commercial portals
Point 5: Keywords; use more general/non-technical words instead.
Response 5: Some terms have been removed and replaced.
Reviewer 3 Report
Utilizing an existing approach (Busker et al. Hydrol. Earth Syst. Sci., 23, 669–690, 2019), temporal variations of the available water in terms of volume were determined for the selected reservoirs within the state of California. Volumetric estimations were based on area-elevation hypsometric relationship, which was achieved by combining water areas from the Global Surface Water (GSW) monthly product and water surface median elevation taken from DEM. Attempts have been made to link drought indicators with changes in historical reservoir water volumes.
The paper is appropriately organised. The methodology has been correctly applied. The interpretation of results and subsequent discussion and conclusions are reasonably clear.
Author Response
We would like to thank the reviewer for the time and effort.
