the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
River hydraulic modelling with ICEsat-2 land and water surface elevation
Monica Coppo Frias
Suxia Liu
Xingguo Mo
Karina Nielsen
Heidi Randall
Liguang Jiang
Jun Ma
Abstract. Advances in geodetic altimetry instruments are providing more accurate measurements, enabling satellite missions to hand over useful data in narrow rivers and streams. Altimetry missions produce spatially dense land and water surface elevation measurements in remote areas where in-situ data is scarce, that can be combined with hydraulic/hydrodynamic models to simulate water surface elevation and estimate discharge. In this study, we combine ICESat-2 land and water surface elevation measurements with a low-parametrized hydraulic calibration to simulate water surface elevation and discharge without the need for a rainfall-runoff model. ICESat-2 provides an opportunity to map river cross-section geometry very accurately with an along-track resolution of 0.7 m using the ATL03 product. These measurements are combined with the inland water product ATL13 to calibrate a steady-state hydraulic model to retrieve unobserved hydraulic parameters, such as river depth or roughness coefficient. The low-parametrized model together with the assumption of steady-state hydraulics enables the application of a global search algorithm for parameter calibration at a manageable computational cost. The model performance is similar to that reported for highly parametrized models, with a root mean square error of around 0.41 m. With the calibrated model, we can calculate water surface elevation time series at any chainage point at any time of an available satellite pass within the river reach, and estimate discharge from water surface elevation. The discharge estimates are validated with in-situ measurements at two available gauging stations. In addition, we use the calibrated parameters in a full hydrodynamic model simulation resulting in a RMSE of 0.59 m for the entire observation period.
-
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
(13809 KB)
-
Supplement
(59 KB)
-
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(13809 KB) -
Supplement
(59 KB) - BibTeX
- EndNote
- Final revised paper
Journal article(s) based on this preprint
Monica Coppo Frias et al.
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-377', Anonymous Referee #1, 05 Aug 2022
The paper is a good example of how to use ICEsat-2 satellite altimetry data in a hydraulic model built without in situ topographic survey of cross-sections. I found the paper interesting and adapt to the HESS journal. I think it deserves to be published after minor changes.
- In the analysis I have only one main concern that relates to the structure of the model without considering large tributaries. My main concern is the fact that you consider the boundary condition as uniformly distributed, but in fact from Figure 4 it is quite evident that there are at least two significant inflows from the two left tributaries (it is not clear whether the right tributary comes upstream or downstream of Jimay station). How did you simulate these significant inflows?
- Perhaps the organization of the article can be revised to avoid repetition and short paragraphs. In particular, the description of the ICEsat-2 dataset and the study area follow the method, but they are actually mentioned again and again to explain the different steps. Therefore, I think the best solution is to describe the material first (satellite and in situ dataset, study area) and then the method so that the reader is able to understand why the two satellite products and the hydraulic model scheme are used.
SPECIFIC COMMENTS:
- Introduction: references on hydraulic simulations should include at least one of the studies conducted by Domeneghetti et al. (2014, 2015, 2020); references on altimetry densification should include the publication of one of the authors Nielsen et al. (2022).
- Lines 119-125: the numbers are rather arbitrary. Please, justify the reason for these thresholds in the main text (93% for the water occurrence; 15 m from the river center-line; 500 m distance between observations and less than 15 times…).
- Lines 132-138: I'm not sure I understand these lines. What is meant by "the reference water surface elevation of the cross-section changes "? If it is a reference, it should be fixed. And what is meant by "the change in flow rate is added to the corresponding depth of the cross section"? How can a flow discharge be added to a depth? Please, rephrase the sentences so that they are clear.
- Line 141: it is not clear why the two products ATL03 and ATL13 are shifted of 41 cm. Is this explained somewhere in the manuscript? If not, please can the authors add the reason of this bias?
- Lines 142-148: Please, explain the concept better. It is not clear why you are removing the red dots in Figure 3a that correspond to the zero change in discharge, or the dot at 400. In fact, I do not understand the logic of these analyses. Perhaps, they deserve more detail in the text.
- Line 212: remove “the”
- Looking at Figure 5 the shape of some cross-sections looks rather unrealistic (c,d,e,f). The river bottom looks high (shallow) and this could affect all the analysis. Do you have information on the topographic survey of some cross sections that could help to understand how much error is in the bottom estimate?
- Please, define all acronyms: e.g. RHS, UPA, Obj
- It is not clear why the paragraph 2.4.2 is described here and what is the role after. Try to explain why you are using the MIKE Hydro model.
- Line 290-294: For a reader who is not thoroughly familiar with the satellite product, this sentence is difficult to understand. Please try to explain what is the difference between weak and strong beam data. Also, since this is a product feature, I think it can be moved to a methods section.
- Line 296-297: please, add the references for these distances (e.g. cross-section chainage and longitudinal distance.
- Line 320-321 the sentence is not clear. Please, reformulate it.
- Line 333: the biggest errors are in the downstream sections because I think the estimated bottom of the river is too high. Can I see some cross-sections in the stretch from 65000 to 68000?
- Figure 13 a: please specify the Depth coming from Mike 11 and Depth coming from ATL13 in the x-axis and y-axis. Are you able to explain the differences between the simulated and satellite observed WSE in the July-August 2020?
- Lines 384-385: do these studies refer to the same study areas? Please, specify.
- Line 439: please clarify this sentence because the paper does not show any comparison with other satellite missions to be stated that it “performs better than previous altimetry missions”
Citation: https://doi.org/10.5194/egusphere-2022-377-RC1 -
AC1: 'Reply on RC1', Monica Coppo Frias, 29 Sep 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-377/egusphere-2022-377-AC1-supplement.pdf
-
RC2: 'Comment on egusphere-2022-377', Anonymous Referee #2, 09 Sep 2022
This work presents the calibration of a 1D hydraulic model on ICEsat-2 altimetric data, for a river portion with unknown bathymetry-friction and given discharge upstream and downstream, which is an interesting topic. The calibrated parameters lead to fair performances in terms of fit to observed water surface elevations (WSE). The parameter space representation and sensitivity analysis are pertinent.
However the scientific novelty of the study and scientific positioning is not sufficiently clear, the authors omit lots of recent works about inverse hydraulic-hydrological modeling with current altimetric and water extent data, and forthcoming SWOT data.
I recommend the authors thoroughly rework the manuscript, rewrite the introduction and sharpen the analysis of results, reoganize some parts to ease the reading. I provide some elements below.This work relies on :
- ICEsat-2 data, preprocessed using water masks and an effective parameterization (from Dingman) for unobserved low flow cross section bathymetry. (which has been used with altimetry and hydraulic modeling in Bjerklie et al. 2018).
- a 1D steady state Saint-Venant shallow water model with 3 spatially uniform parameters (Manning friction, low flow depth, power shape) used in the calibration process. Rerun with calibrated parameters and unsteady solver in the MIKE platform. (Other 1D hydraulic models are used with altimetric data and effective bathymetry parameterization in references provided bellow).
- a global calibration algorithm and a global sensitivity analysis algorithm, both from literature are used.I require clarifications and improvements on these points :
- No real analysis about the hydraulic inverse problem from satellite data, and of the scientfic difficulties related to it as the "bathymetry-fricion" equifinality (Garambois and Monnier 2015), existing elaborate algorithms including variational data assimilation used for high dimensional calibrations and adapted to satellite hydraulics (cf. Larnier et al. 2020, Garambois et al. 2020 and references therein).
- Model derived rating curves (and even stage fall discharge relationships, with WS slope...) have already been presented and thoroughly analyzed in Malou et al. with a hydraulic model calibrated on altimetry and water extents.
- It is not clear to me how much snapshots of WS are used, what about temporal variability ? Is there any nadir altimetric time series available on this study zone? The only validation regarding temporal variability is thus at gauging stations used as boundary conditions for the hydraulic models?
- Hydraulic analysis are not deep enough, regarding forward-inverse hypothesis and resulting misfit wrt observations and river morphological features, regarding also "WS interpolation at any river point" as mentioned in the flowchart of Fig 1. Detailed discussions about hydraulic extrapolation in altimetry context can be found in Pujol et al. 2020, Malou et al. 2021.
- An algorithm enabling Bayesian uncertainty analysis is used "To study the uncertainty of the model", the uncertainty is provided on infered parameters but not on other estimates in the rest of the paper ; analysis are not deep. What is the sensitivity of the parameter inference to first guess, to pdf choices and other calibration algorithm parameters?
- Is "80-180 meter in low flow season" corresponding to a narrow river ? This corresponds to rivers visible with current nadir altimeters and by the future swot mission, which should be properly discussed with a litterature review.
- Clarify UPA_{x} in Eq. 7.
****Bjerklie, D. M., Birkett, C. M., Rover, J. A., Jones, J. W., Fulton, J. W., Garambois, P.-A.. (2018). « Satellite Remote Sensing Determination of River Discharge: Application to the Yukon River Alaska ». Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2018.04.005
Garambois, P. -A, Monnier, J. (2015). « Inference of effective river properties from remotly sensed observations of water surface ». Advances in Water Ressources. (79) 103-120. doi:10.1016/j.advwatres.2015.02.007
Garambois, P.-A., Larnier, K., Monnier, J., Finaud-Guyot, P., Verley, J., Montazem, A. S., Calmant, S. (2020). « Variational estimation of effective channel and ungauged anabranching river discharge from multi-satellite water heights of different spatial sparsity ». Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2019.124409
Larnier, K., Monnier, J., Garambois, P.-A., Verley, J. (2020) « River discharge and bathymetry estimation from SWOT altimetry measurements ». Inverse Problems in Science and Engineering (IPSE). https://doi.org/10.1080/17415977.2020.1803858
Malou T., P.-A. Garambois, A. Paris, J. Monnier, K. Larnier. (2021) Generation and analysis of stage-fall-discharge laws from coupled hydrological-hydraulic river network model integrating sparse multi-satellite data. Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2021.126993
Pujol, L, Garambois, P.-A., Finaud-Guyot, P., Monnier, J., Larnier, K., Mosé, R., Biancamaria, S., Yésou, H., Moreira, D., Paris, A., Calmant, S. (2020). « Estimation of Multiple Inflows and Effective Channel by Assimilation of Multi-satellite Hydraulic Signatures: Case of the Ungauged Anabranching Negro River ». Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2020.125331
Citation: https://doi.org/10.5194/egusphere-2022-377-RC2 -
AC2: 'Reply on RC2', Monica Coppo Frias, 29 Sep 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-377/egusphere-2022-377-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Monica Coppo Frias, 29 Sep 2022
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2022-377', Anonymous Referee #1, 05 Aug 2022
The paper is a good example of how to use ICEsat-2 satellite altimetry data in a hydraulic model built without in situ topographic survey of cross-sections. I found the paper interesting and adapt to the HESS journal. I think it deserves to be published after minor changes.
- In the analysis I have only one main concern that relates to the structure of the model without considering large tributaries. My main concern is the fact that you consider the boundary condition as uniformly distributed, but in fact from Figure 4 it is quite evident that there are at least two significant inflows from the two left tributaries (it is not clear whether the right tributary comes upstream or downstream of Jimay station). How did you simulate these significant inflows?
- Perhaps the organization of the article can be revised to avoid repetition and short paragraphs. In particular, the description of the ICEsat-2 dataset and the study area follow the method, but they are actually mentioned again and again to explain the different steps. Therefore, I think the best solution is to describe the material first (satellite and in situ dataset, study area) and then the method so that the reader is able to understand why the two satellite products and the hydraulic model scheme are used.
SPECIFIC COMMENTS:
- Introduction: references on hydraulic simulations should include at least one of the studies conducted by Domeneghetti et al. (2014, 2015, 2020); references on altimetry densification should include the publication of one of the authors Nielsen et al. (2022).
- Lines 119-125: the numbers are rather arbitrary. Please, justify the reason for these thresholds in the main text (93% for the water occurrence; 15 m from the river center-line; 500 m distance between observations and less than 15 times…).
- Lines 132-138: I'm not sure I understand these lines. What is meant by "the reference water surface elevation of the cross-section changes "? If it is a reference, it should be fixed. And what is meant by "the change in flow rate is added to the corresponding depth of the cross section"? How can a flow discharge be added to a depth? Please, rephrase the sentences so that they are clear.
- Line 141: it is not clear why the two products ATL03 and ATL13 are shifted of 41 cm. Is this explained somewhere in the manuscript? If not, please can the authors add the reason of this bias?
- Lines 142-148: Please, explain the concept better. It is not clear why you are removing the red dots in Figure 3a that correspond to the zero change in discharge, or the dot at 400. In fact, I do not understand the logic of these analyses. Perhaps, they deserve more detail in the text.
- Line 212: remove “the”
- Looking at Figure 5 the shape of some cross-sections looks rather unrealistic (c,d,e,f). The river bottom looks high (shallow) and this could affect all the analysis. Do you have information on the topographic survey of some cross sections that could help to understand how much error is in the bottom estimate?
- Please, define all acronyms: e.g. RHS, UPA, Obj
- It is not clear why the paragraph 2.4.2 is described here and what is the role after. Try to explain why you are using the MIKE Hydro model.
- Line 290-294: For a reader who is not thoroughly familiar with the satellite product, this sentence is difficult to understand. Please try to explain what is the difference between weak and strong beam data. Also, since this is a product feature, I think it can be moved to a methods section.
- Line 296-297: please, add the references for these distances (e.g. cross-section chainage and longitudinal distance.
- Line 320-321 the sentence is not clear. Please, reformulate it.
- Line 333: the biggest errors are in the downstream sections because I think the estimated bottom of the river is too high. Can I see some cross-sections in the stretch from 65000 to 68000?
- Figure 13 a: please specify the Depth coming from Mike 11 and Depth coming from ATL13 in the x-axis and y-axis. Are you able to explain the differences between the simulated and satellite observed WSE in the July-August 2020?
- Lines 384-385: do these studies refer to the same study areas? Please, specify.
- Line 439: please clarify this sentence because the paper does not show any comparison with other satellite missions to be stated that it “performs better than previous altimetry missions”
Citation: https://doi.org/10.5194/egusphere-2022-377-RC1 -
AC1: 'Reply on RC1', Monica Coppo Frias, 29 Sep 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-377/egusphere-2022-377-AC1-supplement.pdf
-
RC2: 'Comment on egusphere-2022-377', Anonymous Referee #2, 09 Sep 2022
This work presents the calibration of a 1D hydraulic model on ICEsat-2 altimetric data, for a river portion with unknown bathymetry-friction and given discharge upstream and downstream, which is an interesting topic. The calibrated parameters lead to fair performances in terms of fit to observed water surface elevations (WSE). The parameter space representation and sensitivity analysis are pertinent.
However the scientific novelty of the study and scientific positioning is not sufficiently clear, the authors omit lots of recent works about inverse hydraulic-hydrological modeling with current altimetric and water extent data, and forthcoming SWOT data.
I recommend the authors thoroughly rework the manuscript, rewrite the introduction and sharpen the analysis of results, reoganize some parts to ease the reading. I provide some elements below.This work relies on :
- ICEsat-2 data, preprocessed using water masks and an effective parameterization (from Dingman) for unobserved low flow cross section bathymetry. (which has been used with altimetry and hydraulic modeling in Bjerklie et al. 2018).
- a 1D steady state Saint-Venant shallow water model with 3 spatially uniform parameters (Manning friction, low flow depth, power shape) used in the calibration process. Rerun with calibrated parameters and unsteady solver in the MIKE platform. (Other 1D hydraulic models are used with altimetric data and effective bathymetry parameterization in references provided bellow).
- a global calibration algorithm and a global sensitivity analysis algorithm, both from literature are used.I require clarifications and improvements on these points :
- No real analysis about the hydraulic inverse problem from satellite data, and of the scientfic difficulties related to it as the "bathymetry-fricion" equifinality (Garambois and Monnier 2015), existing elaborate algorithms including variational data assimilation used for high dimensional calibrations and adapted to satellite hydraulics (cf. Larnier et al. 2020, Garambois et al. 2020 and references therein).
- Model derived rating curves (and even stage fall discharge relationships, with WS slope...) have already been presented and thoroughly analyzed in Malou et al. with a hydraulic model calibrated on altimetry and water extents.
- It is not clear to me how much snapshots of WS are used, what about temporal variability ? Is there any nadir altimetric time series available on this study zone? The only validation regarding temporal variability is thus at gauging stations used as boundary conditions for the hydraulic models?
- Hydraulic analysis are not deep enough, regarding forward-inverse hypothesis and resulting misfit wrt observations and river morphological features, regarding also "WS interpolation at any river point" as mentioned in the flowchart of Fig 1. Detailed discussions about hydraulic extrapolation in altimetry context can be found in Pujol et al. 2020, Malou et al. 2021.
- An algorithm enabling Bayesian uncertainty analysis is used "To study the uncertainty of the model", the uncertainty is provided on infered parameters but not on other estimates in the rest of the paper ; analysis are not deep. What is the sensitivity of the parameter inference to first guess, to pdf choices and other calibration algorithm parameters?
- Is "80-180 meter in low flow season" corresponding to a narrow river ? This corresponds to rivers visible with current nadir altimeters and by the future swot mission, which should be properly discussed with a litterature review.
- Clarify UPA_{x} in Eq. 7.
****Bjerklie, D. M., Birkett, C. M., Rover, J. A., Jones, J. W., Fulton, J. W., Garambois, P.-A.. (2018). « Satellite Remote Sensing Determination of River Discharge: Application to the Yukon River Alaska ». Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2018.04.005
Garambois, P. -A, Monnier, J. (2015). « Inference of effective river properties from remotly sensed observations of water surface ». Advances in Water Ressources. (79) 103-120. doi:10.1016/j.advwatres.2015.02.007
Garambois, P.-A., Larnier, K., Monnier, J., Finaud-Guyot, P., Verley, J., Montazem, A. S., Calmant, S. (2020). « Variational estimation of effective channel and ungauged anabranching river discharge from multi-satellite water heights of different spatial sparsity ». Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2019.124409
Larnier, K., Monnier, J., Garambois, P.-A., Verley, J. (2020) « River discharge and bathymetry estimation from SWOT altimetry measurements ». Inverse Problems in Science and Engineering (IPSE). https://doi.org/10.1080/17415977.2020.1803858
Malou T., P.-A. Garambois, A. Paris, J. Monnier, K. Larnier. (2021) Generation and analysis of stage-fall-discharge laws from coupled hydrological-hydraulic river network model integrating sparse multi-satellite data. Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2021.126993
Pujol, L, Garambois, P.-A., Finaud-Guyot, P., Monnier, J., Larnier, K., Mosé, R., Biancamaria, S., Yésou, H., Moreira, D., Paris, A., Calmant, S. (2020). « Estimation of Multiple Inflows and Effective Channel by Assimilation of Multi-satellite Hydraulic Signatures: Case of the Ungauged Anabranching Negro River ». Journal of Hydrology, https://doi.org/10.1016/j.jhydrol.2020.125331
Citation: https://doi.org/10.5194/egusphere-2022-377-RC2 -
AC2: 'Reply on RC2', Monica Coppo Frias, 29 Sep 2022
The comment was uploaded in the form of a supplement: https://egusphere.copernicus.org/preprints/2022/egusphere-2022-377/egusphere-2022-377-AC2-supplement.pdf
-
AC2: 'Reply on RC2', Monica Coppo Frias, 29 Sep 2022
Peer review completion
Journal article(s) based on this preprint
Monica Coppo Frias et al.
Monica Coppo Frias et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 529 | 208 | 15 | 752 | 28 | 5 | 9 |
- HTML: 529
- PDF: 208
- XML: 15
- Total: 752
- Supplement: 28
- BibTeX: 5
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
The requested preprint has a corresponding peer-reviewed final revised paper. You are encouraged to refer to the final revised version.
- Preprint
(13809 KB) - Metadata XML
-
Supplement
(59 KB) - BibTeX
- EndNote
- Final revised paper