the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
Benchmarking multimodel terrestrial water storage seasonal cycle against GRACE observations over major global river basins
Sadia Bibi
Ashraf Rateb
Bridget R. Scanlon
Muhammad Aqeel Kamran
Abdelrazek Elnashar
Ali Bennour
Ci Li
Abstract. The increasing reliance on global models for evaluating climate and human-induced impacts on the hydrological cycle underscores the importance of assessing their reliability. Hydrological models provide valuable data on ungagged river basins or basins with limited gauge networks. The objective of this study was to evaluate the reliability of 13 global models using the Gravity Recovery and Climate Experiment (GRACE) satellites total water storage (TWS) seasonal cycle for 29 river basins in different climate zones. Results show that the simulated seasonal total water storage change (TWSC) does not compare well with GRACE even in basins within the same climate zone. The models overestimated the seasonal amplitude in most boreal basins and underestimated it in tropical, arid, and temperate zones. In cold basins, the modeled phase of TWSC precedes that of GRACE by up to 2–3 months. However, it lags the GRACE phase by one month over temperate, arid to semi-arid basins. There was good agreement between GRACE and model amplitudes in the tropical zone. With the findings and analysis, we concluded that R2 models with optimized parametrizations have a better correlation with GRACE than the reverse scenario. This signifies an enhancement in the predictive capability of models regarding the variability of TWSC. The seasonal amplitude and phase-difference analysis in this study provide new insights into the future improvement of large-scale hydrological models and TWS investigations.
- Preprint
(6273 KB) - Metadata XML
-
Supplement
(7006 KB) - BibTeX
- EndNote
Sadia Bibi et al.
Status: final response (author comments only)
-
RC1: 'Comment on hess-2023-163', Anonymous Referee #1, 24 Aug 2023
The authors compared the phase and amplitude from models (GHMs and LSMs) and GRACE in global major river basins, with a focus on the differences in different climate zones. This study is a good supplementary to previously published works by Bridget. It provides more information on how GRACE differs from GHMs/LSMs globally. The major limitation of this study is that too much comparison for individual river basins but few in-depth analysis of how these differences come from. My detailed comments include:
- The comparison is mainly implemented for different climate zones. However, as we known, GWS may dominate TWS in many regions, thus the comparison between the models and GRACE is better to be divided into the models with and without groundwater simulations.
- More investigations on the water cycle (e.g., P, ET and Q) and storage (e.g., GWS, SMS) compartments will be helpful for a better understanding on the reasons why GRACE and models differ from each other in the aspects of phase and amplitude.
- Figure 1-4: what the map colors mean?
- Table 1: statistical information on the phase and amplitude derived from models and GRACE needs to be provided as they are the key information. It can be included in Table 1 or summarized in another table.
- Line 83: Full names are need for the abbreviation for R1 and R2 at its first time.
Citation: https://doi.org/10.5194/hess-2023-163-RC1 -
AC1: 'Reply on RC1', Tingju Zhu, 08 Sep 2023
Comment: The comparison is mainly implemented for different climate zones. However, as we known, GWS may dominate TWS in many regions, thus the comparison between the models and GRACE is better to be divided into the models with and without groundwater simulations.
Reply: Thank you for your valuable feedback. We appreciate your input and have taken your suggestion into consideration. As you rightly pointed out, groundwater storage (GWS) can indeed dominate total water storage (TWS) in many regions, like basins which can significantly impact the comparison between model simulations and GRACE data.
In response to your suggestion, we have divided our comparison over five river basins with major underlying aquifers (Congo, Amazon, Orinoco, Ganges-Brahmaputra, and California) into models with and without groundwater simulations. Please see Figure 5 and explanations in the Supplement. The comparison indicates that including a groundwater compartment in GHM/LSM models can apparently improve the presentation of water storage dynamics.
Comment: More investigations on the water cycle (e.g., P, ET and Q) and storage (e.g., GWS, SMS) compartments will be helpful for a better understanding on the reasons why GRACE and models differ from each other in the aspects of phase and amplitude.
Response: We sincerely appreciate your valuable suggestion regarding additional investigations into the water cycle components and storage compartments. Your comments are certainly insightful.
While we understand the potential benefits of delving deeper into parameters such as precipitation (P), evapotranspiration (ET), and streamflow (Q), as well as storage elements like groundwater storage (GWS) and soil moisture storage (SMS), we would like to express our perspective on this matter.
Considering the scope and objectives of the current study, we aimed to focus on specific aspects, i.e. seasonal total water storage dynamics, to maintain clarity in the analysis. Expanding the investigation to include a detailed examination of all these water cycle variables might introduce a level of complexity that could potentially divert attention from the primary research questions. However, we acknowledge that future studies dedicated specifically to the water cycle components and storage compartments could certainly enhance our understanding of the discrepancies between GRACE and the models.
While we aligned with the significance of your suggestion, we believe it would be best suited for a separate investigation, perhaps as an extension of this work or as an independent study. Your guidance is immensely valuable, and we will definitely consider your input for future research endeavors.
Comment: Figure 1-4: what the map colors mean?
Response: We provided the description of base map in captions of Figures 1-4. Base map represents KGClim Climate Zones classification.
Comment: Table 1: Statistical information on the phase and amplitude derived from models and GRACE needs to be provided as they are the key information. It can be included in Table 1 or summarized in another table.
Response: The statistical analysis on amplitude derived from models and GRACE is provided in Table 2 and phase analysis is provided in Figure 5.
Comment: Line 83: Full names are need for the abbreviation for R1 and R2 at its first time.
Response: Full names of R1 and R2 are provided; please see the Supplement (Water Resource Reanalysis tier-1 (R1) and tier-2 (R2)).
-
RC2: 'Comment on hess-2023-163', Anonymous Referee #2, 08 Sep 2023
This study compares the performance of 13 GHMs and LSMs in capturing amplitude and phase of TWSC in global major rivers against GRACE data, including comparisons across climate zones and model version (R1 and R2). This detailed comparison facilitates improved parameterization model process. However, limitations of this study include overly detailed descriptions of the comparisons of different basins so that it is difficult for the reader to get to the point, and the figure lacks summarization.
Major comments:
- The configuration of the modules of each model should be clearly stated, and some differences may be due to missing modules, e.g., snow, permafrost, groundwater, etc., which would also be useful for analyzing the causes of deviations. And, if key modules are missing does it still make sense to compare changes in TWSC between the model (GHMs and LSMs) and GRACE.
- This study did not analyze in depth the causes of amplitude and phase differences, especially 4.1 section
- Line67, “due to human intervention and climate change respectively”, the underestimation is due to anthropogenic interventions and climate change, doesn't that have anything to do with model performance, shouldn't model performance be the main reason?
- Line 89, amplitude and phase of “polar” zone was not analyzed in result section
- Line 139, Why not CSR and JPL on average?
- The difference in the length of the text in parts 3.1 and 3.2 is too large. 3.1 section over-emphasis on basin comparisons.
- The figures are not summarizing enough, too many similar comparisons, e.g., I think Figures 5-8 should be in the Appendix, and the main results should be put in the main text, e.g., the overall results for the different climatic zones in one fig.
- I suggest to add the spatial distribution map of biases in amplitude and phase.
- Figure 1-4 suggests the addition of lines for the GHM and LSM model averages, which facilitates comparison of the two types of models
Minor comments:
- Line 4, “(e.g., the amount and” misses the corresponding right parentheses.
- Line68, “Other studies focused on the seasonal cycle of TWSC to identify" to “Other studies on the seasonal cycle of TWSC focus on identifying” is more suitable? “disparities”, specifically what are the disparities?
- Line 75, “northern basins” is vague, please specifically point
- Line 84, “replicate water storage against the latest release (RL06) of GRACE TWSC.”, this sentence indicates the result? this place is to say what is to be studied
-
CC1: 'Reply on RC2', Sadia Bibi, 15 Sep 2023
This study compares the performance of 13 GHMs and LSMs in capturing amplitude and phase of TWSC in global major rivers against GRACE data, including comparisons across climate zones and model version (R1 and R2). This detailed comparison facilitates improved parameterization model process. However, limitations of this study include overly detailed descriptions of the comparisons of different basins so that it is difficult for the reader to get to the point, and the figure lacks summarization.
Major comments:
- The configuration of the modules of each model should be clearly stated, and some differences may be due to missing modules, e.g., snow, permafrost, groundwater, etc., which would also be useful for analyzing the causes of deviations. And, if key modules are missing does it still make sense to compare changes in TWSC between the model (GHMs and LSMs) and GRACE.
Response: Thank you for your valuable feedback on our manuscript. We appreciate your thoughtful comments and suggestions. In response to your specific points:
We agree that providing a clear and comprehensive description of the configuration of the modules in each model is essential. In our revised manuscript, we included details of key modules within each model in a table, including any specific modules for snow, groundwater, and other relevant components. This will help readers understand the differences and similarities between the models and their potential impact on simulation results. We acknowledge the importance of considering the potential impact of missing key modules in the GHMs and LSMs. While some differences between simulations of the models and GRACE may indeed be due to these missing modules, we believe that the comparison still holds value. Because we aim to assess the agreement and discrepancies between the models and GRACE in terms of Total Water Storage Changes (TWSC) for a better understanding of the limitations of this approach. By highlighting the missing modules in section 4.1, we provided insights into the potential sources of deviations and uncertainties in TWSC estimates.
- This study did not analyze in depth the causes of amplitude and phase differences, especially 4.1 section.
Response: Thank you for your valuable feedback. We appreciate your suggestion to delve deeper into the causes of amplitude and phase differences in Section 4.1. In response to this comment, we expanded the discussion in Section 4.1 and provided a more comprehensive analysis of the factors contributing to the observed amplitude and phase differences.
- Line67, “due to human intervention and climate change respectively”, the underestimation is due to anthropogenic interventions and climate change, doesn't that have anything to do with model performance, shouldn't model performance be the main reason?
Response: Thank you for your comment. We rephrase the lines as “Compared to GRACE-derived TWS trends, Scanlon et al. (2018) revealed that the TWS trends of GHMs were either underestimated or had the opposite sign over numerous basins across the globe”
- Line 89, amplitude, and phase of “polar” zone was not analyzed in result section.
Response: Thank you for your valuable comment. In this study, we primarily concentrated on analyzing the boreal, temperate, arid, and tropical zones, we did not include the polar zone in our analysis. However, we believe that exploring the amplitude and phase of the polar zone could indeed be a valuable avenue for future research to provide a more comprehensive understanding of the subject matter. We will duly consider this suggestion for future studies in this field.
- Line 139, Why not CSR and JPL on average?
Response: We appreciate your comments and the opportunity to clarify our choice of using GRACE data from two data processing centers rather than utilizing the average. It is common in the field of Earth sciences to use data from multiple sources, and it is often encouraged to include data from different processing centers to account for potential biases and uncertainties in the measurements. Our decision to use GRACE data from two processing centers was made to enhance the robustness and reliability of our findings and better capture regional variations. We believe this approach aligns with best practices in the field and contributes to the scientific rigor of our study.
- The difference in the length of the text in parts 3.1 and 3.2 is too large. 3.1 section over-emphasis on basin comparisons.
Response: Thank you for the reviewer's comment regarding the difference in the length of text between sections 3.1 and 3.2. We appreciate your feedback, and we will work to ensure a more balanced and consistent presentation of information in these sections.
To address this concern, we will review and revise Section 3.1 to ensure that it does not over-emphasize basin comparisons and that it aligns more evenly with Section 3.2 in terms of content length. This will help maintain a better structural balance and coherence in the paper while providing equal attention to all relevant aspects of the study.
Your input is valuable, and we will improve the overall flow and readability of our manuscript.
- The figures are not summarizing enough, too many similar comparisons, e.g., I think Figures 5-8 should be in the Appendix, and the main results should be put in the main text, e.g., the overall results for the different climatic zones in one fig.
Response: Thank you for the reviewer's comment regarding the figures in our manuscript. We understand your concern about the number of comparisons and the desire for a more concise summarization. However, we believe that Figures 5-8 are important for understanding the detailed results and patterns in different regions and should remain in the main text.
To address your suggestion for a more concise presentation of the overall results for different climatic zones, we will work on improving the clarity of the figures and their captions to ensure that readers can easily grasp the key findings. This will help strike a balance between providing detailed regional information and presenting a clear overview of the main results.
We appreciate your feedback and are committed to enhancing the presentation of our results to make them more accessible to readers while preserving the important details provided by these figures.
- I suggest to add the spatial distribution map of biases in amplitude and phase.
Response: Thank you for your suggestion to include spatial distribution maps of biases in amplitude and phase. We understand the importance of visualizing these biases for a comprehensive understanding of the results. However, we want to clarify that such maps have already been provided by Schellekens et al. (2017), and our study relies on their analysis in this regard. Including redundant maps in our paper would indeed be repetitive and not add significant new insights to the existing literature.
We appreciate your concern, and to ensure clarity in our paper, we will explicitly reference and acknowledge the work of Schellekens et al. (2017) for the spatial distribution maps of biases in amplitude and phase between the models and GRACE data. This will help readers access the relevant information in the cited source while maintaining the focus of our study on its unique contributions and analyses.
Schellekens, J., Dutra, E., Martínez-de la Torre, A., Balsamo, G., van Dijk, A., Sperna Weiland, F., Minvielle, M., Calvet, J.-C., Decharme, B., Eisner, S., Fink, G., Flörke, M., Peßenteiner, S., van Beek, R., Polcher, J., Beck, H., Orth, R., Calton, B., Burke, S., Dorigo, W., and Weedon, G. P.: A global water resources ensemble of hydrological models: the eartH2Observe Tier-1 dataset, Earth Syst Sci Data, 9, 389–413, https://doi.org/10.5194/essd-9-389-2017, 2017.
- Figure 1-4 suggests the addition of lines for the GHM and LSM model averages, which facilitates comparison of the two types of models
Response: Thank you for your feedback regarding Figure 1-4. We appreciate your suggestion to add lines for the GHM and LSM model averages to facilitate a clearer comparison between the two types of models. We have now incorporated these lines into the figures as per your recommendation. This enhancement should provide readers with a more comprehensive view of the model comparisons and improve the overall clarity of the presentation.
Minor comments:
- Line 4, “(e.g., the amount and” misses the corresponding right parentheses.
Response: We appreciate your suggestion and added parenthesis in the revised manuscript.
- Line68, “Other studies focused on the seasonal cycle of TWSC to identify" to “Other studies on the seasonal cycle of TWSC focus on identifying” is more suitable? “disparities”, specifically what are the disparities?
Response: Thank you for your suggestion. We have revised the sentence as follows: "Other studies on the seasonal cycle of TWSC, such as Zhang et al. (2017), have focused on identifying disparities." The term "disparities" refers to differences or variations in four global numerical model realizations that simulate the continental branch of the global water cycle and GRACE that have been investigated in previous studies.
- Line 75, “northern basins” is vague, please specifically point
Response: Thank you for the suggestion. We have made the requested clarification in the manuscript. Line 75 now reads, "northern high-altitude basins," to provide a more specific description of the geographic region being referred to. This should help eliminate any ambiguity and ensure a clearer understanding for the readers.
- Line 84, “replicate water storage against the latest release (RL06) of GRACE TWSC.”, this sentence indicates the result? this place is to say what is to be studied
Response: We appreciate your feedback. We rephrase the sentence to “Compare high-resolution and more optimally structured R2 models against R1 models and access their ability to simulate TWSC variability and replicate water storage against GRACE TWSC.
Citation: https://doi.org/10.5194/hess-2023-163-CC1
-
RC3: 'Comment on hess-2023-163', Anonymous Referee #3, 19 Sep 2023
Bibi et al. evaluated the reliability of 13 global models using the GRACE TWS for 29 river basins. They conclude that the modeled TWS does not compare well with the GRACE TWS. Authors analyzed amplitude and phase-difference and performed the comparisons based on 5 climate zones (Polar, boreal, temperate, arid, and tropical), 2 set of hydrological model types (LSM and GHM), and 2 sets of R1 and R2 model types. The authors find that R2 models have better correlations with GRACE than R1 models. Though this study provides new insights into the future improvement of large-scale hydrological models, there are some major concerns in this study. By addressing these concerns, the manuscript will better align with the standards of the HESS and provide a more compelling and novel contribution to the field. Please find my detailed review below-
Line 22- It would be easier for readers to understand if the meaning of the term ‘R2 models’ is provided here.
R1 and R2 models are Water Resource Reanalysis tier-1 and tier-2 products which provide a large set of LSMs and GHMs.
R1: 0.5° forced with ERA-Interim data (WFDEI) meteorological reanalysis dataset
R2: 0.25° forced with Multi-Source Weighted Ensemble Precipitation (MSWEP) dataset
Lines 91-102: The authors have used only JPL-M and CSR-M solutions, why not GSFC Mascons as well? The authors did not provide the reason behind using linear interpolation of GRACE TWS data.
Lines 109-110: Please correct the sentence.
Why only amplitude and phase of seasonal cycle of TWS was checked in this study? Why not the trend in the TWS data?
Lines 138-139: Why only GRACE CSR_M seasonal cycle was used to validate the model results? As indicated above why GRACE JPL-M data was not used? Or the mean of the two datasets?
Line 375-: The causes of discrepancies in seasonal amplitudes and phase between models and GRACE TWSC provided in section 4.1 are without any reference. There is no analysis shown to backup the claim. For example, how do the authors know that Model Parameterization is causing the difference in GRACE and model TWS data without doing any analysis and citing any literature? If it is well known then what is the contribution of this study?
Scanlon et al., (2018) already compared the model TWS trends against the GRACE TWS datasets. What are the novel contributions here? Please state them clearly.
Scanlon, B. R., Zhang, Z., Save, H., Sun, A. Y., Müller Schmied, H., Van Beek, L. P., ... & Bierkens, M. F. (2018). Global models underestimate large decadal declining and rising water storage trends relative to GRACE satellite data. Proceedings of the National Academy of Sciences, 115(6), E1080-E1089.
Citation: https://doi.org/10.5194/hess-2023-163-RC3
Sadia Bibi et al.
Sadia Bibi et al.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 371 | 98 | 22 | 491 | 29 | 8 | 9 |
- HTML: 371
- PDF: 98
- XML: 22
- Total: 491
- Supplement: 29
- BibTeX: 8
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1