the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 13 May 2026
C-CWatM v1.0: A high-resolution water resources and river routing model enabling direct linkage to state-of-the-art Earth-system and land‑surface models
Abstract. River routing and human water management are often poorly represented in many Earth-system and land-surface models, not permitting consistent assessments of human-water-climate interactions. In this work, we introduce C-CWatM v1.0 (Climate-CWatM v1.0), a land-surface-driven version of the Community Water model (CWatM) that enables online and offline simulations of river routing, water resource availability, and management. The model operates on standard land-surface output and includes an OASIS3-MCT coupling interface, enabling efficient two-way coupling with Earth-system and land-surface models. In comparison to CWatM, all modules related to land-surface and snow processes have been removed to prevent conflicts with land-surface parameterisations implemented in the coupled models. C-CWatM also offers options for reducing-forcing requirements and simple on-the-fly bias correction using predefined quantile weights. We test C-CWatM in both online and offline settings and evaluate the model performance in offline mode using REMO output across a European domain. C-CWatM reproduces large-scale discharge patterns and hydrological gradients when using non-bias-corrected forcing data without calibration of model parameters. Calibration results in moderate improvements in model performance, while quantile-mapping-based bias correction of runoff significantly enhances model skill. A reduced-forcing version of the model enables rapid simulations based on available climate model output, as demonstrated using readily available EURO-CORDEX output. C-CWatM provides a novel and flexible hydrological and water resources modelling tool for representing river routing and water management in coupled modelling systems, enabling more integrated analyses of climate-water-human interactions.
- Preprint
(9644 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 06 Aug 2026)
- RC1: 'Comment on egusphere-2026-2366', Anonymous Referee #1, 03 Jul 2026 reply
-
RC2: 'Comment on egusphere-2026-2366', Anonymous Referee #2, 07 Jul 2026
reply
Overall assessment
This manuscript presents C-CWatM v1.0, a land-surface-driven version of CWatM designed for offline and online coupling with land surface and Earth system models through an OASIS interface. The manuscript is generally well written, well organized, and the software appears carefully implemented. I believe the model will be useful for researchers interested in coupling existing hydrological routing and water-management capabilities with climate models.
Although I believe this is a solid model development paper that fits GMD's scope, I have significant concerns regarding the novelty and positioning of the work. Much of the hydrological functionality—including river routing, groundwater, reservoirs, lakes, and sectoral water demand—is inherited directly from CWatM, while the principal additions consist of software engineering developments (coupling interface, simplified forcing requirements, calibration workflow, and bias-correction utilities). These are valuable developments, but the manuscript frequently presents them as filling major scientific gaps in Earth system modeling. In my opinion, the presented evidence does not fully support these claims.
Consequently, I encourage the authors to restructure the manuscript to better distinguish between software engineering contributions and scientific advances, and to moderate several statements regarding novelty and gap filling.
Major comments:
1. The manuscript overstates the gap in current ESMs
The Introduction argues that Earth system models generally lack realistic representations of human water management. However, the manuscript itself cites numerous recent examples (e.g., CESM, E3SM, MIROC, CNRM, ORCHIDEE) that already include irrigation, water withdrawals, reservoirs, or groundwater abstractions. Therefore, the remaining gap appears to be the lack of a flexible, reusable coupling framework rather than an absence of human water management itself.
2. Limited validation of human water management components
Although irrigation, groundwater abstraction, reservoir operations, and sectoral water demand are central motivations for the model, the evaluation focuses almost exclusively on river discharge.
The manuscript states that suitable benchmark datasets are unavailable for validating water demand. I do not fully agree with this statement. Several datasets are now available that could provide valuable first-order benchmarks, despite their own uncertainties. For example:
- Huang et al. (2018, HESS) provides globally gridded monthly sectoral water withdrawal estimates that are widely used in the global hydrological modeling community.
- The recently published IRRMIP simulations (Yao et al., 2025) provide harmonized irrigation diagnostics from seven state-of-the-art Earth System Models, offering an excellent benchmark for evaluating irrigation-related processes in a framework specifically designed for coupled climate-water studies.
While none of these datasets represents absolute truth, comparison against independent products would substantially strengthen the manuscript and provide direct evidence supporting the paper’s central claims regarding improved representation of human water management.
3. Address Offline Evaluation Limitations
The authors tested the uncalibrated/calibrated and non-bias-corrected/bias-corrected model against observed discharge in Europe. This is a validation approach for an offline hydrological model, but it tells the reader very little about how stable or accurate the model will be when run dynamically inside an ESM, which is the stated purpose of C-CWatM. I suggest the discussion to explicitly address this limitation of an offline-only validation for a tool built for online coupling.
References:
Huang, Z., Hejazi, M., Li, X., Tang, Q., Vernon, C., Leng, G., Liu, Y., Döll, P., Eisner, S., Gerten, D., Hanasaki, N., and Wada, Y.: Reconstruction of global gridded monthly sectoral water withdrawals for 1971–2010 and analysis of their spatiotemporal patterns, Hydrol. Earth Syst. Sci., 22, 2117–2133, https://doi.org/10.5194/hess-22-2117-2018, 2018
Yao, Y., Thiery, W., Ducharne, A. et al. Irrigation-induced land water depletion aggravated by climate change. Nat Water 3, 1424–1435 (2025). https://doi.org/10.1038/s44221-025-00529-1
Citation: https://doi.org/10.5194/egusphere-2026-2366-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 329 | 142 | 17 | 488 | 25 | 24 |
- HTML: 329
- PDF: 142
- XML: 17
- Total: 488
- BibTeX: 25
- EndNote: 24
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Main comments
The manuscript presents a novel framework that couples CWatM, which represents river routing, human water demand, and water resources management, with Earth system and land surface models through the OASIS3-MCT coupling interface. However, several aspects of the experimental design require further clarification, and the overall organization of the manuscript could be significantly improved. In addition, the Discussion should provide a more comprehensive assessment of the current limitations of the proposed framework and possible directions for future model development.
Detailed comments:
Abstract
Line 5: Define OASIS3-MCT at it first occur.
Line 9: Define REMO at it first occur.
Lines 11-12: Include quantitative performance metrics to support the reported improvements achieved by coupling, calibration, and bias correction.
Line 13: Define EURO-CORDEX at it first occur.
1 Introduction
Lines 35-40: The Introduction places disproportionate emphasis on irrigation, whereas irrigation represents only one component of C-CWatM. The Introduction should instead emphasize the model's capability to represent human water demand and water resources management, which forms the primary motivation for the proposed framework.
Lines 61-63: The manuscript should better emphasize that the primary contribution of C-CWatM is the representation of realistic human water demand and water resources management. The methodology, results, and discussion should focus more on evaluating these capabilities rather than primarily on calibration and bias correction.
Line 66: According to the model description, the coupling is one-way rather than two-way, since information is transferred only from the land surface model to C-CWatM without feedback. The authors should use consistent terminology throughout the manuscript and avoiding expressions such as "fully coupled" or "two-way coupling" to prevent misleading readers.
2 Model concept
Lines 78-81: This section should focus primarily on introducing C-CWatM. The discussion of limitations of other models largely repeats the Introduction and could be shortened.
Lines 81-83: Refer to Figure 1 when introducing the model framework.
Figure 1: Consider separating the online and offline configurations into two labeled subpanels (e.g., panels a and b) and refer to these subpanels consistently throughout the manuscript.
Lines 85-86: Provide the governing equations or appropriate references for the major model components used in this research.
Line 109: The coupling described appears to be one-way rather than two-way because variables are transferred from the land surface model to C-CWatM without feedback. revise the terminology accordingly.
3 Overview of processes
3.1 River routing, reservoirs and lakes
Lines 144-145: Provide the key equations (or references) describing the simplified representation of lakes and the treatment of open-water evaporation as an external forcing.
3.3 Human water use
3.3.1 Irrigation water demand
Line 174: Provide a reference supporting the assumption of maintaining a 50 mm water depth for paddy irrigation.
Lines 177-179: Provide references supporting the simplified irrigation representation adopted in the model.
3.3.3 Water abstraction by source
Lines 194-197: This subsection represents one of the key contributions of C-CWatM. The ability to allocate water demand from different sectors (e.g., irrigation, domestic, industrial, and livestock) among multiple water sources (surface water and groundwater) is one of the major strengths of the model. The authors should expand this description and provide the governing equations or appropriate references for the water allocation scheme.
Line 203: Clarify how the environmental flow threshold is determined, including the data source, parameter values, and corresponding references.
4 Model enhancements and optional features
4.1 The OASIS3-MCT coupling interface
Lines 208-209: Based on the model description, the coupling appears to be one-way, avoid the term "fully coupled" and clarify that OASIS3-MCT transfers information from the land surface model to C-CWatM rather than performing two-way data exchange.
Line 227: Clarify the computational benchmark used to quantify the reported 50% increase in runtime, including the computing platform and the baseline configuration.
Figure 1: The role of the OASIS3-MCT coupling interface should be highlighted more clearly, as it represents one of the key methodological contributions of this study. If any part of the figure is adapted from existing CWatM documentation or other published sources, provide the appropriate citation. In addition, the current figure contains a large amount of secondary graphical information that is not essential to understanding the coupled framework. A simplified conceptual diagram focusing on the interaction between the land surface model, OASIS3-MCT, and C-CWatM would improve readability.
4.2 Optional model calibration
Line 247 and Table 1: Provide the justification or reference for selecting alpha depletion is 0.7 as the newly introduced calibration parameter.
4.3 Bias correction
Lines 257-263: The discussion of bias correction and its limitations is unnecessarily lengthy. This section could be simplified into one or two concise paragraphs while focusing on the bias correction method actually implemented in this research.
Line 275: The real-time bias correction is estimated from reference CWatM simulations rather than observational datasets. This experimental design makes it difficult to distinguish improvements arising from the coupled model itself from those introduced by the bias correction procedure. Please discuss this limitation more thoroughly.
4.4 Reduced forcing requirements
Line 280: Define CMIP and CORDEX at it first occur.
Line 281: Refer to the corresponding panel in Figure 1 when introducing the offline configuration.
5 Data
5.1 Input map
Lines 301-302: Identify the dataset used for field capacity and wilting point, including its spatial resolution. If this dataset is not already included in Table B1, it should be added accordingly.
Figure 2: Add a north arrow and scale bar.
5.2 Forcing data
Table B-1: Clarify how the 3′ DEM was resampled to match the 5′ model grid.
Lines 307-308: Describe how the 0.11° EURO-CORDEX forcing was transformed to the 5′ C-CWatM grid.
Line 323-325: Provide the reference describing the conversion from relative soil saturation (SM relsat) to absolute soil moisture.
5.3 CWatM model run for comparison
Line 328: Clarify the spatial resolution of the MSWX forcing dataset.
6 Test application
Lines 336-337: The description of the CDO bilinear remapping applied to the REMO outputs is part of the forcing data preprocessing and would be more appropriately presented in Sect. 5.2 (Forcing data) rather than in Sect. 6 (Test application). The relevant parts of Sect. 6 should be integrated into Sect. 5.2.
6.1 River discharge and water demand
Lines 342-343 and Figure 3: The manuscript states that the model reproduces the European river discharge patterns well. Figure 3 only presents the spatial distribution of simulated discharge. The authors should add the spatial distribution of the reference discharge together with a difference map (simulated minus reference discharge) to better support this conclusion.
Line 350 and Figure 3: If no observational dataset is available for evaluating water demand, the authors should move these results to the Appendix or Supplementary Material rather than presenting them in the main figure. The main figure should focus on results that directly support the principal conclusions of the manuscript.
Line 359: The experimental design requires further clarification. The coupled C-CWatM simulations are driven by runoff, evaporation and soil moisture simulated by REMO2020, whereas the standalone CWatM reference simulations are driven by MSWX meteorological forcing. Consequently, the differences may arise from both the forcing datasets and the coupled modeling framework, making it difficult to attribute the performance improvements solely to model coupling.
The authors should clarify the rationale for this experimental design and discuss the influence of the different forcing datasets on the comparison. Ideally, the baseline CWatM simulations should be driven by forcing data that are consistent with those used in the coupled simulations to better isolate the effect of the coupled modeling framework.
6.2.1 Validation against observed discharge
Lines 376-377 and Figure 4: To better demonstrate the added value of the coupled model, the authors should include the performance of the stand-alone CWatM against GRDC observations in Fig. 4e, together with a spatial distribution of KGE for the standalone CWatM. Such a comparison would provide a more direct assessment of whether coupling improves model performance relative to the standalone configuration.
Lines 390-393: Further discuss why the largest discrepancies occur over southern and central Europe, whereas the strongest positive biases are mainly found over northern and eastern Europe.
6.2.2 Validation against reference simulations
Lines 416-417: Improved agreement with the reference model does not necessarily imply improved agreement with observations. Since the standalone CWatM runoff is used as the reference for bias correction, the improved consistency between the two models is an expected outcome. These results should be interpreted with greater caution.
Lines 419-420: Because bias correction artificially increases the consistency between the two model configurations, the conclusion that bias correction is more important than calibration is not yet sufficiently supported. The manuscript does not demonstrate whether a similar conclusion would still hold if bias correction were performed using independent observations (e.g., GRDC). If the primary objective of the study is to demonstrate the added value of C-CWatM, the discussion should focus more on the comparison between the coupled and standalone model configurations (e.g., noC-noBC versus C-noBC), rather than primarily emphasizing the improvements introduced by bias correction.
Figure 7: The current figure mainly highlights the benefit of bias correction. The authors should redesign this figure to emphasize the advantages of the coupled model itself, for example by comparing C-CWatM simulations driven by REMO runoff directly against observations and against the standalone CWatM simulations.
6.3 Reduced forcing considering EURO-CORDEX simulations
Figure 8: The authors should compare the C-CWatM simulations forced by the 13 EURO-CORDEX model outputs directly against GRDC observations rather than against the standalone CWatM reference simulations. Such a comparison would provide a more direct assessment of the simulation ability of the coupled model under different climate forcing datasets.
7 Discussion
Lines 455-459: The authors state that the limited availability of GRDC observations prevents their use for bias correction; however, this alone does not fully justify the chosen approach. Bias correction could still be evaluated for basins with sufficient observations, followed by a discussion of the corresponding results. In addition, the manuscript should discuss the implications of using model-simulated runoff, rather than observational data, as the reference for bias correction, particularly with respect to the interpretation of the coupled model performance.
Line 465: The manuscript should provide further discussion of the potential influence of using different meteorological forcing datasets for the standalone CWatM (MSWX) and the coupled C-CWatM (REMO2020). Since differences in model performance may arise from both the forcing data and the coupled modeling framework, the contribution of each source of uncertainty should be discussed more explicitly.
In addition, given that REMO2020 is driven by ERA5 boundary conditions, please justify why the standalone CWatM simulations were not performed using ERA5-based forcing to provide a more consistent baseline for evaluating the coupled modeling framework.
Lines 459-460: The manuscript reports predominantly positive discharge biases over northern, eastern and mountainous regions, whereas negative biases mainly occur over western and southern Europe. The physical mechanisms responsible for these regional differences deserve further discussion.
The Discussion would benefit from a more comprehensive assessment of the current limitations of C-CWatM and possible future improvements. Besides groundwater and irrigation processes, the authors should discuss how the coupling scheme, the choice of forcing datasets, and the bias-correction framework may contribute to the remaining model biases. A broader discussion of potential improvements would provide a more balanced assessment of the strengths and limitations of the coupled modeling framework.