the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 Jan 2026
A coupled surface water–groundwater multi-objective optimization framework for coordinated water–ecosystem–agriculture management in arid inland river basin
Abstract. In arid regions, water is the key link sustaining both production and ecosystems, and its sustainable management is essential for regional security. This study constructs a coupled surface water–groundwater hydrology–agriculture multi-objective optimization model for the mainstream area of the Tarim River Basin, in which the NSGA-III algorithm is applied to optimize four objectives, including agricultural economic benefit per unit of irrigation water (fAB), groundwater level rise (fGL), terminal lake area (fLA), and total agricultural nitrogen load (fTN). Based on the optimized solutions, the trade-offs and synergistic pathways among multiple objectives within the water–ecosystem–agriculture system are systematically evaluated under different hydrological year conditions. The results indicate that significant trade-offs exist among objectives, with fAB showing negative relationships with fTN and fLA, and solutions with higher economic benefits generally accompanied by reduced ecological water supply and increased nitrogen loads. The spatial heterogeneity of the basin necessitates the adoption of differentiated management strategies, whereby upstream areas with relatively stable water availability can sustain higher levels of agricultural production, while midstream and downstream areas are highly sensitive to ecological water constraints and therefore require priority allocation to ecological water use. The optimization results show that cultivated land area should be dynamically adjusted under different hydrological conditions, ranging from 11.3–14.3 × 10⁴ hm2 in wet years, 10.1–13.1 × 10⁴ hm2 in normal years, and contracting to 9.5–11.9 × 10⁴ ha in dry years. The cropping structure is dominated by cotton (69.7 %–75.8 %), with the proportion of high-benefit crops such as vegetables and fruit crops moderately increased in wet years, whereas in dry years the structure shifts toward water-saving crops and high water-consuming crops are appropriately restricted. This study demonstrates that combining multi-objective optimization with spatially differentiated regulation can achieve coordinated management of water resources, ecosystems, and agriculture, and provides an operational decision-making basis for managing water–ecosystem–agriculture systems in arid inland river basins.
- Preprint
(1783 KB) - Metadata XML
-
Supplement
(984 KB) - BibTeX
- EndNote
Status: final response (author comments only)
- CC1: 'Comment on egusphere-2026-55', Nima Zafarmomen, 26 Jan 2026
-
RC1: 'Comment on egusphere-2026-55', Anonymous Referee #1, 30 Jan 2026
Comments for Manuscript egusphere-2026-55
General Comments
This study develops a coupled surface water–groundwater multi‑objective optimization framework to coordinate water‑ecology‑agriculture management in the Tarim River Basin. The work integrates the GSFLOW hydrological model, the SRM snowmelt module, surrogate modeling, and the NSGA‑III algorithm. The methodology is advanced and reasonable, and the research holds clear scientific value and application potential for synergistic water resources management in arid inland river basins. The paper is highly readable, well‑structured, and provides in‑depth analysis. However, the following aspects require further clarification and strengthening. It is recommended for publication after the authors address these points.
Specific Comments
Comment 1. Regarding the construction of the surrogate model, the current description of its input variables (e.g., which key GSFLOW parameters or optimization decision variables are included), the size and dimensionality of the training sample set generated via Latin Hypercube Sampling, as well as the method and specific proportion for dividing the training and test sets, remains insufficient. Additional explanation is needed.
Comment 2. Although the paper mentions that model results still require validation with field data, the discussion of potential real‑world constraints during the implementation of the proposed solutions is not yet sufficient. It is recommended to supplement the analysis by discussing the potential impacts of the Pareto‑optimal solutions-such as the suggested dynamic adjustments to cropland scale and planting structure-on local livelihoods (e.g., residents and the economy), regional agricultural production stability, and food security. This would strengthen the linkage between the research findings and actual management decision‑making.
Comment 3. The paper notes that the estimation of agricultural non‑point source pollution relies on a simplified model, but the explanation of the method's applicability assumptions and limitations could be further deepened. It is suggested to elaborate on the context in which the simplified method is applied in this study (e.g., its computational efficiency advantages in strategic, large‑scale multi‑objective optimization analyses) and to specify its main limitations more clearly, such as the inability to depict the transport and transformation processes of nitrogen in the "soil-groundwater-surface water" system. Building on this, a more targeted discussion could be provided regarding potential future improvements, such as coupling the hydrological model with a biogeochemical process model (e.g., MODFLOW‑MT3DMS) to enhance the precision and reliability of pollutant simulation.
Comment 4. The justification for selecting the compromise solutions (S5, S10, S15) in Section 4.2.2 is currently quite brief. It is recommended to supplement the explanation with the specific criteria and rationale used for choosing these compromise solutions, thereby increasing the feasibility and transparency of the proposed method.
Comment 5. It is suggested to add a technical workflow diagram in the Methods section to clearly illustrate the overall process and data flow among the key steps (e.g., SRM, GSFLOW, surrogate model, NSGA‑III optimization). This would enhance the intuitive understanding of the framework's logic.
Minor Comments
- It is recommended to unify the formatting for numerical ranges throughout the text, for example, consistently using the format "11.3×10⁴ - 14.3×10⁴ hm²."
- The thresholds used in Table 4 for classifying hydrological year types based on the runoff anomaly percentage (P) are not rigorously defined. It is recommended to use symbols such as "≤" to clearly specify the interval boundaries.
- The terminology for describing hydrological year types is inconsistent in the text. It is suggested to standardize the terms, for example, consistently using "wet/normal/dry year" throughout the main text, avoiding mixed usage with terms like "High/Low Flow Year."
- There are errors in the sub‑figure labels in Figure 9. The letters (f)-(l) overlap with (i)-(q), and the number of labeled sub‑figures does not match the number of representative schemes described in the text. Please correct the labels according to the actual content.
Citation: https://doi.org/10.5194/egusphere-2026-55-RC1 -
RC2: 'Comment on egusphere-2026-55', Anonymous Referee #2, 03 Feb 2026
Comments for "A coupled surface water–groundwater multi-objective optimization framework for coordinated water–ecosystem–agriculture management in arid inland river basin" by Chen et al.
General Comments
Chen et al. (2026) develops a coupled surface water-groundwater multi objective optimization framework to coordinate water ecology agriculture management in the Tarim River Basin. The research gap is clear, the methods are novel, and the manuscript is mostly well written. However, several aspects require further clarification to improve its coherence and readability before publication.
Specific Comments
1. Methods
This study adopts multiple methods, but the methodology appears somewhat overloaded and fragmented. It is recommended to briefly introduce the Data in this section and present the detailed information in the form of supplementary materials or appendices. Also, a flowchart is suggested here to clearly illustrate the integration of the adopted methods. Furthermore, the descriptions of the design of the GSFLOW model and the surrogate model are insufficient, and detailed information (e.g. parameters and variables for the surrogate model) is needed.
2. Discussion
The findings of this study need to be further discussed to enhance the value of the manuscript. It is recommended to discuss the advantages and limitations of the multi-objective optimization framework used in this study compared with previous research (e.g. for single methods or combinations of two methods). It would greatly increase the significance of the study, if quantitative comparison can be presented. The discussion could also be strengthened by addressing how the findings of this study could be applied to other regions. This may involve identifying which of the datasets used has the greatest impact on the results, as well as considering how errors from different datasets could limit the applicability of the methodology in other areas.
Specific Comments
Line 23 Please unify the units used in the manuscript, such as hm, km, etc., throughout the text.
Line 51 References should be added here. Also, references are lacked in some parts of the remain Introduction.
Line 83-87 This paragraph is unnecessary.
Line 158 This is the first time GSFLOW appears, and it is recommended to provide a brief introduction. However, since this section is actually about the Data, please briefly introduce GSFLOW in the appropriate place.
Line 261 Please unify the use of these symbols
Line 503 The order of the subfigures in the figure captions does not match their actual order.Citation: https://doi.org/10.5194/egusphere-2026-55-RC2 -
RC3: 'Comment on egusphere-2026-55', Anonymous Referee #3, 20 Feb 2026
This study aims to identify water-ecosystem-agriculture (WEA) critical thresholds of cultivated land expansion and optimal cropping structures under different hydrological conditions. The authors achieve this by constructing a coupled surface water-groundwater hydrology-agriculture multi-objective optimization model for the mainstream area of the Tarim River Basin with the objective to optimize for (i) agricultural benefit per unit of irrigation water, (ii) groundwater level rise, (iii) terminal lake area, and (iv) total agricultural nitrogen load. Their results show significant trade-offs among their objectives. According to their results, solutions that maximize economic output increase the agricultural returns but reduce ecological water availability and increase nitrogen pollution. On the other hand, maximizing ecological or environmental objectives requires moderate reductions in agricultural output, underscoring the need for management strategies that balance multiple objectives.
My overall opinion is that the manuscript, although well written, cannot be published in its current state and needs additional revisions and clarifications. Below, I have listed comments, hoping they may help improve the manuscript’s quality.
Specific Comments
- There are multiple citations within the manuscript that are not added to the reference list. This prevented me from assessing the quality of the data the authors used and from corroborating some of their assumptions and interpretations. This is something the authors need to correct to maintain the overall integrity of their work. I strongly encourage the authors to review the manuscript and verify that all their references are included. Here are some of the ones I found missing.
- Line 59, “Chen (2018)”.
- Line 125, “Liao et al. (2020).”
- Line 148 “Li et al. (2003).”
- Line 156 “Xue et al. (2024)”
- Line 160 “(Save et al., 2016; Save, 2024)”
- Line 162 “Zhong et al. (2019).”
- Line 198 “Chen et al. (2025).”
- Line 306 “Deb and Jain (2014).”
- Line 534 “Taon et al. (2012).”
- Line 593 “Miralles et al. (2011)”
- The paragraphs starting in lines 95, 99, 107, 111, 115, and 120 contain information related to the study area. Is this information calculated by the authors, or does it come from other studies? I encourage the authors to include references to the source of this information for transparency. This includes the reports related to the Ecological Water Conveyance Project (EWCP) that are mentioned but not referenced in the manuscript.
- Clarifications are needed regarding the data and model construction:
- The authors mentioned that the surface water and groundwater systems within the region are closely connected (lines 102-103). Is this a result of this study or of other studies? Please clarify and add proper references.
- The authors derived their hydrogeologic conditions from borehole and regional hydrogeologic maps reported by Li et al. (2003) (citation not presented in the reference list). I have a few questions regarding this data,
- Are the authors looking at confined and unconfined aquifers? Have these aquifers been reported in the area? Is there any groundwater pumping that might alter their model assumptions or the water allocation parameter? I encourage the authors to provide this additional information and add potential limitations.
- According to the authors, the hydrogeologic parameters were constrained based on the primary lithology and parameter ranges reported in previous studies (lines 149-150). However, there is no citation for this statement. I encourage them to provide the necessary references and a table or map containing geologic units and the hydraulic conductivity values selected, as this will enhance the reproducibility of their work.
- They mentioned that their coupled surface water-groundwater model was developed based on the GSFLOW framework proposed by Chen et al. (2025), which is not in the reference list. The authors should add this reference and provide additional clarifications to their modeling approach. Mainly,
- The current groundwater model is discretized into 6 layers. Is the depth between the layers coming from the geologic data? What is the hydraulic conductivity of these layers? How was the depth of the aquifer (i.e., depth to a confining unit) determined for this area?
- I also encourage the authors to provide a figure that shows the boundary conditions they used for their model.
- The authors calibrated their model using 139 monitoring wells from Xue et al. (2024), which is another work not in the reference list. Is this information looking at unconfined aquifers in the area, or is it also looking at confined aquifers? This can have important implications for the model's calibration that depends on the conceptualization of the subsurface structure.
- They also mention that other key parameters, such as infiltration coefficients and aquifer hydraulic conductivity, were calibrated using a Bayesian uncertainty framework (lines 209-210). However, they do not provide the specifics of their calibration scheme or the distribution of values used for calibration. I encourage the authors to provide this additional information.
- They used evapotranspiration data from the GLEAM dataset. However, different crop compositions will increase or decrease evapotranspiration. Did the authors consider this in their optimization process?
- This reviewer understands the limitations behind their complex modeling approach. However, there is not much information in the modeling section about nitrogen loads. Are the authors also including a transport model to move nitrogen through the system? Or do they assume this is a point-source contamination that stays in crop areas? The authors should to clarify how they are handling nitrogen loads in their optimization scheme.
- In the multi-objective optimization model section, the authors provide decision variables, their constraints, and the objective functions. Although there are equations for all of them, I don’t seem to find where the monthly ecological water allocation coefficient () is presented within the modeling scheme. Is this an extraction within the hydrological model? Where is it drawn from? It would be helpful to have a schematic of the optimization process.
- Due to the complexity of their hydrological model, the authors opted to train an RBF-NN as a surrogate model. However, it is unclear what the model's input and output variables are, as well as its structure. I encourage the authors to include this information to improve reproducibility.
- Also, the authors show that the model performs better by giving additional training points, which is expected. However, doing this might also lead the model to overfit. I encourage them to consider using a cross-validation scheme for their surrogate model.
Technical Corrections
Besides the comments above, I have a few technical recommendations for the manuscript.
- The text within Figure 1 is difficult to read; consider increasing the font size.
- In line 96, the authors state “… and a total precipitation below 80 mm.” Is this total “annual” precipitation? Please clarify.
- Correct comma placement in “This dataset ,” (line 137).
- Labels in Figure 2 are difficult to read. I recommend increasing the font size and removing the x-labels from figures (a) through (d), as they are the same; this can give you more space in the figure.
- Table 5 is referenced before Figures 5-7. I recommend changing the order of their presentation in the manuscript.
Citation: https://doi.org/10.5194/egusphere-2026-55-RC3 - There are multiple citations within the manuscript that are not added to the reference list. This prevented me from assessing the quality of the data the authors used and from corroborating some of their assumptions and interpretations. This is something the authors need to correct to maintain the overall integrity of their work. I strongly encourage the authors to review the manuscript and verify that all their references are included. Here are some of the ones I found missing.
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 220 | 83 | 26 | 329 | 38 | 22 | 29 |
- HTML: 220
- PDF: 83
- XML: 26
- Total: 329
- Supplement: 38
- BibTeX: 22
- EndNote: 29
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This paper presents a coupled surface water–groundwater–agriculture multi-objective optimization framework for coordinated management of water resources, ecosystems, and agricultural production in the Tarim River mainstream, an arid inland river basin in China. The authors integrate a basin-scale SRM–GSFLOW hydrological model with a multi-objective optimization model solved using NSGA-III, targeting four competing objectives: maximizing agricultural economic benefit per unit irrigation water (fAB), maximizing groundwater level recovery (fGL), maximizing terminal lake area (fLA), and minimizing agricultural nitrogen load (fTN). Well written and fitted for the publication.
1) Several figures (e.g., Figures 4–7) contain dense scatter plots and parallel coordinate lines that are difficult to interpret at first glance. Adding brief quantitative annotations (e.g., correlation coefficients or key threshold markers) would improve readability and interpretability.
2) The manuscript alternates between the terms WEA and WAE when referring to the water–ecosystem–agriculture system. Please standardize terminology throughout the text for consistency.
3) Although the surrogate model performance is strong (R² =0.98), the paper would benefit from explicitly stating how surrogate prediction errors may influence decision-making, especially near Pareto front extremes where trade-offs are most sensitive.
4) While the study carefully calibrates the coupled SRM–GSFLOW model and validates the surrogate RBF-NN, uncertainty is not explicitly propagated through the optimization results. Key sources of uncertainty, such as climate forcing, groundwater parameters, crop water requirements, and fertilizer coefficients, may significantly affect Pareto fronts and identified “compromise solutions.” I recommend to add this limitation in the conclusion part.
5) I do strongly suggest that the authors consider relevant studies that have explored the "assimilation of Sentinel-derived leaf area index to improve the representation of surface–groundwater interactions in irrigation districts". Citing and briefly discussing such work would strengthen the linkage between the proposed framework and existing literature.