Applicability of the Lumped GR4J Model for Modeling the Hydrology of the Inland Valleys of the Sudanian Zones of Benin

Comments 1: Several figures (e.g., Figures 6–9 and 10–12) contain many subplots and dense information, making them difficult to interpret. Consider adding name for each subplot, improving annotations and captions to enhance clarity.

Response 1: Agree. We have improved the captions for Figures 10–12 to enhance clarity. For example, the caption for Figure 10 has been updated as follows: “Figure 10: Trend analysis of historical climatic parameters (rainfall, reference evapotranspiration, aridity index) and hydrological parameters (total runoff, high flow, runoff coefficients) in the Bahounkpo inland valley. Blue markers represent annual estimated values, and red dashed lines indicate linear regression trends.”

Comments 2: Some repetition is observed in the description of model performance across different sections. These parts could be condensed to improve the overall conciseness of the manuscript

Response 2: Agree. We have reformulated repeated sentences to improve conciseness. For example, the sentence “KGE scores recorded for the highest-performing high-flow simulations indicate marginal improvement over a mean-flow benchmark, reflecting borderline acceptability and necessitating cautious interpretation” has been revised to: “The KGE values for the best high-flow simulations show only limited improvement compared to the mean-flow benchmark and should therefore be interpreted with caution.”

While some similar phrasing remains in Sections 3.1.1 and 3.1.2, it is retained because each instance refers to different results or datasets, ensuring that the interpretation remains context-specific.

Comments 3: In several cases, Kling–Gupta Efficiency (KGE) values close to zero or slightly above −0.41 are described as acceptable, particularly for high-flow simulations. Given that one of the study objectives relates to water resources management and extreme event assessment, a more cautious interpretation of these results is needed.

Response 3: Agree. The phrase “reflecting borderline acceptability” has been removed, and the text now emphasizes that these KGE results, particularly for high-flow simulations, should be interpreted with caution in the context of water resources management and extreme event assessment.

Comments 4: The sensitivity analysis identifies parameter X2 as highly influential across all catchments, but its physical meaning in relation to groundwater exchange and seasonal storage dynamics is only briefly discussed. A deeper hydrological interpretation would strengthen the analysis

Response 4: Agree. The influence of the groundwater exchange coefficient (X2) is now interpreted in hydrological terms in the discussion section, highlighting its role in regulating baseflow and modulating peak runoff. Groundwater exchange coefficient appears as a key driver of runoff response in inland valleys.

Comments 5: Although CHIRPS generally outperforms GPM-IMERG and GSMAP, the discussion of the underlying reasons for these differences remains superficial. The authors should further analyze how rainfall characteristics (e.g., event intensity, temporal resolution, bias in extreme rainfall) and inland valley hydrological responses contribute to the observed performance differences

Response 5: Agree. We noted in the Limitations section that the study does not analyze how rainfall characteristics and inland valley hydrological responses contribute to the observed differences in performance among CHIRPS, GPM-IMERG, and GSMAP. This remains an area for future investigation.

Comments 6: The manuscript is generally well written, but some sentences are long and information-dense. Minor language editing and sentence restructuring would improve readability

Response 6: We thank the reviewer for this comment. After review, we propose to retain the current sentence structure in order to preserve the full meaning and precision of the information throughout the manuscript.

Comments 1: What is the added value from a methodological point of view? What does it bring from a methodological point of view? The model is an intensively studied one having already been applied in those regions. I consider that the novelty of the material consists only in its regional applicative value.

Response 1: We thank the reviewer for this observation. While the GR4J model has been applied in many hydrological studies, including in West Africa, our work presents methodological novelties by demonstrating its reliable application in data-scarce small catchments (<200 Km²), applying a framework to assess the ability of satellite precipitation datasets to drive hydrological simulations at the inland valley scale, and providing a practical approach to support water management under limited data conditions.

Comments 2: An explicit discussion is needed regarding the conceptual limitations of GR4J in basins with intermittent flows, extremely small areas, and/or strong control of local groundwater processes

Response 2: We agree regarding discussions about the conceptual limitations of GR4J in basins with intermittent flows and strong control of local groundwater processes. We note in the discussion section that the lumped structure of GR4J, with a single production store and fixed flow partitioning, may limit its ability to accurately simulate high and low flows, particularly in catchments with intermittent regimes such as inland valleys, where groundwater fluctuations seem to strongly influence runoff generation.

Regarding catchment area, only three sites were included in this study, so any relationship between basin size and model performance should be interpreted with caution. In our results, the model performed better in the smaller catchment (Nalohou) than in the larger one (Lower-Sowé), consistent with observations by [https://doi.org/10.3390/w10111655], which suggest that GR4J is often more suitable for smaller, spatially more homogeneous basins. Conversely, other studies [https://doi.org/10.1038/s41598-025-16429-z] have shown that parsimonious models such as GR4J perform better in larger basins.

Comments 3: The length of the available data is very short, being statistically insufficient for a robust calibration, an analysis of extremes and for identifying and highlighting climate or trend inferences

Response 3: Agree. We acknowledge that the relatively short duration of the available data limits the statistical robustness for extreme event analysis. Nevertheless, the available data are sufficient to evaluate the general hydrological behavior of the inland valleys and to assess the performance of satellite rainfall-driven model simulations, which helps reconstruct the hydrological behavior of the inland valleys according to the methods applied in this study. It is important to note that the databases used in this study are among the most extensive available for inland valleys in Benin

Comments 4: The extremes are estimated with the GEV distribution on simulated, not observed data, thus the uncertainty is under-estimated. The GEV distribution is strongly influenced by the short lengths of the analyzed data series (https://doi.org/10.3390/cli13070152; https://doi.org/10.1029/2012WR012557; https://doi.org/10.1155/2022/54). I ask the authors to introduce an explanatory paragraph regarding this influence of the variability of the observed data lengths and their associated uncertainties.

Response 4: Agree. We note that extremes estimated using the GEV distribution on simulated data may underestimate uncertainty in extreme flow estimates, highlighting the need for caution in their interpretation.

Comments 5: No bias correction is applied, although the authors explicitly recognize this limitation

Response 5: We agree with the reviewer and have acknowledged this limitation in the manuscript. In this study, our objective was to test the applicability of satellite rainfall-driven simulations in inland valleys under limited data conditions, rather than to optimize every methodological aspect, such as bias correction. These observations, which can be integrated into future studies, do not undermine the results obtained here.