| 15 Jun 2026
Decoupling between local precipitation and deep groundwater recharge in a tropical sedimentary basin: Evidence from stable isotopes
Abstract. Groundwater recharge in sedimentary basins is often assumed to be locally controlled, yet increasing evidence suggests the importance of spatially distributed and multi-scale processes. Here, we investigate the controls on groundwater recharge in the Taubaté Sedimentary Basin (southeastern Brazil) using stable isotopes (δ¹⁸O and δ²H), d-excess, and electrical conductivity data from precipitation, springs, and deep wells. Results reveal a clear decoupling between local precipitation and deep groundwater recharge. Most groundwater samples plot below the Local Meteoric Water Line (LMWL) and exhibit reduced regression slopes, indicating the influence of mixing, delayed percolation, and non-conservative recharge processes. Deep groundwater shows low isotopic variability and weak seasonal signals, consistent with longer residence times and regional flow integration. In contrast, spring waters display greater variability and stronger sensitivity to seasonal inputs and near-surface processes. A closer isotopic correspondence between deep groundwater and precipitation from higher-altitude areas suggests that effective recharge occurs preferentially in elevated zones and is subsequently redistributed toward lower basin sectors. Electrical conductivity and d-excess relationships further indicate that shallow groundwater is influenced by evaporation and short flow paths. Together, these findings support a conceptual model of a dual groundwater system, characterized by the coexistence of rapid, shallow flow paths and slower, regionally integrated deep circulation. This study highlights the importance of spatially decoupled recharge and cross-scale groundwater dynamics for understanding hydrological functioning in sedimentary basins.
This manuscript addresses an important hydrological topic: the decoupling of local precipitation and deep groundwater recharge in the Taubaté Sedimentary Basin using stable isotopes and hydrochemical data. The dataset is substantial, and the fieldwork appears to have been conducted rigorously. However, the manuscript currently reads more like a descriptive report of data rather than an in-depth scientific investigation. The conclusions, while plausible, rely heavily on qualitative observations rather than quantitative isotope modeling. To meet the standards for publication in HESS, the authors need to significantly deepen their analysis. Specifically, the manuscript lacks a clear quantitative assessment of recharge sources (e.g., altitude effect calculation, mixing ratios), and the proposed conceptual model is described only in text without a supporting schematic diagram. Furthermore, the number of tables is excessive, and figures require improvement for better visual clarity. Overall, the paper has a strong foundation but requires substantial revision to elevate it from a data report to a mechanistic study. I recommend a Major Revision.
Major comments:
The term decoupling is central to the title but needs more precise definition. Does it mean the recharge water does not come from rain falling on that specific coordinate (spatial decoupling) or that the timing of recharge does not match the season (temporal decoupling)? The manuscript seems to argue for spatial decoupling (recharge from highlands). The authors should clarify this distinction in the Introduction and Discussion to avoid confusion with temporal decoupling (lag time)?
The discussion on the altitude effect is qualitative (Section 4.2). The authors should calculate the altitude gradient (δ18O vs. altitude) using the precipitation data from the upper basin (SFX) and the local area to quantitatively estimate the mean recharge altitude of the deep groundwater.
The manuscript notes that groundwater plots below the LMWL, suggesting evaporation or mixing. A quantitative analysis is needed. The authors should apply a linear mixing model (e.g., IsoSource or similar) or a Craig-Gordon evaporation model to estimate the proportion of mixing or the degree of evaporation loss before infiltration.
Section 4.5 proposes a conceptual recharge model, yet no figure illustrates this. The authors should include a clear, high-quality schematic diagram (conceptual model) that illustrates the dual groundwater system (shallow vs. deep), the regional flow from high-altitude recharge areas to the urban lowlands, and the decoupling mechanism. This figure is essential for readers to visualize the hydrological processes described.
Specific comments:
In Section 3.3, statistical comparisons are made between geological formations. However, Table 1 shows that the Tremembé Formation is represented by only one well (W151). Performing statistical comparisons (ANOVA) with a sample size of n=1 is statistically invalid and misleading. The authors should either exclude this formation from the statistical analysis or explicitly state that statistical tests were not applicable for this unit due to insufficient sampling.
The manuscript distinguishes between deep wells and springs (shallow), implying a distinction based on flow depth. However, there is no detailed table or summary of well depths and screen intervals provided in the main text. A deep well might screen a shallow aquifer, and a spring might represent a perched water table. The authors must provide a summary of the depth and screened formations for all wells in a concise table to justify the deep vs. shallow hydrogeological classification used in the discussion.
The introduction and data processing sections mention the amount effect, yet it is not visually presented or discussed in the results. It is crucial to plot δ18O vs. rainfall amount for the precipitation events to determine if an amount effect exists in this tropical region. If a strong amount effect is present, it explains why light isotopic recharge is linked to heavy rainfall events. This plot and discussion are missing but necessary to support the interpretation of recharge sources.
In Section 4.4, the authors interpret high Electrical Conductivity (EC) in springs (e.g., S7, S10, S15) as an indicator of anthropogenic influence or land-use change. This is a speculative leap without supporting major ion chemistry. High EC can result from natural water-rock interaction or evaporation in semi-confined aquifers.
In the conclusion, the authors mention that well yields have declined by 50%. This is an interesting point but seems disconnected from the isotope analysis results. The authors should discuss whether the observed decoupling (reliance on regional flow from highlands) makes the system more or less vulnerable to climate change compared to a locally recharged system. This would add significant value to the discussion.
There are too many tables in the main text (Tables 1-6). Tables 1 and 2 contain raw data which can be moved to the Supplementary Material. The main text should focus on the results of the analysis (means, statistical significances) rather than listing every sample value.
Tables 4 and 6 present statistical test results (Tukey HSD). These can be summarized succinctly within the text (e.g., “ANOVA revealed no significant difference between formations...”) rather than occupying a full page with a table. Removing these tables will improve the manuscript's flow.
The discussion of why groundwater slopes are lower than the LMWL (Section 3.2 and 4.1) needs to be more rigorous. The authors attribute this to mixing and delayed percolation. Please discuss specific geochemical or physical mechanisms in more detail. For instance, could isotopic exchange with clay minerals or the matrix in the sedimentary basin contribute to the fractionation?
The interpretation of d-excess in relation to EC (Section 3.4) is somewhat superficial. Elaborate on the specific hydrological pathways that lead to high EC but variable d-excess. Is there a clear distinction between water-rock interaction (high EC) and evaporation (low d-excess)?
Figure 1: The cross-section mentioned in the caption is not clearly visible or detailed enough to support the hydrogeological arguments.
Figure 2: The sampling points are difficult to distinguish due to overlap. Please use semi-transparent markers or others so that all 44 sampling locations are visible.
Figure 4: The dual-isotope plots are crucial but are currently cluttered. The regression equations and R2 values are hard to read. Ensure the font size in the legend and axes is legible.
Figure 6: The d-excess vs. EC plot is simplistic. Consider adding 95% confidence ellipses for different groups (e.g., seasons or formations) to highlight differences visually.
There are inconsistencies in the reference list and in-text citations. For example, in the acknowledgments, DeepSeek is cited as DEEPSEEK and DeepSeek. Ensure consistency throughout.
Add more recently published relevant literature. In the Introduction (around Lines 38-44): The text states: “Isotopic analysis of natural waters… has become a fundamental tool for identifying water sources, flow paths, and the hydrological processes that regulate the dynamics of aquatic systems…”. These references of “Chen et al., 2023. https://doi.org/10.1016/j.jhydrol.2023.129149” and “Chen et al., 2024. https://doi.org/10.1016/j.jhydrol.2024.132117”) provide excellent examples to support the introductory statement about identifying flow paths.
In the Introduction (around Lines 44-48): The text discusses the expansion of isotopic approaches in recharge studies and water balance. It is suggested to supplement relevant literature published in the recent five years, especially the latest research progress on isotope hydrology (like Zhang et al. 2026. https://doi.org/10.1016/j.jhydrol.2026.135365). Citing this here highlights the potential for quantitative analysis, which is currently a weakness in your manuscript.
The manuscript frequently uses the first person (e.g., “Here, we investigate”, “Our data”). For a more formal academic tone, please revise these sentences to use the passive voice or impersonal subjects (e.g., “This study investigates”, “The data indicate...”).
Check units in Table 3.
Ensure consistency in terminology throughout the text.