the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 11 Jun 2026
Bedrock water storage regulates seasonal forest sensitivity to climatic water deficit
Abstract. Understanding how subsurface water storage regulates ecosystem responses to hydroclimatic variability is central to ecohydrology, but the extent to which lithology mediates seasonal sensitivity of forests to climatic water deficit (CWD) through soil–regolith water storage remains poorly understood. Using four satellite vegetation metrics and meteorological reanalysis (2000–2023), we quantify seasonal sensitivity to CWD across the hydro-lithological regions of the Qinling Mountains. Results show that forest functional responses to CWD were more pronounced than structural greenness, implying that greenness-based metrics may overlook substantial drought impacts on forest ecosystems. Temperature and precipitation affected drought sensitivity in different ways across seasons and regions. Higher temperatures generally reduced drought sensitivity in spring and summer, especially where soils and bedrock can store more water, but increased sensitivity in autumn. More precipitation increased sensitivity to CWD in summer and autumn, although this effect differed across regions. Bedrock-stored water exerts a dual effect. In summer, it helped reduce drought stress by providing extra water during the growing season. In spring, however, it could increase vegetation sensitivity because it encouraged canopy development, which raised water demand. In regions where access to bedrock water was limited, even a small amount of bedrock-water access was associated with higher sensitivity to CWD. Our findings demonstrate that drought assessment and forest management should account for hydro-lithological properties and bedrock water dynamics.
- Preprint
(2463 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 02 Sep 2026)
- RC1: 'Comment on egusphere-2026-2712', Anonymous Referee #1, 13 Jul 2026 reply
-
RC2: 'Comment on egusphere-2026-2712', Anonymous Referee #2, 12 Aug 2026
reply
The manuscript addresses an important ecohydrological question by examining how hydro-lithological variation and inferred bedrock-water storage may regulate seasonal forest responses to climatic water deficit (CWD) in the Qinling Mountains. The comparison of structural and functional vegetation indicators is potentially valuable, and the attempt to quantify the role of subsurface water storage is particularly interesting. However, I have substantial concerns about the construction and interpretation of the counterfactual bedrock-water analysis, as well as about the interpretation of the regression coefficient β₁ as drought sensitivity. These issues affect several of the manuscript’s central conclusions.
The authors first define ordinary climatic water deficit as CWD = PET − ET (L131–135). To estimate monthly bedrock-water use, the authors assume that when ET exceeds the water supplied by precipitation plus the water remaining in the soil, the missing amount is assigned to bedrock water (Eq. 2). To create a counterfactual scenario without a bedrock subsidy, they define CWDe = CWD + Sbedrock and rerun the regression using CWDe instead of CWD. They then conclude that the difference between the resulting coefficients quantifies the extent to which bedrock water buffers vegetation responses (L254–263).
I see two major issues here. First, to simplify Eq. (2), let A denote the available shallow-water supply represented by the inner term in Eq. (2). In months for which Sbedrock > 0, Eq. (2) reduces to ET − A. Since CWDe is defined as CWD + Sbedrock, and since CWD is defined as PET − ET, Eq. (4) essentially becomes CWDe = (PET − ET) + (ET − A). Thus, in months with Sbedrock > 0, the ET term cancels algebraically and CWDe becomes PET − A, i.e. a deficit index based on PET relative to estimated shallow-water availability. To me, this reads more like an alternative water-deficit metric than a counterfactual representing the absence of bedrock-water access.
Second, the vegetation response variable in the two regression models remains unchanged. Therefore, the dependent variable still represents the actual observed vegetation response, including any effect of bedrock-water access. I do not understand how replacing CWD with CWDe while keeping the same observed vegetation response identifies how vegetation itself would have responded in the absence of bedrock water. The change in the regression slopes demonstrates that the same vegetation observations relate differently to two definitions of water deficit, but that does not by itself identify the causal buffering effect of bedrock water. Yet, the authors use this comparison to infer buffering (L397–404), amplification and structural overshoot”(L406–410, L418–430), and even a shift from energy to water limitation (L414–417).
In addition, the methods state that all response and predictor variables are standardized to zero mean and unit variance after differencing (L203–207). If CWD and CWDe are therefore standardized separately in their respective models, their variances and covariance structures with the other predictors will differ, so a quantitative comparison between the resulting standardized coefficients requires further justification.
Independently of these issues, I also have difficulty with the interpretation of β₁ as drought sensitivity. The methods explicitly state that negative β₁ indicates declining vegetation activity with increasing CWD, whereas positive β₁ indicates increasing vegetation activity (L210–215). The manuscript further suggests that the frequently positive functional response may occur because higher CWD coincides with clearer skies and greater incoming radiation (L283–286). This would mean that a larger positive β₁ cannot straightforwardly be interpreted as greater drought vulnerability, as it may partly reflect covariation between CWD and energy availability. This makes statements such as “greenness-based metrics may overlook substantial drought impacts” in the abstract (L23–25) difficult to support from the coefficient magnitude alone. The authors should therefore explicitly separate energy- and moisture-driven covariation, for example through models incorporating incoming radiation and/or VPD, or through an alternative sensitivity analysis. The manuscript also needs to distinguish throughout between "responsiveness to CWD" and adverse "drought vulnerability."
Finally, the key results are presented without uncertainty quantification, e.g. for the estimated contrasts in β, the marginal effects in Table 1, and especially the reported sensitivity changes between the original and counterfactual models. For example, changes in structural sensitivity from 0.026 to 0.022 and from 0.016 to 0.012 are interpreted as evidence of bedrock-water buffering (L397–401), while elsewhere changes are expressed as relative differences ranging from 0.7% to 24.2% (L418–424). Confidence intervals, or preferably resampling-based uncertainty estimates, are needed to establish whether these differences are distinguishable from estimation uncertainty. Given how small some of the absolute changes in effect size are, I would expect at least some uncertainty intervals to overlap. Any uncertainty analysis should also account for spatial autocorrelation among grid cells rather than treating spatial observations as independent.
Overall, I think the manuscript addresses a worthwhile question and contains potentially useful analyses, but the central mechanistic and causal conclusions currently go beyond what the statistical comparison appears to identify. In particular, the authors should reconsider the interpretation of CWDe as a counterfactual without bedrock-water access, distinguish statistical responsiveness to CWD from adverse drought vulnerability, and provide uncertainty estimates for the reported coefficient differences and marginal effects. Addressing these points would be required to connect between the statistical results and the manuscript’s ecological conclusions.
Citation: https://doi.org/10.5194/egusphere-2026-2712-RC2
Viewed
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 186 | 24 | 9 | 219 | 6 | 7 |
- HTML: 186
- PDF: 24
- XML: 9
- Total: 219
- BibTeX: 6
- EndNote: 7
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
This manuscript focuses on how bedrock water storage regulates seasonal forest sensitivity to climatic water deficit in the Qinling Mountains. The topic has certain scientific significance and is relevant to ecohydrology and critical-zone hydrology. The authors combine multiple remote-sensing vegetation indicators, meteorological data, hydro-lithological classification, ridge regression, and a water-balance framework. However, the manuscript still has several important problems, including insufficient validation of bedrock water estimation, some overinterpretation of statistical results, and limited readability of several figures. Therefore, I recommend major revision.