the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 08 Jul 2025
Distinct mechanisms shaping global surface and root-zone soil moisture
Abstract. Soil moisture (SM) plays a vital role in the global water and carbon cycles, with long-term impacts on ecosystem functioning and vegetation growth. However, under the background of climate change, a decoupling phenomenon have occurred between surface soil moisture (SMsurf) and rootzone soil moisture (SMroot). The variations and primary driving factors of SM across different layers have not been studied comprehensively. Therefore, this study explored the spatiotemporal dynamics of global SMsurf and SMroot from 2001 to 2021. The Random Forest coupled with numerical simulation experiments were applied to measure the influences of climate and vegetation dynamics to SM changes. The Partial Least Squares Structural Equation Modeling was employed to demonstrate the direct and indirect pathways of them to SM variability. Additionally, the copula functions were applied to examine the probability of SM loss under different stress scenarios caused by climate or vegetation changes. The results indicated that SM variation exhibited a significant spatial heterogeneity. Global greening significantly contributed to the increase in SMsurf at a rate of 0.000087 m3/m3/a, while precipitation (Pre) had the most significant impact on replenishing SMroot, with a contribution rate of 0.000117 m3/m3/a. Atmospheric water demand (Ep) was identified as the primary cause of global SM drought, with rates of -0.000089 m3/m3/a and -0.000075 m3/m3/a for SMsurf and SMroot respectively. Although vapor pressure deficit (VPD) had a significant dominant effect in regions with high VPD values, rather than globally, as global positive and negative VPD effects offset each other. Vegetation typically acted as an intermediary variable transmitting the indirect effects of climate factors on SM. Under the extreme scenarios, Precipitation, Standardized precipitation evapotranspiration index and vegetation resulted in the highest probability of SM loss. This research will furnish a theoretical underpinning for global water resource management and hold significant implications for the sustainable development of ecosystems.
- Preprint
(2971 KB) - Metadata XML
-
Supplement
(7352 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-2828', Anonymous Referee #1, 01 Aug 2025
This study analyzes the contributions of climate factors, vegetation dynamics, and drought indices to surface (SMsurf) and root-zone (SMroot) soil moisture variability using Random Forest (RF) and Partial Least Squares Structural Equation Modeling (PLS-SEM), while also evaluating SM loss probabilities under extreme scenarios using copula functions. Although the research offers significant insights into SM mechanisms, several critical issues warrant further clarification and discussion.
- The manuscript suggests that the increase in surface and root-zone SM globally can be attributed to vegetation greening and precipitation, respectively. However, this contrasts with findings from multiple studies (e.g., Deng et al., 2020; Qin et al., 2023; Seo et al., 2025) that document decreasing SM trends globally, particularly in recent decades.
- Figure 5: At the global scale, the PLS-SEM indicates that the “Climate Change” negatively impacts the “Drought Intensification”. Does it mean climate change alleviates drought? I think it’s unseasonable. In addition, the relationships in the second and fourth columns are chaotic, please further clarify it.
- Significant discrepancies exist between RF and PLS-SEM results concerning vegetation’s role in SM dynamics. While RF analysis identifies vegetation as a primary driver of SMsurf increases (Fig. 3a), PLS-SEM shows minimal direct vegetation effects on SMsurf (Fig. 5f-1). Furthermore, RF indicates vegetation contributes to SMroot decreases (Fig. 3a), whereas PLS-SEM suggests vegetation greening promotes SMroot increases (Fig. 5-3).
- Although the study aims to elucidate distinct mechanisms governing surface and root-zone soil moisture, the results reveal remarkably similar spatial patterns and driving factors for both layers (e.g., Figs. 4-7).
Minor Comments
- Figure 3: Please modify the figure caption.
- Line 491: Please check “he”.
- Line 530: Please check “conFigureurations”.
- Line 558: Don’t repeat the definition of an abbreviation (Ep).
- Line 561: Please use the abbreviated form (SPEI).
References
Deng et al., Variation trend of global soil moisture and its cause analysis. Ecological Indicators, 110, 105939 (2020). https://doi.org/10.1016/j.ecolind.2019.105939
Ki-Weon Seo et al., Abrupt sea level rise and Earth’s gradual pole shift reveal permanent hydrological regime changes in the 21st century. Science, 387,1408-1413 (2025). DOI:10.1126/science.adq6529
Qin et al., Continued decline of global soil moisture content, with obvious soil stratification and regional difference. Science of The Total Environment, 864, 160982 (2023). https://doi.org/10.1016/j.scitotenv.2022.160982
Citation: https://doi.org/10.5194/egusphere-2025-2828-RC1 -
AC1: 'Reply on RC1', Yangyang Liu, 05 Aug 2025
Dear Reviewer,
We would like to express our sincere gratitude for your valuable comments and detailed review which enabled us to improve our manuscript. We provide a detailed point-by-point response to the explanations we need to offer and potential improvement plans in the attached pdf file.
We hope you will grant us an opportunity to revise the manuscript. We believe that through these revisions, the quality of the manuscript will be significantly enhanced. If you have further suggestions, we are happy to continue discussing them with you.
Best Wishes,
Zijun Wang, Yangyang Liu
On behalf of the author team
-
RC2: 'Comment on egusphere-2025-2828', Anonymous Referee #2, 24 Aug 2025
This study aims to revealing distinct mechanisms behind changes in surface and root-zone soil moisture (SM) variabilities. However, the basic SM data from GLEAM do not support this study, and the definition for the SM variability is not specific. Therefore, this manuscript is not recommended for publication in HESS.
Major concerns:
(1) The authors tried to analyze the differences of changes in surface and root-zone SM. They used the GLEAM SM data and stated that the GLEAM SM data give surface (0-10 cm) and root-zone (10-250 cm) soil moisture estimations (Lines 136-137). However, this statement is not true. In Martens et al. (2017) (the paper gives details on GLEAM, doi: 10.5194/gmd-10-1903-2017), “The depth of the root zone is a function of the land-cover type and comprises three model layers for the fraction of tall vegetation (0–10, 10–100, and 100–250 cm), two for the fraction of low vegetation (0–10, 10–100 cm), and only one for the fraction of bare soil (0–10 cm)”. Given the root-zone SM includes the surface SM, the differences of changes in surface and root-zone SM are not well supported.
(2) The authors used another two SM datasets (i.e. GLDAS and ERA5-Land) to validate the GLEAM SM data. First, the depths of SM layers between the three datasets are discrepant. Second, the depth of the GLEAM root-zone varies by land-cover types. The direct comparisons among these datasets are very hard. Last, even if comparing these datasets, it is better to compare changes in surface and root-zone SM among these datasets than original values.
(3) In Abstract and across the manuscript, the decoupling phenomenon and difference of changes between surface soil moisture (SMsurf) and rootzone soil moisture (SMroot) are not specified. For example, in Section 3.1 and Figure 1, the authors do not give what is the trend of SM. For the time series in Figure, it seems as monthly time series with annual cycle. How to estimate trend in the monthly time series as m3/m3/a? In Section 2.1.6, the STL method was used to detend trend and seasonal components in SM time series, why in Figure 1, the two components still remained.
(4) As the above mentioned, the STL method was used to detend trend and seasonal components in SM time series, and then, the residual component is used for analyzing SM variability. Here, how to define SM variability and at which time scale? Usually, standard deviation is used to define variability (see doi: 10.1126/sciadv.adm9732).
(5) Too many statistical approaches were used in this study, leading to these results dizzy and giddy. These methods includes Seasonal and Trend Decomposition using Loess, Mann-Kendall test, Random Forests, Partial Least Squares Structural Equation Modeling, Wavelet coherence analysis, and Copula functions. Some of the results from these methods are not shown in the main text.
(6) Readability. The manuscript has a bad readability because of many statistical approaches, many variables, and many climatic zones. Especially, the globe is divided into many climatic zones. However, the results varies with these zones, without consistent performance. Maybe, the climatic zones do not have strong reasonable.
(7) Seasonality. As we know, SM in many regions has seasonality. Changes in SM vary by seasons, and reasons behind the changes vary by seasons. For example, in mid- and high-latitude regions of the Northern Hemisphere, drivers behind winter and summer SM are very different. The authors clearly overlooked this issue.
(8) Physical mechanisms. The authors used Random Forests, and Partial Least Squares Structural Equation Modeling to quantify drivers and reveal influencing pathways. However, the analyses of results are shallow, like day-to-day account. For example, in Figures 2-4, the clear spatial heterogeneity between these drivers and between regions makes reader hard to understand which is the dominant driver and pathway, and why.
(9) Figures. Many of the figures in main text and supporting information lack aesthetics and clear description. For example, in Figure 1, the font size is too small and the caption expressed very vague. In Figure 6e, h, and others, the colourbars loss the ability to reflect spatial changes.
Citation: https://doi.org/10.5194/egusphere-2025-2828-RC2 -
AC2: 'Reply on RC2', Yangyang Liu, 10 Sep 2025
Dear Reviewer,
Thank you very much for carefully reviewing our manuscript and helping us improve its quality. You believe that this manuscript is not recommended for publication in HESS, and we feel very regretful and sad. However, regarding the comments you put forward, we can make sufficient revisions and reasonable explanations. The core issue with our manuscript is that it fails to highlight the decoupling phenomenon between SMsurf and SMroot. However, we have figured out solutions: we will adopt new methods and, instead of only conducting analysis on a global scale, focus on regions with obvious decoupling. In summary, our revision plan can be summarized as follows:
- We will use five widely used SM products: GLEAM, GLDAS, ERA5-Land, MERRA-2, and CFSR, as well as the mean value of multiple products for subsequent analysis. Instead of merely conducting product comparison and validation, we will simultaneously focus on the calculation results of various products and discuss possible inconsistencies.
- We will pay special attention to the decoupling regions of SMsurf and SMroot, and the calculation methods include: Pearson correlation coefficient, lag cross-correlation analysis, ratio of coefficient of variation between SMsurf and SMroot, etc. In the subsequent analysis, we will focus on these decoupling regions and conduct in-depth discussions.
- We will remove or improve the previous RF or PLS-SEM methods. In the revised manuscript, we will calculate PLS-SEM at the pixel scale and determine the main impact paths pixel by pixel; or use causal Shapley analysis based on machine learning models to analyze the impact paths of variables on SM and the relative importance of variables.
- Regarding the above results, we will focus on areas with different vegetation types, different climate zones, as well as regions where SMsurf and SMroot are significantly decoupled, and pay attention to the results where SMsurf and SMroot have large differences in their responses to driving factors. Therefore, we will make great efforts in the presentation and discussion of the results, closely centering on the theme of the decoupling of SMsurf and SMroot.
We believe that through our careful revisions, we will definitely be able to help us understand the decoupling phenomenon between SMsurf and SMroot, as well as help us unravel the different driving mechanisms of SMsurf and SMroot.
We provide a detailed point-by-point response to the explanations we need to offer and potential improvement plans in the attached pdf file.
We hope you will grant us an opportunity to revise the manuscript. We believe that through these revisions, the quality of the manuscript will be significantly enhanced. If you have further suggestions, we are happy to continue discussing them with you.
Best Wishes,
Zijun Wang, Yangyang Liu
On behalf of the author team
-
AC2: 'Reply on RC2', Yangyang Liu, 10 Sep 2025
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 381 | 45 | 14 | 440 | 16 | 7 | 9 |
- HTML: 381
- PDF: 45
- XML: 14
- Total: 440
- Supplement: 16
- BibTeX: 7
- EndNote: 9
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1