Preprints
https://doi.org/10.5194/egusphere-2025-4820
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
https://doi.org/10.5194/egusphere-2025-4820
© Author(s) 2025. This work is distributed under
the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
21 Oct 2025
| 21 Oct 2025
Status: this preprint is open for discussion and under review for Hydrology and Earth System Sciences (HESS).
The influence of vapor pressure deficit changes on global terrestrial evapotranspiration
Abstract. Vapor pressure deficit (VPD) is increasingly recognized as the primary driver of uncertainty in future global evapotranspiration (E) trends. Accurately characterizing the spatiotemporal dynamics of VPD and clarifying its mechanisms of influence on terrestrial E are crucial for improving water-use efficiency, optimizing ecosystem structure and function, and addressing the challenges of global climate change. Previous studies, however, have largely concentrated on the physiological regulation of vegetation transpiration (Et) at the micro scale. Here, we integrate multi-source remote sensing products and reanalysis datasets spanning 1981–2020 to quantitatively disentangle the contributions of VPD to E and assess its role in shaping global terrestrial evapotranspiration. Our results demonstrate that: (1) across 60.7 % of the global land surface, E increased with rising VPD, while in arid regions with limited soil moisture the effect was generally weak; (2) VPD regulates E primarily by modulating Et, with elevated VPD directly enhancing transpiration; (3) the regulation of E by VPD exhibits a clear climatic gradient: arid zones (1.31 kPa) > humid zones (0.32 kPa), and the tropical (0.79 kPa) > temperate (0.68 kPa) > cold (0.28 kPa) > polar (0.07 kPa). By elucidating the dominant pathways and regional heterogeneity of VPD–E interactions at the global scale, this study strengthens the mechanistic understanding of the coupled warming–atmospheric aridity–water flux system. These findings provide quantitative constraints for predicting terrestrial water-cycle changes under global warming and offer scientific evidence to support targeted climate adaptation strategies worldwide.
How to cite. Miao, Y., Zhu, G., Wang, Y., Huang, E., Wang, Q., Gun, Y., Zheng, Z., Yang, J., Li, W., and Liu, Z.: The influence of vapor pressure deficit changes on global terrestrial evapotranspiration, EGUsphere [preprint], https://doi.org/10.5194/egusphere-2025-4820, 2025.
Publisher's note: Copernicus Publications remains neutral with regard to jurisdictional claims made in the text, published maps, institutional affiliations, or any other geographical representation in this paper. While Copernicus Publications makes every effort to include appropriate place names, the final responsibility lies with the authors. Views expressed in the text are those of the authors and do not necessarily reflect the views of the publisher.
Download & links
Download & links
- Preprint
(2219 KB) - Metadata XML
- BibTeX
- EndNote
Status: open (until 31 Mar 2026)
Comment types: AC – author | RC – referee | CC – community | EC – editor | CEC – chief editor
| : Report abuse
- RC1: 'Comment on egusphere-2025-4820', Vagner Ferreira, 27 Feb 2026 reply
Viewed
Total article views: 461 (including HTML, PDF, and XML)
Cumulative views and downloads
(calculated since 21 Oct 2025)
| HTML | XML | Total | BibTeX | EndNote | |
|---|---|---|---|---|---|
| 250 | 185 | 26 | 461 | 27 | 22 |
- HTML: 250
- PDF: 185
- XML: 26
- Total: 461
- BibTeX: 27
- EndNote: 22
Viewed (geographical distribution)
Total article views: 478 (including HTML, PDF, and XML)
Thereof 478 with geography defined
and 0 with unknown origin.
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
1
Latest update: 28 Feb 2026
Yuxin Miao
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Remote Sensing Application Center of Ministry of Agriculture and Rural Affairs, Lanzhou Branch, Lanzhou, 730070, Gansu, China
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Gansu Key Laboratory of Oasis Resource Environment and Sustainable Development, Lanzhou, 730070, Gansu, China
Remote Sensing Application Center of Ministry of Agriculture and Rural Affairs, Lanzhou Branch, Lanzhou, 730070, Gansu, China
Yuhao Wang
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Gansu Key Laboratory of Oasis Resource Environment and Sustainable Development, Lanzhou, 730070, Gansu, China
Remote Sensing Application Center of Ministry of Agriculture and Rural Affairs, Lanzhou Branch, Lanzhou, 730070, Gansu, China
Enwei Huang
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Gansu Key Laboratory of Oasis Resource Environment and Sustainable Development, Lanzhou, 730070, Gansu, China
Remote Sensing Application Center of Ministry of Agriculture and Rural Affairs, Lanzhou Branch, Lanzhou, 730070, Gansu, China
Qingyang Wang
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Remote Sensing Application Center of Ministry of Agriculture and Rural Affairs, Lanzhou Branch, Lanzhou, 730070, Gansu, China
Yani Gun
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Remote Sensing Application Center of Ministry of Agriculture and Rural Affairs, Lanzhou Branch, Lanzhou, 730070, Gansu, China
Zhijie Zheng
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Remote Sensing Application Center of Ministry of Agriculture and Rural Affairs, Lanzhou Branch, Lanzhou, 730070, Gansu, China
Jiangwei Yang
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Gansu Key Laboratory of Oasis Resource Environment and Sustainable Development, Lanzhou, 730070, Gansu, China
Wenmin Li
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Gansu Key Laboratory of Oasis Resource Environment and Sustainable Development, Lanzhou, 730070, Gansu, China
Ziwen Liu
College of Geography and Environmental Science, Northwest Normal University, Lanzhou 730070, Gansu, China
Comprehensive Observatory of Ecological Environment in Shiyang River Basin, Northwest Normal University, Lanzhou, 730070, Gansu, China
Gansu Key Laboratory of Oasis Resource Environment and Sustainable Development, Lanzhou, 730070, Gansu, China
Short summary
We examined how rising air dryness influences water movement from land to air. Using global data from 1981 to 2020, we found that in most regions, greater dryness increased plant water loss, though the effect was weaker in dry soils. The impact varied by climate, strongest in dry zones and weakest in polar regions. These findings improve predictions of water cycle changes under climate change and support strategies for adaptation.
We examined how rising air dryness influences water movement from land to air. Using global data...
General comments:
The proposed ms “The influence of vapor pressure deficit changes on global terrestrial evapotranspiration” integrates multi-source remote sensing and reanalysis data over a 40-year period (1981 to 2020) to analyze VPD-evapotranspiration relationships at the global scale. The ms’s most notable contributions are the quantification of climate-zone-specific VPD thresholds using piecewise regression and GAM, and the application of Structural Equation Modelling to disentangle direct and indirect VPD pathways on E components. However, several issues, such as the propagation of ERA5 biases into SEM path coefficients (see my comment n. 9), and others are listed below, require further clarification/improvement before potential publication of the ms in HESS.
Specific comments:
1. Lines 27–29: The climatic-gradient thresholds are reported inconsistently. In the Abstract, the arid-zone threshold is given as 1.31 kPa, whereas in Section 4.1 and the Conclusion (Line 468) it is given as 1.67–1.68 kPa. The same discrepancy applies to the temperate zone (0.68 kPa in the Abstract; 1.67–1.68 kPa referred to in Section 4.3, Line 407). The authors must reconcile these values, clarify whether the Abstract reports the GAM-derived or piecewise-derived threshold, and ensure consistency throughout.
2. Section 2.2.2: The "space-for-time" moving-window approach to compute dE/dVPD is borrowed from land-use-change literature, where the logic is well-established for discrete cover transitions. Applying it here to derive VPD sensitivity assumes that spatial VPD gradients at 5 × 5 km are driven solely by biophysical feedbacks, rather than by mesoscale atmospheric dynamics. Please provide justification for this assumption for the global domain, especially in complex terrain or coastline-adjacent pixels. Furthermore, the authors should discuss the robustness of this assumption and, if possible, offer a cross-validation against pixel-level temporal regression (dE/dVPD from time series at each pixel).
3. Line 164: The moving-window size is stated as "5 × 5 km based on previous studies," but no citation is provided, and the dataset native spatial resolutions (ERA5 ~31 km, GLEAM ~25 km, MODIS MCD12C1 ~5.5 km) are substantially coarser than 5 km. This creates a fundamental mismatch: in practice, a 5 × 5 km window over ERA5 or GLEAM data contains a single pixel. The authors need to explicitly state the actual grid resolution at which the analysis was conducted and reconcile it with the stated window size.
4. Line 270: The global mean sensitivity of 293.27 ± 62.28 mm·hPa⁻¹·yr⁻¹ is a central quantitative result, yet the units are unconventional. Sensitivity (dE/dVPD) is dimensionally [mm / hPa], not [mm·hPa⁻¹·yr⁻¹]. The inclusion of yr⁻¹ suggests this may be derived from annual trend magnitudes rather than a pure partial derivative. Unless I have misunderstood the definition of what quantity this represents, e.g., is it a sensitivity coefficient, a trend-normalised ratio, or a regression slope?
5. Lines 296-303: The statement that "VPD rose abruptly in 1998 (0.034 hPa yr⁻¹)" reports a trend rate, not an abrupt change, and the phrasing conflates trend magnitude with event-driven anomaly. Additionally, Figure 3b shows VPD peaking at 7.839 hPa in 2010, the connection between the 1998 El Niño and the 2010 peak is not straightforward and warrants more rigorous treatment, potentially with ENSO-index partial correlation.
6. Lines 362-364: The SEM model is stated to explain 77% of the variance in E, and this is presented as a single global metric. However, SEM was presumably fitted globally on spatially aggregated, or grid-cell mean data. The authors should clarify the unit of analysis (pixel-level, regional mean, or annual global mean) and include model fit statistics (Fisher’s C, AIC) in the main text rather than only referencing Figure 5.
7. Lines 375-378: The finding that LAI increases markedly under high VPD (standardised coefficient = 0.90) but has only a weak direct effect on Et (standardised coefficient = 0.07) is counterintuitive and deserves deeper mechanistic discussion. A strong VPD–LAI relationship with near-zero LAI–Et effect suggests that either collinearity in the SEM is absorbing the LAI pathway into the VPD direct effect, or that the LAI variable used (GIMMS V1.2 monthly) does not resolve the sub-seasonal dynamics at which VPD–stomatal coupling operates. Please, make sure to address this explicitly.
8. Lines 406-410: In the discussion of the SM–atmospheric water vapor feedback under high VPD in arid regions, the mechanism is described qualitatively without connecting it to the quantitative results established earlier (e.g., the SEM path coefficients or the dE/dVPD spatial distribution). The discussion would be very interesting if the authors could map the regions where VPD already exceeds the 1.67-1.68 kPa threshold (as identified in Section 4.1) onto the negative-sensitivity regions shown in Figure 2a to demonstrate that the feedback mechanism is already active.
9. Lines 431-436: The acknowledgement that ERA5 overestimates SM in arid regions (Kokkalis et al., 2024) is important, but its consequence is underplayed. For example, given that the SEM assigns the VPD–SM path a standardized coefficient of -0.84, the strongest single path in the entire model, a systematic positive bias in SM would directly attenuate this coefficient, inflating the apparent direct effect of VPD on E. This has a directionally predictable effect on the ms's central conclusions and should be discussed with explicit reference to the SEM coefficients.
10. Lines 444-449: The authors acknowledge that piecewise regression assumes a single breakpoint. Yet in the Conclusion (Lines 466-469), a full table of single-threshold values per climate zone is presented as a primary finding. The authors should either present evidence that the VPD–E response in each zone is indeed well described by a single breakpoint (e.g., by showing that a two-breakpoint model does not significantly improve fit) or explicitly qualify the reported thresholds as approximations of the dominant transition, avoiding overstatement of their precision.