the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 05 Dec 2025
Drivers of diurnal and seasonal dynamics of triple oxygen isotopes in atmospheric water vapor and precipitation at a Mediterranean forest site
Abstract. Triple oxygen isotopes are a powerful tracer of hydrological processes, yet their variability in atmospheric water vapor and the processes driving them remain poorly understood. We present a one-year record of triple oxygen and hydrogen isotopes of atmospheric water vapor (V) measured at four heights below and above a downy oak forest canopy at the AnaEE platform O3HP in the French Mediterranean. This vapor dataset is complemented by isotope data from rainfall and groundwater, as well as monthly measurements of stomatal conductance and transpiration. Our results demonstrate that 17O-excessV is principally driven by evaporation processes. Seasonal variations in 17O-excessV ranging from 33 ± 9 per meg in winter to 25 ± 6 per meg in summer, reflect evaporative conditions in oceanic moisture sources. Diurnal variations, particularly pronounced in summer, with daytime maxima around 33 ± 6 per meg and nighttime minima around 16 ± 7 per meg, are linked to local evapotranspiration and isotope exchange between leaf waters and the atmosphere. On a monthly scale, precipitation is generally close to isotope equilibrium with atmospheric water vapor, except in summer when rain re-evaporation occurs. At event scale, large deviations from isotope equilibrium can occur due to raindrop evaporation and incomplete re-equilibration. Our findings enhance the mechanistic basis for interpreting precipitation isotopes in paleoclimate context, improves the robustness of isotope-based model evaluation, and highlights the potential of 17O-excess for better understanding of land-atmosphere water exchange across diverse climate and vegetation contexts.
- Preprint
(7986 KB) - Metadata XML
-
Supplement
(4652 KB) - BibTeX
- EndNote
Status: final response (author comments only)
-
RC1: 'Comment on egusphere-2025-5879', Anonymous Referee #1, 29 Dec 2025
The manuscript by Voigt et al. provides an important observation dataset of triple oxygen isotopic composition of water vapor, precipitation and surface/groundwater for one year in a forest in southern France. I believe this valuable dataset could be of interest to a broad scientific community working with stable water vapor isotopes. I also believe this is one of the very few studies focused on variations of the triple oxygen isotope composition in atmospheric water vapor. However, I have some concerns regarding this study, most of them are actually related to the measurement of 17O-excess in water vapor. That said, I'd like to suggest the manuscript to be deeply revised before continuing. Hope my comments can provide useful insights in this regards.Major comment #1: Measurements of 17O-excess in atmospheric water vapor using Picarro CRDS analyzers are challenging. Even under optimal conditions, averaging for long time is required to discriminate a meaningful signal from instrumental noise. For L2140 analyzers, the typical 1-second Allan deviation is on the order of ~0.1–0.2 ‰ for both δ18O and δ17O, which results into ~100 per meg uncertainty in 17O-excess . Achieving a precision of ~10 per meg therefore generally requires 10–20 minutes of averaging.These numbers are indicative, as each L2140 analyzer has specific performance. However, the key point is that such averaging reduces uncertainty only if the noise is white. Atmospheric water vapor at canopy scale is unlikely to exhibit white-noise characteristics on ~70 minute timescales due to e.g. turbulence and boundary-layer dynamics. In the presence of autocorrelation or colored noise, time averaging does not automatically reduce uncertainty as √N; instead, Allan deviation typically reaches a plateau or even increases at longer integration times in these conditions.This limitation likely explains why measurements of 17O-excess in atmospheric water vapor are relatively rare and why cryogenic trapping followed by off-line analysis can still be a solution. In the context of this paper, the effective uncertainty of the reported 17O-excess values in the present study is likely underestimated, particularly for diurnal variations and vertical gradients between measurement heights that are of similar magnitude to the expected noise level.I therefore recommend that the authors clarify (i) the noise structure of their vapor measurements, (ii) how the optimal averaging time was determined, and (iii) how effective uncertainty accounting for non-white noise and autocorrelation was estimated. The authors report that precision was determined using a Monte Carlo simulation. It is not clear on which assumptions the simulation is based on. Does the simulation account for uncorrelated noise and ~static signal (such as the one obtained by the A0211 vaporizer?). Or the simulation also accounts for e.g. turbulence noise spectrum?I believe this is an important point to address, since the 17O-excess variability the authors observe is about the same order of magnitude of the uncertainty. This comment does not apply to the liquid measurement of precipitation, well and spring water.Major comment #2: The authors report water vapor observations using a multiple-inlet system (0.4 m, 1.5 m, 3.5 m, and 12.5 m). However, several key aspects of the sampling configuration are insufficiently described. In particular, the total length of the inlet lines is not stated, and it is not clear whether the individual lines are flushed continuously when not connected to the Picarro via the port distribution manifold. It is also unclear whether the analyzer is connected directly to the inlet distribution manifold or whether an intermediate volume is present (e.g. a buffer or the A0211 vaporizer). Moreover, the reported analyzer flow rate (~0.4 mL/min) is low for such analyzers. My concern is that the combination of potentially unflushed inlet lines and low analyzer flow rates have impact on residence time and memory effects within the sampling system, resulting into signal smoothing and carryover between successive inlets. This is especialy true when switching between heights characterised by different humidity and isotopic composition. With the information reported by the authors, it is difficult to assess whether the (lack of) observed variability between inlets reflect true atmospheric variability or are partly influenced by sampling-system artefacts. A long-term averaging (e.g. daily or even seasonal) might also have flattened any difference in isotopic composition of water vapor between the inlets at different heights. I recommend that the authors provide a more detailed description of the inlet configuration and flow scheme, e.g. with a diagram, and, if possible, include quantitative tests or estimates of memory effects associated with inlet switching.Major comment #3: The statistical analysis of air-parcel origins based on back-trajectory frequency does not directly identify moisture sources of precipitation or boundary-layer water vapor origin. In particular, in Section 3.6 (lines 315–320), the attribution of moisture sources remains qualitative, as the analysis considers only the geographic origin of air masses and not the actual moisture exchange of air parcels with the ocean surface (evaporation) or land (evapotranspiration).This distinction is important because air-mass origin does not necessarily correspond to moisture source, strongly limiting boundary layer water vapor mass-balance calculation/speculations. During atmospheric transport, water vapor undergo several processes, to mention a few: condensation, re-evaporation, mixing, local evapotranspiration, etc. All of these processes substantially modify the isotopic composition of water vapor. Consequently, caution is required when attributing isotopic signatures to specific moisture sources based solely on back trajectories, particularly for δ18O and δD, whose variability is influenced by several factors such as local temperature and convective activity. In this context, d-excess and likely 17O-excess (as suggested by the manuscript itself) represents a more robust parameter for investigating moisture source conditions.A more statistically and physically consistent approach would be to combine back trajectories with isotopic observations at the O3HP site, for example using concentration weighted trajectories (implemented in the HYSPLIT framework employed by the authors) which have proven effective in previous studies in the Mediterranean region (e.g. Salamalikis et al., 2015). Another option would be to use a Lagrangian moisture diagnostics that explicitly identify moisture uptake along trajectories (e.g. following Sodemann et al., 2008, as mentioned by the authors in the conclusion). The latter would provide a more rigorous basis for source attribution and mass-balance. Such approach have been successfully applied to disentangle the drivers of water vapor d-excess variability in continental boundary layers (e.g. Aemisegger et al., 2014).At present, the message conveyed in lines 321–329 is unclear. The interpretation would benefit from rephrasing and acknowledging the limitations of the source attribution method used in this study. A similar caution applies to discussion section 4.1.Finally, given the proximity of the study site to the Mediterranean Sea, one might also expect a stronger influence from the western Mediterranean and central Europe. The seasonal source decomposition presented by Sodemann and Zubler (2010; their Fig. 8) could serve as a useful reference framework for contextualizing the results (although the latter paper is focused on precipitation only).The authors should clarify the distinction between air-mass pathways and moisture sources, and edit the interpretation accordingly.
- Aemisegger, F., Pfahl, S., Sodemann, H., Lehner, I., Seneviratne, S. I., & Wernli, H. (2014). Deuterium excess as a proxy for continental moisture recycling and plant transpiration. _Atmospheric Chemistry and Physics_, _14_(8), 4029–4054. https://doi.org/10.5194/acp-14-4029-2014
- Salamalikis, V., Argiriou, A. A., & Dotsika, E. (2015). Stable isotopic composition of atmospheric water vapor in Patras, Greece: A concentration weighted trajectory approach. _Atmospheric Research_, _152_, 93–104. https://doi.org/10.1016/j.atmosres.2014.02.021
- Sodemann, H., Schwierz, C., & Wernli, H. (2008). Interannual variability of Greenland winter precipitation sources: Lagrangian moisture diagnostic and North Atlantic Oscillation influence. _Journal of Geophysical Research Atmospheres_, _113_(3), 1–17. https://doi.org/10.1029/2007JD008503
- Sodemann, H., & Zubler, E. (2010). Seasonal and inter-annual variability of the moisture sources for alpine precipitation during 1995-2002. _International Journal of Climatology_, _30_(7), 947–961. https://doi.org/10.1002/joc.1932
Minor comments:- Abstract L20: Please clarify whether the authors refer to evaporation or evapotranspiration.- L104-115: I believe three mixing ratio sensitivity experiments are sufficient, assuming no change in the experimental setup during the study period. Please also report how may calibrations were performed (daily, weekly, etc). Was the water vapor Picarro also calibrated for the mixing ratio?- L138: please correct 0.4 - 10 m agl.- L168-169: The sentence is unclear as currently written. Please rephrase.- I suggest framing the geographical location of the study area more clearly in the context of the Mediterranean basin and the Atlantic Ocean. This information is currently discussed later in the manuscript (see Fig. 4), but it would be helpful to introduce it earlier, together with the description of the study site. It would be nice to have a single Figure showing the geographical location of the study area and a photo of the study site. Given that surrounding vegetation and landscape seem to be important actors in isotope processes at the ecosystem scale, visual context would be valuable. More on this, I have googled the coordinates and the observatory seems to be on a mountain slope. Do the authors expect wind direction or local circulation (e.g. mountain–valley breeze systems) to influence local moisture sources or boundary-layer dynamics?- A schematic representing the measurement setup is missing. A simple diagram illustrating the measurement setup would greatly improve clarity.- Modeling the isotopic evaporation flux (L180): if the authors refer to the exponent n controlling the ratio of water vapor isotopolues diffusivities in air (see, e.g eq. 17 in Horita et al., 2008) it should be noted several works show smaller effective values of n, following e.g. Duetsch et al. (2025, Figure 2) and experimental evidences in references therein.- Horita, J., Rozanski, K., & Cohen, S. (2008). Isotope effects in the evaporation of water: A status report of the Craig-Gordon model. _Isotopes in Environmental and Health Studies_, _44_(1), 23–49. https://doi.org/10.1080/10256010801887174
- Duetsch, M., Fairall, C. W., Blossey, P. N., & Fiorella, R. P. (2025). _A new theoretical framework for parameterizing nonequilibrium fractionation during evaporation from the ocean_. https://doi.org/10.22541/essoar.174785883.32412797/v1
- Figure 1 c: evidence of fractionated precipitation during summertime (negative d-excess)- L284: For δ18O and δD might be of interest of comparing with similar studies that were performed for long time-intervals in different climate settings such as Chen et al., 2024, Deshpande et al. 2010.- Chen, M., Gao, J., Luo, L., Zhao, A., Niu, X., Yu, W., Liu, Y., & Chen, G. (2024). Temporal variations of stable isotopic compositions in atmospheric water vapor on the Southeastern Tibetan Plateau and their controlling factors. _Atmospheric Research_, _303_, 107328. https://doi.org/10.1016/j.atmosres.2024.107328
- Deshpande, R. D., Maurya, A. S., Kumar, B., Sarkar, A., & Gupta, S. K. (2010). Rain-vapor interaction and vapor source identification using stable isotopes from semiarid western India. _Journal of Geophysical Research Atmospheres_, _115_(23), 1–11. https://doi.org/10.1029/2010JD014458
- Figure 4: The caption length should be reduced and discussion of the results should be moved from the caption to the main text (sect. 3.4)- L342-346: The discussion of NAO phases and water vapor isotope composition is interesting but only briefly mentioned. It is worth noting that other studies identified strong links between NAO phases, weather regimes and ground-level water vapor and precipitation isotope composition e.g.: Deininger et al. (2016) for precipitation over continental Europe in winter, Zannoni et al. (2019) for water vapor at ground level in the Mediterranean basin. A short expansion or contextualization would strengthen this section.- Deininger, M., Werner, M., & McDermott, F. (2016). North Atlantic Oscillation controls on oxygen and hydrogen isotope gradients in winter precipitation across Europe; implications for palaeoclimate studies. _Climate of the Past_, _12_(11), 2127–2143. https://doi.org/10.5194/cp-12-2127-2016
- Zannoni, D., Steen-Larsen, H. C., Stenni, B., Dreossi, G., & Rampazzo, G. (2019). Synoptic to mesoscale processes affecting the water vapor isotopic daily cycle over a coastal lagoon. _Atmospheric Environment_, _197_(March 2018), 118–130. https://doi.org/10.1016/j.atmosenv.2018.10.032
- L346-347: Water vapor isotopic composition measured near the surface can be representative of the boundary layer, but its variability may be modulated by ground roughness/homogeneity, convection, mixing, climate and weather regime, time of the day among the most important. Aircraft and tall-tower measurements provide strong evidence in this regard: see e.g. Salmon et al. (2019), Griffis et al. (2016)- Griffis, T. J., Wood, J. D., Baker, J. M., Lee, X., Xiao, K., Chen, Z., Welp, L. R., Schultz, N. M., Gorski, G., Chen, M., & Nieber, J. (2016). Investigating the source, transport, and isotope composition of water vapor in the planetary boundary layer. _Atmospheric Chemistry and Physics_, _16_(8), 5139–5157. https://doi.org/10.5194/acp-16-5139-2016
- Salmon, O., Welp, L. R., Baldwin, M., Hajny, K., Stirm, B., & Shepson, P. (2019). Vertical profile observations of water vapor deuterium excess in the lower troposphere. _Atmospheric Chemistry and Physics_, _19_(17), 11525–11543. https://doi.org/10.5194/acp-19-11525-2019
- L391-394, L407-409, Figure 5.i: what the 17O-excess composition of the seawater source should be to agree with observed atmospheric water vapor value?- L427: compounded forms such as "kppmv" are not recommended under metrology standards. Just reports "thousands of ppmv" or simply "< 4000 ppmv"- L428-430: Rayleigh distillation assumes isotopic equilibrium. These conditions are applicable inside a cloud (100% RH), but not for most of the conditions reported in this study.- L590 (data availability) PANGAEACitation: https://doi.org/10.5194/egusphere-2025-5879-RC1 -
RC2: 'Comment on egusphere-2025-5879', Anonymous Referee #2, 08 Jan 2026
Summary:
The authors present one of the first sets of continuous measurements of triple oxygen and hydrogen isotopes in atmospheric water vapor from multiple heights within and above a forest canopy for one year. They seek to identify the processes that change the isotopic values at the diurnal and seasonal time scales, as the title clearly indicates. They conclude that local evapotranspiration during the daytime and water vapor exchange with leaf water at night control the diurnal cycles. They conclude that seasonal variability is consistent with ocean moisture source signals changing environmental conditions, temperature and relative humidity. Also, the paper presents an analysis showing that the isotopic signatures of the potential oceanic moisture sources are not sufficiently different to use the observations to detect source region changes at the event or seasonal scale. However, they do show that vapor 17O-excess mean values are inconsistent with common assumptions made for ocean evaporation isotopic composition and suggest that more 17O-excess sea water observations are needed to resolve this discrepancy.
General comments:
Overall, this study is valuable because 17O-excess measurements in vapor are rare. For that reason, it will be of interest to the research community to quantify the ranges of variability observed. The paper is written clearly and the figures are helpful. However, there are several omissions the authors make in their interpretations and other edits that would improve the quality of the manuscript.
- The interpretation of the diurnal isotopic variability neglects the influence of vertical mixing within the PBL that has been identified in many previous studies of d18O, dD, and deuterium-excess. See Griffis et al., 2016 and references therein. It is reasonable to believe that this process is also a major influence on 17O-excess.
- Additionally, the influence of nighttime water exchange between the surface and the atmosphere does not have to be confined to leaf waters. Soil water vapor exchange may also play a role as noted by Berkelhammer et al. 2013.
- In the conclusions, it’s stated that 17O-excess can separate evaporation and transpiration signals. D-excess has this potential too, but neither is shown in this dataset. Please provide support or remove that conclusion.
- Please provide more description of how uncertainty in 17O-excess and deuterium-excess are quantified. See specific comments below.
- As reviewer #1 mentions, the moisture source region discussion is rather qualitative, and I agree with many of their points.
Specific comments:
Section 2.1: Include canopy LAI, tree stem density, or other metric of closed vs. open canopy so that vertical mixing with the boundary layer can be put into context.
Line 37: unclear meaning: ‘recharge conditions during lake evaporation (Surma et al., 2015, 2018)’
Line 66: Do these site acronyms mean anything? ‘AnaEE in natural experimental platform O3HP’
Section 2.2: 5 L/min continuous flushing of inlet lines with 70 min switching is likely sufficient for d18O and dD, but I’m not familiar with typical memory timescales for 17O-excess. While I don’t expect d17O memory to be longer than d18O, a demonstration of step-change response curves or discussion of equilibration time for this system would be appreciated.
Line 115: ‘Precision was better than 0.1 ‰, 0.2 ‰, 1.8 ‰ and 14 per meg, and 0.9 ‰ for δ17O, δ18O, δ2H, 17O-excess, and d-excess, respectively.’ How was precision quantified? Averaging over a defined time period? Propagation of error of d18O and dD would result in a d-excess uncertainty of 3.4 permil, much greater than the authors’ estimate of 0.9 permil.
Line 191: ‘The average wind speed at 10 m height was 0.3 m s-1’ How is the stdev determined?
Line 195: ‘The largest day-night differences occurred near the ground and in summer (Fig. A1-A3). During the day, q and Tair were highest near the ground, whereas RH was vertically homogenous (Fig. A1-A3). At night, the Tair vertical gradient was inverted, with the lowest values occurring near the ground, while q showed no vertical variation, resulting in decreasing RH from the surface to 10 m agl (Fig. A1-A3).’ Assuming this refers to summer conditions, wouldn’t it be unusual for air temp to be cooler near the surface than above when the surface usually radiates heat at night?
Section 3.6: what frequency of observations was this analysis based on? Monthly means or hourly resolution?
Fig 6: 17O-excessv is different for summer and winter. Driven by RHsst differences. The process label ‘E’ is not described in the text.
Section 4.1: These patterns are consistent with many other continuous water vapor isotope studies including d-excess and citations should be included. The novelty here is adding on 17O-excess and comparing how it’s similar/different from d-excess as it’s sensitive to many of the same influences.
Line 371: ‘thus the isotope compositions of water vapor evaporated from these sources are *predicted to be* similar.’ I suggest qualifying this statement because later the authors mention uncertainty in the inputs for calculating these estimates.
Line 420-430: could use improved citations and is misleading in present form. Berkelhammer et al., 2013 and references therein attribute diurnal variability to increased vertical mixing or PBL entrainment during the daytime as well as vapor exchange with the surface (leaf water and soil water) at night. The authors minimize this process because PBL air is drier, but nevertheless, it has been previously shown to be important and is likely contributing here also. Some well-documented entrainment influence citations like Griffis et al., 2016 and Welp et al., 2012 should be included.
Line 430: The authors argue that Rayleigh distillation is not expected to lead to 17O-excess enrichment in the lower free troposphere. Xia et al., 2023 shows theoretical deuterium-excess and 17O-excess signals increase as Rayleigh distillation intensifies.
Line 440-443: Ice/snow formation in cloud is likely not in isotopic equilibrium with cloud vapor due to super saturation effects Dutsch et al., 2019 and Xia et al., 2023.
Line 467: main conclusions are mostly asserted by theory rather than using observations as an independent test.
Line 473: The findings in this paper contrast previous findings that mediterranean moisture source has a high deuterium-excess compared to North Atlantic, both in theoretical predictions and observations of downwind vapor and precipitation. What makes this study different? Different assumptions about conditions in the source regions? Perhaps ERA5 doesn’t accurately capture near-surface RH differences? Does monthly resolution mute signal? Etc.
Fig 2: d-excess deserves error bars also
Fig 3: What frequency of data was used for the monthly means? 15-min or hourly, etc. Is the plot a smoothed spine fit through discrete values?
Fig A8: please include d-excess vs 17O-excess correlation here or somewhere in the paper
Fig A10: There are more statistics that could be helpful for the reader. The spread of back trajectories and/or the frequency of each cluster over the year.
Table B2: How is the SD of veq determined. The precip is collected monthly and event scale during some months, so are different methods used?
References:
Berkelhammer, M, J Hu, A Bailey, et al. “The Nocturnal Water Cycle in an Open-Canopy Forest.” Journal of Geophysical Research-Atmospheres 118, no. 17 (2013): 10,225-10,242. https://doi.org/10.1002/jgrd.50701.
Dütsch, Marina, Peter N. Blossey, Eric J. Steig, and Jesse M. Nusbaumer. “Nonequilibrium Fractionation During Ice Cloud Formation in iCAM5: Evaluating the Common Parameterization of Supersaturation as a Linear Function of Temperature.” Journal of Advances in Modeling Earth Systems 11, no. 11 (2019): 3777–93. https://doi.org/10.1029/2019MS001764.
Griffis, Timothy J, Jeffrey D Wood, John M Baker, et al. “Investigating the Source, Transport, and Isotope Composition of Water Vapor in the Planetary Boundary Layer.” Atmospheric Chemistry and Physics 16, no. 8 (2016): 5139–57. https://doi.org/10.5194/acp-16-5139-2016-supplement.
Xia, Zhengyu, Jakub Surma, and Matthew J. Winnick. “The Response and Sensitivity of Deuterium and 17O Excess Parameters in Precipitation to Hydroclimate Processes.” Earth-Science Reviews 242 (July 2023): 104432. https://doi.org/10.1016/j.earscirev.2023.104432.
Citation: https://doi.org/10.5194/egusphere-2025-5879-RC2
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 173 | 92 | 23 | 288 | 28 | 28 | 29 |
- HTML: 173
- PDF: 92
- XML: 23
- Total: 288
- Supplement: 28
- BibTeX: 28
- EndNote: 29
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1