the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 06 Mar 2026
Soil–atmosphere water vapor exchange in semi-arid Northwest China: New insights from fiber-optic relative humidity sensing
Abstract. Soil–atmosphere water vapor exchange in arid and semi-arid regions is a key process in near-surface hydrology, reflecting the dynamic coupling of surface energy and moisture. In this study, a novel fiber-optic sensing technique was employed to measure vertical water vapor fluxes across the soil–atmosphere interface in a semi-arid region of the Loess Plateau, Yanan, China. The observations captured vapor flux dynamics across a 7-mm dry soil layer beneath the interface (hereafter referred to as Flux Layer soil) and a 10-mm molecular diffusion layer in the air above it (Flux Layer air), revealing how meteorological factors modulate near-surface vapor transport. Solar radiation enhanced vapor fluxes primarily by increasing the vapor pressure deficit (VPD), with Flux Layer soil exhibiting a slower response than Flux Layer air. This lag was most pronounced in winter, reaching up to 120 minutes. During rainfall, fluxes in both layers declined sharply as VPD dropped to near zero. Following precipitation, Flux Layer air recovered rapidly, driven by surface evaporation, while Flux Layer soil increased more gradually due to the progressive drying of subsurface moisture. Structural equation modeling based on 5657 observations revealed distinct influence pathways: Flux Layer air was more sensitive to solar radiation, air temperature, and VPD, while Flux Layer soil was predominantly governed by VPD. These findings advance the quantitative understanding of near-surface vapor transport mechanisms and improve insight into the coupled feedbacks governing soil–atmosphere interactions under variable climatic conditions.
- Preprint
(1441 KB) - Metadata XML
-
Supplement
(788 KB) - BibTeX
- EndNote
Status: open (until 23 Apr 2026)
-
RC1: 'Comment on egusphere-2026-1143', Anonymous Referee #1, 08 Apr 2026
reply
-
AC2: 'Reply on RC1', Junyi Guo, 20 Apr 2026
reply
We sincerely thank the reviewer for the positive, careful, and constructive evaluation of our manuscript. We are encouraged that the reviewer finds the study interesting and potentially valuable for improving the understanding of soil–atmosphere water vapour exchange in semi-arid regions. We also greatly appreciate the reviewer’s detailed and helpful suggestions regarding the description of the study site, the field setup, calibration and validation procedures, figure clarity, and the explicit discussion of study limitations. We agree that these aspects are not presented with sufficient clarity in the current manuscript. In the revised manuscript, we will substantially improve the description of the study site, move key methodological information from the Supplementary Material into the main text, clarify the calibration and validation procedures as well as possible sensor-related uncertainties, correct the inconsistency in the time range shown in Figure 2, and add a dedicated discussion of the limitations and applicability of the study. We believe that these revisions will improve the clarity, transparency, and overall quality of the manuscript. Comment 1 The site description lacks information on vegetation cover and land use. Figure S1 should also be moved from the Supplementary Material into the main text. Response: We thank the reviewer for this helpful suggestion. We agree that the current manuscript does not provide sufficient information on vegetation cover and land use around the study site. Although the main analysis in this manuscript focuses on bare-soil conditions, we agree that the surrounding vegetation and land-use context should be described more clearly, because these factors may influence soil–atmosphere exchange and the interpretation of the results. In the revised manuscript, we will add a fuller description of vegetation cover and land use around the site. We will also clarify whether the bare-soil monitoring area was naturally bare or prepared for experimental purposes. In addition, we agree that the site context would be easier for the reader to understand if the study-site figure were included in the main text. Therefore, we will move the current Figure S1 from the Supplementary Material into the main manuscript under the study-site section. Comment 2 The FBG system is central to the study, but key information is currently left in the Supplementary Material. Figure 1 is difficult to follow and should be improved. Response: We thank the reviewer for this important comment. We agree that the FBG system is the central methodological approach of this study and that the current manuscript relies too heavily on the Supplementary Material for key technical information. In the revised manuscript, we will move a more complete description of the FBG monitoring system into Section 2.2 of the main text so that readers can follow the experimental design and sensing principle more easily. We also agree that Figure 1 should be made clearer. In the revised manuscript, we will revise Figure 1 to improve its readability and make the sensor configuration, probe layout, and measurement concept easier to understand. Comment 3 The sentence referring to previous sensor evaluations is unclear. Please specify what evaluations are being referred to. Response: We thank the reviewer for pointing this out. We agree that the current wording is too general and does not clearly explain what is meant by “previous evaluations.” In the revised manuscript, we will rewrite this sentence to specify more clearly which performance characteristics had been evaluated previously. In particular, our earlier calibration and performance assessment showed that the OR-FBG hygrometer exhibited low hysteresis, with a hysteresis error of ±1.87%. The response time during humidification was approximately twice that during dehumidification. After repeated humidity cycles, only a slight shift in the humidity-sensitive calibration curve was observed, and the repeatability error remained within ±2.43%, indicating good repeatability. We will make clear that these statements are based on our previous calibration and sensor-performance evaluation work. Comment 4 Please clarify the separation distance among the three identical sets of FBG sensors and whether they were installed on the same soil type. Response: We thank the reviewer for this useful comment. We agree that the present manuscript does not explain the arrangement of the three FBG profiles clearly enough. In the revised manuscript, we will clarify that the three sensor profiles were installed less than 1 m apart and that all of them were deployed in loess, while also describing the differences in their near-surface microenvironmental conditions more explicitly. We will also revise the wording so that the relationship between these profiles and the intended assessment of spatial variability is described more clearly. Comment 5 The experimental setup section contains important gaps, especially regarding calibration, validation, and the treatment of potential sensor biases. Response: We thank the reviewer for highlighting this important methodological issue. We agree that the current manuscript does not describe the calibration, validation, and treatment of possible sensor-related uncertainties with sufficient clarity. In the revised manuscript, we will provide a more explicit description of the calibration and validation procedures used in this study and clarify how possible uncertainties and sensor-related biases were considered and handled. Because relative humidity is strongly temperature dependent, calibration was carried out using eight sealed humidity-control boxes containing different saturated salt solutions, which were placed in a split-type temperature-controlled chamber with an accuracy of 0.1 °C. By precisely controlling the chamber temperature, each saturated salt solution was allowed to reach a stable two-phase equilibrium, thereby providing the corresponding reference relative humidity at a given temperature. To verify humidity stability inside each control box, an electronic hygrometer was introduced, and the condition was considered stable when the readings fluctuated within the instrument error range. To minimize possible bias caused by vertical gradients inside the box, the probe of the electronic hygrometer and the FBG hygrometer were installed at the same height. Small openings in the lid were used to introduce the probes, and these openings were sealed to maintain airtight conditions during the tests. The performance evaluation of the coated FBG hygrometer included both feasibility tests and sensor-characterization tests. In the feasibility tests, humidity sensitivity and temperature sensitivity were examined. For the humidity-sensitivity test, the chamber temperature was fixed at 25 °C, and the FBG hygrometer was sequentially exposed to eight saturated salt environments in order of increasing relative humidity and then decreasing relative humidity to complete one full humidification–dehumidification cycle. The wavelength response was recorded every 5 s until stabilization. To evaluate temperature–humidity coupling, the same procedure was repeated over a temperature range from 10 to 60 °C at 10 °C intervals. For the temperature-sensitivity test, the temperature was increased stepwise from 10 to 60 °C under several fixed relative humidity conditions (32.8%, 57.7%, 75.3%, and 84.3% RH), and the corresponding wavelength changes were recorded. In the revised manuscript, we will summarize these procedures more clearly and explain how they support the evaluation of sensor accuracy, hysteresis, repeatability, and possible sources of uncertainty. Comment 6 Vegetation effects are only briefly mentioned in the Supplementary Material, and Figure S4 is difficult to relate to the bare-soil setup. More context is needed, and site photographs would also be helpful if possible. Response: We thank the reviewer for this constructive suggestion. We agree that the relationship between the bare-soil analysis in the main text and the supplementary material, especially Figure S4, is not sufficiently clear in the current version. In the revised manuscript, we will improve the description of the different monitoring conditions and clarify more explicitly how the supplementary figures relate to the main bare-soil setup. We also agree that additional visual context would help readers better understand the site conditions and field arrangement. Therefore, we will improve the presentation of the site setup and, where possible, include photographic information or equivalent visual clarification to make the field conditions easier to interpret. Comment 7 There is a discrepancy in Figure 2: the date range shown is 2022–2025, whereas the monitoring period described in the manuscript is August 2022 to June 2023. Response: We thank the reviewer for noting this inconsistency. This is an error in the current figure labeling. The monitoring period used in this study is August 2022 to June 2023, and we will correct the date range shown in Figure 2 accordingly in the revised manuscript. Comment 8 The manuscript does not explicitly state the limitations of the study. Limitations should be discussed, including the focus on bare-soil surfaces, the less-than-two-year monitoring period, the limited number of vertical measurement positions, and the methodological boundary relative to conventional flux measurements. Response: We thank the reviewer for these valuable suggestions. We fully agree that the limitations of the study should be stated more explicitly. In the revised manuscript, we will add a dedicated paragraph discussing the main limitations and applicability boundaries of the present study. Specifically, we will clarify that: (i) the present analysis focuses on bare-soil conditions, while only preliminary comparative observations under vegetated and shaded conditions are shown in the Supplementary Material, which limits the direct generalization of the results to vegetated landscapes; (ii) the monitoring period extends from August 2022 to June 2023 and therefore does not capture longer-term climatic variability over multiple years; (iii) measurements were made at only three vertical positions, which provide very fine but still limited vertical resolution; and (iv) the observations are specific to this fiber-optic approach and are not directly equivalent to conventional flux measurements such as eddy covariance. We agree that making these limitations more explicit will improve the transparency of the manuscript and strengthen the interpretation of the results and conclusions. Comment 9 The reviewer would better appreciate the Results and Conclusions after clearer site information is provided, especially regarding land use, land cover, and related variables that may affect the interpretation. Response: We thank the reviewer for this comment and agree with this point. We recognize that a clearer presentation of the site characteristics, including land use, land cover, and related contextual variables, will help readers better interpret the reported results and conclusions. In the revised manuscript, we will improve the site description accordingly and ensure that the relationship between site conditions and the interpretation of the results is stated more clearly. We sincerely appreciate the reviewer’s constructive and encouraging comments. We believe that these suggestions will significantly improve the clarity, transparency, and overall quality of the manuscript. In the revised version, we will incorporate the above changes carefully and systematically.
Citation: https://doi.org/10.5194/egusphere-2026-1143-AC2
-
AC2: 'Reply on RC1', Junyi Guo, 20 Apr 2026
reply
-
RC2: 'Comment on egusphere-2026-1143', Anonymous Referee #2, 19 Apr 2026
reply
Review of the manuscript titled ‚Soil–atmosphere water vapor exchange in semi-arid Northwest China:
New insights from fiber-optic relative humidity sensing’ by Guo et al. for potential publication in HESS:
This manuscript attempts at quantifying the water vapor transfer in the soil-air interfacial layer by means of a vertical profile from 3 point based-sensors employing Fiber Bragg Grating for temperature and humidity. Observational records from a semi-arid site in China span about 10 months and include some rudimentary ancillary measurements of climate elements as an upper aerial boundary condition, without including observations of the lower within-soil deeper boundary. The temporal resolution of the fiber-based quantities is 10 min, that of the climate elements 60 min. Results are presented for short hand-picked case studies during clear-sky conditions and post rain events for the bare soil subsite across seasons. Structural equation modeling is employed to statistically determine environmental drivers of the evaporative flux across the soil and air layers.
Summary evaluation:
This study addresses a substantial and well-known, yet underexplored knowledge gap for evaporation at the air-soil interface at scales larger than the molecular, but smaller than the eddy covariance scale: all water vapor has to pass through the shallow interfacial boundary dominated by pore-to-pore gas diffusional transport within the soil and the directly adjacent laminar and viscous sublayers in the lowest 1 cm of the air. Here, the mass transport is dominated by molecular diffusion (within soil to 1mm above soil in air) but turbulence is already starting to take hold (1 to 10 mm in air). Observations of the evaporative turbulent flux from eddy covariance near the surface at roughly 2m above ground layer (agl) and all land-surface models often neglect this conceptual ‘bottleneck’ of the mass and energy transports. It is an extraordinarily difficult to study and understand interfacial transport layer. Hence, I commend the authors for attempting at experimentally quantifying the evaporational flux across this shallow, but critical interface. The fiber-based optic sensors is a fresh sensing approach the authors have been pursuing for the past years with success (Guo et al. 2021, 2022, 2025). While I was and am excited about the potential of this sensing technology to provide new observations-based insights and contribution to fill our knowledge gap, I do not think the presented results of observations at the field scale are credible or novel. In fact, I have (i) large doubts about the experimental design and (ii) am concerned the statistical, shallow evaluation of the findings without any uncertainty analysis and/ or critical evaluation of the sensor performance. As far as I can tell, the presented findings for the clear sky and post-rain events do not hold any conceptual novelty either. While (ii) can be improved by post-field analysis and discussion, I am afraid the impact of (i) cannot be remedied but call for a new, more carefully designed field experiment to provide evidence for the physical meaningfulness of the observations. Maybe the authors have additional experimental data at other sites or subsites not suffering from the same shortcomings which can be used to address my concerns, but none of it is mentioned in the main manuscript or supplements. The latter do contain a glimpse of observations from a vegetated canopy, which raise even more questions and concerns.
I summary, I cannot recommend this manuscript for publication based on the concerns listed below.
Major remarks:
- Vertical scales of sensors and the soil-air interface, and observer effects: I do appreciate the inherent difficulties in studying the mass and energy transport across the shallow but critical interface, but even a simple comparison of vertical length scales across the 17mm thick (7mm in the soil, 10 mm in the air) layer raised major concerns that evaluated water vapor gradients, temperatures and distances (Eq. 1) are meaningful and free of observer effects: The diameter of the fiber-based Bragg Grating (FBG) sensor is 6.4mm (Figure 1b in current manuscript, identical to Fig.1 in Guo et al., 2025 in WRR, which – by the way – raises copyright concerns even if authors overlap to a large degree), leaving the radius half of it at 3.2mm. Given the vertical dimension of the soil layer of 7mm and assuming the depth of the horizontally oriented FBG sensor is determined at its center (see Fig. 1a of current manuscript), the enclosed mineral soil layer in between the lowermost and middle FBG sensor is Δzsoil= 7 mm – 2*(3.2mm) = 0.6 mm. The same computation for the air layer yields Δzair= 10 mm – 2*(3.2mm) = 3.6 mm air space between the middle and uppermost sensor. Hence, the sketch in Fig. 1a does not do the length scales any justice. These very small length scales of the studied media (mineral soil, air) in relation to the thickness of the sensors (ratio of 0.6/6.4 \approx 1:10, and 3.6/6.4 approx 1:2) inevitably lead to potentially dominant observer effects as the heat and water vapor conducting properties, i.e. the thermal and water vapor diffusivities- defined as the ratio of conductivity to capacity – of the powder-sintered stainless steel housing become more important that those of the targeted soil and aerial layers. Or, the heat and water vapor are conducted around or preferentially through the sensors, an ‘lens’ effect well-known in the soil community. I am sure the authors must have thought about this earlier when designing the experiment. Even from a practical point of view: how are sensors so carefully placed at such tiny spacings of Δz= 0.6 and 3.6 mm and more importantly maintained across the annual cycle with frost heave, snow fall, ablation, wind erosion (bare soil), etc? Were distances monitored? Were the sensors laterally displaced to avoid such observer effects? If yes, what was their separation and how isotropic and homogeneous were soil conditions? A related question is why results from only one, or the ensemble average of all 3 bare soil replicate subsites is presented as an average, and the inherent spatial variability not even mentioned or discussed?
- Energy balance and radiative artifacts of FBG sensors exposed to sunlight: The middle and uppermost FBG sensors were exposed to radiative shortwave and longwave forcing and (laminar to turbulent) airflow during snow-free periods during the day, and longwave forcing and airflow at night or when covered with snow (well, depending on snow depth even shortwave forcing. While vapor pressure readings of polyimide-coated side (PI-FBG) of the sensor was compensated for temperatures measured at the uncoated side (B-FBG), the i) within-sensor separation between these surfaces is larger (soil layer) or at the same scale (air layer) compared to the vertical separation distances computed in comment (a), and ii) the sensor surely suffers from substantial radiative heating or cooling beyond/ below air temperature, the computed differences and gradients (differences across separation distance) are likely to be unphysical. The middle and uppermost FBG sensor is also subject to convective cooling, since turbulence can reach into the viscous sublayer when surface momentum roughness lengths are small (for bare soil on the scale of millimeters, i.e. at the scale of the upper half of the middle FBG sensor sticking out of the soil). The authors need to present a full energy balance of these FBG sensors including the radiative and convective terms in addition to their conductive transport analyses to eliminate artifacts and estimate measurement uncertainty. The lack of time lag between the peak in diurnal temperature traces at 10 mm, 0 mm and -7mm in Fig. 2 b) indicate the dominance of the radiative artifacts and support my concerns. The height of the measurements of ‘AirT’ from the weather station was never mentioned, so the reader cannot say anything about vertical gradients between 1cm and 2m (?) agl. I am concerned that ‘AirT’ heats up substantially earlier than any of the FBG sensors, which suggest topographic sheltering or obstruction at low sun angles at sunrise.
- The findings are presented for selected clear-sky and post-rainfalls events for one 3 point-based profile, or as the ensemble average across all three – why? No information about how and why these periods were selected is provided, which is important for estimating their representativeness. Simply stating that it was (e.g. ln 172, Section 3.2) is insufficient. Why do the authors present only case studies and do not attempt at computing a surface water balance given the novelty of their observations? Such an analysis would allow for quantifying the systematic error of their, or conventional flux observational and/ or modeling approaches. Comparison of their surface evaporation with that from EC would be tremendously insightful and yield potential shortcomings of either approach. Note that my concerns raised in comment (b) are largest for calm (wind speeds close to zero) and clear skies conditions, so why not contrast evaporational fluxes from their technique with those for cloudy (close to 8/8) and high wind speeds to minimize the artifacts? Were vertical temperature and vapor gradients for these days beyond the detection limit of the FBG sensors? Did they yield fluxes in accordance or contrary to our current process understanding and expectations?
- Time scales of lags, sensors response time, and averaging: Guo et al. (2025, WRR) states in their Table 1 that the FBG rH probe suffers from substantial hysteresis and has response times of 1370 s during ascending (wetting) and 750 s during descending (drying) humidity levels. It is unclear if these time scales on the order of 23 and 13 min, respectively, is the effective (i.e. 4.5 fold) or simple (1) time to reach (1-(1/e)) = 63% change. If it was the time constant (simply, 63%), then the effective sensors response time is the 4.5fold, i.e. on the order of 100 and 60 min. Too long to evaluate any processes less than these time scales. Please clarify. Given that FBG observations were averaged over 10 min (but maybe collected more frequently and simply stored as averages), but ancillary climate elements at 60 min averaging (which is very unusual), how can time lags between solar radiation and FBG sensors be reported at time scales of 30 min? (Fig. 4)? Were data interpolated? If yes, how? Were lags only identified for sunny days, or also for cloudy days? I would like to see a histogram across all days. Was the hysteresis in sensor response to rH considered in the discussion of lags?
- Lack of novelty of current findings and uninspired statistical analysis: even when setting all the experimental questions and concerns aside for a moment – and assuming the presented findings are meaningful: what do we learn from this study? Evaporation increases with net radiative forcing, becomes limited (aET/pET declining) with decreasing water availability, the soil evaporation responds lagged compared to the air, and snow and frozen soil moisture inhibit gas phase transport. We know all this already. Actually, if their sensors worked correctly, I think a detailed study of water vapor transport under snow could potentially be really exciting as snow is permeable to water vapor (if no ice layers are present) and transmittent to solar radiation, which may create substantial vapor pressure deficits in the interfacial layer – just to mention one potential science nugget.
Minor comments: Given the major remarks, I limit my comments to the most important ones.
- Please use terminology of ‘air and soil layers’, not Flux Layer_air and _soil.
- Ln 27 and later: avoid using wording ‘complexity’ without definition of context and clear definition of its metric.
- Ln 34: Evaporation is relevant not only in arid environments, but everywhere.
- Ln 38: how can the flux be constrained by the same flux itself? Revise.
- Section 1: we have a fairly good understanding of the macroscopic forcing of ET at t the stan scales, which is incorporated in the classic textbook equations of Penman, various forms of Penman-Monteith, Priestley-Taylor, so why not mention this an cite only recent semi-empirical studies? Demonstrating the command of the classic textbook knowledge helps shaping the knowledge gap and precisely formulating the research questions.
- Ln 50: anything is comparable, little is similar.
- Ln 89: use correct terminology of ‘climate elements’. Not meteorological factors.
- Ln 103: eliminate any mentioning and data from the canopy experiment also in the supplement: given you report largest temperatures for these subsites, the lack of convectional cooling by momentum absorption of the leaves and trapping of longwave radiation indicated that your bare soil observations are effected by these processes (see my major comment b).
- Ln 144: Use uniform citation formatting.
- (1) and others: define all symbols (also applied to supplement).
- Ln 140: what is a ‘vaild data point’ according to your definition? This is important to know as your filter data for unknown plausbility.
- Ln 155: the correct measurable thermodynamic quantity is enthalpy, not ‘heat’.
- Fig 2: why the diurnal asymmetry in solar forcing? Colors in panel b and c are difficult to discern, where was VPD (d) measured? In soil or air flux layer, or at weather station height? Combine e and f in one subpanel for comparison.
- Ln 180: show VPD for all 4 heights (FBG, weather station) since VPD in air should decline sharply after sunset.
- Ln 188ff: potentially super exciting, but meager discussion at the moment
- Ln 195: I sense a contradiction: define sharp vs gradual rates, then compare effects of cloud to that of sunset. Potential effects of radiative heating of sensors? wind speed?
- Fig .3: remove shading of lines.
- Section 5: please also compare and discuss classic ET models against your findings, the main forcings are identical. Not very surprising – so what did we learn? Why all this effort?
Citation: https://doi.org/10.5194/egusphere-2026-1143-RC2 -
AC1: 'Reply on RC2', Junyi Guo, 20 Apr 2026
reply
We sincerely thank RC2 for the very careful, critical, and technically insightful assessment of our manuscript. We fully recognize the seriousness of the concerns raised. In particular, we understand that the main reservations focus on: (1) the geometric scale mismatch between the sensor size and the very shallow soil–air interfacial layer, and the resulting potential observer effects; (2) possible radiative and convective artifacts affecting the exposed FBG sensors; (3) the interpretation of short-timescale lags in light of sensor response time and differing averaging intervals; and (4) whether the current manuscript demonstrates sufficient conceptual novelty beyond the established understanding of evaporation controls. We greatly appreciate the depth and professionalism of these comments and agree that, in its present form, the manuscript does not yet address these issues with the clarity, caution, and completeness required. At the same time, we respectfully believe that these concerns, while substantial, can be addressed through a major revision. In the revised manuscript, and in our detailed point-by-point response, we will provide clearer methodological clarification, additional analyses, and substantial textual revisions to address these issues more rigorously and transparently. Comment 1. Geometric scale mismatch, potential observer effects, and spatial variability RC2’s concern: The sensor size may be too large relative to the very shallow soil–air interfacial layer, which raises the possibility of observer effects. In addition, the current manuscript does not clearly explain the monitoring layout or the treatment of spatial variability. Response: We thank RC2 for this important comment. We agree that the current manuscript does not describe the sensor configuration and monitoring layout clearly enough, which may lead to confusion regarding sensor spacing, measurement representativeness, and the treatment of spatial variability. To clarify, the 7 mm and 10 mm vertical distances used in this study are defined as the geometric distances between the centers of adjacent sensing elements, that is, from fiber to fiber. This definition is consistent with both the spatial positioning of the sensing elements and the flux calculation method. In terms of sensing principle, the recorded signal mainly reflects the local condition in the immediate vicinity of the optical fiber, rather than a simple volumetric average within the probe. The optical fiber itself has a diameter of approximately 250 μm, which is much smaller than the 7 mm and 10 mm center-to-center spacing between adjacent sensing elements. In addition, the fiber is positioned at the center of the protective probe, which is why the spacing was defined based on sensing-element centers. The probe housing is gas-permeable and mainly serves to protect the fiber; accordingly, we did not redefine the spacing using a “net distance” obtained by subtracting the outer probe diameter. At the same time, we acknowledge the core of RC2’s concern: because the probe diameter is on the same order as the monitored layer thickness, the possible influence of the protective housing and supporting structure on the local temperature and humidity fields cannot be fully excluded in the present manuscript. In the revised manuscript, we will add a dedicated discussion of this issue and define more carefully the applicability and interpretive limits of the method for observations within the millimeter-scale soil–atmosphere interfacial layer. Regarding spatial variability, we also recognize that the present manuscript may give the impression that the three sensor sets were replicate bare-soil profiles. In fact, although the profiles had the same sensor configuration, they were deployed under different near-surface microenvironmental conditions. The main text focused on the bare-soil profile, whereas the other profiles were shown only in the Supplementary Material and were not discussed systematically. In the revised manuscript, we will clarify the monitoring layout and local conditions of each profile, avoid potentially misleading averaged expressions, and add a more explicit discussion of spatial differences, representativeness, and study boundaries. Comment 2. Possible radiative and convective artifacts affecting the exposed probes RC2’s concern: The exposed probes may be affected by shortwave radiation, longwave radiation, and local airflow, and the current manuscript does not sufficiently discuss how such effects may influence the measurements and interpretation. Response: We thank RC2 for highlighting this important issue. We agree that the present manuscript does not discuss this source of uncertainty sufficiently, and that the temperature compensation used in this study could be misunderstood as a complete correction for all thermal disturbances. In our system, PI-FBG is used to sense humidity-related signals, whereas B-FBG provides temperature information to correct the temperature cross-sensitivity of PI-FBG. The two fibers are positioned very close to each other inside the probe, as shown in Figure 1b, so that the compensation can effectively reflect the local thermal state around the humidity-sensitive element. However, this compensation is specifically intended to correct temperature cross-sensitivity in humidity measurement; it does not fully remove all thermal influences associated with the energy balance of the exposed probe. Thus, while the correction improves the reliability of humidity inversion, it does not by itself rule out possible radiative heating/cooling and local convective effects. We therefore agree that the middle and upper exposed probes may be influenced by radiation and local airflow, and that the derived vapor pressure differences and flux magnitudes should be interpreted with appropriate caution. In the revised manuscript, we will clarify the compensation principle, the relative positions of the two fiber types inside the probe, and the scope of applicability of this correction, and we will add a dedicated discussion of possible radiative and convective artifacts. We will also clarify the measurement height and usage of the meteorological AirT record, and explicitly state that AirT from the weather station was used only as a background meteorological reference, not as a direct measurement point for millimeter-scale gradient inversion. Regarding RC2’s comment on the limited visible phase difference among temperature peaks at 10 mm, 0 mm, and −7 mm in Figure 2, we agree that this point deserves careful discussion. However, the absence of a pronounced lag among temperature peaks alone is not sufficient to conclude that the observations are primarily controlled by radiative artifacts. At the millimeter scale, temperature signals generally respond faster than moisture- or vapor-related signals, and the lag in vapor flux is more closely related to the delayed response of humidity-related signals than to the lag of temperature peaks alone. In the revised manuscript, we will further distinguish the temporal behavior of temperature and humidity/vapor responses, revisit Figure 2, and add comparative analyses under different radiative conditions, such as sunny versus cloudy periods and daytime versus nighttime conditions. Comment 3. Short-timescale lags, sensor response time, and differing averaging intervals RC2’s concern: The present manuscript may over-interpret short-timescale lag relationships without sufficiently clarifying the role of sensor response time and the differing temporal resolutions of FBG and meteorological data. Response: We thank RC2 for raising this important issue. We agree that the current manuscript does not explain this aspect sufficiently. The previously reported response times of 1370 s for humidification and 750 s for dehumidification were obtained under controlled laboratory conditions for a step change in relative humidity from 31.3% to 61.6%. In contrast, field temperature and humidity variations are typically continuous and gradual. Therefore, these laboratory-derived response times should not be directly equated with the actual lags observed under field conditions. Under the 10 min temporal resolution used in this study, the sensor response characteristics are more appropriately understood as a limitation on the precise quantification of minute-scale absolute lag values, rather than as a basis for rejecting the identification of relative response sequencing among variables. Regarding RC2’s question about reporting 30 min-scale lags when 10 min FBG data and 60 min meteorological data coexist, we clarify that the meteorological data were linearly interpolated to a unified 10 min temporal resolution prior to lag analysis. We will state this explicitly in the revised Methods section and will correspondingly tighten the wording used to describe lag precision. In addition, we acknowledge that the lag discussion in the present manuscript is mainly based on selected representative periods. In the revised manuscript, we will define more clearly the sample range used for lag analysis and, where possible, include more comparisons across multiple periods and conditions, such as sunny versus non-sunny periods, different seasons, and different environmental settings. This will allow the lag discussion to be presented more cautiously and with clearer interpretive boundaries. Comment 4. Conceptual novelty beyond the established understanding of evaporation controls RC2’s concern: The conceptual novelty of the manuscript is not yet communicated clearly, and the current presentation may make the contribution appear too close to already established understanding of evaporation controls. Response: We thank RC2 for questioning the conceptual novelty of the manuscript. We understand the concern that, if the main findings are summarized only as well-known statements about radiation control, water limitation, slower soil response, and suppression under frozen conditions, the contribution may appear insufficiently novel. In our view, however, the main issue lies in how the novelty was expressed in the present manuscript rather than in the absence of substantive new information. The novelty of this study does not primarily lie in proposing an entirely new evaporation mechanism. Rather, it lies in advancing direct observation into the millimeter-scale soil–atmosphere interfacial layer, which is extremely difficult to access with conventional methods, and in using fiber-optic relative humidity and temperature sensing to achieve continuous in situ observations within this critical layer. In this sense, the key contribution of the study is to provide direct observational evidence from the interfacial “bottleneck” layer itself, rather than merely restating macro-scale evaporation controls. More specifically, we see the novelty of this work in three aspects. First, it provides rare cross-seasonal continuous observations from the millimeter-scale soil–atmosphere interfacial layer. Second, it resolves how macro-scale controls are expressed through directly observed temperature–humidity dynamics and soil-layer versus air-layer differences. Third, it helps bridge classical evaporation theory with direct interface-layer observations. In the revised manuscript, we will sharpen this statement of novelty, reduce wording that could be interpreted as claiming an entirely new evaporation law, and emphasize more clearly the manuscript’s contribution in terms of continuous in situ observation, differentiated layer responses, and scale-bridging interpretation. Overall, we accept RC2’s reminder that the novelty of the current manuscript has not yet been communicated clearly enough. However, we do not believe that the manuscript lacks substantive novelty. Rather, we believe that its main innovation lies in extending evaporation research from conventional near-surface macro-scale observations to continuous in situ observation within the millimeter-scale soil–atmosphere interfacial layer, thereby providing direct evidence, stratified response characteristics, and a scale-bridging perspective on vapor exchange processes within this key layer. We would therefore be very grateful for the opportunity to submit a substantially revised version that addresses RC2’s concerns more fully and presents the contribution of the study in a more rigorous, transparent, and focused manner.
Citation: https://doi.org/10.5194/egusphere-2026-1143-AC1
Viewed
| HTML | XML | Total | Supplement | BibTeX | EndNote | |
|---|---|---|---|---|---|---|
| 117 | 79 | 14 | 210 | 27 | 14 | 15 |
- HTML: 117
- PDF: 79
- XML: 14
- Total: 210
- Supplement: 27
- BibTeX: 14
- EndNote: 15
Viewed (geographical distribution)
| Country | # | Views | % |
|---|
| Total: | 0 |
| HTML: | 0 |
| PDF: | 0 |
| XML: | 0 |
- 1
Clarity is required in the description of the study site and in the explanation of how the field set-up was conducted. Additional detailed comments are in the pdf document