the Creative Commons Attribution 4.0 License.
the Creative Commons Attribution 4.0 License.
| 15 May 2026
Hybrid methodology for optimised water vapour mixing ratio profiles from Raman lidar measurements
Arlett Díaz-Zurita
Daniel Pérez-Ramírez
David N. Whiteman
Onel Rodríguez-Navarro
Víctor M. Naval-Hernández
Jorge Muñiz-Rosado
Soledad Fernández-Carvelo
Jesús Abril-Gago
Ana del Águila
Pablo Ortiz-Amezcua
Juan Antonio Bravo-Aranda
María José Granados-Muñoz
Juan Luis Guerrero-Rascado
Manuel Antón
Javier Vaquero-Martínez
Inmaculada Foyo-Moreno
Jose Antonio Benavent-Oltra
Lucas Alados-Arboledas
Download
- Final revised paper (published on 15 May 2026)
- Preprint (discussion started on 19 Jan 2026)
Interactive discussion
Status: closed
-
RC1: 'Comment on egusphere-2025-5035', Anonymous Referee #2, 02 Feb 2026
-
AC1: 'Reply on RC1', Arlett Díaz Zurita, 10 Apr 2026
Response to Referee # 2 Comments
Hybrid methodology for optimised water vapour mixing ratio profiles from Raman lidar measurements
Referee Report
The manuscript represents a substantial contribution to scientific progress within the scope of this journal, introducing a new method that combines correlative measurements of collocated precipitable water vapour and Numerical Weather Prediction data to reconstruct the profile within the incomplete overlap region of a lidar system, improving the calculation of the calibration constants and the water vapour mixing ratio profiles. The scientific approaches, the applied method, the results, and conclusions are well discussed and well structured. However, the English language needs some small corrections, mostly due to typos. So, I have some suggestions to improve the text
Response
We sincerely thank the referee for his/her time and suggestions, which have greatly contributed to improving the quality of this study. All the points raised have been carefully considered, and the corresponding revisions have been incorporated into the manuscript. Below we provide detailed, point-by-point responses to each comment (in blue).
Scientific Questions:
- Based on your research, the GNSS operates in most weather conditions, but the clear-sky conditions are preferred for optimum radiosonde comparisons. Can the Hybrid method be applied for a partly cloudy sky?
We appreciate the referee’s suggestion. Numerous previous studies have shown that the most suitable conditions for accurate water vapour Raman lidar calibration are those with a high signal-to-noise ratio (SNR) in the lidar measurements. A high SNR is more feasible under night-time and clear-sky conditions, as these minimise daylight background noise, signal contamination, and cloud-induced biases in lidar retrievals. All these issues are particularly critical for our water vapour Raman lidar; it should be noted that the achievable SNR depends strongly on specific system characteristics, such as laser power, optical configuration, and detector performance. Therefore, the hybrid methodology can be applied under partly cloudy conditions but depends explicitly on these system characteristics.
The hybrid methodology relies on the accurate computation of precipitable water vapour (PWV) from lidar measurements. Thus, it is essential that the lidar measurements adequately represent the entire atmospheric profile used for PWV integration. While clear-sky conditions are mainly to avoid limitations associated with noisy profiles during vertical integration, the methodology can also be applied under partly cloudy conditions, if the cloud base is located above the altitude range used for PWV integration and if the lidar SNR remains sufficiently high. Its applicability in the presence of low clouds depends on the size, frequency, and distribution of cloud-free gaps. Provided these gaps are sufficiently large and frequent, the lidar can acquire measurements with adequate SNR, enabling reliable vertical integration and temporal averaging.
In our revised manuscript, we will include minor modifications to clarify this point.
- Lines 230-235. Have you considered getting the aerosol contribution from the other inversion algorithm? For example, the GRASP code.
Response
We thank the referee for the suggestion. GRASP could be used to retrieve the aerosol concentration profile and, consequently, estimate the aerosol extinction profile. However, this requires accurate backscattered signals, which may not always be available (Lopatin et al., 2013).
Due to lidar overlap issues in backscattered signal at 355 nm, we preferred to use the approach proposed in our previous study (Díaz-Zurita et al., 2025), in which an approach based on AOD from AERONET sun photometer measurements was applied. Although not ideal, this method minimises the effect of systematic uncertainties. It is also well suited for very large databases, such as the one used in our study. Furthermore, the required symmetry in sky radiances for GRASP inversions, including sun photometry measurements, implies that many partly cloudy days are rejected.
We would also like to mention that in previous analyses we explored model-based approaches, such as CAMS, and simulated the aerosol effect using a step-function aerosol profile, assuming a well-mixed aerosol layer confined to the atmospheric boundary layer (ABL) with an approximately homogeneous vertical aerosol distribution. The resulting calculations closely matched lidar inversions. However, in the presence of decoupled aerosol layers above the ABL—which frequently occur at our station—the step-function approach is less accurate. Further details can be found in Díaz-Zurita et al. (2025).
Díaz-Zurita, A., Naval-Hernández, V. M., Whiteman, D. N., Rodríguez-Navarro, O., Muñiz-Rosado, J., Pérez-Ramírez, D., Alados-Arboledas, L., and Navas-Guzmán, F.: Sensitivity Analysis of the Differential Atmospheric Transmission in Water Vapour Mixing Ratio Retrieval from Raman Lidar Measurements, Remote Sensing, 17, 3444, 2025. https://doi.org/10.3390/rs17203444
Lopatin, A., Dubovik, O., Chaikovsky, A., Goloub, P., Lapyonok, T., Tanré, D., & Litvinov, P. (2013). Enhancement of aerosol characterization using synergy of lidar and sun-photometer coincident observations: The GARRLiC algorithm. Atmospheric Measurement Techniques, 6, 2065–2088. https://doi.org/10.5194/amt-6-2065-2013
- Regarding Table 4, why separate it into two periods, since the sample size seems insufficient (N=31) for estimating the statistical parameters? There are slight differences between k1, k2, and k3 in the second period for all ranges, and between k3 in the first and the entire period for all ranges.
Response
We appreciate this valuable question. The dataset was separated into two periods due to significant instrumental changes in the Raman lidar configuration, which importantly affect the calibration constant retrievals and, consequently, the water vapour mixing ratio profiles. It should be noted that, due to the different upgrades of the lidar system, there were changes in the region affected by incomplete overlap. Specifically, before May 2017 the system exhibited a significantly larger incomplete overlap region (around 700 m agl), whereas after June 2017, optical realignments and configuration upgrades reduced this region to around 300 m agl. In addition, the molecular reference channel was modified, changing from a vibrational–rotational nitrogen Raman channel at 387 nm to a pure rotational Raman configuration (nitrogen and oxygen at 354 nm), which directly affects the calibration constants.
Separating the dataset into two periods allows us to assess the sensitivity of the different calibration methods to these instrumental changes. Although the total number of radiosonde-lidar simultaneous measurements is limited (N = 31), this separation enables an evaluation of the impact of the system upgrades on the calibration performance and on the relative behaviour of the different calibration approaches. The observed behaviour is physically consistent: differences between K1, K2 and K3 are larger in the first period and become smaller and more stable in the second period, most likely due to improved system optimisation that resulted in a smaller overlap region. This result is illustrated in Figs. 2 and 7, where the calibration constants exhibit significant temporal variability, which can be explained by the modifications introduced in the Raman lidar optical configuration and the larger incomplete overlap region during the first period. Similar variability in calibration constants associated with changes in system design has also been reported for other lidar systems within the EARLINET network (e.g. Stachlewska et al., 2017).
In our revised manuscript, we will clarify why two different periods were used in the analyses.
Stachlewska, I. S., Costa-Surós, M., and Althausen, D.: Raman lidar water vapor profiling over Warsaw, Poland, Atmos Res, 194, https://doi.org/10.1016/j.atmosres.2017.05.004, 201.
- Is it possible to calculate a profile of calibration constants instead of one calibration constant, since the precipitable water vapour varies in height?
Response
We appreciate this question. The calibration constant in a Raman lidar system for water vapour measurement depends on the characteristics of the lidar system and is consequently height-independent proportionality factor. It accounts for the fractional volume of nitrogen in the atmosphere, the ratio of molecular masses, the range-independent calibration constants of the molecular reference and water vapour channels, and the range-independent Raman backscatter cross sections. Any observed variability in this calibration constant arises from uncertainties in its determination rather than from a true altitude dependence (Whiteman et al., 1992).
Whiteman, D. N., Melfi, S. H., & Ferrare, R. A. (1992). Raman lidar system for the measurement of water vapor and aerosols in the Earth's atmosphere. Applied Optics, 31(16), 3068–3082. https://doi.org/10.1364/AO.31.003068.
Technical Corrections
We again appreciate the referee for their efforts in pointing out these technical issues. We have carefully implemented the corrections in the revised manuscript.
-
AC1: 'Reply on RC1', Arlett Díaz Zurita, 10 Apr 2026
-
RC2: 'Comment on egusphere-2025-5035', Anonymous Referee #3, 01 Apr 2026
Dear Authors,
Congratulations on your excellent article. It's a pleasure to see how you've contributed to one of the outstanding problems: the calibration of lidar systems using hybrid methods involving different sensors. The article is well-written and clear, the methodology is robust, and the results are well-interpreted.
I only have a few minor comments, mostly about the figures.
1) L4: please define the MULHACEN (if it is ana Acronimum) for the reader the first time that it is used.2) L59: propably is De Rosa et ala. (2020)
3) L115 It would be better to introduce a a much more readable and immediate table with the system characteristics, instead the
Please specify better the type of conditions under which you operated and whether the statistical results are confirmed for all conditions or only in particular conditions.( Maybe in the some Table or in cocnlusion!!): I think the reader may be interested in understanding whether the uncertainty between Calibration of a Lidar System between Radiosondes and ERA5 also depends on meteorological conditions.
Citation: https://doi.org/10.5194/egusphere-2025-5035-RC2 -
RC3: 'Reply on RC2', Anonymous Referee #3, 10 Apr 2026
Dear Authors,
Congratulations on your excellent article. It's a pleasure to see how you've contributed to one of the outstanding problems: the calibration of lidar systems using hybrid methods involving different sensors. The article is well-written and clear, the methodology is robust, and the results are well-interpreted.
I only have a few minor comments.
1) L4: please define the MULHACEN (if it is ana Acronimum) for the reader the first time that it is used.2) L59: propably is De Rosa et ala. (2020)
3) L115: It would be better to introduce a much more readable and immediate table with the system characteristics, instead of a list of characteristics written in the text, or possibly both, simply by adding a table before point 2.4 (Model Data).
Please specify better the type of conditions under which you operated and whether the statistical results are confirmed for all conditions or only in particular conditions.( Maybe in the some Table or in cocnlusion!!): I think the reader may be interested in understanding whether the uncertainty between Calibration of a Lidar System between Radiosondes and ERA5 also depends on meteorological conditions.Citation: https://doi.org/10.5194/egusphere-2025-5035-RC3 -
AC2: 'Reply on RC3', Arlett Díaz Zurita, 10 Apr 2026
Response to Referee #3 Comments
Hybrid methodology for optimised water vapour mixing ratio profiles from Raman lidar measurements
Dear Authors,
Congratulations on your excellent article. It's a pleasure to see how you've contributed to one of the outstanding problems: the calibration of lidar systems using hybrid methods involving different sensors. The article is well-written and clear, the methodology is robust, and the results are well-interpreted.
I only have a few minor comments.
Response
Thank you very much for taking the time to review this manuscript. We sincerely appreciate your comments. Your feedback has been very valuable in helping us clarify important points and enhance the overall quality of the manuscript. Our detailed, point-by-point responses to your specific comments are provided below (in blue).
Specific comments:
1) L4: please define the MULHACEN (if it is ana Acronimum) for the reader the first time that it is used.
Response
Thank you for this helpful suggestion. MULHACEN is not an acronym and does not correspond to any technical aspect of the lidar technique. Rather, it is the given name of the first multiwavelength Raman lidar operated at the UGR urban station (Spain). Our lidar system is a commercial model by Raymetrics S.A., Greece (LR331D400). To avoid any confusion, we have removed the acronym MULHACEN from the revised manuscript.
2) L59: propably is De Rosa et ala. (2020)
Response
Thank you very much for pointing this out. We appreciate your careful reading and helpful comment. This was a LaTeX compilation error, and we have corrected it to De Rosa et al. (2020) in the revised manuscript.
3) L115 It would be better to introduce a much more readable and immediate table with the system characteristics, instead of a list of characteristics written in the text, or possibly both, simply by adding a table before point 2.4 (Model Data).
Response
Thank you very much for this valuable suggestion. We have introduced a new table summarising the main characteristics of the MULHACEN Raman lidar system in Section 2.2 to improve the clarity and readability of the manuscript.
4) Please specify better the type of conditions under which you operated and whether the statistical results are confirmed for all conditions or only in particular conditions.( Maybe in the some Table or in cocnlusion!!): I think the reader may be interested in understanding whether the uncertainty between Calibration of a Lidar System between Radiosondes and ERA5 also depends on meteorological conditions.
Response
Thank you very much for this relevant comment. The validation analysis of the calibrated water vapour profiles presented in this study is based on 31 simultaneous Raman lidar and radiosonde (RS) profiles, obtained during nighttime clear-sky conditions, which are typically associated with more stable atmospheric conditions. Although both the integrated column and the hybrid methods generally provide good calibration constants and show good agreement with radiosonde measurements, the results in Table 5 do not show statistically significant differences between the two methods. However, it is important to note that Precipitable Water Vapour (PWV) ranged from 0.5 to 25 mm, which corresponds to the typical range of minimum and maximum PWV registered at the Granada station (Navas-Guzman et al., 2014) and these extremes correspond to very different meteorological conditions.
Meteorological conditions can, however, influence the behaviour of water vapour profiles in the incomplete lidar overlap region and, consequently, in the PWV computation. Figure 5 illustrates that, within the incomplete overlap region, significant differences arise: when the atmospheric boundary layer (ABL) was well mixed (e.g., 19 May 2016, Fig. 5b, associated with anticyclones), both methods provide adequate profiles. By contrast, when the conditions were not well mixed (panels a, c, and d, associated with dissipating cold fronts and upper-level troughs), the hybrid method shows the best agreement, with SDs of 0.9, 1.0, and 0.4 g/kg, compared to 1.7, 1.4, and 0.3 g/kg when assuming constant values. In these cases, significant differences in calibration constants were observed: K₂ = 172 ± 3 g/kg and K₃ = 164 ± 2.4 g/kg on 25 July 2011, and K₂ = 93.5 ± 1.7 g/kg and K₃ = 88.7 ± 1.6 g/kg on 25 July 2016. This demonstrates that the hybrid methodology allows a reliable estimation of the vertical distribution of water vapour in the lower layers and of the calibration constant, under both stable and unstable conditions. The individual cases analysed (Figs. 5 and 8) include a variety of synoptic situations, such as anticyclones, dissipating cold fronts, upper-level troughs, extratropical low-pressure systems with associated cold fronts, and strong winds at mid- and upper levels (300 and 200 hPa, approximately 9.5–12 km asl).
Furthermore, the validation of the ERA5 model against radiosondes (Table 4 and Fig. 4) shows that the model can reproduce the behaviour of profiles in the incomplete overlap region, both day and night, across a wide range of meteorological conditions (148 profiles analysed from 2011–2023), which suggests that the applicability of the hybrid methodology is largely independent of meteorological conditions. In particular, Figure 8 shows that month-to-month variations of the calibration constants remain within the uncertainties of the method, except during periods with significant changes in system configuration. The low standard deviations in the monthly values further support the robustness and feasibility of the method.
The main objective of this analysis is not to establish a statistical dependence of calibration uncertainty on specific meteorological conditions, which would require a much larger RS dataset, but to demonstrate the robustness of the hybrid method under different atmospheric situations, enabling the reconstruction of the incomplete overlap region (Fig. 5 and 8) without assuming constant values, which are not always appropriate. This approach allows obtaining a calibration constant for each uncalibrated profile, provided the signal-to-noise ratio is adequate.
Navas-Guzmán, F., Fernández-Gálvez, J., Granados-Muñoz, M. J., Guerrero-Rascado, J. L., Bravo-Aranda, J. A., and Alados-Arboledas, L. (2014). Tropospheric water vapour and relative humidity profiles from lidar and microwave radiometry. Atmospheric Measurement Techniques, 7, 1201–1240. https://doi.org/10.5194/amt-7-1201-2014
All the points commented on here will be implemented in the revised manuscript, particularly in the conclusion section, as the referee suggests.
-
AC2: 'Reply on RC3', Arlett Díaz Zurita, 10 Apr 2026
-
RC3: 'Reply on RC2', Anonymous Referee #3, 10 Apr 2026
GENERAL COMMENTS
The manuscript represents a substantial contribution to scientific progress within the scope of this journal, introducing a new method that combines correlative measurements of collocated precipitable water vapour and Numerical Weather Prediction data to reconstruct the profile within the incomplete overlap region of a lidar system, improving the calculation of the calibration constants and the water vapour mixing ratio profiles. The scientific approaches, the applied method, the results, and conclusions are well discussed and well structured. However, the English language needs some small corrections, mostly due to typos. So, I have some suggestions to improve the text.
SPECIFIC COMMENTS
Based on your research, the GNSS operates in most weather conditions, but the clear-sky conditions are preferred for optimum radiosonde comparisons. Can the Hybrid method be applied for a partly cloudy sky?
Lines 230-235. Have you considered getting the aerosol contribution from the other inversion algorithm? For example, the GRASP code.
Regarding Table 4, why separate it into two periods, since the sample size seems insufficient (N=31) for estimating the statistical parameters? There are slight differences between k1, k2, and k3 in the second period for all ranges, and between k3 in the first and the entire period for all ranges.
Is it possible to calculate a profile of calibration constants instead of one calibration constant, since the precipitable water vapour varies in height?
TECHNICAL CORRECTIONS