The BASIL Raman lidar system entered the
International Network for the Detection of Atmospheric Composition Change
(NDACC) in 2012. Since then, measurements have been carried out routinely on a
weekly basis. This paper reports specific measurement results from this
effort, with a dedicated focus on temperature and water vapour profile
measurements. The main objective of this research effort is to provide a
characterisation of the system performance. The results illustrated in this
publication demonstrate the ability of BASIL to perform measurements of the
temperature profile up to 50 km and of the water vapour mixing ratio profile
up to 15 km, when considering an integration time of 2 h and a vertical
resolution of 150–600 m; the mean measurement accuracy, determined based on
comparisons with simultaneous and co-located radiosondes, is 0.1 K (for the temperature profile) and 0.1 g kg
Raman lidar measurements are compared with measurements from additional
instruments, such as radiosondes and satellite sensors (IASI and AIRS),
as well as with model reanalyses data (ECMWF and ECMWF-ERA). We focused our
attention on six case studies collected during the first 2 years of system operation (November 2013–October 2015). Comparisons between
BASIL and the different sensor/model data in terms of the water vapour mixing
ratio indicate biases in the altitudinal interval between 2 and 15 km that are always within
Water vapour is the most important atmospheric greenhouse gas, and its
increasing tropospheric concentration is primarily driven (although indirectly) by human activities. Increasing concentrations of
Despite the well recognised importance of having accurate tropospheric and
stratospheric water vapour and temperature profile measurements, datasets
of these variables and their long-term variability are limited, especially
in the UTLS region. Quality water vapour measurements in the UTLS region are
provided by radiosondes or balloon-borne frost-point hygrometers. The latter
is considered to be the most accurate water vapour sensor for the low
humidity levels found in the UTLS region (Vomel et al., 2007). However, the
global radiosonde network, which includes
All of the above weather- and climate-related issues call for highly accurate measurements of both the water vapour and temperature profiles throughout the troposphere and stratosphere, with a specific focus on the UTLS region. These motivations pushed the Network for the Detection of Atmospheric Composition Change (NDACC), formerly the international Network for the Detection of Stratospheric Change (NCSC), to include water vapour and temperature lidars among its ensemble of instruments in the early 2000s. NDACC, which originally focused on the long-term monitoring of stratospheric ozone changes, has progressively broadened its priorities to include the monitoring of other atmospheric species and assessing their impacts on the stratosphere and troposphere. Atmospheric composition changes have a significant impact on the atmospheric thermal structure and this makes atmospheric temperature measurements of paramount importance for NDACC.
The University of BASILicata Raman lidar system (BASIL) entered NDACC in
November 2012. The primary contribution of BASIL to NDACC is the provision of
accurate routine measurements of the vertical profiles of both the water vapour
mixing ratio and temperature. Water vapour profile measurements by BASIL
cover the altitudinal interval from the surface up to
In the present research work, we illustrate and discuss temperature and water vapour profile measurements from BASIL with the purpose of assessing system performance in terms of measurement bias. Specific measurement examples are considered for this effort, which are compared with measurements from other instruments, such as radiosondes and satellite sensors (IASI and AIRS), and with model reanalyses data (ECMWF and ECMWF-ERA).
The paper outline is as follows: Sect. 2 gives a brief description of the Raman lidar set-up and its operation schedule in the framework of NDACC; Sect. 3 describes the additional profiling sensors and model data involved in the present intercomparison effort; Sect. 4 illustrates the different lidar techniques considered to measure atmospheric thermodynamic variables; Sect. 5 defines the statistical quantities used in the intercomparison for the assessment of the measurement performance; Sect. 6 illustrates the intercomparison results and provides an assessment of the performance of the sensors and models considered; and Sect. 7 summarises all of the results reported and illustrates some possible future developments of the present study.
The Network for the Detection of Atmospheric Composition Change (NDACC) became operational in 1991. It includes more than 70 globally distributed, ground-based remote-sensing research stations for the observation of the physical and chemical state of the upper troposphere and stratosphere as well as their changes and for assessing the impact of these changes on global climate. Trends in the chemical and physical state of the atmosphere can be detected based on the collection of long-term databases. NDACC includes approximately 25 ground-based lidar systems distributed worldwide, which are routinely operated for the monitoring of atmospheric temperature, ozone, aerosols, water vapour and polar stratospheric clouds. To extend its research, NDACC has also established formal collaboration agreements with eight other major research networks (De Mazière et al., 2018), namely the AErosol RObotic NETwork (AERONET), the Baseline Surface Radiation Network (BSRN), the Advanced Global Atmospheric Gases Experiment (AGAGE), the Global Climate Observing System (GCOS) Reference Upper-Air Network (GRUAN), the National Aeronautics and Space Administration (NASA) Micro-Pulse Lidar Network (MPLNET), the Halocarbons and other Trace Species Network (HATS), the Southern Hemisphere Additional Ozonesonde Network (SHADOZ) and the Total Carbon Column Observing Network (TCCON).
A fundamental aspect of NDACC is represented by the high standard of quality of the data collected; we demonstrate, based on the results illustrated in this paper, that this standard is also reached by BASIL. Measurements of vertical profiles of atmospheric temperature, water vapour mixing ratio and the particle backscattering coefficient at 354.7 nm from BASIL are included in the NDACC database. BASIL is the only lidar system within the network that provides simultaneous and co-located measurements of these three atmospheric variables, with the data for these three variables being ingested into the NDACC repository and made available to the NDACC community.
BASIL in situated in Potenza, Italy (40
The main characteristics of the BASIL Raman lidar system.
In addition to a higher accuracy and better vertical resolution, a further advantage of lidar techniques with respect to traditional passive remote sensors is represented by the accurate characterisation of the random uncertainty affecting the measurements, which is available for all altitudes and each individual profile. This is determined from the signal photon number based on the application of Poisson statistics. The application of Poisson statistics to lidar signals is correct when dealing with lidar echoes acquired in both photon-counting and analogical mode. In the latter case, analogical lidar signals must first be converted into “virtual” counts. Considering an integration time of 5 min and a vertical resolution of 150 m, measurement precision at 10 km is typically 5 % for the water vapour mixing ratio and 1 K for temperature for night-time measurements. A detail description of the system set-up has been provided in several previous publications (e.g. Di Girolamo et al., 2009a, b).
The Infrared Atmospheric Sounding Interferometer (IASI), onboard the polar
orbiting MetOp satellite series, is a nadir-viewing Fourier transform
spectrometer measuring the radiation emitted from Earth's atmosphere in the thermal
infrared region (3.2–15.5
The Atmospheric Infrared Sounder (AIRS), launched aboard NASA's Aqua EOS
satellite in 2002, is a hyper-spectral sensor including 2378 infrared
channels and 4 visible/near-infrared channels, covering the
spectral interval from 3.7 to 15.4
The combined use of this ensemble of sensors allows for the provision of global
coverage, as well as accurate and high-resolution measurements of atmospheric
temperature and humidity profiles. Temperature profiles are measured in the
troposphere and stratosphere under clear-sky conditions, with an accuracy of 1 K and a horizontal resolution of 50 km. The vertical resolution is 1 and 4 km for tropospheric and stratospheric measurements respectively. Tropospheric
humidity profiles are measured under cloud-free conditions, with a vertical
resolution of 2 km and an accuracy of 15 and 50 % in lower and upper
troposphere respectively. The Aqua satellite is located on a
sun-synchronous orbit with a nominal altitude of 705 km and an
orbiting period of 98.8 min, corresponding to
Reanalysis from the European Centre for Medium-Range Weather Forecasts
(ECMWF) are also considered in this intercomparison effort. Two distinct
reanalysis products are considered: ERA-15 (ECMWF, 2006), covering the
15-year period from December 1978 to February 1994, hereafter referred to as
ECMWF; and ERA-40 (Uppala et al., 2005), hereafter referred to as ECMWF-ERA, which was
originally intended to cover a 40-year period but finally included a
45-year period from 1957 (International Geophysical Year) to 2002. This
latter reanalysis makes use of a larger ensemble of archived data, which
was not available at the time of the original analyses. The horizontal
resolution of the dataset is
Raman lidar measurements of the water vapour mixing ratio profile have been
extensively reported in the literature (Whiteman et al., 1992; Whiteman, 2003).
The approach makes use of the roto-vibrational Raman lidar signals from
water vapour and nitrogen molecules at the two Raman-shifted
wavelengths
In the recent past, temperature lidar measurements have become more and more important in weather and climate studies. Several lidar techniques have been demonstrated to be effective for routine measurements (Behrendt, 2005), including the rotational Raman technique (Behrendt and Reichardt, 2000) and the integration technique (Hauchecorne and Chanin, 1980; Hauchercorne et al., 1992) among others. The rotational Raman technique, especially if implemented in the UV domain, allows for the measurement of temperature profiles typically up to the lower stratosphere, whereas the integration technique is successfully used to measure temperature profiles throughout the stratosphere and mesosphere.
The Raman lidar system considered in the present paper performs simultaneous temperature measurements using both the rotational Raman technique (up to approximately 25 km) and the integration technique (from 20 km up to approximately 50 km), with a partial superimposition of the two sounded ranges in the altitudinal interval from 20 to 25 km and with no contamination of the elastic signals due to signal-induced noise effects. To the best of our knowledge, these measurements represent the first successful demonstration of the simultaneous application of both the rotational Raman and integration lidar techniques in a single instrument in the ultraviolet spectral region, i.e. in the region where the simultaneous exploitation of these two techniques has the highest potential.
Rotational Raman lidar measurements of the atmospheric temperature profile
rely on the use of the rotational Raman backscattered signals from nitrogen
and oxygen molecules within two narrow spectral regions encompassing
rotational lines from these two species with opposite sensitivity to
temperature changes: rotational lines that are closer to the laser
wavelength
Atmospheric temperature measurements are obtained from the ratio of the
signal including low quantum number
The location of the rotational Raman signals' centre wavelengths
The atmospheric number density profile,
Once
This algorithm can be applied starting from a reference maximum altitude,
hereafter identified using the symbol
The availability of simultaneous and co-located measurements of the water
vapour mixing ratio and temperature profiles, as is the case for BASIL,
makes the determination of the relative humidity profile straightforward.
Relative humidity is defined as the ratio, expressed as a percentage, between
the water vapour partial pressure profile
In order to assess the performance of the different profiling sensors and
models considered in the study, an appropriate statistical analysis has to
be carried out based on the estimation of specific statistical quantities.
Specifically, for each sensor/model pair, the percentage mean bias and
root-mean-square (RMS) deviation profile between two sensors/models, can be
determined using the following expressions (Behrendt et al., 2007a, b,
Bhawar et al., 2011):
For all intercomparisons reported in this paper, the bias and RMS deviation are
computed over 500 m width altitudinal intervals (
The index
BASIL was approved to enter NDACC in November 2012 and started operations
shortly afterwards. However, routine measurements on a weekly basis started
only 1 year later. In this paper we report measurements performed during
the 2-year period from 7 November 2013 to 5 October 2015. During this time
interval BASIL collected 385 h of measurements distributed over 80 d.
Lidar measurements are compared with model reanalyses (ECMWF and ECMWF-ERA40), satellite data (IASI and AIRS), and radiosondes from CNR in Tito, Italy. Figure 1
shows the location of BASIL (40
The location of BASIL (red dot) and the footprint of AIRS (green square centred on the green dot) and IASI (pink square centred on the pink dot), the IMAA-CNR radiosonde launching facility (purple dot) and the grid point of ECMWF ERA-15 and ECMWF ERA-40 (blue square centred on the blue dot). Map data ©2018 Google.
For the purpose of this paper, we focused our attention on six case studies collected during the first 2 years of system operation, namely 7 November, 19 December 2013, 9 October, 27 November 2014, and 2 and 9 April 2015. While a larger dataset could have been chosen, we decided to focus our attention on clear-sky cases only. In fact, clear-sky conditions represent the most suitable conditions for both water vapour and temperature measurements with Raman lidar, with water vapour profile measurements extending up to the UTLS region and temperature profile measurements extending up to 50 km. An appropriate assessment of the measurement performance based on a sensor/model intercomparison effort requires the sensors to be operated under clear-sky conditions, which is not always the case for the Raman lidar or the two passive space sensors IASI and AIRS. More specifically, the BASIL Raman lidar system does not have an all-weather measurement capability, which implies that the system is shut down in the case of precipitation. Additionally, BASIL – and this is true for all lidar systems – cannot penetrate thick clouds, as the laser beam is completely extinguished at optical thicknesses of around 2. Acceptable Raman lidar performance is still possible above thin clouds, with optical thickness less than 0.3. Thus, for the purposes of the present intercomparison effort, even the presence of high cirrus clouds makes case studies ineligible for the comparison. In other case studies, IASI and/or AIRS data were characterised by a very poor quality and unrealistic biases, which forced us to remove them from the intercomparison effort. After April 2015, the laser experienced a period during which the emitted power was reduced, possibly as a results of an unidentified internal optical misalignment. This caused a decline in the lidar's performance, which prevented us from considering measurements carried out after April 2015 within this intercomparison effort.
Vertical profiles of the water vapour mixing ratio mean bias and RMS
deviation for the 11 comparisons of BASIL with the radiosondes available in
the period from 9 October 2014 to 7 May 2015:
The Raman lidar has been calibrated based on an extensive comparison with
the radiosondes launched from the nearby IMAA-CNR station, which is only
8.2 km from the Raman lidar. Launched radiosondes are manufactured by
Vaisala (model RS92-SGP). For the purpose of determining the calibration
constant,
For the purpose of determining the calibration constant,
Water vapour mixing ratio profile as measured by BASIL over the
time period from 17:00 to 19:00 UTC on 7 November 2013, as well as the closest profiles (in
time) from IASI (at 19:29 UTC), AIRS (at 14:09 UTC) and the ECMWF (ERA-15 and ERA-40, at 18:00 UTC) model
reanalysis
A very similar procedure was applied to calibrate temperature measurements.
However, for the purpose of determining the calibrating constants
The constancy of the calibration constant was verified over the 2-year measurement period, appearing quite stable, as neither short-term or long-term time variations were revealed. Ageing of transmitter/receiver components was verified to not produce any appreciable variation in the calibration coefficients.
With respect to the water vapour mixing ratio measurements, above the
planetary boundary layer and up to 8.5 km (Fig. 2a), the mean bias is
found not to exceed
The above-specified uncertainties affecting the water vapour measurements are in
agreement with those reported for a variety of other Raman lidars operated
within the framework of NDACC. Specifically, Whiteman et al. (2012) reported a 5 %
uncertainty in the upper troposphere based on an extended comparison of the
NASA-GSFC Raman lidar system ALVICE with Vaisala RS92 radiosondes. For the
Maïdo Lidar on Réunion island, Dionisi et al. (2015) reported a
relative difference below 10 % in the low and middle troposphere (2–10 km) based on a comparison with 15 co-located and simultaneous Vaisala RS92
radiosondes. The upper troposphere, up to 15 km, is found to be
characterised by a larger spread (approximately 20 %), attributed to the
increasing distance between the two sensors. Leblanc et al. (2012) reported
water vapour mixing ratio profile measurements from the JPL Raman lidar at the Table Mountain Facility (California); these lidar have demonstrated the capability to
cover the region from
Vertical profile of atmospheric temperature as measured by BASIL
over the time period from 17:00 to 19:00 UTC on 7 November 2013 as well as the closest profiles (in time) from IASI (at 19:20 UTC), AIRS (at 14:09 UTC) and
the ECMWF (ERA-15 and ERA-40, at 18:00 UTC) model reanalysis
Figure 3a illustrates the mean water vapour mixing ratio profile measured by
BASIL on 7 November 2013 over the time interval from 17:00 to 19:00 UTC. The
vertical resolution of the data is 150 m from the surface up to 6 km, 300 m
between 6 and 8 km, and 600 m above 8 km. The water vapour mixing ratio
profile from BASIL reaches an altitude of approximately 15 km, with the capability
of measuring humidity levels as low as 0.003–0.004 g kg
Figure 3b shows the time evolution of the water vapour mixing ratio over a 6 h time interval from 16:00 to 22:00 UTC on this same day (7 November 2013).
The figure is a succession of 72 consecutive 5 min averaged profiles. For
the purpose of reducing signal statistical fluctuations, a vertical
smoothing filter was applied to the data, finally achieving an overall
vertical resolution of 150 m. The vertical smoothing filter considered is a
simple central moving running average computed using equally spaced data
(vertical step
Figure 4a illustrates the mean atmospheric temperature profile measured by
BASIL on 7 November 2013 over the same time interval considered in Fig. 3a. The measurement is based on the use of the rotational technique up to 20 km and the integration technique above 20 km. The combined use of these two
techniques allows for temperature profile measurements up to typically 50–55 km. In the
altitudinal region exploited using the rotational Raman technique, the
vertical resolution is 150 m from the surface up to 6 km and 600 m above this
altitude. The integration technique is applied downward, initialising the
algorithm at an altitude of 55 km. As mentioned above, although the boundary
value of
Again, the closest (in time) temperature profiles from the IASI (at
19:20 UTC) and AIRS (at 14:09 UTC) sensors and the ECMWF and
ECMWF-ERA40 (at 18:00 UTC) model reanalysis are also illustrated in Fig. 4a. The agreement
between BASIL and the different sensors/models is very good. Specifically,
deviations between BASIL and AIRS/IASI are lower than 2 K from the surface up
to 40 km and lower than 3–5 K above this altitude. Deviations between BASIL and the ECMWF
analyses (ERA-15 and ERA-40) also do not exceed 2 K all the way up to 50 km.
It is noteworthy that deviations between BASIL and the other
sensors/models may be the result of the random and systematic uncertainties
affecting the different sensors, as well as of the different air masses
sounded by the different sensors or encompassed in the different grid
points. However, it is to be added that temperature measurements by lidar
frequently reveal temperature fluctuations associated with the propagation
of internal gravity waves (Di Girolamo et al., 2009a). These fluctuations have
amplitudes that increase with increasing altitude and can be as large as 5–15 K
(Chanin et al., 1994; Zhao et al., 2017). Consequently, deviations between BASIL and the
other sensors/models are possibly associated in part with the effects of
gravity waves. The mean bias of BASIL vs. AIRS, IASI, ECMWF and ECMWF-ERA40
is 1.05, 0.83, 0.41 and
Vertical profile of relative humidity as measured by BASIL over
the time period from 17:00 to 19:00 UTC on 7 November 2013 and the
closest profiles (in time) from IASI (at 19:20 UTC), AIRS (at 14:09 UTC) and
the ECMWF (ERA-15 and ERA-40, at 18:00 UTC) model reanalysis
Figure 4b shows the evolution of the atmospheric temperature profile over the same 6 h time interval considered in Fig. 3b. Again, the figure is a succession of 72 consecutive 5 min averaged profiles. In this case, for the purpose of obtaining sufficiently high signal statistics, a vertical resolution of 150 m was considered. It is to be noticed that, despite the short integration time, the strong signal intensities, in combination with favourable clear weather conditions, allow for an altitude of 50 km to be reached. The tropopause region and its fluctuations are clearly visible in the figure.
Accurate relative humidity (RH) measurements are of paramount importance to determine cloud and aerosol radiative properties and related microphysical processes. RH has been demonstrated to have a critical influence on aerosol climate forcing (Pilins et al., 1995). Aerosol hygroscopic growth at high relative humidity levels may significantly influence the aerosol direct effect on climate (Wulfmeyer and Feingold, 2000). As described in Sect. 4.3, RH profiles are obtained from the simultaneous and independent measurements of the water vapour mixing ratio and temperature profiles carried out by BASIL. Figure 5a illustrates the mean atmospheric relative humidity profile measured by BASIL on 7 November 2013 over the same time period considered in Figs. 3a and 4a. The agreement between BASIL and the different sensors/models is good, with deviations not exceeding 10 % up to 15 km.
Figure 5b shows the time evolution of relative humidity over the same 6 h interval considered in Figs. 3b and 4b, the present figure is again a succession of 72 consecutive 5 min averaged profiles with a vertical resolution of 150 m. It is to be noticed that, despite the short integration time, an altitude of 15 km is reached, with measurements revealing a RH variability in the UTLS region that is systematically larger than the random uncertainty affecting the Raman lidar measurements.
The performance of the different profiling sensors and models considered in the present study are assessed using a dedicated statistical analysis. Specifically, for each sensor/model pair and each case study, the relative bias and root-mean-square (RMS) deviation profiles are determined in terms of both the water vapour mixing ratio and temperature.
The overall number of all possible sensor/model pairs is 15, which is the
maximum number of pairs possible when five sensors/models are available. More
specifically, these are BASIL vs. radiosondes (RS), BASIL vs. IASI, BASIL
vs. AIRS, BASIL vs. ECMWF, BASIL vs. ECMWF-ERA, RS vs. IASI, RS vs. AIRS, RS
vs. ECMWF, RS vs. ECMWF-ERA, IASI vs. AIRS, IASI vs. ECMWF, IASI vs. ECMWF-ERA, AIRS vs. ECMWF, AIRS vs. ECMWF-ERA and ECMWF vs. ECMWF-ERA. The
altitudinal intervals considered in the computation of the bias and RMS
deviation profiles may vary for the different sensor/model pairs depending on
the vertical coverage of the sensor considered, with the selection
being driven by the sensor with lower coverage. In this regard, we have to
recall that BASIL measurements of the water vapour mixing ratio and
temperature profile extend up to 15 and 50 km respectively, whereas the
radiosondes considered in the present study, which are those launched from
the nearby IMAA-CNR station, provide profiles extending up to
Time intervals for all sensors/models for all case studies considered.
For all intercomparisons reported in this paper, we computed bias and RMS
deviation considering vertical intervals of 500 m (i.e.
Figure 6 shows the water vapour mixing ratio bias and RMS deviation profiles
for all sensor/model pairs. As expected, the bias shows higher values in the
ABL (in the range between
For all sensor/model pairs, the bias shows values lower than
For all sensor/model pairs, the percentage bias shows values in the range of
Figure 7 illustrates the temperature bias and RMS deviation profiles for all
sensor/model pairs. The bias of BASIL vs. the radiosondes is in the range of
Except for a few points, bias values are within
So far, we have reported and discussed the mutual bias and RMS deviation
profiles between different sensors/models, highlighting the altitude
variability of these quantities. However, in order to assess sensors' and
models' performance, it is often preferable to use a single bias
The vertically averaged mean bias,
The weight
The vertically averaged absolute mean bias,
Vertical profiles of the water vapour mixing ratio mean bias and RMS
deviation for all sensor/model pairs: bias
Table 3 includes the vertically averaged mean mutual bias,
The value of
Water vapour mixing ratio vertically averaged mean and absolute mean bias and the RMS deviation values for all of the sensor/model pairs considered.
Vertical profiles of the temperature mean bias and RMS deviation for
all sensor/model pairs:
Values of the percentage
Overall bias affecting the water vapour profile
Table 4 includes the vertically averaged mean mutual bias and RMS deviation
values,
Temperature vertically averaged mean and absolute mean bias and RMS deviation values for all considered sensor/model pairs.
Making use of the available statistics of comparison results, an approach is
considered to determine the overall bias values for all sensors/models
involved in this intercomparison effort. This approach, originally proposed
by Behrendt et al. (2007a, b), can be applied when there is at least one sensor
whose measurements are comparable with all other sensors/models. For this
purpose we considered the BASIL Raman lidar. Assuming equal weight on the
data reliability of each sensor/model, an estimate of the overall bias
affecting all sensors/models is obtained by imposing the condition that the summation of
all mutual biases between the sensor/model pairs is equal to zero. The choice of
attributing equal weight to the data reliability of each sensor/model is
driven by the awareness that none of them can be assumed a priori to be more accurate than
the others and, thus, by the assumption that the closest profile to a
reference profile can be obtained by taking the mean of all of the available
profiles. Based on this approach, the overall bias affecting the water vapour
profile data from BASIL, the radiosondes, IASI, AIRS, ECMWF and ECMWF-ERA40
is estimated to be
The same approach was applied to determine the overall bias for temperature profile data from
all of the sensors/models involved in this intercomparison effort. In this case,
as we have previously identified a significant systematic uncertainty
affecting AIRS measurements, this sensor was excluded from the summation of
all mutual biases between sensor/model pairs. Thus, assuming equal weight on
the data reliability of all of the other sensors/models, the overall bias affecting
temperature profile data from BASIL, the radiosondes, IASI, AIRS, ECMWF and
ECMWF-ERA40 is found to be 0.19, 0.22,
Case studies illustrated in this paper demonstrate the ability of BASIL to
perform temperature profile measurements up to 50 km and water vapour mixing
ratio profile measurements up to 15 km, considering an integration time of 2 h and a vertical resolution of 150–600 m, with a measurement accuracy of
0.1 K and 0.1 g kg
The possibility of assessing the overall bias values for all of the sensors/models
included in this intercomparison effort was also exploited, benefiting from
the circumstance that the BASIL Raman lidar could be compared with all other
sensor/model data. The overall bias affecting water vapour/temperature
profile data from BASIL, the radiosondes, IASI, AIRS, ECMWF and ECMWF-ERA40
was estimated to be
The present study allows us to gain confidence in the high quality of the water vapour and temperature profiling carried out by BASIL and included in the NDACC database as well as in the possibility of using long-term records of these measurements for monitoring atmospheric composition and thermal structure changes and, ultimately, for climate trend studies.
The ECMWF data used in this study were
obtained from the ECMWF data server:
Paolo Di Girolamo designed the experiment and carried out the measurements with Benedetto De Rosa and Donato Summa. Benedetto De Rosa and Donato Summa developed the data analysis algorithms, and Benedetto De Rosa carried out the data analysis. Benedetto De Rosa and Paolo Di Girolamo prepared the paper with contributions from Donato Summa.
The authors declare that they have no conflict of interest.
This work was made possible by support from the Italian Ministry for Education, University and Research (grant no. OT4CLIMA). We thank the GCOS Reference Upper-Air Network (GRUAN) for providing radiosonde data. Special thanks are given to IMAA-CNR of Tito Scalo and the site representative Fabio Madonna.
This paper was edited by Ulla Wandinger and reviewed by two anonymous referees.