High-resolution atmospheric water vapor measurements with a scanning di ff erential absorption lidar

Introduction Conclusions References Tables Figures


Introduction
Water vapor is the most important greenhouse gas and plays a key role in Earth's weather and climate.Accurate observations of water vapor (WV) in the lower troposphere from the surface to the entrainment zone at the top of the convective atmospheric boundary layer (ABL) are essential for improving weather forecasting (e.g., Crook 1996;Dierer et al., 2009) and reanalyses (Bengtsson et al., 2004).Furthermore, better parameterizations of land-surface and turbulent transport processes in the ABL Introduction

Conclusions References
Tables Figures

Back Close
Full are essential for improved regional climate projections (Warrach-Sagi et al., 2013;Kotlarski et al., 2014).Lidar with its ability to measure water vapor with high temporal and spatial resolution and high accuracy has the means to investigate the relevant atmospheric processes for these improvements.
For sounding water vapor with lidar, the Raman technique (e.g., Melfi et al., 1969;Whiteman et al., 1992;Behrendt et al., 2002;Hammann et al., 2014) or the differential absorption lidar (DIAL) technique (Schotland, 1974;Browell et al., 1979;Wulfmeyer and Walther, 2001a, b) can be applied.The water vapor DIAL of University of Hohenheim (UHOH) is a ground-based instrument which allows for 3-dimensional measurements of water vapor fields in the atmospheric boundary layer with high resolution and accuracy (Behrendt et al., 2009(Behrendt et al., , 2011(Behrendt et al., , 2012)).For instance in the COPS field campaign 2007 (Wulfmeyer et al., 2008(Wulfmeyer et al., , 2011;;Behrendt et al., 2013), the UHOH DIAL demonstrated its high spatial and temporal resolution and range for studying the preconvective environment (Behrendt et al., 2011) and convection initiation (Corsmeier et al., 2011) as well as its high accuracy in comparison to other in-situ sensors and remote sensing systems (Bhawar et al., 2011).Furthermore, its scanning capability was shown within the DFG Transregio 32 project in 2009 (Behrendt et al., 2009) for studying the performance of turbulence parameterizations in the convective atmospheric boundary layer (Wulfmeyer et al., 2014a).
In spring 2013, the UHOH DIAL was operated within the project "High Definition of Clouds and Precipitation for advancing Climate Prediction" (HD(CP) 2 ).HD(CP) 2 was initiated for improving the understanding of turbulent and mesoscale processes resulting in formation of clouds and precipitation.According to Stevens and Bony (2013), advanced formulation of these processes in climate models is critical for improving the predictions.Therefore in April and May 2013 the "HD(CP) 2 Observation Prototype Experiment" (HOPE) was organized near research center Jülich, in western Germany.Goal of the experiment was to produce a dataset of atmospheric measurements for a model evaluation at the scale of the model simulations.The new HD(CP) Introduction

Conclusions References
Tables Figures

Back Close
Full provide simulations with a horizontal grid spacing of 100 m over domains of 100 × 100 up to 1000 km × 1000 km.
For HOPE, three supersites were set up in a triangular configuration with distances of about 4 km between each other.All these sites were equipped with in-situ and remote sensing instruments to measure atmospheric parameters.With the different instruments, the temporal and spatial heterogeneity of the convective boundary layer (CBL) was investigated concerning WV, temperature, and wind fields as well as the distribution of aerosol particles and clouds.The high-resolution fields of the measured thermodynamic parameters are also used to derive higher-order moments of turbulent fluctuations (Muppa et al., 2014;Behrendt et al., 2014) as well as sensible and latent heat fluxes (Wulfmeyer et al., 2014b).The results shall contribute to an improved understanding of the relations between surface properties and boundary layer characteristics, shallow cumulus convection, as well as convection initiation.
In the following, we explain the DIAL technique and describe sensitivity studies of the water-vapor absorption cross-section in Sect.

Theory
DIAL measures the number density of a trace gas, e.g.water vapor, with two backscatter lidar signals.One signal is tuned to a wavelength with high absorption of the trace gas (S on (r): online signal) and the other signal to a wavelength with low absorption (S off (r): offline signal).r is the range from the lidar system to the scattering volume.
The exponential decrease of the intensity of a beam passing through a gas with extinction coefficient α gas is described by Lambert-Beer's law with the transmitted power S(r), the initial intensity S 0 .α gas is related to the absorption cross-section σ gas and the number density N gas of the trace gas by α gas (r) = N gas (r) σ gas (r).From these two equations, one gets the Schotland approximation of the DIAL equation (Schotland 1966(Schotland , 1974) with σ eff (r) = σ on (r) − σ off (r) for the effective cross-section.Online and offline wavelengths should be chosen close to each other because the Rayleigh extinction coefficient, the particle extinction coefficient, and the backscatter coefficient for online and offline wavelengths cancel when taking the ratio of the signals in Eq. (2).System parameters which are constant with range r cancel because of the derivation.Thus, for DIAL measurements no calibration needs to be applied but only the values of the absorption cross-sections at online and offline wavelength σ on and σ off are needed for measuring N gas .Equation (2), however, assumes that changes of the backscatter spectrum in the considered range cells can be neglected.Otherwise, a correction Introduction

Conclusions References
Tables Figures

Back Close
Full has to be applied for the Rayleigh-Doppler effect (Ansmann and Bösenberg 1987), which becomes significant at strong particle backscatter gradients.Within the analyses presented here this effect was not critical as confirmed also by comparisons with radiosoundings.
In the following, we concentrate on the accuracy of the determination of σ eff in Eq. ( 2).
σ eff depends on atmospheric conditions, i.e. ambient temperature, air pressure, and humidity.These parameters are height-dependent.Due to this, σ eff depends on r.The calculation of σ eff in our data analysis procedure is described in more detail in Sect.2.2.

Water vapor absorption cross-section
The water vapor number density N WV (z) varies within several orders of magnitude in the Earth's atmosphere.Also temperature T (z) and air pressure p(z) change with height and time.These three parameters influence the water-vapor absorption crosssection σ WV (λ, T , p, N WV ) at a certain wavelength λ.In this section we are using wave numbers ν = λ −1 in cm −1 instead of wavelengths as it is usual in spectroscopy.All given wave numbers and wavelength throughout this paper are based on the vacuum wavelength.σ WV is calculated as a sum over broadened absorption lines at a specific wave number νWV,k .The absorption cross-section σ k of each absorption line k can be described by the line strength I k and by the line shape V k according to (3) The line strength I k depends on the temperature T and the ground-state energy E B,k at the standard temperature T 0 and follows the equation wavelengths with increasing height.The line shape V k can be calculated as convolution of a Lorentzian profile (pressure broadening) and a Gaussian profile (Doppler broadening) with with the full width at half maximum (FWHM) of pressure broadening being: where γ air,k and γ WV,k are the coefficients for air broadening and self broadening at standard pressure p 0 , the FWHM of Doppler broadening being: with the velocity of light c, the gas constant R, and the molar mass of water vapor M WV , and the pressure shift of the line center being: with ∆ν 0,k being the line shift.Finally, the water vapor absorption cross-section is the sum over all absorption lines with their contribution at a certain wave number

Conclusions References
Tables Figures
In Fig. 1 the absorption cross-section was plotted for the wave number region from 12 210.0 to 12 230.0cm −1 (wavelengths between 819.0 to 817.7 nm).This region is used by us because it presents a good compromise between possible Ti:sapphire laser radiation with high power, the tuning range of available seed lasers with high stability and a large range of suitable water-vapor absorption cross sections.The parameters for the calculation, e.g.ground-state energy E B,k , can be accessed from high accuracy laboratory measurements collected in databases, e.g.HIgh-resolution TRANsmission molecular absorption database (HITRAN) described by Rothman et al. (2009Rothman et al. ( , 2013)).
Figure 2 presents the difference of the absorption cross-sections between the two compilations of HITRAN 2008and HITRAN 2012(Rothman et al., 2009, 2013) at ground level and for an altitude of 10 km for the same wave number range as in Fig. 1.The line parameters do partly differ significantly (up to 2 to 4 %) so that still corresponding uncertainties remain until today.As explained above (Eq.2), errors in the absorption cross-section -especially at the online wavelength -transfer into measurement errors with the DIAL technique.As we aim at accuracies of better than 5 %, care has to be taken in selecting a database as accurate as possible for the absorption cross-section parameters.We believe that HITRAN 2012 fulfills this demand and used this database for analyzing the HOPE data.Introduction

Conclusions References
Tables Figures

Back Close
Full

Sensitivity analyses
σ eff is shown in Fig. 3 for several heights between ground level and 10 km altitude.The atmospheric conditions for the calculation were taken here from the US standard atmosphere (NASA, 1976).With increasing height, the absorption lines become narrower (Eqs.6 and 7).Additionally, the absorption line peaks are shifted to larger wavelengths with increasing height due to the pressure shift ∆ν in Eq. ( 8).
It is clear that the US standard atmosphere (NASA, 1976) does not represent the real atmosphere.But it can be used to study the sensitivity of the water vapor DIAL measurements with respect to uncertainties of the absorption cross-section.As discussed by Wulfmeyer and Walther (2001a, b), a relative error in the effective absorption crosssection σ eff is transferred directly to the relative error of the number density of the DIAL measurements according to: with N being the number density of WV.
The sensitivity of σ eff to temperature, pressure and humidity with height is shown in Fig. 4. The sensitivities to temperature and pressure uncertainties increase with height while those to relative changes in humidity decrease.The profiles are nearly symmetric to zero.The sensitivities to temperature show the opposite sign than to pressure.
To reduce the uncertainties in our data analysis procedure, we modify the standard atmosphere with temperature and pressure values at the DIAL site.If nearby radiosonde measurements are available, we use these if the deviations to the standard atmosphere are significant.Temperature profiles with higher spatial and temporal resolution and closer proximity to the DIAL observations than from radiosoundings, can be obtained from RRL (Behrendt, 2005;Radlach et al., 2008;Hammann et al., 2014).The relative deviations due to humidity uncertainties (Fig. 4c) are small compared to those of temperature and pressure.They decrease with height because of the decreasing amount of humidity.To reduce this uncertainty even further, one can start the water Introduction

Conclusions References
Tables Figures

Back Close
Full vapor retrieval with the standard-atmosphere humidity profile and iterate with the calculated humidity profile measured by the DIAL.
With, e.g., a pressure uncertainty of 1 % and a temperature uncertainty of 1 K, the total uncertainty of the WV DIAL data analysis procedure is smaller than 1.4 % up to 3 km and 1.8 % up to 10 km altitude.This is sufficient for most applications and exceeds the development goals of our DIAL system, which are summarized in Wagner et al. (2013).

Wavelength selection
To improve the performance of the measurements, one can choose transmitter online wavelengths with different absorption cross-section values for different atmospheric conditions, different measurement ranges of interest, and different measurement modes, e.g. for vertical or scanning measurements.
The UHOH DIAL system is able to operate at wavelengths between 817.7 and 819.0 nm which is equivalent to wave numbers between 12 229.42 and 12 210.01cm −1 .
The main limitation for this range is set by the injection seeders (Späth et al., 2013).
As mentioned above, Fig. 1 shows the WV absorption cross-section spectrum for this range.
With distributed feedback (DFB) laser diodes as injection seeders, we reach a laser frequency stability that allows choosing wavelengths not only at the peak but also on the slope of an absorption line and thus select different values for the absorption crosssection (Späth et al., 2013).
The optimum absorption cross-section σ on depends on the optical thickness τ of the atmosphere in the maximum measurement range of interest R. τ is defined by When assuming that σ off σ on .Ramsberger and Gordley (1978) and Ismail and Browell (1989) estimated that an optimum one-way optical thickness of τ ≈ 1 minimizes the measurement errors for up to R.More accurate simulations revealed an optimum optical thickness in the range cell interval considered for each measurement bin: Wulfmeyer and Walther (2001a, b) demonstrated that present single-shot pulse energies of laser transmitters permit a target differential optical thickness of ∆τ ≈ 0.05.This means that the maximum range R with τ ≈ 1 can be separated in about 20 measurement intervals with optimum signal-to-noise ratio.
To estimate τ for a certain range R in the simulations, we assume a homogenous atmosphere with constant pressure, humidity and temperature, so that N WV and σ on become range-independent and the optical thickness simplifies to τ = N WV σ on R. With the condition τ = 1 we get then an equation for the absorption cross-section of For HOPE, we selected the absorption line at a wave number of 12 223 cm −1 .Figure 5 shows a zoom of the absorption cross-section of Fig. 1 into the wave number range of 12 220 to 12 223.5 cm −1 .Differently to Fig. 1, a temperature of 0 • C, a relative humidity rh of 80 %, and an air pressure of 1000 hPa resulting in a water vapor number density of 3 × 10 26 m −3 (absolute humidity of 8.95 g m −3 ) were used for the calculation.On the slope of the line, different absorption cross-sections σ on can be selected (marked in Fig. 5).For offline, a minimum of the absorption cross-section spectrum at a wave number of 12 220.99cm −1 is chosen.
In Table 1, the 6 example absorption cross-sections marked in Fig. 5 are listed together with the range R at which an optical thickness τ of 1 is reached (Eq.12).As one can see, we can select absorption cross sections which cover about one order of magnitude in values and thus optimum ranges between 1 and 10 km in our example atmosphere.Clearly, this helps in optimizing the sensitivity of the DIAL for different atmospheric conditions and ranges of interest.Introduction

Conclusions References
Tables Figures

Back Close
Full In Fig. 6 the UHOH DIAL with its components is depicted.The system can be divided in three parts: (1) the high-power laser transmitter, (2) the transceiver unit, and (3) the receiver with the data acquisition system.The laser transmitter of the UHOH DIAL is based on an unidirectional Ti:sapphire ring laser (Schiller, 2008;Wagner et al., 2011Wagner et al., , 2013) ) which is end-pumped with the frequency-doubled radiation of a pulsed diode-pumped Nd:YAG laser (Ostermeyer et al., 2005).Frequency-control is realized with the injection seeding technique (Wulfmeyer, 1995;Wulfmeyer et al., 1995Wulfmeyer et al., , 1996Wulfmeyer et al., , 1998;;Barnes and Barnes, 1993;Barnes et al., 1993;Späth et al., 2013).For the measurements described in this paper, the previous transmitter configuration (Wagner et al., 2013) had been modified in the following aspects: The pump laser has undergone a full redesign and consists now of a unidirectional ring oscillator in triangle configuration.The Nd:YAG laser power was set to up to 38 W at 1064 nm resulting in up to 20 W at 532 nm used for pumping the Ti:sapphire laser.The resonator of the Ti:sapphire laser has also been changed.Now, four resonator mirrors form a dynamically-stable ring resonator in bow-tie configuration (Metzendorf et al., 2012).In the HOPE experiment, a new seed laser system with two frequency-stabilized DFB laser diodes as injection seeders (Späth et al., 2013) was operated for the first time as part of the transmitter.For switching between online and offline wavelengths between single shots, a 2x1 optical fiber switch was used.
It meets the requirements concerning time and crosstalk (Späth et al., 2013) and is much easier to align than the electro-optical deflector used before.The deflector was still implemented and served as a backup.The laser transmitter was operated at two output-power levels depending on the transceiver configuration.The specifications of the Ti:sapphire laser for both configurations are listed in Table 2.
The UHOH DIAL has two configurations for transmitting the laser pulses to the atmosphere and receiving the backscattered lidar signals.In configuration 1 (see Fig. 6), the laser is coupled into an optical fiber and sent out to the atmosphere via a 20 cm Introduction

Conclusions References
Tables Figures

Back Close
Full transmitting telescope.The laser output power is restricted to 2 W in this configuration because the fiber accepts only up to this power before being damaged.The power transmitted into the atmosphere is then 1.6 W.
The backscattered photons are collected with an 80 cm receiving telescope with a focal length of 10 m.The light path behind this telescope contains five mirrors.These keep the optical axes on the scanner axes (Coudé concept).All mirrors are highly reflective at 820 nm.The following beam reducers and a lens focus the light on the detector.To block the daylight background, an interference filter (LTIF) is used.Its transmission curve has a full width at half maximum (FWHM) of about 1 nm around our laser wavelengths.For the measurements in autumn 2013 in Stuttgart-Hohenheim, the daylight-blocking interference filter LTIF was exchanged with one with higher transmission than during HOPE: while the transmission at the online wavelength was 50 % in spring 2013, it became > 90 % in autumn 2013.An avalanche photo diode (APD) detects the signals.The transmitting and receiving telescopes are mounted together on a scanner unit which allows for 3-dimensional measurements.The scanner unit can be operated with speeds between 0.1 to 6 • s −1 .The control software of the scanner provides predefined modes like range-height indicator (RHI) and plan-position indicator (PPI) scans in a single direction as well as different types of volume scans.
In configuration 2 (see Fig. 6), the Ti:sapphire laser beam is guided with a highreflection flip mirror (HRFM) to a beam steering mirror (BSM), which reflects the laser pulses to the atmosphere in vertical direction.The laser beam diameter is enlarged 5-fold with a beam expander, which reduces the beam divergence by a factor of 5 to below 0.1 mrad.In addition to the large, vertically pointing 80 cm telescope, a smaller 20 cm Schmidt-Cassegrain telescope can be used in this configuration for improving the detection of the near-field signals.To block the daylight background, a 30 nm broad IF (STIF) is implemented in front of the detector which is mounted to this telecope.For the measurements presented here, however, only data from the large telescope were used.As discussed above, the overlap effects cancel (see Eq. 2) when the system is well aligned.The minimum for the range of the remaining overlap effects is found at Introduction

Conclusions References
Tables Figures

Back Close
Full about 500 m in our case.In configuration 2, the full output power available from the Ti:sapphire laser of currently 4 W (Table 2) was sent to the atmosphere.The data acquisition system (DAQ) (14 bit transient recorder MI.4032, Spectrum GmbH, Germany) records the atmospheric backscatter signals with typically 10 MHz sampling rate resulting in a range resolution of the raw signals ∆R of 15 m.We record the backscatter signals of each laser shot, which gives us maximum flexibility later when analyzing the data.The DAQ is triggered by the Ti:sapphire laser pulse detected with a photo diode using leakage light behind a high-reflection mirror in the transmitter setup.The online/offline sorting is executed with a hardware signal from the online/offline switch of the seeder system.Together with the lidar signals, elevation and azimuth angles of the telescope are stored with each pulse.Within the WV data analysis process, averages in time and range are applied.These are typically 10 s in time for the online and offline raw data.Afterwards several range bins of data are used for the Savitzky-Golay fit (Savitzky and Golay, 1964) to determine the derivative in Eq. ( 2).

Measurement examples
In the following we show vertical measurements taken in autumn 2013 in Stuttgart-Hohenheim (southern Germany, 48 • 42 40.76N, 9 • 12 15.21E) and some highlights of scanning measurements of the HOPE campaign of spring 2013 near Jülich, Western Germany.These examples illustrate the capabilities of the UHOH DIAL in resolving humidity fields in the height range for which the system was optimized: the atmospheric boundary layer and the lower free troposphere (see Sect. 2.4).

Vertical measurements
For vertical WV measurements we use configuration 2 with maximum laser output power of 4 W.In the following, we present measurements made on 3 October 2013 in Stuttgart-Hohenheim.Within 9 h from 09:36 to 18:39 UTC, 5 radiosondes were Introduction

Conclusions References
Tables Figures

Back Close
Full launched at the lidar site (type RS92-SGP of Vaisala).Figure 7 shows the time-height cross-section of the absolute humidity.For the retrieval the online and offline raw signals were averaged over 10 s.For the WV retrieval we calculated the derivation in Eq. ( 2) with a Savitzky-Golay (SaGo) window length ∆R of 135 m (Savitzky and Golay, 1964) but keep the vertical step size of 15 m.
For the intercomparisons with the radiosonde data (RS), we averaged the DIAL data over a 20 min period around the radiosonde launch times to reduce the sampling differences.Intercomparisons of WV profiles for the height range from 0.5 up to 4.5 km for the five cases of 3 October 2013 are shown in Fig. 8.
For the noise error analysis, we use the technique of Lenschow et al. (2000), which distinguishes between instrumental and atmospheric noise by extrapolating the autocovariance function to the zero lag (see also Wulfmeyer et al., 2010).Figure 9 shows the instrumental noise error profiles in a height range of 0.5 to 3 km.Also for the noise analysis, 20 min periods around the radiosonde launching times were used.The noise error profiles mostly compare like expected for the different background conditions in the course of the day (local noon was at 11:12 UTC).∆q/q is smaller than 0.1 % up to 2500 m AGL (above ground level) for all cases; and smaller than 0.5 % up to 3000 m AGL But interestingly, the noise errors at 14:40 UTC are smaller than at 13:07 and 16:08 UTC.The reason for the lower noise level was a change in the presence of clouds here.On this day, there were mostly clouds present at altitudes of 6 to 8 km which scattered the sunlight so that the background light was higher than in non-cloudy conditions.These clouds were not present around 14:40 UTC so that the background signal was smaller at that time.Thus we can state, that the noise errors shown in Fig. 9 are mostly larger than those under ideal cloud-free conditions.It is interesting to scale the noise data (Fig. 9) to other resolutions with (Ismail and Browell, 1989) ∆q q ∝ (∆t) −0.5 (∆R) −1.5 .( 13) Introduction

Conclusions References
Tables Figures

Back Close
Full Wulfmeyer and Walther (2001a, b) showed that this equation still holds in the presence of background noise and speckle errors.We get a relative noise error up to 2500 m a.g.l.
< 1 % for 10 s time resolution instead of 0.1 % for 20 min resolution (factor 120 0.5 ).This is very important for turbulence measurements.We calculated the bias between DIAL and RS data (Behrendt et al., 2007a, b) with q DIAL (z) and q RS (z) being the absolute humidity measured with the DIAL and the RS at height z, evaluated in the height interval between z 1 and z 2 with N Z data points.
The profiles of the absolute and relative bias in 500 m intervals for all five intercomparisons as well as for their mean profiles were plotted in Fig. 10.The data cover a height range between 0.5 and 3 km, above the noise in the DIAL data becomes too large.The overall bias between the DIAL and the RS is only (−0.04 ± 0.11) g m −3 or (−1.0 ± 2.3) % which satisfies our development goal (Table 3).From the mean profile we can see an indication for dryer values measured by the DIAL than by the RS in higher altitudes; we plan to investigate this point further in the future by more extensive intercomparisons.

Scanning measurements
To capture the horizontal and vertical WV field in the HOPE domain, scanning measurements with the UHOH DIAL were performed mainly in the directions of the other two Introduction

Conclusions References
Tables Figures

Back Close
Full supersites.We used the DIAL in configuration 1.The three supersites were oriented in a triangle with distances of about 4 km. Figure 11 shows a sketch of the measurement area and the triangular scan pattern.The first part of the scan was made in RHI mode in the direction of LACROS (Leipzig Aerosol and Cloud Research Observations System).It followed a PPI scan with low elevation from LACROS to JOYCE direction (Jülich ObservatoY for Cloud Evaluation).Then again an RHI scan back to the vertical was made in direction of JOYCE.Finally, the scanner turned back to the start position staring vertically.Due to the position of the sun, only RHI scans in the direction of JOYCE or LACROS were possible in the afternoon and evening.Depending on the focus of the IOP, slow scans with scan speeds of 0.1 to 0.3 • s −1 or fast scans with a scan speed of 6 • s −1 were made.When scanning fast, the DIAL captured also fast changes in the atmosphere: one RHI scan from vertical to horizon then took only 15 s.The whole triangular pattern of HOPE was consequently performed in only 45 s. Figure 12 shows as example rangecorrected offline backscatter signals of an RHI scan in the direction of JOYCE on 14 April 2013.Three scans in a row were plotted in Fig. 12a for the near-range (3 km vertically and 4 km horizontally).In Fig. 12b the data is shown up to 14 km vertically and 8 km horizontally.For these plots, no time averaging was applied.The backscatter signals of each offline shot with 125 Hz are shown.In the near-range plots, one can see fascinating changes in the boundary layer structures between the scans -similar to what has been presented already by use of the volume-imaging lidar (Piironen and Eloranta, 1995) and the Raman-shifted Eye-safe Aerosol Lidar (REAL system) (Mayor and Spuler, 2004).The atmospheric dynamics captured here are mainly due to the presence of gravity waves above the ABL top and breaking Kelvin-Helmholtz waves in the interfacial layer.The interactions of these processes has certainly influence on the water-vapor entrainment flux and will be a key topic of investigations in the future.In the far-range, clouds at an altitude of 5 km and at altitudes of 10 to 12 km were detected.For the retrieval of WV averaging over several online and offline laser pulses is usually performed to achieve water vapor data with lower statistical uncertainties.Water vapor Introduction

Conclusions References
Tables Figures

Back Close
Full measurements of the UHOH DIAL based on such fast scans are currently analyzed and will be presented later.In Fig. 13 slower scanning measurements in the directions of LACROS and JOYCE on IOP 5 on 20 April 2013 between 06:03 and 06:27 UTC are presented.The time resolution is 10 s which results in an angle resolution of 1.5 • .One of these RHI scans took 10 min.The scanning data have been gridded to a horizontal-vertical grid in order to display the atmospheric structures appropriately.Figure 13a and b shows the rangecorrected backscatter signal over a horizontal range of 0.7 to 4.2 km up to an altitude of 2 km (25 m grid).Due to the longer path of the laser beam through the aerosol-laden air in lower elevation and the corresponding higher extinction, the signal decrease is stronger in the horizontal than in the vertical direction.The WV data were plotted with a grid resolution of 50 m in the same range (Fig. 13c and d).Because of higher extinction of the online signal in the moist boundary layer, the noise in a horizontal profile is larger than in a vertical profile.Thus we use for larger distances larger SaGo window lengths ∆R to reduce the noise and to keep the angle resolution.Up to a distance of 1.3 km we used ∆R = 285 m, up to 2.5 km of ∆R = 615 m, up to 3.0 km of ∆R = 915 m, and up to 4.2 km of ∆R = 1995 m.The WV field in the two scanning directions shows several similarities but also significant differences revealing the heterogeneities of the ABL in the HOPE region.Up to 500 m altitude, the moist growing CBL is seen.The residual boundary layer above can be separated indeed in several layers which differ in humidity and in altitude.While the humidity in the CBL and the layer around 1.5 km is similar for both scan directions, the layer at 1 km shows larger differences between the scans but also within each scan.

Summary
The DIAL technique was reviewed and the sensitivity of the absorption cross-section concerning temperature, pressure and humidity was investigated.It turned out that uncertainties due to relative changes of temperature and pressure have different sign.Introduction

Conclusions References
Tables Figures

Back Close
Full The total uncertainty is smaller than 1.4 % up to 3 km and smaller than 1.8 % up to 10 km altitude for uncertainties of temperature by 1 K and of pressure by 1 %.The uncertainty due to relative changes of humidity is small and decrease with height.
The UHOH DIAL was operated within HOPE over 180 h on 18 IOP days to capture the water vapor fields by vertical pointing observations and in slow and fast scanning mode.In this paper we presented the DIAL setup and demonstrated the scanning capability with examples of slow scanning measurements observing the 3-dimensional water-vapor field and fast scanning backscatter measurements showing the aerosol structure in the boundary layer and clouds up to altitudes of 14 km.An intercomparison of the WV profiles with radiosondes showed a very small bias between the instruments of only (−0.04 ± 0.11) g m −3 or (−1.0 ± 2.3) % for heights between 0.5 and 3.0 km.The noise analysis gave less than 0.1 % instrumental noise for altitudes up to 2.5 km and less than 0.5 % up to 3 km.Introduction

Conclusions References
Tables Figures

Back Close
Full  Full Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 2. In Sect.3, the setup of the UHOH DIAL is presented.Results of vertical measurements from Stuttgart-Hohenheim and scanning measurements from the HOPE campaign are shown in Sect.4Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | 4) with k B for the Boltzmann constant.The absorption lines are broadened due to pressure and Doppler effects.Additionally the absorption line peaks are shifted to larger Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper | Screen / Esc Printer-friendly Version Interactive Discussion Discussion Paper | Discussion Paper | Discussion Paper | Discussion Paper |

Figure 1 .Figure 2 .Figure 5 .Figure 10 .Figure 12 .Figure 13 .
Figure 1.Water vapor absorption cross-sections for T = 15 • C, p = 1013.25 hPa and N WV = 1.7 × 10 23 m −3 (ground level of NASA 1976 standard atmosphere) was plotted in the tuning range of the DFB seed lasers.The grey bars indicate the online and offline wave number regions used throughout this work.

Table 3 .
Intercomparisons between WV DIAL and radiosonde profiles for the measurements on 3 October 2013 in Stuttgart-Hohenheim with respect to absolute and relative bias for the height range of 0.5 to 3 km.