LOWER-FREE TROPOSPHERIC OZONE DIAL MEASUREMENTS OVER ATHENS, GREECE

A compact ozone differential absorption lidar (DIAL) was implemented at the Laboratory of Laser Remote Sensing of the National Technical University of Athens (NTUA), in Athens, Greece. The DIAL system is based on a Nd:YAG laser emitting at 266 nm. A high-pressure Raman cell, filled with D2, was used to generate the λON and λOFF laser wavelength pairs (i.e., 266-289 nm and 289-316 nm, respectively) based on the Stimulated Raman Scattering (SRS) effect. The system was run during daytime and nighttime conditions to obtain the vertical profile of tropospheric ozone in the Planetary Boundary Layer (PBL) and the adjacent free troposphere.


INTRODUCTION-TROPOSPHERIC OZONE
Ozone plays a significant role in the tropospheric chemistry as the major precursor of the hydroxyl radicals (OH), which in turn is the primary removal agent for most of the atmospheric pollutants [1]. In addition, ozone as a greenhouse gas influences the radiative balance of the troposphere and consequently the climate change [2].
Moreover, high ozone levels over urban areas contribute to the formation of photochemical smog while can cause health problems to humans and plants [3].
The ozone budget in the troposphere and its concentration changes are controlled by both chemical and dynamical processes: photochemical production in the Planetary Boundary Layer (PBL), downward transport from the stratosphere and horizontal transport from a source region [4].
Furthermore, ozone in the free troposphere can be easily transported from a source region to another, due to its relatively longer lifetime and the higher wind velocities, as detected over the Eastern Mediterranean [8][9][10].

Methodology
By applying the well-known lidar equation [12] at the two wavelengths (λ ON and λ OFF ), we get the ozone concentration n O3 [4]: (1) where, P i = number of photons received at channel i (i=ON or OFF wavelength), Δσ i = σ(λ ON )-σ(λ OFF ) for O 3 and σ(λ) = absorption cross section of ozone (cm 2 ) at wavelength λ.
The corresponding statistical error (ε 1 ) if given by the following equation [4]: and P bi (z) is the atmospheric and electronic background noise at channel i, Δz is the range resolution of the measurement, and N(z) is the number of laser shots (related to the integration time).
The corresponding systematic error (ε 2 ), if we omit the laser beam absorption by other gases, is given by the following equation [4]: where, β(λ i ,z) is the atmospheric backscatter coefficient at wavelength λ i (i=ON or OFF) at range z, Δα RAY (λ,z)= α RAY (λ ON ,z)-α RAY (λ OFF ,z) and Δα aer (λ,z)=α aer (λ ON ,z)-α aer (λ OFF ,z), where α RAY (λ i ,z) and α aer (λ i ,z) are the extinction coefficients of atmospheric molecules and particles, respectively, at wavelength λ i (i=ON or OFF) at range z. In Fig. 1 we present simulations of the systematic error ε 2 , in a polluted urban atmosphere (with an intense presence of aerosols), for three wavelength pairs (266-289 nm, 289-299 nm, 289-316 nm), a range resolution of 300 m and a typical lidar ratio (LR), LR(λ,z)=α aer (λ,z)/β aer (λ,z), of 40 sr. From this figure we see that we can derive the ozone profiles within the first 2 km (typically within the Planetary Boundary Layer) with a systematic error ε 2 <10%, if we use the 266-289 nm wavelength pair.

Experimental setup
To produce the ON and OFF wavelength we use the Stimulated Raman Scattering (SRS) effect by optically pumping, with the 4 th harmonic of a Nd:YAG laser (17-20 mJ/pulse at 266 nm), a 1-m long high-pressure Raman cell filled with a Raman active gas (D 2 at 14 atm).  (289 and 316 nm, respectively) and 266 nm [13].
Previous experiments [13] have shown that at a pressure of 14 bar (D 2 ) we can achieve an energy conversion efficiency at the exit of the Raman cell of the order of 40% at both the 1 st and 2 nd Stokes wavelength (289 and 316 nm, respectively), and of ~18% at the fundamental wavelength 266 nm (Fig.  2). By adding H 2 in the Raman cell, the wavelength of 299 nm, can also be generated [13].     (Fig. 3). This telescope is optically coupled to a discriminating spectrometer through a pure SiO 2 optical fiber [14]. The backscattered lidar signals between 266-316 nm are spectrally separated by a Czerny-Turner grating spectrometer (4960 lines/mm) with a spectral resolution better than 0.5 nm (Fig. 4). Four concave mirrors are used to focus the lidar signals at the photocathode of compact photomultiplier tubes (PMTs) at four respective wavelengths: 266, 289, 299 and 316 nm. High performance transient recorders (Licel GmbH) are used to digitize the received lidar signals (12 bits -40 MHz -8192 signal bins) at 266, 289 and 316 nm. The typical number of averaged lidar signals is 4000, corresponding to 6 min time resolution.

EXPERIMENTAL RESULTS
In Fig. 5, we present the vertical profile of the ozone number density (moles/cm 3 ) obtained during daytime (10:15-14:21 UTC) over Athens, on 02 December 2016, based on a combination of two-wavelength pairs: 266-289 nm (inside the PBL, from ground up to 2 km height) and 289-316 nm (in the free troposphere). This profile has been corrected for the atmospheric backscatter and for the molecular and particulate differential extinction (cf. equation 4). The latter one has been based on aerosol measurements obtained by our DIAL system at 316 nm and at 355 nm by the collocated multi-wavelength Raman lidar of NTUA (EOLE system). The corresponding vertical profile of the relative humidity (RH) obtained by radiosonde (cf. Fig. 5 a) High values of absolute vorticity and winds field over Athens (observed a few hours before) at 500 hPa surface (cf. Fig. 6), b) Total ozone data showing a cut-off low (denoted by a brown circle) event [5][6][7][9][10] with high ozone values over Greece (cf. Fig. 7).
c) Strong downward winds around 2 km height (skew-T diagram not shown).

CONCLUSIONS
In this paper we presented a typical vertical profile of the ozone number density obtained over Athens, using a two-wavelength pair DIAL system (266-289 nm and 289-316 nm), which coincided with a deep STT event [9][10][11], observed down to 2.1 km height.
Systematic ozone measurements are to be performed over Athens, starting on February 2017, during daytime and nighttime.

ACKNOWLEDGMENTS
The AutoCAD simulation of Figure 4 has been kindly provided by Dr. Jacques Porteneuve (IPSL, CNRS, France). Total ozone map has been provided by the WMO Ozone mapping center (LAP-AUTH, Greece). Radiosonde data has been obtained by the Hellenic National Meteorological Service. Absolute vorticity map and winds field, at 500 hPa surface, have been provided by the University of Wyoming (USA).