CADEX and beyond: Installation of a new Polly XT site in Dushanbe

. During the 18-month Central Asian Dust Experiment we conducted continuous lidar measurements at the Physical Technical Institute of the Academy of Sciences of Tajikistan in Dushanbe between 2015 and 2016. Mineral dust plumes from various source regions have been observed and characterized in terms of their occurrence, and their optical and microphysical properties with the Raman lidar Polly XT . Currently a new container-based lidar system is constructed which will be installed for continuous long-term measurements in Dushanbe. house the lidar system, redun-dant air condition and power supply and a roof platform for further instruments.


Introduction
Mineral dust from Central Asia is one of several major contributors to global aerosol. However, profiling observations of dust occurrence, mobilization, transport, of its mixing with anthropogenic pollution, and also of its interaction with clouds are rare in these regions although such measurements are strongly needed to correctly address the dust effects in regional and global dust models. Therefore, the Central Asian Dust EXperiment (CADEX) was initiated in which in-situ and remote lidar measurements of dust over Dushanbe were performed from 2015 to 2016 in a statistical manner [1,2]. The portable Raman lidar Polly XT [3] used for the campaign is part of Polly-NET [4]. At the same time, the idea of two new permanent Polly-NET stations in the eastern Mediterranean and Central Asian dust regions was born, which led into the project to install two Polly XT lidar systems in Cyprus and Tajikistan (PoLiCyTa). In this presentation we want to summarize some results from the CADEX campaign and present the current status of the new permanent lidar system in Dushanbe.

Measurements
During CADEX near-ground and lofted dust layers have been observed frequently above the measurement location in Dushanbe, Tajikistan [2]. An example of such a nearground dust layer is shown in Fig. 1. On 2 July 2016 after 20:30 UTC until about 3 July 2016 08:00 UTC the lidar signal in the lowest 2 km was strongly enhanced. Thereafter, most likely a dilution process by the developing convective layer took place and the signal intensity decreased. Figure 2 shows the vertical profiles of the optical properties measured on 2 July 2016 21:40 -23:19 UTC. Dust was present up to 4.5 km height which can be seen from * Corresponding author: ronny@tropos.de the particle backscatter and extinction coefficients and the elevated particle linear depolarization ratio (PLDR) values. Above 3 km height the depolarization ratios decreased only slightly down to 25% which indicates a nondust fraction of the aerosol composition.
The polarization-lidar photometer networking (POLIPHON) method [5][6][7] further allowed us to separate the dust and non-dust fractions in terms of mass concentrations by help of the measured PLDR. In order to apply POLIPHON we assumed the PLDR and lidar ratio of 31% and 40 sr, respectively, for mineral dust and 5% and 50 sr, respectively, for non-dust particles. Furthermore, an extinction-to-mass conversion factor of 2.08 µg m −3 Mm was applied for dust and of 0.56 µg m −3 Mm for non-dust to derive the separated mass concentrations. Fig. 2f shows almost pure dust except between 3 and 4.5 km height where a minimal non-dust fraction below 10 µg m −3 was present. A maximum dust mass concentration of 450 µg m −3 at 1.5 km height was calculated. The calculation of the integrated dust mass yielded 1.3 g m −2 .
A summary of the mean optical properties measured by lidar and the AOTs and Ångström exponent from AERONET are presented in Tab. 1. The lidar-derived AOT was derived from integration of the profiles of the extinction coefficient from 0 -5 km height. Because of the overlap effect these profiles were extended towards the ground by means of the backscatter coefficient times the lidar ratios of S 355 (1.1 km) = 42.8 sr and S 532 (1.1 km) = 40.6 sr below a height of 1.1 km (c.f. Fig. 2b). The two AERONET measurements temporally closest to this lidar measurement indicate that the dust arrived during night time in between the two data points. The AOT at 500 nm wavelength increased and the Ångström exponent from the 440 -840 nm wavelength range decreased drastically overnight. Therefore, comparison with the lidar measurements at night-time are difficult. Nevertheless the AOT calculated from the lidar profiles are in reasonable agreement with the AERONET measurement about 2 h after the lidar observation.
Based on the 18-month data set during CADEX, we retrieved 268 profiles of optical properties during nighttime when it was possible to apply the Raman lidar method. Sixty-eight of these profiles were classified as dusty using a criterion of a PLDR larger Table 1. Averaged optical properties for the dust case in Fig. 1 and 2 derived from lidar and sun photometer. Top: Mean values of PLDR (δ) and lidar ratio (S ) at 355 and 532 nm wavelength, AOT (τ), and Ångström exponent (Å) . Bottom: AOT at 340, 380 and 500 nm wavelength and Ångström exponent from the 440 -870 nm wavelength range from AERONET Level 2 from before and during the dust event.
Polly XT

New Polly XT lidar in Dushanbe
Within the project PoLiCyTa a new version of Polly XT is currently developed. Polly XT are a series of automatic multiwavelength Raman polarization lidars developed by TROPOS over the last 15 years [3,8]. Those systems are all part of Polly-NET [4]. The measured data are uploaded to a central server (http://polly.tropos.de) where they are all automatically processed and plotted by the same data routines. In this way data from different systems is comparable, quality assured, and in the future held available for a broader science community. The new lidar system will be installed in a 20-ft sea container which is already constructed. Figure 3 shows a sketch of this container. An automatic hatch is foreseen with the ability to clean the roof window, a power generator will be applied to guarantee operation during power cuts, and a roof platform is integrated to give space for future instrumentation.
The main improvement for the new systems within PoLiCyTa is the application of a 100-Hz diode-pumped Nd:YAG laser system (Spitlight DPSS-250, InnoLas Laser GmbH, Krailling, Germany) in contrast to the 20-Hz flashlamp pumped system before which needed maintenance every two to three months. According to the specifications of this laser, maintenance is only required after two years of operation.
The lidar will have eight detection channels on a 30-cm far-range telescope and four channels on a near-range telescope. In this way it will be possible to measure profiles of the backscatter coefficient at three wavelength (1064, 532, 355 nm), the extinction coefficient and the depolarization ratio at two wavelengths (532, 355 nm), and the water-vapor mixing ratio from 150 m to 15 km height.

Outlook
In 2019 a new automatic Polly XT lidar system will be installed in Dushanbe as a collaboration between TROPOS and the Physical Technical Institute of the Academy of Sciences of Tajikistan to extend the measurements started during CADEX. With the operation of this system we will aim for the site of Dushanbe to be part of the European Research Infrastructure for the observation of Aerosol, Clouds, and Trace gases (ACTRIS). Furthermore, we will support the satellite missions Aeolus and EarthCare of the European Space Agency with calibration and validation data by this new ground station.