Iaoos Observations of Aerosols and Clouds in the High Arctic by Autonomous Drifting Lidar Platforms

New drifting platforms have been deployed within the French project IAOOS (Ice-Atmosphere-Ocean Observing System) in the Arctic since 2014. Radiation and meteorological parameters are measured at the surface and profiles of aerosol and cloud properties are obtained with autonomous backscatter lidar systems. These platforms are indeed equipped for ocean-ice-atmosphere studies over the Arctic to better understand processes and interactions controlling sea-ice changes [1]. As stations in the Arctic are sparse, they can also be used as reference measurements for satellite observations. They are deployed in the Arctic almost every year and allow to perform regular measurements of the vertical structure and optical properties of the atmosphere in complement to satellite observations. Other data on snow, ice and ocean are simultaneously measured. Comparisons were made with CALIPSO/CALIOP observations. Measurements on the atmosphere are presented and results are discussed. 1. IAOOS OBSERVATIONS The upper part of the IAOOS platforms is dedicated to atmospheric measurements. It includes an integrated meteorological station, radiometric sensors, a backscatter lidar (possibly including polarization separation) and a camera. Figure 1 presents the mounting of the platforms. Additional temperature measurements are performed in ice, and upper ocean from a multiple temperature sensor chain [2]. Fig. 1: Implementation of the instrumentation on the IAOOS platform Other sensors are adapted for temperature, salinity, gases and nutrients for measurements in the ocean [1]. Data are taken at regular intervals (up to 8 times per day) over a given period of time (minutes to an hour) Several platforms can be deployed each year, depending on the opportunities by the collaborations with groups performing ocean campaigns in Arctic, or the possibility to deploy from the Russian Barneo camp in spring at the North Pole. Field measurements are starting in spring or summer and may last up to two years as the platforms are drifting over the Arctic, depending on where they are placed and problems that may occur. Figure 2 : Trajectories of all deployed IAOOS platforms over high Arctic since 2013, showing the transpolar drift of sea-ice. Depth of Artic ocean correspond to shades of gray. Land is reported as white areas (courtesy N. Sennéchael, LOCEAN/SU). The sea-ice drift is such as most of the ice exits through the Fram straight between Greenland and Svalbard Islands. Some moves to the Beaufort gyre north of Canada in western Arctic. This can be seen in Fig. 2 presenting the trajectories of all the buoys that have been deployed. Platforms are usually destroyed when they reach the limit of ice EPJ Web Conferences 237, 05007 (2020) ILRC 29 https://doi.org/10.1051/epjconf/202023705007 © The Authors, published by EDP Sciences. This is an open access article distributed under the terms of the Creative Commons Attribution License 4.0 (http://creativecommons.org/licenses/by/4.0/). and ocean, in summer. Some have been crashed in ridges formed by ice floes, some by bears playing with them. Table 2 lists the deployments of platforms equipped with lidars since the beginning of the project, their operation and their duration of operation. Observations are taken over several months and some platforms have provided more than 1,5 year of data, some much less. The deployment is done during ocean cruises, from various ice-breakers in the frame of collaborative international efforts, most particularly with Korea (Araon), Norway (Lance), Germany (Polarstern) and from the North pole station (Barneo). Platform name Lidar type Year Deploy Barneo N-ICE N-ICE N-ICE Pstern Araon Barneo Barneo Barneo Araon Barneo Barneo Barneo Oden Location N. Pole Svalbard Svalbard N. Pole Siberia Siberia N. Pole N. Pole N. Pole Siberia N. Pole N. Pole N. Pole N. Pole 0peration IAOOS 4 back 2014 4/2014-11/2014 IAOOS 7 back 2015 1/2015-4/2015 IAOOS 9 Polar 2015 1/2015-4/2015 IAOOS10 IAOOS 12 back Polar 2015 2015 4/2015/6-2015 IAOOS 16 IAOOS 20 IAOOS 21 IAOOS 23 back back Polar back 2015 2016 2016 2017 9/2015-7/2016 10 days* 10 days* 4/2017-12/2017 IAOOS 24 back 2017 4/2017-11/2017 IAOOS 25 back 2017 8/2017-11/2018 IAOOS 26 back 2018 4/2018IAOOS 27 back 2018 4/2018IAOOS 28 back 2018 7/2018-8/2018* Table 1. IAOOS Lidar Platforms deployed since 2014 Icing of the window remains an issue, due to the discontinuous operation, although heating is applied at the end of summer. 2. IAOOS ATMOSPHERIC SYSTEM 2.1 Lidar The laser systems were designed to be eye-safe (regarding the NF EN 60825-1 norm). Energy was therefore limited, and high operation frequency (up to 10 kHz) has been selected. The time duration of the sequence was chosen as a compromise between the optimization of the signal to noise ratio, the atmospheric variability, and the sampling need. Vertical sampling resolution is 15m, but signals are further averaged before transmission (see Table 2). A typical 10 minutes averaging sequence for each profile was adopted. Four sequences a day were aimed at to allow a sampling of cloud properties compatible with inputs/outputs of meteorological model analyses. The optical head was developed in France at INSU by Division Technique, who also designed the control system and interface. Integration of the lidar system was performed by the French company Cimel, that also designed the photon counting sampling and averaging electronics. Backscattering lidar systems can be implemented with and without polarization discrimination. A similar overall design is used for both systems, including a polarizing device and a polarization splitter at the emission and at the reception of the polarized system. A two-channel reception system is used for dual polarization signal analysis. Detectors are avalanche photodiode from Excelitas used in Geiger mode. Ancillary data are acquired such as those from accelerometers to inform about the tilt angles of the buoy linked to ridge formation and sea-ice breaking. All data are transferred to IPEV (http://iaoos.ipev.fr) using Iridium satellite link. To reduce data transmission costs, vertical averaging is performed in progressively increasing altitude bins (Table 2).


IAOOS OBSERVATIONS
The upper part of the IAOOS platforms is dedicated to atmospheric measurements. It includes an integrated meteorological station, radiometric sensors, a backscatter lidar (possibly including polarization separation) and a camera. Figure 1 presents the mounting of the platforms. Additional temperature measurements are performed in ice, and upper ocean from a multiple temperature sensor chain [2]. Other sensors are adapted for temperature, salinity, gases and nutrients for measurements in the ocean [1]. Data are taken at regular intervals (up to 8 times per day) over a given period of time (minutes to an hour) Several platforms can be deployed each year, depending on the opportunities by the collaborations with groups performing ocean campaigns in Arctic, or the possibility to deploy from the Russian Barneo camp in spring at the North Pole. Field measurements are starting in spring or summer and may last up to two years as the platforms are drifting over the Arctic, depending on where they are placed and problems that may occur. The sea-ice drift is such as most of the ice exits through the Fram straight between Greenland and Svalbard Islands. Some moves to the Beaufort gyre north of Canada in western Arctic. This can be seen in Fig. 2 presenting the trajectories of all the buoys that have been deployed. Platforms are usually destroyed when they reach the limit of ice and ocean, in summer. Some have been crashed in ridges formed by ice floes, some by bears playing with them. Table 2 lists the deployments of platforms equipped with lidars since the beginning of the project, their operation and their duration of operation. Observations are taken over several months and some platforms have provided more than 1,5 year of data, some much less. The deployment is done during ocean cruises, from various ice-breakers in the frame of collaborative international efforts, most particularly with Korea (Araon), Norway (Lance), Germany (Polarstern) and from the North pole station (Barneo). Icing of the window remains an issue, due to the discontinuous operation, although heating is applied at the end of summer.

IAOOS ATMOSPHERIC SYSTEM 2.1 Lidar
The laser systems were designed to be eye-safe (regarding the NF EN 60825-1 norm). Energy was therefore limited, and high operation frequency (up to 10 kHz) has been selected. The time duration of the sequence was chosen as a compromise between the optimization of the signal to noise ratio, the atmospheric variability, and the sampling need. Vertical sampling resolution is 15m, but signals are further averaged before transmission (see Table 2). A typical 10 minutes averaging sequence for each profile was adopted. Four sequences a day were aimed at to allow a sampling of cloud properties compatible with inputs/outputs of meteorological model analyses.
The optical head was developed in France at INSU by Division Technique, who also designed the control system and interface. Integration of the lidar system was performed by the French company Cimel, that also designed the photon counting sampling and averaging electronics. Backscattering lidar systems can be implemented with and without polarization discrimination. A similar overall design is used for both systems, including a polarizing device and a polarization splitter at the emission and at the reception of the polarized system. A two-channel reception system is used for dual polarization signal analysis. Detectors are avalanche photodiode from Excelitas used in Geiger mode. Ancillary data are acquired such as those from accelerometers to inform about the tilt angles of the buoy linked to ridge formation and sea-ice breaking. All data are transferred to IPEV (http://iaoos.ipev.fr) using Iridium satellite link.
To reduce data transmission costs, vertical averaging is performed in progressively increasing altitude bins (  Altitude of cloud and aerosol layers can be determined directly, whereas forward lidar signal inversion is used to determine optical depths and scattering ratio or extinction profiles.

Met and radiation
As seen in Fig. 1 IAOOS platforms are equipped with several meteorological sensors (pressure, temperature, moisture and wind (from sonic anemometer)), and infrared brightness temperature measurement in the 8-12 µm region. Lidar is also used as a radiometer to measure the solar radiance scattered at nadir by the sky above. A large bandwidth solar radiometer is also mounted on some platforms. A low cost camera is used to take views of the platform environment during sunlight seasons.

IAOOS ATM. OBSERVATIONS
An example of time series obtained on the atmosphere, results obtained by the B24 platform set up at the North Pole in 2017 are shown in Figure 3. The variation of the lidar signal background is reported in the upper panel. It is a standard measure of IAOOS, used as a proxy for the downward diffuse solar flux. The sky's brightness temperature TB (which indicates the presence of cloud when TB is high) and the altitude-time cross-section of the attenuated backscattering signal. Areas of strong positive correlations (high particle scattering, high TB, high diffuse flux) are seen, marking the presence of dense clouds (red rectangles). Areas of small scattering, low TB and small diffuse fluxes marking clear-air atmospheric cases (blue rectangles). Between the two, all types of cloud structures can be observed. This information is used for the determination of radiative flux at the surface and the analysis of the surface energy balance which leads to the melting of snow. The TB measurements provide an additional constraint on the optical depth and type of the clouds. Lidar measurements show that the occurrence of supercooled boundary layer clouds can be significantly high. They are predominating in spring, as observed for all deployment years. Aerosols are frequently observed in winter [2,3]. depolarizing. Systematic trajectories have been performed for all aerosol layers identified by CALIOP. The analysis of backward trajectories indicates that below 2 km aerosols originate mostly from Russia and Europe, while above 2 km significant sources are also North America and the Pacific areas. Fig. 4 reports trajectories for dust and polluted dust.

IAOOS AND CALIPSO OBSERVATIONS
Results from IAOOS drifting platforms have been compared to CALIOP observations for a few cases of CALIOP identified dust types. (IAB~7.3x10 -3 sr -1 ), and CALIOP identified the presence of aerosols both along backward and forward trajectories from the IAOOS position. On February 14th, the dominant subtypes identified by CALIPSO were dusty marine at 1 km and polluted dust at 3 km. Rather highdepolarization ratios (larger than 10%) in both layers are indicative of the presence of nonspherical particles. Assuming that clear air is observed on this IAOOS profile above 5 km, as suggested by an extended range of altitude where SRatt is low and almost constant, one can determine that the two-way transmission due to particle scattering and absorption is about 60% (see Fig. 5). The IAB retrieved for the 14 February IAOOS profile combined with the measured transmission allows to estimate an average IAOOS lidar ratio (LR) of 27(±6) sr at 800 nm, for the aerosol layer between 0 and 5 km, considering measurement uncertainties of ±10% for both transmission and IAB and multiple scattering correction [2]. It corresponds to a 532 nm lidar ratio of 29 (±8) sr assuming the spectral dependence proposed by [4]. This value better corresponds to ice clouds (LR~25-30 sr at 532 nm) possibly including clean continental particles (LR~35 sr). Pure desert dust, polluted dust or polluted particles/smoke, or Arctic haze having much larger LR values [4].

CONCLUSION
Observations provided by IAOOS platforms deployed in the Arctic allow to characterize super-cooled boundary layer clouds frequently observed in spring and elevated aerosol layers in winter. Comparisons with space-borne CALIOP observations lead to suggest that the aerosol classification in Arctic may be biased by the occurrence of diamond dust.