---have very low RH values compared to the surrounding air which has direct implica-tions in cloud seeding. If cloud seeding has been successfully done and a cloud has been formed then under favourable atmospheric conditions, precipitation falls out from the bottom of the cloud. In the presence of isolated dry layers, this precipitation quickly evaporates or sublimates before reaching the ground. Thus, the effectiveness of cloud seeding is linked, among others, to the presence of dry layers. During the one-year campaign, we found that dry layers were present most of the time above the measurement site, showing higher frequency during the winter months (more than 60 % of the time). The average altitude that these layers reside is at 5 km on average and their geometrical depth is about 1 km.
Figure 7. Example of dry layers at UAE measurement size during OASIS campaign. Diurnal RH profiles retrieved by a PollyXT Raman lidar at 19th of November 2019. The asterisks show top and bottom of the detected layers. The retrieval is limited to nighttime observations
hence from around 3 until 15 UTC the data gap is shown with white color (see Section 3).
4.2 Optical properties of pollen particles
In Paper II, we explored the capabilities of ground-based multi-wavelength Ra-man lidars to retrieve optical properties of atmospheric pollen particles. Despite the allergenic effects of pollen, current lidar aerosol classification schemes do not include a pollen aerosol particle category as there is still no comprehensive and validated characterization of their atmospheric optical properties. In this study, we focused on birch and mixtures of birch and spruce pollen, taking advantage of the distinct polli-nation periods of these pollen types in the rural lidar site of Vehmasmäki in Kuopio,
---Finland. Figure 9 shows a case study of lidar-derived optical properties and meteor-ological conditions where both birch and spruce pollen was measured in ground-level.
Figure 9. Example of lidar optical properties during a birch-spruce event. From left to right:
Backscatter coefficient (Bsc. Coef.) and extinction coefficient (Ext. Coef.), lidar ratio (Lidar ra-tio) and linear depolarization ratio (Depol. Rara-tio), Angstrom exponents (A), relative humidity (RH) ad temperature. The available wavelengths are marked for each optical parameter in each panel.
Using a Hirst-type volumetric air sampler (Hirst, 1952), we were able to measure the pollen type and concentration near the ground and further link this information to the lidar observations. In the first period (5 to 9 of May 2016), high concentrations of birch pollen were measured in-situ with a maximum pollen concentration of 3700 grains per m3. During the second period (12 to 15 of May 2016), birch was still the dominant type in the collected samples, but a significant contribution of spruce was also evident. These two pollen types exhibit different shapes (near spherical versus non spherical) and therefore lidar derived optical properties resulted in contrasting values (Table 2). The retrieved pollen LRs are characteristic for dust and dust–smoke mixtures (Tesche et al., 2011), hence the characterization of pollen particles using the LR alone is rather problematic. On the contrary, δp values of 10 % and 26 % for birch and mixture or birch and spruce pollen, respectively, can be used in aerosol classifi-cation schemes. Nonetheless, these values fall in the same range of dust and biomass burning aerosol mixtures or dust mixtures with marine contribution (Groß et al.,
---2011). As a result, pollen is currently misclassified as dusty mixtures in automatic aerosol classification schemes (e.g. Omar et al., 2009). Thus, to separate pollen from other aerosol types a minimum of two depolarization wavelengths must be available as well as, backward air mass trajectories and dust and biomass-burning aerosol sources must be considered from auxiliary methods.
Table 2. Lidar ratio (LR) and linear particle depolarization ratio (δp) for birch and birch-spruce mixture at 532 nm wavelength.
Pollen type LR (sr) δp (%)
Birch 52 ± 12 10 ± 6
Spruce contaminated 62 ± 10 26 ± 7
Paper II shows the potential of ground-based lidar measurements to detect pol-len in the atmosphere. Nevertheless, there are several chalpol-lenges which need to be addressed in order to improve the characterization of optical properties of airborne pollen. Currently, the contribution of pollen to other aerosol particles, such as an-thropogenic pollution, has not been separated using well known methodologies (Tesche et al., 2009). This implies that the pure extensive properties of birch and spruce particles have not been fully defined. Moreover, laboratory studies on birch and spruce have reported much higher δp values than in the present study. This can be attributed to three reasons. The first reason is that laboratory measurements study pollen optical properties in dry conditions in contrast to lidar observations which are subject to ambient RH. For example, it has been shown that dry birch pollen particles become less spherical causing depolarization. The second reason can be related to pollen orientation. Previous studies have shown that pollen particles exhibit certain orientation when airborne. Thus, the non-spherical pollen particles and therefore the observed δp can be sensitive to the viewing angle; in the case of laboratory measure-ments such thing might have not been considered. Currently, theoretical optical properties of pollen particles under ambient conditions from scattering simulations do not exist. The third reason can be related with aging in the atmosphere. Big parti-cles are likely to collect smaller ones and can potentially change their initial shape.
Although this is not so likely in our relatively clean conditions.