Most precipitation over continents originates in clouds containing ice (Mülmenstädt et al. 2015; Heymsfield et al. 2020), but melts into raindrops or drizzle drops on the way to the ground. The precipitation reaching the ground is mainly formed in mixed-phase clouds in which supercooled cloud droplets and ice crystals coexist. In the mixed-phase temperature regime (from −38° to 0°C), ice crystals form via heterogeneous nucleation with the help of suitable aerosol particles, called ice nucleating particles. The presence of ice crystals in a supercooled liquid cloud can rapidly change its further development because the mixture of cloud droplets and ice crystals is thermodynamically unstable due to the lower saturation vapor pressure over ice than over liquid water. As a result, ice crystals can grow rapidly at the expense of surrounding cloud droplets if the water vapor pressure drops below water saturation (Fig. 1; see also Korolev 2007). This so-called Wegener–Bergeron–Findeisen process (Wegener 1911; Bergeron 1935; Findeisen 1938) can lead to the glaciation of supercooled clouds and, subsequently, to the formation of precipitation.
Overview of the cloud seeding experiments performed during CLOUDLAB: A seeding UAV releases seeding particles into a supercooled low stratus cloud, which initiates ice formation through heterogeneous ice nucleation. The newly formed ice crystals can grow by diffusion and induce the transition from the liquid to the ice phase through the Wegener–Bergeron–Findeisen process. The seeded patch is characterized by a measurement UAV, a tethered balloon system, a vertical-pointing cloud radar, and a scanning cloud radar that performs sector scans downstream of the main site to observe the evolution of the seeded patch.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
Based on knowledge of the Wegener–Bergeron–Findeisen process, scientists, companies, and governments have actively tried to intervene in cloud microphysics since the 1940s (Vonnegut 1947) by seeding clouds with glaciogenic material to enhance precipitation (Bruintjes 1999; French et al. 2018; Rauber et al. 2019). The basic hypothesis of glaciogenic cloud seeding is that the introduction of additional ice nucleating particles into a supercooled cloud induces ice nucleation and thus converts supercooled droplets into ice crystals, which grow to precipitation size. Glaciogenic cloud seeding is most effective at temperatures higher than −10°C, because of the low concentration of naturally occurring ice nucleating particles at these temperatures. Subsequently, these ice crystals grow through efficient diffusional and collisional growth mechanisms (Fig. 1) eventually to sizes large enough to sediment to the ground as precipitation. Even though the basic mechanism of glaciogenic cloud seeding is understood, there are still major gaps in our knowledge about the exact pathway of the formation of ice crystals and their growth due to the wide variety of ice shapes and density (Flossmann et al. 2018). In addition, the efficiency of cloud seeding is still heavily debated, mainly because it is difficult to assess how the unseeded cloud would have evolved, i.e., a control experiment is missing in the real atmosphere (e.g., Tessendorf et al. 2012).
The large number of hydrometeors present in a cloud makes it impossible to simulate the evolution of a cloud by explicitly tracking all cloud particles in a model. Instead, numerical weather prediction and climate models must rely on cloud microphysics parameterizations. Major issues of these microphysics schemes are related to uncertainties in microphysical growth rates, especially for ice crystals (Morrison et al. 2020).
The rates of individual microphysical processes are often studied in the laboratory, where experiments can be repeated under controlled conditions. For example, Ryan et al. (1976) and Takahashi et al. (1991) measured the diffusional growth rate of freely suspended ice crystals in a vertical supercooled cloud tunnel at water saturation and temperatures ranging from −21° to −3°C for growth times between 30 and 180 s (Ryan et al. 1976) and between 3 and 30 min, respectively (Takahashi et al. 1991). Bailey and Hallett (2004) measured the ice growth rate by vapor diffusion in ice supersaturation and at temperatures between −20° and −70°C using a horizontal static diffusion chamber. Studies conducted within cloud chambers are crucial to advance our understanding of individual microphysical processes. However, the constraints of their finite dimensions limit the spatial and temporal scales that can be explored (Shaw et al. 2020).
Field measurements capture the microphysical state in a real-world scenario where a variety of atmospheric processes act at the same time (e.g., Korolev et al. 2003). Thus, field observations provide important information on the variability of microphysical cloud properties that occur in nature. However, since every cloud is different, a generalization of the observations requires a large number of cloud cases. Moreover, it is often not possible to obtain individual microphysical process rates from field observations because of the occurrence of multiple processes under unconstrained conditions and the limited capability of repeated sampling of the same cloud. To overcome this problem, Heymsfield et al. (2011) performed multiple penetrations through wave clouds to estimate growth rates of ice crystals at temperatures between −32° and −20°C. Nevertheless, the statement of Gunn (1952) from 70 years ago that “Progress in cloud physics has been seriously limited due to the impossibility of conducting controlled cloud experiments on a sufficiently large scale” still holds largely true today (Shaw et al. 2020).
The CLOUDLAB project aims to overcome some of these limitations by performing targeted cloud seeding in supercooled stratus clouds using a multirotor uncrewed aerial vehicle (UAV). While previous seeding experiments have been conducted mainly to increase precipitation (Bruintjes 1999; Flossmann et al. 2019; French et al. 2018; Rasmussen et al. 2018; Rauber et al. 2019) or mitigate hail damage (Dessens 1986; Dessens et al. 2016; Haupt et al. 2018; Marwitz 1973), the CLOUDLAB project employs seeding as a method to introduce targeted and controlled aerosol perturbations into a well-constrained natural environment (i.e., persistent stratus clouds) to advance our knowledge of the underlying ice formation and growth processes. Seeding is restricted to a small patch of a supercooled stratus cloud to compare it to the initial state (unseeded baseline) and repeated at high frequency in a cloud system that is in an approximate steady state with reasonably simple geometry and well-characterized boundary conditions. Stratus clouds are the ideal natural laboratory because they are the least dynamic cloud type, have relatively little horizontal variability, and can exist for long periods as a metastable mixed-phase cloud (up to a few days over the Swiss Plateau; Scherrer and Appenzeller 2014). Furthermore, Rauber et al. (2019) state in a review of wintertime orographic cloud seeding that the clearest seeding signatures were obtained in shallow orographic clouds containing supercooled cloud droplets and a few ice crystals, where the seeding signal clearly stands out against the background. The seeding experiments here are complemented with simulations of a numerical weather prediction model to gain a deeper insight into the impact of seeding on the cloud structure and dynamics. Having unambiguous seeding signals in the data provides excellent opportunities to constrain and validate numerical simulation, for example, to quantify the precipitation enhancement of seeding (Xue et al. 2022). Since the measurements obtained during CLOUDLAB are conducted within 2–10 min after seeding, the simulations will focus on the initial stages of ice crystal growth and precipitation formation. This article provides an overview of the scientific objectives, experimental design, and first results of the CLOUDLAB project.
The CLOUDLAB project
The CLOUDLAB project was designed to study the rates of ice crystal formation and growth through microphysical processes in natural clouds under well-constrained environmental conditions. To achieve this, CLOUDLAB combines the use of stratus clouds, glaciogenic seeding, and UAVs in a unique way:
- 1)CLOUDLAB uses supercooled stratus clouds that frequently occur over the Swiss Plateau during wintertime to perform glaciogenic seeding experiments. The persistence of stratus clouds allows repeating seeding experiments under similar environmental conditions.
- 2)CLOUDLAB uses two multirotor UAVs: one to introduce localized and controlled aerosol perturbations into supercooled stratus clouds (seeding UAV) and one to measure the induced perturbations (measurement UAV). Unlike seeding with crewed research aircraft, UAVs enable flexible and precise control of the seeding location and repeating of seeding experiments at high frequency.
- 3)CLOUDLAB uses state-of-the-art in situ and ground-based remote sensing instrumentation to thoroughly characterize the seeding-induced microphysical changes. While in situ observations provide single-particle information in unprecedented detail, remote sensing instruments collect continuous observations over a large cloud area.
- 4)CLOUDLAB uses the ICON numerical weather prediction model in large-eddy mode 1) to reproduce and extrapolate the results of the performed cloud seeding experiments, thereby contextualizing these findings within a larger framework, and 2) to perform variations of seeding experiments that are not feasible in the field.
- 5)CLOUDLAB encompasses several field campaigns in the wintertime to obtain a comprehensive observational dataset of seeded and unseeded clouds. As of the time of this writing, 2023, two field campaigns have been conducted and additional campaigns will occur in the following years. The repetitive introduction of targeted and controlled perturbations in a quasi-steady-state cloud environment allows disentangling the contributions of different processes and directly inferring microphysical process rates. The dataset will be used to refine the retrieval techniques of various remote sensing instruments and to improve the ice phase parameterizations of numerical weather prediction models, and with that, precipitation forecasts.
To achieve these objectives, a sophisticated design of seeding experiments and the synergistic use of in situ and remote sensing instrumentation are essential to generate, detect, and track the seeding signature (Fig. 2 and Table 1).
Overview of the instrumentation installed at the main site during the 2022/23 field campaign, including a tethered balloon system, a seeding and measurement UAV, and a large set of ground-based remote sensing devices.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
Instrumentation deployed during the CLOUDLAB campaigns (TBS = tethered balloon system; LACROS = Leipzig Aerosol and Cloud Remote Observations System). Instrumentation marked with an asterisk symbol (*) was not deployed during the 2021/22 campaign.
Experimental design.
Seeding experiments are carried out over the Swiss Plateau near Eriswil, Switzerland. The experimental zone consists of a restricted airspace of 4-km radius, which includes a main site at its center (47°04′14″N, 7°52′22″E) and several UAV launch sites (Fig. 3). One seeding experiment consists of one flight and involves the following two steps: First, the seeding UAV injects ice nucleating particles into supercooled stratus clouds upstream of the main site to induce ice formation and subsequent growth processes. Second, related microphysical changes are measured downwind of the seeding location at the main site by in situ and ground-based remote sensing instrumentation (Fig. 2 and Table 1), including a scanning 35-GHz cloud radar and a tethered balloon system. Due to the flexibility of the platforms and systems involved, the seeding and measurement locations can be easily modified on a single-experiment basis and adjusted to varying wind conditions.
Overview of the experimental area during the CLOUDLAB field campaigns. The main site (blue circle) is located at the center of the experimental area and is surrounded by different UAV launch sites (black triangles). For illustrative purposes, the seeding experiment SM005 (Table 2) is depicted, highlighting the seeding location (blue cross), the seeded plume of ice crystals (white hexagons), the scan pattern of the scanning polarimetric cloud radar (green shaded area), and the prevailing wind speed and direction.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
The seeding flares release particles containing a mixture of silver iodide, silver chloride, an ammonium salt, and a potassium salt as seeding agent (Cloud Seeding Technologies 2021, personal communication), which can initiate freezing at temperatures below −5°C (DeMott 1995; Marcolli et al. 2016). The seeding material is introduced from the seeding UAV 1–4 km upwind of the main site. Both UAVs have an on-demand propeller heating system to reduce icing in supercooled cloud conditions. The seeding UAV is equipped with up to two burn-in-place flares (Cloud Seeding Technologies, Zeus MK2), which can be ignited sequentially to produce a continuous seeding plume. A single flare burns 200 g of material, of which approximately 10% is the ice-active seeding agent, releasing a plume over 5–6 min. For a successful seeding experiment, the seeding plume must pass over the main site, which requires a precise alignment of the seeding and measurement location with respect to the prevailing wind direction. The seeding location is determined prior to each seeding experiment based on wind measurements from regularly launched radiosondes, UAV profiles, and wind profilers (Table 1), and is limited by the dimensions of the restricted airspace (radius: 4 km; altitude: ground up to 2.2 km MSL) and the maximum flight time of the UAV (15–20 min). Multiple UAV launch sites were established in the vicinity of the field site (Fig. 3) to ensure that a suitable launch site is available for the most common wind directions. During the CLOUDLAB campaign 2021/22, the seeding material was released as a point source upwind of the main site. Other seeding patterns, such as horizontal flight tracks perpendicular to the prevailing wind, were implemented during the second CLOUDLAB field campaign (2022/23).
At the main site, a suite of in situ and remote sensing instruments (Table 1) measure the microphysical, dynamical, and meteorological properties of nonseeded and seeded clouds. Measurements before and after seeding are used to characterize the background conditions and variations thereof. Seeding signatures are studied using vertically pointing and scanning polarimetric cloud radars (Table 1) with a focus on the equivalent radar reflectivity (referred to hereafter as radar reflectivity) and linear depolarization ratio, which is the ratio of the power received in the orthogonal channel to that received in the transmission channel, indicating particle sphericity and orientation. While vertically pointed measurements provide information on cloud microphysical properties, scans along and across the prevailing wind direction are useful to obtain information about the dimensions, spatial heterogeneity, and temporal evolution of the seeding signal (Kollias et al. 2014). Additional remote sensing capabilities are provided by means of the mobile Leipzig Aerosol and Cloud Remote Observations System (LACROS), which runs a comprehensive set of multiwavelength, scanning, and polarimetric Doppler-capable cloud radar and lidar systems (Table 1; Radenz et al. 2021). While the full capabilities of LACROS are available for all CLOUDLAB campaigns from 2022/23 onward, only a 94-GHz vertical-pointing cloud radar was provided in the 2021/22 campaign. These synergistic observations are combined by Cloudnet (Illingworth et al. 2007; Tukiainen et al. 2022) to derive standardized and quality-controlled vertical profiles of cloud properties (e.g., liquid and ice water content, hydrometeor type).
The remote sensing observations are complemented by collocated in situ measurements, which characterize the microphysical properties in more detail and can be used as a validation dataset for the remote sensing retrievals. More specifically, a tethered balloon system (TBS; HoloBalloon; Ramelli et al. 2020), equipped with a microphysics, aerosol, and meteorological instrument package, is deployed during the CLOUDLAB campaigns (from campaign 2022/23 onward). The holographic imager (HOLIMO; Table 1) mounted on the tethered balloon system provides detailed information on the shape and number size distribution of cloud droplets and ice crystals and thus yields insights into the growth of ice crystals and the depletion of the liquid phase induced by seeding. Additionally, precipitation is measured at the main site with two Particle Size Velocity (PARSIVEL2) disdrometers and one 2D Video Disdrometer (2DVD). Aerosol properties are continuously monitored at the main site (Table 1) to examine background particle concentrations in the absence of seeding. During intensive observation periods, the aerosol size distribution is also measured with a Portable Optical Particle Spectrometer (POPS) on board the tethered balloon system and the measurement UAV to sample the seeding plume.
Characterization of the seeding plume
The dispersion and particle size distribution of the seeding plume were characterized by the measurement UAV equipped with a POPS in a series of clear-sky seeding experiments. Here, we present the results of one dispersion experiment performed on 25 March 2022, where the seeding UAV hovered 500 m above the main site (1,420 m MSL) while two flares were burning sequentially for 6 min each. At the same time, the measurement UAV flew horizontal transects across the plume (i.e., perpendicular to the wind direction) downwind of the seeding UAV to characterize the dispersion of the plume.
At a distance of 350 m downwind of the plume (corresponding to 50 s after emission), POPS measured particle concentrations of up to 104 cm−3 inside the plume, whereas the concentration measured in the background atmosphere was one order of magnitude lower at around 7 × 102 cm−3 (Fig. 4a). The mean horizontal dispersion of the plume was around 50 m at 350 m downwind, though in single transects, the plume widths were smaller (20–30 m) and the position of the maximum concentration varied due to fluctuations in wind conditions. During further propagation, turbulence will disperse the seeding plume even further. A thorough analysis of the CLOUDLAB dispersion experiments, focusing on the horizontal and vertical dispersion of the seeding particles, will be presented in an upcoming CLOUDLAB publication (Miller et al. 2023).
Characterization of the dispersion and particle size distribution of the seeding plume. (a) Horizontal dispersion of the plume at 350 m downwind from the seeding location (i.e., after 50 s with the wind speed of 7 m s−1). The particle concentrations measured during transects across the plume (each color and marker represent an individual transect) and the mean concentration (blue line) are shown. The shaded band represents the 75% confidence interval. (b) Mean particle size distribution of the plume (solid blue line) and the background (dashed gray line) measured by the POPS mounted on the measurement UAV during the dispersion experiment shown in (a). Background concentrations exceeding the plume concentrations at the larger sizes (>800 nm) are likely a result of sampling uncertainty caused by the low concentration of such large particles.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
POPS measured particles in the size range between 130 and 2,000 nm during the clear-sky seeding experiment (Fig. 4b). The majority of the seeding particles had a diameter below 200 nm. Due to the ammonium and potassium salts contained in the flare material, the seeding particles are expected to be highly hygroscopic and thus expected to act as cloud condensation nuclei if they are exposed to supersaturated conditions. According to previous studies, seeding particles can nucleate ice at temperatures below −5°C through various freezing mechanisms (DeMott 1995; Marcolli et al. 2016). However, it is unclear whether the seeding particles first activate into cloud droplets and then nucleate ice through immersion freezing, or whether the seeding particles collide with existing cloud droplets and cause contact freezing. In upcoming CLOUDLAB publications, we will examine which ice nucleation mechanism is dominant in the temperature range between −10° and 0°C (e.g., immersion or contact nucleation) and investigate the seeding efficiency for different environmental conditions.
Cloud seeding experiments
Over the course of the first two CLOUDLAB field campaigns (2021/22, 2022/23), a total of 55 seeding experiments were conducted in supercooled stratus clouds over the Swiss Plateau. While the first campaign served as a proof-of-concept for upcoming campaigns, clouds covering a wide range of environmental conditions were sampled during the 2022/23 campaign: seeding temperature (from –10° to –5°C), wind speed (2–14 m s−1), liquid water path (0–300 g m−2), and time of observation after seeding (2–10 min). Here we present first findings of the CLOUDLAB seeding experiments performed within persistent supercooled stratus clouds (case 1) and turbulent updraft structures (case 2) (Table 2).
Mission parameters and meteorological conditions during the cloud seeding experiments conducted on 26 Feb 2022 (SM005) and 25 Jan 2023 (SM058–SM060). Temperature data were obtained from the seeding UAV, wind data from the seeding UAV (SM005) and wind profiler (SM058–SM060).
Case 1: Persistent low stratus clouds.
Between 21 and 28 January 2023, the weather situation in Europe was characterized by a low pressure system over Italy and a high pressure ridge that extended from the Atlantic to eastern Europe (Fig. 5a) causing a Bise situation (i.e., cold northeasterly wind; Wanner and Furger 1990). The northeasterly wind brought cold and moist air masses toward Switzerland and led to the formation of a persistent low stratus cloud. These conditions persisted over a week and provided an optimal environment for conducting a total of 35 successful seeding experiments.
Synoptic weather situation and radiosonde profile on (top) 25 Jan 2023 and (bottom) 26 Feb 2022. (a),(c) ERA5 data (Hersbach et al. 2023): Temperature at 850 hPa (°C; shaded) and geopotential height at 500 hPa (dam; black lines). The white star indicates the location of the main site. Vertical profile of the temperature (solid black line), potential temperature (dashed black line), relative humidity (solid blue line), and wind speed and direction (wind barbs) measured by the radiosonde launched from the main site (b) at 1209 UTC 25 Jan 2023 and (d) at 0608 UTC 26 Feb 2022. The extent of the cloud is depicted by the gray-shaded area and the seeding height is indicated by the pink dotted line.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
On the morning of 25 January 2023, three consecutive seeding experiments (SM058–SM060, Table 2) were conducted in a low stratus cloud with a cloud-top temperature of −6.8°C (Fig. 5b). During each of these missions, the seeding UAV flew four 400-m legs perpendicular to the wind direction at a height of 1,300 m MSL (i.e., seeding temperature of −5.5°C). The three missions were carried out at different distances between the seeding location and the main site (2, 2.5, and 3 km, see Table 2), while the other experimental parameters were kept constant. The background properties of the low stratus cloud remained approximately constant during the three missions and were characterized by a liquid water content of 0.25–0.3 g m−3, a cloud droplet concentration of 400–500 cm−3 and an ice water content of below 0.001 g m−3.
The microphysical changes caused by seeding were detected by the vertically pointing 35-GHz cloud radar and the tethered balloon system at the main site. During the passage of the seeding plume, the radar reflectivity increased from the natural background of −25 dBZ (i.e., typical values for stratus clouds), up to 0 dBZ (i.e., typical values for mixed-phase clouds) (Fig. 6). Seeding signals were observed over the whole vertical extent of the cloud and for the whole burn time of the seeding flare (5 min). This large increase in the radar reflectivity (>20 dBZ, i.e., two orders of magnitude) relative to the background reflectivity can be related to the seeding-induced microphysical changes, namely, the formation and growth of ice crystals. Indeed, high concentrations of ice crystals were simultaneously observed by the tethered balloon system during the passage of the seeding plume (detected by the increase in aerosol concentration) (Fig. 7). The holographic imager measured ice crystal number concentrations of up to 2,000 L−1 (mainly columns) during SM059, which was five orders of magnitude higher than the background ice crystal concentrations (<0.01 L−1). The observation of such high concentrations of ice crystals is rare in natural clouds, even in regions with strong secondary ice production (Korolev et al. 2020). Interestingly, despite the short life span of the ice crystals (4–6 min), some ice crystals indicate signs of aggregation and riming (Fig. 7). The lower concentrations observed at the beginning of SM059 (1057–1059 UTC, Fig. 7) can be attributed to the tethered balloon only measuring the edge of the seeded area (Fig. 6).
Radar reflectivity measured at the main site by the vertically pointing 35-GHz cloud radar during the seeding missions SM058–SM060 conducted on 25 Jan 2023. The ignition times of the seeding flare (black dots), the balloon height (black line), and the time period of the expected seeding signals (black bar on top) are highlighted. The time periods of the expected seeding signals are estimated based on the wind speed and the horizontal distance between the seeding and measurement location (Table 2).
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
Aerosol (black line, in size range 115–3,370 nm) and ice crystal (red line) concentrations of seeding mission SM059 measured by the optical counter POPS and the holographic imager HOLIMO, respectively, mounted on the tethered balloon. A randomly selected sample of ice crystal images is shown in the red box.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
With increasing distance between the seeding location to the main site, the radar reflectivity increased from −14.96 dBZ (SM059: 2 km) to −13.03 dBZ (SM058: 2.5 km) to −11.17 dBZ (SM060: 3 km) (Fig. 6, mean radar reflectivities computed after applying a threshold of ≥−20 dBZ). Assuming a wind speed of 8 m s−1 (Table 2), the ice crystals had an additional two minutes to grow to larger sizes by diffusion in SM060 compared to SM059, resulting in a higher radar reflectivity. The radar echo descended to lower altitudes with increasing distance downwind of the seeding location (Fig. 6). This pattern is consistent with observations made by French et al. (2018) where it is attributed to the larger ice crystal size and higher ice crystal fall velocity. Upcoming CLOUDLAB publications will investigate the ice crystal growth rates in the temperature range between −10° and 0°C in more detail by combining collocated in situ and ground-based remote sensing observations obtained during the CLOUDLAB campaigns.
Case 2: Turbulent updraft structures.
On 26 February 2022, the synoptic weather situation over Europe was characterized by a high pressure system located over France and a low pressure system over Scandinavia (Fig. 5c). A small-scale disturbance brought cold air masses toward Switzerland and led to the formation of a supercooled stratus cloud with a cloud-top temperature of −10°C above our target area (Figs. 5d, 8).
Radar reflectivity and Doppler velocity measured by the vertically pointing 94-GHz cloud radar during the seeding mission SM005 conducted on 26 Feb 2022. The ignition time of the seeding flare (black dot) and the time period of the expected seeding signal (black bar on top) are highlighted. The time period of the expected seeding signal is estimated based on the wind speed and the horizontal distance between the seeding and measurement location (Table 2). Due to stationary seeding and fluctuations in wind direction, which prevented perfect alignment between the seeding and measurement location, the seeding signal is visible for about 25% of the flare burning time in the vertical radar measurements. The gray box indicates the area of the generating cell considered in Fig. 10b (purple markers).
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
One seeding mission (SM005) was carried out in the low stratus cloud at 0805 UTC (Table 2). The cloud had turbulent cloud-top structures (i.e., cloud-top generating cells) with vertical variation in the cloud-top height of up to 300 m and updraft velocities up to 2 m s−1 (Fig. 8b). During SM005, seeding material was released in the turbulent structures near cloud top at an altitude of 1.8 km MSL and at a seeding temperature of −8°C. Stationary seeding was performed, i.e., the seeding UAV hovered for 5 min 24 s at the seeding height (Table 2).
Figure 8 highlights the microphysical changes induced by cloud seeding, as detected by the vertically pointing 94-GHz cloud radar located 2.3 km downwind of the seeding location. Seeding signatures were identified by regions of enhanced radar reflectivities (up to −15 dBZ) compared to the natural background radar reflectivity (−30 dBZ). The seeding signal also exceeded the radar reflectivity (−23 dBZ) observed inside the generating cells (e.g., Fig. 8a at 0756 and 0803 UTC) within which natural ice production occurred. The seeding signal appeared 4 min after the initiation of seeding in the radar echo and was detected up to 200 m above the seeding altitude (1.8 km MSL) as a result of turbulent updraft structures near cloud top (Fig. 8b).
The seeding signatures were also tracked further downstream (around 4 km downstream of the seeding location) using a scanning 35-GHz cloud radar (Fig. 9). Sector scans indicate regions of enhanced radar reflectivities (between −20 and −10 dBZ), which stand out clearly from the background cloud (−30 dBZ). The dimensions of the enhanced radar echo ranged between 500 and 800 m (horizontal extent) and between 200 and 400 m (vertical extent). The seeding signal appeared 6–7 min after the ignition of the seeding flare in the sector scans. The radar reflectivity measured by the scanning radar (after 6–7 min) was 2–5 dBZ higher compared to the radar reflectivity measured by the vertically pointing cloud radar (after 4 min) (Fig. 10a), in agreement with the longer diffusional growth time and thus larger particle size. The linear depolarization ratio of −29 dB measured by the cloud radars (Fig. 10b) indicated the presence of oblate (i.e., platelike) particles (Myagkov et al. 2016; Bühl et al. 2016).
(a) Radar reflectivity, (b) radar reflectivity with a −20-dBZ threshold filter applied, and (c) linear depolarization ratio LDR measured by the scanning 35-GHz cloud radar during the seeding mission SM005 conducted on 26 Feb 2022. One sectorial scan at 40° elevation lasting for 1.5 min (0811:26–0812:57 UTC) is shown.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
(a) Height-resolved radar reflectivity and (b) linear depolarization ratio (LDR) during the seeding experiment SM005. From the vertically pointing radar (triangles), the seeding signal is shown in orange and the background in light blue. From the scanning radar (circles), the seeding signal is shown in red and the background in dark blue. The purple color indicates the radar reflectivity measured inside generating cells, as indicated by the gray box in Fig. 8. LDR is only shown if the signal is above the noise level. Shaded areas indicate standard errors and markers indicate the mean of the respective 50-m height interval.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
Higher radar reflectivities were measured with increasing altitude during SM005 (Fig. 10a), likely because cloud particles were exposed to different ambient conditions. Cloud particles that were lifted to higher altitudes might have experienced higher supersaturation and thus grew to larger sizes, resulting in higher radar reflectivity. In upcoming CLOUDLAB publications, we will investigate how generating cells and other cloud inhomogeneities influence ice crystal growth.
Cloud seeding in numerical weather model
In CLOUDLAB, the numerical weather prediction model ICON 2.6.5 is operated in large-eddy mode (Zängl et al. 2015) using the existing two-moment scheme (Seifert and Beheng 2006), which tracks the number concentration and mass mixing ratio of hydrometeors. An additional immersion freezing parameterization for the seeding particles was implemented (Marcolli et al. 2016). The seeding experiments conducted on 25 January 2023 were represented by large-eddy simulations in a nested setup with a final horizontal resolution of 130 m and 80 vertical levels. The simulated domain is 40 km × 30 km with the main site in the center. The model simulation was initialized at 0000 UTC 25 January 2023 with COSMO-1 analysis at 1-km horizontal resolution, which combines model simulations with observational data, as initial and boundary conditions.
The meteorological conditions in the morning of the experiments were characterized by northeasterly to easterly winds, low clouds, and a temperature inversion (Fig. 5b). The model predicted low clouds with a lower cloud top and a weaker temperature inversion than was observed (Fig. 11a). The cloud-top temperature was found to be around −4°C, which is 2°–3°C warmer than observed (Fig. 11a). We introduced seeding particles along a leg of 400 m with a concentration of 106 m−3 s−1 from 1050 to 1055 UTC mimicking the seeding setup of experiment SM059. The lower cloud top enforced a slightly lower seeding height (≈1,200 m, Fig. 11a, red cross) than in the field (1,300 m). Figure 11b shows the vertical profile and the horizontal extent of the ice concentration 5 min after seeding ended. The reference simulation without any seeding particles did not predict any ice crystals (not shown). The seeding simulation shows ice crystal concentrations up to 9 L−1 with the maximum at the seeding height. The increase in ice concentration is not constrained to the seeding height but is visible across 300 m in the vertical. Additionally, it is clear that the seeded ice crystals are moving downstream of the seeding location toward the main site (see inset in Fig. 11b). However, the slightly warmer temperatures in the model resulted in lower ice activation of the seeding particles into ice crystals (9 L−1) compared to the field observations (up to 2,000 L−1, Fig. 7). These results show that the observed seeding signals can be represented and reproduced in numerical weather prediction models if the environmental conditions are simulated correctly. Future work aims to improve parameterizations of ice crystal formation and growth processes by constraining and validating the numerical simulation with the CLOUDLAB observations obtained in the field.
Model simulation of the seeding experiment SM059 conducted on 25 Jan 2023. Seeding particles were released for 5 min along a 400-m transect perpendicular to the wind direction at 1.2 km MSL. (a) Cross section of cloud cover along mean wind direction (black arrow) at the starting time of seeding. The isotherms (colored lines), the seeding location (red cross), the main site (blue dot), and the model topography (gray) are shown. (b) Vertical profile and horizontal cross section (inset, dimension: 3.5 km × 2.5 km) of the maximum ice concentration 10 min after seeding started.
Citation: Bulletin of the American Meteorological Society 104, 11; 10.1175/BAMS-D-22-0178.1
Summary and outlook
This paper introduces the CLOUDLAB project and highlights the potential of targeted glaciogenic cloud seeding of supercooled stratus clouds with UAVs for studying microphysical ice processes systematically. The first findings of the CLOUDLAB project are as follows:
- 1)We demonstrated that a multirotor UAV can be used for targeted glaciogenic cloud seeding. The UAV can safely fly into supercooled cloud regions, burn a flare without being damaged, and has enough battery capacity to complete a seeding mission (10–15 min) while heating the rotors. Flexibility in the launch location of the UAV allowed the seeding patch to pass over the main site and be measured by the ground-based remote sensing and in situ instrumentation.
- 2)The dispersion of the seeding plume was characterized by a second multirotor UAV equipped with an optical particle counter. Although most of the particles had a diameter below 200 nm, particles up to 1,000 nm were observed inside the seeding plume.
- 3)The microphysical changes induced by the seeding were detected by a vertically pointing and a scanning cloud radar 4–10 min after igniting the seeding flare. High radar reflectivity (from −20 to 0 dBZ), which was clearly above the background cloud (−30 dBZ), was measured over a region spanning 500–800 m. The measured radar reflectivities increased with increasing seeding distance, in accordance with the longer growth time. Simultaneously with the increase in radar reflectivity, high aerosol and ice crystal (up to 2,000 L−1) number concentrations were measured by the instrumentation mounted on the tethered balloon system.
- 4)The simulation of the seeding experiment SM059 in the weather prediction model ICON in large-eddy mode showed an increase in ice concentration (9 L−1) after injecting particles along a 400-m leg mimicking the setup of the field experiment. The ice concentration was lower than observed, which can be attributed to the warmer temperatures present in the model leading to a lower activation of seeding particles. These findings show the potential of model sensitivity studies with strong experimental constraints.
The case studies presented here serve as a proof-of-concept that the CLOUDLAB seeding approach is feasible: 1) the seeding material released by the UAV creates a seeding signal that is large enough to be detected by the remote sensing and in situ instrumentation, and 2) the persistent nature of low stratus clouds allows for repeating seeding experiments under similar and well-constrained environmental conditions and for performing microphysical growth studies by only changing one experimental parameter in a controlled manner (e.g., seeding distance). The aim of future CLOUDLAB campaigns is to conduct seeding experiments with high frequency and with varying experimental parameters (e.g., residence time) to explore microphysical ice processes (e.g., diffusional growth rate) under different environmental conditions (e.g., seeding temperatures). The goal is to use this dataset to improve the microphysics scheme in weather prediction models to increase precipitation forecast skills. The unique combination of remote sensing and in situ data will improve our process understanding of ice formation and growth in mixed-phase clouds.
Acknowledgments.
This project has received funding from the European Research Council (ERC) under the European Union’s Horizon 2020 research and innovation programme (Grant Agreement 101021272 CLOUDLAB). The deployment of the 94-GHz radar in the 2021/22 campaign was part of a Transnational access project that is supported by the European Commission under the Horizon 2020–Research and Innovation Framework Programme, H2020-INFRAIA-2020-1, ATMO-ACCESS Grant Agreement 101008004. The deployment of LACROS in campaigns from 2022/23 onward has received funding from the European Union’s Horizon 2020 research and innovation programme under Grant Agreement 654109 and from the Deutsche Forschungsgemeinschaft (DFG; German Research Foundation)—Project 408027490.
We acknowledge access to Piz Daint at the Swiss National Supercomputing Centre, Switzerland, under Project ID s1144.
The authors acknowledge the support of MeteoSwiss, namely, Daniel Leuenberger and the COSMO team, for providing weather forecasts which helped us decide on suitable conditions. Furthermore, we thank Maxime Hervo and his team for providing and installing the wind profiler. Also, a special thanks goes to Stephanie Westerhuis, who supported us in the model setup and facilitated access to COSMO data at MeteoSwiss.
Additionally, we are grateful to Meteomatics, namely, Lukas Hammerschmidt, Philipp Kryenbühl, Daniel Schmitz, and Remo Steiner for the iterative development of the UAV flight pattern and for the modification of the UAV to mount and ignite the flares and to integrate the POPS instrument.
The authors thank Jürg Wildi and Philip Bärtschi from v2sky for the efficient handling of the application for the flight permits and Santiago Llucià, Jeroen Kroese, and Judith Baumann from the Federal Office of Civil Aviation (FOCA) for efficient communication during the approval process.
Furthermore, we thank Mathias Bauer and Hendrik Brast from Metek for support before, during, and after the field campaign, such as delivering and installing the scanning radar, as well as helping with the application for concessions.
We are indebted to the Swiss Army and the Gütergemeinde Hinterdorf Eriswil, who allowed us to conduct the experiment at their locations, providing a base with sufficient power. We thank Stefan Minder for the maintenance of the base and Eriswil firefighters for their help when in need.
Data availability statement.
The data from the vertically pointing remote sensing instruments are available on the CLOUDNET site: https://cloudnet.fmi.fi/site/eriswil. Data from the in situ instruments, the radar scans, and the model simulations and the scripts for the figure are available on the ETH reasearch collection: 10.3929/ethz-b-000638794.
References
Bailey, M., and J. Hallett, 2004: Growth rates and habits of ice crystals between −20° and −70°C. J. Atmos. Sci., 61, 514–544, https://doi.org/10.1175/1520-0469(2004)061<0514:GRAHOI>2.0.CO;2.
Bergeron, T., 1935: On the physics of cloud and precipitation. Proc. Fifth Assembly of the Int. Union of Geodesy and Geophysics, Lisbon, Portugal, IUGG, 156–178.
Bruintjes, R. T., 1999: A review of cloud seeding experiments to enhance precipitation and some new prospects. Bull. Amer. Meteor. Soc., 80, 805–820, https://doi.org/10.1175/1520-0477(1999)080<0805:AROCSE>2.0.CO;2.
Bühl, J., P. Seifert, A. Myagkov, and A. Ansmann, 2016: Measuring ice- and liquid-water properties in mixed-phase cloud layers at the Leipzig Cloudnet station. Atmos. Chem. Phys., 16, 10 609–10 620, https://doi.org/10.5194/acp-16-10609-2016.
David, R. O., and Coauthors, 2019: Development of the Droplet Ice Nuclei Counter Zurich (DRINCZ): Validation and application to field-collected snow samples. Atmos. Meas. Tech., 12, 6865–6888, https://doi.org/10.5194/amt-12-6865-2019.
DeMott, P., 1995: Quantitative descriptions of ice formation mechanisms of silver iodide-type aerosols. Atmos. Res., 38, 63–99, https://doi.org/10.1016/0169-8095(94)00088-U.
Dessens, J., 1986: Hail in southwestern France. II: Results of a 30-year hail prevention project with silver iodide seeding from the ground. J. Climate Appl. Meteor., 25, 48–58, https://doi.org/10.1175/1520-0450(1986)025<0048:HISFIR>2.0.CO;2.
Dessens, J., J. L. Sánchez, C. Berthet, L. Hermida, and A. Merino, 2016: Hail prevention by ground-based silver iodide generators: Results of historical and modern field projects. Atmos. Res., 170, 98–111, https://doi.org/10.1016/j.atmosres.2015.11.008.
Engelmann, R., and Coauthors, 2016: The automated multiwavelength Raman polarization and water-vapor lidar PollyXT: The next generation. Atmos. Meas. Tech., 9, 1767–1784, https://doi.org/10.5194/amt-9-1767-2016.
Findeisen, W., 1938: Die kolloidmeteorologischen Vorgänge bei der Niederschlagsbildung. Meteor. Z., 55, 121–133.
Flossmann, A. I., M. Manton, A. Abshaev, R. Bruintjes, M. Murakami, T. Prabhakaran, and Z. Yao, 2018: Peer review report on global precipitation enhancement activities. WMO Research Rep. WWRP 2018-1, 126 pp., https://library.wmo.int/doc_num.php?explnum_id=9945.
Flossmann, A. I., M. Manton, A. Abshaev, R. Bruintjes, M. Murakami, T. Prabhakaran, and Z. Yao 2019: Review of advances in precipitation enhancement research. Bull. Amer. Meteor. Soc., 100, 1465–1480, https://doi.org/10.1175/BAMS-D-18-0160.1.
French, J. R., and Coauthors, 2018: Precipitation formation from orographic cloud seeding. Proc. Natl. Acad. Sci. USA, 115, 1168–1173, https://doi.org/10.1073/pnas.1716995115.
Gunn, R., 1952: A vertical shaft for the production of thick artificial clouds and for the study of precipitation mechanics. J. Appl. Phys., 23, 1–5, https://doi.org/10.1063/1.1701971.
Haupt, S. E., R. M. Rauber, B. Carmichael, J. C. Knievel, and J. L. Cogan, 2018: 100 years of progress in applied meteorology. Part I: Basic applications. A Century of Progress in Atmospheric and Related Sciences: Celebrating the American Meteorological Society Centennial, Meteor. Monogr., No. 59, Amer. Meteor. Soc., https://doi.org/10.1175/AMSMONOGRAPHS-D-18-0004.1.
Hersbach, H., and Coauthors, 2023: ERA5 hourly data on pressure levels from 1940 to present. C3S CDS, accessed 2 June 2023, https://doi.org/10.24381/cds.bd0915c6.
Heymsfield, A. J., P. R. Field, M. Bailey, D. Rogers, J. Stith, C. Twohy, Z. Wang, and S. Haimov, 2011: Ice in Clouds Experiment—Layer Clouds. Part I: Ice growth rates derived from lenticular wave cloud penetrations. J. Atmos. Sci., 68, 2628–2654, https://doi.org/10.1175/JAS-D-11-025.1.
Heymsfield, A. J., C. Schmitt, C.-C.-J. Chen, A. Bansemer, A. Gettelman, P. R. Field, and C. Liu, 2020: Contributions of the liquid and ice phases to global surface precipitation: Observations and global climate modeling. J. Atmos. Sci., 77, 2629–2648, https://doi.org/10.1175/JAS-D-19-0352.1.
Illingworth, A. J., and Coauthors, 2007: Cloudnet: Continuous evaluation of cloud profiles in seven operational models using ground-based observations. Bull. Amer. Meteor. Soc., 88, 883–898, https://doi.org/10.1175/BAMS-88-6-883.
Kollias, P., N. Bharadwaj, K. Widener, I. Jo, and K. Johnson, 2014: Scanning ARM cloud radars. Part I: Operational sampling strategies. J. Atmos. Oceanic Technol., 31, 569–582, https://doi.org/10.1175/JTECH-D-13-00044.1.
Korolev, A., 2007: Limitations of the Wegener–Bergeron–Findeisen mechanism in the evolution of mixed-phase clouds. J. Atmos. Sci., 64, 3372–3375, https://doi.org/10.1175/JAS4035.1.
Korolev, A., G. A. Isaac, S. G. Cober, J. W. Strapp, and J. Hallett, 2003: Microphysical characterization of mixed-phase clouds. Quart. J. Roy. Meteor. Soc., 129, 39–65, https://doi.org/10.1256/qj.01.204.
Korolev, A., and Coauthors, 2020: A new look at the environmental conditions favorable to secondary ice production. Atmos. Chem. Phys., 20, 1391–1429, https://doi.org/10.5194/acp-20-1391-2020.
Marcolli, C., B. Nagare, A. Welti, and U. Lohmann, 2016: Ice nucleation efficiency of AgI: Review and new insights. Atmos. Chem. Phys., 16, 8915–8937, https://doi.org/10.5194/acp-16-8915-2016.
Marwitz, J. D., 1973: Hailstorms and hail suppression techniques in the U.S.S.R.—1972. Bull. Amer. Meteor. Soc., 54, 317–325, https://doi.org/10.1175/1520-0477(1973)054<0317:HAHSTI>2.0.CO;2.
Miller, A. J., and Coauthors, 2023: Two new multirotor UAVs for glaciogenic cloud seeding and aerosol measurements within the CLOUDLAB project. Atmos. Meas. Tech. Discuss., https://doi.org/10.5194/amt-2023-157.
Morrison, H., and Coauthors, 2020: Confronting the challenge of modeling cloud and precipitation microphysics. J. Adv. Model. Earth Syst., 12, e2019MS001689, https://doi.org/10.1029/2019MS001689.
Mülmenstädt, J., O. Sourdeval, J. Delanoë, and J. Quaas, 2015: Frequency of occurrence of rain from liquid-, mixed-, and ice-phase clouds derived from A-Train satellite retrievals. Geophys. Res. Lett., 42, 6502–6509, https://doi.org/10.1002/2015GL064604.
Myagkov, A., P. Seifert, M. Bauer-Pfundstein, and U. Wandinger, 2016: Cloud radar with hybrid mode towards estimation of shape and orientation of ice crystals. Atmos. Meas. Tech., 9, 469–489, https://doi.org/10.5194/amt-9-469-2016.
Radenz, M., and Coauthors, 2021: Hemispheric contrasts in ice formation in stratiform mixed-phase clouds: Disentangling the role of aerosol and dynamics with ground-based remote sensing. Atmos. Chem. Phys., 21, 17 969–17 994, https://doi.org/10.5194/acp-21-17969-2021.
Ramelli, F., A. Beck, J. Henneberger, and U. Lohmann, 2020: Using a holographic imager on a tethered balloon system for microphysical observations of boundary layer clouds. Atmos. Meas. Tech., 13, 925–939, https://doi.org/10.5194/amt-13-925-2020.
Rasmussen, R. M., and Coauthors, 2018: Evaluation of the Wyoming Weather Modification Pilot Project (WWMPP) using two approaches: Traditional statistics and ensemble modeling. J. Appl. Meteor. Climatol., 57, 2639–2660, https://doi.org/10.1175/JAMC-D-17-0335.1.
Rauber, R. M., and Coauthors, 2019: Wintertime orographic cloud seeding—A review. J. Appl. Meteor. Climatol., 58, 2117–2140, https://doi.org/10.1175/JAMC-D-18-0341.1.
Ryan, B. F., E. R. Wishart, and D. E. Shaw, 1976: The growth rates and densities of ice crystals between −3°C and −21°C. J. Atmos. Sci., 33, 842–850, https://doi.org/10.1175/1520-0469(1976)033<0842:TGRADO>2.0.CO;2.
Scherrer, S. C., and C. Appenzeller, 2014: Fog and low stratus over the Swiss Plateau—A climatological study. Int. J. Climatol., 34, 678–686, https://doi.org/10.1002/joc.3714.
Seifert, A., and K. D. Beheng, 2006: A two-moment cloud microphysics parameterization for mixed-phase clouds. Part 1: Model description. Meteor. Atmos. Phys., 92, 45–66, https://doi.org/10.1007/s00703-005-0112-4.
Shaw, R. A., and Coauthors, 2020: Cloud–aerosol–turbulence interactions: Science priorities and concepts for a large-scale laboratory facility. Bull. Amer. Meteor. Soc., 101, E1026–E1035, https://doi.org/10.1175/BAMS-D-20-0009.1.
Takahashi, T., T. Endoh, G. Wakahama, and N. Fukuta, 1991: Vapor diffusional growth of free-falling snow crystals between −3 and −23°C. J. Meteor. Soc. Japan, 69, 15–30, https://doi.org/10.2151/jmsj1965.69.1_15.
Tessendorf, S. A., and Coauthors, 2012: The Queensland Cloud Seeding Research Program. Bull. Amer. Meteor. Soc., 93, 75–90, https://doi.org/10.1175/BAMS-D-11-00060.1.
Tukiainen, S., E. O’Connor, and K. Anniina, 2022: CloudnetPy, v1.34.0. Zenodo, https://doi.org/10.5281/zenodo.6669042.
Vonnegut, B., 1947: The nucleation of ice formation by silver iodide. J. Appl. Phys., 18, 593–595, https://doi.org/10.1063/1.1697813.
Wanner, H., and M. Furger, 1990: The Bise—Climatology of a regional wind north of the Alps. Meteor. Atmos. Phys., 43, 105–115, https://doi.org/10.1007/BF01028113.
Wegener, A., 1911: Thermodynamik der Atmosphäre. J. A. Barth, 331 pp.
Xue, L., and Coauthors, 2022: Comparison between observed and simulated AgI seeding impacts in a well-observed case from the SNOWIE field program. J. Appl. Meteor. Climatol., 61, 345–367, https://doi.org/10.1175/JAMC-D-21-0103.1.
Zängl, G., D. Reinert, P. Rípodas, and M. Baldauf, 2015: The ICON (Icosahedral Non-Hydrostatic) modelling framework of DWD and MPI-M: Description of the non-hydrostatic dynamical core. Quart. J. Roy. Meteor. Soc., 141, 563–579, https://doi.org/10.1002/qj.2378.
