Validation Report: Drone-Based, Glaciogenic Cloud Seeding in the Kenai Peninsula

Executive Summary

Rainmaker has successfully validated the impact of drone-based, glaciogenic cloud seeding operations in the Kenai Peninsula, AK on August 22–23, 2026. Over a three-hour period, it executed seven coordinated seeding missions, responding to a strengthening low pressure system spinning in the Gulf of Alaska that brought northeasterly flow. From these flights, seven unique seeding signatures were identified on radar utilizing the PAHG NEXRAD. These seeding signatures were identified within weakening, yet ongoing, natural background precipitation. Using a QPE ensemble method similar to Friedrich et al. (2020), Rainmaker estimates these seeding signatures created 45—65 acre-feet of net new precipitation, with a mean value of 57.6 acre-feet.

These results serve as a critical proof of concept for glaciogenic cloud seeding technology. To date, this technology remains the best studied and most extensively validated form of precipitation enhancement. For this reason, it has been deployed in the Colorado River Basin, throughout the American West, and around the world. Cloud seeding supports agricultural, industrial, municipal, and environmental applications, including restoration of the Great Salt Lake.

Rainmaker plans to continue refinement of its cloud seeding platform in the upcoming winter season.

Background

Atmospheric Conditions

Model guidance placed the target region north of a strengthening closed 500 mb low with stacked surface low pressure system over the Gulf of Alaska (Figure 1). 500 mb flow remained relatively weak (0–20 knots). Soundings further showed winds in the lower half of the atmosphere remained below 20 knots with wind direction generally northeasterly (Figure 2). These observed wind speeds were well within the operational limits of Rainmaker’s Elijah drone.

Soundings launched from the seeding site placed the average freezing level near 1,874 m MSL, with the -5°C isotherm near 2,813 m and the -15°C isotherm near 4,756 m (Figure 2), corresponding to the temperature range deemed suitable for glaciogenic cloud seeding with AgI (e.g., Geerts and Rauber 2022). The elevated freezing level means surface precipitation fell as rain in the valleys and as snow over higher mountainous terrain. Multi-radar multi-sensor system (MRMS) precipitation estimates indicate light precipitation with less than 5 mm of rain falling over the western part of the Kenai Peninsula during the focused seeding period from 06:18 to 09:54 UTC 23 August 2026 (Figure 3).

Ceilometer observations co-located with the seeding site suggest several different cloud layers with bases as low as 1,300 m MSL and as high as 3,500 m MSL (Figure 4). This suggests a varying cloud depth over which we would expect ice to grow between seeding altitude and the freezing level. Sounding moisture profiles indicate nearly saturated environments throughout the atmospheric column (Figure 2), suggesting no lack of clouds. GOES satellite imagery indicates colder upper-level clouds were present near the seeding site with cloud tops near -27°C (Figure 1). These cloud top temperatures are warmer than -40°C where homogeneous ice nucleation is expected, making natural ice less abundant and clouds more suitable for targeted glaciogenic seeding (e.g., Rauber et al. 2019).

A vibrating-wire supercooled liquid water (SLW) and ice concentration sensor (Anasphere UWC2) mounted on launched soundings observed consistent mixed phase conditions (ice and supercooled liquid water both present) below ~600 mb (~ -12°C) to the -5°C level that should be conducive to seeding with AgI depending on exactly how much natural ice is present (Figure 2). Note that large uncertainty exists in these SLW measurements, but even weak increases in SLW may indicate pockets of SLW existed in cloud.

Overall, atmospheric conditions appeared suitable for glaciogenic seeding. At targeted seeding levels, sounding temperatures were near -10 °C with relative humidity near 83%, placing the layer well below the -5°C minimum AgI activation temperature and likely in cloud. Observed winds near 10 knots were also comfortably below the 45 knot drone limit. The main controlling factor is how much natural ice was present as SLW was only observed intermittently and the satellite-observed cloud top was ice, so the environment was favorable but not uniformly SLW-rich.

Satellite — GOES IR (enhanced) + GFS 500 mb heights / winds

Figure 1. Modeled atmospheric conditions and satellite observations looping over the seeding analysis period at 30 min cadence. GOES ABI satellite imagery is from the NOAA public AWS archive via goes2go; 500 mb fields are from the GFS model through anvil.gfs. White/black contours show geopotential height, black barbs show wind (kt), and the star marks the seeding site. Valid time and analysis/forecast status appear in each frame.

Soundings (Skew-T sequence)

Figure 2. Soundings from launches over the seeding analysis period are plotted as Skew-T diagrams in the left panel. Relative humidity and supercooled liquid water/ice concentrations are plotted in the right panel.

Precipitation over the seeding window — MRMS reflectivity and QPE accumulation

Figure 3. Seeding window 06:18 – 09:54 UTC 23 Aug 2026 (flare releases ±20 min). Left: MRMS lowest-altitude reflectivity, with times of active flare dispersion annotated in red; ★ = site. Right: MRMS PrecipRate accumulated from the window start, integrated at the native 2-min cadence.

Ceilometer backscatter + detected flights

Figure 4. CL31 backscatter (log10) over time and height in meters MSL; black = cloud base. Drone tracks in gray, flare-dispersal (seeding) samples in red.

Seeding Flights & Timeline

Across the seven coordinated seeding missions, 19 red flares were released by drones EL-151 and EL-153 between 06:38:18 and 09:33:41 UTC (Table 1). Each coordinated mission is a grouping of flares that are released in quick succession, in this case either by a single or both drones. Missions 1 and 2 each consisted of four flares released nearly simultaneously by both drones, while Mission 3 used three flares between the two drones. Missions 4–7 were conducted solely by EL-151 and consisted of two flares per mission. Each flare dispersed approximately 19.7 g of AgI over a nominal 3.5-min seeding period, for a total of approximately 374.3 g of AgI. The first three missions were conducted near 3.57–3.66 km MSL, after which release altitudes increased substantially: Missions 4–6 were conducted near 4.21–4.27 km MSL. Mission 7 spanned a somewhat broader range, with its two releases occurring near 3.78 and 4.18 km MSL.

Thermodynamic conditions at release altitude were favorable for AgI activation throughout the sequence, with all median release temperatures falling within the approximate -5 to -15°C activation window. The first three missions occurred in the warmer portion of this range. EL-153 sampled median temperatures near -7°C during these missions, while EL-151 encountered colder conditions near -9 to -10°C. The increase in seeding altitude beginning with Mission 4 placed subsequent flare releases in substantially colder conditions and closer to the more favorable end of the AgI activation range. Median temperatures during Missions 4–6 were approximately -13 to -13.6°C, approaching the -15°C region where AgI ice-nucleating activity is optimal. Mission 7 remained favorable but was somewhat more variable because of the difference in release altitude, with median temperatures of approximately -9.9°C at 3.78 km MSL and -13.0°C at 4.18 km MSL. Thus, although all seven missions occurred within the nominal AgI activation window, the later, higher-altitude missions—particularly Missions 4–6—were conducted under colder and more favorable activation conditions.

Relative humidity was also consistently high during the releases, although systematic differences were evident between the two drone locations during the first three missions. EL-151 at seeding location AK36 sampled median relative humidity (RH) values of approximately 94%, compared with roughly 80–83% for EL-153 at seeding location AK52. Once operations shifted exclusively to EL-151 for Missions 4–7, median RH remained consistently high at approximately 92–93%. Taken together, the temperature and humidity observations indicate that the seeded air was generally favorable for AgI activation throughout the experiment, with conditions becoming particularly favorable during the higher-altitude releases later in the sequence.

Table 1. Information on the individual flare releases and drone telemetry + pressure, temperature, and humidity data from the seven coordinated missions during this IOP.

Validation Methodology

Seeding Signature Identification

Our validation framework is designed to identify “seeding signatures,” or visible patterns on radar which result from cloud seeding operations and are distinguishable from natural background precipitation. Historically, this type of attribution has been challenging because operations take place in evolving natural clouds, rather than in isolation. Rainmaker’s approach builds on methodology developed in the National Center for Atmospheric Research’s 2017SNOWIE campaign, which combined radar, aircraft, and environmental observations to trace the physical formation and evolution of seeded precipitation (French et al. 2018; Tessendorf et al. 2019; Friedrich et al. 2021; Zaremba et al. 2024). This “physical validation” approach enables near real-time identification of seeding signatures and quantitative precipitation estimation (QPE).

For each seeding mission, the analysis begins with the precise location, altitude, and time of AgI release by the Elijah drone. Wind observations from drone telemetry and, where available, nearby radiosondes are used to project an expected downwind region in which a seeding plume could develop. This trajectory is treated as a search corridor to identify an initial reflectivity enhancement, but it is not a deterministic plume path. Seeded ice particles are subject to vertical wind shear, particle growth, terrain-induced atmospheric motions, and other factors influencing the trajectory of falling hydrometeors. Thus, the observed radar signature could depart from a simple horizontal extrapolation of the wind at seeding altitude.

Quality-controlled radar volumes are examined before, during, and after each release to determine whether a new reflectivity enhancement develops within this physically plausible region. The WSR-88D (PAHG NEXRAD) provides the principal observational basis for this analysis, with quality control used to minimize contamination from terrain and other nonmeteorological echoes. Radar detection is particularly useful because AgI-induced ice growth can produce particles sufficiently large to generate a coherent reflectivity enhancement, as demonstrated in previous seeding experiments (French et al. 2018; Friedrich et al. 2021; Zaremba et al. 2024).

We note that timing and location alone are not sufficient for attribution. Candidate seeding signatures must also exhibit physically consistent motion, vertical development, lifecycle, and, most importantly, repeatability within the same atmospheric conditions. Features should generally advect with the ambient wind rather than remain terrain-locked, and they should develop vertically in a manner consistent with seeded ice growth and sedimentation. Coherent growth and persistence across successive radar volumes provide stronger evidence than isolated reflectivity enhancements, particularly in weak precipitation environments.

Repeatability across separate releases within the same operation further strengthens attribution when signatures occur with similar timing, motion, and morphology. Environmental and UAS observations provide supporting context, including release temperatures favorable for AgI activation (Vonnegut 1947; DeMott 1995; Marcolli et al. 2016; Miller et al. 2025), evidence of supercooled liquid water from icing on the drone, and satellite-observed cloud changes where available(Henneberger et al. 2023; Ramelli et al. 2024; Omanovic et al. 2026). However, these observations support, but do not replace, radar evidence.

Once attributed to seeding, each signature is manually tracked using polygons drawn independently for each radar volume and elevation angle, preserving its observed three-dimensional structure. Tracking begins at first confident detection and continues until the feature dissipates, exits useful radar coverage, merges with other precipitation, or can no longer be reliably isolated. These polygons are then used to calculate seeding signature characteristics as well as quantitative precipitation estimation (QPE), which is detailed below.

Quantitative Precipitation Estimation (QPE)

Liquid-equivalent precipitation associated with each seeding signature is estimated using a radar-based QPE workflow. A composite reflectivity field is constructed using the lowest usable filtered reflectivity radar gate at each location, excluding beams with substantial terrain blockage. Because the lowest valid elevation varies with range and azimuth, signature polygons for QPE specifically are redrawn on the composite reflectivity field.

Reflectivity is converted to liquid-equivalent snowfall rate using 27 relationships of the form

Z = aSᵇ

where a is varied between 100 and 500 every 50 and b from 2.0 to 2.2 every 0.1. This ensemble approach is informed from Friedrich et al. (2020), and spans a range of published snow relationships (Matrosov et al. 2009; Wolfe and Snider 2012), accounting for variability in particle size, habit, density, and fall characteristics. Each relationship is applied through the full accumulation period, and precipitation is integrated within the evolving signature footprint to produce an ensemble of precipitation values.

The ensemble spread reflects uncertainty associated with the choice of Z-S relationship, but does not represent total QPE uncertainty. Additional uncertainties include radar calibration and beam geometry, vertical hydrometeor evolution (e.g., sublimation or evaporation below the radar beam), polygon placement, and possible contamination from weak natural precipitation. The latter is mitigated by allowing a subtraction of a baseline reflectivity value from the entire reflectivity field before QPE calculations, the choice of which for this is detailed in the results section further below.

Validation Results

Radar-Based Seeding Signatures

Figure 5 shows an animation of the 0.88° tilt from PAHG along with the manually-drawn polygons isolating the signatures. Signature A is attributable to coordinated mission 1 in Table 1, flown by drones EL-151 and EL-153, during which four red flares were released between 06:38:18 and 06:38:37 UTC. The resulting signature was first observed at 07:18:40 UTC, 40.4 min after release, and remained detectable through 08:14:45 UTC, corresponding to 56.1 min across 9 radar volumes. It was observed at three tilts (0.48°–1.32°), spanning 1.76–3.67 km MSL at ranges of 108.1–123 km from PAHG. The radar-derived vertical extent should be interpreted as a minimum estimate of the true vertical extent. The signature top likely fell between the 1.32° and 1.8° tilts, while its lower boundary likely extended below the lowest available 0.48° tilt and was therefore not resolved by the radar observations. The signature reached a maximum observed extent of 16.4 km2 at 07:53:43 UTC. Its disappearance from radar does not necessarily indicate dissipation, as low-level radar coverage progressively diminished as the signature, and those that followed, advected farther downwind from PAHG.

Figure 5. Filtered reflectivity from the PAHG radar at the 0.88° tilt, with in-flight drone locations and MSL altitudes annotated. The seven time-relevant seeding-signature polygons are labeled Signatures A–G.

Signature B followed coordinated mission 23 (flares 51–54), again flown by EL-151 and EL-153, with four red flares released between 07:06:54 and 07:07:10 UTC. It was first detected at 07:39:42 UTC, 32.8 min after release, and persisted in the radar observations until 09:17:50 UTC, yielding the longest observed duration of the seven signatures: 98.1 min across 15 volumes. B was observed at three tilts (0.48°–1.32°), spanning 1.70–3.83 km MSL and ranges of 105–133.5 km. Its maximum observed extent of 41.2 km2 occurred at 08:35:46 UTC, the largest of the seven signatures.

Signature C was associated with coordinated mission 24 (flares 55–57), during which EL-151 and EL-153 released three red flares between 07:36:48 and 07:36:57 UTC. First detected at 08:07:44 UTC, 30.9 min after release, C remained observable through 09:31:52 UTC, a period of 84.1 min across 13 volumes. It was detected at three tilts (0.48°–1.32°), spanning 1.72–4.00 km MSL at ranges of 105.1–132.3 km. Its maximum observed extent was 39.8 km2 at 09:17:50 UTC, second only to B.

The character of the observations changed somewhat with Signature D, associated with coordinated mission 25 (flares 58 and 59). EL-151 released two red flares between 08:19:46 and 08:19:51 UTC, and the resulting signature was detected 23.0 min later, at 08:42:47 UTC. D remained observable until 10:13:56 UTC, spanning 91.2 min and 14 volumes. Unlike A–C, it was detected at all four available tilts (0.48°–1.8°), extending from 1.65 to 4.36 km MSL at ranges of 102.4–128.8 km. This was the greatest vertical extent observed among the seven signatures. Its maximum horizontal extent, 35.4 km², occurred at 09:45:53 UTC.

Signature E, associated with coordinated mission 26 (flares 60 and 61), followed a similar pattern. EL-151 released two red flares between 08:43:25 and 08:43:30 UTC, with the signature first appearing at 09:03:49 UTC, only 20.4 min later. It remained detectable through 10:20:57 UTC, spanning 77.1 min across 12 volumes. As with D, E was observed at all four tilts (0.48°–1.8°), spanning 1.66–4.35 km MSL at ranges of 102.1–124.8 km. Its maximum observed extent of 30.4 km² occurred at 10:13:56 UTC, near the end of its observed lifetime.

Signature F was associated with coordinated mission 27 (flares 62 and 63), flown by EL-151, with two red flares released between 09:07:24 and 09:10:03 UTC. F was first detected at 09:24:51 UTC, 17.4 min after release, representing the shortest release-to-detection interval of the seven signatures. It remained observable through 10:27:57 UTC, spanning 63.1 min across 10 volumes. Like D and E, F was detected at all four tilts (0.48°–1.8°), occupying heights of 1.67–4.21 km MSL at ranges of 100.5–120 km. Its maximum observed extent was 16.9 km² at 10:06:56 UTC.

Finally, Signature G resulted from coordinated mission 28 (flares 64 and 65), during which EL-151 released two red flares between 09:29:52 and 09:33:41 UTC. The signature was first observed at 09:59:55 UTC, 30.1 min after release, and remained detectable until 10:41:57 UTC, giving an observed duration of 42.0 min across 7 volumes. In contrast to D–F, G was detected at only three tilts (0.48°–1.32°), spanning 1.74–3.51 km MSL at ranges of 105.8–117.9 km. Its maximum observed extent was 9.4 km² at 10:27:57 UTC, the smallest of the seven signatures.

Taken together, the seven signatures show systematic differences in both detection time and observed spatial evolution. The interval between flare release and initial radar detection decreased from 40.4 min for A to 17.4 min for F before increasing to 30.1 min for G. At least part of this progression appears related to the surrounding precipitation environment: the earlier signatures were embedded within somewhat heavier natural precipitation, making their initial identification more difficult. The largest horizontal extents occurred for B and C, which reached 41.2 and 39.8 km², respectively, whereas A, F, and especially G remained considerably smaller. A corresponding difference is apparent in the vertical observations. D, E, and F were the only signatures detected at all four available tilts and reached tops above 4.2 km MSL, while A and G remained below 3.7 km MSL. Importantly, the observed lifetimes and maximum extents should not necessarily be interpreted as measures of the signatures’ complete physical lifetimes or ultimate sizes. Low-level radar coverage decreased as the signatures advected farther from PAHG, and E, F, and G reached their maximum observed extents at or near the end of their respective observational records. Their subsequent evolution therefore cannot be determined from the available radar observations.

Quantitative Precipitation Estimation (QPE)

Precipitation was estimated over the period from 06:21:21 to 10:55:58 UTC on 23 August 2026, encompassing 40 radar volumes and 4.69 h of integration time. Precipitation rates were derived using a 27-member Z–S ensemble of the form

Ze = aSᵇ

where a ranged from 100 to 500 in increments of 50 and b ranged from 2.0 to 2.2 in increments of 0.1. The range of ensemble estimates represents uncertainty associated with ice-crystal characteristics and the subsequent conversion of ice-equivalent radar reflectivity to liquid-water precipitation (Friedrich et al. 2020). Importantly, this range does not account for uncertainty associated with the evolution of precipitation between the lowest radar beam (~1800–2200 m MSL) and the surface.

For each radar column, Ze was obtained from composite reflectivity using the lowest unblocked radar gate, defined here as having less than 30% beam blockage. Because non-negligible natural precipitation surrounded the seeding signatures, 5 dBZ was subtracted from all radar gates prior to conversion to liquid-equivalent precipitation using the Z–S relationship. Applying this 5-dBZ adjustment within the seeding signatures serves two purposes. First, it provides an estimate of the precipitation enhancement above the surrounding natural precipitation that would otherwise have occurred in the absence of seeding. Second, it reduces sensitivity to spatial uncertainty in the manually delineated signature polygons. If surrounding natural precipitation was inadvertently included within a polygon, subtracting 5 dBZ, representative of the background precipitation intensity, minimizes its contribution to the resulting QPE.

Although precipitation reached the surface as rain, with surface temperatures near +10°C, the lowest radar beam (0.48°) was approximately 1,800 m MSL at the seeding location and increased in altitude as the signatures advected SSW away from PAHG. Consequently, during periods when the seeding signatures were producing precipitation, the lowest radar observations generally sampled hydrometeors above the freezing level and therefore primarily represented frozen precipitation. This supports the use of the same Z–S coefficients employed in the SNOWIE precipitation quantification (Friedrich et al. 2020). Although not shown here, this interpretation is also consistent with the NEXRAD hydrometeor classification algorithm, which predominantly classified hydrometeors within the seeding signatures as ice crystals, dry snow, or wet snow. We also note from soundings and ceilometer data that the atmosphere from seeding level down to the surface was naturally saturated and precipitating, meaning that while uncertainties due to evaporation / sublimation between the surface and the lowest beam are still present, large overestimations of precipitation production due to this uncertainty are highly unlikely.

Across the 27-member QPE ensemble, the mean precipitation volume attributed to seeding was 57.62 acre-feet (71,072 m³), with a 5th–95th percentile range of 41.70–89.35 acre-feet. The 25th, 50th, and 75th percentiles were 45.69, 52.97, and 64.79 acre-feet, respectively. Signatures B and C, both of which persisted for more than an hour, each contributed a mean QPE exceeding 14 acre-feet and together accounted for more than 50% of the total mean QPE across all seven signatures (Table 2; Figure 6). Following B, the mean precipitation contribution per signature decreased monotonically through the remainder of the sequence.

Table 2. QPE characteristics of each observed seeding signature.

Figure 6. Cumulative QPE of each observed seeding signature over time and summed.

The four-panel QPE loop further illustrates the spatial evolution of the seeded precipitation (Figure 7). The signatures repeatedly trained over one another in relatively quick succession, while each subsequent signature advected slightly more southward relative to westward than its predecessor. The resulting wet area, defined by accumulated seeded precipitation exceeding 0.01 mm, covered 249.53 km². Within this footprint, mean ensemble QPE indicates that the core of the affected area received more than 1.0 mm of precipitation attributable to seeding.

Figure 7. (A) Composite reflectivity, (B) instantaneous precipitation rate, (C) total accumulated precipitation, and (D) time series of polygon-integrated precipitation volume and spatial area exceeding 0.01 mm. Brown shading in (D) represents ensemble uncertainty, while vertical gray shading indicates active seeding periods.

References

DeMott, P. J., 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.

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.

Friedrich, K., and Coauthors, 2021: Microphysical characteristics and evolution of seeded orographic clouds. J. Appl. Meteor. Climatol., 60, 909–934, https://doi.org/10.1175/JAMC-D-20-0206.1.

Geerts, B., and R. M. Rauber, 2022: Glaciogenic Seeding of Cold-Season Orographic Clouds to Enhance Precipitation: Status and Prospects. Bull. Amer. Meteor. Soc., 103, E2302–E2314, https://doi.org/10.1175/BAMS-D-21-0279.1.

Henneberger, J., and Coauthors, 2023: Seeding of supercooled low stratus clouds with a UAV to study microphysical ice processes: An introduction to the CLOUDLAB project. Bull. Amer. Meteor. Soc., 104, E1962–E1979, https://doi.org/10.1175/BAMS-D-22-0178.1.

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.

Matrosov, S. Y., C. Campbell, D. Kingsmill, and E. Sukovich, 2009: Assessing snowfall rates from X-band radar reflectivity measurements. J. Atmos. Oceanic Technol., 26, 2324–2339, https://doi.org/10.1175/2009JTECHA1238.1.

Miller, A. J., and Coauthors, 2025: Quantified ice-nucleating ability of AgI-containing seeding particles in natural clouds. Atmos. Chem. Phys., 25, 5387–5407, https://doi.org/10.5194/acp-25-5387-2025.

Omanovic, N., D. Bötticher, C. Fuchs, and U. Lohmann, 2026: Glaciogenic seeding-induced hole-punch clouds and their sensitivity to the clouds’ background state. Atmos. Chem. Phys., 26, 5345–5353, https://doi.org/10.5194/acp-26-5345-2026.

Ramelli, F., and Coauthors, 2024: Repurposing weather modification for cloud research showcased by ice crystal growth. PNAS Nexus, 3, pgae402, https://doi.org/10.1093/pnasnexus/pgae402.

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.

Tessendorf, S. A., and Coauthors, 2019: A transformational approach to winter orographic weather modification research: The SNOWIE Project. Bull. Amer. Meteor. Soc., 100, 71–92, https://doi.org/10.1175/BAMS-D-17-0152.1.

Vonnegut, B., 1947: The nucleation of ice formation by silver iodide. J. Appl. Phys., 18, 593–595, https://doi.org/10.1063/1.1697813.

Wolfe, J. P., and J. R. Snider, 2012: A relationship between reflectivity and snow rate for a high-altitude S-band radar. J. Appl. Meteor. Climatol., 51, 1111–1128, https://doi.org/10.1175/JAMC-D-11-0112.1.

Zaremba, T. J., R. M. Rauber, L. Di Girolamo, J. R. Loveridge, and G. M. McFarquhar, 2024: On the radar detection of cloud seeding effects in wintertime orographic cloud systems. J. Appl. Meteor. Climatol., 63, 27–45, https://doi.org/10.1175/JAMC-D-22-0154.1.