Providing charge emission for cloud seeding aircraft

In water-stressed regions, cloud seeding is sometimes used to encourage or enhance rainfall.¹ Conventional seeding materials include salt or silver iodide, released from an aircraft with an underwing burning flare, ignited in flight. Our recent work demonstrates that modifying or generating the electric charge on water drops in clouds provides a possible alternative methodology, as charge can encourage droplet coalescence, expediting raindrop growth.² Droplet size spectra may be influenced more generally,^3,4 as charge on droplets reduces their evaporation rates, and, through Rayleigh explosions,⁵ may even lead to droplet disintegration.

Air in the lower atmosphere contains positive and negative cluster ions at concentrations of about 100 to 1000 cm^{−3 6} at the surface, generated by radioactivity and cosmic rays. At flight altitudes of 3–5 km, where the loss of ions by aerosol collection is less than that at the surface but at which height cosmic ray ionization has not begun to significantly increase,⁷ the ion concentrations become 1500–2000 cm⁻³.⁸ Some charging of atmospheric aerosols and any water droplets present will result from attachment of the ions, in a system that resembles weak dusty plasma. By introducing further ions artificially, a new route to influence the behavior of natural droplet systems becomes possible. The charge release can be achieved through corona emission at atmospheric pressure, using a high voltage source with suitable emitter electrodes, deployed either at the surface⁹ or aloft.¹⁰ Release at altitude allows the critical regions of developing clouds to be targeted.

Testing the effectiveness of these approaches requires a suitable airborne platform to provide access to developing clouds. A first step is to establish reliable and controlled charge emission at altitude from such a platform. The use of a crewed aircraft is evaluated here, using a Beechcraft King Air C90 seeding aircraft, operated by the National Center for Meteorology (NCM) in the United Arab Emirates (UAE). Cloud seeding aircraft such as this are typically modified to carry a set of flares in racks below their wings, Fig. 1(a). The conventional flares release material after a brief (timescale ∼1 min) period of combustion, electrically ignited by the pilot from the aircraft’s cockpit. No meteorological information is recorded by these aircraft.

In this paper, devices with a physical form compatible with the conventional flare mounting rack are demonstrated to provide a straightforward method for charge emission in flight. Section II outlines the concept of the devices, and Sec. III summarizes their deployment in a series of test flights. Section IV provides case study data from a clear air test flight. The results and applications are discussed further in Secs. V and VI.

The existing mounting tubes for the conventional cloud-seeding flares provide a volume that can be repurposed for a charge emission device, removing the need for further structural change to the cloud-seeding aircraft. A charge releasing flare—or “flare emitter”—has been developed¹¹ to occupy the same space, and physically compatible with the fixings used for conventional flares [Fig. 1(b)]. Each flare tube position includes a single electrical connection to the pilot’s cockpit controls. This is used to ignite the conventional flare by supplying a short pulse of current to heat a fine wire within the flare. This same connection and initiation pulse are used to trigger the flare emitter.

As the flare emitter is expected to be deployed by individuals whose primary skillset is associated with conventional flares, they are designed to be stand-alone devices, able to be operated without training. To provide data with which they can be subsequently evaluated, the flare emitters contain their own basic meteorological sensors and data logging to record operating parameters and conditions.

The flare emitter operates by corona emission, delivering a regulated high voltage (∼3 kV) to a carbon brush electrode, which is found empirically to be sufficient in the laboratory at the surface. As the breakdown electric field for air decreases with pressure,¹² the voltage required will be reduced. This high voltage, selectable for either polarity, is generated from internal batteries using a compact 6 V mains transformer in reverse and twelve stages of Cockroft–Walton voltage multipliers (Fig. 2). Electronic component choices in the design phase were restricted to standard parts due to pandemic-related supply problems.

Figure 2(a) provides a block diagram of the systems within the flare emitter. An Arduino Nano is programmed to control the system and record and timestamp internal pressure, temperature, and relative humidity (“PTU”) measurements, with the battery voltage, emitter voltage, and electrode current. The data values obtained are written to an internal microSD card for subsequent analysis. Emission current monitoring is achieved by using an infra-red LED as a current sensor. This provides optical isolation and, through an analog feedback method, compensation for non-linearities and temperature variations.¹³  Figure 2(b) shows the physical arrangement.

Flights carrying the flare emitters on a King Air aircraft were undertaken in August and September 2023, from Al Ain airport in the UAE (24.25°N, 55.61°E). Positive and negative flare emitters were mounted in the flare tubes that were nearest to the aircraft fuselage, alongside conventional flares, and grounded to the aircraft chassis. To operate the flare emitter, a trigger pulse from the cockpit selector switch initiated the microcontroller. As this pulse was generated in the cockpit with a direct connection to the aircraft systems, the signal was optically isolated at the flare emitter for safety and to reduce noise carried back to the cockpit. After the trigger pulse was received, the flare emitter operated for 120 s, with the high voltage activated for the first 60 s only. During the 120 s of operation, data from the sensors and internal monitoring were recorded on the microSD card at a 10 Hz sampling rate.

During the campaign, the flare emitters were operated on the runway and in flight. Runway activations were made on six different days (on 14, 28, and 31 Aug and 3, 4, and 5 Sept) and in-flight activations on two flights (28 Aug and 16 Sept). Emitter voltages recorded from the negative emitter on these occasions are summarized in Fig. 3: the positive flare emitter voltage was unfortunately insufficient to generate reliable corona and was not considered further. Figure 3(a) shows that, despite the regulation system implemented in the circuitry, the negative emitter voltage varied. This is investigated further in Fig. 3(b). Emitter voltages are plotted against the battery voltage, separately for those obtained in flight and those obtained at the surface.

An effect of declining battery voltage on emitter voltage is apparent in both sets of data. If, as is evident from Fig. 3(b), the emitter voltage data are restricted to those when the battery voltage was 6 V or greater, the emitter voltages are similar in both the aloft and surface cases (see also Table I). Regulation of the emitter voltage therefore becomes less effective as the battery ages. This is because the regulation works by temporarily deactivating the oscillator stage whenever the emitter voltage is above a prescribed operating value, which reduces power consumption and enhances reliability by reducing voltage stress on components.¹⁴ If this operating value is not reached due to a weak battery, the regulation ceases and the emitter voltage will decline as the battery voltage declines. This indicates a need for additional multiplier stages in a revised design.

The detailed operation of the flare emitter during a flight was investigated during the flight on 16th September. This day was chosen for clear air conditions to remove complicating effects due to cloud. The surface relative humidity at 06 UTC was <45% and decreased with height (see also Fig. S1).

The flight departed from Al Ain at about 05 UTC and traveled southwest to a holding position over the southeast of the UAE, where several loops of level flight at 3 km altitude were made [Fig. 4(a)]. The flare emitters were operated intermittently during this period of level flight [Fig. 4(b)], with some variation in air speed [Fig. 4(c)]. Flare emitter operations during the flight were identified from samples of the internal Real Time Clock in the data file generated. The aircraft returned to Al Ain, landing at 0630 UTC.

Each operation of the flare emitter during the flight lasted 60 s, with data recorded for a further 60 s after which the system went back into standby mode. Figure 5 summarizes the data obtained during the seven in-flight runs of the flare emitter on 16th Sep, 2023. Figure 5(a) shows that the battery supply was initially above 6 V, and decreased by almost 1 V through the series of activations. Figure 5(b) shows that the electrode voltage was initially steady but that it decreases proportionally with the declining battery voltage. Figure 5(c) shows that a median emitter current of >3 µA was sustained, also declining with the deterioration in the battery voltage.

More details of the flare emitter operation for the full 120 s of run 5 is provided in Fig. 6. Figure 6(a) shows the rapid onset and switch-off of the high voltage on the electrode during the first 60 s. It is clear that there is variability in the high voltage during the operating time. There is a decline in the mean voltage of about 9%, which is likely to be associated with the regulation matters mentioned earlier due to the driver oscillator operating continuously and draining the battery. Figure 6(b) shows the measured electrode current during the same interval. As for the electrode voltage, there is variability, with a central tendency of ∼3 µA, as also evident from the median value for run 5 given in Fig. 5(c). Figure 6(c) extracts the voltage–current relationship, from the decay in emitter voltage at the end of the operating time. Although these data points will have been smoothed by the measurement system, this shows the transition for 1 kV or more on the electrode, to an emission current plateau at ∼3 µA.

In Fig. 7(a), histogram of the emitter currents from run 5 is given, which emphasizes that a modal value is well defined. Figure 7(b) provides a spectrum derived from the same data, showing that the variability is essentially white across the range of frequencies available with 10 Hz sampling.

Monitoring of the emitter voltage in flight demonstrates that the emitter current is highly variable. This is associated with the corona discharge process and ion emission, rather than the measuring system, as it is absent when the high voltage is switched off. Although there is variability in the high voltage supply, variability is a fundamental expectation of corona discharge in general,^15,16 and it is also known to be affected by air speed,¹⁷ which in this case is substantial (∼100 ms⁻¹).

Despite the observed variability, a central tendency of emitter current from Figs. 5(c) and 6(b) is apparent and ∼3 µA. Assuming emission occurred into 1 m², the charge released by a single flare at 120 ms⁻¹ would be 1.5 × 10¹¹e m⁻³. This can be compared with atmospheric ion concentrations at the flight level of ∼10⁹ m⁻³.⁸ Hence, the artificial introduction of ions is likely to lead to ion concentrations about two orders of magnitude greater than the natural background.

This series of experiments demonstrates that the flare emitter concept provides a device that can release charge under pilot control during a flight and that is straightforwardly fitted to an existing cloud seeding aircraft. Furthermore, the devices provide a quantitative indication of the charge emitted. By deploying a set of multiple flare emitters on a single aircraft, ion concentrations that are considerably enhanced over the natural background can be readily delivered into developing clouds in a controlled way.

Co-located meteorological data for the case study flight is provided as an atmospheric profile in the supplementary material.

This material is based on work supported by the National Center of Meteorology, Abu Dhabi, UAE, under the UAE Research Program for Rain Enhancement Science (UAEREP). The highlight image obtained in flight and Fig. 1(b) were provided by NCM.

The authors have no conflicts to disclose.

R. Giles Harrison: Conceptualization (equal); Investigation (equal); Methodology (equal); Supervision (equal); Writing – original draft (equal); Writing – review & editing (equal). Ahmad A. Alkamali: Data curation (equal); Investigation (equal); Resources (equal); Writing – review & editing (equal). Veronica Escobar-Ruiz: Data curation (equal); Resources (equal); Software (equal); Writing – review & editing (equal). Keri A. Nicoll: Conceptualization (equal); Funding acquisition (equal); Methodology (equal); Supervision (equal); Writing – review & editing (equal). Maarten H. P. Ambaum: Funding acquisition (equal); Project administration (equal); Supervision (equal); Writing – review & editing (equal).

Data is available from the University of Reading data archive,¹⁸ at https://doi.org/10.17864/1947.001337. The co-located meteorological data was obtained from Copernicus Climate Change Service (C3S), as ERA5 hourly data on single levels from 1940 to present. Climate Data Store (CDS), DOI: 10.24381/cds.adbb2d47 (Accessed 20th Feb 2024).