A Novel Drone Sampling Method for Lower Atmospheric Fungal Spores

Abstract

Novel and practical methods are always sought across all disciplines; within bioaerosol research, portable, lightweight, and low-cost sampling pumps are few and far between. Fungal spores, key components of bioaerosols, have attracted attention due to their negative effects on human populations, agricultural systems, and ubiquitous nature. In terms of spatial scales, fungal spores across vertical gradients are frequently overlooked and in cases where atmospheric samples are collected, they are often a large distance away from the ground, occurring hundreds or thousands of meters into the atmosphere, which also requires substantial expenses for specialist apparatus. Here, we have utilized a drone and low-cost equipment to produce a new sampling method that can efficiently collect fungal spores and bridge the gap between ground sampling and atmospheric sampling, and sample in areas such as forest canopies or at building rooftop heights, in which planes, helicopters, or other UAVs may not be able to safely or practically maneuver. Additionally, we have created a novel approach to utilizing a drone for bioaerosol sampling during rain events, which, to our knowledge, is the first of its kind, opening up the possibilities for much needed comparisons of fungal spores in varying weather conditions.

1. Introduction

Bioaerosols are organic particles that are present within the atmosphere and have gained much interest in recent decades [1,2]; an important component of bioaerosols are fungal spores [3]. Traditionally, fungal spore research focused on allergenic fungi and plant pathogens of agricultural crops [4,5] due to their negative influences on human populations. More recently, fungal spores have been found to play important roles in ice nucleation, a significant process early on in cloud formation [6]. Typically, fixed positioned, large and heavy equipment has been implemented for bioaerosol sampling, such as Hirst volumetric samplers, although these methods are informative and beneficial, they can lack versatility in terms of portability and/or economically costly. Moreover, complications arise when aiming to distinguish clear distinctions of bioaerosol microbial communities between varying heights. It is known that differences in height can result in variable fungal spore communities but currently the comparisons are either between ground level and building rooftops or ground level to hundreds/thousands of meters into the air [7,8]. Rooftops were designated as proxies for “higher” samples [9,10] but the roofs themselves are liable to contain microbes, which may not be free floating and found within the air, so sampling free from surfaces is much sought after. Previously, researchers have effectively sampled lower atmospheric areas, i.e., below ~1000 m from ground level, with the use of helicopters, planes, and balloons [7,11,12]; however, the cost for such equipment severely limits the number of researchers that are capable of utilizing them, as well as regulations which can restrict minimum flying heights due to obvious dangers such as the risk of collisions or practically being incapable of flying within enclosed areas such as forest/woodland or urban sites. Therefore, there is a clear gap in vertical sampling for fungal spore communities between the ground and lower atmosphere and a promising way of sampling at lower atmospheric altitudes with easy maneuverability lies with drones.

Drone technology or Unmanned Aerial Vehicles (UAVs) are modifiable and capable of maneuvering into difficult to reach areas. Drones are utilized for multiple tasks including habitat assessment, mapping, insect sampling [13], and microbial collection from both the air and water [14]. The modifiability of drones to employ various sensors controlled by Arduino technology to analyze atmospheric aerosols was also explored with insightful impacts, clearly demonstrating the utility and versatility of employing drones in aerosol research [15]. Drones were also applied when trying to capture bioaerosols with ice-nucleating abilities to good effect; Bieber et al. [16] were able to modify DJI drones with an impingement device where they were able to capture bioaerosols from multiple heights into a liquid medium, which were then analyzed for ice nucleation ability. Despite their advantages, drone usage in an ecological context remains underutilized. In terms of drone usage for fungal spore collection, early research focused upon basic passive methods, implementing agar plates attached to wings of an UAV [17]. Subsequent colony forming units (CFUs) were then counted and quantified, but commonly it has not been possible to measure the volume of filtered air, in addition to only a small percentage of fungi being culturable, thus important information has been missed.

Furthermore, many UAVs are incapable of flying during wet weather; thus, current data regarding bioaerosols captured by UAVs is largely biased toward dry weather [11]. In the context of fungal spores, this sampling bias is problematic as meteorological factors such as precipitation and relative humidity are often noted as being important drivers of fungal spore release for a range of taxa, typically due to the turgor pressure required for spore release [18]. The physical impact of raindrops can cause what is known as “Splash dispersal” [19], which promotes the release of spores, as seen in a range of fungi such as Puffballs Lycoperdon [20]. At higher atmospheric heights, rainfall may also have a “washing effect” in which spores are captured in falling raindrops and brought to the ground surface, decreasing atmospheric spore abundance [12,21]. In various environments, rainfall can be unpredictable and vary in terms of intensity and duration; drones provide the opportunity of being able to be set up and ready for atmospheric sampling in a matter of minutes. The ability to move away from roof sampling as proxies for “higher atmospheric samples” may also prove beneficial as an air sample away from surfaces can be collected, potentially leading to a less biased assumption of fungal flora at specific heights. Additionally, advances in drone technology are producing possibilities of novel sampling methods at affordable prices, which makes them easily accessible. The aim of our research was to compare the collection efficiency of the proposed drone method to a standard aerosol sampling pump to provide an affordable, lightweight, portable, and practical alternative to current bioaerosol sampling methods, highlighting the applicability of drones in an ecological context, and to produce an effective drone sampling method during rainy periods.

2. Materials and Methods

2.1. Pump Design and Mechanism

A novel bioaerosol sampling device, targeting fungal spores, was created, comprising readily available, low-cost equipment including a programmable Arduino Nano and Arduino UNO R3 (open-source hardware, Arduino, Monza, Italy) to automatically control the on/off function of a Flextailgear Tiny Pump (Changzhou Flextail Technology, Co., Ltd., Changzhou, Jiangsu, China). Overall, 6 Flextailgear Tiny Pumps were used, each with a maximum battery runtime of approximately 15 min. The structure of the components is outlined in Figure 1. A 25 mm diameter glass fiber filter, GF/D, with a pore size of 2.7 μm (Whatman, Co., Ltd., Maidstone, UK) was used as the collection medium. The mesh structure and pore size allowed for the efficient collection of a range of fungal spore sizes as well as maintaining an effective flow rate. Two different polypropylene filter holders were used, Swinnex (Sigma-Aldrich, St. Louis, MO, USA) and Cole-Parmer (Vernon Hills, IL, USA). Despite both filters having a diameter of 25 mm, the outlet diameter differed (Swinnex—1.5 mm, Cole Parmer—2 mm), resulting in differing flow rates for each filter holder. Flow rates of the drone pump were measured using a flow meter (TSI 4100 series, TSI Incorporated, Shoreview, MN, USA) before and after each flight and averaged, ranging from 7.1 L/min to 5.6 L/min before flights and 6.7 L/min to 5.2 L/min after flights using a Swinnex filter. For Cole-Parmer filter holders, the flow rate before flights ranged from 5.2 L/min to 4.3 L/min and 4.8 L/min to 3.9 L/min after flights. The apparent variability in flow rate is likely due to the slight differences in inlet size between the filter holders and variation in suction strength between pumps. New GF filters were always utilized for checking flow rates. Please refer to the Supplementary Materials to see the weights and prices of each component for the proposed sampling design.

The Arduinos were programmed using open-source software (Arduino IDE 1.8.19.), causing the SG90 servo to press the on button of the pump after 1 min and then switch off the pump after 10 min, with a total running time of 11 min. The one-minute delay allowed sufficient time for the drone to be flown to the desired heights for sampling. Arduinos were chosen due to their extremely lightweight, low-cost, simple programming features, and ease of use when attaching components. The requirement of a timer function with the ability to control a small SG90 servo indicated that the Arduino was adequate for this purpose. Please refer to the Supplementary Materials to see the input code for the Arduinos.

2.2. Drones

Two drones were used: a DJI Phantom 3 Standard (DJI JAPAN, Co., Ltd. Tokyo, Japan) during dry weather and a Poweregg X (PowerVision Inc., Beijing, China) for samples collected during rainfall. Both were modified to allow for the attachment of the pump; cable ties and cotton string were used to firmly hang the pump from the drone body (Figure 2a,b). Both drones were registered and adhered to national drone laws.

The DJI GO application was installed onto an android device to view live video from the drone whilst flying and also utilizing the app’s flight information (i.e., drone height and distance from take-off, battery consumption). For the Poweregg X, the PowerVision app was also installed. Notably, Poweregg X has a waterproof case, which can be equipped to the body to allow for flying during rainfall. Both drones were fitted with control filter holders during each flight, and the controls consisted of a GF filter placed within a filter holder hanging from the body of the drone. The control was to test whether flying to and from the desired height would cause spores to be inadvertently collected.

2.3. Study Site and Sampling

The locations of drone sampling were predominantly based within Kakuma Campus, Kanazawa University, Japan, as well as a coastal area (Figure 3a,b &

Table 1. Locations of sampling sites and land type and dates of sampling.

Site	Coordinates	Type	Sampling Date	Heights Sampled (m)	Elevation (Meters Above Sea Level)
KU Kakuma Campus Site 1 (Bamboo)	N36° 32.822′ E136° 42.173′	Forest	April 2024–December 2024	Ground, ~10 m, 50 m	127
KU KC Site 2 (MD Forest)	N36° 32.858′ E136° 42.050′	Forest	April 2024–November 2024	Ground, ~8 m, 50 m	144
KU KC Site 3 (Open)	N36° 33.013′ E136° 41.903′	Forest	May 2024– October 2024	Ground, ~11 m, 50 m	132
KU KC Building 1	N36° 32.678′ E136° 42.277′	Urban	July 2024–October 2024	Roof	148
Kanazawa, Coastal area	N 36° 37.106′ E136° 36.316′	Coastal, urban	19 May 2024	Ground 20 m, 50 m	8

). The forested area of Kakuma Campus is host to a range of plants, bamboo, and trees, including Sasa (Sasa palmate), Japanese Cedar (Cryptomeria japonica), Oak (Quercus serrata and Q. variabilis), and various Acers, which are known to form symbiotic bonds with many ectomycorrhizal fungi such as Amanita, Boletus, and Russula species, and thus boasts a rich and diverse fungal flora found throughout the year. Due to the diverse fungal flora of the surrounding area, it was thought to be an important source of fungal spores in the surrounding atmosphere. Leaves of multiple plant species were also examined and found to host a range of fungal spores. The seaside environment was located next to the Sea of Japan, with sparse vegetation and located within an urban area, thus providing an environment with opposing conditions, topography, and vegetation in comparison to Kakuma forest.

To test for collection efficiencies between sampling methods, a number of drone flights occurred at the same height as a Sibata pump (MINIPUMP MP-Sigma500NII, SIBATA Scientific Technology Holdings, Ltd., Saitama, Japan). Multiple flights were performed beside a building roof whilst a Sibata pump sampled at the same height on the roof surface. In the forest, Sibata pumps were hoisted to below canopy height with the use of a rope and set to start at specific times with a timer setting function, the drone then sampled simultaneously at the same height as the Sibata pump below canopy level. Drone samples were carried out using a hand launch, which consisted of turning on the drone propellers whilst holding onto the legs/body of the drone and subsequently taking flight, thus avoiding contact between the drone and surrounding surfaces (Figure 4). In most cases, flights (samples) were carried out 3 times sequentially, due to having 3 drone batteries. After each flight, the drone battery was replaced, and the pump was replaced with a fully charged pump and new filter holder already containing a GF filter.

Drone samples comprised of ~12 min flights, with the pump collecting bioaerosols for 10 min, hovering at specific heights (measured as above ground level in meters) both below and above the forest canopy; the sampling heights ranged from 8 m to 50 m. A maximum height of 50 m was chosen as it was high enough to be clearly identifiable as “above canopy” but also low enough to return the drone back to ground level quickly, thus maximizing flight time. It is also a height that other previously established sampling methods such as the use of helicopters or planes would be unable to operate in, highlighting the practicality of drone sampling.

2.4. Collection Efficiency and Spore Quantification

Initially, the total area of GF filters was observed under a microscope at a magnification of 200× and 400× to identify any spores that were collected, which were then photographed. Spore abundances were quantified microscopically by counting spores along the diameter of the GF filter at 400× magnification and equating the estimated number of spores per m³ of filtered air. The collection efficiencies of the drone method were compared to spore counts obtained from the Sibata pump samples, which sampled at a flow rate of 5 L/min for 10 min. For equations relating to spore quantification and spore estimation, please refer to the Supplementary Materials.

3. Results

Overall, 49 flights (dry = 43; rain = 6) were carried out between April and December 2024. The collection efficiency of drone sampling was found to be as effective as a Sibata pump when sampling at the same height (paired t-test, t = 1.6376, df = 24, p > 0.1) (Figure 5a). More spores were collected overall at ground level by the Sibata pumps when compared to the drone samples hovering below the canopy (paired t-test, t = 3.8281, df = 7, p < 0.05) (Figure 5b) or above the canopy (paired t-test, t = 3.409, df = 15, p < 0.05) (Figure 5c). The drone sampled a diverse range of fungal spores during dry and wet weather (Figure 6). Fungal spore abundance above the forest canopy remained fairly consistent throughout May until October as well as along the rooftop (Figure 7). Above the canopy, the coastal seaside area produced the lowest spore abundance (Figure S3).

4. Discussion

Overall, the drone collected fungal spores in all environments with a collection efficiency comparable to that of a commonly used portable sampler (MINIPUMP MP-Sigma500NII) when sampled at the same height, showcasing the effectiveness of the proposed drone sampling method. Although the range of heights are relatively short, the environments and contexts in which the drone was flown are important, as the drone was easily able to sample in thick forest vegetation just below the canopy. These areas are difficult or impossible for a range of commonly used aircrafts to safely maneuver in. The subsequent differences in spore abundance between ground sampling and below/above canopy are unlikely to be due to sampling bias as Sibata pumps and drone sampling have similar collection efficiencies at the same height. Therefore, other factors must be influencing spore abundances; for example, gravitational effects may cause spores to settle on the ground along with limited dispersal ability for various spores to be able to reach below/above the canopy [22]. The ground also harbors a greater range of fungal substrates (i.e., soil, plant litter, decaying matter) when compared with the tree canopy and thus there are likely to be more fungal fruiting bodies found at ground level; during rainy periods, the mechanical impacts of rain may be an additional causal factor for differences between heights [23].

In terms of spores collected, notable fungal spore types include Arthrinium, Nigrospora, and Cladosporium; these were common throughout the entire sampling period and were collected from all heights from the forest as well as the building roof. It appears that a majority of spores collected were plant pathogens. From July onwards, more spores with spiny ornamentation began appearing but we were unable to accurately identify the spores from morphology alone. Ganoderma spores appeared only in forest samples, whereas Spegazzinia spore collection was restricted to roof samples, which may suggest that certain spore genera may be affiliated with specific environments; this is logical as many fungi are known to be host-specific. The proposed method can be used in multiple environments at a range of heights as well as varying weather conditions and may be able to help indicate the importance of land cover and vegetation in relation to fungal spores; for example, in the current study, the low number of spores from the coastal area may be indicative of the importance of the supply of spores from the land where the lack of vegetation of land cover may limit the diversity and abundance of wind-dispersed fungal spores. In terms of wider practices, multiple drones could potentially be used across a range of spatial scales to sample simultaneously for allergenic and pathogenic fungal spores. Being able to collect seasonal samples both above and below forest canopies as well as within agricultural areas (i.e., above crops) may provide answers as to when and which fungal pathogens are present. Knowing when pathogenic fungi are most prevalent may help agricultural communities decide when to plant specific crops, and forestry commissions may also be benefited by similar information in relation to the protection of trees, which are susceptible to multiple fungal infections. Spore calendars based on height are currently lacking and would also further our understanding of how fungal spore communities differ and change spatiotemporally along a vertical gradient. Although sampling times are limited to 10 min, the known proximation for the volume of air sampled allows quantitative estimates to be predicted; this may be helpful for meteorologists creating atmospheric models as data regarding atmospheric fungal spores is lacking.

A noteworthy element was an Arduino as a program controlling the start and end of the attached pump, which ensures that spores are collected at the desired heights without contamination, confirmed by no fungal spores being found within the controls, and also allows for flexibility in terms of sampling duration. For example, in cases where the drone may be required to sample a considerable distance away from the pilot, the code can be changed to start after x number of minutes to allow researchers enough time to fly to desired areas. The total weight of the waterproof drone sampling pump was ~193 g for the Poweregg X and ~210 g for the DJI Phantom 3 Standard, payloads that can be carried by a range of hobbyist drones, perhaps widening the choice of drones that could be used with the proposed method, although there were no issues with stability, handling, or flight performance of either drone with the attached pump, it should also be noted that the additional weight reduces flight time and so it is suggested to have multiple batteries available in order to increase overall sampling time.

The ability of drones to collect visual data should also be taken note of as it can provide important evidence of multiple factors such as weather, light conditions, and topography. The ability to capture high resolution data with the use of sensors combined with the mobility of drones was outlined in research focusing on aerosol pollutants [24] and opens up the possibility of sampling with multiple drones in tandem, each with a differing target focus, i.e., organic and non-organic aerosols, thus providing a real time, multidisciplinary, and comprehensive overview of aerosols within the air. Last but not least, to our knowledge, this is the first time a drone was utilized to collect fungal spores during rain. This is a major advancement to drone-based bioaerosol sampling as it would allow researchers to compare spore abundances in varying weather conditions. Precipitation can have important effects on the fungal spores present within the atmosphere through splash dispersal and washing out effects [12,18,19,21,23] and may also be important for fungal spores which have ice-nucleating capabilities; therefore, the proposed sampling method may be able to further provide important information regarding these aspects. In terms of capturing ice-nucleating particles, the pump device could potentially be modified to allow for impingement-based sampling, collecting particles directly into a liquid medium to then be further analyzed by freezing experiments [17]. To further assess the utility for the proposed method, it is important to attempt DNA extractions from samples to understand whether further analysis can be carried out by this sampling method as well as to understand the mixing or movement of disturbed air due to the drone itself as this may influence the number and type of spores collected.

To summarize, the low-cost and practicality of the proposed method will allow researchers to sample fungal spores across a diverse range of habitats along a vertical gradient and provide an important advantage of being able to sample during rain, which we believe is an important step into understanding fungal spore dynamics during wet weather. As we successfully collected fungal spores with the use of a drone across all heights and habitats sampled, we suggest that the modifiability and ease of use of a drone can provide researchers with a suitable low-cost option when analyzing bioaerosols from lower atmospheric heights. It also offers the freedom of sampling in various locations due to its portability. It is highly likely that drone technology will continue to advance in terms of battery life; therefore, longer sampling times may be possible along with increased payload capabilities, allowing for larger more powerful pumps to be carried. The proposed method is a practical advancement to existing bioaerosol methods and one that has great potential to be modified and made increasingly efficient.