Use of Participatory sUAS in Resilient Socioecological Systems (SES) Research: A Review and Case Study from the Southern Great Plains, USA

Abstract

Since the publication of the seminal work People and Pixels: Linking Remote Sensing and the Social Sciences, the call to “socialize the pixel” and “pixelize the social” has gone largely unheeded from a truly participatory research context. Instead, participatory remote sensing has primarily involved ground truthing to verify remote sensing observations and/or participatory mapping methods to complement remotely sensed data products. However, the recent proliferation of relatively low-cost, ready-to-fly small unoccupied aerial systems (sUAS), colloquially known as drones, may be changing this trajectory. sUAS may provide a means for community participation in all aspects of the photogrammetric/remote sensing process, from mission planning and data acquisition to data processing and analysis. We present an overview of the present state of so-called participatory sUAS through a comprehensive literature review of recent English-language journal articles. This is followed by an overview of our own experiences with the use of sUAS in a multi-year participatory research project in an agroecological system encompassing a tri-county/tri-state region in the Southern Great Plains, USA. We conclude with a discussion of opportunities and challenges associated with our experience.

1. Introduction

Land use/land cover change (LULCC) is a pressing global environmental issue, resulting in landscape transformations, biodiversity loss [1], alterations in ecological processes [2], reductions in ecosystem services [2], and contributions to global climate change [3]. Such changes are often most acute in agroecosystems [4], where poor agriculture land management practices can lead to severe environmental and ecological degradation, such as soil erosion [5], deforestation [6], aquifer drawdown [7], and woody plant encroachment [8]. However, while agriculture may be the proximate cause of LULCC, governmental policy and institutional change [9], economic markets [10], livelihood preferences [11], as well as formal and informal governance regimes, serve as underlying, often invisible forces driving agricultural practices [7,12].

Sustainable land management, such as with agriculture, needs to be conceptualized as a socioecological system (SES), beyond land itself, which has both human and natural components. As such, SES resilience faces a fundamental challenge: the activities that contribute to environmental changes are often undertaken to support the livelihoods of those within these systems [13]. Land use practices are often driven by short-term needs rather than long-term sustainability. Understanding the complex multifaceted drivers of land cover change in SES is paramount for sustainable land management.

One approach to detangling these multifaceted drivers of land cover change in agroecosystems is to increase stakeholder involvement in the co-production of knowledge. Socioecological systems research is becoming increasingly reliant on citizen scientists and participatory research [8] to complement and supplant traditional top-down research approaches. Citizen scientists and other stakeholders (e.g., government officials and extension officers) in agroecosystems can provide unique perspectives on land use histories and the effects of governmental policy on landscape trajectories based on lived experiences, personal perceptions, and generational folk knowledge. The integration of this local, experiential, and traditional knowledge into the research ensures an equitable process that is mutually beneficial to both participants and researchers [14].

Such participatory approaches have roots in education and the social sciences [15] and involve researchers and participants working together towards a common objective. In contrast to traditional research approaches, participatory research stresses a disruption in traditional power relationships between the researcher/researched, emphasizing a bottom-up and community-engaged approach by involving stakeholders in the research process and the creation of user-friendly deliverables [9]. While fostering community engagement, participatory research simultaneously assists the public to become more informed decision-makers.

In the following sections, we present an overview of participatory research in the geospatial sciences. Through a comprehensive literature review, we argue that small unoccupied aerial systems (sUAS) offer new opportunities and challenges to develop a truly participatory approach to remote sensing. To further illustrate this, we present a case study of our own research using participatory sUAS to foster SES resilience in the Southern Great Plains, USA (SGP).

2. Literature Review

2.1. Participatory Research in the Geospatial Sciences

Participatory approaches within the geospatial sciences have a rich history, dating back at least to the late 1980s [16] with the adoption of sketch mapping in rural and indigenous communities (e.g., [17,18]). Prior to these initiatives, cartographic production was primarily limited to skilled practitioners, often operating in a position of power relative to the local communities being mapped. Due to a map’s role in shaping perceptions of the world [19], the power relationships between those producing and those “within” the map were reinforced through map adoption and use [20]. By integrating local communities into the collection, analysis, and sharing of spatial information, a new inclusive, bottom-up approach to modern mapping took hold (albeit not without criticisms, e.g., Bryan [21]). Cartography was once again supplanted by the more inclusive and ubiquitous mapmaking; that is, individuals and communities could choose to “map” rather than “be mapped” [22], or what others have called counter-mapping [23].

As geospatial technologies proliferated and aerial photographs and satellite imagery become more accessible, many citizen-based mapping approaches were instituted, expanding the number of potential participants and improving the accuracy of the collected data [21]. From this, “participatory GIS” (PGIS) was born, combining participatory mapping, geospatial technologies, communication, and advocacy with participatory research [16]. In its infancy, PGIS focused largely on the co-production and interpretation of spatial data with the goal of empowering less privileged groups in society [24]. As the availability of user-friendly web mapping platforms grew and smartphones became ubiquitous, the collaborative production of a broad range of spatial data, referred to as geo-crowdsourcing [25], volunteered geographic information (VGIS) [26], and/or neogeography [27] took hold.

Attempts to link the social sciences with remote Earth observations date to the mid-1990s, when the National Resource Council (NRC) and NASA organized a workshop to bring social and remote sensing scientists together. Subsequently, a seminal volume, Pixel and People: Linking Remote Sensing and Social Science [28] was produced. The overarching goals were to encourage mixed methods researchers and social scientists to use remotely sensed data and to foster collaboration between remote sensing experts and social scientists [29]. Despite attempts to “socialize the pixel” and “pixelize the social” [30], the analysis and interpretation of remotely sensed data, even within a social context, remains a largely a top-down approach [31].

Participatory remote sensing involves ground-level observations to verify (i.e., ground truth) remote sensing data products and/or participatory mapping methods to complement remotely sensed data products. For instance, Zaehringer et al. [32] combined participatory maps, local knowledge of land use history, and remotely sensed imagery to map annual land cover change in three countries. Sawant et al. [33] utilized satellite imagery in conjunction with ground-based, crowdsourced data to estimate crop phenology, while both Hodbod et al. [34] and Delgado-Aguilar et al. [35] combined participatory mapping with classified remote sensing derived land cover data to evaluate ecosystem services. Recently, Bennett and colleagues [36] pushed for a more robust co-created approach via “critical remote sensing”.

Another example involves our SES research in the SGP, e.g., [9,12,37,38]. As part of a long-term (15+ year) project on drought resilience and land governance, we developed LULCC products to explore the relationship between drought, woody plant encroachment (WPE), land tenure policy [37], and the growth of center pivot agriculture [7]. Though the remote sensing analysis indicated increases in WPE, as well as differential expansion and contraction of center pivot agriculture, we sought to co-produce our findings with on-the-ground stakeholders. Through a series of adaptive co-management community meetings [9], we shared remote sensing-derived LULCC maps with stakeholders. Participants interpreted, explained, and co-produced the results. This helped us to better understand the socioecological factors that have contributed to the observed changes.

2.2. The Role of sUAS in Socioecological Systems (SES) Research

Despite increasing community involvement in the co-production process, the acquisition and processing of remote sensing products have remained largely non-participatory. Perhaps this is on account of the nature of acquiring and processing remotely sensed imagery. It may also be due to the specific training of many remote sensing professionals, which often takes place outside of participatory training. However, the growth of team-based, transdisciplinary, and convergent research funding opportunities, along with the recent surge in relatively low-cost, ready-to-fly small unoccupied aerial systems (sUAS) may be changing this.

sUAS are playing an increasingly important role in SES research, particularly sustainable agriculture. For instance, sUAS have been used to calculate the crop canopy leaf area index [39], estimate crop evapotranspiration rates [40], identify crop water stress [41], identify nuisance species [42] and pest management strategies [43], and assess crop health [44]. Additionally, sUAS have been used to assist smallholder farmers in farming operations, reducing food insecurity, and improving farming efficiency [45]. Moreover, sUAS-based imagery, used in conjunction with traditional satellite imagery, has been used to enhance agricultural practices and increase yields [46]. sUAS have also been used in precision ranching, for instance, for igniting prescribed fires [47] and rangeland pest management [48].

In 2015, the U.S. Federal Aviation Administration (FAA) mandated registration of all sUAS weighing more than 0.55 pounds (250 g) and less than 55 pounds (25 kg). Subsequently, more than 1.37 million sUAS users in the U.S. have registered sUAS. Given the ubiquity of these tools and the wide range of applications in SES, a new frontier of participatory sUAS is possible. Perhaps the most obvious avenue is crowdsourcing drone imagery collection or, so-called voluntary sUAS (VsUAS).

For example, the Fly4Fall campaign [49] called on drone pilot volunteers to capture 360 panoramic imagery of plant phenology around the globe during fall senescence. To participate, volunteers needed a compatible DJI sUAS, a mobile device running iOS, and a free app called Hangar 360 for DJI. Qualifying volunteers could then launch the Hangar 360 for DJI app in an appropriate area, which would pilot the sUAS to 300 feet (100 m) and capture a series of images from multiple directions and angles. Volunteers could then upload the images to Hangar for processing and email a link of the processed images to the researchers [50]. Following Fly4Fall, the project coordinator applied the same crowd-sourced sUAS principles to map the devastating Camp Fire in Sierra Nevada Foothills in California. Over several days in 2018, 16 teams of pilots and visual observers mapped structural damage throughout the town of Paradise, CA. With assistance from industry, the VsUAS project was able to rapidly produce digital orthomosaics for search and rescue crews from each flight and 360 panoramas from ancillary flights [51].

Several other VsUAS initiatives are noteworthy. The Interdisciplinary Citizen-Based Coastal Remote Sensing for Adaptive Management (IC-CREAM) program, for instance, trains volunteers in the Great Lakes region to pilot quadcopters and collect geospatial data related to coastal change and vulnerability [52]. Similarly, Purcino et al. [53] recruited citizen scientists to use UAS to measure seasonal-scale beach erosion in Victoria, Australia.

Nonetheless, sUAS have some limitations, especially when compared to traditional remote sensing platforms. For instance, due to their short battery life, flight range restrictions, advanced maneuverability needs, and consistent data collection methodologies, collection of consistent sUAS data over large areas is difficult. To this end, Alwateer et al. [54] have proposed a scripting framework to help carry out crowd-powered drone services. Another initiative, SOAR, seeks to monetize VsUAS imagery by providing a decentralized platform for content creators to share nadir sUAS-derived aerial images, videos, and other content with consumers of said products. SOAR’s goal is to develop what the project creators call an aerial “super-map” of the world [55].

2.3. Participatory sUAS

While VsUAS is a subset of participatory sUAS, it involves the collection and dissemination of drone imagery by pilots themselves. A more robust definition of participatory sUAS involves stakeholder and researcher involvement in multiple aspects of a sUAS-based project, from project design to analysis, to achieve common objectives. Based on this proposed framework, we envision stakeholder participation in all phases of the sUAS process, from mission planning to data analysis (Figure 1). Despite the examples above, the published literature on sUAS for use in this narrower definition of participatory research remains scant.

In fall 2023, we conducted a survey of recent English-language peer-reviewed literature using a sensu stricto definition of participatory sUAS (see Figure 1), namely community involvement in multiple aspects of the research including (but not limited to) site selection, data acquisition, data interpretation, and/or data analysis (see Vargas-Ramírez et al. [56] for a review of “community drones”; and Sauls et al. [57] for an additional review). We excluded grey literature, as well as conference proceedings. Using a series of keywords and keyword combinations, such as “drones”, “participatory”, “unoccupied aerial systems” (and multiple variants thereof), “sUAS”, “UAV”, we discovered a small number (<30) articles, of which relatively few met our senso stricto criteria. Of these, we selected twelve to review in more depth (Table 1).

The resulting articles mostly involved socioecological analysis of coupled human and natural systems, whether for disaster risk management, e.g., [58,59,60], agroecosystem analysis, e.g., [61,62], urban or land use planning, e.g., [63,64,65,66], landscape archeology, e.g., [67], or environmental justice, e.g., [68,69]. The studies occurred in multiple locations, primarily involved multirotor aircraft, and primarily involved participation in imagery interpretation, but lacked a truly bottom-up approach in all aspects of sUAS imagery collection and analysis.

Table 1. Overview of the reviewed participatory sUAS literature.

Publication	Topic	Platform	Participatory Approach
Brandt et al. [58]	Disaster Risk Management	n/a	Data interpretation
Colloredo-Mansfeld et al. [61]	Agroecosystem Analysis	Multirotor	Data acquisition and interpretation
Dinko, Nyantakyi-Frimpong [62]	Agroecosystem Analysis	Multirotor	Site selection, data acquisition and interpretation
Kleinschroth et al. [63]	Urban or Land Use Planning	Fixed Wing and Multirotor	Site selection and interpretation
Larrain et al. [67]	Landscape Archeology	Multirotor	Site selection
Li and Deliberty [59]	Disaster Risk Management	Multirotor	Data interpretation
Luo et al. [64]	Urban or Land Use Planning	Multirotor	Data interpretation
Naufal et al. [65]	Urban or Land Use Planning	Fixed Wing	Data interpretation
Paneque-Gálvez et al. [68]	Environmental Justice and Sustainability	Fixed Wing and Multirotor	Site selection, data acquisition and interpretation
Radjawali et al. [69]	Counter-mapping	Multirotor	Data acquisition and interpretation
Saputra et al. [60]	Disaster Risk Management	Multirotor	Data interpretation
Skondras et al. [66]	Urban or Land Use Planning	Multirotor	Data interpretation

Table 1. Overview of the reviewed participatory sUAS literature.

Publication	Topic	Platform	Participatory Approach
Brandt et al. [58]	Disaster Risk Management	n/a	Data interpretation
Colloredo-Mansfeld et al. [61]	Agroecosystem Analysis	Multirotor	Data acquisition and interpretation
Dinko, Nyantakyi-Frimpong [62]	Agroecosystem Analysis	Multirotor	Site selection, data acquisition and interpretation
Kleinschroth et al. [63]	Urban or Land Use Planning	Fixed Wing and Multirotor	Site selection and interpretation
Larrain et al. [67]	Landscape Archeology	Multirotor	Site selection
Li and Deliberty [59]	Disaster Risk Management	Multirotor	Data interpretation
Luo et al. [64]	Urban or Land Use Planning	Multirotor	Data interpretation
Naufal et al. [65]	Urban or Land Use Planning	Fixed Wing	Data interpretation
Paneque-Gálvez et al. [68]	Environmental Justice and Sustainability	Fixed Wing and Multirotor	Site selection, data acquisition and interpretation
Radjawali et al. [69]	Counter-mapping	Multirotor	Data acquisition and interpretation
Saputra et al. [60]	Disaster Risk Management	Multirotor	Data interpretation
Skondras et al. [66]	Urban or Land Use Planning	Multirotor	Data interpretation

Nonetheless, several articles truly exemplified the potential of participatory sUAS. Colloredo-Mansfeld and colleagues [61] utilized high-resolution sUAS imagery to study local agricultural practices in the Galapagos Islands. The authors involved participants in multiple aspects of the sUAS process, from site selection to final image interpretation. The authors worked with the participants to identify areas they found most significant or exhibited the greatest diversity in land cover types and targeted these for sUAS imagery collection. The researchers then printed the processed orthomosaics on A3 (297 mm × 420 mm) paper and used these in interviews with the participants to classify land cover and document invasive species.

Paneque-Gálvez and colleagues [68] utilized a participatory approach in sUAS data collection to foster environmental justice and sustainable development in Central and South America. Because both legal and illegal activities pose threats to indigenous territories, the authors trained indigenous communities in the use of sUAS for mapping territorial boundaries and documenting activities on indigenous lands.

Radjawali and colleagues [69] demonstrated that cheap, custom built sUAS could be utilized for counter-mapping purposes and to monitor illegal activities on indigenous lands. As a result, various NGOs and community groups recognized the potential uses of these technologies to aid counter-mapping initiatives and began to use the technologies. Indeed, traditional counter-mapping strategies relied on relatively low resolution, freely available satellite imagery coupled with community-based sketches. These products often could not compete with officially sanctioned cadastral maps, especially when seeking formal recognition by state agencies and courts. The use of sUAS enabled these communities to capture features on the ground that were unobtainable using previous counter-mapping techniques. Moreover, the use of sUAS enabled the capture of information over relatively large areas over short periods of time at an affordable cost to the affected communities.

Ultimately, while our survey of the literature demonstrates that the currently published research on the participatory sUAS remains somewhat limited, sUAS have already proven to be a valuable tool in various facets of participatory research. In particular, researchers, such as Theuerkauf et al. [52], have established protocols to involve stakeholders into all phases of sUAS data collection and analysis, here for coastal monitoring, while others, such as Colloreda-Mansfeld et al. [61] have demonstrated that sUAS coupled with local agroecological knowledge can promote sustainable socioecological systems. These examples underscore the broad applicability of integrating sUAS into a range of participatory studies. To demonstrate this further, we will present an overview of our own use of sUAS in a mixed-method, multi-year participatory research study involving agroecosystems during times of drought in the Southern Great Plains, USA. We present the opportunities and challenges to assess the success and limitations of our project.

3. Case Study

3.1. Participatory sUAS: An Example from the Southern Great Plains

As part of a multi-year, mixed-methods project in a tri-county/tri-state area (Figure 2) in the SGP, we sought to evaluate socioecological resilience in agroecosystems experiencing recurrent, prolonged, and intense drought. Here, we defined SES as coupled human and natural systems which are inextricably linked and exert influence on one another [70]. Operating under the premise that sustainable management solutions are more effectively developed and adopted using participatory and citizen science approaches, we used a variety of such approaches to identify and test how agricultural and water policies impact individual land and water decisions. Within this context, we implemented a participatory sUAS approach involving land managers in the identification, collection, and analysis of sUAS imagery and derived products (see framework in Figure 1).

3.1.1. Background

Drought is a recurrent cyclical phenomenon in the semi-arid SGP, including the infamous 1930s Dust Bowl [71]. More recently, the region experienced an extended period of abnormally dry conditions lasting from ~2000 to 2022, with brief abatements in between [72]. Drought can cause shifts in ecosystem structure and function and drive ecosystems beyond thresholds of resilience, impacting ecosystem services and triggering feedbacks in the coupled human and natural system [73]. In semi-arid grassland environments, drought-induced reductions in herbaceous cover may provide opportunities for woody plant seedling establishment [8]. Drought, coupled with overgrazing, can reduce fuel loads, altering fire regimes and fostering conditions for additional woody plant encroachment (WPE) [74].

Against this backdrop of drought and WPE, land managers must make decisions on how to balance their own economic livelihoods with sustainable land management practices. Additional factors, such as land tenure [37] and common resource policies [7] can influence the decision-making processes [9], leading to differential LULCC and socioecological resilience. Moreover, while WPE has been extensively studied, uncertainty remains about its rates, dynamics, and spatial extent [75]. Methods to quantify WPE have traditionally involved the use of aerial photography and satellite imagery. However, the temporal resolution of the former is often too coarse to adequately determine the rates and extents of WPE, while the spatial resolution of the latter is too coarse to detect WPE until it has already developed into a significant problem (Figure 3).

Coupled with recurrent, persistent drought and differential vulnerability/resilience, WPE requires land managers to make complex land use decisions based not only on current biophysical conditions, but also on other factors such as market conditions and state, regional, and local-level regulations. However, land managers often lack sufficient tools to effectively monitor and gauge these ecological changes. Consequently, we proposed using participatory sUAS to monitor fine scale landscape dynamics.

3.1.2. sUAS Surveys

As part of the larger project in the study area (Figure 2), our project team conducted 135 household surveys with landowners and managers, balanced between a mix of farmers and ranchers, beginning in spring 2018. Participants were recruited through an open, participatory and citizen science enrollment approach, with an eye on geographically stratifying to capture the diversity of ecological transitions. To link LULCC to agricultural and water policies at multiple spatial scales, we sought to map detailed LULCC at the plot/household level using sUAS and collected the data in concert with the household surveys.

During the household surveys, participants were asked if they would like to participate in the ancillary sUAS project (an experimental control group of up to n = 45) involving repeat aerial photography of two 10-acre (4 ha) plots on their property to be flown at an altitude of 150 ft (~46 m). We presented this as an opportunity to provide high-resolution aerial imagery of portions of their property to aid in land management decisions and to analyze fine-scale LULCC. Thirty-four producers volunteered to participate. Flights were conducted in tandem with the household survey in 2018 and 2019 (Figure 4; 15 in Cimarron County, 15 in Union County, and 4 in Las Animas County). Scheduled repeat flights were hampered by the global COVID-19 pandemic, and were not completed until 2023, when they took place in conjunction with repeat household surveys, but with a much smaller sample given the pandemic and related attrition.

Each participant was provided a book of imagery with plat-level boundaries and asked to identify two plots on their property to fly. With an eye to understanding environmental perceptions and folk knowledge, we asked participants to identify one plot exhibiting ecological degradation (e.g., WPE or severe erosion) and the other limited perturbations. The selection of the plots, though, was left to the discretion of the participants to ensure sustained engagement as well as user-friendly deliverables. We then provided an IRB consent form (AS-18-24) covering the procedures, potential risks to participants, confidentially assurances, and participant rights; and a sampling protocol form to capture relevant information about each plot, such as the geographic location, vegetation cover types, disturbances, land use types, and land use and land management history (Figure 5).

Either with the participants’ direct assistance or based on the areas they identified, we created automated flight plans (Figure 6) using DroneDeploy, a cloud-based sUAS mapping platform and associated mobile application. Because the flights were conducted in concert with household surveys, we did not generally have advanced notice of flight locations, so flight plans had to be created in situ. For each plot, we sketched a ~10 acre (4 ha) flight plan and set a launch point, altitude (150 ft; (~46 m)), side overlap (70%), front overlap (75%), flight direction (variable), and mapping flight speed. Due to landscape heterogeneity, flight plan sizes, shapes, and orientations varied.

At a date and time designated by the participant, we visited the flight sites. We always had at least two project members, an FAA-licensed remote pilot-in-command to fly the sUAS and a visual observer to keep constant eye contact on the sUAS during the flights. The landowners/managers were also invited to participate in the flights as much or as little as personal preference allowed (Figure 7). Flights were conducted in the summer of 2018, though we returned to the study area in the summer of 2019 to complete several flights we were unable to complete the previous year. All initial flights used a DJI Phantom IV Pro. Follow-up flights conducted in 2023 utilized a DJI Mavic 3E.

3.1.3. Post-Processing and Image Distribution

Each flight produced ~150 individual georeferenced stills (though this value varied due to the irregular shapes and sizes of sampled areas). We used Pix4D Cloud, a suite of photogrammetry and sUAS mapping software, to create individual orthomosaics, point clouds, and digital surface models for each flight (n = 68). We set up an ArcGIS Enterprise Server to host the processed orthomosaics and created a customized, secure, and user-friendly web mapping application (web app) for each participant using the ArcGIS API for Javascript. Each web map (n = 34) featured the two orthomosaics captured on the participant’s property and custom co-developed tools for image interaction. These include basic navigation tools; basemap selection; printing options; linear and areal measurement tools; location search; and annotation tools (placemarkers, lines, and various polygonal drawings) (Figure 8). Additionally, we wrote help documentation to assist the users, many of whom have limited technological prowess. We sent an email to each participant in April 2019 providing a link, username, and password to their web app. After distribution, we continued to update the web app to make it more user-friendly based on conversations with producers at annual Soil and Water Conference District (SWCD) annual meetings and other interactions. While COVID-19 hampered repeat flights, for the few properties that we completed repeat flights in 2023, we updated the web apps to also include the repeat orthomosaics.

4. Results and Discussion

4.1. Opportunities and Challenges

Based on our experience, we offer the following observations on the opportunities and challenges of participatory sUAS. We conclude by presenting an overview of lessons learned in hopes of benefiting future work involving participatory sUAS.

4.1.1. Mission Planning

Participatory sUAS mission planning provides opportunities to involve stakeholders in ways not afforded in traditional remote sensing. Land managers provide a unique perspective on the land use history, as well as challenges they face as producers. Additionally, few will know the land as well as those who have worked on it for years. Through direct stakeholder engagement during the household surveys and recruitment phase of the project, we were able to identify the areas of interest to the producers; ecological issues surrounding the areas to be surveyed, and detailed land use histories. Prior to each flight, we asked participants to identify the areas they wanted us to fly desired data products; areas to avoid; species of concern; land tenure; and land use practices encompassing the past 10 years at the time of the flights (2009–2018; see Figure 5). It would not be possible to ascertain such ancillary information using traditional remote sensing methods, even with participant involvement in ground truthing and/or sketch mapping.

Nonetheless, land manager availability in these agroecosystems occasionally posed a challenge from a participatory perspective. The producers with whom we worked kept long work hours, especially during the growing season, and are often located in remote, inaccessible locations. Coordinating with them was often challenging because it would require them to readjust their schedules to accommodate our flights. This often delayed flights and was sometimes overcome by the land manager allowing us access to the property without their physical presence. Other flights took place on public or rented lands and required multiple permissions.

Our goal was to involve the land managers in as much of the data acquisition process as possible. While all participants selected the flight areas and provided insight into the land use histories, due to both regulations and our IRB protocol, we always conducted the flights ourselves. Nonetheless, we sought land manager participation in all aspects of the project. We did this for several reasons, e.g., to ensure stakeholders felt a degree of ownership of the data products and to gain additional insight into the portions of the property on which we were conducting the aerial surveys.

Because of the geographic extent of the study area and the necessity of scheduling flights based on land manager availability, we often had to conduct flights at suboptimal times. To ensure appropriate top lighting of Earth’s surface for vertical aerial photography, images should be acquired under clear sky conditions with the sun in a relatively high position to reduce the effects of shadows [76]. However, land managers were often not available at optimal times and image acquisition times varied from late morning to late afternoon (see Figure 7). Additionally, many participants were located tens to hundreds of kilometers from one another. To complete the maximum number of flights in a short time, this required traversing long distances and conducting flights when possible.

Other atmospheric conditions, such as cloud cover, wind, turbulence, and precipitation can inhibit the time of flights and the sUAS data accuracy [76]. Heavy winds can cause relatively light-weight sUAS to fly off-course, leading to blurred images, poor flight performance, and even catastrophic hardware failure. Most sUAS manufacturers will provide specifications for appropriate flight conditions. DJI’s safety guidelines for the Phantom 4 Pro caution against use when wind speeds exceed 10 m/s or 22 mph [77]. In the semi-arid SGP, these wind speeds were occasionally exceeded, resulting in delayed or canceled flights.

The remoteness also posed issues when designing the flight plans. The flight planning software we utilized, DroneDeploy, required an internet connection to construct a flight plan (once the plan is saved to a mobile device, connectivity is no longer required to execute the flight plan). Some of the participants had Wi-Fi and permitted us to connect to design the flight plans while we were conducting household surveys. When Wi-Fi was unavailable, we had to rely on cellular networks for connectivity. Within some portions of the study area, such coverage was often unreliable. To compensate for this, we met with the landowners, determined the areas in which to fly, constructed the flight plans offsite, and returned later to execute the flights.

4.1.2. Stakeholder Engagement and Social Capital

A necessary component of participatory research is stakeholder engagement. Our project team began to work with affected communities in the study area almost fifteen years ago. Through the course of related projects, all of which involved varying degrees of stakeholder outreach, we fostered positive relationships with community members. This enabled us to gain valuable social capital. Additionally, this social capital afforded us opportunities to work with participants in capacities that go beyond the traditional research paradigms [8], including participation in county fairs and soil and water conservation district meetings. Through this, we established strong bonds with the participants, which helped us to maintain relationships even through a pandemic.

Further, opportunities to be directly in the field with producers are enlightening. One graduate student received ample training on grass species biodiversity while working with a producer at the flight site. While our household surveys generally happen around the kitchen table, being outside with producers often reveals connections between complex human–environment processes or additional information about their property that we would not otherwise have known. Producers often feel more at ease outside, where they are in their element. The property walkabouts and time spent in a vehicle together can be a bonding experience. This relaxed atmosphere can produce enthusiastic conversations about land management that can be as revealing as the images produced from the endeavor itself.

Nonetheless, gaining and maintaining trust between a community and academic researchers is a challenge of participatory research [78]. In rural communities, building relationships can be particularly challenging. Additionally, the nature of the research can inadvertently foster this distrust. As early as 30 years ago, researchers (e.g., [29,79]) raised privacy and other ethical concerns regarding the acquisition and use of geospatial data. The ubiquity of sUAS, whether used for government, academic, or recreational purposes, has further compounded these concerns. Though we only operated with express informed, written consent of participants, guaranteeing both privacy and anonymity, some potential participants indicated reticence about the use of sUAS on their property.

4.1.3. Hardware Failure

We encountered several other technological problems throughout the course of the project, including the loss of two aircraft. In both instances, the failure was related to a communication link error between the controller and aircraft resulting in the aircraft crashing during the automated flight plan.

These crashes underscore a critical issue when using sUAS in participatory and other research. Though there is a large sUAS recreational market and many regard sUAS as little more than toys, there are multiple hazards involved in the operation of sUAS, some of which can result in property damage, injury, and even death. Researchers wishing to utilize sUAS should ensure they have appropriate training, are operating in accordance with appropriate jurisdictional regulations, have informed, written consent from any participants who have been duly apprised of potential hazards; and are operating in a manner that mitigates risk.

4.2. Lessons Learned: The Pursuit of a Truly Participatory sUAS

Despite challenges, we believe our participatory sUAS approach accomplished its primary goals, which were to involve land managers in the process of selecting areas for aerial survey, to provide participants with high-resolution aerial imagery, and to provide tools to assist in sustainable land management, especially in relation to WPE and recurrent drought. We see this as a first step toward developing participatory sUAS in the study area. We hope to encourage land managers to adopt similar technologies to help crowdsource high resolution, multi-temporal imagery throughout this drought-prone region.

Our experiences, both the opportunities and challenges, provide a perspective from which we can discuss the lessons learned and, ultimately, promote a truly participatory sUAS to foster resilient socioecological systems. Based on our review of the current literature on participatory sUAS, we found that, despite the growing interest in using sUAS in participatory studies, few have truly achieved a senso stricto participatory sUAS. Nonetheless, the ubiquity of off-the-shelf, ready-to-fly sUAS, the availability of relatively easy-to-use desktop and cloud-based mission planning and data processing software, evolving regulatory requirements, and new tools to analyze sUAS-derived imagery, this could change. To this end, we propose to incorporate stakeholder involvement in all aspects of the sUAS process (see Figure 1), including mission planning, data collection, data processing, and data analysis. While we recognize that barriers to stakeholder engagement may continue to exist each step of the process, such as a regulatory requirements, lack of training, costs, and other factors, we believe that it is possible to engage stakeholders in all phases of a participatory sUAS project. Through this, we can finally achieve the goals to “socialize the pixel” and “pixelize the social” [30].

5. Conclusions

The rapid proliferation of sUAS is quickly opening new avenues of scientific research in an array of disciplines. Within SES research, we envision a number of opportunities using sUAS, especially incorporating both crowd-sourced (VsUAS) imagery and other participatory approaches. Unlike traditional participatory remote sensing in which participants typically are only involved in ground-level observations to verify (ground truth) remote sensing data products and/or ancillary participatory mapping methods, use of sUAS in participatory research provides a means for participant involvement in all aspects of the research, from project design and data collection to data processing and analysis (assuming proper training and resource accessibility). Our research in the SGP demonstrates the feasibility of incorporating stakeholder involvement in fine scale sUAS land cover mapping and provides insights from which other participatory sUAS projects can develop. Underpinning this approach is a desire to address critical environmental concerns and build community resilience by involving the very stakeholders whose livelihoods are dependent on the sustainable use and development of their natural resources.