iNaturalist as a Tool to Improve Agroecological Biodiversity Knowledge: An Ecuadorian Case Study

Introduction

Defined as public participation in scientific research that involves non-professional individuals in scientific endeavors (Kobori et al. 2018), citizen science has the potential to promote social empowerment, scientific literacy, the democratization of science, collaborative systems, and global citizenship (Finquelievich and Fischnaller 2014; Fritz et al. 2019; Wu and Hsu 2024). It can also contribute to the development of public policy and environmental governance (Kobori et al. 2016; Schade et al. 2021). As an exercise in democratizing knowledge production, citizen science has expanded successfully in recent decades, primarily in countries of the Global North, while practice and benefits in the Global South are only beginning to emerge (D’Onofrio, Arza, and Actis 2024).

Citizen science and agroecology share an approach centered on active public participation. Agroecology promotes the creation of sustainable food systems by engaging both producers and consumers (Hanappe et al. 2016). Similarly, citizen science enables people to contribute to the planning, monitoring, and evaluation of agroecosystems. However, its application within agricultural research remains relatively underexplored (van de Gevel, van Etten and Deterding 2020; Ebitu et al. 2021; Ryan et al. 2018).

Given that agriculture is a major driver of global biodiversity loss, addressing information gaps on biodiversity in cultivated landscapes is critical for improving our understanding of these impacts and for guiding conservation and restoration priorities (Remelgado, Levers, and Cord 2025). Integrating agroecology with citizen science can help address challenges related to fertilization, pest and pollinator management, land use, irrigation, and crop yields (Mourad et al. 2020). In addition, citizen science can drive shifts in practices and knowledge associated with agroecology and biodiversity conservation (Gerits et al. 2024), and thereby advance agricultural practices that support biodiversity conservation (HLPE 2019).

Nevertheless, several barriers limit stronger connections between agroecology and citizen science. A major challenge is the limited participation of farmers in citizen science projects (Mourad et al. 2020).

Numerous farmer-oriented applications exist (e.g., Plantix, Bioleaf, Adama Bullseye, Pmapp; Mendes et al. 2020), that focus on crop protection and diagnostics, nutrition and fertilization, irrigation, plant development, harvesting, general farm management, or field information systems. However, very few integrate environmental monitoring or biodiversity assessment components (Ruck et al. 2024). In contrast, one of the most widely used and successful citizen science platforms for biodiversity monitoring worldwide is iNaturalist and its associated application, Seek (Wittmann, Girman and Crocker 2019). In Latin America, 9 of the 17 countries operate their own national iNaturalist nodes to promote citizen science activities (iNaturalist 2025). These nodes play an important role in establishing the platform’s legitimacy and promoting its local adoption (Grattarola et al. 2024; Riquelme 2024).

Key features of iNaturalist—such as ease of use, open access, versatility, and automation—have facilitated the rapid development of exchange networks among users (López-Guillén et al. 2024). An additional benefit is the ability to identify pollinators without lethal sampling methods, thereby avoiding ethical concerns linked to specimen collection (Turley et al. 2024). Although iNaturalist activities cannot replace the expertise of professional specialists such as botanists, they remain an important data source when researchers carefully evaluate the contributed records (López-Guillén et al. 2024). Interactions among and between youth participants and professional taxonomists affiliated with museums and herbaria can further strengthen the use of the platform (Echeverria et al. 2021). Despite the large number of projects hosted on iNaturalist, relatively few scientific publications actively use the available data. Martínez-Sagarra et al. (2022) reported that 86% of iNaturalist projects in Spain serve exclusively educational or recreational purposes, while only two percent focus on scientific research. (In contrast, the data from iNaturalist on GBIF are used more than any other data provider for scientific publications). In 2023, the lead author of this article conducted an exploratory search within the global iNaturalist Projects section using the prefixes “agri” and “agro,” identifying 291 projects. After applying additional filters, including a minimum number of participants, location within agricultural areas, and a threshold of at least 10 observations per project, only 110 projects (38%) worldwide were found to directly link biodiversity observations to agriculture, agroecology, or agricultural environments.

In Ecuador, a megadiverse country, most iNaturalist initiatives have historically concentrated on documenting wild biodiversity. Starting with only four records in 2008, iNaturalist now reports more than 2.3 million observations from Ecuador, with exponential growth in recent years (iNaturalist 2025). The Ecuadorian national iNaturalist node (iNaturalistEC) is coordinated by the National Biodiversity Institute (INABIO), which has promoted the platform since 2018 (INABIO 2019). Most data contributions have been collected through “bioblitzes,” intensive citizen science events promoted by INABIO and other initiatives to enrich global biodiversity databases (Meeus et al. 2023).

Here, we present Agri Andes Ecuador, an agricultural biodiversity monitoring project initiated in July 2019 that actively engages students, farmers, and the general public in iNaturalist-based data collection. In Ecuador, only a small number of studies have emphasized community-based monitoring approaches involving both citizens and scientists (Medrano-Vizcaíno et al. 2023). Artificial intelligence platforms such as eBird, ePlants, and Ictio have expanded opportunities for environmental monitoring and for documenting environmental crimes (Jones et al. 2022; Páez-Vacas et al. 2023; Ulloa 2023). For our study, it is therefore significant that iNaturalist does not rely exclusively on artificial intelligence for species identification, and instead actively promotes collaboration between participants and experts (Meeus et al. 2023).

Our study focuses on rural regions of the Ecuadorian Andes, typically characterized by crop rotation systems involving potatoes, corn, quinoa, barley, lupins, flowers, and diverse vegetables, alongside livestock grazing and agroforestry practices (Struelens, Mina, and Dangles 2024). Our current approach aims to establish a biodiversity baseline in agroecosystems located 2,000 m above sea level, enabling the investigation of ecological patterns and processes as long-term data accumulate. Beyond its ecological objectives, this study advances citizen science in agroecology by demonstrating that participatory platforms can generate robust data in understudied agricultural landscapes. By engaging students, farmers, and expert identifiers, it highlights the capacity of citizen science to bridge information gaps, enhance taxonomic resolution, and inform decision-making. This aligns with emerging frameworks positioning citizen science as a tool for agroecological assessment and transformation (Mupangwa et al. 2025), particularly through context-specific knowledge co-production. Our findings provide empirical evidence that citizen science is not only a viable research approach but also a mechanism to foster learning, strengthen participation, and support transitions toward more sustainable agricultural systems.

The objectives of this study were to demonstrate the value of citizen science using the iNaturalist platform to document biodiversity in the agricultural landscape by: (1) examining iNaturalist projects in Ecuador and exploring biases in studies of agricultural diversity (2) reporting the diversity of wild plants, insects, spiders and birds within agricultural systems, (3) evaluating the improvement in the identification of a complex taxonomic class such as insects, and (4) identifying plant-insect ecological relationships (especially with respect to insects that are capable of pollination and plants that can provide these insects with the resources they need) from observations generated in the project from July 2019 to December 2023. Our ultimate goal is to develop citizen science-supported tools that generate useful data to inform decisions in support of the agroecological transition of food systems, to conserve non-cultivated biodiversity in crop systems and the ecosystem services they provide.

Methods

The Agri Andes Ecuador project

Agri Andes Ecuador was established in 2019 as part of the Agroecological Pest Management (AMIGO) project, funded by the McKnight Foundation and implemented in collaboration with local farmers and universities in Ecuador (Agri Andes Ecuador 2019). The initiative was conceived by the lead authors of this article, who are affiliated with the Pontifical Catholic University of Ecuador.

Between July 2019 and December 2023, the authors used iNaturalist to document birds, spiders, insects, and plants across predominantly agricultural landscapes in the central–northern Andes of Ecuador, at elevations above 2,000 m. The goal was to discover, learn about, and share information on these organisms and their interactions. The study area included the provinces of Carchi, Imbabura, Pichincha, Cotopaxi, Tungurahua, Bolívar, Chimborazo, Cañar, Azuay, and Loja, excluding national parks, reserves, moorland grasslands, and volcanic areas (Figure 1a). This region is characterized primarily by a mosaic of smallholder agricultural farms and expanding commercial floriculture, interspersed with remnants of Andean shrubland and moorland, as well as growing suburban and urban areas (Curatola Fernández et al. 2015; Guarderas, Smith, and Dufrene 2022) (Figure 1b).

Figure 1

Results

iNaturalist projects in Ecuador

The analysis showed that almost all of the analyzed projects had a generalist approach in terms of the diversity of species they intended to record. Of the 379 projects, 51% collected information at the country level and did not define a specific geographic area. The remaining projects focused on specific areas such as paramos (high altitude grasslands), protected areas, national parks, reserves, etc. Of the 379 projects, 59% were focused on flora and fauna in general, but only one (Agri Andes Ecuador) focused on agricultural landscapes with farmland biodiversity (Figure 2).

Figure 2

Dynamics of the Agri Andes Ecuador project

Agri Andes Ecuador has created a community of students, researchers and farmers who have contributed their records to the continued growth of this local initiative (see Supplemental Table 2). The community was 43% men and 57% women; and 93% of participants were high school/university students, while 7% were teachers, farmers, or professionals from other fields. The growth of the project is illustrated by the strong correlation between the number of observers and the number of observations (r = 0.97; p value = 0 < 0.001) in Figure 3a. This trend is further supported by Figure 3b, which shows a correlation between the number of observations and the number of identifiers (r = 0.98; p value = 0 < 0.001).

Figure 3

Between July 2020 and November 2023, 27 training sessions were conducted, including virtual and in-person classes, as well as bioblitz workshops and competitions. Thanks to these activities (especially the bioblitzes), some trained students mobilized and recorded observations across several farmers’ fields. A total of 99 hours were invested in these activities. Between April and December 2021, training activities consisted mainly of classes, most of which were virtual because of the pandemic. Following these activities, observation numbers increased by between 10.5% and 16.6% (on average, 13.5%). An average of 2,000 observations per month was recorded in the four bioblitzes carried out (December 2020, May 2022, November 2022, and May 2023), ranging from 918 to 3,082 observations per month.

Interest in iNaturalist motivated faculty at participating universities to incorporate workshops, thesis projects, and teaching activities based on the application. The average number of records before the interventions was 318 per month. After the bioblitz and the classes, this figure increased to 745 per month. Overall, bioblitzes increased the slope of the number of observations over time by 200 to 459% compared with pre-bioblitz trends (Supplemental Table 3). Between the third and fourth years of the project (June 2022 to June 2023), the number of observers increased by 53.4%.

Between July 2019 and December 2023, Agri Andes Ecuador recorded 32,457 observations with 2,630 species, of which 51.5% were plants, 39.3% insects, 5.6% birds, 2.6% arachnids, and 0.9% other animals. 22.5% of the observations were identified to “research grade” level, while the rest of the records are still in taxonomic categories such as genus, family, and observations marked as “needs identification”.

The implementation of the Agri Andes Ecuador project allowed interaction with more than 1,900 identifiers (expert iNaturalist users mainly categorized as curators) and Agri Andes Ecuador project members (mainly ecuadorian researchers and university students) (Figure 3b). Of the top 20 identifiers who contributed the most to Agri Andes Ecuador, eight (40%) are Ecuadorian identifiers and naturalists, many of them with a regular identification activity within iNaturalist. Seven (35%) are curators in the iNaturalist platform, from three continents and nine countries (Germany, Brazil, Canada, Denmark, Ecuador, USA, Spain, Korea, and Mexico).

iNaturalist in the identification of diversity associated with agricultural fields

Plants represented the largest group recorded through citizen science, with 16,715 observations and 1,163 identified species, 99% of which were non-cultivated and associated with agricultural landscapes. Insects were the second most abundant group (12,762 records), mostly identified at the genus level, reflecting typical taxonomic challenges in citizen science datasets. Arthropod identifications were predominantly resolved at the genus level, with Bombus as the most frequent genus and Apis mellifera as the most recorded species. Birds (1,846 records) and spiders (860 records) were less represented. These patterns highlight both the potential and limitations of participatory platforms in generating biodiversity data across taxa (Figure 4).

Figure 4

iNaturalist as a tool to improve taxonomic identification in insects

Among insects recorded through citizen science, Diptera was the most represented order (4,686 records; 34%), followed by Hymenoptera (21%) and Coleoptera (17%). When considering observations with at least three agreeing identifiers, Hymenoptera showed the highest taxonomic resolution, with 550 records identified at species level, followed by Diptera (349) and Coleoptera (277) (Figure 5a).

Figure 5

Taxonomic improvement varied among orders, reflecting differences in detectability and expertise availability. All groups showed progress from initial to final identifications, demonstrating the effectiveness of community-based validation (see Supplemental Figure 1). Maximum identification levels (MIL) differed across orders, although some records remained unidentified even at the family level, highlighting persistent knowledge gaps (see Supplemental Figure 2).

More than 50% of records in Hymenoptera, Diptera, and Coleoptera reached higher taxonomic resolution (tribe/genus level) (Figure 5a,b,c). A. mellifera accounted for 40% of species-level identifications within Hymenoptera, indicating a bias toward conspicuous species. In contrast, Hemiptera showed lower initial and final identification levels. (Figure 5d). These results illustrate both the strengths and limitations of citizen science in improving taxonomic knowledge through distributed expertise.

Ecological relationships between insects’ pollinators and plants

From 494 citizen science photographic records, 21 insect species across 4 orders were identified interacting with 27 plant families. Hymenoptera dominated (71%), with A. mellifera, Bombus robustus, and Bombus funebris as the most frequent species, reflecting observer bias toward conspicuous pollinators. Diptera with species such as Eristalis tenax and Eristalis bogotensis, accounted for 13% of photographic records analyzed. These insects were associated with 24 plant families (89%), mainly Fabaceae, Asteraceae, Brassicaceae, and Malvaceae. Fabaceae was the most represented plant family (39% of records) (Figure 6). These findings demonstrate the capacity of citizen science imagery to document plant–insect interactions, while also highlighting taxonomic and detection biases inherent to participatory data collection.

Figure 6

Discussion

iNaturalist projects in Ecuador have addressed a variety of objectives, often at the country level. In cases where geographical scope was indicated, they have primarily focused on reporting biodiversity in protected areas, national parks, urban environments, or other non-agricultural areas. Only one study (our Agri Andes Ecuador project) was conducted within agricultural systems. This highlights a critical gap, but also a conceptual opportunity to position citizen science within agroecology, a field that explicitly integrates biodiversity conservation and crop habitat management (HLPE 2019). Agricultural landscapes remain underrepresented in biodiversity monitoring despite being key sites of ecological interaction and transformation (Remelgado, Levers, and Cord 2025). From a citizen science perspective, this underrepresentation reflects a broader tendency to prioritize “charismatic” or protected ecosystems, rather than productive landscapes where biodiversity is often more contested and less visible.

In this context, our project contributes to an emerging body of work that situates citizen science not only as a data collection tool, but as a knowledge co-production process embedded in socioecological systems. Martínez-Sagarra et al. (2022) have shown that iNaturalist projects in Spain led by local associations, involving researchers in project design, identification support, and data analysis, attracted more participants when the purpose and scientific value were clearly articulated. In Agri Andes Ecuador, partner universities incorporated citizen science tools into formal training programs and degree projects, which extend beyond the interactivity described by Echeverría et al. (2021). This integration suggests a shift from contributory citizen science models toward more collaborative or co-creative approaches, where participants are not only observers but also learners and interpreters of ecological processes.

This distinction is important from a theoretical standpoint. Citizen science has often been categorized along a spectrum, from contributory (data collection) to collaborative and co-created models (Bonney et al. 2009). Our project aligns more closely with the latter, as its goal is not only to report observations but to generate knowledge relevant for conservation and agricultural decision-making in underrepresented systems (Remelgado, Levers, and Cord 2025). In this sense, citizen science becomes a mechanism for linking ecological monitoring with evidence-based management and policy processes, particularly in contexts where formal scientific data are scarce.

At the same time, the project demonstrates that the quality and impact of citizen science depend heavily on the strength of interaction networks between scientists and participants. As highlighted by Henning et al. (2023), community networks can lead to significant scientific discoveries, such as the rediscovery of species previously thought extinct. In our case, the recording of (Pteroplatus luculentus), an insect with no recent reports of living specimens, illustrates how agricultural landscapes managed without pesticides can serve as reservoirs of overlooked biodiversity. This finding reinforces the idea that citizen science can reveal “hidden” ecological dynamics, particularly when it is embedded in territories undergoing alternative management practices.

Bioblitzes were identified as one of the most effective strategies for expanding participation (see also Martinez-Sagarra et al. 2022), increasing observations, and raising awareness. Their effectiveness aligns with previous studies emphasizing the role of gamification and collective events in motivating participation (Bowser et al. 2013; Meeus et al. 2023; Potsikas et al. 2023). However, our experience also highlights a limitation often overlooked in citizen science literature: Participation does not automatically translate into sustained engagement or data quality. Variations in duration, geographic scope, and institutional involvement may dilute sampling effort, particularly in projects that rely on heterogeneous participant groups such as students, schools, and universities.

This points to a key tension in citizen science: the balance between scale and depth. While large-scale participation increases data volume, it may also reduce consistency and engagement. Dimson and Gillespie (2023) suggest that projects with lower intrinsic appeal, such as those focused on agricultural systems, require additional support, feedback, and institutional mediation. Our findings support this view, as approximately three-quarters of participants did not remain actively engaged. This suggests that citizen science in agricultural contexts may require more intensive facilitation and feedback loops, particularly when working with communities that are not traditionally engaged in biodiversity monitoring.

Technological barriers also played a significant role. While smartphones facilitated participation among younger users (see Stevenson, Merrill, and Burn [2021] for a project with students at the University of Massachusetts, Boston), they limited engagement among farmers, particularly those with low connectivity or limited device capacity. This digital divide is consistent with findings from other contexts (Dehnen-Schmutz et al. 2016), and highlights the importance of adapting citizen science tools to local conditions. The positive reception of the Seek application suggests that offline-capable tools with intuitive interfaces can reduce barriers to participation (Campbell et al. 2023). However, the effectiveness of such tools depends on the availability of local biodiversity data, pointing to a recursive challenge: citizen science depends on data availability, but also contributes to it.

Data quality remains a central concern in citizen science. Our results confirm previous findings that identification accuracy can be variable, particularly in taxonomically complex groups such as insects (López-Guillén et al. 2024; Gardiner and Roy 2022). While plants were easier to record and often overrepresented, insects posed greater challenges due to their size, diversity, and the limited taxonomic expertise among participants. Nevertheless, methodological adaptations, such as the use of magnifying lenses and video recordings linked to youtube (Mina 2023), improved data quality and facilitated expert validation. These practices illustrate how citizen science can evolve through iterative methodological innovation, rather than relying solely on standardized protocols.

Importantly, citizen science platforms such as iNaturalist also provide opportunities to integrate and value local and experiential knowledge. In our project, the inclusion of local names and uses for plants and insects enriched the dataset and expanded its interpretative value (e.g., moscabeja, pakunka, and piliquitze; Cárdenas et al. 2025). This aligns with broader discussions on the role of citizen science in recognizing diverse knowledge systems and challenging the dominance of purely scientific classifications. From this perspective, citizen science is not only about data collection but also about epistemic inclusion, allowing different ways of knowing to coexist and inform ecological understanding.

The focus on pollinators revealed patterns consistent with global observations, including biases toward larger and more visible species (Turley et al. 2024). While such biases are often seen as limitations, they also reflect the ways in which perception, visibility, and cultural familiarity shape participation in citizen science. Rather than attempting to eliminate these biases entirely, it may be more productive to recognize them as part of the situated nature of observation, and to design complementary strategies that capture less visible taxa.

Despite these challenges, our findings demonstrate that monitoring insect–plant interactions does not require highly specialized expertise or equipment (Lewinsohn et al. 2022). Through guided observation and training, participants were able to document ecological interactions and identify generalist pollinators across multiple plant families. This suggests that citizen science can play a key role in democratizing ecological knowledge, making it accessible and actionable for non-specialists. In agricultural contexts, this has important implications for teaching–learning processes, as it enables farmers to reinterpret ecological interactions and potentially adjust management practices (Puig 2021; Curilén and Luciano 2023).

From a broader perspective, the Agri Andes Ecuador project illustrates how citizen science can contribute to transformative processes in socioecological systems. Rather than simply generating data, the project facilitated the creation of a community of practice, supported by training activities and institutional partnerships. However, the limited participation of farmers indicates that further work is needed to strengthen engagement and ensure that citizen science initiatives are inclusive and context-sensitive.

Overall, integrating citizen science into agricultural systems requires moving beyond a purely technological or data-driven approach. It involves recognizing citizen science as a social, epistemic, and ecological process, where participation, learning, and knowledge production are intertwined. In this sense, the potential of citizen science in agroecology lies not only in its ability to monitor biodiversity, but also in its capacity to reshape how people perceive, value, and manage ecological interactions within agricultural landscapes.

Despite active agroecological transitions, ecological data remain scarce in Ecuadorian agroecosystems and elsewhere around the world, where citizen science projects are crucial to documenting non-crop biodiversity. Engaging farmers and communities can strengthen monitoring capacities, democratize knowledge, and inform policies that link food sovereignty, sustainability, and biodiversity conservation to the country’s broader social and economic transformations (Saenz-Lituma 2025).

Conclusions

Considering the impacts of agriculture on biodiversity, citizen science applications such as iNaturalist hold considerable potential for expanding ecological monitoring within cultivated landscapes. Within the Agri Andes Ecuador project, the development of an active community of naturalists was supported through training activities and public events, even under the constraints of the COVID-19 pandemic, which likely limited participation and slowed project implementation. Beyond generating biodiversity records, our findings indicate that citizen science functioned as a process of knowledge co-production, enabling participants to observe, interpret, and engage with ecological interactions, particularly plant–insect relationships within agricultural contexts. This suggests that citizen science can contribute not only to data generation but also to shaping how ecological processes are understood by different actors.

However, participation was uneven, with limited engagement from farmers, highlighting the need for more context-sensitive approaches that address barriers such as connectivity, digital literacy, entomoliteracy, and time availability. Tools such as the Seek application showed potential to lower these barriers by enabling participation among users with limited experience using mobile technologies. The project’s key strengths include the integration of naturalist training into university curricula, and the presence of a national iNaturalist node, both of which supported sustained engagement and a sense of shared ownership across geographically distributed participants. Bioblitz events were effective in increasing observations and fostering interaction, although more standardized protocols may improve data consistency and long-term engagement. Overall, this study highlights the potential of citizen science to link biodiversity monitoring, social learning, and environmental education, while also underscoring the importance of designing inclusive and context-adapted approaches for agricultural systems.

Supplementary Files

The supplementary files for this article can be found as follows:

Data Accessibility Statement

The data are summarized in Agri Andes Ecuador. The analyzed data is available directly from the Agri Andes Ecuador project website.

Ethics and Consent

Our team confirms that training included a presentation of the iNaturalist platform, where participants agree to terms of use and data-sharing policies upon registration, constituting informed consent. Only publicly available data were used, and no personal information was accessed or stored. Participation of schoolchildren and farmers was voluntary. No data or information that could personally identify participants were collected.

Author Contributions

All authors conceptualized the idea and defined the questions and methodological approach. Diego Mina carried out the analyses, assisted by Olivier Dangles and Rommel Montufar, and led the writing. Mayra Coro actively contributed to the workshops/bioblitz and data collection, especially data on plant-insect ecological relationships. All authors contributed to the writing, participated in the discussion and agreed on the final version of the manuscript.