Citizen science delivers high-value biosecurity surveillance and reporting capability

Introduction

Effective mitigation of invasive species centres on our capacity to prevent, detect and respond to their spread in a timely manner. Invasive species are those alien or native species that, when introduced, become established and cause harm to human and environmental values (Diagne et al. 2021). Early detection of an invasive species of biosecurity concern entering the environment is a crucial capability of an effective biosecurity system, and often essential if eradication is to be possible. Australia is a global leader in biosecurity, having some of the world’s most stringent and recognisable biosecurity measures. They are particularly exemplified by strong national border-control systems including biosecurity at points of entry. These elaborate and robust systems recognise that the heavy cost of incursions to agriculture, forestry, health, environment, and Aboriginal and Torres Strait Islander peoples’ connection to Country is significant, and has been widely demonstrated (Hoffmann and Broadhurst 2016; Carnegie and Nahrung 2019). Australia is home to some of the most consequential introduced species worldwide because of the vulnerability of our ecosystems and a history of both accidental and deliberate introductions. The introduction of rabbits (Oryctolagus cuniculus), European carp (Cyprinus carpio), cane toads (Rhinella marina) and Paterson’s curse (Echium plantagineum) have cost billions in impact and millions in control measures (Hoffmann and Broadhurst 2016; Bradshaw et al. 2021). The risk of new incursions increases annually as a consequence of increasing volumes of trade, people movement, changes in land use, climate change, and the expanding spread of invasive species around the globe. Once established, the spread of invasive species is very difficult to curtail (Hulme et al. 2023).

Participatory science, or citizen science, spans a broad range of scientific disciplines, from chemistry to space to human health (Pocock et al. 2017; Roger et al. 2023). However, the majority of citizen-science projects globally are biodiversity related (Theobald et al. 2015). This holds true in Australia and translates to citizen scientists contributing the fastest-growing source of open-source biodiversity data in Australia (Roger et al. 2023). Despite the potential for citizen science to benefit environmental biosecurity, such as first-in-country reporting of exotic invasive species (Welvaert and Caley 2016; Poland and Rassati 2019; Douch and Poupa 2021; Pocock et al. 2024), its contribution is much less than that of native species reporting (Meentemeyer et al. 2015). Many biosecurity agencies globally are adopting the narrative that biosecurity is a ‘shared responsibility’, but there are few concrete ways to engender mass participation (Diprose et al. 2022). Given Australia’s expansive citizen-science ethos and national commitment to a strong biosecurity system, leveraging already engaged biodiversity-focused citizen-science communities in environmental biosecurity is a natural progression (Probert et al. 2022).

Enabled in part through technology, there is potential for citizen science to contribute to monitoring the spread of species already in Australia and reporting new incursions. Indeed, Carnegie and Nahrung (2019) found that 71% of detections of exotic forest species in Australia occurred via general rather than active surveillance. That 71% consisted of 24% detections made by the public, compared with 35% by researchers and 12% by industry. Post-border, away from the first points of entry such as seaports and airports, there are often limited or rudimentary surveillance systems (Vall-llosera et al. 2017; Poland and Rassati 2019). As such, biosecurity departments often rely on public awareness and participation for the detection and reporting of exotic invasive species. Such biosecurity-targeted communications are typically directed at agricultural communities, despite the role urban communities can play. Urban environments are most likely where pests will be first found and where invasive species are more likely to thrive (Cadotte et al. 2017; Gaertner et al. 2017). An informed general citizen surveillance network has the potential to greatly contribute to Australia’s overall environmental biosecurity system (Douch and Poupa 2021). However, critical to demonstrating the value of citizen science to biosecurity is linking public contributions to management and decision-making, and addressing the accuracy of the data provided (Pocock et al. 2024).

Concerns around using citizen science in the context of environmental biosecurity arise due to criticisms mainly with observer error and uneven spatial coverage (Crall et al. 2010; Pocock et al. 2024). The consequences of a false positive observation can have sizeable implications for interstate and international trade; however, under normal conditions these reports can be quickly corrected. Whereas some studies have found that the public perform poorly at identifying pest species, others have reported that with the proper training citizens are capable of providing diagnostic information with high accuracy, as reviewed by Thomas et al. (2017). Meentemeyer et al. (2015) noted that techniques and approaches exist (such as training, statistical methods, and targeted projects) that reduce these concerns, and access to technology in the form of collection and submission tools can increase accuracy. Pocock et al. (2024) explored solutions to some of the main challenges with citizen science in detail and included training, verification at submission, analysis adjustment and good design. Ultimately, the value of a data set should be assessed in terms of its purpose, including its usage, timeliness, relevance, and accuracy, meaning that the utility of citizen science needs to be assessed dependent on its use-case. For environmental biosecurity, the system should be weighted towards minimising, or eliminating, false negatives. False negatives can leave an invasive species incursion undetected in the environment with the possibility of highly significant negative consequences, whereas a false positive would waste only the time taken to confirm the sighting was false. Systems that connect public observations with experts provide an element of oversight and quality-control integral for citizen science to reach its potential for biosecurity application (Douch and Poupa 2021). Crucially, these systems offer a platform for constant public review and amendment, which is a limitation of some physical collections. To prevent duplication and ensure a sufficient critical mass of users, projects need to either use existing tools or adopt approaches that leverage them. Platforms that can utilise big-data aggregation and pull data in from a variety of existing citizen-science infrastructure have much to offer the sector (Crall et al. 2010).

The Atlas of Living Australia (ALA) is Australia’s national open biodiversity data service. The ALA collates environmental data from over 850 data providers such as museums, herbaria, government monitoring programs, research projects, Indigenous knowledge, and citizen-science projects and platforms (Belbin et al. 2021). Citizen-science data are a critical and growing source of data for the ALA, with approximately half of all species occurrence records now being derived from citizen science (Roger et al. 2023). In 2023, the ALA formally established a Biosecurity Alerts Service. The service works by performing weekly queries of the ALA database for new occurrences matched against lists provided by biosecurity management agencies. Alerts can be geographically limited to national, state and territory, and local government areas, as well as within defined spatial areas such as national park estates. Users can also select specific genera for alerts and elect to exclude certain species within a genus. Critically, the majority of alerts sent are a result of citizen-science observations aggregated by the ALA.

Here, we detail the design and development of the Biosecurity Alerts Service of ALA. We provide information on how the service was developed and its impact (over a 12-month period). We provide links to the open-access code, and report on key metrics of success related to users of the service and species detected. Finally, we include case studies detailing how the alerts are being used and evaluate the alerts in terms of representativeness. Our aim is to present an example of how citizen science is being used for environmental biosecurity surveillance in Australia and showcase the value of large data aggregators such as the ALA, as a connector between citizen science and decision-making.

Methods

Results

Number of users

After a period of 12 months, the ALA sent biosecurity alerts to a total of 32 users. Users who provided these lists were distributed across Commonwealth, state and territory agriculture and management agencies, and a single local government authority (Fig. 1).

Fig. 1.

Number of users from each jurisdiction.

A total of 1442 unique records triggered biosecurity alerts between January 2023 and December 2023 inclusive. Arthropods were the most common phylum detected, representing >50% of alerts (Fig. 2).

Fig. 2.

Number of alerts per phylum, of species from user lists, that occurred from January 2023 to December 2023.

The majority of the alerts were generated from users of the global citizen-science platform iNaturalist, followed by the Australian-built reporter apps FeralScan and NatureMapr, with remaining contributions from local area-provided datasets (Fig. 3).

Fig. 3.

Data sources per data provider that generated user notification alerts. iNaturalist is the data provider generating most of the occurrence records for the alerts.

So as to assess the proportion of taxa being alerted on relative to the total number of taxa submitted by users on each list, we amalgamated all listed taxa into one list. We were then able to compare the number of taxa the ALA issued an alert for, against the total number of taxa on the amalgamated list (Fig. 4). However, it needs to be noted that although a few lists contain bacteria and viruses (Fig. 4), users seem to be excluding them from their provided lists by default, perhaps recognising the limitations of the ALA being able to issue an alert on them.

Fig. 4.

Representativeness of the biosecurity alerts comparing the number of taxa the ALA issued an alert for (list %) against the total number of taxa on an amalgamated list for all taxa (occurrence %).

The Biosecurity Alert Service is being used in a variety of ways (outlined below). Here, we have provided case studies of how land management agencies have used the information from one occurrence record.

Discussion

Preventing the spread or introduction of invasive species is usually considered the most cost-effective management option (Hoffmann and Broadhurst 2016; Vall-llosera et al. 2017). This is best achieved through pre-border and post-border surveillance. The early detection of an invasive species that has entered the environment is crucial for limiting invasive species impacts and provides the best opportunities to eradicate the incursion. Vall-llosera et al. (2017) found that the majority of reports concerning the invasive rose-ringed parakeet (Psittacula krameria) in Australia were not reaching agencies with an oversight or legislative responsibility for intervention. Vall-llosera et al. (2017) suggested that there is a disconnect between communities and biosecurity agencies. Linking observational records made by the public on open access platforms to biosecurity agencies greatly expands the capacity of surveillance to detect and report invasive species throughout Australia. For example, the biosecurity alerts are alerting management when detections of red imported fire ants (Solenopsis invicta) are detected and thus could play a critical role in reporting southerly incursions. Similarly, the Service is alerting management authorities only when cane toads (Rhinella marina) are recorded outside of their known distribution, to assist management in active control areas along the Queensland–New South Wales border. The Biosecurity Alerts Service connects community observations with decision-making with good representation from biosecurity agencies as users help resolve this disconnect (Fig. 1). The next phase of the Biosecurity Alerts Service will involve more active promotion and communication of the benefits of the Service, including expansion into more Local Government areas, state and federally managed national parks and forests, and increasing representation from states and territories such was the Northern Territory and South Australia.

The majority of biosecurity alerts were generated from the iNaturalist platform (Fig. 3). iNaturalist is the most popular biodiversity-recording platform in Australia in terms of the number of users (Mesaglio and Callaghan 2021). Use of iNaturalist in the context of biosecurity has a few specific benefits. First, iNaturalist is driven by a machine-learning algorithm that provides suggested identifications based on location and image. This algorithm is trained on a global database of species images; even if a species is not present in Australia, it is still likely to be identified within the platform. The algorithm also helps improve the quality of identifications. As the number of correct identifications increases for a species, this algorithm will continue to improve because of the additional image-training data. Further, every iNaturalist observation can be viewed by an online community of experts (including knowledgeable amateurs), meaning that often there is someone on the platform familiar with organisms that are foreign pests and diseases in Australia (Diprose et al. 2022). iNaturalist also has a large community of users; the resulting large volume of observations warrants a weekly data ingest by the ALA, and therefore increases the likelihood that observations will reach management authorities on a timely basis, which can increase the probability of eradication (Byers et al. 2015). This was most recently displayed when the ALA sent out an alert for a highly invasive species, the western conifer seed bug (Leptoglossus occidentalis), to management authorities. On receiving the alert, management authorities got in contact with the observer via the platform, requesting further information (including a specimen) and were looking into the supplier of the shipped product the species was recorded in. The species has not been confirmed, because a specimen was not able to be collected; however, the user is on alert to assist authorities should one be recorded in the future. The specimen recorded in the shipment was dead, and there is no indication that it had spread.

Although the ALA receives data from more than 850 data providers, citizen science represents the bulk of the alert observations (Fig. 3). This is because the volume of data received from non-citizen-science data providers, as well as limitations (such as staff capacity) from data providers, result in these records typically not being able to meet the time-critical nature required for biosecurity alerts (recorded in the past 30 days). The current array of biodiversity and biosecurity reporter apps and datasets in Australia provides some redundancy in the reporting system and also creates unwanted complexity. The primary contribution of the ALA stems from the aggregation of data from these platforms, which helps address the siloed and fragmented nature that is often characteristic of biosecurity datasets (Aronsson et al. 2023). The value of the alerts service is that it makes it easy for biosecurity agencies to become aware of the diversity of records and taxa being reported (Fig. 2) and, thereby, enables them to make a time-critical response if required. This is a powerful and beneficial use of near real-time citizen-scientist and public surveillance. A central element of a successful network of volunteers, such as the citizen-science cohort, is having a sense of community and shared purpose that is recognised and actively supported. Mechanisms to facilitate the building and maintenance of the citizen-science reporter community will be important for the further development of the biosecurity surveillance system (Probert et al. 2022). Presently, each email encourages users to use the citizen-science platforms to reach out to contributors, acknowledging the value of their observations. We are also planning a large outreach campaign to communicate the effectiveness of the system, to further acknowledge the contribution.

Biosecurity agencies that receive the alerts are using the information in several ways. First, they are reviewing the emails, sending observations for additional scrutiny and verification (if required) by internal experts, and undertaking management action as required. Management action has included undertaking site visits to confirm the reported sighting, eradication (where applicable) and targeted sampling (often with additional community involvement). Second, they are contacting platform users directly to request additional information or, at times, a physical sample if possible. Third, they are correcting false positive identifications on citizen-science platforms. This serves the dual purpose of improving the data associated with species of concern and helping users with improved verification and identification of target species, thereby increasing community expertise. Such interaction within the online platforms is likely to have a positive impact on citizen scientists as they can see the benefits of their engagement. Demonstrating impact from citizen science is a critical element of successful citizen-science programs because it provides recognition of the citizen scientists’ contribution and involves them in the objectives of the system (von Gönner et al. 2023).

The establishment of the ALAs Biosecurity Alerts Service occurred simultaneously with other alert tools overseas. These include an early warning system developed by the Finnish Biodiversity Information Facility (FinBIF), which provides alerts to biosecurity coordinators at two different authorities when an occurrence record of biosecurity concern enters the data warehouse (Morris and Piirainen 2023). Similarly, a tool initially developed for Belgium LIFE RIPARIAS is an early alert system that harvests records from the Global Biodiversity Information Facility (GBIF) and allows users to visualise records on a map and view the original report. Users can create email alerts, so that they are notified of certain species on the basis of a set checklist of available species (Noé et al. 2022). The ALA Biosecurity Alert Service differs from these examples through active partnership with management agencies. The ALA generates alerts on the basis of management agency species lists (via a curation and species matching process) and tailors the alert email received to management needs such as a set spatial area of interest. These steps have linked management agencies and open-access environmental biosecurity data.

Despite the success of the biosecurity alerts, there remain some noted limitations with what can be delivered. We looked at representativeness to gain an understanding of how well the alerts were performing. High detections of arthropods reflect the high number of arthropods on user’s alert lists (Figs 3, 4). Although we found that the alerts were overall representative of taxonomy, Chordata was over-represented as well as was Basidiomycota. In contrast, Ascomycota was under-represented in reports compared with frequency of alert list taxa. There are a few users who included bacteria and viruses on their list of species of interest; however, the majority of species on the lists are for plants, animals and fungi (Fig. 2). We believe this is because users are aware that reports from citizen scientists are essentially limited to species that can be observed with the naked eye (or acoustic recording), and therefore we see negligible inclusion and reporting for viruses, bacteria and protozoa (see Caley et al. 2020 for a discussion on how physical features predict reporting probability). We are unable to determine whether there are more users who would like to be notified of these taxa but have excluded them from their lists because they are aware of the limitations. It would be interesting to explore how users are choosing the range of taxa on their provided lists. Another issue is how host–pathogen information is currently displayed. Presently, there is no easy way of associating a disease such as myrtle rust caused by Austropuccinia psidi with a record of a host species (a plant in the Myrtaceae family). Both host and pathogen records are equally important to researchers and management. We also know that we have many images of wildlife with disease such as sarcoptic mange, but currently have no way to differentiate the host animals from a record of a pathogen. Presenting this information in a digestible format is a future priority.

Consistent taxonomy has long been recognised, if undervalued, as a fundamental requirement of biosecurity (Wittenberg and Cock 2001). We found consistent taxonomy to be an essential element to a successful alerts system based on name-matching. Biosecurity agencies and international organisations such as the Centre for Agriculture and Bioscience International (CABI) or the Global Register of Introduced and Invasive Species (GRIIS) compile biosecurity species lists. Although often internally consistent, these include synonyms and invalid name combinations as an inevitable consequence of being practical management lists (Pagad et al. 2018). Furthermore, name synonyms are often lost in global databases when taxonomic backbones (a hierarchy of levels used to arrange scientific names) are updated. This makes the process of list-matching to aggregated datasets curated on the basis of taxonomic name databases challenging. Careful editing of user lists is essential to ensure consistent interpretation for alert purposes. For example, the invasive freshwater mussel Corbicula (Corbicula) fluminea is listed as a priority biosecurity species in Australia. Corbicula (Corbicula) maroubra was a putative Australian species described in 1943 that was in Australian checklists. Initially, our name-matching was assigning C. fluminea as a junior synonym to C. maroubra on the basis of Australian checklists. It was only manual detective work that established that C. maroubra was the junior synonym and that there had been several historical C. fluminea incursions in the 1940s that had gone unnoticed by biosecurity agencies because of the taxonomic discrepancy. Other issues include the issue that national name checklists may not include international species or international checklists can sometimes use alternate taxonomic hierarchies. For example, the house crow (Corvus splendens) is regarded as an invasive species in Australia. Although the taxonomy is straight-forward, adding a species to the taxonomic hierarchy used in the ALA resulted in the name being placed against Corvidae automatically (because the species is not found in Australian checklists). This had the unintended consequence of initially showing all Corvidae records as invasive alerts, until the name was manually placed in the hierarchy.

Acquiring additional datasets to enhance the siloed and fragmented nature of biosecurity data remains an ongoing priority. However, many management agencies rightly express concern about making their data publicly available. This can be for a myriad of reasons, including concerns over hunting, records on private land, or a highly traded sensitive species being made public (Roger et al. 2023). The next step to addressing these concerns will be establishing improved governance and systems to manage sensitive datasets, which will ensure that something is discoverable via appropriate metadata, but accessibility is limited via an approvals system. Gridding of sensitive data as opposed to obfuscation or truncation will also help resolve issues around private land (Roger et al. 2023). In 2023, Australia released a national framework for sharing restricted-access data (Atlas of Living Australia 2023), which is motivating large biodiversity-data infrastructures to also establish frameworks for managing sensitive species data. We see managing environmental biosecurity data as an extension of the process detailed in this framework.

Conclusions

Although citizen science has exploded in terms of participation and data production over the past decade (Chandler et al. 2017), presenting evidence of impact in the form of informing decision-making and management remains challenging (von Gönner et al. 2023). However, so as to maintain motivated and enthused citizen-science communities, evidence of impact in terms of the value to science and environmental protection is essential (Roger and Kinsela 2023; von Gönner et al. 2023). The biosecurity alerts are providing a method to link citizen-science data with management and biosecurity agencies. As a result, there is building awareness of the value of publicly generated and accessible data, and the benefit of adopting more flexible surveillance approaches (Pocock et al. 2024). Given that biosecurity challenges are only set to increase over the coming years (Diprose et al. 2022; Hulme et al. 2023), recognising that communities can play an increasingly important role in surveillance will be crucial. It is the coupling of citizen science with new technologies such as artificial intelligence (Ceccaroni et al. 2019), drones and sensors (Schacher et al. 2023) that will likely have the largest impact in expanding biosecurity-surveillance capability over the medium term. Equally important will be the need to build and support reporting communities and ensure that their contributions to biosecurity and biodiversity conservation are appropriately acknowledged.

Supplementary material

Supplementary material is available online.