Capitalizing on opportunistic citizen science data to monitor urban biodiversity: A multi-taxa framework

Highlights

• Understanding species-specific responses to urbanization is crucial for conservation of urban biodiversity.

• Citizen science data are valuable for monitoring urban biodiversity.

• We provide a framework for tracking and monitoring urban biodiversity across multiple taxa.

Abstract

Monitoring urban biodiversity is increasingly important, given the increasing anthropogenic pressures on biodiversity in urban areas. While the cost of broad-scale monitoring by professionals may be prohibitive, citizen science (also referred to as community science) will likely play an important role in understanding biodiversity responses to urbanization into the future. Here, we present a framework that relies on broad-scale citizen science data –– collected through iNaturalist –– to quantify (1) species-specific responses to urbanization on a continuous scale, capitalizing on globally-available VIIRS night-time lights data; and (2) community-level measures of the urbanness of a given biological community that can be aggregated to any spatial unit relevant for policy-decisions. We demonstrate the potential utility of this framework in the Boston metropolitan region, using >1000 species aggregated across 87 towns throughout the region. Of the most common species, our species-specific urbanness measures highlighted the expected difference between native and non-native species. Further, our biological community-level urbanness measures –– aggregated by towns –– negatively correlated with enhanced vegetation indices within a town and positively correlated with the area of impervious surface within a town. We conclude by demonstrating how towns can be ‘ranked’ promoting a framework where towns can be compared based on whether they over- or under-perform in the urbanness of their community relative to other towns. Ultimately, biodiversity conservation in urban environments will best succeed with robust, repeatable, and interpretable measures of biodiversity responses to urbanization, and involving the broader public in the derivation and tracking of these responses will likely result in increased bioliteracy and conservation awareness.

Introduction

We are currently facing the 6th mass extinction event in the Anthropocene, and biodiversity is increasingly at risk from various anthropogenic pressures (Ceballos and Ehrlich, 2002). Monitoring how biodiversity responds to both threats (e.g., pollution, habitat loss, invasive species, climate change, and other anthropogenically-derived pressures) as well as interventions for enhancement (e.g., habitat restoration, green infrastructure) is essential to understand how best to preserve and manage our collective biodiversity. Biodiversity plays a key role in regulating ecosystem processes, and as acts as an ecosystem service in itself, subject to valuation (Mace et al., 2012). This, combined with the increased recognition that human well-being is positively linked with increased biodiversity highlight the necessity of monitoring changes in biodiversity (Davies et al., 2019). But current funding for conservation science is failing to keep pace with the increased necessity to fully understand and monitor biodiversity change in response to varied anthropogenic pressures (Bakker et al., 2010). So, how then can we monitor biodiversity cost-effectively, with the aim of understanding how biodiversity responds to anthropogenic changes?

Broad-scale citizen science or community science projects likely provide necessary data to monitor biodiversity into the future (Bonney et al., 2009; Chandler et al., 2017; McKinley et al., 2017). Citizen science –– the collaboration between members of the public, regardless of citizen status in a particular jurisdiction, with professional scientists –– projects are increasingly used in natural resource management, ecology, and conservation biology (McKinley et al., 2017). And the number of such projects is simultaneously increasing (Pocock et al., 2017). For example, citizen science data have been used to increase the accuracy and specificity of threat levels of endemic birds in the Western Ghats (Ramesh et al., 2017), identify the important role temperature plays in sexual coloration in a dragonfly (Moore et al., 2019), identify new records and range extensions (Rosenberg et al., 2017), and quantify biodiversity changes in space and time (Cooper et al., 2014). These are only a select few examples. Despite the increasing prevalence of citizen science data (Pocock et al., 2017), there is still reluctance to fully adapt such data in wide-spread monitoring of biodiversity (e.g., Burgess et al., 2017). This is, in part, likely due to the biases generally associated with citizen science data (Boakes et al., 2010). Such biases include increased sampling on weekends (Courter et al., 2012), taxonomic preferences for ‘charismatic’ fauna and flora (Ward, 2014), and generally skewed data collections to areas with large human populations (Kelling et al., 2015). This latter bias is generally problematic for any citizen science project with semi-structured or unstructured data collection (Kelling et al., 2019).

While this sampling bias towards urban areas can limit our inferences surrounding biodiversity in natural, remote regions (Callaghan et al., 2020a), it offers opportunities to better understand urban ecological and conservation questions (Cooper et al., 2007) and can complement biases ecologists have in sampling predominantly protected areas (Martin et al., 2012). Indeed, citizen science data have recently been leveraged to understand many aspects of urban ecology (e.g., Boukili et al., 2017; Li et al., 2019; Leong and Trautwein, 2019). And citizen science data may provide a relatively cost-effective method to monitor biodiversity in urban areas (Callaghan et al., 2019b), including private lands which are often only accessible to private landowners (e.g., Li et al., 2019). This is critical, given the fact that urbanization is an intense anthropogenic pressure, and habitat loss and fragmentation associated with urban land transformation has generally negative impacts on biodiversity (Cincotta et al., 2000; McKinney, 2006). Further, the importance of fully using citizen science data in urban areas is made clear because: (1) urban areas are where many people experience nature, and thus involving urban residents in citizen science projects can have flow-on effects for conservation (Lepczyk et al., 2017), because people are more likely to take conservation action when they have direct experiences with nature (i.e., the pigeon paradox; Dunn et al., 2006); (2) citizen science biodiversity research provides education benefits to participants (Jordan et al., 2011) with the potential to increase bioliteracy, benefitting biodiversity inside and outside of cities (Ballard et al., 2017); (3) urban areas can act as vessels for conservation (Dearborn and Kark, 2010); and (4) urban areas can even protect threatened species (Ives et al., 2016).

Given the importance of understanding urban biodiversity, and the potential for citizen science data to enhance this understanding and increase bioliteracy, the use of citizen science data needs to be validated to better understand how these data can be used in future monitoring of urban biodiversity. By increasing the bioliteracy of participants in citizen science projects a positive feedback cycle can be initiated, leading to an increase in the quality of the data (i.e., people become better at identifying and finding specific species) as the project continues. Many people have quantified the relationship between citizen science data and ‘professional’ data (Kosmala et al., 2016; Aceves-Bueno et al., 2017). But most comparisons have been from semi-structured citizen science datasets (e.g., eBird). Opportunistic citizen science projects (e.g., iNaturalist) likely have their own sets of biases (Brown and Williams, 2018), but are showing promise in helping to deduce patterns of biodiversity across urban environments (Leong and Trautwein, 2019; Li et al., 2019). The development of repeatable and robust methods that harness the power of citizen science data may not only help monitor biodiversity responses to urbanization but potentially help bridge the translation gap from science to urban planning and conservation action (Norton et al., 2016).

iNaturalist is one of the most popular global biodiversity recording platforms with over 33 million observations of 250,000 species made by more than 800,000 observers. Moreover, data from iNaturalist is the second most downloaded source of data from the Global Biodiversity Information Facility. Here, we use opportunistic (i.e., generally collected in an unstructured format) iNaturalist data from the metropolitan region of Boston, USA to detect and understand patterns in biodiversity across an urban to rural gradient. Urban environments differ from natural landscapes in many ways, and efforts to understand these differences (e.g., land use, fragmentation, disturbance) often rely on land use analyses (e.g., Pearse et al., 2018; Li et al., 2019; Leong and Trautwein, 2019). A global dataset of night-time lights has allowed for an approach to analyze the response of organisms to urbanization on a continuous scale, and has thus far been used to understand patterns in urban bird biodiversity at local and regional scales using eBird citizen science data (Callaghan et al., 2019a, Callaghan et al., 2019b, Callaghan et al., 2020b). Here we look to test whether opportunistically-collected iNaturalist data can similarly help to detect patterns in biodiversity across urbanization gradients, scaling from species-specific responses to town-specific measures of the urbanness of the biological community within that town. Our approach highlights how directed efforts of sampling such as the City Nature Challenge hold potential for building both a robust dataset to understand patterns of biodiversity responses to urbanization and increase public awareness of surrounding urban biodiversity.

First, we assess the sampling biases of participants contributing opportunistic citizen science iNaturalist observations, as it pertains to a continuous gradient of urbanization –– defined using night-time lights –– available to sample across. We hypothesized that the degree of urbanization in a town would be positively correlated with the degree of urbanization of the observations in that town (i.e., more urban towns would have more urban observations). We then use these citizen science data to assign species-specific measures of urban tolerance, defined as the median night-time lights value of all observations for a species. From this, we produce town-specific measures of the urbanness of the collective species found therein, defined as the median of all species-specific measures of urban tolerance. We hypothesized that the relationship between the underlying degree of urbanization in a town and the cumulative town-specific urban tolerance of the species found therein would be positively correlated. We then demonstrate how these town-specific measures of urbanness can be used in an ecological context by showing the relationship between the town-specific urbanness and its ecological attributes (i.e., tree cover, impervious surface, and enhanced vegetation index). And lastly, we provide a forward-looking approach to compare individual planning units (e.g., towns) among one another in regards to the "urbanness" of their biodiversity. Ultimately, we highlight a framework that is robust and uses globally-available datasets (i.e., VIIRS night-time lights and iNaturalist citizen science data) to better understand how to fully realize the potential of citizen science data to understand urban biodiversity. Because of the ubiquity of iNaturalist data in cities and availability of a global night-time lights data set, we expect this approach can be successfully applied to increase awareness of and manage urban environments worldwide.