Building a Database using Unconventional Sources: Squirrels of India

Squirrels, like most other small mammals, have been poorly documented in the Indian subcontinent, which deters us from understanding species declines and prioritizing research on sensitive taxa (McKinney 1999, Koprowski Nandini 2008). They are diverse their habits morphology perform important ecological roles, including seed dispersal, pollination, regulating plant growth. also form a significant prey base for many predators. Squirrels respond strongly to pressures around them, urbanization, habitat modification, climate change, can thus act as model study systems (Sol et al. 2013). The first step how such changes is by gaining knowledge of where they occur. rapid spread internet connectivity access mobile technology across India allow large-scale secondary data ways that were not possible decade ago. We created pan-India database 30 squirrels using primary (from fieldwork) sources (museum records, published gray literature), citizen science portals (six sources), social media platforms (14 sources). use increasing exponentially India, yet these remain largely unexplored source harvesting biodiversity information. Given low public awareness squirrel species, we expected high error rates with contributors' assignment identity. A key images calls was used while gathering maintain consistency team. pipeline collection curation created, all volunteers project trained collection. To ensure verification identity, (photographs, audio, video data) collected when or cross-checked site. Some (iNaturalist, Project Noah) script-based search-based downloads bulk records without media. Each record sites manually checked confirm identification. On (Fig. 1), species-wise searches performed (using common scientific names) within each platform. For (photograph, video) downloaded along location, date, observer, relevant notes. entry entered into researchers (12 over two years, volunteer interns). then verified one four curators (more less-familiar species). Duplicates removed curator, who compares species-specific multiple sources. location curated, georeference added unavailable original post. All imported Google Sheets, map tool Geocode Awesome Table obtain latitude longitude places. occasions, contacted observers details before an finalized. Over grew include 24,170 approximately 14,000 files, team working 2200+ hours. About 48% (12,035) occurrence came sources, followed 30% (7375) traditional 22% (4660) 1). examined temporal trends bias three assess under-representation based body size, activity period, body-color, International Union Conservation Nature (IUCN) Red List status, range type. majority tree 2), flying ground squirrels. This likely because diurnal more abundant, hence easier compared cryptic nocturnal. restricted higher elevations Himalayas, making them difficult record. There are, however, differences regions India. Based quick examination urban areas, reflecting either (concentrated human densities) response urbanization. Funambulus palmarum pennantii known be abundant areas densities, might reflected number records. In contrast, seem less anthropogenic disturbance. Therefore, recording fine-scale occurrences this group crucial understand species' responses landscape modifications Our changing world has greatly improved combining, harmonizing, analyzing large amounts heterogeneous (Hampton availability accurate enables studies address questions at increasingly spatial scales stronger inference predictive models, which, turn, yield biological insights (Lewis 2018).