Challenges and benefits of using unstructured citizen science data to estimate seasonal timing of bird migration across large scales

PLOS ONE RESEARCH ARTICLE Challenges and benefits of using unstructured citizen science data to estimate seasonal timing of bird migration across large scales Nadja Weisshaupt ID1*, Aleksi Lehikoinen2, Terhi Mäkinen1, Jarmo Koistinen1 1 Finnish Meteorological Institute, Helsinki, Finland, 2 Finnish Museum of Natural History, University of Helsinki, Helsinki, Finland a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 OPEN ACCESS Citation: Weisshaupt N, Lehikoinen A, Mäkinen T, Koistinen J (2021) Challenges and benefits of using unstructured citizen science data to estimate seasonal timing of bird migration across large scales. PLoS ONE 16(2): e0246572. https://doi. org/10.1371/journal.pone.0246572 Editor: Sergio Rossi, Universite du Quebec a Chicoutimi, CANADA Received: October 20, 2020 Accepted: January 21, 2021 Published: February 4, 2021 Copyright: © 2021 Weisshaupt et al. This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited. Data Availability Statement: Data cannot be shared publicly because it is third-party data. Data underlying the results presented in the study are available from the online bird portal Tiira (https:// www.tiira.fi/) with prior consent of BirdLife Finland (https://www.birdlife.fi/), and the Museum of Natural History, Helsinki, Finland (https://luomus. fi). The authors did not receive special access privileges to the data that others would not have. Funding: NW, JK and TM have received funding as part of the GloBAM project (https://globam. * Abstract Millions of bird observations have been entered on online portals in the past 20 years either as checklists or arbitrary individual entries. While several hundred publications have been written on a variety of topics based on bird checklists worldwide, unstructured non-checklist observations have received little attention and praise by academia. In the present study we tested the suitability of non-checklist data to estimate key figures of large-scale migration phenology in four zones covering the whole of Finland. For that purpose, we analysed 10 years of ornithological non-checklist data including over 400 million. individuals of 115 bird species. We discuss bird- and human-induced effects to be considered in handling nonchecklist data in this context and describe applied methodologies to address these effects. We calculated 5%, 50% and 95% percentile dates of spring and autumn migration period for all species in all four zones. For validation purposes we compared the temporal distributions of 43 bird species with migration phenology from standardized long-term ringing data in autumn of which 24 species (56%) showed similar medians. In a model approach, nonchecklist data successfully revealed latitudinal migration progression in spring and autumn. Overall, non-checklist data proved to be well suited to determine descriptors of migration phenology in Northern Europe which are challenging to attain by any other currently available means. The effort-to-yield ratio of data processing was commensurate to the outcomes. The unprecedented spatiotemporal coverage makes non-checklist data a valuable complement to current migration databases from bird observatories. The basic concept of the present methodology is applicable to data from other bird portals, if combined with local field ornithological knowledge and literature. Species-specific descriptors of migration phenology can be potentially used in climate change studies and to support echo interpretation in radar ornithology. Introduction Technological advances in the past 20 years have enabled the creation of online bird portals dedicated to the collection of casual daily visual and acoustic field observations. Prominent PLOS ONE | https://doi.org/10.1371/journal.pone.0246572 February 4, 2021 1 / 17 PLOS ONE science/). This project is funded through the 20172018 Belmont Forum and BiodivERsA joint call for research proposals (https://www.biodiversa.org), under the BiodivScen ERA-Net COFUND programme, and with the funding organisations Swiss National Science Foundation (SNF 31BD30_184120), Belgian Federal Science Policy Office (BelSPO BR/185/A1/GloBAM-BE), Netherlands Organisation for Scientific Research (NWO E10008), Academy of Finland (aka 326315) and National Science Foundation (NSF 1927743). The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. Competing interests: No authors have competing interests. Migration phenology from unstructured citizen science data examples of such online databases are the Ornitho, Observation and BirdTrack platforms in Europe (for an overview see [1]) or the US-American eBird. The general idea of these portals is to report sightings of birds, with species, number of birds and location as minimum information required and optionally additional details, such as the birds’ age, sex, behaviour, observation time and so on. Through the widespread network of observers, millions of observational reports accumulate across vast areas throughout the year, providing an unprecedented trove of information. This wealth of observations could be potentially used for scientific research about avifaunal dynamics, as a complement to traditional monitoring programs through bird surveys and ringing by volunteers and scientists [2]. Despite more than 300 publications based on eBird data (https://ebird.org/about/publications/), the use of such bird repositories other than eBird, is still hesitant in the scientific community. Hesitation to use these data portals is often linked to quality concerns and data availability [3, 4], but probably also with unfamiliarity with and unawareness of the data portals. Quality concerns relate mainly to variable observer expertise and heterogeneity in data collection, socalled observer bias, which can lead to false positive or negative records [2, 5]. Observer bias is a well-known issue in observational field data, also those collected by professionals, and can originate from the observers’ level of ornithological knowledge and experience, species attractiveness (e.g. rare vs. abundant species, first appearance in spring), hearing and visual capacity, motivation and dedication depending on environmental conditions and so on [6–8]. Theoretically, many of these observer biases can be controlled prior to data analysis by a careful study design, preparation, and training of the samplers before the field work starts [9]. Various publications and manuals address potential and actual pitfalls of using citizen science data in general, also outside ornithology, and instruct on how to obtain homogeneous data and reliable results [10, 11]. However, in case of bird portals, where data is gathered continuously and unawarely of some potential research in (...truncated)