To boldly go where no volunteer has gone before: Predicting volunteer activity to prioritize surveys at the landscape scale

Ayesha I T Tulloch, Karen Mustin, Hugh P. Possingham, Judit Szabo, Kerrie A. Wilson

    Research output: Contribution to journalArticlepeer-review

    90 Citations (Scopus)

    Abstract

    Aim: To identify the relationships between volunteer bird survey effort and motivations in order to prioritize investment in future surveying activities. Location: South-west Western Australia, a global biodiversity hotspot.

    Methods: We developed nine hypotheses for volunteer motivations to predict the probability of a bird survey being undertaken anywhere in the landscape using data from the New Atlas of Australian Birds. We then established three goals for surveying in the study region: (1) equal representation of surveys across the landscape, (2) surveys stratified by habitat type and (3) representation of surveys in protected areas. We developed a function to estimate the benefit of investing in professional surveys, given the probability of a volunteer survey taking place and the survey goal, and calculated the cost of meeting a surveying goal with and without accounting for the probability of cells not being surveyed by volunteers.

    Results: A model combining the location of protected areas, location of previous records of threatened species and habitat diversity was the strongest predictor of the probability of a volunteer bird survey being conducted. Each surveying goal resulted in different areas being prioritized for future surveying, indicating the importance of setting clear objectives before undertaking broad-scale monitoring or surveying activities. If our primary goal is stratified protected area representation in surveys, there are huge cost savings if only protected areas with a 70% predicted probability of not being surveyed by volunteers are selected for professional surveys.

    Main conclusions: Professional sampling in survey gaps is required to reduce bias in volunteer-collected datasets. Using models of volunteer behaviour, we can identify areas unlikely to be surveyed. If these areas are important for the project objective, then we can either provide incentives for volunteers or carry out professional surveying. These analyses are best carried out before data collection commences.

    Original languageEnglish
    Pages (from-to)465-480
    Number of pages16
    JournalDiversity and Distributions
    Volume19
    Issue number4
    DOIs
    Publication statusPublished - Apr 2013

    Fingerprint

    Dive into the research topics of 'To boldly go where no volunteer has gone before: Predicting volunteer activity to prioritize surveys at the landscape scale'. Together they form a unique fingerprint.

    Cite this