An integrated analysis of the utility of Citizen Science for understanding trends in species distribution and abundance: The Great Kererū Count as a case study
Understanding the distribution and abundance of species over large spatial extents is key for undertaking evidenced-based conservation. Further understanding of the key factors that determine this range is imperative for understanding how past and future actions have, and will continue to have, wide-ranging impacts on species populations. Citizen Science presents an unprecedented opportunity to gather data at spatial extents and temporal scales not possible through traditional scientific methods. Accurate inference into species ecology based upon these data allow for the conservation and restoration of species throughout the globe. However, obtaining accurate inference from Citizen Science is notoriously difficult due to a number of biases.
Here, I have conducted a multifaceted analysis into the distribution and abundance of the native New Zealand Wood Pigeon (Hemiphaga novaeseelandiae, kererū) based upon data gathered from a large-scale Citizen Science project, The Great Kererū Count (GKC). Analyses allowed me to interrogate both the ecology of the kererū and the utility of Citizen Science data for providing inference into species’ ecology. To account for the biases and inaccuracies discovered through this process, and to capitilise on the useful aspects of the data, a relatively new method of analysis was conducted, where multiple, independent datasets are jointly modelled in an Integrated Species Distribution Model. A number of methods for joining datasets were assessed before the most robust method was applied to a dynamic ecology problem, investigating the effect of a mega-mast on the national kererū population.
Data from The Great Kererū Count were found to be useful for naïve estimation of behavioural and preference characteristics of kererū. Kererū behaved differentially in urban and rural areas, with a greater incidence of congregations and displays in the latter. Citizen Scientists estimates of the frequency of kererū visits to an area were accurate but their assessments of trends over time were not. Incorporating data from another, more structured Citizen Science project, the New Zealand Bird Atlas, into an Integrated Species distribution Model significantly reduced error and improved model fit compared with a GKC-only model. The national kererū population was found to be stable, if not increasing, but dependent on both indigenous forest and ongoing predator control. Application of a modified model illustrated the effects of a nationwide mega-mast driving kererū into higher elevation forests and away from the irrupting introduced mammalian predators. Despite this, the national kererū population remained stable.
This thesis represents the most up-to-date assessment of the national kererū population. Results support the classification of kererū as ‘Not Threatened’ but conservation dependent. Ensuring the national kererū population remains stable will require ongoing protection from introduced predators, and restoration of key kererū habitat. The modelling methods used here have been described overseas but this thesis represents the first application of them in New Zealand. The unique advantages of these modelling techniques allowed for the description of temporal and spatial changes in kererū populations in response to national scale events. With the multitude of conservation dependent, and threatened, species in New Zealand, robust, and timely assessment of population trends, and the factors impacting them, are imperative. This thesis demonstrates the efficacy of Integrated Species Distribution Models for this task in the New Zealand context.