Evaluating the usefulness of published data for estimating key parameters required in modelling global avian extinction risk

Abstract

Despite the best efforts of conservationists worldwide, species extinction risks continue to rise. It is predicted that under intermediate climate warming scenarios 15-37% of species will be committed to extinction by 2050. This, coupled with limited funding and resources, means conservation management must be prioritised. Population viability analysis (PVA) models can help prioritisation by providing estimates of extinction risks for species. However, at present the availability of avian life history data and population data is limited, which makes this analysis challenging. Therefore, the aim of this thesis is to collate and calculate the necessary data so PVA models can be run for all bird species of the world.
We begin by looking at what density data is available for species because these underpin much of our understanding of the extinction risks, as they are directly linked to population sizes and population sizes are known to be highly correlated with extinction. We collate field densities for approximately 30% of all avian species and then implement a Generalised Linear Model (GLM) to calculate densities for the remaining species. In total, densities are modelled for 8,541 species with a 37% accuracy. We then use these densities, along with distribution polygons and habitat data, to calculate population sizes for 6,206 species with a 55% accuracy. Finally, as survival estimates are a key demographic parameter to include in PVA models, we calculate these for 5,291 species with a 36% accuracy.
Having calculated densities, population sizes and survival rates for over half of the worlds birds, we conclude that this is a huge step forward in being able to calculate extinction risks for many species. However, we highlight throughout that accuracy could be improved with more data collection, and fundamentally some data are still crucially missing if we want to run PVA models. Therefore, we suggest further research should aim to collect more avian data, such as fecundity, so simple PVA models can be run. For those species with the highest extinction risks we suggest even more data is collected, so more complex models, which include the effects of stochasticity, genetics and climate change can be run. We believe if robust and reliable data can be collected and included in PVA models, the results would be truly informative and insightful for conservation management and prioritisation.