Data-driven approaches to support amphibian and reptile conservation

The ability to effectively monitor and respond to declines in global biodiversity is often constrained by a pervasive 'data crisis' in conservation: the paucity of available, high quality, and re-usable biodiversity data. Alongside data collection, models can accelerate scientific understanding by revealing patterns and processes that are difficult to observe and measure directly, particularly for threatened and cryptic species with limited data. Globally, amphibians and reptiles face high extinction risks, yet they are underrepresented in conservation science and policy. Currently, the status and management needs for species in the United Kingdon (UK) are not well understood due to limitations in available data. In this thesis, I explore data-driven approaches to improve the evidence base for UK amphibians and reptiles, focusing on data integration and modelling. A scoping review and network analysis revealed a dynamic yet patchy data landscape, with limitations in the re-usability of datasets for national-scale assessments. Focussing on a widespread species with low detectability - the slow-worm (Anguis fragilis) - I developed an integrated, multi-season occupancy-detection model which combined data from structured surveys, community science programmes, and online databases. The integrated model outperformed single-dataset models in terms of predictive precision and accuracy, but there were only marginal benefits over simpler, pooled models. Slow-worm occupancy showed increasing trends from 2007-2020, although this may be confounded by variation in sampling effort. Accounting for imperfect detection produced realistic and useful predictions of the species range, identifying environmental drivers of slow-worm distribution. A process-based model to simulate landscape-use by the European adder (Vipera berus) incorporated expert knowledge and landcover data to produce spatially explicit estimates of relative abundance. Validation against independent data collected from 257 monitoring sites provided the first quantitative evidence that anthropogenic pressures may have limiting effects on some adder populations. I present two demonstrations of how the model could be used to support spatial conservation planning at different spatial scales. A synthesis of insights from a participatory stakeholder workshop examined the barriers and opportunities to improving data flow across the monitoring network. Stakeholder integration, enhanced technical capacity, and sustained funding mechanisms emerged as critical to ensure that monitoring translates into positive conservation outcomes. Collectively, the results demonstrate that co-designed, data-driven approaches can strengthen conservation science and policy at a range of spatial and temporal scales.