Abstract
This thesis explores novel methodologies for monitoring remote populations of albatrosses, one of the most threatened groups of birds globally, facing pressures both at sea and on land. Their extensive oceanic range and inaccessible offshore breeding colonies make regular and accurate population monitoring difficult, limiting our understanding of demographic trends and impeding effective conservation measures. Recent technological advancements have begun to overcome these challenges, offering new opportunities for long-term passive data collection via remote cameras and increased accessibility to aerial surveys at rugged breeding colonies using uncrewed aerial vehicles (UAVs). However, these tools introduce their own bias, which must be addressed to produce robust insights that are comparable with historical data. To support this integration, I develop and apply two methodological frameworks. First, I process five years of trail camera images into capture histories of eggs and chicks and develop Bayesian survival models to estimate stage-specific survival throughout the breeding cycle. Second, I design an adjustment pipeline for population survey data, accounting for spatial coverage, nest occupancy, and survey timing, to harmonise disparate survey methods and metrics into cohesive population trends. Using Salvin’s albatross (Thalassarche salvini) as a case study, I demonstrate the potential of these methods. I find low breeding success at 35% (95% credible interval: 23–48%), largely driven by poor egg survival (53%; 95% CI: 34–70%) during incubation. Chick survival was more moderate (67%; 95% CI: 58–75%) across the guard and post-guard stages. For long-term trends, I compiled and adjusted all available population estimates from 1977 to 2025 and fitted a state-space model. Results suggest a 43.7% decline in the global breeding population, from 87,776 pairs (95% CI: 43,437–144,483) in 1977 to 49,953 pairs (95% CI: 39,507–63,166) in 2025. This research shows that with careful integration of modern technologies and statistical approaches, we can substantially improve population assessments of hard-to-monitor species like Salvin’s albatross. These methods are readily adaptable to other albatross species and offer valuable tools for advancing seabird conservation.