Advancing methods for estimating and modeling breeding bird distributions for conservation and management

Abstract

Effective conservation depends on reliable inference about where species occur and how they respond to environmental change, yet both goals are often undermined by limitations in monitoring design and the scale-dependence of habitat relationships. In this thesis, I address two complementary challenges: (1) whether occupancy estimates for mobile species can be made comparable across variable point-count protocols, and (2) how multi-scale habitat conditions shape distributions of a disturbance-dependent game bird within managed landscapes. First, utilizing an individual-based movement model for Wood Thrush (Hylocichla mustelina), I simulated populations across a range of densities and survey protocols and evaluated whether post hoc asymptotic modeling could recover consistent instantaneous and daily occupancy. Protocol variation substantially altered occupancy estimates, and asymptotic approaches did not reliably standardize results across designs; although asymptotic regression reduced bias in daily occupancy, improvements were inconsistent and accompanied by reduced precision, underscoring the need for standardized survey design. Second, I developed a 30-m species distribution model for Northern Bobwhite (Colinus virginianus) across 11 Alabama Wildlife Management Areas using breeding-season surveys from 2008–2010 and 2024–2025. A multi-scale modeling workflow that screened predictors, refined random forest models, and validated results with independent data produced strong predictive performance (AUC = 0.864). Bobwhite occurrence was most strongly associated with intermediate-scale shrubland and was consistently constrained by broader-scale closed-canopy deciduous forest, revealing a mismatch between the scales of key habitat drivers and the finer scales at which management is typically implemented. Together, these chapters demonstrate that reliable conservation inference depends on both rigorous, standardized monitoring and spatial models that capture scale-dependent habitat relationships, providing a strong foundation for long-term biodiversity assessment and adaptive management.