University of Southampton
Sampling effort and uncertainty assessment in capture recapture studies
Abstract
dc:description.abstractThis thesis examines the optimal allocation of sampling effort in capture-recapture studies, with the aim of improving population size estimation. Sampling effort is a critical component of study design, yet it is often determined without a systematic approach. To address this gap, a simple, practical framework is developed using theoretical insights, basic mathematical computations, and computer simulations.<br/>The analysis begins with the Lincoln-Petersen model, which involves two primary capture occasions, each comprising multiple sub-occasions. When capture probabilities are either constant or follow a consistent pattern, distributing sub-occasions equally across primary occasions yields the most accurate estimates. However, when capture probabilities vary and are known in advance, optimal allocation is better achieved through numerical methods such as the Newton-Raphson algorithm and simple estimation techniques that incorporate prior information. These targeted strategies outperform equal allocation, particularly when capture probabilities fluctuate significantly over time.<br/>The investigation then extends to the Schnabel model, which focuses on determining the optimal number of capture occasions. To account for unobserved individuals, the model incorporates zero-truncated count data. In cases where closed-form solutions are unavailable, the Expectation-Maximisation algorithm is employed to estimate parameters. The hierarchical structure is expanded to scenarios involving multiple capture occasions. When capture probabilities remain stable or change predictably, uniform sampling effort remains effective. However, in contexts where capture probabilities decline over time, allocating greater effort in later occasions leads to more accurate population estimates by compensating for reduced detectability.<br/>The thesis also provides practical recommendations for real-world applications where resources are limited. The proposed methods support informed decisions about sampling effort, avoiding reliance on arbitrary or overly conservative designs. By clarifying the relationship between detectability, study design, and estimation precision, the framework enables more efficient planning.
Degree
thesis:*- Name dc:type.qualificationname
- Ph.D.
- Level dc:type.qualificationlevel
- doctoral
- Grantor dc:publisher.institution
- University of Southampton
- Year dc:date.issued
- 2025
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Chin, Su Na
- Advisors dc:contributor.advisor
-
- Böhning, Dankmar
- Overstall, Antony
- ORCID person:identifier.orcid
Rights
- Language dc:language
- English, eng