Skip to main content

Statistical Development of Ecological Removal Models

Zhou, Ming (2018) Statistical Development of Ecological Removal Models. Doctor of Philosophy (PhD) thesis, University of Kent,. (KAR id:70240)

Language: English
Click to download this file (3MB) Preview
[thumbnail of 47thesis_corrected_MZ (1).pdf]
This file may not be suitable for users of assistive technology.
Request an accessible format


Removal sampling is commonly used to estimate abundance of populations in which captured individuals are permanently removed from a study area. The classic removal model (Moran, 1951) assumes a constant capture probability and all animals are available for detection throughout the study, which results in a simple geometric decline of counts of removed individuals over time. However, the real data collected from some species exhibit unexpected fluctuations in the number of captured animals. The work in this thesis is driven by real data on common lizards, Zootoca vivipara and great crested newts, Triturus cristatus, where existing approaches may give rise to misleading conclusions.

When modelling removal data it is crucial to account for imperfect availability in the population, as individuals could sometimes temporarily become undetectable at

study area, or emerge from an area outside the study. This thesis deals with three

aspects of removal modelling: (i) We develop a robust design multievent removal

modelling (RMER framework) which allows considerable flexibility in estimating temporary emigration as well as capture probability and the size of populations. We

also consider the effect of sparse data and investigate the use of modelling different

sources of data in conjunction with the removal data (Besbeas et al, 2002). (ii) The

estimation of temporary emigration or population renewal for removal data relies on

the use of the robust design (Zhou et al. 2018). However, there are many removal

data which lack the robust design structure. Motivated by the analysis of a data

set of common lizards collected under standard sampling protocol, we develop and

evaluate the use of penalised maximum likelihood estimation to allow populations to be open to new individuals via birth/arrival for data sets without the robust design. (iii) We use four criteria to explore study design aspects of removal data with the robust design, including the trade-off in survey effort allocation between primary periods and secondary periods for a fixed level of total sampling effort. The models we propose can account for temporary emigration or new arrivals of individuals during removal sampling and represent a step forward with respect to current modelling approaches and will guide wildlife management.

Item Type: Thesis (Doctor of Philosophy (PhD))
Thesis advisor: McCrea, Rachel
Thesis advisor: Matechou, Eleni
Thesis advisor: Cole, Diana
Divisions: Divisions > Division of Computing, Engineering and Mathematical Sciences > School of Mathematics, Statistics and Actuarial Science
SWORD Depositor: System Moodle
Depositing User: System Moodle
Date Deposited: 21 Nov 2018 13:10 UTC
Last Modified: 01 Dec 2021 00:00 UTC
Resource URI: (The current URI for this page, for reference purposes)
  • Depositors only (login required):

Total unique views for this document in KAR since July 2020. For more details click on the image.