Use this URL to cite or link to this record in EThOS: http://ethos.bl.uk/OrderDetails.do?uin=uk.bl.ethos.705917
Title: Population mapping using census data, GIS and remote sensing
Author: Firoozi Nejad, Behnam
ISNI:       0000 0004 6062 0068
Awarding Body: Queen's University Belfast
Current Institution: Queen's University Belfast
Date of Award: 2016
Availability of Full Text:
Full text unavailable from EThOS. Please contact the current institution’s library for further details.
Abstract:
This thesis assesses approaches to population surface modeling by pulling together the benefits of reference gridded population data with local regression procedures and geographically weighted regression. This study provides a more detailed assessment of surface modelling accuracy than was achieved in any previous studies to assess factors which explain errors in the predictions. The primary aim of this thesis is to evaluate Martin’s (1989) population surface modeling approach and also design and implement a method using secondary data, suitable for application in England and Wales. This research is based on the idea that population data presented for a single zone could be redistributed in the zone using local parameters such as housing density. A weighted sum performs the spatial redistribution. The thesis also aims to make use of remote sensing (RS) data and image processing techniques such as maximum likelihood classification and normalised difference vegetation index to identify (un) populated cells. The potential of Landsat images and RS data analysis is assessed particularly for countries where high quality land use data are not readily obtainable, and their generation is not feasible in the near future. This thesis focuses on the identification of unpopulated cells, rather than populated units, using RS data. Case studies make use of data from Northern Ireland (NI), and Jonkoping in southern Sweden. The outcomes indicate the impact of population density, population variance, and resolution of source zones on the accuracy of population allocation to grid cells using Martin’s (1989) model. The results show significant accuracy in prediction to 100m cells using an alternative approach based on settlement data for NI and this is recommended as an alternative method for England and Wales. It also concluded that there the potential to generate population surfaces using Landsat data for areas where local residential data are not easily accessible.
Supervisor: Not available Sponsor: Not available
Qualification Name: Thesis (Ph.D.) Qualification Level: Doctoral
EThOS ID: uk.bl.ethos.705917  DOI: Not available
Share: