The application of multivariate cluster analysis in the assessment of volcanic social vulnerability

Willis, Iain Stuart (2016) The application of multivariate cluster analysis in the assessment of volcanic social vulnerability. PhD thesis, Birkbeck, University of London.

cp_Thesis_lS_WILLIS_2015.pdf - Full Version

Download (10MB) | Preview
Print Copy Information:


The 20th Century was characterized by increasing human population settlement in volcanically active regions of the world. This continued growth, particularly in less developed nations, has led to an increasing exposure of households and communities more predisposed to the social and physical risks a disaster could present. This thesis proposes a new methodology for the identification, targeting and assessment of these socially vulnerable communities. Drawing from specific examples of Mount Vesuvius (Italy) and Guagua Pichincha (Ecuador), multivariate statistics are applied to population census data to characterise the frailties and assumed coping capacity of different neighbourhood types to volcanic risk. Using cluster analysis and geodemographic discriminatory techniques, results show that communities more pre-disposed to the social and economic pressures of a disaster can be identified using this method. This approach looks to enhance upon current disaster risk metrics that tend to focus on single or cumulative risk scores, rather than seeking to define the behavioural traits and attitudinal perceptions of a neighbourhood. The peripheral and often informal barrios around Quito, Ecuador are shown to be highly susceptible to volcanic social vulnerability, whilst the Campania province around Vesuvius, Italy, highlights that the greatest risk to community resilience is associated with the high density settlements along the coastal towns near the volcano. The complex nature and site-specific characteristics of volcanic hazards, as well as the cultural landscape in which a volcanic eruption takes place are found to be key determinants in all aspects of disaster reduction. Vulnerability indicators, as defined in previous studies of disaster response are often independent of each other, and in many cases, non-transferrable in different cultural settings. Similarly, vulnerability and risk perception are as much a consequence of culture and state as they are of geographical setting and the physical characteristics of a volcanic eruption. Whilst caution is advised on the application and treatment of vulnerability metrics for mitigation, examples are provided as to how a neighbourhood classification systems methodology can be practically applied for disaster risk reduction. The output of this thesis is proposed as being of direct use to disaster risk managers (DRM), civil authorities and NGOs as an alternative tool in community outreach, exposure management, disaster mitigation and disaster preparedness plans. The contribution is also discussed in the wider context of disaster risk reduction measures, recent conceptual frameworks, and ongoing global initiatives such as the United Nations’ Hyogo Framework for Action and its intended replacement, HFA2.

Item Type: Thesis (PhD)
Additional Information: Date of PhD award confirmed as 2016 by registry
Copyright Holders: The copyright of this thesis rests with the author, who asserts his/her right to be known as such according to the Copyright Designs and Patents Act 1988. No dealing with the thesis contrary to the copyright or moral rights of the author is permitted.
School/Department: School of Social Sciences, History & Philosophy > Department of Geography
Depositing User: ORBIT Editor
Date Deposited: 04 May 2016 10:49
Last Modified: 04 May 2016 10:58

Actions (ORBIT staff only)
View Item View Item