Indicators for the assessment of social resilience in flood-affected communities - A text mining-based methodology

Sci Total Environ. 2020 Nov 20:744:140973. doi: 10.1016/j.scitotenv.2020.140973. Epub 2020 Jul 15.


This work turns the social resilience concept into a practical and tangible set of dimensions and indicators for social resilience assessment. It further provides an analysis of the social resilience concept in the context of flood risk governance. Floods are a worldwide recurring phenomenon that causes severe social, economic and environmental losses. In the context of global change, it is very difficult to accurately predict extreme events that may increase disaster frequency; hence the implementation of social resilience is essential to lessen the losses. Indeed, the right balance between natural and social factors and indicators is yet to be found. Social resilience has been debated extensively for decades, both in scientific and political contexts. It has been a concern in disaster risk reduction and risk governance fields, both of which have strived to implement it. The enlarged conceptual discussion regarding this topic has resulted in some indicator-based assessments that hardly reflect the conceptual discussion developed so far. These indicator-based approaches still lack accurate inclusion of social dynamics and the capacity to learn from experience. In order to contribute to a comprehensive approach (concept and methods) for assessing social resilience to floods, the evolutionary resilience concept (Davoudi, Simin; Shaw, Keith; Haider, L. Jamila; Quilnlan, Allyson E; Petterson, Garry D.; Wilkinson, Cathy; Fünfgeld, Hartmut; McEvoy, Darryn; Porter, 2012) was considered as a reference in this work, as it can include dimensions that are difficult to evaluate (non-static time and learning-capacity in multi-dimensional systems). This work addresses the challenge of a conceptual overview of social resilience to include key factors and indicators. Our methodology uses text mining, experts' surveys and bibliography reviews to generate an indicators database. The contribution of this article to the scientific debate on social resilience assessment is twofold. First, the key-concepts, words and expressions in this field are identified, which provides the basis to build a comprehensive and coherent analytical framework. Secondly, an original indicators database is proposed in line with that framework. The results of a text mining-based methodology and an online survey, involving experts from different countries, show that four of the six dimensions of the indicators database refer to social aspects of risks (Individuals, Society, Governance, and Built Environment), while the remaining two refer to the Environment and Disaster. The results obtained so far suggest the need for a next step aiming to validate the dimensions and the indicators of this database through its application to real case studies.

Keywords: Experts' opinions; Floods; Resilience indicators; Social resilience; Text mining.

MeSH terms

  • Data Mining
  • Disasters*
  • Floods*
  • Humans
  • Surveys and Questionnaires