Understanding people's perceptions of climate change impacts on rural water resources in Bangladesh : an integrated modelling approach

Bangladesh is highly vulnerable to the impacts of climate change on water resources through extreme weather events such as floods, droughts, storms, and the inundation of coastal regions. These extreme weather events threaten human life, water security and crop production. Changes in the observed frequency and intensity of these events are believed to be the consequence of changes in climate. Bangladesh government organisations have undertaken initiatives to achieve the sustainability of water resources and to adapt to, and mitigate, the impacts of climate change on water resources, but these initiatives have focused largely on the scientific perspective. This thesis explores the social dimensions of local peoples' understanding of the impacts of climate change and interactions with local water resources and environment. The research examines the nature of the link between social-cultural perceptions about current environmental impacts and historical social-cultural traditions in Bangladesh. It develops an integrated water resource management model that incorporates this socio{u00AD}cultural knowledge with other factors (for example, economic and environmental aspects). The thesis integrates two main tools and methods to analyse water resources management in rural areas in Bangladesh. Social cognitive aspects are investigated using the Social Network Analysis (SNA) method to better understand social-cultural perceptions and patterns of social ties for environmental knowledge sharing among participants in communities in Bangladesh. The Bayesian Network (BN) approach (probability based graphical models that deal with uncertainties) is used in this thesis to integrate knowledge from the natural and human sciences, and to represent interrelationships between social, cultural, economic factors and environmental aspects to identify the critical variables influencing water resources management. The analyses and models developed in this thesis were based on semi-structured surveys undertaken in two rural villages in Bangladesh: Bogra and Meherpur. The surveys elicited people's observations of the impacts of climate change on water resources, their patterns of sharing environmental knowledge and the relationships between social beliefs and perceptions about the causes of extreme weather. People are mostly connected with a few influential actors within and outside the community for environmental knowledge sharing. These social connections are to some degree influenced by peoples' social and cognitive factors, expected future benefit (e.g. expected outcome), social identity, social position and status. Results showed that people have various cultural and traditional views about the negative impacts of extreme weather events on water resources due to the impacts of climate change. Among the perceptions about the impacts of climate change, 'Divine retribution' is most commonly identified by participants. Local interpretations of the impacts of climate change on water resources are linked to people's historical traditions, and multi faith religions, and a lack of contemporary environmental education (e.g. formal and informal education). Of the people who believe that the incidence of extreme weather events is increasing, more people in Bogra think that 'various religious values from the past' contribute to other people's belief in 'Divine retribution', compared to Meherpur where more people think that 'lack of contemporary environmental context in education' contributes to this belief. Cultural and traditional views and beliefs may influence people's willingness to adopt climate change strategies for water resources. However, this is not deeply investigated due to the limited timeframe and resources of this study. Bayesian networks modelling in this thesis provided different observations under various scenarios and identified the most influential variables (factors) in the system. Environmental impact context and policy context are influential variables in both locations. In Bogra the national authority variable is important, while in Meherpur the local authority variable is important. This model can potentially guide policy actions towards better water resources management. The study experienced a number of challenges that could be addressed in future research. Due to the small sample size, sub-group analysis was limited and statistical significance could not be tested. The study could therefore be extended to more locations with a larger sample population. The thesis used semi-structured questions and as a result it became difficult to compare and group people's comments in neat categories. More structured surveys in larger sample populations might produce statistically significant results and would be better suited to the development of BN conditional probability tables using learning algorithms. The framework could be further tested and revised to identify other factors that potentially influence the outcome