Multi-scale assessment of risks to environmental hazards in coastal area of Bangladesh

Abstract

The coastal zone of Bangladesh is highly vulnerable to extreme climate and environmental hazards. The severity or consequences of hazard depends on its nature along with the exposure of human and natural systems and vulnerability. Vulnerability and risk assessments help devise proper risk management plans focusing on reducing or modifying components of risks. But vulnerability and risk assessment needs to be carried out at appropriate scale as the components of risk vary with scale. The objective of this study was to conduct coastal risk and vulnerability assessment for three different scales: regional scale (coastal zone excluding Chittagong and Cox’s Bazar districts), sub-regional scale (Koyra upazila, Khulna district) and local scale (Bagali union, Koyra upazila). Central to the methodology was the use of IPCC AR5 approach, which defines ‘risk’ as an outcome of the interaction between ‘hazards’, ‘exposure’ and ‘vulnerability’. Respective indicators were selected based on secondary literature, expert opinion, data availability and local stakeholders’ opinion. Relative importance of indicators and domains were determined by Principal Component Analysis (PCA) for regional scale analysis. Analytical Hierarchy Process (AHP) was used for smaller scale analyses. Multi-scale assessment allows stakeholder derived and weighted parameters, which are more relevant at the local scale. For example, Sundarbans dependent livelihood and early warning system were identified as more relevant by local stakeholders. Also, regional assessment tends to mask important local scale processes, which are captured by local scale assessment. For example, Koyra upazila was found to have almost zero erosion at regional scale, whereas it is the biggest concern in all unions of the upazila. This has considerable policy implications in the context of prioritizing investments for hazard mitigation. Variations in weights against indicators and vulnerability domains were found across different scales, as local peoples’ perceptions captured in AHP differed from PCA derived weights. For example, education domain was ranked first (weightage 0.1990) at the regional scale whereas housing and infrastructure was ranked first at the sub-regional and local scales (weightage 0.3368). Similarly, livelihood domain received much higher weightage (ranked third with weightage value 0.1351) at the smaller scales because of the very high livelihood dependency on agriculture and Sundarbans dependent activities. In contrast, livelihood was ranked bottom at the regional scale (weightage 0.0213). These have implication in that investment at the local scale ought to have also major focus on housing and infrastructure improvement and livelihood improvement. The scale of assessment can determine the outcome; an area can be regarded as highly vulnerable at one scale, but differently at another scale, and an area with moderate multi- hazard value at one scale may have different multi-hazard value at another scale. For example, in Koyra upazila, multi-hazard, exposure and vulnerability values are 38.39, 24.35 and 66.51 respectively in regional scale of analysis while the same are 63.126, 44.495 and 52.884 in sub-regional scale assessment, respectively. The high multi-hazard value in Koyra is because of better representation of hazards and use of stakeholder derived weightage at the smaller scales. Hence, investments would also receive priority at the local scale for Koyra in terms of hazard mitigation. In summary, the results highlight the value of a nested approach to index development and reinforces the importance of scale in determining policy response to vulnerability. The regional scale indices could be used in regional development planning in the context of reduction of risk components. At the local level, the detail of the indices makes it possible to identify specific investments/ interventions. The indices at the three scales may allow users to make tradeoffs appropriate to their needs by choosing the most suitable index scale and thus try to ensure better utilization of resources.