Jump to Main Content
Risk assessment based on fuzzy synthetic evaluation method
- Akter, Marin, Jahan, Momtaz, Kabir, Rubaiya, Karim, Dewan Sadia, Haque, Anisul, Rahman, Munsur, Salehin, Mashfiqus
- The Science of the total environment 2019 v.658 pp. 818-829
- expert opinion, experts, humans, risk, risk assessment, uncertainty
- The IPCC fifth assessment report envisions risk of climate-related impacts as an outcome of the interaction of climate-related hazards with the vulnerability and the exposure of human and natural systems. This approach relies heavily on human perception, via expert opinions. As experts decide appropriate placement of an indicator in any of the exposure, sensitivity or adaptive capacity domains, several risk maps can potentially be created for the same study area. There is thus some degree of uncertainty in selecting the most appropriate and representative risk map from the several alternatives created by IPCC methods. On the other hand, Fuzzy Synthetic Evaluation (FSE) method, when used to assess risk, can handle this uncertainty much better, as there is no need to distribute indicators among different domains. In FSE, a specific indicator can either increase (positive sign) or decrease (negative sign) a risk, following a simple binary logic. This does not require any expert opinion and thus is free from subjective perception. In this study, risk maps are generated and compared by applying FSE method and two IPCC methods, as outlined in the third and fifth assessments (TAR and AR5). A variant of AR5 risk map is created by interchanging one indicator from the exposure domain to the sensitivity domain. It is found that risk zones are created with statistically significant difference when different IPCC methods are applied. This makes it uncertain to judge a specific risk map by a specific IPCC method as a true risk map. This uncertainty does not exist in FSE method as there is only one risk map where indicators are placed with certainty by following a simple binary logic for a known hazard domain. Hence, this risk map may be considered as the true risk map for the given set of indicators.