Abstract:
Flood vulnerability index has long been used as an essential tool to quantify the sufferings of the flood victims of Bangladesh. Enormous studies highlighted vulnerability to flood to be dependent on the physical exposures to the location and socioeconomic conditions of the community. But very little is known about the varying spatial scales of vulnerability index within the same socioecological system and the varying vulnerability of people across gender. To address this drawback, this research conducted vulnerability analyses in three distinct land features (i.e., mainland, attached char, and island char) of Dewanganj Upazila at Jamalpur district and assessed how vulnerability differs here between men and women. A mixed method was applied in collection of both qualitative and quantitative primary data, including 120 household surveys, 7 focus group discussions, 13 key informant interviews (KII) with officials, and 8 in-depth interviews with flood victims. The collection of secondary data has encompassed various sources, including census data obtained from the Bangladesh Bureau of Statistics, as well as rainfall and river discharge data acquired from the Bangladesh Water Development Board. First, location specific gender wise priority indicators were selected out of a total of 39 vulnerability indicators. Subsequently, indicators’ values were obtained from the 120 household surveys and corresponding weightages were estimated via two methods, namely, Analytic Hierarchy Process (AHP) and Principal Component Analysis (PCA). Finally, gendered vulnerability indices for male (M-VI) and female (F-VI) and gender-blind vulnerability indices (Gb-VI) were generated for each place using both vulnerability concepts proposed in the Fifth Assessment Report (AR5) and Fourth Assessment Report (AR4). The results show that women prioritise flood depth, distance from river, source of drinking water, dependent population, shelters, and so on, whereas men prioritise a different set of indicators, such as unemployment rate, income, migration, and so on, by ranking them >=4 out of 5. Interestingly, indicator prioritisation differs throughout space. For example, on a scale of 1-5, the indicator 'river erosion' receives < =3 in the mainland but 5 in attached char and island char. Gb-VI is always lower than F-VI and greater than M-VI in both AR5 and AR4 vulnerability concepts estimation. AHP with consideration of all indicators (32 indicators under AR5 and 39 indicators under AR4 concepts) and PCA with selective indicators (24 indicators under AR5 and 29 indicators under AR4 concepts) demonstrates that females consistently exhibit greater vulnerability compared to males within same geographical locations. Women of island char are the greatest vulnerable group in AHP (3.06 times more vulnerable under AR5 and 37% vulnerable under AR4 concepts), while women of attached char are the highly vulnerable group in PCA (2.56 times more vulnerable under AR5 and 32% vulnerable under AR4 concepts). Men from the mainland are the least vulnerable, independent of the methodologies or equations used. M-VI identifies men are 1.33 and 1.17 times more vulnerable under AR5 concepts in AHP and PCA, respectively and 22% and 24% vulnerable under AR4 concepts in AHP and PCA, respectively. Based on the findings from multiple locational vulnerability assessments, it was noted that the vulnerability index derived through the AHP method has a higher level of correspondence with data obtained through KII compared to estimates obtained through PCA. The reason for the elimination of significant indicators through sensitivity analysis in PCA, as opposed to the consideration of all indicators in AHP, might be attributed to this discrepancy. The AHP has demonstrated excellent performance in assessing the existing vulnerability index, leading to the conclusion that it is a suitable methodology for estimating future vulnerability. Based on the projected changes in rainfall, population density, female population density, proportion of dependent population, and life expectancy rate, which are estimated to be 11%, 19%, 1.67%, 14.7%, and 11% respectively, it is anticipated that the vulnerability of the specified region will undergo a 36% increase by the year 2050, as indicated by the estimation derived from AR5 vulnerability concept. In a similar vein, the computation derived from the vulnerability concept outlined in AR4 anticipates a 38% increase in vulnerability during the identical temporal span. The findings of this research will provide valuable insights for policymakers in implementing flood precaution measures that are tailored to certain locations and genders. The study has identified and prioritised indicators that can serve as a guide for the implementation of gender-focused services in other flood-affected communities.