Abstract

Flood events have been generating great risks and intensifying the challenges of water management in coastal megacities. Instead mitigating the impact of climate change, improving the resilience has an expanding scope of application in environmental science, covering climate change, risk and disaster management [1]. In the past decades, metrological and hydrological causes have been the main drivers of disasters [2], which makes vulnerability assessment methods such as Flood vulnerability index (FVI) clear development pathways [3].This study is built on the household survey data in HCMC (Ho Chi Minh City, Vietnam) in the framework of the DECIDER project (DECisions for Adaptive Pathway Design and the Integrative Development, Evaluation and Governance of Flood Risk Reduction Measures in Transforming Urban-Rural-Systems). It aimed at creating a framework to interpret social-economic attributes of flood vulnerability with physical features of household. As an essential part of the influencing factor for social vulnerability, the data is regionally intrinsic and mostly accessible only by field survey. Proxy variables were obtained to conduct contextual analysis based on remote sensing images, environmental risk estimates, as well as elevation data, in which study concludes that this method can contribute to identifying the key indicators and optimize the social vulnerability assessment to be more efficient [4]. In this study, 17 socioeconomic indicators only accessible from survey data were weighted using the Principal Component Analysis (PCA), and further aggregated into the FVI. Then 11 physical proxy indicators were collected from field inspection, remote sensing data and environmental flood risk estimates, and trained machine learning models to predict FVI. The AdaBoost model identified the most important physical indicators and the model was able to predict the test data with a MAE of 0.089 but small R2. Another decision tree model, however, was overfitted and yielded a moderate accuracy (~0.4) and further machine learning classification models were also applied on both eleven indicators and selected indicators for each case but no obvious difference showed among these models. Therefore, the socioeconomic FVI could be predicted with physical proxy variables with AdaBoost accurately, but more featured data should be acquired and model rendering can be done in the future for a better prediction model, especially for regional prediction with the scale of households and community.  References[1] O’Brien, K. Global environmental change II. Progress in Human Geography 2012, 36, 667–676, doi:10.1177/0309132511425767.[2] Birkmann, J.; Teichman, K. von. Integrating disaster risk reduction and climate change adaptation: key challenges—scales, knowledge, and norms. Sustain Sci 2010, 5, 171–184, doi:10.1007/s11625-010-0108-y.[3] Balica, S.F.; Wright, N.G.; van der Meulen, F. A flood vulnerability index for coastal cities and its use in assessing climate change impacts. Nat Hazards 2012, 64, 73–105, doi:10.1007/s11069-012-0234-1.[4] Ebert, A.; Kerle, N.; Stein, A. Urban social vulnerability assessment with physical proxies and spatial metrics derived from air- and spaceborne imagery and GIS data. Nat Hazards 2009, 48, 275–294, doi:10.1007/s11069-008-9264-0.

Talk to us

Join us for a 30 min session where you can share your feedback and ask us any queries you have

Schedule a call

Disclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.