| Abstrak/Abstract |
Land use changes and rainfall can trigger inundation. This study aimed to observe the dynamics of spatial patterns of temporal flood vulnerability due to rainfall and land use changes using the Maximum Entropy (MaxEnt) Model. Flood vulnerability was assessed using 12 environmental variables, including elevation, slope, aspects, curvature, Topographic Wetness Index (TWI), drainage density, distance from rivers, distance from roads, soil texture, soil aggregates, rainfall, and land use. Rainfall and land use were dynamic variables analyzed in 2013-2023. Past flood occurrence points were obtained using the participatory mapping method. Temporal flood vulnerability mapping in 2013, 2018, and 2023 showed the influence of elevation, Topographic Wetness Index (TWI), and distance from rivers, which were very dominant. Typically, the flood vulnerability pattern formed showed a percentage of moderate (13%), high (17%), and very high (5%) class areas consistently clustered in the southern region. An interesting finding is that rainfall changes have a more significant influence (7.2%), causing the dynamics of high and very high-class vulnerability patterns, compared to the influence of land use changes (0.4%). MaxEnt's flood vulnerability prediction accuracy is classified as very good, as evidenced by its AUC values of 0.835 in 2013, 0.819 in 2018, and 0.824 in 2023. Finally, the findings showed that the accuracy of the MaxEnt Model is classified as very good, so it can be extrapolated globally with similar regional typologies. |
