Integrating LULC Prediction and Hydrodynamic Modelling for Urban Flood Hazard Assessment: A Case Study of Bagjola, Kolkata

Urban flooding presents a complex challenge driven by urbanization, changing land use patterns, and climate variability, particularly in monsoon-dominated developing countries like India. According to the WMO, the world is rapidly urbanizing, especially in the flood plains. As a result, the global population is at risk, with the number of people living in flood-prone areas rising by 24% from 58 million to 86 million between 2000 and 2015. This study develops a comprehensive methodology for flood hazard mapping in the Bagjola area( Kolkata, India), integrating Markov Chain and Cellular Automata (Markov-CA) Land Use and Land Cover (LULC) prediction models with hydrodynamic simulations. The Markov-CA model predicted the 2050 land use changes after employing the land use data from 1990, 2005, and 2020. The predicted LULC changes were incorporated into the calibrated and validated MIKE+ hydrodynamic model (2020 data) for flood simulations, and the simulation results depicted the flood hazard maps, highlighting vulnerable areas under different land use scenarios.

The Markov-CA model achieved a Kappa coefficient of 0.84, indicating reasonably good agreement. The results reveal a growing trend of urbanization in the Bagjola Canal region, with 6.1% of vegetation and 29.06% of barren land projected to be urbanized by 2050 compared to 2020. Compared to the situations observed in 2020, under the future scenarios for 2050, the total flood hazard area is expected to increase by 15-40% in the Bagjola Canal region.

This methodology provides a practical framework for assessing the spatial impacts of urbanization on flood risk, offering valuable insights for urban planning and flood management in rapidly developing regions. The resulting floodplain and hazard maps can assist local municipal bodies in preparing flood mitigation and evacuation plans and serve as a criterion for property insurance evaluations.