Future flood susceptibility mapping under climate and land use change

Area of Science:

• Environmental Science
• Hydrology
• Climate Science

Background:

• Floods pose significant natural hazards, causing widespread damage.
• Understanding the interplay of climate change and land use/land cover (LULC) changes is vital for effective flood management.
• Urbanization and climate shifts are altering flood patterns, necessitating updated risk assessments.

Purpose of the Study:

• To develop flood susceptibility maps incorporating climate change and LULC dynamics.
• To assess flood risks associated with urbanization and projected climate shifts.
• To provide insights for sustainable flood risk management strategies.

Main Methods:

• Applied three machine learning models: XGBoost, Random Forest (RF), and Support Vector Machine (SVM), optimized with Particle Swarm Optimization.
• Utilized factors such as river distance, digital elevation model, precipitation, and LULC for susceptibility mapping.
• Employed the CA-MARKOV model for land use projections and analyzed future precipitation under SSP126 and SSP585 scenarios using General Circulation Models.

Main Results:

• The Random Forest model demonstrated superior performance in mapping flood-prone areas.
• High-risk flood zones were identified, covering 20% of the Kashkan watershed, predominantly in built-up areas.
• Land use projections indicate significant expansion of urban areas by 2050.
• Future climate scenarios, particularly SSP585, suggest an increase in moderate and high flood risk areas.

Conclusions:

• Flood susceptibility is significantly influenced by proximity to rivers, topography, precipitation, and LULC.
• Urban expansion exacerbates flood risks, particularly in high-risk zones.
• Climate change under future scenarios will likely expand areas affected by moderate and high flood risks, demanding proactive adaptation measures.