Comparative assessment of flood resilience in urban and rural environments: a case study of Chennai City and Wardha River Sub-basin, India

Flooding is one of the most destructive natural disasters, with devastating impacts on human lives, infrastructure, and economies. The increasing frequency and severity of extreme weather events, as highlighted by the 2024 Global Risks Report, emphasize the need for effective flood management strategies. In India, a country highly vulnerable to floods, both urban and rural areas face distinct challenges. This thesis explores the contrasting dynamics of urban and rural flooding, focusing on Chennai as a case study for urban flooding and the Wardha River sub-basin for rural flooding. By examining the role of land use changes, urbanization, and hydrological factors, it provides insights into the differing mechanisms, impacts, and management strategies required for these environments. In urban areas like Chennai, flooding is predominantly driven by rapid urbanization, unplanned development, and the proliferation of impervious surfaces. These factors increase surface runoff and reduce infiltration, amplifying flood risks during extreme rainfall events. Using the InVEST-UFRM model, this research quantifies the effects of land use and rainfall changes on runoff and flood resilience. Findings highlight that land use transformations, particularly the reduction of water bodies, have a more significant impact on runoff than rainfall variability. The study estimates potential infrastructure damage of up to 10 billion USD and underscores the need for sustainable urban planning, improved drainage systems, and advanced flood forecasting to mitigate risks. In contrast, rural flooding in the Wardha River sub-basin is influenced by hydrological responses to land use changes and rainfall patterns. This thesis employs the NRCS-CN (The Natural Resources Conservation Service-Curve Number) method integrated with Google Earth Engine to model runoff dynamics, demonstrating how deforestation and land conversion exacerbate surface runoff. Landscape fragmentation analysis using machine learning and the FRAGSTATS programme reveals substantial forest loss and habitat degradation between 2010 and 2025, compromising ecological connectivity and increasing flood risks. Hypothetical scenarios project further increases in runoff under continued deforestation, highlighting the need for forest conservation and sustainable land management to mitigate rural flood impacts. The thesis also evaluates flood susceptibility mapping techniques, comparing statistical- FR (Frequency Ratio) and SEI (Shannon’s Entropy Index) and machine learning models across varying spatial resolutions. In urban areas, extreme gradient boosting (XGB) outperformed other methods, providing robust flood susceptibility predictions. In rural settings, the study demonstrated the effectiveness of GIS-based Frequency Ratio (FR) and Shannon’s Entropy Index (SEI) models in accurately delineating flood-prone zones. Both models achieved high levels of predictive performance, as evidenced by their Area Under the Curve (AUC) values exceeding 0.96, highlighting their reliability in flood susceptibility mapping and risk assessment. By integrating hydrological modelling, machine learning, and spatial analysis, this research distinguishes the drivers and impacts of urban and rural flooding, providing a comprehensive framework for targeted flood risk management. It emphasizes the critical need for region-specific strategies—combining sustainable land use planning, advanced modelling techniques, and proactive policy interventions—to address the unique challenges of urban and rural flood resilience effectively.

Author dc:creator

• Sharma, Asheesh

Advisor dc:contributor.advisor

• Hinkelmann, Reinhard

Licence dc:rights.uri

• http://rightsstatements.org/vocab/InC/1.0/

Language dc:language.iso

en

Identifier URI

https://doi.org/10.14279/depositonce-24626

OAI identifier oai:identifier

oai:depositonce.tu-berlin.de:11303/25802