Mr Shashank Kumar Anshu will present his APS as per the details below:

Date: 25th February 2026 (Wednesday)  

Time:   1500 - 1700 hrs.    

Venue: C-TARA Conference Room No. 1   

Title:  A Multi-Dimensional Approach for Landslide Risk Assessment combining Dynamic Hazard Modelling and Response Strategies in Western Ghats of Maharashtra, India    

Guide: Prof. Parmeshwar D. Udmale   

RPC Members:  Prof. Pankaj Sekhsaria, Prof. RAAJ Ramsankaran   

Abstract:

Landslides pose a significant threat to human life, infrastructure, and the environment, disrupting ecosystems and undermining socio-economic stability. The United Nations Agenda 2030 emphasizes disaster risk reduction as a global priority, recognizing that hazards such as landslides intensify existing vulnerabilities. The 2023 UN SDG Report indicates that over 1.30 billion people reside in landslide-prone areas worldwide, with millions affected annually. India ranks among the five most vulnerable countries, recording more than 200 landslide-related fatalities each year. Approximately 12.6% of the country’s land area and nearly 30% of its population are exposed to landslide risk, particularly in the Himalayas and the Western Ghats. Maharashtra is one of the most severely affected states within the Western Ghats. Its complex geology, intense monsoonal rainfall, and anthropogenic slope modifications substantially increase landslide susceptibility. Moreover, climate change and rapid land-use alterations have expanded hazard exposure into previously stable areas. The interaction between geomorphic processes and growing socio-economic pressures renders landslide risk increasingly complex. However, existing approaches to landslide assessment remain largely static, with limited integration of temporal variability and governance-linked decision frameworks. As a result, an integrated risk assessment framework that combines dynamic hazard modelling with institutional response mechanisms remains underdeveloped.

This research aims to address this gap by developing a multidimensional framework for landslide risk assessment that integrates a dynamic hazard model and scenario-based response strategies. It aims to integrate the landslide models into governance strategies to support planning and decision-making. As an initial step, a detailed analysis of hydro-meteo-geological factors responsible for landslides has been done, and a high-resolution susceptibility map has also been prepared using an ensemble learning model. The ensemble approach was adopted to better capture non-linear relationships between landslide occurrence and multiple conditioning factors. The results indicate that high- and very-high-susceptibility zones are concentrated along the Western Ghats escarpment and the Konkan belt, with some new areas identified across Vidarbha and the Amravati region. An explainable AI, TreeSHAP, is also used to improve transparency and understanding of model behavior for susceptibility analysis. This susceptibility assessment provides a spatial basis for further hazard modeling and risk assessment.

Keywords: Landslide susceptibility mapping.  Ensemble machine learning, Spatio-temporal hazard analysis, Disaster risk reduction, Policy and institutional frameworks.