Roquia Salam

Explainable AI-driven Digital Twin-Enabling Technologies for Rainfall-Induced Landslide Forecasting in Bangladesh

All over the world, a significant number of extreme landslides occur during the rainy season in mountainous areas. These rainfall-induced landslides are recurrent and catastrophic disasters, with long-lasting impacts on people's lives, livelihoods, critical infrastructure, and sustainable development. Several landslide forecasting systems are operational globally to provide real-time forecasts of upcoming landslide events to primary stakeholders, enabling affected populations to take proactive measures to reduce risk. Thus, these forecasting systems help minimize the overall loss and damage caused by landslides.

Developing these operational forecasting systems requires extensive data from various sources, such as meteorological, hydrological, geological, lithological, and sensor data. However, many regions affected by rainfall-induced shallow landslides lack such forecasting systems, leading to significant losses and damages every year. These areas are often data-sparse, lacking the necessary data mentioned above. Therefore, there is a need to develop landslide forecasting systems for these data-sparse regions.

Explainable artificial intelligence (AI) can help identify the appropriate factors from the available data that contribute to landslide events. Digital twins enable the step-by-step execution of all necessary components for developing a forecasting system, from continuous data collection to user interaction. This research will utilize both explainable AI and digital twin technologies to develop a forecasting system for rainfall-induced shallow landslides. The south-eastern part of Bangladesh has been selected as a case study.

Selected publications

• Salam, R., & Islam, A. R. M. T. (2020). Potential of RT, bagging and RS ensemble learning algorithms for reference evapotranspiration prediction using climatic data-limited humid region in Bangladesh. Journal of Hydrology, 590, 12524.
• Salam, R., Islam, A. R. M. T., Pham, Q. B., Dehghani, M., Al-Ansari, N., & Linh, N. T. T. (2020). The optimal alternative for quantifying reference evapotranspiration in climatic sub-regions of Bangladesh. Scientific reports, 10(1), 1-21.
• Salam, R., Islam, A. R. M., & Islam, S. (2019). Spatiotemporal distribution and prediction of groundwater level linked to ENSO teleconnection indices in the northwestern region of Bangladesh. Environment, Development and Sustainability: A Multidisciplinary Approach to the Theory and Practice of Sustainable Development, 22(5), 4509-453.
• Salam, R., Ghose, B., Shill, B. K., Islam, M. A., Islam, A. T., Sattar, M. A., Alam, G.M.M., & Ahmed, B. (2021). Perceived and actual risks of drought: household and expert views from lower Teesta River Basin of Northern Bangladesh, Natural Hazards, 108, 2569–2587.
• Salam, R., Islam, A. R. M. T., Shill, B. K., Alam, G. M., Hasanuzzaman, M., Hasan, M. M., ... & Shouse, R. C. (2021). Nexus between vulnerability and adaptive capacity of drought-prone rural households in northern Bangladesh. Natural Hazards, 106(1), 509-527.

Experience

• Research Associate (October 2021–September 2022) – Centre for Genocide Studies (CGS) at the University of Dhaka (DU), Bangladesh

Qualifications

• MSc in Geospatial Technologies, Westfälische Wilhelms-Universität, Universitat Jaume I, Universidade Nova de Lisboa, 2022–2024
• MSc in Disaster Management, Begum Rokeya University, 2019–2020
• BSc in Disaster Management, Begum Rokeya University, 2014–2019

Achievements and awards

• UCL Research Excellence Scholarship (RES), awarded in 2023 from UCL
• Erasmus Mundus Joint Master Scholarship, awarded in 2022 by the European Commission
• Bangabandhu Merit Scholarship, awarded in 2021 from Begum Rokeya University