Announcement of Ph.D. Dissertation/M.S. Thesis Oral Defense
Student: Chathuranika Imiya Mudiyanselage Advisor: Dr. Dalya Ismael
Building: Kaufman Hall Room: 136
Day: Monday, 13th April 2026 Time: 1:30 PM – 3:00 PM
Dissertation/Thesis Title: Advancing Environmental Risk Assessment Through Hydro-Climatological Modeling, Geospatial Analytics, and Machine Learning
Abstract: Environmental systems are increasingly affected by climate variability, environmental hazards, and limitations in hydro-climatological data availability, creating significant challenges for accurate environmental risk assessment and sustainable resource management. This dissertation develops an integrated analytical framework that combines hydro-climatological modeling, geospatial analysis, statistical data reconstruction, remote sensing, and machine learning techniques to enhance environmental monitoring, prediction, and risk assessment. A central component of this research focuses on improving the reliability and completeness of hydro-climatological datasets used in environmental analysis. Satellite-based precipitation products were evaluated and bias-adjusted to improve their accuracy for hydrological applications in data-limited environments, while statistical precipitation reconstruction techniques were assessed across contrasting physiographic regions to understand how terrain characteristics influence data reconstruction performance and hydrological modeling reliability. These analyses demonstrate the importance of region-specific approaches for improving precipitation datasets used in environmental modeling. The dissertation also investigates climate change impacts on watershed hydrology using hydrological modeling and future climate projections, revealing that changing climate conditions may alter precipitation patterns, hydrological processes, and streamflow dynamics, thereby increasing variability in water availability and the likelihood of hydrological extremes. Comparative analysis of watersheds in different climatic zones further highlights that climate change effects on hydrology vary significantly depending on regional climate regimes and watershed characteristics, emphasizing the need for location-specific adaptation strategies. Beyond hydrological processes, the research expands environmental risk assessment through geospatial modeling of climate-driven hazards, including wildfire susceptibility analysis under future climate scenarios to identify landscapes where climate change may increase wildfire risk. In addition, groundwater potential mapping using geographic information systems, remote sensing data, and multi-criteria decision analysis was conducted to identify areas suitable for sustainable groundwater development and resource protection. The dissertation also demonstrates the application of deep learning techniques for hydrological forecasting by implementing recurrent neural network architectures to predict river discharge using hydro-climatological time-series data, showing that machine learning approaches can effectively capture temporal patterns in watershed behavior and support operational water resource management. This research demonstrates that integrating hydro-climatological modeling, geospatial analytics, and machine learning provides a comprehensive framework for advancing environmental risk assessment by improving environmental data reliability, enhancing predictive modeling capabilities, and supporting climate-informed decision-making for water resources, hazard mitigation, and sustainable environmental management in regions experiencing increasing environmental variability and climate change.