123 - Advancing Climate Risk Analysis using Machine Learning and Statistical Assessments for Hydro-Climatic Extremes in Southeast Texas and Southwest Louisiana

The Southeast Texas and Southwest Louisiana region is increasingly recognized as a climate hotspot, vulnerable to a range of extreme climatic events. Historical instances from devastating droughts to frequent hurricanes and tropical storms leading to high flood events, underscore the urgency of understanding climate extremes in this area. The study focuses on quantitative assessment of hydro-climatological extremes, evaluating both their severity and frequency across historical watershed data and downscaled future prediction scenarios. This approach serves as a robust analytical framework for climate risk evaluation. Building on observed historical flow and various drought indices such as the Standardized Precipitation Index (SPI), Standardized Precipitation Evapotranspiration Index (SPEI), Standardized Runoff Index (SRI), and Standardized Soil Water Index (SSWI), the study develops new drought and flood vulnerability indices tailored to the region's unique climatic challenges. Machine learning methods are employed to conduct and validate the analysis on historical data, ensuring model reliability and robustness. Statistical tests, including trend assessments, change point detection and stationarity tests are performed for severity and magnitude of drought and flood. Once validated, these machine learning models are applied to future projection data, leveraging parameters available in downscaled Global Circulation Models (GCM), especially CMIP5 and CMIP6. This multi-tiered approach combining machine learning-algorithms, specialized climate indices, and rigorous statistical tests not only amplifies the study's immediate relevance but also equips planers and decision makers with predictive capabilities essential for long-term climate risk evaluation and management.