This body of work applies natural language processing, explainable machine learning, and multi-source environmental information to improve drought impact monitoring and assessment. The goal is to capture impacts that are difficult to observe directly through conventional hydro-meteorological datasets alone.