Beyond One-Size-Fits-All: Evaluating Flash Drought Detection Methods and Impact Assessment Across CONUS Catchments

Presentation at the American Geophysical Union General Meeting 2025.
Abstract: Flash droughts develop rapidly within days to weeks, challenging existing early warning systems through their abrupt onset and compound effects with wildfires and heatwaves. Despite growing attention, the absence of standardized definitions and detection methods hampers effective monitoring and response, as well as the assessment of their  societal and environmental impacts. This study addresses these gaps through a comprehensive 40-year catchment-scale analysis across the Contiguous United States (CONUS) using six common flash drought indicators. We quantify detection consistency, assess inter-method agreement, and evaluate operational trade-offs between single- and multi-indicator approaches.  Our results reveal substantial inconsistencies across flash drought detection methods, even among those using similar variables. Multi-indicator frameworks improve robustness but may miss events due to strict thresholds, while single-indicator methods may lead to over-detection.  These trade-offs highlight the critical importance of aligning detection strategies with regional hydroclimatic contexts and observed impacts. To bridge this gap, we propose a new impact-based monitoring framework using Natural Language Processing (NLP) to extract and classify flash drought impacts from media and newspaper reports. We advocate for moving beyond a one-size-fits-all framework toward region- and sector-specific detection protocols that support more adaptive and responsive water management systems. The planned classification and impact database will establish the foundation for an operational, impact-based monitoring system that identifies flash droughts and anticipates their consequences, providing actionable information for decision-makers and the public. By validating significant events against observed societal impacts, this research will contribute to building a robust validation framework that enhances detection methods through real-world impact data.