Spatiotemporal prediction and attribution of groundwater storage anomaly using enhanced hybrid deep learning modeling with uncertainty quantification

Key Points

• Groundwater storage anomalies were effectively predicted with high accuracy using advanced deep learning techniques, enhancing forecasting abilities.
• The modeling approach integrates uncertainty quantification to evaluate prediction reliability, providing insights into potential groundwater resource management.
• Using hybrid deep learning models, changes in groundwater storage over time and space were assessed, demonstrating effective anomaly attribution.
• The findings highlight the need for robust modeling methods in environmental monitoring, which may improve management strategies for water resources.