Water and land management, for example to limit the discharge of pollutants into water resources, relies on accurate model predictions of subsurface water flow and solute transport at scales of hillslopes and catchments. Issues at these scales arise from unknown heterogeneity of subsurface properties and simplified representation of the flow and transport dynamics. Within hydrologic models these simplifications require closure relationships (representing the aggregated small-scale physics) that vary in time and space. Limited measurement capabilities in the subsurface and difficult experimentation at the hillslope scale and beyond hamper the search for these closure relationships, that are needed for better water and land management policies. This project aims at discovering the nature of these closure relationships and will make an important step towards an improved understanding directly at the hillslope scale. This project will examine the closure relationships in a hillslope scale experimental system and explore its generalizability to real-world catchments. The hillslope scale study will take advantage of the exceptional measurement capability of the Landscape Evolution Observatory (LEO) hillslopes at Biosphere 2, The University of Arizona. Data collected from recent experiments at the LEO hillslopes will be utilized to find and explain temporally changing closure relationships for two different levels of model complexity: (i) the system-scale closure relationships, e.g., transit time distributions and the storage-discharge relationship, and (ii) the simplified process-based representation of coupled unsaturated and saturated zone of the hybrid-3D hillslope hydrological model (h3D). The project will gain novel insights into the closure relationships, e.g., their temporal variability and their connection to other hydrologic variables, by combining recent methodological advances, namely (i) new data-based approaches to determine transit time distributions and the storage-discharge relationship, (ii) data assimilation, and (iii) machine learning methods. The transfer to real-world cases will be tested on a well-equipped zero-order basin. This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.