Search for: All records

Large-scale hydrologic models are increasingly being developed for operational use in the forecasting and planning of water resources. However, the predictive strength of such models depends on how well they resolve various functions of catchment hydrology, which are influenced by gradients in climate, topography, soils, and land use. Most assessments of hydrologic model uncertainty have been limited to traditional statistical methods. Here, we present a proof-of-concept approach that uses interpretable machine learning techniques to provide post hoc assessment of model sensitivity and process deficiency in hydrologic models. We train a random forest model to predict the Kling–Gupta efficiency (KGE) of National Water Model (NWM) and National Hydrologic Model (NHM) streamflow predictions for 4383 stream gauges in the conterminous United States. Thereafter, we explain the local and global controls that 48 catchment attributes exert on KGE prediction using interpretable Shapley values. Overall, we find that soil water content is the most impactful feature controlling successful model performance, suggesting that soil water storage is difficult for hydrologic models to resolve, particularly for arid locations. We identify nonlinear thresholds beyond which predictive performance decreases for NWM and NHM. For example, soil water content less than 210 mm, precipitation less than 900 mm yr−1, road density greater than 5 km km−2, and lake area percent greater than 10 % contributed to lower KGE values. These results suggest that improvements in how these influential processes are represented could result in the largest increases in NWM and NHM predictive performance. This study demonstrates the utility of interrogating process-based models using data-driven techniques, which has broad applicability and potential for improving the next generation of large-scale hydrologic models. 

The interaction between climate and the hydrologic cycle is complex due to intricate feedback mechanisms that can have multiple impacts on key hydrologic variables. Under a changing climate, it is becoming increasingly important for undergraduate engineering students to have a better understanding of climate and the hydrologic cycle to ensure future engineering systems are more climate resilient. One way of teaching undergraduate students about these key interactions between climate and the hydrologic cycle is through numerical models that mimic these relationships. However, this is difficult to do in an undergraduate engineering course because these models are complex, and it is not feasible to devote class time and resources to teaching students the knowledge base required to run and analyze these numerical models. In addition, the recent COVID-19 pandemic required a rapid change to flexible teaching methods that can be implemented in online, hybrid, or in-person courses. To overcome these limitations, a backward design and constructive alignment approach was used to develop an active learning module in the HydroLearn framework that allows students to explore the connection between snow processes and streamflow and how this will change under different climate scenarios using numerical models and analysis. This learning module provides learning activities and tools that help the student develop a basic knowledge of snow formation and terminology, snow measurements, numerical models of snow processes, and changes in snow and streamflow under future climate. This module is particularly innovative in that it uses Google Colabs and an interactive user interface to facilitate the students' active learning in an environment that is accessible for all students and is sustainable for continued use and adaptation. This paper describes the approach, best practices and lessons learned in developing and implementing this active learning module in a remote and in-person course. In addition, it presents the results from motivation and student self-assessment surveys and discusses opportunities for improvement and further implementation that have implications for the future of hydrologic education.