Interactive Learning in Hydrological Modelling With a Web-Based Tool

Web apps offer an effective approach for teaching complex concepts through active learning by removing technical barriers. We present a web-based application of a well-known hydrological bucket-type model (HBV-light) that enables students to explore modelling concepts based on data provided for a sample of 700 global catchments from the CARAVAN dataset. The app, built using R and Shiny, allows users to change model parameters, parameterise the model using Monte Carlo simulation or a genetic algorithm, compare multiple model runs and analyse objective functions, including both the Nash–Sutcliffe Efficiency and the Non-parametric Kling–Gupta Efficiency. Users can select sites through an interactive map interface, download model outputs and visualise results without specialised software or computing resources. The app supports various learning objectives across different course levels, from basic parameter sensitivity analysis to advanced exploration of model structure and catchment behaviour. The app is hosted on Consortium of Universities for the Advancement of Hydrological Sciences Inc.'s shiny server and the code is available on GitHub. Initial classroom trials indicate the app effectively supports understanding of hydrological modelling concepts by the students.