Understanding and predicting fluxes of water and nutrients across watersheds and river basins are important for future water security. New-generation models capable of resolving processes across a range of spatio-temporal scales are needed to accurately predict watershed function. Such models need access to high-resolution and diverse data accounting for variations in topography, vegetation, subsurface properties, and biogeochemical gradients. Moreover, these models should be able to simulate stresses on global water resources from climate change, extreme weather, and land-use change. Machine learning models, process-based models, and hybrid models are all promising avenues for simulating watershed-scale processes. Here, we invite contributions that (1) integrate diverse and multi-scale data into watershed models at desired spatio-temporal resolutions, (2) model watershed-scale processes using process-based, data-driven, and hybrid approaches, (3) quantify uncertainty in observations and model outputs, and (4) incorporate emerging high-performance techniques such as 5G, edge computing, and quantum computing.