Estimating parameters for distributed hydrological models is challenging, especially in regions with limited data. Here the authors employ variational autoencoders to generate optimal parameter ...