Hydrological Process Surrogate Modelling and Simulation with Neural Networks

摘要: Environmental sustainability is a major concern for urban and rural development. Actors stakeholders need economic, effective efficient simulations in order to predict evaluate the impact of development on environment constraints that imposes Numerical simulation models are usually computation expensive require expert knowledge. We consider problem hydrological modelling simulation. With training set consisting pairs inputs outputs from an off-the-shelves simulator, show neural network can learn surrogate model effectively efficiently thus be used as model. Moreover, we argue model, although trained some example terrains, generally capable simulating terrains different sizes spatial characteristics.