PNNL

Abstract

Watershed model parameters, such as the subsurface permeability field, are difficult or expensive to measure directly at the spatial extent and resolution required by mechanistic watershed models. Within a watershed, stream discharge measured at stream gauges is usually available with historical record, and thus can be used to infer soil and geologic properties using inverse modeling. However, very few studies have applied inverse modeling assisted by machine learning to infer subsurface permeability. In this study, we perform machine learning (ML)-assisted hydrologic inverse modeling to estimate soil and geologic permeability using observed hydrographs. We train three deep neural networks (DNNs) model architectures to predict subsurface permeability from simulated stream discharges using an fully-distributed, integrated surface-subsurface hydrologic model. We show that the permeability with larger spatial coverage has better predictability. Compared to traditional ensemble smoother method, DNNs show stronger performance in accurately estimating the subsurface permeability. Our results show that stream discharge may be used to inverse subsurface permeability utilizing the state-of-art DNNs and integrated watershed models.