Authors: Ziyu Li – Colorado School of Mines
Title: Incorporating a differentiable version of CFE into Neural Hydrology to train CFE parameters for use in NextGen
Presentation Type: Poster
Abstract: Like any successful modeling framework, NextGen requires parameter estimation to achieve high-quality calibrated model simulations. This crucial task can be challenging even with sufficient observations to constrain local model parameters and far more difficult when estimating parameters across large domains – a process known as parameter regionalization. In recent years, the use of machine learning in differentiable hydrologic modeling has shown potential. Here, a neural network learns to predict model parameters from meteorological forcings and geophysical catchment attributes by updating its weights using gradient-based optimization to minimize a loss function that quantifies the discrepancy between the conceptual model’s simulations and observations. Such a model trained over a large set of basins at once will learn regional hydrological behaviors and can be used for parameter regionalization. In this collaborative CIROH project between CSM and U. Alabama, we apply this approach for NextGen’s CFE model by embedding a differentiable version (dCFE) into the Neural Hydrology (NH-dCFE) platform for training and extracting the trained LSTM to use in CFE parameter regionalization across CONUS. This presentation describes this effort, including the validation of NH-dCFE to dCFE and CFE, successes in training NH-dCFE models, and the challenges encountered. We also offer recommendations on strategies to advance this parameter estimation approach in the future.