Simulation-Based Inference for Parameter Estimation of Complex Watershed Simulators

Abstract. High-resolution, spatially distributed process-based (PB) simulators are widely employed in the study of complex watershed processes and their responses to a changing climate. However, calibrating these simulators to observed data remains a significant challenge due to several persistent issues including: (1) intractability stemming from the computational demands and complex responses of simulators, which renders infeasible calculation of the conditional probability of parameters and data, and (2) uncertainty stemming from the choice of simplified model representations of complex natural hydrologic processes. Here we demonstrate how Simulation-Based Inference (SBI) can help address both these challenges for parameter estimation. SBI uses a learned mapping between parameter space and observed data to estimate parameters for generation of calibrated model simulations. To demonstrate the potential of SBI in hydrologic modelling, we conduct a set of synthetic experiments to infer two common physical parameters, Manning's coefficient and hydraulic conductivity, using a representation of a snowmelt-dominated catchment in Colorado, USA. We introduce novel deep learning (DL) components to the SBI approach, including an 'emulator' as a surrogate for the process-based simulator to rapidly explore parameter responses. We also employ a density-based neural network to represent the joint probability of parameters and data without strong assumptions about its functional form. While addressing intractability, we also show that where uncertainty about model structure is significant, SBI can yield unreliable parameter estimates. Approaches to adopting the SBI framework to cases where model structure(s) may not be adequate are introduced using a performance-weighting approach.