Methods to improve run time of hydrologic models: opportunities and challenges in the machine learning era

TL;DR

The paper investigates how machine learning can speed up physics-based hydrologic models. It surveys opportunities such as surrogate modeling, adaptive sampling, and physics-informed or hybrid ML approaches, alongside challenges like data quality, validation, interpretability, transferability, and extreme-event performance. Concrete methods to improve run time include dimensionality reduction with $PCA$ and $EOF$, parallel computing for training and calibration, and surrogate models or hybrid couplings (e.g., WetSpa-LSTM, Muskingum-Cunge neural routing). The findings suggest substantial potential for rapid hydrologic forecasting and decision-making, though careful evaluation, scalability, and robust validation are required for widespread adoption.

Abstract

The application of Machine Learning (ML) to hydrologic modeling is fledgling. Its applicability to capture the dependencies on watersheds to forecast better within a short period is fascinating. One of the key reasons to adopt ML algorithms over physics-based models is its computational efficiency advantage and flexibility to work with various data sets. The diverse applications, particularly in emergency response and expanding over a large scale, demand the hydrological model in a short time and make researchers adopt data-driven modeling approaches unhesitatingly. In this work, in the era of ML and deep learning (DL), how it can help to improve the overall run time of physics-based model and potential constraints that should be addressed while modeling. This paper covers the opportunities and challenges of adopting ML for hydrological modeling and subsequently how it can help to improve the simulation time of physics-based models and future works that should be addressed.