Replication Study: Enhancing Hydrological Modeling with Physics-Guided Machine Learning

Mostafa Esmaeilzadeh,Melika Amirzadeh

2024-02-22

Abstract:Current hydrological modeling methods combine data-driven Machine Learning (ML) algorithms and traditional physics-based models to address their respective limitations incorrect parameter estimates from rigid physics-based models and the neglect of physical process constraints by ML algorithms. Despite the accuracy of ML in outcome prediction, the integration of scientific knowledge is crucial for reliable predictions. This study introduces a Physics Informed Machine Learning (PIML) model, which merges the process understanding of conceptual hydrological models with the predictive efficiency of ML algorithms. Applied to the Anandapur sub-catchment, the PIML model demonstrates superior performance in forecasting monthly streamflow and actual evapotranspiration over both standalone conceptual models and ML algorithms, ensuring physical consistency of the outputs. This study replicates the methodologies of Bhasme, P., Vagadiya, J., & Bhatia, U. (2022) from their pivotal work on Physics Informed Machine Learning for hydrological processes, utilizing their shared code and datasets to further explore the predictive capabilities in hydrological modeling.

Machine Learning

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the limitations of current hydrological models in predicting monthly runoff and actual evapotranspiration. Specifically, although traditional physics - based models can provide process understanding, there may be biases in parameter estimation; while data - driven machine learning algorithms, although highly accurate in result prediction, ignore the constraints of physical processes, resulting in insufficient physical consistency. Therefore, this study proposes a physics - informed machine learning (PIML) model, aiming to combine the process understanding of conceptual hydrological models and the prediction efficiency of machine learning algorithms to improve prediction performance and ensure the physical consistency of outputs. The PIML model embeds machine learning algorithms into the key steps of the hydrological model, replacing empirical equations to determine the relationships between input variables (such as potential evapotranspiration PET, precipitation P, groundwater level GW and soil moisture SM), intermediate variables (such as actual evapotranspiration ET) and target variables (such as runoff Q at a specific gauging station). This method not only retains the interpretability of physical models and the advantages of physical information selection, but also can identify complex non - linear relationships, thereby improving the generalization ability and prediction accuracy of the model. Through the application in the Anandapur sub - basin, this study verifies the superior performance of the PIML model in predicting monthly runoff and actual evapotranspiration. Especially when dealing with long - sequence time data, it shows better effects than individual conceptual models or machine learning algorithms. This indicates that the PIML model has important application potential in hydrological forecasting and prediction.

Replication Study: Enhancing Hydrological Modeling with Physics-Guided Machine Learning

Enhancing predictive skills in physically-consistent way: Physics Informed Machine Learning for hydrological processes

Physics-aware Machine Learning Revolutionizes Scientific Paradigm for Machine Learning and Process-based Hydrology

Assessment of monthly runoff simulations based on a physics-informed machine learning framework: The effect of intermediate variables in its construction

A Physical Process and Machine Learning Combined Hydrological Model for Daily Streamflow Simulations of Large Watersheds with Limited Observation Data

Physics-Guided Machine Learning for Scientific Discovery: An Application in Simulating Lake Temperature Profiles

Physics Guided Machine Learning Methods for Hydrology

Generalizability of deep-learning-based emulation of hydrodynamic flood models

Comparison of novel physics-guided machine learning models with empirical equations for predicting longitudinal dispersion coefficient in diverse natural river systems

Using a physics-based hydrological model and storm transposition to investigate machine-learning algorithms for streamflow prediction

Physics-enhanced machine learning models for streamflow discharge forecasting

A Review of Physics-Informed Machine Learning in Fluid Mechanics

Distributed Hydrological Modeling With Physics‐Encoded Deep Learning: A General Framework and Its Application in the Amazon

A New Benchmark on Machine Learning Methodologies for Hydrological Processes Modelling: A Comprehensive Review for Limitations and Future Research Directions

When Physics Meets Machine Learning: A Survey of Physics-Informed Machine Learning

A Critical Review of Physics-Informed Machine Learning Applications in Subsurface Energy Systems

Physics-guided Deep Learning for Rainfall-Runoff Modeling by Considering Extreme Events and Monotonic Relationships

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

Can artificial intelligence and data-driven machine learning models match or even replace process-driven hydrologic models for streamflow simulation?: A case study of four watersheds with different hydro-climatic regions across the CONUS

Performance evaluation of ML techniques in hydrologic studies: Comparing streamflow simulated by SWAT, GR4J, and state-of-the-art ML-based models

Developing a Physics-Informed Deep Learning Model to Simulate Runoff Response to Climate Change in Alpine Catchments