Towards CONUS-Wide ML-Augmented Conceptually-Interpretable Modeling of Catchment-Scale Precipitation-Storage-Runoff Dynamics

2025-10-02

AI Summary

Conventional large-sample hydrological modeling often lacks physical interpretability, prioritizing predictive accuracy over mechanistic understanding of the precipitation–storage–runoff process. Method: We propose a machine learning–enhanced physically conceptual framework centered on the Mass-Conserving Perceptron (MCP), which intrinsically embeds theoretical guidance and hard physical constraints—particularly mass conservation—into its architecture. MCP is coupled with LSTM units to capture temporal dynamics and integrated with attribute masking (e.g., snow regime, forest cover, climate zone) to enable cross-regional mechanistic identification and generalizability assessment. Results: Evaluated across U.S. watersheds, MCP achieves predictive performance comparable to pure data-driven LSTM models while substantially improving process-based interpretability and physical consistency—e.g., realistic storage–discharge relationships and mass-balance adherence. This work establishes a new paradigm for interpretable, physics-informed hydrological modeling that bridges data-driven flexibility with process understanding.

Technology Category

• Machine Learning: Transparent, Interpretable, Explainable ML
• Natural Language Processing: Interpretability, Analysis, and Evaluation of NLP Models
• Cognitive Modeling & Cognitive Systems: Conceptual Inference and Reasoning

Application Category

• User Modeling, Personalization and Recommendation: Explainable and interpretable methods for personalization
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs

Abstract

While many modern studies are dedicated to ML-based large-sample hydrologic modeling, these efforts have not necessarily translated into predictive improvements that are grounded in enhanced physical-conceptual understanding. Here, we report on a CONUS-wide large-sample study (spanning diverse hydro-geo-climatic conditions) using ML-augmented physically-interpretable catchment-scale models of varying complexity based in the Mass-Conserving Perceptron (MCP). Results were evaluated using attribute masks such as snow regime, forest cover, and climate zone. Our results indicate the importance of selecting model architectures of appropriate model complexity based on how process dominance varies with hydrological regime. Benchmark comparisons show that physically-interpretable mass-conserving MCP-based models can achieve performance comparable to data-based models based in the Long Short-Term Memory network (LSTM) architecture. Overall, this study highlights the potential of a theory-informed, physically grounded approach to large-sample hydrology, with emphasis on mechanistic understanding and the development of parsimonious and interpretable model architectures, thereby laying the foundation for future models of everywhere that architecturally encode information about spatially- and temporally-varying process dominance.

Problem

• Developing interpretable hydrological models for precipitation-storage-runoff dynamics
• Evaluating model complexity across diverse hydro-geo-climatic conditions
• Comparing physically-grounded models with data-driven ML approaches

Innovation

• ML-augmented physically-interpretable catchment-scale models
• Mass-Conserving Perceptron architecture for hydrological modeling
• Theory-informed approach balancing performance and mechanistic understanding