Talha et al. (2026) Watershed-agnostic modeling: An explainable, distributed, data-driven hydrologic model

Identification

• Journal: Environmental Modelling & Software
• Year: 2026
• Date: 2026-09-01
• Authors: Muhammad Talha, Amir Pouyan Nejadhashemi
• DOI: 10.1016/j.envsoft.2026.107149

Research Groups

• Department of Biosystems and Agricultural Engineering, Michigan State University
• Department of Computer Science and Engineering, Michigan State University
• Department of Plant, Soil and Microbial Sciences, Michigan State University

Short Summary

This study introduces a data-driven framework that combines Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) layers to estimate next-day discharge across a dense network of Michigan gauges. The method achieves predominantly 'good' to 'very good' Nash-Sutcliffe and Kling-Gupta efficiencies at most stations.

Objective

• Develop an explainable, distributed, data-driven hydrologic model that can estimate streamflow without the need for watershed delineation or other hydrologic preprocessing.

Study Configuration

• Spatial Scale: Watershed-agnostic modeling across a dense network of Michigan gauges.
• Temporal Scale: Next-day discharge estimation using multi-day gridded weather layers.

Methodology and Data

• Models used: Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) layers.
• Data sources: Multi-day gridded weather layers, decade-long streamflow records from Michigan gauges.

Main Results

• The model achieves predominantly 'good' to 'very good' Nash-Sutcliffe and Kling-Gupta efficiencies at most stations.
• Gradient-based saliency shows that the model focuses on hydrologically plausible source regions.

Contributions

• This study provides skillful, transparent, and efficient streamflow guidance suitable for basin-scale water resource decision support.
• The method offers a digital analog to distributed process models, eliminating the need for watershed delineation or other hydrologic preprocessing.

Funding

• This research was funded by [project/program name] (reference code: [insert reference code]).

Citation

@article{Talha2026Watershedagnostic,
  author = {Talha, Muhammad and Nejadhashemi, Amir Pouyan},
  title = {Watershed-agnostic modeling: An explainable, distributed, data-driven hydrologic model},
  journal = {Environmental Modelling & Software},
  year = {2026},
  doi = {10.1016/j.envsoft.2026.107149},
  url = {https://doi.org/10.1016/j.envsoft.2026.107149}
}