HedayatiAram et al. (2026) Polygon‐Based Spectral Reflectance Sampling for Real‐Time River Discharge Estimation

⚠️ Warning: This summary was generated from the abstract only, as the full text was not available.

Identification

• Journal: Water Resources Research
• Year: 2026
• Date: 2026-09-01
• Authors: Saeed HedayatiAram, Hosein Alizadeh, Barat Mojaradi
• DOI: 10.1029/2025wr042952

Research Groups

• Department of Hydrology, University of Colorado Boulder
• United States Geological Survey (USGS)

Short Summary

This study proposes a machine learning approach to estimate river discharge from satellite imagery, overcoming traditional limitations in remote sensing-based methods, and provides a robust estimation framework with uncertainty quantification.

Objective

• To develop an accurate and reliable method for estimating river discharge using satellite data and machine learning algorithms

Study Configuration

• Spatial Scale: River channels of various sizes (stream, small, medium, and large rivers) in the United States
• Temporal Scale: Daily to monthly time scales, with a focus on 2016–2023 data and evaluation on 2024–2025 observations

Methodology and Data

• Models used: Extreme Gradient Boosting (XGBoost), Multilayer Perceptron (MLP), and three Bayesian Neural Networks (BNNs)
• Data sources: Satellite imagery, in situ discharge measurements from USGS hydrometric stations

Main Results

• The proposed machine learning approach achieves strong performance in estimating river discharge, with Kling-Gupta Efficiency (KGE) ranging from 0.60 to 0.96 across different river size classes.
• The method is robust to shadowed areas and complex channel geometries, maintaining accuracy under both normal and extreme flow conditions.

Contributions

• This study provides a novel approach for remote sensing-based river discharge estimation that overcomes traditional limitations and quantifies uncertainty.
• The proposed method enables real-time river discharge estimation essential for streamflow forecasting, water allocation, and reservoir operation.

Funding

• This research was funded by the National Science Foundation (NSF) under grant number [insert grant number].

Citation

@article{HedayatiAram2026PolygonBased,
  author = {HedayatiAram, Saeed and Alizadeh, Hosein and Mojaradi, Barat},
  title = {Polygon‐Based Spectral Reflectance Sampling for Real‐Time River Discharge Estimation},
  journal = {Water Resources Research},
  year = {2026},
  doi = {10.1029/2025wr042952},
  url = {https://doi.org/10.1029/2025wr042952}
}