Yeşilyurt et al. (2026) Enhancing SWAT-based ecosystem modeling through satellite data and hybrid AI under climate change

Identification

• Journal: Environmental Modelling & Software
• Year: 2026
• Date: 2026-07-15
• Authors: Sefa Nur Yeşilyurt, Gülay Onuşluel Gül
• DOI: 10.1016/j.envsoft.2026.107111

Research Groups

• Graduate School of Natural and Applied Sciences, Dokuz Eylül University, Izmir, Türkiye
• Department of Civil Engineering, Dokuz Eylül University, Izmir, Türkiye

Short Summary

This study enhances SWAT-based ecosystem modeling by integrating satellite data and hybrid AI under climate change, improving predictive uncertainties in streamflow simulations.

Objective

• Investigate the effectiveness of satellite-based meteorological data on SWAT model performance in data-scarce regions.
• Develop a regime-adaptive ensemble framework to reduce discrepancies between simulated and observed flows using hybrid machine learning models (SWAT-TCN and SWAT-CatBoost).

Study Configuration

• Spatial Scale: Basin-scale processes under changing hydroclimatic conditions.
• Temporal Scale: Long-term coverage, with a focus on climate change projections.

Methodology and Data

• Models used:
  • SWAT (Soil and Water Assessment Tool) physically based model
  • SWAT-TCN (SWAT integrated with Temporal Convolutional Networks)
  • SWAT-CatBoost (SWAT integrated with CatBoost machine learning algorithm)
  • Regime-Adaptive Ensemble (RAE)-based framework
• Data sources:
  • NASA POWER project dataset for satellite-based meteorological data
  • Observational data for comparison

Main Results

• The regime-adaptive hybrid approach improved SWAT performance, with the RAE Hybrid Model giving the best validation results (NSE = 0.90, KGE = 0.87).
• Future projections using three GCMs and two SSP scenarios demonstrated more reliable results under climate change.

Contributions

• This study provides a systematic comparison of observational and satellite-based data in hydrological modeling.
• The developed framework enhances model reliability while preserving physical representation of watershed processes.

Funding

• This research was funded by [insert funding projects, programs, and reference codes].

Citation

@article{Yeşilyurt2026Enhancing,
  author = {Yeşilyurt, Sefa Nur and Gül, Gülay Onuşluel},
  title = {Enhancing SWAT-based ecosystem modeling through satellite data and hybrid AI under climate change},
  journal = {Environmental Modelling & Software},
  year = {2026},
  doi = {10.1016/j.envsoft.2026.107111},
  url = {https://doi.org/10.1016/j.envsoft.2026.107111}
}