Enhancing SWAT-based ecosystem modeling through satellite data and hybrid AI under climate change


Yeşilyurt S. N., Onuşluel Gül G.

ENVIRONMENTAL MODELLING & SOFTWARE, cilt.204, ss.107111, 2026 (SCI-Expanded, Scopus)

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 204
  • Basım Tarihi: 2026
  • Doi Numarası: 10.1016/j.envsoft.2026.107111
  • Dergi Adı: ENVIRONMENTAL MODELLING & SOFTWARE
  • Derginin Tarandığı İndeksler: Applied Science & Technology Source, Academic Search Ultimate (EBSCO), Engineering Source (EBSCO), Scopus, Science Citation Index Expanded (SCI-EXPANDED), Compendex, Environment Index, Geobase, Greenfile, INSPEC, Public Affairs Index
  • Sayfa Sayıları: ss.107111
  • Erzincan Binali Yıldırım Üniversitesi Adresli: Hayır

Özet

As climate change intensifies, reliable ecosystem models are becoming important for sustainable water management. This study aims to improve SWAT-based ecosystem modeling by minimizing predictive uncertainties in SWAT simulations and providing more reliable streamflow predictions. To assess data source effects, the SWAT model was driven by observational and satellite-based meteorological data, demonstrating the effectiveness of satellite inputs in data-scarce regions. To reduce discrepancies between simulated and observed flows, SWAT-TCN and SWAT-CatBoost models were integrated into a Regime-Adaptive Ensemble (RAE)-based SWAT + TCN–CatBoost framework. Model performance was evaluated using statistical metrics and TOPSIS and VIKOR, while SHAP was used for interpretation. The results indicate that regime-adaptive hybrid approach improved SWAT performance, with RAE Hybrid Model giving the best validation results (NSE = 0.90, KGE = 0.87). Subsequently, the framework was applied to future projections using three GCMs and two SSP scenarios, demonstrating more reliable results under climate change.