Akın (2026) Hybrid Deep Learning for Climate-Driven Atmospheric Irrigation Potential Forecasting: A Case Study for Ankara

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

Identification

• Journal: Gazi University Journal of Science Part A Engineering and Innovation
• Year: 2026
• Date: 2026-03-31
• Authors: Murat Akın
• DOI: 10.54287/gujsa.1831069

Research Groups

Not explicitly mentioned in the provided text.

Short Summary

This study developed a hybrid LSTM-XGBoost residual model to forecast monthly atmospheric irrigation potential for Ankara, Türkiye, achieving stable forecasts with a Root Mean Square Error of 24.4 mm and a Coefficient of Determination of 0.87.

Objective

• To analyze and forecast monthly atmospheric irrigation potential for Ankara, Türkiye, under variable climate conditions using a hybrid residual model.

Study Configuration

• Spatial Scale: Ankara, Türkiye (a single location).
• Temporal Scale: Daily climate data from 2001–2024 used for training, monthly atmospheric irrigation potential analyzed, and model tested on the 2022–2024 period.

Methodology and Data

• Models used: Hybrid residual model combining a Long Short-Term Memory (LSTM) network with an Extreme Gradient Boosting (XGBoost) algorithm. Reference evapotranspiration (ET₀) calculated using the Food and Agriculture Organization (FAO) 56 Penman–Monteith equation.
• Data sources: Daily climate data from the National Aeronautics and Space Administration Prediction of Worldwide Energy Resources (NASA POWER) archive for 2001–2024.

Main Results

• The hybrid model achieved a Root Mean Square Error (RMSE) of 24.4 mm, a Mean Absolute Error (MAE) of 20.3 mm, and a Coefficient of Determination (R²) of 0.87 when tested on the 2022–2024 period.
• The hybrid design delivered more stable forecasts compared to standalone models, despite moderate accuracy gains.
• Feature importance analysis revealed that Month Sine and last-step ET₀ had the strongest influence on forecasts, followed by relative humidity, wind speed, and shortwave radiation.

Contributions

• Introduction of a novel hybrid residual model combining LSTM and XGBoost for forecasting atmospheric irrigation potential.
• Demonstration of the model's effectiveness and stability in a semi-arid region (Ankara, Türkiye).
• Identification of key climatic and seasonal features influencing atmospheric irrigation potential, providing practical insights for water management.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Akın2026Hybrid,
  author = {Akın, Murat},
  title = {Hybrid Deep Learning for Climate-Driven Atmospheric Irrigation Potential Forecasting: A Case Study for Ankara},
  journal = {Gazi University Journal of Science Part A Engineering and Innovation},
  year = {2026},
  doi = {10.54287/gujsa.1831069},
  url = {https://doi.org/10.54287/gujsa.1831069}
}