özden (2025) Data-Driven Decision Support in Environmental Management: Hybrid GNN-PINN Modeling of Subsurface Soil Temperature

Identification

• Journal: Mendeley Data
• Year: 2025
• Date: 2025-12-05
• Authors: özden, cevher
• DOI: 10.17632/kwj724c62f

Research Groups

• Cukurova Universitesi

Short Summary

This paper describes a comprehensive dataset of daily meteorological and subsurface soil temperature records from 15 stations across Turkey, spanning five years, specifically designed to support research in environmental decision support systems and the development of hybrid Physics-Informed Neural Networks (PINNs) and Graph Neural Networks (GNNs).

Objective

• To provide a rigorously preprocessed and structured dataset for researchers developing advanced machine learning models (PINNs and GNNs) for soil temperature forecasting, thermal diffusion modeling, agricultural optimization, and climate generalization analysis under extreme seasonal conditions.

Study Configuration

• Spatial Scale: 15 distinct regional meteorological stations across diverse climatic and geographic zones in Turkey.
• Temporal Scale: Five-year period from January 1, 2020, to December 31, 2024, with daily resolution.

Methodology and Data

• Models used: The dataset is intended for training and benchmarking hybrid Physics-Informed Neural Networks (PINNs) and Graph Neural Networks (GNNs).
• Data sources: Daily meteorological records obtained from the Turkish State Meteorological Service (MGM) via the MEVBIS platform.
  • Subsurface Data: Daily mean, minimum, and maximum soil temperatures recorded at five distinct depths: 0.05 m, 0.10 m, 0.20 m, 0.50 m, and 1.00 m.
  • Surface Data: Daily mean 2-meter air temperature (K).
  • Geospatial Metadata: Station-specific latitude, longitude, and altitude (m).
  • Data Processing: Temporal filtering to remove corrupted entries and sequences with excessive missing values; min-max normalization of temperature values to the range [0, 1]; structuring data to support depth encoding as a continuous input variable.

Main Results

• The primary result is a comprehensive, preprocessed, and structured dataset suitable for advanced machine learning applications in environmental management.
• The dataset includes daily soil temperature profiles at multiple depths and surface air temperature, along with geospatial metadata for 15 stations across Turkey.
• The data is prepared to facilitate detailed visual and statistical analysis across seasons and locations, supporting the development and benchmarking of hybrid GNN-PINN architectures.

Contributions

• Provides a unique, multi-depth soil temperature and meteorological dataset specifically curated and preprocessed for the development of hybrid GNN-PINN models.
• Enables research in data-driven environmental decision support systems, precision agriculture, and advanced thermal diffusion modeling.
• Offers a valuable resource for analyzing climate generalization under varying seasonal and geographic conditions within Turkey.

Funding

• Not explicitly stated in the provided text.

Citation

@article{özden2025DataDriven,
  author = {özden, cevher},
  title = {Data-Driven Decision Support in Environmental Management: Hybrid GNN-PINN Modeling of Subsurface Soil Temperature},
  journal = {Mendeley Data},
  year = {2025},
  doi = {10.17632/kwj724c62f},
  url = {https://doi.org/10.17632/kwj724c62f}
}