Babaei et al. (2025) Geospatial analysis of Lake Urmia’s drying: predicting land surface temperature changes using remote sensing and machine learning

Identification

• Journal: Advances in Space Research
• Year: 2025
• Date: 2025-10-10
• Authors: M. Babaei, Fatemeh Rajaei, Hamid Siroosi, Nasrin Alamdari
• DOI: 10.1016/j.asr.2025.10.013

Research Groups

• Department of Environmental Sciences, Faculty of Science, University of Zanjan, Zanjan, Iran
• FAMU-FSU College of Engineering, United States

Short Summary

This study investigates the impact of Lake Urmia's desiccation on regional cooling effects and predicts future land surface temperature changes from 1986 to 2035 using remote sensing and machine learning. It found a drastic reduction in Lake Urmia's area, leading to a significant increase in regional land surface temperature and a diminished cooling effect, with projections indicating further warming.

Objective

• To investigate the impact of Lake Urmia’s desiccation on regional cooling effects from 1986 to 2035.
• To predict future land surface temperature changes around Lake Urmia using remote sensing and machine learning.

Study Configuration

• Spatial Scale: Region around Lake Urmia, specifically within 5 km and 30 m buffers.
• Temporal Scale: 1986 to 2035, with specific analysis periods in 1986, 1992, 1998, 2004, 2010, 2016, 2023, and predictions for 2029 and 2035.

Methodology and Data

• Models used:
  • Object-oriented-pixel based land use classification
  • ANN-CA (Artificial Neural Network - Cellular Automata) algorithm for future land use change prediction
  • Machine learning regression algorithms for Land Surface Temperature (LST) prediction: Random Forest, Extra trees, Decision tree, Ridge, Lasso, Linear Regression, Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and XGBoost.
• Data sources:
  • Remote sensing images from Landsat 5, 7, and 8 satellites.
  • Environmental and human variables (used in machine learning models).

Main Results

• The area of Lake Urmia decreased from 5,151,740,400 square meters (1986) to 646,203,600 square meters (2023), and is predicted to reach 513,092,700 square meters by 2035.
• The land surface temperature class above 50 degrees Celsius, which covered 43.03 % of the region in 2023, is predicted to increase to 69.22 % of the entire region by 2035.
• The maximum cooling effect distance of the lake was 1050 meters in 1998, which drastically decreased to 90 meters by 2023.

Contributions

• Provides a comprehensive geospatial analysis of Lake Urmia's desiccation impact on regional cooling effects over a long temporal scale (1986-2035).
• Quantifies the historical and predicted future changes in Lake Urmia's area and its direct correlation with increasing regional land surface temperatures.
• Utilizes a combination of remote sensing and multiple machine learning algorithms for robust prediction of land use and land surface temperature changes.
• Highlights the significant reduction in the lake's cooling effect, offering critical insights into the environmental consequences of its desiccation.

Funding

Not explicitly mentioned in the provided text.

Citation

@article{Babaei2025Geospatial,
  author = {Babaei, M. and Rajaei, Fatemeh and Siroosi, Hamid and Alamdari, Nasrin},
  title = {Geospatial analysis of Lake Urmia’s drying: predicting land surface temperature changes using remote sensing and machine learning},
  journal = {Advances in Space Research},
  year = {2025},
  doi = {10.1016/j.asr.2025.10.013},
  url = {https://doi.org/10.1016/j.asr.2025.10.013}
}