Serkendiz et al. (2025) Machine learning and geographic information systems-based framework for multidimensional analysis of cascading drought impacts using remote sensing and in-situ data

Identification

• Journal: The Science of The Total Environment
• Year: 2025
• Date: 2025-09-16
• Authors: Hıdır Serkendiz, Hasan Tatlı, Emre Özelkan, Mahmut Çetin
• DOI: 10.1016/j.scitotenv.2025.180504

Research Groups

• Canakkale Onsekiz Mart University, Department of Geography, Canakkale, Turkiye
• Department of City and Regional Planning, Faculty of Architecture and Design, and Risk Management of Natural Disasters Program, School of Graduate Studies, Canakkale Onsekiz Mart University, Canakkale, Turkiye
• Department of Agricultural Structures and Irrigation, Faculty of Agriculture, University of Cukurova, Adana, Turkiye

Short Summary

This study proposes a multidimensional framework to assess cascading drought impacts on the agricultural sector, demonstrating its application in the Konya Closed Basin. It reveals severe groundwater depletion coinciding with intensified drought periods and a significant conversion of over 510,000 hectares of irrigated land to non-irrigated areas between 1990 and 2018, highlighting maladaptive agricultural practices.

Objective

• To propose and apply a multidimensional conceptual framework for assessing the cascading impacts of drought on the agricultural sector.
• To characterize drought dynamics and land use transitions in a drought-sensitive agricultural region (Konya Closed Basin) by integrating remote sensing, ground-based observations, and machine learning methods.

Study Configuration

• Spatial Scale: Konya Closed Basin, central Türkiye.
• Temporal Scale: 1990–2018 (for land use change and drought dynamics analysis).

Methodology and Data

• Models used:
  • Statistical trend analyses (Mann-Kendall)
  • Machine learning algorithms for land cover classification: Support Vector Machines (SVM), Artificial Neural Networks (ANN), Random Forests (RF)
• Data sources:
  • Remote sensing indicators: Normalized Difference Vegetation Index (NDVI), Normalized Difference Water Index (NDWI), Land Surface Temperature (LST), land cover data.
  • Ground-based observations: Precipitation, temperature, groundwater levels.
  • Drought-related indicators (used as input variables for ML): NDVI, NDWI, LST, Palmer Drought Severity Index (PDSI).

Main Results

• A multidimensional conceptual framework was developed to assess cascading drought impacts on the agricultural sector.
• Severe groundwater depletion was found to coincide with intensified drought periods in the study area.
• Over 510,000 hectares of irrigated land were converted to non-irrigated areas between 1990 and 2018.
• The expansion of water-intensive crops despite ongoing drought and groundwater decline indicates maladaptation within the agricultural sector.
• The study successfully integrated remote sensing, Geographic Information Systems (GIS), and machine learning methods for multidimensional drought impact analysis.

Contributions

• Development of a novel multidimensional conceptual framework for a comprehensive assessment of cascading drought impacts on the agricultural sector.
• Integration of diverse data sources (remote sensing, in-situ observations) and advanced analytical techniques (GIS, machine learning, statistical trend analysis) to provide a holistic understanding of drought dynamics and their socio-ecological consequences.
• Quantification of significant land use changes (conversion of irrigated to non-irrigated land) and identification of maladaptive agricultural practices in a critical drought-sensitive region.

Funding

• Not specified in the provided paper text.

Citation

@article{Serkendiz2025Machine,
  author = {Serkendiz, Hıdır and Tatlı, Hasan and Özelkan, Emre and Çetin, Mahmut},
  title = {Machine learning and geographic information systems-based framework for multidimensional analysis of cascading drought impacts using remote sensing and in-situ data},
  journal = {The Science of The Total Environment},
  year = {2025},
  doi = {10.1016/j.scitotenv.2025.180504},
  url = {https://doi.org/10.1016/j.scitotenv.2025.180504}
}