Bindajam et al. (2025) Advanced predictive modelling of urban expansion and land surface temperature dynamics using multi-scale machine learning approaches

Identification

• Journal: Advances in Space Research
• Year: 2025
• Date: 2025-10-24
• Authors: Ahmed Ali Bindajam, Javed Mallick, Hoàng Thị Hằng, Chander Kumar Singh
• DOI: 10.1016/j.asr.2025.10.080

Research Groups

• Department of Architecture, College of Architecture and Planning, King Khalid University, Abha, Kingdom of Saudi Arabia
• Department of Civil Engineering, College of Engineering, King Khalid University, Abha, Kingdom of Saudi Arabia
• Department of Energy and Environment, Analytical and Geochemistry Laboratory, TERI School of Advanced Studies, New Delhi, India
• Department of Environmental Science, School of Vocational Studies and Applied Sciences, Gautam Buddha University, UP-201312, India

Short Summary

This study systematically quantifies urban expansion, assesses its morphological and thermal impacts, and forecasts future land surface temperature (LST) dynamics in Lucknow, India, from 1991 to 2021, revealing a significant LST increase linked to urban morphology.

Objective

• To systematically quantify urban expansion, assess its morphological and thermal impacts, and forecast future Land Surface Temperature (LST) dynamics in Lucknow, India, from 1991 to 2021.

Study Configuration

• Spatial Scale: Urban area of Lucknow, India.
• Temporal Scale: Analysis period from 1991 to 2021; forecast for 2031.

Methodology and Data

• Models used: Random Forest (RF) classifiers, Mono-Window Algorithm (MWA) for LST retrieval, Morphological Spatial Pattern Analysis (MSPA), Polynomial regression models (degrees 2–4) for LST forecasting.
• Data sources: Multi-decadal Landsat satellite data, including six spectral bands (Blue, Green, Red, NIR, SWIR1, SWIR2), three indices (NDVI, NDBI, NDWI), and eight GLCM texture metrics extracted from NIR and SWIR1 bands.

Main Results

• Land Use and Land Cover (LULC) models achieved high accuracy with test accuracies ranging from 86.12 % to 88.13 % and kappa values from 0.8562 to 0.8741.
• The mean Land Surface Temperature (LST) increased by approximately 2.5 °C per decade.
• By 2021, 55 % of the study region exceeded an LST of 30 °C.
• Morphological Spatial Pattern Analysis (MSPA) indicated growing fragmentation in agricultural and open land (e.g., loss of core and bridge elements) and densification in built-up zones.
• Forecasted LST for 2031 predicted the highest values in built-up cores (34.30 °C) and bridges (32.51 °C), and the lowest in core agriculture (25.77 °C).
• The study confirmed that landscape form and composition strongly regulate urban thermal patterns.

Contributions

• Presents an integrated modeling pipeline that links machine learning classification, thermal retrieval, and morphological analysis to anticipate urban heat trajectories.
• Provides critical insights for climate-adaptive land use planning and resilient urban design.
• Systematically quantifies urban expansion and its morphological and thermal impacts using multi-scale machine learning approaches.

Funding

• Not specified in the provided text.

Citation

@article{Bindajam2025Advanced,
  author = {Bindajam, Ahmed Ali and Mallick, Javed and Hằng, Hoàng Thị and Singh, Chander Kumar},
  title = {Advanced predictive modelling of urban expansion and land surface temperature dynamics using multi-scale machine learning approaches},
  journal = {Advances in Space Research},
  year = {2025},
  doi = {10.1016/j.asr.2025.10.080},
  url = {https://doi.org/10.1016/j.asr.2025.10.080}
}