Spatiotemporal Modeling of Vegetation Response to Climate Change and Drought Using Multiscale Geographically Weighted Regression (MGWR): An Applied Study of Qasrok Subdistrict for the Period 2000–2025

Authors

  • Ali Sulaiman Arzik Al-Karbouli General Directorate of Education in Anbar Anbar / Iraq
  • Haneen Harith Ahmed Jabr Al-Iraqia University / College of Arts / Department of Geography Baghdad / Iraq

DOI:

https://doi.org/10.66026/c2ywh048

Keywords:

Temporal Change, Spatial Variation, Vegetation Cover, Drought, PDSI, MGWR.

Abstract

The two researchers aim to investigate the nature of the spatial and temporal variation in vegetation response to climatic and environmental factors in Qasrok Subdistrict, Shekhan District, Duhok Governorate, during the period 2000–2025. The study identifies the nature of the relationships between the Normalized Difference Vegetation Index (NDVI) and rainfall, temperature, the Palmer Drought Severity Index (PDSI), the Normalized Difference Moisture Index (NDMI), and elevation. It also examines changes in the direction, strength, and spatial scales of these relationships between the two study periods. The analysis is based on remote sensing data, climatic data, and a Digital Elevation Model (DEM). A total of 340 fixed spatial locations were used for the analysis, with Multiscale Geographically Weighted Regression (MGWR) employed as the principal model to identify the spatial variations in the relationships between vegetation cover and the explanatory factors.

The results showed that the spatial mean of the NDVI increased from 0.366 in 2000 to 0.429 in 2025. The mean NDMI increased from 0.093 to 0.121, while the mean temperature increased from 16.45°C to 19.38°C. In contrast, the mean rainfall decreased from 220.23 mm to 213.13 mm. The MGWR model demonstrated improved explanatory performance, with the coefficient of determination (R²) increasing from 0.913 in the OLS model to 0.931 in 2000, and from 0.923 to 0.951 in 2025. The model also revealed clear differences in the spatial scales of the relationships and their changes between the two periods. NDMI was the most stable variable in terms of the direction of its effect and spatial significance, whereas rainfall and elevation exhibited greater changes. The sign of the rainfall coefficient changed at 219 locations, while the sign of the elevation coefficient changed at 312 locations. Temperature maintained a negative relationship, whereas both PDSI and NDMI maintained positive relationships with vegetation cover.

References

Downloads

Published

2026-11-09