Khan et al. (2026) Mapping agricultural drought hotspots in Pakistan: a remote sensing-based climate–vegetation nexus

Identification

• Journal: Theoretical and Applied Climatology
• Year: 2026
• Date: 2026-03-01
• Authors: Wisal Khan, Mohamad Hidayat Bin Jamal, Mohammed Magdy Hamed, Mohd. Khairul Idlan Muhammad, Shamsuddin Shahid
• DOI: 10.1007/s00704-026-06064-7

Research Groups

• Department of Water and Environmental Engineering, Faculty of Civil Engineering, Universiti Teknologi Malaysia (UTM), Johor Bahru, Malaysia
• Centre for River and Coastal Engineering (CRCE), Universiti Teknologi Malaysia (UTM), Johor Bahru, Malaysia
• Government of Khyber Pakhtunkhwa, Public Health Engineering Department, Peshawar, Pakistan
• Construction and Building Engineering Department, College of Engineering and Technology, Arab Academy for Science, Technology and Maritime Transport (AASTMT), Giza, Egypt
• Regional Climate Change Center, National Center for Meteorology, Jeddah, Saudi Arabia
• Environmental and Atmospheric Sciences Research Group, Scientific Research Center, Al-Ayen University, Thi-Qar, Nasiriyah, Iraq

Short Summary

This study analyzes agricultural drought dynamics in Pakistan from 2001 to 2023 using multisensor remote-sensing indices, revealing spatially heterogeneous and seasonally structured drought occurrences with northern regions being resilient and southern/western regions highly vulnerable, necessitating region- and season-specific adaptation strategies.

Objective

• To analyze agricultural drought dynamics in Pakistan using multisensor remote-sensing indices (TCI, VCI, VHI, NDVI, LST, and rainfall) from 2001 to 2023.
• To map spatiotemporal drought hotspots and severity across all provinces of Pakistan.
• To investigate the relationships and lagged correlations among land surface temperature (LST), rainfall, and vegetation health (NDVI).
• To assess annual and seasonal trends in drought indices (TCI, VCI, VHI), LST, rainfall, and NDVI.

Study Configuration

• Spatial Scale: National scale (Pakistan), covering all provinces, with data resampled to a uniform 10 km spatial resolution.
• Temporal Scale: 2001–2023 (23 years), with data aggregated to 16-day intervals for analysis.

Methodology and Data

• Models used:
  • Vegetation Condition Index (VCI)
  • Temperature Condition Index (TCI)
  • Vegetation Health Index (VHI)
  • Spearman's Rank Correlation Coefficient (for non-linear, monotonic relationships and lagged correlations)
  • Hamed–Rao bias-corrected pre-whitening Mann–Kendall (BCPW-MK) test (for trend analysis)
  • Sen’s slope estimator (for trend magnitude)
  • VHI-based five-class drought categorization (No Drought, Mild, Moderate, Severe, Extreme)
• Data sources:
  • MODIS Normalized Difference Vegetation Index (NDVI) (MOD13A2) at 1 km spatial resolution, 16-day temporal resolution (2001–2023).
  • MODIS Land Surface Temperature (LST) (MOD11A1) at 1 km spatial resolution, daily temporal resolution (2001–2023).
  • Global Satellite Mapping of Precipitation (GSMaP) Gauge-calibrated Near-Real-Time product (GSMaPGaugeNRT, Version 6) at 10 km spatial resolution, daily temporal resolution (2001–2023).

Main Results

• Drought occurrence is spatially heterogeneous and exhibits a seasonal structure across Pakistan.
• Northern high-altitude provinces (Azad Kashmir and Gilgit–Baltistan) are relatively resilient, with over 80% of their area free from drought and less than 5% exposed to extreme drought.
• Southern provinces (Sindh, Balochistan, and Punjab) exhibit extensive vulnerability, with up to 99% of Punjab and approximately 92–94% of Sindh and Balochistan experiencing mild to severe droughts, while extreme drought affected 78–92% of their respective areas at least once.
• Climatic controls diverge by region: in the highlands, vegetation is temperature-limited with strong positive NDVI–LST correlations (annual medians ~ 0.7–0.8) persisting across multiple 16-day lags; in arid lowlands, vegetation is water-limited, exhibiting negative NDVI–LST correlations (~ –0.4 to –0.6) but positive rainfall–NDVI relationships (~ 0.4–0.7) during the monsoon.
• Lag analysis indicates rapid greening within 16 days in humid regions, while heat-stressed plains recover only after 64–96 days.
• Trend analysis reveals positive decadal gains in vegetation indices, with VCI and VHI increasing by 0.05–0.10 per decade and NDVI by ~ 0.05 per decade in Punjab.
• LST trends reveal spring cooling (~ –1.5 to –2.0 °C per decade) but localized summer warming in Balochistan (+ 1.0 °C per decade).
• Rainfall trends indicate intensified monsoon precipitation, particularly in Punjab and Azad Kashmir (by 70–150 mm per decade).
• Overall, mild drought is ubiquitous, moderate drought is common, and severe-to-extreme drought is regionally concentrated in the south and west.

Contributions

• Provides the first integrated, multi-sensor assessment of agricultural drought dynamics and their climatic drivers across Pakistan for the period 2001–2023.
• Introduces a comprehensive, nationwide framework uniquely combining long-term, multi-source satellite and climate datasets with explicit lag-correlation analysis of climate variables and vegetation indices.
• Delivers the first high-resolution spatiotemporal hotspot mapping of drought severity across all provinces, rigorously capturing both immediate and delayed vegetation responses to drought drivers at seasonal and interannual scales.
• Quantifies spatial hotspots and elucidates the timing and intensity of drought impacts across Pakistan, providing actionable insights for climate-resilient agricultural management and policy targeting.

Funding

• The Ministry of Higher Education Malaysia
• Universiti Teknologi Malaysia

Citation

@article{Khan2026Mapping,
  author = {Khan, Wisal and Jamal, Mohamad Hidayat Bin and Hamed, Mohammed Magdy and Muhammad, Mohd. Khairul Idlan and Shahid, Shamsuddin},
  title = {Mapping agricultural drought hotspots in Pakistan: a remote sensing-based climate–vegetation nexus},
  journal = {Theoretical and Applied Climatology},
  year = {2026},
  doi = {10.1007/s00704-026-06064-7},
  url = {https://doi.org/10.1007/s00704-026-06064-7}
}