Climate-Driven Vegetation Vulnerability in Pakistan, 2001-2023

A 23-year record from 2001 to 2023 assesses vegetation vulnerability across Pakistan's four major provinces using satellite-derived indices. The study combines kernel NDVI and Vegetation Health Index with climate data and interpretable machine learning to quantify climatic drivers. Author Kaleem Mehmood published the findings under a CC-BY-4.0 license in 2026.

Use Cases

• Assessing long-term vegetation health trends based on kNDVI and VHI indices
• Quantifying the relative influence of precipitation and temperature on vegetation using SHAP analysis from Random Forest models
• Identifying regions of persistent vegetation decline versus short-term recovery based on the dual-index framework
• Informing drought-resilient afforestation and rangeland rehabilitation strategies based on province-scale vulnerability patterns

Strengths

• Covers a 23-year time series from 2001 to 2023.
• Analyzes data across four defined provincial regions of Pakistan.
• Employs a dual-index methodology combining kNDVI and VHI for assessment.
• Uses interpretable machine learning (Random Forest with SHAP) to quantify climatic drivers.

Limitations

• The dataset is very small at 33.8 KB, suggesting limited raw data or a summary document.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count is unknown, which may limit suitability assessment for large-scale analysis.

Provenance

Source

figshare, author Kaleem Mehmood.

Collection Method

Analysis of Landsat-based kernel NDVI and Vegetation Health Index, combined with climate data and machine learning.

Time Range

2001 to 2023

Freshness

Last updated 2026-05-26 06:02:04.

Geography

Pakistan, specifically provinces Balochistan, Khyber Pakhtunkhwa–Gilgit–Baltistan–AJK, Punjab, and Sindh.