Enhanced NDVI prediction accuracy in complex geographic regions by integrating machine learning and climate data—a case study of Southwest basin

The normalized difference vegetation index (NDVI) is a vital metric for assessing vegetation growth, yet accurate prediction remains challenging, particularly in regions with complex geographic and climatic conditions. Machine learning methods offer promise but are often hindered by sensitivity to m...

Full description

Saved in:
Bibliographic Details
Main Authors: Zehui Zhou, Jiaxin Jin, Bin Yong, Weidong Huang, Lei Yu, Peiqi Yang, Dianchen Sun
Format: Article
Language:English
Published: Elsevier 2025-05-01
Series:International Journal of Applied Earth Observations and Geoinformation
Subjects:
Online Access:http://www.sciencedirect.com/science/article/pii/S1569843225001451
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:The normalized difference vegetation index (NDVI) is a vital metric for assessing vegetation growth, yet accurate prediction remains challenging, particularly in regions with complex geographic and climatic conditions. Machine learning methods offer promise but are often hindered by sensitivity to model structure, input parameters, and training samples. To address these limitations, this study developed an NDVI time-series prediction optimization model, LSKRX, which integrates multiple machine learning algorithms with local geographic and climatic data. Using the Southwest Basin of China as a case study, dominant climatic factors were identified through sub-basin analysis, and machine learning models were constructed to link NDVI with these factors. The LSKRX model demonstrated significant improvements in prediction accuracy compared to single-model approaches, with the most notable enhancement in BIAS. Spatially, the model’s predictions aligned closely with observed values, particularly in the middle and lower reaches of the Yarlung Zangbo River. The model performed exceptionally well in winter (CC: 0.964) and summer (CC: 0.918) and achieved optimal accuracy in alpine regions at altitudes of 4000–5000 m (CC: 0.900). By leveraging the strengths of multiple machine learning models, the LSKRX model enhances NDVI prediction reliability under complex mountainous and alpine conditions, providing a robust tool for precise ecological assessment and management.
ISSN:1569-8432