Spatiotemporal pattern and prediction model of normalized difference vegetation index in the Yellow River Source Zone

Area of Science:

• Ecology
• Environmental Science
• Remote Sensing

Background:

• The Yellow River Source Zone is a vital ecological barrier for the Yellow River Basin.
• It plays a crucial role in water conservation, climate regulation, and biodiversity protection.

Purpose of the Study:

• To analyze the spatiotemporal variations and driving mechanisms of the normalized difference vegetation index (NDVI) in the Yellow River Source Zone from 2000-2020.
• To construct a prediction model for NDVI changes.

Main Methods:

• Integrated MODIS remote sensing images, meteorological, hydrological, and terrain data.
• Employed Sen slope method, partial correlation analysis, variance inflation factor analysis, interaction detection, random forests, and geographically weighted random forests models.

Main Results:

• Annual NDVI increased significantly by 18.9% (0.0028·a^-1) from 0.3301 to 0.3924 between 2000-2020.
• NDVI showed an increasing trend from northwest to southeast.
• Wind speed, precipitation, and minimum temperature were identified as primary drivers of NDVI changes, with complex interactions.

Conclusions:

• The geographically weighted random forest model significantly outperformed the random forest model in predicting NDVI, achieving R²=0.976.
• This highlights the model's effectiveness in capturing spatial heterogeneity and local driving mechanisms.
• The study provides valuable insights into vegetation dynamics and their drivers in the Yellow River Source Zone.