[Contrastive analysis on soil alkalinization predicting models based on measured reflectance and TM image...

Area of Science:

• Soil Science
• Remote Sensing
• Agricultural Science

Background:

• Soil alkalinization is a significant environmental issue in arid and semi-arid regions.
• Alkaline soils can negatively impact crop yields and soil health.
• Monitoring and managing soil alkalinization is crucial for sustainable agriculture.

Purpose of the Study:

• To evaluate the level of soil alkalinization using spectral reflectance data.
• To compare the accuracy of soil pH prediction models based on measured and satellite-derived reflectance.
• To assess the potential of remote sensing for detecting changes in alkalinized soils.

Main Methods:

• Collected soil pH and Vis-NIR (Visible and Near-Infrared) spectral reflectance data in Qitai oasis, Xinjiang.
• Developed multivariate linear regression models relating soil pH to measured and TM (Thematic Mapper) image reflectance.
• Analyzed the influence of vegetation cover using NDVI (Normalized Difference Vegetation Index) on TM reflectance models.

Main Results:

• A significant positive correlation was observed between soil pH and spectral reflectance.
• Soil alkalinization, characterized by hardening, exhibited good spectral response.
• Measured reflectance models showed higher accuracy in predicting soil pH compared to direct TM reflectance models.
• Removing vegetation effects with NDVI improved the accuracy of TM reflectance models.

Conclusions:

• Spectral reflectance, particularly Vis-NIR, is a viable indicator for assessing soil alkalinization.
• Remote sensing data, when properly processed (e.g., NDVI correction), can effectively monitor soil alkalinization.
• The hardening state of alkalinized soil influences model accuracy, highlighting the need for further refinement.