[Black soil organic matter predicting model based on field hyperspectral reflectance].

Area of Science:

• Agricultural remote sensing
• Soil science
• Quantitative spectroscopy

Background:

• Soil organic matter (OM) is crucial for soil health and agricultural productivity.
• Accurate and rapid measurement of OM is needed for effective soil management.
• Remote sensing offers potential for large-scale soil monitoring.

Purpose of the Study:

• To develop rapid soil organic matter (OM) measurement methods using hyperspectral reflectance.
• To enhance the application of remote sensing in agriculture and terrestrial ecosystem studies.
• To build a predictive model for black soil OM content.

Main Methods:

• Collected field hyperspectral reflectance data in visible/near-infrared bands.
• Analyzed spectral characteristics and influencing factors, focusing on OM.
• Derived spectral indices and built a predictive model using differential coefficient of logarithmic reflectance reciprocal (DCLRR).
• Validated model predictability and stability using RMSE and R2.

Main Results:

• Significant spectral differences in black soil reflectance were observed below 1250 nm, especially below 1000 nm.
• Soil OM content significantly influences reflectance curve shape and spectral features.
• The DCLRR at 1260 nm showed the strongest correlation with OM content.
• The developed OM predictive model achieved R2 of 0.71 and RMSE of 0.42.

Conclusions:

• Hyperspectral remote sensing, particularly using DCLRR at 1260 nm, is effective for rapid OM assessment in black soil.
• The developed model demonstrates good stability and predictability for practical application.
• This method supports improved agricultural production and management through quantitative soil analysis.