Different Particle Sizes of Soil Organic Matter on The Spectral Predictionr

    As an important component of soil, soil organic matter (SOM) plays a decisive role in soil structure formation and soil quality improvement. Understanding SOM dynamic changes is the basic condition for monitoring soil fertility, realizing precision agriculture and ensuring sustainable agricultural development. Traditional field sampling and laboratory chemical analysis methods are time-consuming and inefficient, which are difficult to meet the real-time data requirements of current agricultural development. The spectroscopic method is low-cost, fast and pollution-free, and can be easily implemented in laboratory and field conditions. However, the spectral reflectance of soil surface with different particle size has great spatial heterogeneity. Moreover, the spectral reflectance of soil measured by the spectrometer is a comprehensive result of the soil surface in a specific area. Therefore, the uncertainty of the composition of different particles on the soil surface makes the spectral determination fluctuate, thus affecting the spectrum of soil samples.
    In order to determine the best standard for sample preparation, the authors conducted field investigations with field spectrometer to clarify the surface composition characteristics of uniform and mixed samples of different sample sizes and their influence on SOM prediction, and compared and analyzed the prediction accuracy of different particle size estimation models.
    From the average spectral reflectance of soil samples with different particle sizes, it can be seen that the spectral reflectance increases gradually with the decrease of particle size. When the mixed sample and uniformsample have the same total reflectance, the composition characteristics of soil surface particles are different. The smaller the particle size is, the smaller the heterogeneity of soil surface is. Increasing the interval of soil surface spectral reflectance is helpful to distinguish the subtle differences between soil samples with different SOM contents. The decrease of soil particle size increased the correlation between the whole band and SOM. Mixed samples can improve the surface reflectance of soil samples, and increase the correlation with SOM due to the complementation of large and small particles.