Evaluating soil parameters with visible spectroscopy

Using spectroscopic techniques, correlation analysis and regression were conducted between soil parameters and visible reflectance spectra of raw soil samples. The samples were collected in a cornfield with lower moisture content and a forage field with higher moisture content. Five soil parameters, soil moisture, SOM, NO₃-N, EC and pH, were analyzed by conventional method in a laboratory. Visible reflectance spectra of the samples were measured by a spectrophotometer. The spectral variable used was the first derivative of visible reflectance. Results showed that all five soil parameters could be evaluated by visible reflectance. In the cornfield with lower moisture (<30% db), a single linear regression model was available for moisture estimation, and a multiple exponential model was recommended to evaluate NO₃-N. However, it failed to establish the evaluating model of SOM because of its small variation, and it was difficult to obtain the estimation model of EC and pH because of lower moisture content. In the forage field with higher moisture (>70% db), a multinomial model with single variable was better for moisture estimation, and a single linear model and two multiple linear models were recommended to evaluate SOM, EC and pH respectively. For NO₃-N, a multiple exponential model was also recommended. It would be necessary to study the correlation between visible reflectance and soil parameters of raw soil samples with middle moisture (30%～70% db).