Soil properties prediction on the basis reflectance spectroscopy using plsr modeling

This study aimed to apply the partial least square regression algorithm to estimate different soil parameters. Surface soil samples were collected and analyzed for some parameters (i.e. pH, sand, silt, clay, and CEC) using the conventional methods of soil analysis. Hyperspectral signatures of soil samples were collected in the range of Vis-NIR spectra (350-2500nm) using the analytical spectroradiometer device (ASD) in the laboratory conditions. The PLSR model was applied to soil spectra and soil parameters’ data to develop the calibration and validation models. The obtained results showed that sand and CEC soil parameters recorded an excellent predictability with R2 values 0.91 and 0.86, and RPD values 3.00 and 2.41, respectively. The rest soil parameters such as pH, silt and clay were having moderate predictability whereas R2 values were 0.68, 0.50, and 0.62, and RPD values were 1.84, 1.41 and 1.70, respectively. The diffuse reflectance spectroscopy integrated with multivariate regression models such as PLSR could successfully estimate soil parameters with good prediction. This technique was found to be promising for soil parameters’ prediction. It saves time, effort, chemicals, and many soil parameters that can be estimated simultaneously.