Agricultural soil fertility properties in the near infrared spectrum

Abstract

Soil spectrum in the near infrared (NIR) wavelength region can be used to reveal fertility properties which is related to plant cultivations. The main purpose of this presented paper is to study the soil spectrum in the NIR region and its related to the fertility properties in form of heavy metals like Fe and Cu. Soil samples were obtain from several land-use including agriculture, mining and ground field. Near infrared spectrum of soil samples were acquired in wavelength range from 1000 to 2500 nm. Prediction models used to determine Fe and Cu were built by means of partial least squares regression (PLSR) followed by leverage cross validation. Prediction performance was evaluated using coefficient of determination (r2) and ratio of prediction to deviation (RPD). The results showed that both Fe and Cu can be revealed simultaneously using the NIR spectrum with maximum r2 and RPD indexes were 0.93 and 3.86 for Fe and 0.71 and 1.88 for Cu prediction respectively. Based on the achieved results, it may conclude that soil fertility properties can be revealed simultaneously and rapidly using mear infrared spectral data.