Mapping the Variability of Soil Texture-Based on Vis-NIR Proximal Sensing

Abstract

Soil texture is one of the soil properties influencing most physical, chemical, and biological soil processes. Information on soil texture is important to support the agronomic decisions for farm management. The problem is how to provide reliable, fast and inexpensive information of soil texture in numerous soil samples and repeated measurement. The objective of this research was to generate the soil texture map based on laboratory Vis-NIR (Visible - Near Infra-Red) spectroscopy and inverse distance weighted (IDW) interpolation method. An ASD Fieldspec 3 with a spectral range from 350 nm to 2500 nm was used to measure the soil reflectance. Pipette method was used to measure the silt, clay and sand fractions. The partial least square regression (PLSR) was performed to establish the prediction model of soil texture. The predicted values were mapped and showing the information of spatial and temporal variability of soil texture.
 Keywords: Vis-NIR, spectroscopy, soil texture, PLSR, IDW

The manifestation of VIS-NIRS spectroscopy data to predict and map soil texture in the Triffa plain (Morocco)

“…The main objective of this study, compared with other studies carried out by Curcio et al (2013) and Virgawati et al (2018), in order to predict the soil texture, is to assess the feasibility of the data of the VIS-NIRS spectroscopy to predict and map the soil texture (clay, silt, and sand) in the semi-arid region of Triffa plain, using the PLSR method, and to show the influence of the texture of the soil samples used to make the prediction on the coefficient of determination.…”

“…The advantage of this method lies in the sample preparation, which requires only a few minutes for each sample compared to the traditional laboratory method, and it is a useful tool for large-scale digital soil mapping, in which the number of soil samples is enormous, while the cost of soil analysis is expensive (Lazaar et al, 2019). Currently, the VIS-NIRS spectroscopy represents an alternative method for predicting a large range of soil physical properties such as the texture, structure, and bulk density (Emerson, 1995;Gomez et al, 2008;Leone et al, 2012;Curcio et al, 2013;Virgawati et al, 2018). Concerning the use of VIS-NIRS spectroscopy to predict soil texture, Rossel & Webster (2012) and Lacerda et al (2016) found that the prediction is robust for the sand and clay fractions but is not reliable for the silt fraction.…”