A spectral soil quality index (SSQI) for characterizing soil function in areas of changed land use

• We developed two soil quality indices using laboratory and spectral measures.
• We examined changes in land-use using soil quality indices.
• Reflectance spectroscopy is a reliable diagnostic tool for assessing soil quality.
• The Spectral Soil Quality Index can predict land-use degradation.
• Spectral Soil Quality Index and the Soil Quality Index are complementary methods.

During the last several decades, a large proportion of the planet's terrestrial surface has transformed from natural ecosystems to human-dominated systems. These land-use dynamics affect ecosystems' soil quality. The current study was conducted at the fringe of the northern Negev Desert, Israel, and strived to assess and compare the soil quality in three different land-use types (afforestation, traditional grazing, and agro-pastoral) that were changed from managed to unmanaged or vice versa (e.g., shrubland was transformed to a planted forest; pastoral grazing to natural shrubland with no grazing; and agro-pastoral to abandoned agricultural). The overall aim of this research is twofold: (1) to evaluate by reflectance spectroscopy the changes in 14 soil physical, biological, and chemical properties and their derived soil quality index (SQI) in the changed land uses; and (2) to develop a spectral soil quality index (SSQI) toward applying the technique of reflectance spectroscopy as a diagnostic tool of soil quality. To achieve these objectives, several mathematical/statistical procedures, consisting of a series of operations, were implemented, including a principal component analysis (PCA), a partial least squares-regression (PLS-R), and a partial least squares-discriminate analysis (PLS-DA). The PLS-R's most suitable models successfully predicted soil properties (R2 > 0.80; ratio of performance to deviation (RPD) > 2.0), including sand–silt–clay content, NH4, NO3−, and pH. Moderately well-predicted soil properties (0.50 < R2 < 0.80; RPD > 2) were residual water, soil organic matter, electric conductivity, and potassium. Poor validation (R2 < 0.50; RPD < 2) results were obtained for potential active carbon, phosphorus, and hydraulic conductivity. In addition, the PLS-R model predicted the SQI in the changed land uses. The correlations between the predicted spectral values of the calculated SQI ranged 0.65 < R2 < 0.81 with RPD > 2. The PLS-DA model was used to develop the SSQI model. The correlations between the SSQI and the SQI ranged 0.66 < R2 < 0.74 in the different land uses. This study underscores the potential application of reflectance spectroscopy as a reliable diagnostic screening tool for assessing soil quality. The classification of soils into spectral definitions provides a basis for a spatially explicit and quantitative approach for developing the SSQI. The SSQI can be used to assess hot spots of change in areas of land-use changes and to identify soil degradation.