Alternative Comprehensive Soil Quality Assessment Using a Combined Method of VNIR Spectroscopy and Traditional Soil Analytical Methods.

Visible and near-infrared reflectance spectroscopy (VNIRS) is a rapid and non-destructive method to predict various soil properties. Past study has shown that the combination of VNIRS, measured penetration resistance, and measured soil chemical values (pH and EC) can estimate soil quality (SQ) indicators (R² > 0.80; ratio of performance to deviation (RPD) > 2.00) of soil organic matter, active carbon, Ca, Mg, cation exchange capacity, clay, water content at permanent wilting point, and water content at field capacity with good accuracy in a small region in Kenya. However, its application in evaluating multidimensional SQ assessment for a large region (i.e. Northeastern USA) has not been well studied. In this study, VNIRS (350-to-2500 nm; 1 nm resolution) was employed to analyze 1200 air-dried soil samples from a wide range of farms across the Northeastern USA for the analysis of sixteen Cornell SQ indicators. This study aimed to build rigorous partial least squares regression (PLSR) models for the region using various spectral preprocessing methods (first derivative reflectance, standard normal variate transform, and standard normal variate transform with detrending), combinations with measured values (penetration resistance, pH and EC), and also soil texture, soil series or soil order specific models. This study proved that VNIRS can be incorporated to a standard soil-testing laboratory in the Northeastern USA to reduce the cost and time for a suite of SQ assessment. Moreover, the obtained SQ indicator values can be used as site-specific information to simulate various biophysical processes such as on-farm carbon and nitrogen cycling.