Combined Proximal and Remote Sensing Methods for Rapid Soil Salinity Quantification.

Worldwide, salt affected soils are common in arid and semiarid climates. Since the 1950s, electrical conductivity (EC) measurement has been widely used as a quantitative measure of soil salinity. However, newer technologies [e.g., remote sensing (RS), visible near infrared diffuse reflectance spectroscopy (VisNIR DRS), portable x-ray fluorescence (PXRF) spectrometry] have the potential to provide salinity data on soils more quickly and inexpensively than traditional techniques. This study evaluated two saline playas in West Texas, USA featuring 165 soil samples collected at 0-5 cm. Sampling was conducted such that upland soils from playa annuli and playa floors were collected, representing a wide variation in soil salinity. Samples were subjected to PXRF and VisNIR DRS scanning under laboratory conditions, then compared to Landsat spectral data and traditional laboratory analyses of salinity (e.g., 1:5 v/v suspensions). Data showed a broad range of EC (0.028 to 43.41 dS m-1). PXRF analyses showed that both Cl and S were significantly and positively correlated with log10 transformed EC. VisNIR partial least squares prediction models produced strong residual prediction deviations (RPDs) of 2.49-2.91. Validation statistics of Savitzky–Golay support vector regression outperformed all other VisNIR models tested with an RPD of 3.1. Taken individually, each technique produced acceptable predictive results. However, synthesis of the three techniques resulted in superior predictability with an RPD of 3.35. Given that PXRF, VisNIR DRS, and RS data are quick, inexpensive, and easy to collect, their combination in predictive models offers strong potential for providing soil salinity measurements comparable to laboratory derived data. As such, this approach has the potential for widespread application in salt affected desert and coastal areas, and areas of anthropogenic salinity impact.