Characterization of Soil Reaction (pH) Via Portable x-Ray Fluorescence Spectrometry.

Soil reaction (pH) has profound influence on plant nutrient availability, plant vigor, and is a key indicator of many soil chemical properties. Soil pH determinations have been made traditionally using colorimetric or electrometric methods. While accurate, these methods are laborious, time consuming, require sample modification and often times are destructive in nature. This research focused on the use of portable x-ray fluorescence (PXRF) spectrometry for soil pH determinations which could potentially overcome the problem of making soil pastes from lithified horizons (e.g., Bsm, Bkkm), bedrock (R), or permafrost (Cf). The advantage of PXRF lies in the rapid, easy and in-situ determination of elemental data. We originally hypothesized that the soil exchange complex of arid and humid lands would be dominated by Ca & Mg, and Al & Fe, respectively. Elemental data from the PXRF was then used as a proxy for determining soil pH. Two datasets representing a wide variety of soil pH were evaluated both by PXRF and standard laboratory methods. Datasets were divided into modeling and validation datasets and simple and multiple linear regression were used to develop predictive models. In addition to PXRF elemental data, auxiliary input data such as sand, clay, and organic matter content were also used to enhance the relationship. The best predictive model was provided with multiple linear regression with auxiliary input data where R² and RMSE were found to be 0.825 and 0.541, respectively. Pure PXRF elemental data provided a quality prediction generating an R² and RMSE of 0.772 and 0.685, respectively with multiple linear regression. Simple linear regression seemed to be ineffective in producing significant model predictions. Correlation analysis was used to validate the developed models. Summarily, PXRF can be used as a proxy for soil pH prediction with reasonable accuracy and without the requirement of sample destruction.