Comparing Visible–Near‐Infrared Spectroscopy and a Pedotransfer Function for Predicting the Dry Region of the Soil‐Water Retention Curve

Abstract

© 2019 The Author(s). The soil-water retention curve (SWRC) at the dry end, also known as soil water vapor sorption isotherms, is essential for the modeling of water vapor transport, microbial activity, and biological processes such as plant water uptake in the vadose zone. Measurement of detailed soil water vapor sorption isotherms (WSIs) can be time consuming. Therefore, we propose rapid, inexpensive methodologies (visible-near-infrared spectroscopy [vis-NIRS] and a pedotransfer function [PTF]) to predict the Campbell-Shiozawa (CS) model parameters to obtain the WSIs. Water vapor sorption isotherms were measured on 144 soil samples with a vapor sorption analyzer. The CS semi-logarithmic-linear function anchored at a soil-water matric potential of -106 cm H2O (log|-106| = pF 6) was fitted to the measured data because it accurately characterizes the WSIs. Thereafter, a vis- NIRS calibration model and a PTF, based on clay and organic C contents, were developed and used to predict the two reference CS model parameters (a and W6). Both parameters were predicted with a reasonable degree of accuracy using vis-NIRS and the PTF (for α, RMSE values of 0.0041 and 0.0025, and for W6, RMSE values of 0.0042 and 0.0034 for vis-NIRS and the PTF, respectively). Based on the predicted a and W6 values, the predicted WSIs compared closely with the measured isotherms for individual soil samples from each field. At the field scale, the vis-NIRS model performed marginally better than the PTF. Thus, it is evident that the use of vis-NIRS or PTFs provides a relatively inexpensive approach to predicting soil water sorption isotherms.