Scope to map available water content using proximal sensed electromagnetic induction and gamma-ray spectrometry data

Abstract

Dryland cotton farmers need to carefully manage the available water content (AWC) in the root-zone. To reduce cost of mapping, a digital soil mapping (DSM) approach might be useful. The purpose of this study was seeing if we could create DSM of the AWC at the field scale. In the first instance, digital data in the form of proximal sensed electromagnetic (i.e EM) and gamma-ray (γ-ray) spectrometry data were collected along with elevation. Using a grid-based sampling design, 52 soil sample locations were selected and sampled; topsoil (0–0.3 m), subsurface (0.3–0.6 m) and subsoil (0.6–0.9 m). In the laboratory, the field capacity (FC) and permanent wilting point (PWP) were determined using a pressure plate apparatus. The digital and soil (i.e. FC and PWP) data were then calibrated with multiple linear regression (MLR) and stepwise-MLR models compared for each depth (e.g. adjusted coefficient of determination [adj-R2]). The stepwise-MLR were shown to be superior. Subsequently, the DSM of FC and PWP were examined in terms of prediction agreement (Lin’s concordance). Leave-one-out cross validation results showed substantial 1:1 agreement between measured and predicted FC (0.80) and PWP (0.79) at all depths. The final DSMs of FC and PWP, and by difference AWC, were indicative of the two main geomorphological/geological units, with large AWC (> 0.3 m3 m−3) associated with the clay rich Vertosols and small AWC (< 0.15 m3 m−3) the loamy soil of the Pilliga Sandstone. Owing to the short scale variation and the fact that the local alluvium and Pilliga Sandstone were close together and the fact that the digital data varied in different ways, larger confidence interval (CI) values were evident where these areas were juxtaposed. Nevertheless, the final DSM allows the farmer to better manage AWC.