|Torbert, Henry - Allen|
Submitted to: Transactions of the ASABE
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: 12/1/2004
Publication Date: 10/1/2005
Citation: Akbar, M.A., Kenimer, A.L., Searcy, S.W., Torbert III, H.A. 2005. Soil water estimation using electromagnetic induction. Transactions of the ASAE. 48(1):129-135. Interpretive Summary: Recently, methods have been developed to remotely measure soil electrical conductively without disturbing the soil using instruments that measure electromagnetic induction. This study examined whether these methods to measure soil electrical conductivity could be used to estimate soil water content in real-time. Two different published salinity models, which relate soil electrical conductively to soil water content, were examined for application to real-time soil water estimation. Field data were collected at two sites representing a range of soil types in Central Texas. Both models show promise for use for real-time, non-invasive soil water content estimation using apparent electrical conductivity, but additional testing is needed.
Technical Abstract: Two published salinity models (designated the Rhoades and Mualem - Friedman models) were examined for application to real-time soil water estimation using apparent soil electrical conductivity. Field data were collected at two sites representing a range of soil types in Central Texas: high shrinking-swelling Vertisols in Temple (the Heiden Clay site) and clay loam soils at the Texas A&M University Research Farm near College Station, TX (the Westwood Scl site). The Rhoades-Corwin model developed for the Heiden Clay site yielded an R2 of 0.72 following calibration, predicted soil water within ± 0.02 g g -1 during validation, and was deemed generally applicable for real-time soil water estimation. The Rhoades-Corwin model developed for the Westwood Scl site gave an R2 of 0.65 following calibration but could not be validated at the site and therefore was not considered applicable for real-time soil water estimation. A modified version of the Rhoades-Corwin model yielded a calibrated R2 of 0.91 at the Westwood Scl site with validation predictions within ± 0.02 g g-1. The Mualem-Friedman model predicted soil water within ± 0.05 g g-1 at the Heiden Clay site and was considered appropriate for real-time soil water estimation. At the Westwood Scl site, the Mualem-Friedman model could not be evaluated since saturation data were not available. Both models show promise for use for real-time, non-invasive soil water content estimation using apparent electrical conductivity, but additional testing is needed.