Publication Date: 2018.
Soil available water capacity (AWC) is an important factor affecting soil productivity in semiarid and subhumid environments and is mainly determined by the soil textural composition. As the soils of these environments usually present fairly uniform textures across depth, we hypothesized that it would be possible to accurately estimate the whole-profile AWC using surface information. Our objective was to test this hypothesis in the Argentine Semiarid Pampas. Information was collected from 152 sites where AWC was measured in 20 cm layers up to a depth of 140 cm or up to the upper limit of the petrocalcic horizon, when present. In each case, whole profile AWC was estimated using a one-step and a two-step approach, comparing multiple regression and artificial neural networks as modeling techniques. Both modeling methods were effective (R2 > 0.76, P < 0.05), however the former was chosen as no special software is required to run it, thus favoring simplicity. Models showed a strong interaction between surface AWC and soil depth and a simple nomogram was developed to estimate whole-profile AWC. Sampling and laboratory efforts should be significantly reduced using the model proposed in this paper.
Author affiliation: de Paepe, Josefina. Universidad de Buenos Aires. Facultad de Agronomía; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina
Author affiliation: Angel Bono, Alfredo. Instituto Nacional de Tecnología Agropecuaria. Centro Regional La Pampa-San Luis. Estación Experimental Agropecuaria Anguil; Argentina
Author affiliation: Alvarez, Roberto. Universidad de Buenos Aires. Facultad de Agronomía; Argentina. Consejo Nacional de Investigaciones Científicas y Técnicas; Argentina
Repository: CONICET Digital (CONICET). Consejo Nacional de Investigaciones Científicas y Técnicas