Geostatistical interpolation based ternary diagrams for estimating water retention properties in soils in the Center-South regions of Brazil

Abstract The aim of this study was to develop ternary diagrams to estimate water retention in the soil θ-33 kPa and θ-1500 kPa corresponding to the field capacity and wilting point, respectively and available soil water (ASW) using geostatistical approach. The data were extracted from the physical-water base of irrigated areas linked to the Irriga® System belonging to the Rural Engineering Department of the Federal University of Santa Maria. The database provides information regarding the physical characterization of soils in irrigated areas in the states of Rio Grande do Sul, Parana, Goias, Mato Grosso, Sao Paulo, Minas Gerais and Bahia. The data included 11,193 observations covering texture (sand, silt and clay), water retention properties (θ-33 kPa and θ-1500 kPa) and structural properties (bulk density and total porosity). The measured values of the water content in the soil were displayed in the ternary diagram according to the coordinates provided by the particle size distribution determined in the soil samples. Estimates of water content (cm³ cm−³) using the kriging method were performed using only the particle size distribution (sand, silt and clay). To obtain the spatial distribution of soil water retention values in the texture triangle, the use of the ordinary kriging estimator was proposed. The uncertainty analysis resulted in an Root mean square error (RMSE) below 0.048 cm3 cm-3 when comparing the interpolated water content and observed in the values of θ-33 kPa and θ-1500 kPa, with the subdivisions by ranges of bulk density having the lowest uncertainties, highlighting the importance of subdividing the database by classes based on structural properties. Ternary diagrams can act as useful and simplified tools to estimate these properties just by applying the particle size distribution.