Deriving and validating pedotransfer functions for some calcareous soils

Summary

The unsaturated soil hydraulic properties are needed for many different applications in soil hydrology. Pedotransfer functions (PTFs) have proven to be useful to indirectly estimate these parameters from more easily obtainable soil data. Until now no studies have been conducted to derive or verify PTFs for calcareous soils, which hydraulically may not behave the same as non-calcareous soils. The objectives of this study were to assess the influence of soil CaCO₃ on the soil water retention characteristics of some calcareous soils, and to derive PTFs for these soils. Two data sets were used to derive and evaluate the established PTFs. Data set 1 containing 220 samples was employed as a calibration set for multiple linear regression. An independent data set containing 55 soil samples from a different location served to verify the derived PTFs. No significant difference in accuracy was found between the PTFs with and without CaCO₃ in terms of estimating specific soil water retention values or the van Genuchten soil hydraulic parameters. Compared with the Rosetta PTFs of Schaap et al. (2001), the derived point and parametric PTFs provided better accuracy with average RMSE values of 0.028 and 0.107 cm³ cm⁻³, respectively.

Introduction

Estimates of the unsaturated soil hydraulic properties are needed for many agronomic, environmental and engineering applications involving water flow and contaminant transport processes in the vadose zone. A broad array of laboratory and field methods are currently available for direct measurement of the hydraulic properties (e.g., Dirksen, 2000, Dane and Topp, 2002). Unfortunately, most or all methods are too time consuming and costly for most applications (Wösten et al., 2001). Furthermore, due to high temporal and spatial variability in the hydraulic characteristics, a large number of samples is generally required to accurately characterize field conditions.

An alternative to direct measurement is the use of indirect methods in which the hydraulic properties are estimated from more easily measured or more readily available soils data (Bouma, 1989, Wösten et al., 2001). Methods that may be used for this purpose include statistical pore-size distribution models, inverse models and especially pedotransfer functions (PTFs). PTFs (Bouma and van Lanen, 1987, Bouma, 1989) are based on the premise that the hydraulic properties can be related to basic soil properties such as the particle size distribution, bulk density and/or organic matter content. From a prediction point of view, PTFs can be distinguished into point and parametric PTFs. Point PTFs predict the water content at pre-defined soil water pressure heads, while parametric PTFs assume that the hydraulic properties can be described adequately with a hydraulic model containing a limited number of parameters.

Iran, as many other countries, does not have sufficient soil hydraulic data for deriving local PTFs for relevant agricultural or environmental modeling applications. However, several large databases such as UNSODA (Leij et al., 1996), HYPRES (Lilly, 1997, Wösten et al., 1999), WISE (Batjes, 1996) and the USDA-NRCS pedon soil characterization database (Soil Survey Staff, 2010a) have been established to facilitate analyses of the soil hydraulic properties. Consequently, it is attractive to use such databases to develop PTFs for regions with less available data. Although the advantages of existing PTFs are well discussed in the literature (e.g. Tietje and Tapkenhinrichs, 1993, Espino et al., 1995, Wösten et al., 1995, Minasny et al., 1999), their applicability to calcareous soils has not been addressed thus far (Homaee and Farrokhian Firouzi, 2008). Also, practical applications of most PTFs is often hampered by their very specific data requirements.

Hydrologists in countries with insufficient soil hydraulic or taxonomic data are often confronted with situations where one or more PTF input parameters is not available. The Rosetta package (Schaap et al., 2001) implements five hierarchical PTFs to predict the WRC, as well as the saturated and unsaturated hydraulic conductivity. The hierarchy in PTFs allows prediction of van Genuchten hydraulic parameters and the saturated hydraulic conductivity using limited (soil textural class only) to more extended (texture, bulk density, and one or two water retention points) input data. This makes Rosetta quite useful for countries having data limitations.

Calcareous soils are widely distributed over the world and physically may not behave the same as non-calcareous soils. Calcareous soils contain 5% or more (by volume) inorganic carbon or CaCO₃ equivalent (Soil Survey Staff, 1999). Soils of this type are the dominant formation in many areas, particularly in arid and semiarid regions. According to the FAO, about 65% of soils in Iran are calcareous (FAO/UNDP, 1972). Inorganic carbon in soils commonly exists in the form of the carbonate minerals calcite (CaCO₃), dolomite [Ca, Mg(CO₃)₂], and as magnesium calcites (Ca_1−xMg_xCO₃). Other less common forms are aragonite (CaCO₃) and siderite (FeCO₃). Carbonates in soils may be of primary (inherited from parent materials) or secondary (pedogenic) origin. Secondary carbonates usually consists of aggregates made up of silt- and clay-sized calcite crystals, which are easily identified using grain mount techniques. Large crystals of calcite or dolomite are mostly of primary origin (Doner and Lynn, 1989).

Qualitative and quantitative knowledge of Ca and Mg carbonates is useful in many agronomic and environmental studies, such as for investigations of root growth, water movement, contaminant transport, soil pH (Nelson, 1982) and the nature of the exchange complex (Goh and Mermut, 2006). Studies that considered the effects of calcium carbonate precipitation on reductions in pore space suggest that neither water retention nor the hydraulic conductivity is affected significantly by calcium carbonate precipitation (Frenkel et al., 1978). However, from water retention and adsorption energy points of view, carbonates of the clay size can be treated as silt particles (Soil Survey Staff, 2010b). Thus, the hydraulic properties of calcareous soils may well differ from non-calcareous soils. Also, calcium carbonate is routinely removed from soil when soil texture is determined. The properties of soils with large CaCO₃ contents hence may well change during standard hydraulic analyses, which may further complicate the use of PTFs for non-calcareous soils. Hence, the applicability of available PTFs, such as those in the Rosetta package, to calcareous soils remains largely untested.

The objectives of this study were to (1) derive and verify pedotransfer functions of several typical calcareous soils, and to test whether the inclusion of soil CaCO₃ content as an input parameter would improve the accuracy of the PTFs, and (2) to investigate if the Rosetta PTFs, which were not derived specifically for calcareous soils, can be applied to calcareous soils.