Abstract. A conceptual model based on the assumption that soil structure evolves from a uniform random fragmentation process is proposed to define the water retention function. The fragmentation process determines the particle size distribution of the soil. The transformation of particles volumes into pore volumes via a power function and the adoption of the capillarity equation lead to an expression for the water retention curve. This expression presents two fitting parameters only. The proposed model is tested on water retention data sets of 12 soils representing a wide range of soil textures, from sand to clay. The agreement between the fitted curves and the measured data is very good. The performances of the model are also compared with those of the two-parameter models of van Genuchten [1980] and Russo [1988] for the water retention function. In general, the proposed model exhibits increased flexibility and improves the fit at both the high and the low water contents range. IntroductionThe solution of the flow equation of water in soils requires the expression of two soil hydraulic characteristics, the water retention curve (WRC) and the hydraulic conductivity function (HCF). The WRC describes the relationship between the soil capillary head, ½, and the volumetric water content, 0. The HCF describes the relationship between the unsaturated hydraulic conductivity, K, and 0. Different models permit defining the HCF in terms of the WRC [Mualem, 1986]. Therefore the WRC can be considered to be one of the most fundamental hydraulic characteristics of a soil. The experimenfal determination of the WRC is tedious and time consuming. Therefore the WRC is not always present in the usual data sets presenting the basic properties of soils. When it is available, it is a discrete representation of a limited number of volumetric water content-capillary head points within the range of the water matric tensions under interest. Consequently, intensive efforts were and are still invested in developing mathematical functions to be fitted to the available set of measured points in order to provide a continuous expression of the WRC. This study is a contribution to this effort. It assumes that the soil particle size distribution (PSD) stems from a fragmentation process and derives the void size distribution (VSD) from the PSD. Approaches inwhere ½o is the capillary head at the inflection point and Se, the effective saturation degree, is defined as A different approach was to develop expressions for the WRC starting from the particle size distribution (PSD) of the soil. Arya and Paris [1981] developed a model to predict the WRC of a soil from its PSD, bulk density, and particle density on the basis of the similarity in shape between the WRC and the PSD of a soil. They proposed the relationship between the mean pore radius, r, and the mean particle radius, R:where e is the void ratio, n is the number of spherical particles with radius r, and a is an empirical constant ranging from In terms of the similarity hypothesis used, a...

of shrinkage curves on undisturbed core samples allowed improvement of the knowledge of soil shrinking behav-The availability of methods for quasi-continuous measurements of ior, and to develop shrinkage curve models with differsoil shrinkage curves allowed the development of new models. The ent sets of parameters, as reviewed in Braudeau et al. exponential (XP) model allows the calculation of the volume of two pore phases in the soil, namely macro and micropore volumes. The (1999). micropore volume is identified with the pore volume of the soil clay Shrinkage curves generally present a typical sigmoid matrix according to the assumption that the maximum swelling of the shape, with linear and curvilinear parts separated by clay matrix (MS) is the point of minimum water content of the structransition points as drawn in Fig. 1. Similar to clay paste, tural shrinkage (Point D). This is discussed using undisturbed and a SL and an AE are observed on structured soil samples repacked soil samples with various clay contents and clay types. The (e.g., McGarry and Malafant, 1987; Tariq and Durnford, slope of the shrinkage curves as a function of equivalent saturated-1993b). However, the slope of the shrinkage curve for pore radius show a transition in pore type around a 10-m pore radius, water content greater than AE, called normal shrinkage where smaller and more deformable pores start to desaturate. This or basic shrinkage as discussed by Mitchell (1992), is corresponds to the fitted D point and is close to the size of the largest generally not equal to one (e.g., Lauritzen and Stewart, pores in the clay matrix or clay-silt phase reported in the literature. The calculated micropore volume and micropore swelling properties

No abstract

Soil compaction affects soil physical properties and, eventually, crop production. A severe drop in the productivity of the state of Parana, southern Brazil, was observed due to soil compaction. Two oxisols from this region, a Haplic Acrothox from the site of Cascavel and a Haplic Eutrothox from the site of Palotina, presenting different compaction behaviors in the field, are studied under laboratory conditions. Uniaxial compressive pressures, from 50 to 1000 kPa, are applied to soil samples at different initial matric potentials, varying from −0.1 to −1000 kPa. The bulk density of the Palotina soil is always higher than that of the Cascavel soil and is the highest when the initial matric tension is − 32 kPa. Differences in pH, cation‐exchange capacity, organic matter, and clay particle thickness also tend to explain the different compaction behaviors. A model of the soil bulk density increase during compaction is proposed and compared with a multiplicative model and a logarithmic model. The performances of the proposed and the multiplicative models are practically similar and better than those of the logarithmic model. The major advantage of the proposed model is that it has one fitting parameter less than the multiplicative model. Compaction affects the soil water retention curves for the whole range of matric tensions, up to − 100 MPa. An approach that allows the evaluation of the hydraulic conductivity functions of the compacted samples is proposed. Applied to the Brooks and Corey relationship, the main drying curves of the compacted samples are well reproduced using one fitting parameter only.

Summary The long‐term effects of intermittent flooding on soil properties were studied in field experiments on a Vertisol cropped with rice in Senegal. The dominant clay minerals were smectite and kaolinite. When the soil was reduced after flooding, its cation exchange capacity (CEC) increased to twice that of its oxidized, unflooded state. Mössbauer spectroscopy showed an increase in smectite structural FeII upon reduction, which explained a part of the increase in CEC. The rest of the increase was attributed to the removal of iron oxyhydroxide coatings by reductive dissolution. The reduction and dissolution of oxides under the field conditions were substantiated by analysis of the surfaces of vermiculites buried in the Ap horizons of the cropped and the non‐cropped soils. The redox‐induced CEC changes were found to be reversible after 22 cycles of rice cropping. Nevertheless, the structural Fe and free Fe contents of the rice field Ap horizon were less than those of soil in uncropped neighbouring land, suggesting that inundation induced weathering and eluviation of the minerals. The observed changes in CEC and related redox reactions may substantially modify proton, anion and cation balances in intermittently flooded soils.