|Guber, Andrey - VISITING SCI., ARS|
|Shein, Eugeny - MOSCOW STATE UNIV.|
Submitted to: Development of Pedotransfer Functions in Soil Hydrology
Publication Type: Book / Chapter
Publication Acceptance Date: December 22, 2003
Publication Date: July 14, 2004
Citation: Pachepsky, Y.A., Rawls, W.J., Guber, A., Shein, E. 2004. Soil aggregates and water retention. Elsevier, Amsterdam. Development of Pedotransfer Functions in Soil Hydrology. 30(8):143-150. Interpretive Summary: Soil, when dried and sieved, breaks into aggregates of various sizes. Aggregate size distribution has been shown to characterize soil structure, and spatial arrangement of pores is soil pore space. Because the spatial arrangement of pores is directly related to soil's ability to retain and transmit water, aggregate size distributions should provide useful information about soil hydraulic properties. Although such hypothesis seems to be plausible, results of its testing have never been reported. The objective of this work was to test such hypothesis. We used a substantial data set containing water retention data, aggregate size distribution and primary particle size distribution for 142 samples of fine-textured soils, such as loams, silt loams, silty clays, and clays. Shape-related parameters of soil water retention curves were selected to examine whether and how aggregate size distribution is related to water retention. The exploratory technique of regression trees was used to see whether splits of the data set in subsets with similar water retention can be made based on aggregate size distribution and particle size distribution of individual samples. The total content of small aggregates and the total content of large aggregates were the most important splitting parameters. This indicates that aggregate size distributions, if available in soil databases or feasible to measure, can be useful in estimating parameters of soil water retention from other soil properties.
Technical Abstract: Aggregate size distributions are used to characterize soil structure. Soil structure is known to affect soil water retention. We hypothesized that aggregate size distributions could yield parameters useful for estimating soil water retention. To test this hypothesis, we assembled a data base of 142 samples for which dry aggregate size distributions for 10 size fractions, particle size distributions for 6 size fractions, and water retention were measured. The dataset included samples of loam, silt loam, silty clay loam and silty clay soils, represented with 11, 28, 47, and 56 samples respectively. The van Genuchten equation was fitted to gravimetric water retention data. Regression trees were used to examine the effect of aggregate size distributions and particle size distributions on van Genuchten parameters alpha, n and saturated water content. Cumulative distributions rather than contents of separate size fractions were used in data analysis. Parameters of the aggregate size composition appeared to be important to split the dataset into homogeneous subsets using regression trees. Contents of either small aggregates (<0.25 mm, <0.5 mm, <1 mm) or large aggregates (>7 mm, >10mm) were the grouping variables in most cases for parameters of the van Genuchten equation. Rsults demonstrate that water retention reflects a complex interaction of texture and structure jointly affecting soil water retention. This indicates that aggregate size distributions, if abailable in soil databases or feasible to measure, can be useful in estimating parameters of soil water retention from other soil properties.