EVALUATION OF THE ARYA-PARIS MODEL FOR ESTIMATING WATER RETENTION CHARACTERISTICS OF LIGNITIC MINE SOILS.

Mine soil materials may be viewed as man-made systems that consist of spatially disordered soil and sediment components, which are in an initial stage of soil development. A question is whether methods and approaches developed for natural soils may also be used for such artificially created soil materials. The applicability of the Arya and Paris pedotransfer function to obtain hydraulic properties from the particle size distribution and bulk density was tested for lignitic mine soil material of the Lusatian Lignite Mining District in eastern Germany. The scaling factor [alpha] in this model was evaluated by (i) fitting of the water retention curves estimated with the Arya-Paris model (APM) to measured water retention data and (ii) interpretation of [alpha] as a fractal dimension of the pore channels and derivation of this fractal dimension from the fractal dimension of the particle size distribution. The two tested fractal approaches resulted in relatively inaccurate predictions of the water content. The use of a single fitted [alpha] value for each depth yielded [alpha] values between 1.05 and 1.47. Because of the inability of the APM to account for residual water contents in this sandy soil material, a correction was applied. The cumulative mass fraction fractal method did not improve the estimation in comparison with the retention curves calculated with a constant [alpha] value of 1.38. The closest fits with the data were obtained by using a variable [alpha] value that depended on the particle size. The accuracy of the predictions of the APM in the higher suction range could be improved (lower mean deviations and root mean square deviations of water content) by using a linear water content-dependent correction factor. Better estimates of water retention in the relatively dry range may be significant for simulation of water budgets of mine spoil sites in the Lusatian Mining District.