JWSS - Journal of Water and Soil Science

Soil water retention curve (SWRC) reflects different states of soil moisture and describes quantitative characteristics of the unsaturated parts of the soil. Direct measurement of SWRC is time-consuming, difficult and costly. Therefore, many indirect attempts have been made to estimate SWRC from other soil properties. Using pedotransfer functions is one of the indirect methods for estimating SWRC. The aim of this research was to assess the effect of using soil particles percentage in comparison with the geometric characteristics of soil particles on the accuracy of the pedotransfer equations of SWRC and the critical point of it. Accordingly, 54 soil samples of Isfahan province from seven texture classes were used. The most suitable functions for estimating SWRC, parameters of van Genuchten and Brooks-Corey equations, and the critical point of SWRC were selected based on statistical indices. The results indicated that the pedotransfer equations fitted the SWRC data well and the outputs from them were in a good agreement with the independent (validation) SWRC data. The results revealed that using soil particles percentage (sand and clay), bulk density and organic matter content in the point estimation of SWRC was better than applying geometric properties of the soil particle diameter. On the other hand, in the estimation of parametric and critical point of SWRC, using the geometric properties of soil particle diameters resulted in more satisfactory results, as compared with using the soil particles percentage. The NRMSE values indicated that the accuracy of the pedotransfer equations in the lower matric head was greater than that of the higher matric head.

Nowadays, the Beerkan computational algorithms (BESTslope and BESTsteday) are known as the suitable indirect methods for estimating soil saturated hydraulic conductivity (Ks) and sorptivity (S), as well as the scale parameter (hg) in van Genuchten soil water retention equation through the data collected in the Beerkan infiltration experiment and other required data. The purpose of this study was to compare these algorithms in estimating Ks and S, as well as the soil water content corresponding to the suctions of 33 kPa, 100 kPa, 200 kPa, 300 kPa and 1500 kPa. For this purpose, a total of 40 Beerkan infiltration experiments were carried out in Sistan dam research field. From all Beerkan experiments, 30 tests in loam and sandy loam textures having a relative error less than 5.5% (Er <5.5%) were selected for further analysis. The statistical criteria RMSE, ME and ωr2 were used to compare the measured and estimated water content values at each suction. The results showed that the BESTsteday algorithm, which had a more simple calculating process than the main algorithm (i.e. BESTslope), could provide the Ks and S values and the soil water content of the near field capacity with an acceptable accuracy. The model performance in estimating water content corresponding to the 1500 kpa suction head (i.e. _θfc) was not acceptable for both algorithms. Moreover, the relative error of estimating soil water content (Er(_h,θ)) was decreased gradually by an increase in clay %.