Comparing field-saturated soil hydraulic conductivity determined by the single-ring pressure infiltrometer and bottomless bucket methods

Reliable estimates of field-saturated soil hydraulic conductivity, K_fs, are necessary for characterizing and modelling flow and solute transport. However, field determination of K_fs is challenging since this parameter is sensitive to the applied measurement method and the underlying assumptions. Infiltration experiments provide a relatively simple and low-cost way to determine K_fs, but different devices and analysis frameworks can yield method-dependent estimates. In this context, it is necessary to benchmark simple field techniques against more established methods and to quantify how the choice of the method influences estimation of K_fs before such techniques can be widely adopted.
The aim of this investigation was to compare the K_fs values obtained with two single-ring infiltrometers methods. In particular, the classical single-ring pressure infiltrometer (PI) and the bottomless bucket (BB) method were applied in three Mediterranean agricultural fields in Italy under summer dry conditions. The soils were classified as sandy-loam at two sites and clay at the third one.
At each site, 15 PI runs with two consecutively applied ponded depths of water, H (H₁ = 5 cm, H₂ = 10 cm), and 15 BB falling-head runs (H ranging repeatedly from 10 to 1 cm) were performed. The data obtained with the PI were analyzed using two different approaches. In particular, the Two-Ponding-Depth (TPD) approach was applied for estimating both K_fs and the site-specific sorptive number, α*. The One-Ponding-Depth (OPD) approach was also applied by using the site-specific α* value and averaging the K_fs estimates for the two ponded depths of water. The site-specific α* parameter was also used to analyze the data obtained with the BB method.
The K_fs distribution was predominantly log-normal for all developed datasets whereas the α* distribution was normal. The α* values were consistent with expectations based on soil texture for the three sites (α* ≈ 4.7 m^-1 in the finer soil and 9.7–16.4 m^-1 in the coarser soils).
At each site, mean K_fs values differed by up to 1.2 times in the two sandy loam soils (70-85 and 249-293 mm h^-1) and by up to 1.9 times in the clay soil (150-279 mm h^-1). Coefficients of variation ranged from 27% to 130%, depending on soil type and measurement method, with both extremes observed in the clay soil. Differences among soils depended on the experimental method (PI, BB) but not on the PI data analysis approach.
In conclusion, a satisfactory correspondence between the PI and BB methods can be expected in sandy-loam soils but less in clay soils. In any case, the differences between the two methods seem more appreciable with reference to K_fs data variability than the mean value of this soil hydrodynamic parameters.