Tillage practices have a significant influence on the soil hydro-physical properties. The objective of this work was to evaluate the effect of tillage on the α (a scaling factor) and n (a pore size distribution parameter) van Genuchten soil water retention curve parameters during an 18-month long fallow period in a semiarid dryland. Three different tillage systems employed during 23 years of trials were compared: conventional (CT), reduced (RT) and no-tillage (NT). Measurements of soil bulk density (ρb) and the soil water retention curve θ(ψ) were performed at 0–10, 10–20 and 20–30cm soil depths. The θ(ψ) was determined with the Time Domain Reflectometry (TDR)-pressure cells at the following pressure heads: 0.5, 1.5, 3, 10, 50, 100, 500 and 1500kPa. From these data, α, n and the SDexter index were evaluated. The 0–40cm depth soil volumetric water content, θ, was also measured in the field using the TDR technique. Compared with CT and RT, NT had the highest θ values during all the fallow period. No significant influence of soil depth on θ(ψ) was observed in all tillage treatments at each sampling date. Although under consolidated soil conditions no significant differences in ρb and the water content at saturation (θs) were observed among tillage treatments, NT had the highest and lowest values of ρb and θs, respectively. The loosening of soil due to tillage practices in CT and RT significantly decreased ρb and increased θ at the wet-end section of θ(ψ). Post-tillage rainfall resulted in significant decreases in θs, α and the maximum value of the pore size distribution (PSDmax). The different soil structure created by mouldboard ploughing (CT) and chiselling (RT) explained the higher PSDmax under RT than CT. The most important changes in θ(ψ) followed the first copious effective rainfall events (>10mm) after tillage. These facts enabled the soil to recover the pre-tillage water retention curve shapes and the van Genuchten parameters pre-tillage values. Effective rainfall events in the late fallow had a minor effect on the water retention curve. Although tillage tended to increase n, this change was not significant. The SDexter index, which was also affected by tillage, was greater than 0.035 during all the fallow period, indicating good soil physical quality.

Determination of saturated hydraulic conductivity, Ks, and the shape parameters α and n of the water retention curve, θ(h), is of paramount importance to characterize the water flow in the vadose zone. This work presents a modified upward infiltration method to estimate Ks, α and n from numerical inverse analysis of the measured cumulative upward infiltration (CUI) at multiple constant tension lower boundary conditions. Using the HYDRUS‐2D software, a theoretical analysis on a synthetic loam soil under different soil tensions (0, 0–10, 0–50 and 0–100 cm), with and without an overpressure step of 10 cm high from the top boundary condition at the end of the upward infiltration process, was performed to check the uniqueness and the accuracy of the solutions. Using a tension sorptivimeter device, the method was validated in a laboratory experiment on five different soils: a coarse and a fine sand, and a 1‐mm sieved loam, clay loam and silt‐gypseous soils. The estimated α and n parameters were compared to the corresponding values measured with the TDR‐pressure cell method. The theoretical analysis demonstrates that Ks and θ(h) can be simultaneously estimated from measured upward cumulative infiltration when high (>50 cm) soil tensions are initially applied at the lower boundary. Alternatively, satisfactory results can be also obtained when medium tensions (<50 cm) and the Ks calculated from the overpressure step at the end of the experiment are considered. A consistent relationship was found between the α (R2 = 0.86, p < 0.02) and n (R2 = 0.97, p < 0.001) values measured with the TDR‐pressure cell and the corresponding values estimated with the tension sorptivimeter. The error between the α (in logarithm scale) and n values estimated with the inverse analysis and the corresponding values measured with pressure chamber were 3.1 and 6.1%, respectively. Copyright © 2016 John Wiley & Sons, Ltd.

Correct estimation of the soil-water retention curve (WRC) is of paramount importance to characterise the hydraulic behaviour of soils. This paper studies the influence of two different soil-wetting processes (waterlogging soil, WP; capillary rise to saturation, CRP) on the estimate of the WRC. The two procedures were applied on undisturbed loam soil samples with three degrees of soil structure: (i) consolidated soils under conventional tillage (CT), reduced tillage (RT) and no tillage (NT); (ii) freshly tilled soil under CT and RT; and (iii) CT and RT after secondary tillage plus some intense rainfalls events. WRCs were estimated with time-domain reflectometry (TDR) pressure cells and volumetric water content was measured at saturation conditions (for the WP method) and at pressure heads of 0.5, 1.5, 3, 5, 10, 50, 100, 500 and 1500 kPa. The same cores were used to determine the soil bulk density (ρb), which was subsequently used to estimate the saturated water content under CRP. The ρb value of the consolidated soil under NT was significantly higher (P < 0.001) than under CT and RT. No effect of the wetting process on the WRC of consolidated soils was observed. Only the freshly tilled soil samples under RT were significantly affected by the wetting process. In these cases, the water draining after WP collapsed the more unstable soil macropores and increased the volume of the smaller ones. However, this effect was minimised by the CRP method, which prevented the collapse of the more unstable soil pores. This work demonstrates that the soil-wetting process may have an important effect on the characterisation of the water-holding capacity on freshly tilled soils.

Determination of saturated hydraulic conductivity, K s , and the van Genuchten water retention curve θ(h) parameters is crucial to evaluate the unsaturated soil water flow. The aim of this work is to present a method to estimate K s , α and n from numerical analysis of an upward infiltration process at saturation (Cap0), with (Cap0 + h) and without (Cap0) an overpressure step (h) at the end of the wetting phase, followed by an evaporation process (Evap). The HYDRUS model as well as bruteforce search method were used on theoretical loam soil parameter estimation. The uniqueness and the accuracy of the solutions from the response surfaces K s-n, α-n and K s-α were evaluated at different scenarios. Numerical experiments showed that only the Cap0 + Evap and Cap0 + h + Evap scenarios were able to univocally estimate the hydraulic properties. The method gave reliable results in sand, loam and clay loam soils.

Estimation of the soil–water retention curve, θ(h), on undisturbed soil samples is of paramount importance to characterise the hydraulic behaviour of soils. Although a method of determining parameters of the water retention curve (α, a scale parameter inversely proportional to mean pore diameter and n, a measure of pore size distribution) from saturated hydraulic conductivity (Ks), sorptivity (S) and the β parameter, using S and β calculated from the inverse analysis of upward infiltration (UI) has been satisfactorily applied to sieved soil samples, its applicability to undisturbed soils has not been tested. The aim of the present study was to show that the method can be applied to undisturbed soil cores representing a range of textures and structures. Undisturbed soil cores were collected using stainless steel cylinders (5cm internal diameter×5cm high) from structured soils located in two different places: (1) an agricultural loam soil under conventional, reduced and no tillage systems; and (2) a loam soil under grazed and ungrazed natural shrubland. The α and n values estimated for the different soils using the UI method were compared with those calculated using time domain reflectometry (TDR) pressure cells (PC) for pressure heads of –0.5, –1.5, –3, –5, –10 and –50kPa. To compare the two methods, α values measured with UI were calculated to the drying branch of θ(h). For each treatment, three replicates of UI and PC calculations were performed. The results showed that the 5-cm high cylinders used in all experiments provided accurate estimates of S and β. Overall, the α and n values estimated with UI were larger than those measured with PC. These differences could be attributed, in part, to limitations of the PC method. On average, the n values calculated from the optimised S and β data were 5% larger than those obtained with PC. A relationship with a slope close to 1 fitted the n values estimated using both methods (nPC=0.73 nUI+0.49; R2=0.78, P<0.05). The results show that the UI method is a promising technique to estimate the hydraulic properties of undisturbed soil samples.