Search

• A new method to simultaneously estimate soil hydraulic properties is presented. • K s , α and n are estimated from inverse analysis of an upward infiltration curve. • The method is evaluated on different synthetic and experimental soils. • Accurate estimates of K s , α and n are obtained in the theoretical curves. • The method allows accurate estimates of K s and n in all experimental soils. This paper presents a new laboratory method to simultaneously estimate K s and α and n parameters of the van Genuchten (1980) θ (h) from the inverse analysis of an upward infiltration curve measured in a 5-cm high soil column. The method was evaluated on synthetic 1D infiltration curves generated for a theoretical loamy sand, loam and clay soil. The influence of the soil initial condition on the inverse analysis was also studied. Next, an optimization method was presented and tested on eight theoretical soils (from loamy sand to clay). The method was subsequently applied to experimental curves measured on five sieved soils (from sand to clay) packed in 5-cm high and diameter cylinders. The K s , α and n values estimated from the inverse analysis of the experimental curves were compared to those measured by Darcy and the pressure cell method (PC). The initial soil tension, h i , which had an important influence on the optimization, was fixed to −6.0·105 cm. The optimization method resulted robust and allowed accurate estimates of the actual hydraulic parameters. A close to one relationship (R2 = 0.99) was observed between the theoretical K s , α and n and the corresponding values obtained with the inverse analysis. Regarding to the experimental soils, significant relationships close to one were obtained between K s and n (R2 > 0.98) estimated from inverse analysis and those measured with Darcy and PC. A non-significant relationship with slope away from one was found for α. [ABSTRACT FROM AUTHOR]

Highlights • This work analyses the influence of the β on the estimate of K and S. • Early infiltration times (i.e. 100 s) promoted good estimations for S. • Longer infiltrations (i.e. 1.000 s) were required to estimate accurately K s. • Only very long infiltration (i.e. 10.000 s) allowed defining the actual β value. • S and K s can be accurately estimated using a constant β. Abstract Soil water infiltration can be described with the quasi-analytical Haverkamp et al. (1994) equation, defined by the hydraulic conductivity (K s), sorptivity (S) and the β parameter. K s and S are commonly estimated from the transient cumulative infiltration curve, using a constant β value. The objective of this work was to study the influence of β on the estimation of K s and S. The study was first performed on synthetic 1-D infiltration curves generated at different infiltration times for loam sandy, loamy and silty soils, and next extrapolated to a 3-D loam synthetic soil and on 10 infiltrations curves measured on the field with a disc infiltrometer on different types of soils. The infiltration measurements lasted between 600 and 900 s. The results showed that, while early infiltration times (i.e. 100 s) promoted good estimations for S , longer infiltrations (i.e. 1.000 s) were required to estimate accurately K s. Only very long infiltration (i.e. 10.000 s) allowed defining the actual β value. A similar behaviour was observed for the 3-D infiltration measurements. Except for β , significant relationships (R2 = 0.99) were obtained between the K s and S of the three theoretical soils and the corresponding values calculated by optimizing the three hydraulic parameters on 2.000 s 1-D infiltration curves. The large confidence interval observed in β (between 0.3 and 2.0) resulted from the fact that β had a small effect on the infiltration curve. A significant relationship (R2 = 0.99) was also obtained between the K s and S optimized from the experimental curves using a β = 1.1 and the corresponding values obtained by simultaneous optimization of the three hydraulic parameters. These results demonstrated that S and K s can be accurately estimated using a constant β and that downward infiltration is not an appropriate procedure to estimate β. [ABSTRACT FROM AUTHOR]

Summary In-situ measurements of soil hydraulic properties on covered soil surfaces (i.e. vegetated or residue covered surfaces) are of paramount importance in many agronomic or hydrological researches. These soil parameters are commonly estimated with the tension infiltrometry technique. This paper presents a portable and modified design of the hood infiltrometer (MHI) that, unlike to the original hood infiltrometer, allows estimating the soil hydraulic properties from the transient cumulative infiltration curve. The MHI consists of a water-supply reservoir attaches to a hat-shaped base placed on the soil surface. The base of the hat is closed by a system of sticks and a malleable material ring. To test the viability of this new design, the hydraulic conductivity ( K s ) estimated with MHI in a loam soil using the multiple head approach was compared to the corresponding values calculated from the transient infiltration curve analysis. Next, the MHI was tested on three different soils at saturated conditions, and the sorptivity ( S ) and K s estimated by the transient infiltration curve analysis were compared to the corresponding values obtained with a disc infiltrometer (DI). An additional field experiment was performed to compare the hydraulic properties measured with MHI on a bare soil and a soil covered with plants. Results demonstrated that this design allows hermetically closing the base of the hat without disturbing the soil surface. The K s estimated with the multiple head approach was not statistically different ( p = 0.61) to that obtained with the transient infiltration curve analysis. No significant differences between the K s ( p = 0.66) and S ( p = 0.50) values estimated with DI and MHI were observed. The S values measured with MHI on the covered soil surface were significantly higher than that measured on the adjacent bare soil. These results indicate that MHI can be a viable alternative to estimate the hydraulic properties of covered soils from the measured transient infiltration curve. [ABSTRACT FROM AUTHOR]

Abstract: This paper describes a new method (NSQE) to estimate soil hydraulic properties (sorptivity, S, and hydraulic conductivity, K) from full-time cumulative infiltration curves. The technique relies on an inverse procedure involving the quasi-exact equation of Haverkamp et al. (1994). The numerical resolution is described and the sensitivity of the method is theoretically evaluated, showing that the accuracy of the estimates depends on the measured infiltration time. A new procedure to detect and remove the effect of the contact sand layer on the cumulative infiltration curve is also given. The method was subsequently compared to the differentiated linearization procedure (DL), which calculate K and S from the simplified Haverkamp et al. (1994) equation, valid only for short to medium times. A total of 264 infiltration measurements performed with a 10cm diameter disc under different soil conditions were used. Compared to the DL procedure, field measurements showed that the NSQE method allowed better estimates of soil hydraulic properties, independently on the infiltration noise and the presence of contact sand layer. Overall, although comparable S values were estimated with both methods, the longer infiltration times allowed by the proposed method made this procedure more accurate estimations of K. In conclusion, the NSQE method have shown to be a significant advance to accurate estimate of the soil hydraulic properties form the transient water flow. [Copyright &y& Elsevier]

• This work analyzes the influence of layered soils on K s and S estimates. • The Sequential Infiltration Analysis (SIA) method was proposed. • SAI was tested on synthetic and laboratory columns and experimental soils. • Estimates of K s and S improved when the SIA method was applied. • Method also allowed determining the thickness of the upper soil layer. The soil sorptivity, S , and saturated hydraulic conductivity, K s , can be estimated from the inverse analysis of a cumulative infiltration curve using the quasi-exact implicit (QEI) formulation or its corresponding 4-Terms (4T) approximation. Although these models consider the soil as homogeneous media, there is no information about how heterogeneous profiles can affect the inferred soil properties. This work analyzes the influence of layered soils on K s and S estimates using QEI and 4T models, and designs a new procedure for treating infiltration curves measured on layered soil profiles. The Sequential Infiltration Analysis (SIA) method considers a sequence of increasing time series from the cumulative infiltration data to estimate K s and S , and its corresponding RMSE as a function of the number of samples used. A procedure to estimate the thickness of the upper uniform soil layer from the estimated wetting front advance (WFA) is also reported. The SIA method was applied on: (i) synthetic homogeneous profiles of loam soil and six layered profiles involving a 1, 2 and 3 cm thickness loam layer over silty or sandy loam soils, respectively, (ii) stratified laboratory soil columns, and (iii) 20 experimental infiltrations performed in a semiarid region of North-Eastern Spain. Similar results were found between QEI and 4T models for all cases. Erroneous estimates of K s and S were observed when the total infiltration time series was considered for the analysis, regardless of the presence of soil layering. In opposite, estimates improved when the SIA method was applied to the layered systems. The SIA method exploits the fact that the RMSE increases when the wetting front reaches the interface between the soil layers. Such increase allows: (i) detection of the soil heterogeneity, (ii) determination of the infiltration time, t o , required for the wetting front to reach the lower layer, and, (iii) accurate estimates of the upper layer K s and S along with its thickness. Laboratory experiments on layered soils and field measurements demonstrated that the SIA method could be satisfactorily applied on different curves with contrasting shapes and magnitudes. Although soil layering encountered on most field samplings restricted the treatment of the observed infiltrations to short-medium times, the SIA method allowed robust estimates of K s and S. These results indicate that the proposed method is a promising tool for characterizing the hydraulic properties of layered and heterogeneous soil profiles. [ABSTRACT FROM AUTHOR]

Summary Based on the analysis of Haverkamp et al. (1994), this paper presents a new technique to estimate the soil hydraulic properties (sorptivity, S , and hydraulic conductivity, K ) from the full-time cumulative infiltration curves. The proposed method, which will be named as the Numerical Solution of the Haverkamp equation (NSH), was validated on 12 synthetic soils simulated with HYDRUS-3D. The K values used to simulate the synthetic curves were compared to those estimated with the NSH method. A procedure to detect and remove the effect of the contact sand layer on the cumulative infiltration curve was also developed. A sensitivity analysis was performed using the water level measurement as uncertainty source and the procedure was evaluated considering different infiltration times and data noise (e.g. air-bubbling in the infiltrometer). The good correlation between the K used in HYDRUS-3D to model the infiltration curves and those obtained by the NSH method ( R 2 = 0.98) indicates this technique is robust enough to estimate the soil hydraulic conductivity from complete infiltration curves. The numerical procedure to detect and remove the influence of the contact sand layer on the K and S estimates resulted to be robust and efficient. A negative effect of the curve infiltration noise on the K estimate was observed. The results showed that infiltration time was an important factor to estimate K . Smaller values of K or lower uncertainty required longer infiltration times. In a second step, the technique was tested in field conditions on 266 different soils at saturation conditions, using a 10 cm diameter disc infiltrometer. The NSH method was compared to the standard differentiated linearization procedure (DL), which estimates the hydraulic parameters using the simplified Haverkamp et al. (1994) equation, valid only for short to medium times. Compared to DL, NSH was considerably less affected by the infiltration bubbling and the contact sand layer, and allowed more robust estimates of K and S . Although comparable S values were obtained with both methods, the NSH technique, which is not limited to short times, resulted in more accurate and robust estimates for K . This paper demonstrates the NSH method is a significant advance to estimate of the soil hydraulic properties from the transient water flow. [ABSTRACT FROM AUTHOR]