Testing laboratory methods to determine the anisotropy of saturated hydraulic conductivity in a sandy-loam soil

Anisotropy, a (the log of the ratio of horizontal to vertical conductivity, log10(Kh/Kv)), of saturated soil hydraulic conductivity, Ks, affects transport processes in soil but is not routinely measured, probably because practical and validated methods are lacking. The objective of this investigation was to determine the effects of different constant-head laboratory and sampling procedures on anisotropy of saturated hydraulic conductivity measurements. The sequence of Ks measurements was varied (vertical conductivity, Kv, first, then horizontal, Kh, second and vice versa) for an experimental set-up considering five variables: 1) water ponding type (Mariotte or siphon); 2) saturation state prior to experiment start (unsaturated or saturated); 3) experiment duration (long or short); 4) sample geometry or extraction (cube or core); and 5) sample volume. The Mariotte, unsaturated/saturated, long experiment for a single soil cube resulted in unreliable mean anisotropy results, where a differed in sign. Generally, the sequence of measurements had a negligible impact on a for a siphon, saturated, short experiment for a soil cube. Furthermore, different a were obtained by varying the undisturbed soil sample collection procedure (cube vs. core). The conclusion of this investigation was that using a siphon and a short-duration run on an initially saturated cube of soil encased in foam is generally expected to yield reliable bi-directional Ks results. However, a check of the independence of the estimated anisotropy on the order of measurements for the sampled soil is recommended. An alternative procedure to determine a mean anisotropy for an area of interest would be to measure Kv and Kh on different soil cubes. Finally, the ratio between the mean Kh and Kv results varied from a not statistically significant factor of 1.02 to a statistically significant factor of 1.95 during the one-year investigation period (five sampling dates). Therefore, anisotropy of this sandy–loam soil varied with time but it was always low or negligible.