Submitted to: Soil Science
Publication Type: Peer Reviewed Journal
Publication Acceptance Date: July 30, 2000
Publication Date: November 1, 2000
Interpretive Summary: Agronomists, agrometeorologists, hydrologists, hydrogeologists, and engineers are often expected to give or use predictions of soil ability to conduct water when little or no relevant measurement results are available. Such hydraulic conductivity predictions can be done using more readily available soil data. It has long been known that the hydraulic conductivity is closely related to the ability to retain water in soils. However, the water transport in soils is affected by soil connectivity and tortuosity that manifest themselves in water retention to a much lesser extent than in hydraulic conductivity. Therefore, empirical parameters are needed to reflect the effect of pore connectivity and tortuosity on hydraulic conductivity. Since the hydraulic conductivity happens to be one of the most time- and labor-consuming soil properties to measure, the less the number of the empirical parameters, the fewer the measurements that need to be done. This work employs the simple hypothesis that connectivity and tortuosity effects depend on pore radii. The resulting model that has only one empirical parameter was successfully tested with the data from the large database, Unsaturated Soil Hydraulic Database. We attempted to relate the empirical connectivity/tortuosity parameter to soil texture, bulk density, and parameters of the water retention curve. The accuracy appeared to be low, and only about 50 percent of variation could be explained by the aforementioned soil properties. Therefore, measurements of the unsaturated hydraulic conductivity appear to be unavoidable, but the number of these measurements can be substantially reduced as only one parameter has to be estimated from these measurements.
Estimating unsaturated hydraulic conductivity (UHC) often relies on using water retention characteristics. Because the water retention curves do not provide data about the pore connectivity, an empirical correction is used in capillary bundle models that are fitted to UHC data. The majority of authors applied the macroscopic correction expressed as a function of water rcontent. A microscopic correction term, expressed as a function of pore radius, was proposed in literature but never applied to soil data. The purpose of this work was to apply the 'hydraulic conductivity-water retention' model with the microscopic connectivity correction to a large data set to see what accuracy could be achieved and whether it is possible to relate the connectivity parameters to some readily available soil properties. Data for 147 soil horizons were extracted from the Unsaturated Soil Hydraulic Database. Water retention and hydraulic conductivity data were in the range of capillary pressures more than 5 kPa and from 5 to 200 kPa, respectively. Sand, clay, and silt contents, as well as bulk density values, were reported for all those samples. The model provided an accurate approximation, so the root mean square error in estimated log10k was less than 0.12 and 0.3 in 50 percent and 90 percent of all cases, respectively. Two parameters of the model appeared to be correlated, so only one parameter was needed. The polynomial-network dependence of the connectivity parameter on soil texture, bulk density, or parameters of the water retention curve was not accurate and explained only about 50 percent of variation. The fact that only one empirical parameter is needed to describe the UHC permits the reduction of the number of measurements.