1 Soil physics and Soil resources, Faculty of Agricultural Sciences, Aarhus University, Aarhus University2 Department of Agroecology and Environment, Faculty of Agricultural Sciences, Aarhus University, Aarhus University3 Institute of Civil Engineering, University of the Philippines-Diliman, Quezon City, Philippines4 Institut for Kemi, Miljø og Bioteknologi5 Department of Agroecology - Soil Physics and Hydropedology, Department of Agroecology, Science and Technology, Aarhus University6 University of Arizona, Tucson, AZ7 Saitama University, Saitama8 Department of Agroecology - Soil Physics and Hydropedology, Department of Agroecology, Science and Technology, Aarhus University
Accurate description of the soil-water retention curve (SWRC) at low water contents is important for simulating water dynamics, plant-water relations, and microbial processes in surface soil. Soil-water retention at soil-water matric potential of less than -10 MPa, where adsorptive forces dominate over capillary forces, has also been used to estimate the soil specific surface area (SA, m2 kg-1). In the present study, the dry part of the SWRC was measured by dewpoint potentiameter on 41 Danish soils covering a wide range of clay (CL, kg kg-1) and organic carbon content (OC, kg kg-1). It was found that measurements by traditional pressure plate apparatus generally overestimated water contents at -1.5 MPa (plant wilting point). The 41 soils were classified into four textural classes based on the so-called Dexter index n (= CL/OC), and the Tuller-Or (TO) general scaling model describing the water film thickness at a given soil-water matric potential (< -10 MPa) was evaluated. Only for low organic soils with n > 10, the estimated SA from the dry soil-water retention was in good agreement with the SA measured using ethylene glycol monoethyl ether (SA_EGME). A strong relationship between the ratio of surface area estimates and the Dexter index n was observed and applied as an additional scaling function in the TO model to rescale the soil-water retention curve at low water contents across soil textural classes. However, the TO model still overestimated the water film thickness at potentials approaching oven dry condition (~ -800 MPa). Alternatively, the linear Campbell-Rossi-Nimmo (CRN) model for pF (where pF = log (negative of matric potential; cm H2O)) versus volumetric water content (m3 m-3) showed accurate fits to data for all 41 soils from -10 to -800 MPa, and yielded high correlations between CRN model slope (1/B) and SA and CL. We, therefore, recommend to use the empirical CRN model for predicting the dry part of the water retention curve (-10 to -800 MPa) from measured soil texture or surface area and, in perspective, to modify the more conceptual TO model to obtain better descriptions in the very dry range (-300 to -800 MPa).
Main Research Area:
ASA, CSSA, and SSSA 2010 International Annual Meetings