A database containing about 800 soil profiles located in a 7-km grid covering Denmark has been used to develop a set of regression equations of soil water content at pressure heads –1, -10, -100 and –1500 kPa versus particle size distribution, organic matter, CaCO3 and bulk density. One purpose was to elaborate equations based of soil parameters available in the Danish Soil Classification's texture database on particle size distribution and organic matter. It was also tested if inclusion of bulk density or CaCO3 content (in CaCO3-containing samples) as predictors or grouping in surface and subsurface horizons or textural classes improved the regression equations. Compared to existing Danish equations based on much fewer observations the accuracies of the new equations were better. The equations also predicted the soil water content at the measured pressure heads more accurate than the pedotransfer functions developed in HYPRES (Hydraulic Properties of European Soils). Introducing bulk density as a predictor improved the equation for pressure head –1 kPa but not for lower ones. The grouping of data sets in surface and subsurface horizons or in textural classes did not improve the equations. Based on the equations a set of van Genuchten parameters for soil types in the Danish Soil Classification was elaborated. The prediction of soil water content, especially at pressure head –1 kPa, is more accurate using these van Genuchten parameters than using the pedotransfer functions developed in relation to the HYPRES database from a broad range of European soils.
|Status||Udgivet - 2005|