brown forest soils, chernozems, conditional inference framework (cforest), field capacity, hygroscopic water content, meadow soils, saturated water content, soil water retention, wilting point
Abstract
Relationship between easily available soil properties and soil water retention at given matric potentials were analysed on brown forest soils, chernozems and meadow soils of Hungarian Detailed Soil Hydrophysical Database (Hungarian acronym: MARTHA). We studied the influence of soil properties displayed on the 1:10000 scale Hungarian soil maps on soil water retention at -0.1, -33, -1500 and -150000 kPa. Continuous (particle size distribution, organic matter content, calcium carbonate content and pH) and category type (ordinal: soil texture, ordinal type information on organic matter content, calcium carbonate content and pH; nominal: soil subtype classes) variables were used in the analyses. The relationships was analysed with random forest method based on conditional inference trees (cforest). Water retention of different soil types was characterized. Importance of soil properties in the prediction of soil water content varies according to soil type and matric potentials.
