Bulk densities of Ghanaian forest soils in relation to other physico-chemical soil parameters

Prediction of soil bulk density (BD) requires taking several representative volumetric soft samples which is often laborious and difficult, particularly for wet and stony mineral soils. Alternative empirical models to predict BD under teak (Tectona grandis) plantations from physico-chemical soil properties, which are easier to obtain, are presented. Core samples were taken from the 0-20 cm (A-horizon) and 20-40 cm (B-horizon) soil layers from 28 soil pedons in Bosomoa, Tain II and Yaya forest reserves in The High Forest Zone of Ghana. Multiple linear regression techniques were used to develop statistical relationships to estimate BD for the A-horizon. The data were grouped according to their soil taxonomy classification. BD was negatively correlated with soil organic matter content (SOM) (r = -0.85, p -3 for Haplic Acrisols (BDa) and Haplic Ferralsols (BDf) respectively. The following predictive equations were derived for Haplic Ferralsols and Haplic Acrisols: BDf = 1.96-0.12 pH + 0.03 clay-0.05 SOM (r2 = 0.55, Se = 0.18, n =18) and BDa = 1.75 - 7.21 Arcsin (SOM) - 0.17 pH + 0.32 In (silt) (r2 = 0.92, Se= 0.05, n =10) (BD in g cm-3, and clay, SOM and silt in per cent). In general, the models suggested that soil properties such as SOM, pH, silt and clay content can provide a reliable alternative to determine BD.