Estimating stone and boulder content in forest soils - Evaluating the potential of surface penetration methods

Given the amount and variation of stones and boulders in forest soils, it is remarkable that little attention has been paid to this property in the analysis and modelling of soil processes. This study aims (a) to investigate the stoniness (relative volume of stones and boulders) of representative forest soils in central Sweden, (b) to evaluate the possibilities to predict soil stoniness with non-destructive surface penetration methods, particularly the rod penetration method suggested by Viro in 1952. The content of stones and boulders down to 0.4 m depth in forest till were quantified in 200 systematically distributed pits. The mean content of stones and boulders was 0.50 with a range of 0.14-0.78. Penetration depths were useful as predictors only for cases of stoniness below 0.50 where 75% of the Variation could be explained. The general evaluation of surface penetration methods revealed that large errors in very stony soils cannot be avoided. The major source of error is inherent in the method and not caused by sampling. This conclusion may lower expectations that ground-penetrating radar systems will become useful for predicting stoniness.