Spatial variability of soil phosphorus in the Fribourg canton, Switzerland

Publication date: April 2014
Source:Geoderma, Volumes 217–218
Author(s): Aurélien Roger , Zamir Libohova , Nicolas Rossier , Stéphane Joost , Alexandra Maltas , Emmanuel Frossard , Sokrat Sinaj
Phosphorus (P) is the second essential nutrient for plant growth but can become an ecological and economical concern in case of over-fertilization. Soil P dynamic is influenced by many parameters like soil physical–chemical properties and farming practices. A better understanding of the factors controlling its distribution is required to achieve best management of P in cropping systems. In Switzerland, the FRIBO network was launched in 1987 and consists of 250 sites covering a wide diversity of soils (Cambisols, Gleysols, Rendzinas, Lithosols, Luvisols, Fluvisols) and three different land uses (cropland, grassland and mountain pasture) across the Fribourg canton. A spatial investigation of the different P forms (total, organic and available) for the FRIBO network led to the following main conclusions: (i) The P status in agricultural soils was significantly different among the three land uses encountered, with the highest mean values of available P found in croplands, from 2.12 (CO2 saturated water extraction) to 81.3mg.kg⁻¹ (acetate ammonium+EDTA extraction); whereas total P was more abundant in permanent grasslands (1186mg.kg⁻¹), followed by mountain pastures (1039mg.kg⁻¹) and croplands (935mg.kg⁻¹). This full characterization of the soil P status provides important data on P distribution related to soil properties and land use. (ii) Environmental variables such as altitude, slope, wetness index or plan curvature, derived from the digital elevation model (DEM) only explained a small part of the spatial variation of the different P forms (20 to 25%). Thus, the geostatistical analyses revealed that land use plays a significant role in soil P distribution. Improved predictions of the spatial distribution of P-related forms at landscape scales are needed and would require additional data points and variables such as parent material, soil types and terrain attributes.