全球变化知识资源中心

Soil organic matter plays a crucial role in soil health and represents one of the key functions for determining soil suitability for crop production. Recently, intensive agricultural production and climatic changes have led to a decline in organic matter level in soils. This paper is to provide the most accurate spatial predictor using different interpolation methods in order to evaluate in detail the status of organic matter in agricultural soils in the Osijek-Baranja County, Croatia. We applied three different interpolation methods, including inverse distance weighting, ordinary kriging and empirical Bayesian kriging. A total number of 9099 soil samples from 0-30 cm layer were compiled and analysed in the laboratory. The average value of soil organic matter in the study area was 2.66% with moderate variability (CV = 30.62%). The best fit variogram model is exponential in the direction of 20 and its spatial variability indicates that soil organic matter varies widely under pedogenetic and soil management practices. Empirical Bayesian kriging method was the most precise (RAISE = 0.457), followed by ordinary kriging (RAISE = 0.466) and inverse distance weighting (RAISE = 0.476). The investigated area shows a heterogeneous spatial pattern of soil organic matter content, with levels below 3% found mostly in western and south-western parts of county.