Soil organic carbon changes in landscape units of Belgium between 1960 and 2000 with reference to 1990

The present study quantifies changes in soil organic carbon (SOC) stocks in Belgium between 1960, 1990 and 2000 for 289 spatially explicit land units with unique soil association and land‐use type, termed landscape units (LSU). The SOC stocks are derived from multiple nonstandardized sets of field measurements up to a depth of 30 cm. Approximately half of the LSU show an increase in SOC between 1960 and 2000. The significant increases occur mainly in soils of grassland LSU in northern Belgium. Significant decreases are observed on loamy cropland soils. Although the largest SOC gains are observed for LSU under forest (22 t C ha^?1 for coniferous and 29 t C ha^?1 for broadleaf and mixed forest in the upper 30 cm of soil), significant changes are rare because of large variability. Because the number of available measurements is very high for agricultural land, most significant changes occur under cropland and grassland, but the corresponding average SOC change is smaller than for forests (9 t C ha^?1 increase for grassland and 1 t C ha^?1 decrease for cropland). The 1990 data for agricultural LSU show that the SOC changes between 1960 and 2000 are not linear. Most agricultural LSU show a higher SOC stock in 1990 than in 2000, especially in northern Belgium. The observed temporal and spatial patterns can be explained by a change in manure application intensity. SOC stock changes caused by land‐use change are estimated. The SOC change over time is derived from observed differences between SOC stocks in space. Because SOC stocks are continuously influenced by a number of external factors, mainly land‐use history and current land management and climate, this approach gives only an approximate estimate whose validity is limited to these conditions.