Collaboratrices et collaborateurs

Total organic carbon (TOC) stocks in soil are of central importance for agricultural productivity, climate regulation, and soil ecological functions. They are strongly influenced by environmental conditions and agricultural management practices, yet changes in TOC stocks often only become detectable over decades. Long-term monitoring programs such as the Swiss National Soil Monitoring Network (NABO) enable reliable detection of such changes. While no changes in topsoil (0–20 cm) TOC have been observed at more than 100 NABO sites over the past 30 years, data on subsoil dynamics remain scarce. This gap reflects the later introduction of systematic subsoil sampling and inconsistencies in sampling procedures, depth intervals, TOC analysis, and bulk density determination compared with the first NABO monitoring period (1985–1989). The aims of this study were therefore to (i) harmonize data from the first (1985–1989) and seventh (2015–2019) monitoring period for profile-wide TOC stock calculations, (ii) quantify land use-specific changes in TOC stocks, (iii) assess the influence of pedoclimatic factors on TOC stocks and their temporal dynamics, and (iv) identify data limitations and provide recommendations for future monitoring. TOC stocks are calculated from measured TOC content and bulk density. Nearly complete datasets on TOC content and bulk density were available for 58 cropland, grassland, orchard/vineyard, and forest sites for both monitoring periods. To ensure comparability, four harmonization steps were implemented: (1) TOC contents determined with different analytical methods were standardized using a generalized linear model; (2) missing bulk density values were estimated with a newly developed pedotransfer function; (3) the depth distributions of the soil parameters were standardized to 60 cm using mass-preserving spline interpolation; and (4) a regression model was developed to correct method-specific differences in bulk density between the monitoring periods. Changes in the harmonized TOC stocks were analyzed in relation to land use and site conditions using linear models. Finally, the minimal detectable difference, pedoclimatic effects, and uncertainties from the harmonization steps were quantified. Average TOC stocks across the first and seventh monitoring period were 62 and 55 t TOC ha⁻¹, respectively, in the topsoil (0–20 cm), and 57 and 50 t TOC ha⁻¹, respectively, in the subsoil (20–60 cm). Topsoil TOC stocks in croplands declined significantly over time, whereas changes in subsoil and in other land uses were not statistically significant. Depending on land use and soil depth, changes would have needed to be up to 18 times larger to be statistically significant. The topsoil in grassland had 1.5 times higher TOC stocks than cropland—likely due to lower soil disturbance and greater root biomass—while forests and orchards/vineyards showed intermediate values. No significant differences in subsoil TOC stocks were found between land uses. Topsoil TOC stocks in cropland, grassland, and forest sites were positively correlated with clay content, pointing at greater stabilization of organic matter in fine-textured soils. In the subsoil, TOC stocks correlated positively with soil pH, elevation, or slope, depending on land use. Changes in TOC stocks between monitoring periods were primarily negatively correlated with initial TOC stocks, suggesting gains in soils with initially low carbon and losses in soils with initially high carbon. Uncertainties in data harmonization were 7–9% for gap-filling and method conversions, but only ~1% for depth interpolation. For future monitoring of TOC stocks, differences in subsoil depth intervals are considered uncritical when spline-based interpolation is applied. However, consistent analytical methods for TOC content and bulk density, as well as quantification of small-scale variability and stone content, are essential for robust, profile-wide estimates of TOC stocks. To better evaluate management effects on cropland TOC and support sustainable soil management in Switzerland, future monitoring should systematically include management data alongside pedoclimatic factors.