Comparing soil inventory with modelling: Carbon balance in central European forest soils varies among forest types

Forest soils represent a large carbon pool and already small changes in this pool may have an important effect on the global carbon cycle. To predict the future development of the soil organic carbon (SOC) pool, well-validated models are needed. We applied the litter and soil carbon model Yasso15 to 1838 plots of the German national forest soil inventory (NFSI) for the period between 1985 and 2014 to enables a direct comparison to the NFSI measurements. In addition, to provide data for the German Greenhouse Gas Inventory, we simulated the development of SOC with Yasso15 applying a climate projection based on the RCP8.5 scenario. The initial model-calculated SOC stocks were adjusted to the measured ones in the NFSI. On average, there were no significant differences between the simulated SOC changes (0.25 ± 0.10 Mg C ha^-1 a^-1) and the NFSI data (0.39 ± 0.11 Mg C ha^-1 a^-1). Comparing regional soil-unit-specific aggregates of the SOC changes, the correlation between both methods was significant (r² = 0.49) although the NFSI values had a wider range and more negative values. In the majority of forest types, representing 75% of plots, both methods produced similar estimates of the SOC balance. Opposite trends were found in mountainous coniferous forests on acidic soils. These soils had lost carbon according to the NFSI (-0.89 ± 0.30 Mg C ha^-1 a^-1) whereas they had gained it according to Yasso15 (0.21 ± 0.10 Mg C ha^-1 a^-1). In oligotrophic pine forests, the NFSI indicated high SOC gains (1.36 ± 0.17 Mg C ha^-1 a^-1) and Yasso15 much smaller (0.29 ± 0.10 Mg C ha^-1 a^-1). According to our results, German forest soils are a large carbon sink. The application of the Yasso15 model supports the results of the NFSI. The sink strength differs between forest types possibly because of differences in organic matter stabilisation.