Forest soils store a substantial amount of carbon, often more than the forest vegetation does. Estimates of the amount of soil carbon, and in particular estimates of changes in these amounts are still inaccurate. Measuring soil carbon is laborious, and measurements taken at a few statistically unrepresentative sites are difficult to scale to larger areas.We combined a simple dynamic model of soil carbon with litter production estimated on the basis of stand parameters, models of tree allometry and biomass turnover rates of different biomass components. This integrated method was used to simulate soil carbon as forest stands develop. The results were compared with measurements of soil carbon from 64 forest sites in southern Finland.Measured carbon stocks in the organic soil layer increased by an average of 4.7 AE 1.4 g m À2 a À1 with increasing stand age and no significant changes were measured in the amount of carbon in mineral soil. Our integrated method indicated that soil carbon stocks declined to a minimum 20 years after clear-cutting and the subsequent increase in the soil carbon stock (F/H À 1 m) was 5.8 AE 1.0 g m À2 a À1 averaged over the period to next harvesting ( $ 125 years). Simulated soil carbon accumulation slowed down considerably in stands older than 50 years. The carbon stock measured (F/H À 1 m) for the study area averaged 6.8 AE 2.5 kg m À2 . The simulated carbon stock in soil was 7.0 AE 0.6 kg m À2 on average.These tests of the validity of the integrated model suggest that this method is suitable for estimating the amount of carbon in soil and its changes on regional scales.

No abstract

No abstract

No abstract

No abstract