Open access publications in the SLU publication database

Abstract

We used two datasets of 14C analyses of archived soil samples to study carbon turnover in O horizons from spruce dominated old-growth stands on well-drained podsols in Scandinavia. The main data set was obtained from archived samples from the National Forest Soil Inventory in Sweden and represents a climatic gradient in temperature. Composite samples from 1966, 1972, 1983 and 2000 from four different regions in a latitude gradient ranging from 57 to 67oN were analysed for 14C content. Along this gradient the C stock in the O horizon ranges from 2.1 kg m-2 in the north to 3.7 kg m-2 in the southwest. The other data set contains 14C analyses from 1986, 1987, 1991, 1996 and 2004 from the O horizons in Birkenes, Norway. Mean residence times (MRT) were calculated using a two compartment model, with a litter decomposition compartment using mass loss data from the literature for the three first years of decomposition and a humus decomposition compartment with a fitted constant turnover rate. We hypothesized that the climatic gradient would result in different C turnover in different parts of the country between northern and southern Sweden. The use of archived soil samples was very valuable for constraining the MRT calculations, which showed that there were differences between the regions. Longest MRT was found in the northernmost region (41 years), with decreasing residence times through the middle (36 years) and central Sweden (28 years), then again increasing in the southwestern region (40 years). The size of the soil organic carbon (SOC) pool in the O horizon was mainly related to differences in litter input and to a lesser degree to MRT. Because N deposition leads both to larger litter input and to longer MRT, we suggest that N deposition contributes significantly to the latitudinal SOC gradient in Scandinavia, with approximately twice as much SOC in the O horizon in the south compared to the north. The data from Birkenes was in good agreement with the Swedish dataset with MRT estimated to 34 years.