Forest management in temperate forests strongly affects carbon stocks in the organic layer but not in the mineral soil

Abstract

Litterfall is an important pathway by which organic carbon in forest ecosystems is transferred from plants to soil. The amount and chemical composition of litterfall, the rate of litter decomposition and the extent of litter incorporation into the soil are relevant for the formation of the organic layer and soil organic matter. However, the interaction between forest management, litterfall, organic layer formation and mineral soil carbon stocks has rarely been considered together. Since 2015, we have been monitoring the above-ground litter input with a total of 750 litter traps in 150 forests in 3 German regions (Schorfheide-Chorin, Hainich-Dün and Swabian Alb). In 2017, 2021 and 2023, the organic layer and the topsoil (0-10 cm) were sampled at 14 sampling points per plot. The silvicultural management intensity index (SMI) introduced by Schall and Ammer (2013) was used to quantify forest management intensity at all plots. Piecewise structural equation modeling (SEM) showed that C stocks in the organic layer were positively related to the proportion of conifers and negatively related to timber harvesting. There was also a negative relationship between C stocks in the organic layer and mineral soil pH. These relationships were mediated by litter C input, litter CN ratio and earthworm biomass. While C stocks in the organic soil layer were positively related to litter C input and litter CN ratio, we found a negative relationship with earthworm biomass. In our model, these effects explained 46 % of the variance in carbon stocks in the organic layer. In addition, the study region explained a further 18% of the variance. A comparison of the relative importance of the different factors influencing carbon storage in the organic layer showed that tree species and litter CN ratio were by far the most strongly associated with carbon stocks in the soil organic layer. The effects of timber harvest and soil pH, mediated by litter C mass and earthworm biomass, were significant but comparatively weak. A direct influence of pH on the carbon stock of the organic soil layer could not be detected. In contrast to the organic layer, we found no direct or indirect influence of forest management intensity on mineral soil OC stocks in 0-10 cm. While litter C mass and litter CN ratio were not related to OC stocks in the mineral soil, we found a very strong positive correlation between clay content, soil pH and OC stocks in the mineral soil (p<0.001). We also found a weak correlation between dithionite extractable iron and mineral soil OC stocks (p<0.05). All three factors together explained 67% of the variance in mineral soil OC stocks (0-10 cm), and the study regions explained a further 11% of the variance. Our results show that forest management and especially tree species selection are strongly reflected in the organic layer. However, the regional gradient of forest management was not large enough to be visible in the mineral soil, where carbon stocks are mainly determined by parent material and associated clay and iron oxide contents.