Interaction of land use history, earthworms, soil chemistry and tree species on soil carbon distribution in managed forests in Vermont, USA

Abstract

Forest soils store a globally important pool of carbon (C) and reforestation has the potential to increase this pool. Past land use, tree species composition, and current management are known to affect carbon storage rates and amounts. Another important factor, especially in the vertical distribution of soil C, is the presence or absence of earthworms. We investigated these interactions in eighteen managed forest sites in Vermont that have varied land use history (cultivation, pasture, woodlot, continuous woodland) and a range of time since reforestation from 60 to 100+ years. Soils were sampled to about 1 m, or shallower if bedrock was reached, and both C and exchangeable cation pools quantified. Earthworms were enumerated to a depth of 20 cm and identified by species. Basal area by tree species was also measured and aboveground C calculated. Detailed land use history was documented using archival sources. Forest type was primarily northern hardwood, with a few lowland spruce-fir sites included. Eleven different species of earthworms were found, representing all ecotypes, and numbers ranged from 0 (at 8 of the 18 sites) to 319 individuals/m2. Carbon pools in the forest floor (1.5–30.1 Mg/ha) were greater with higher site elevation (range of 154–651 masl) and also greater with lower numbers of earthworms. Besides elevation, the best predictor of earthworm presence or absence, and overall numbers, was the soil exchangeable calcium (Ca) pool in the Oa and A horizons (range of 9.4–1288 cmolc/m2). Because soil exchangeable Ca and aluminum (Al) pools were inversely related, higher exchangeable Al was associated with lower earthworm numbers. In the mineral soil, the C pool (82.1–210.8 Mg/ha) was best explained by the thickness of the B horizons with the exchangeable Al pool as a positive, secondary factor. The thickness of the A horizon (0–16.3 cm), and its contribution to the entire profile C pool (0–54.7 Mg/ha), decreased relative to past land use in the order: cultivated < pasture < woodlot. No earthworm metric was strongly related to the C pool in either the mineral soil or the full soil profile. Dramatic differences in the vertical distribution of carbon were related to past land use and earthworm presence, which in turn was negatively correlated with elevation and the presence of conifers. While it is clear that earthworms and prior land-use have a long-lasting effect on the forest floor C pool, it is less clear whether or not there is an effect on full profile C sequestration over time.