Long-Term Impacts of Agriculture on Soil Carbon and Nitrogen in New England Forests

Abandonment and reforestation of agricultural lands has been a major influence on the landscape of eastern North America. Cultivation and soil amendments can dramatically alter soil nutrient pools and cycling, yet few studies have examined the long-term (>50 yr) influence of pasturing and cultivation on soil processes in the forests that develop after abandonment. Twelve forested sites at Harvard Forest in central New England were compared 90–120 yr after abandonment from agricultural use. We measured soil carbon (C), nitrogen (N), and phosphorus (P); light fraction C, N, and δ15N; microbial chloroform-N; net N mineralization and nitrification; nitrification potential; and culturable nitrifiers on sites with differing land-use history and vegetation. The sites had similar soil series and topography but were arrayed along a soil disturbance gradient from permanent woodlots (selective logging but no mineral soil disturbance) to formerly pastured sites (clearcut and grazed but no deep >10 cm soil disturbance) to formerly cultivated sites (cleared-with-plow horizon 15–20 cm thick). Mineral soil C (0–15 cm soil depth) was very similar among all sites, but the forest floor C was lower in the cultivated sites than in the woodlots of both stand types. Mineral soil in cultivated sites contained 800 kg N/ha and 300 kg P/ha more than woodlots, a relative increase of 39% for N and 52% for P. The cultivated soils had lower C:N and C:P ratios, largely driven by higher soil N and P. The light fraction appeared to be more sensitive to prior land use than the bulk soil organic matter. The C content and C:N ratio of light fraction were lower in cultivated soils, which suggests that input and/or turnover of organic matter pools of relatively recent origin remain altered for a century after abandonment. Similar δ15N for the light and heavy fraction organic matter pools in cultivated soils suggests that cultivation accelerates the mineralization of humus N, increasing the proportion of N available for plant uptake, resulting in a convergence of the light and heavy fractions. The N pool in the woodlot soils may have been subject to preferential losses of small amounts of 14N over a longer time period, resulting in a more pronounced divergence between the light fraction (reflecting recent plant inputs) and the mineral-associated heavy fraction (more recalcitrant). Nitrification was strongly influenced by land-use history, with highest rates in formerly cultivated sites. In contrast, soil net N mineralization and chloroform-N were more strongly influenced by present vegetation. Nitrifying bacteria were relatively abundant in all pastured and cultivated sites; however, only the cultivated hardwood sites, with lowest C:N ratios (16–18), had substantial net nitrification. Historical manure inputs may explain the more rapid soil net nitrification rates, by decreasing soil C:N ratios and thus reducing nitrate immobilization in the mineral soil of cultivated sites. Regionally, 65% of the land area was pastured, and a proportion of the nutrients obtained from grazing was transferred to the cultivated croplands, which comprise ≤15% of the land area. Pastures generally had intermediate nutrient ratios and N transformations but were often more similar to woodlots, which suggests that plowing and amendments, rather than forest clearance, have the greatest impact on soil organic matter and nutrients. The influence of land-use history on soil N and P and nitrification rates was more dramatic in hardwood sites, which indicates that characteristics of the recovering vegetation and/or changes in plant community composition associated with prior land use are important factors in the rate of recovery. Our findings lead to the surprising conclusion that 19th century agricultural practices decreased forest floor nutrient content and ratios, and increased nitrifier populations and net nitrate production for approximately a century after abandonment. Consideration of site history clearly deserves more attention in the design of field experiments, and in our understanding of patterns of element distributions and transformations in dynamic forested landscapes.