Recovery of ecosystem carbon pools 35 years after whole-tree and stem-only clearcutting a red spruce – balsam fir forest in north-central Maine, USA

Long-term field studies are critical to obtain credible knowledge on the effects of forest management. We tested the effect of whole-tree (WTH), stem-only (SOH) and no harvest (REF) on carbon (C) pools of a spruce-fir forest in Maine, USA in the long-term Weymouth Point Study Area, consisting of two adjacent watersheds with similar site conditions, species composition, and disturbance history. Harvest was conducted on one watershed in 1981. The other was left as an uncut late-successional reference. Fieldwork took place on various occasions from 1979 to 2016, on 22–31 plots. Thirty-five years after harvest, the average living biomass C-pool was significantly smaller for WTH than SOH on poorly drained (PD) soils, with no significant difference on somewhat poorly to moderately well-drained (SP-MW) soils, and no significant differences for deadwood, forest floor and mineral soil C. When including pre-harvest above-stump living biomass and forest floor C as co-variates for plots with available estimates, the deadwood C-pool was significantly smaller after WTH compared to SOH, as was forest floor C on PD but not on SP-MW soils. In 2016, living biomass, deadwood, and forest floor C-pools on the harvested watershed recovered to 62–67 %, 38–39 %, and 55–64 % of the corresponding C-pools for REF, while the mineral soil C-pool was 115–124 % of that of REF. Interpretation of C recovery is sensitive to the reference. Harvest left a long-term legacy on the landscape, with living biomass C on a trajectory towards fast recovery, while dead organic matter C pools recover at a slower pace.