The importance of replication of CO2 flux measurements in forest clearcuts

Stand-replacing disturbances, such as harvesting, have a major impact on the exchange of carbon dioxide (CO₂) between forested land and the atmosphere. The former forest CO₂ sinks become net CO₂ sources due to the continued respiratory losses of CO₂ and to the significantly reduced photosynthetic uptake of CO₂. The duration and magnitude of this carbon loss has now been quantified for many single sites representing ecosystems worldwide through Fluxnet. However, differences in the characteristics of harvested sites can influence vegetation recovery and respiratory fluxes of such stands and replicated observations are required to quantify the possible differences which can arise within an ecosystem following a stand-replacing disturbance. This study uses data from the well-studied Fluxnet-Canada Douglas-fir chronosequence on Vancouver Island, where the most mature site recently reached harvesting age and was commercially harvested. We compare the first two years following harvesting of eddy-covariance (EC) measurements of CO₂ exchange at this recently clearcut harvested site (HDF11) with those from a previously clearcut harvested Douglas-fir site (HDF00) in the chronosequence. The weather conditions during the years used in this analysis were similar and within the climate normals for the area. Half-hourly energy-balance closure was greater than 80% in both clearcuts. Our results for the first two years post-harvest show that both clearcuts were large annual carbon sources (i.e., net ecosystem productivity (NEP) < 0) with HDF11 being a much stronger source. NEP at HDF11 recovered from -1000 g C m^-2 yr^-1 in the first year to -700 g C m^-2 yr^-1 in the second year while HDF00 recovered from -620 g C m^-2 yr^-1 to -520 g C m^-2 yr^-1. Vegetation recovery was slower at HDF11 with a gross primary productivity (GPP) of 130 g C m^-2 yr^-1 in the first year and 385 g C m^-2 yr^-1 in the second year, while HDF00 had a GPP of 220 and 530 g C m^-2 yr^-1 in the respective years. Ecosystem respiration (Re) in the respective years was greater at HDF11 (1130 and 1085 g C m^-2 yr^-1) than at HDF00 (840 and 1050 g C m^-2 yr^-1). Chamber measurements at HDF11 indicated that soil respiration was 88% of Re. Re followed a logistic relationship with soil temperature at 2-cm depth at HDF11, whereas it followed an exponential relationship at HDF00. During the first year at HDF11, there was no relationship between GPP and incoming photosynthetically active radiation, but the recovery could be seen in the second year with a well-defined rectangular hyperbolic relationship. HDF00 showed well-defined rectangular hyperbolic relationships for both years due to earlier vegetation recovery. Our results show the danger of generalizing on the effects of harvesting on the forest carbon balance from measurements of CO₂ fluxes in a single clearcut.