PS 48-80
Reconciling biometric and eddy flux estimates of forest carbon dynamics during disturbance and recovery
Accurate estimates of carbon (C) pools and fluxes during disturbance and recovery cycles are central to our understanding of forest C dynamics. We compared net ecosystem production (NEP) and C accumulation measured in forest census plots over a 10 year period to net exchange of CO2 (NEE) measured from above-canopy towers using eddy covariance before and after defoliation by Gypsy moth (Lymantria dispar) and prescribed fire in oak- and pine-dominated stands in the New Jersey Pinelands, USA. Forest census plots encompassed measurements at two different temporal and spatial scales; five census plots located within 100 m of the flux tower were measured for annual tree and understory biomass increment, monthly litterfall, and litter decomposition to calculate NEP, and 12 or 16 USFS Forest Inventory and Analysis (FIA) type plots in a 1 km2 grid centered on each flux tower were measured every two years to calculate net C accumulation.
Results/Conclusions
Before disturbance, annual NEP in tower plots averaged 84 ± 15% and 73 ± 22% of NEE, but C accumulation in FIA-type plots averaged only 70 ± 19% and 15 ± 31% of NEE at the oak and pine stands, respectively. Defoliation by Gypsy moth at the oak and pine stands resulted in differences of 338 and 183 g C m-2 between annual NEP and NEE, while annual NEP and NEE differed by only 38 g C m-2 when the pine stand was burned in a prescribed fire. Carbon accumulation in the FIA-type plots and NEE differed by 214 and 141 g C m-2 during disturbance at the oak and pine stands, respectively. Over longer integration times, biometric and cumulative NEE estimates converged; differences between NEP and NEE were 121 and 106 g C m-2, and differences between C accumulation and NEE were 116 and 116 g C m-2 at the oak and pine stands, respectively. Differences between approaches at longer integration times were less than average annual NEE during pre-disturbance periods. Our results highlight the importance of long-term measurements for quantifying forest C dynamics during disturbance and recovery cycles.