A method for deriving net primary productivity and component respiratory fluxes from tower-based eddy covariance data: A case study using a 17-year data record from a Douglas-fir chronosequence

Conventional gap-filling procedures for eddy covariance (EC) data are limited to calculating ecosystem respiration (R_E) and gross ecosystem productivity (P_G) as well as missing values of net ecosystem productivity (F_NEP). We develop additional postprocessing steps that estimate net primary productivity (P_N), autotrophic (R_a), and heterotrophic respiration (R_h). This is based on conservation of mass of carbon (C), Monte Carlo (MC) simulation, and three ratios: C use efficiency (CUE, P_N to P_G), R_a to R_E, and F_NEP to R_E. This procedure, along with the estimation of F_NEP, R_E, and P_G, was applied to a Douglas-fir dominated chronosequence on Vancouver Island, British Columbia, Canada. The EC data set consists of 17 site years from three sites: initiation (HDF00), pole/sapling (HDF88), and near mature (DF49), with stand ages from 1 to 56 years. Analysis focuses on annual C flux totals and C balance ratios as a function of stand age, assuming a rotation age of 56 years. All six C balance terms generally increased with stand age. Average annual P_N by stand was 213, 750, and 1261g Cm^-2 yr^-1 for HDF00, HDF88, and DF49, respectively. The canopy compensation point, the year when the chronosequence switched from a source to a sink of C, occurred at stand age ca. 20 years. HDF00 and HDF88 were strong and moderate sources (F_NEP=-581 and -138g Cm^-2yr^-1), respectively, while DF49 was a moderate sink (F_NEP=294g Cm^-2yr^-1) for C. Differences between sites were greater than interannual variation (IAV) within sites and highlighted the importance of age-related effects in C cycling. The validity of the approach is discussed using a sensitivity analysis, a comparison with growth and yield estimates from the same chronosequence, and an intercomparison with other chronosequences.