Eight years of carbon dioxide exchange above a mixed forest at Borden, Ontario

Abstract

This paper summarizes results from 8 years (1996–2003) of eddy covariance-based ecosystem CO 2 exchange measurements at the Borden Forest Research Station (44°19′N, 79°56′W). The site represents a mid-latitude, 100-year-old, mixed deciduous and coniferous forest dominated by red maple, aspen and white pine. The years 1996 and 1997 were relatively cold, had a late spring and received below average photosynthetic photon flux density (PPFD). This contrasts with an early spring, warmer soil and air temperatures during 1998–1999, and with distinctly wet year of 2000 and dry years of 2001–2003. The combination of early spring, warmer air and soil temperature and relatively high level of PPFD was associated with higher net ecosystem productivity (NEP) that peaked during 1999. Photosynthetic capacity was reduced and NEP showed a mid-growing season depression during the dry years of 2001–2003. Annual average ecosystem respiration (R) determined from a light response model was ∼30% less than R derived from a logistic respiration equation, relating night time CO 2 flux and soil temperature. However these independently determined R values were well correlated indicating that the site is unaffected by fetch and spatial heterogeneity problems. Based on the combined 8 years of growing season daytime data, an air temperature of 20–25 °C and a vapor pressure deficit (VPD) of 1.3 kPa were found to be the optimal conditions for CO 2 uptake by the canopy. Over the 1996–2003 period, the forest sequestered carbon at an average rate of 140 ± 111 gC m −2 y −1. The corresponding gross ecosystem photosynthesis (GEP) and R over this period were 1116 ± 93 gC m −2 y −1 and 976 ± 68 gC m −2 y −1, respectively. The annual carbon sequestration ranged from 19 gC m −2 in 1996 to 281 gC m −2 in 1999. However, these estimates were sensitive to frictional velocity threshold ( u * TH ) used for screening data associated with poor turbulent mixing at night. Increasing u * TH from 0.2 m s −1 (based on the inflection point in the nighttime CO 2 flux vs. u * relationship) to 0.35 m s −1 (determined using a u * TH selection algorithm based on change-point detection) modified the 8-year mean NEP estimate from 140 ± 111 gC m −2 y −1 to 65 ± 120 gC m −2 y −1. Both approaches show that the Borden forest was a low to moderate sink of carbon over the 8-year period.