Annual Variations in Ecosystem Respiration during an El Niño/La Niña Cycle in a Coniferous West Coast Forest

Eddy-correlation CO₂ fluxes measured above a 51-year-old Douglas-fir stand were used to infer annual variations in ecosystem respiration (ER). The stand is located in the Coastal Western Hemlock biogeoclimatic zone that is characterized by yearly precipitation values around 1000 mm and a dry period in summer. The three years of measurements analyzed cover an El Niño/La Niña cycle (1998/1999) and an 'average' year (2000). To compute annual ER, first, a straight line was fitted to the relationship between soil temperature and the log-transformed net ecosystem exchange (NEE) at night. Annual totals were then computed by using the fitted relationships to calculate half-hourly ER throughout the year and summing these values. The effect of applying standard quality control procedures suggested in the literature (stationarity, friction velocity thresholds, flux thresholds) on the estimated annual ER was investigated. Confidence limits of the annual ER were calculated from the confidence limits of the fitted relationships. While this uncertainty does not represent the actual measurement error, it serves as a gauge of the significance of removing values from the data set. Only the removal of cases of low friction velocity proved significant, i.e. the annual estimate converged when the threshold was increased. The temperature response of ER was strongest during the comparatively wet and cold La Niña year (1999) and similar during the El Niño and the average years. Due to the generally lower temperatures in 1999, however, annual ER (1700 g C m^-2 y^-1) was slightly less than in 1998 and 2000 (1870 and 1800 g C m^-2 y^-1, respectively). NEE was -330, -390, and -400 g C m^-2 y^-1 in 1998, 1999, and 2000, respectively. Apparently, the slightly higher ER during the El Niño year (1998) led to a corresponding reduction of the carbon uptake. On the other hand, the reduction of ER in the La Niña year (1999) was accompanied by a reduction in gross ecosystem photosynthesis, leaving NEE virtually unchanged compared to 2000, the 'average' year.