The direct and indirect effects of inter-annual meteorological variability on ecosystem carbon dioxide exchange at a temperate ombrotrophic bog

The link between inter-annual variability (IAV) in ecosystem CO2 exchange and meteorological conditions was evaluated using a decade (June 1998 to October 2008) of eddy covariance (EC) and meteorological data recorded at the Mer Bleue peatland, a large ombrotrophic bog located in Ontario, Canada (45.40°N lat., 75.50°W long.). Year-to-year variability in meteorology was particularly marked by prolonged growing periods, warmer soil temperature and greater water table drawdown during the summer of 1998–1999 and 2001–2003 periods. Ecosystem response to short- and long-term environmental alterations may be direct, where the response can be predicted as a linear function of the changing variables (constant slope) and/or indirect, in which case, a change in the rate of a particularly ecosystem response can be expected (variable slope). In this study, the potential direct and indirect effects of a decadal meteorological variability on IAV of ecosystem CO2 exchange were evaluated using a constant and variable slope multiple regression models. Indirect effect of IAV that results from a functional change in ecosystem response was identified and quantified as significant year-to-year changes in the slopes of the regression (i.e. as a difference between a constant and variable slope model) between weekly means of daytime net ecosystem productivity (NEPday) or nighttime ecosystem respiration (Rnight) and the following meteorological variables: photosynthetic photon flux density (PPFD), water table depth (WTD), air (TA) and soil temperature (TS), relative humidity (RH), vapor pressure deficit (VPD) and wind speed(WS). Results of the analysis illustrated that the direct effects of meteorological variation (seasonal and inter-annual) together explained ∼67% of the total variance in both NEPday and Rnight over the 10 year period, with seasonal (week-to-week) variations explaining most of the total variance in NEPday (48.2%) or Rnight (61.2%). Indirect effects (functional change) explained only 5.3% and 3.4% of the total variance in NEPday and Rnight over the 10 year monitoring period, respectively. The relationship between NEPday and TA, RH and that of Rnight and TS were identified as the sources of the slope changes. This level of indirect effect of meteorological variation at Mer Bleue bog is smaller than that found previously at some upland sites and may be a general characteristic of undisturbed peatland ecosystems. Hence, at present variations in CO2 exchange at this bog are largely tuned to the cyclic seasonal and inter-annual fluctuations in meteorology.

Research highlights▶ Effects of meteorology on inter-annual CO2 flux variability was studied by multiple linear regression. ▶ Functional change was detected as year-to-year variation in slope of CO2 flux vs. meteorology relationship. ▶ Direct effects of meteorology explained the bulk of variance in the bog's CO2 flux pattern. ▶ Effects of functional change on the bog's CO2 flux was small but significant over 10 year study period. ▶ Findings highlight importance of functional change in long-term climate change studies.