Linking interannual variability of carbon exchange in an old-growth forest to seasonal and interannual variations in water availability

Carbon fluxes have been estimated at the top (70 m) of a 400-500 year old coniferous forest canopy in southern Washington using eddy covariance techniques since July 1998. Eddy covariance measurements show overall net uptake of carbon by the old-growth forest ecosystem, though the data also indicate high interannual variability of carbon exchange. Three years (1999, 2000, and 2002) have shown significant carbon uptake by the forest sequestering 204 g Cm^-2year^-1, 74 g Cm^-2year^-1, and 80 g Cm^-2year^-1, respectively. 2003 was at equilibrium with carbon exchange and 2001 was a significant source year, with a loss of 49 g Cm^-2year^-1, for the old-growth forest canopy.

Here we analyze the interannual variability of carbon exchange for this forest by closely examining interannual and seasonal variations in precipitation and soil moisture. Our data indicate that carbon exchange for this old-growth forest is closely linked to rainy-season precipitation and any subsequent periods of water shortage during dry-season summer months, when water lost through evaporation and transpiration is greater than precipitation. Strong seasonal changes in water availability partially explain seasonal to interannual variations in carbon exchange. Gross ecosystem productivity (GEP) was moderately correlated with monthly precipitation (R² = 0.5) and with evapotranspiration (R² = 0.6). Net ecosystem exchange (NEE) was moderately correlated with annual precipitation (R² = 0.4), but the correlation increased significantly with the addition of a six month lag for precipitation (R² = 0.8 – 0.9). The correlations between daily NEE and soil moisture are less certain, though soil moisture data during the winter months of 1999 were found to explain 10 to 15 percent of the variability in daily NEE.

Links between carbon exchange and precipitation, and to a lesser degree soil moisture, suggest that water availability may be an important factor determining whether or not the forest becomes an annual carbon sink, source or is at equilibrium. Seasonal to interannual variability in precipitation appears to influence the timing of the transition from photosynthesis dominance to respiration dominance within the canopy. The timing of this seasonal transition in the forest canopy ranged from DOY 145 (2001) to DOY 200 (1999). A very dry autumn and winter in 2000-2001 (lower 5th percentile, 1949-2003) accelerated the seasonal transition in 2001, while much wetter conditions during autumn and winter of 1998-1999 (upper 95th percentile, 1949-2003) delayed the seasonal transition in spring 1999.