Annual ecosystem respiration (Reco) and gross ecosystem exchange (GEE) are independent processes, yet they are highly correlated, suggesting that Reco is a fixed fraction of GEE, i.e. a flat-tax. Reco and GEE are derived from observations of net ecosystem exchange (NEE), so the estimation of GEE may include uncertainties due to autocorrelation. Soil CO2 efflux (Fs) is a large contributor to Reco (30-80%) and may serve as a proxy to estimate the biotic and abiotic controls of Reco. Longleaf pine savannas are model ecosystems for studying this flat-tax hypothesis, as they are found on a large edaphic and hydrologic range of conditions. Additionally, they are subject to the periodic disturbance of fire. We measured Fs, air temperature (Tair) and soil water content (SWC) at three sites (mesic, xeric, and intermediate) in a longleaf savanna in southwestern Georgia, USA, from 2009 to 2014 (which included three prescribed burns). We hypothesized that Fs would be driven by NEE and controlled by Tair and SWC. In addition, Fs should increase with increasing physiological activity. We also hypothesized that Fs would decrease during fire, due to reduced respiring material on the forest floor, and substantially increase post-burn, owing to plant regrowth in the understory.
Results/Conclusions
Results indicated that Fs was positively correlated with Tair and NEE, but negatively correlated with SWC. We did not find significant differences in Fs at short timescales. At annual timescales, Fs was significantly higher at the mesic site, in contrast to the xeric site. Fs at the mesic site increased proportionally with GEE, resulting in similar fractions at both sites, in support of a flat-tax. Differences in Fs were largely driven by changes in SWC and Tair, indicating abiotic factors such as soil saturation inhibiting microbial processes, and low temperatures reducing enzymatic activities. Fs rates changed with variation in NEEday driven by SWC and Tair, suggesting that the rate of ecosystem carbon uptake plays an important role in determining Fs. Fs significantly increased 28 days following fire, which may be attributed to nutrient pulses encouraging plant regrowth in the understory. Our results indicate that annually, Fs was a consistent large proportion of Reco (~50%). Fs observations could therefore help to inform the estimation of Reco and hence increase the precision of flux partitioning of eddy covariance data. This research improves our understanding of the flat-tax of Fs and its environmental drivers, enhancing our understanding of the carbon cycle.