Landscape structure control on soil CO2 efflux variability in complex terrain: Scaling from point observations to watershed scale fluxes

We investigated the spatial and temporal variability of soil CO2 efflux across 62 sites of a 393-ha complex watershed of the northern Rocky Mountains. Growing season (83 day) cumulative soil CO2 efflux varied from similar to 300 to similar to 2000 g CO2 m(-2), depending upon landscape position, with a median of 879.8 g CO2 m(-2). Our findings revealed that highest soil CO2 efflux rates were observed in areas with persistently high soil moisture (riparian meadows), whereas lower soil CO2 efflux rates were observed on forested uplands (98% of watershed area). Furthermore, upslope accumulated area (UAA), a surrogate measure of the lateral redistribution of soil water, was positively correlated with seasonal soil CO2 efflux at all upland sites, increasing in explanatory power when sites were separated by the major aspects of the watershed (SE/NW). We used the UAA-soil CO2 efflux relationship to upscale measured CO2 efflux to the entire watershed and found watershed-scale soil CO2 efflux of 799.45 +/- 151.1 g CO2 m(-2) over 83 days. These estimates compared well with independent eddy covariance estimates of nighttime ecosystem respiration measured over the forest. We applied this empirical model to three synthetic watersheds with progressively reduced complexity and found that seasonal estimates of soil CO2 efflux increased by 50, 58, and 98%, demonstrating the importance of landscape structure in controlling CO2 efflux magnitude. Our study represents an empirical quantification of seasonal watershed-scale soil CO2 efflux and demonstrates that UAA (i.e., landscape position) and drainage patterns are important controls on the spatial organization of large-scale (similar to km(2)) soil CO2 efflux, particularly in semiarid, subalpine ecosystems.