Soil respiration variability across a soil moisture and vegetation community gradient within a snow-scoured alpine meadow

Abstract

The alpine tundra landscape is a patchwork of co-mingled ecosystems that vary due to meso-topographical (<100 m) landscape position, shallow subsurface heterogeneity, and subsequent soil moisture availability. This results in hotspots of biological activity, variable carbon cycling over short horizontal distances, and confounds predictions of the alpine tundra response to forecasted environmental change. To advance our understanding of carbon cycling within snow-scoured alpine meadows, we characterized the spatio-temporal variability of soil respiration (R S) from 17 sites across a broadly representative soil moisture and vegetation gradient, within the footprint of ongoing eddy covariance measurements at Niwot Ridge, Colorado, USA. Chamber-based R S samples were collected on a weekly to bi-weekly basis over three complete growing seasons (2011–2013), and a soil moisture threshold was used to integrate the data into dry, mesic, and wet tundra categories. In every year, measured R S was greatest from mesic tundra, followed by wet and then dry tundra locations. Increasing soil moisture invoked a bidirectional R S response from areas of dry and mesic tundra (directly proportional) compared to wet tundra (inversely proportional), and the optimum R S conditions were between 0.30 and 0.45 m3 m−3 soil moisture, which mainly coincided with soil temperatures below 8 °C. We also developed simple models to predict R S from concurrent measurements of soil moisture and temperature, and from nighttime eddy covariance measurements. Both models were significant predictors of R S in all years and for all ecosystem types (where applicable), but the models did not adequately capture the intra-seasonal R S variability. The median cumulative growing season R S flux ranged from 138.6 g C m−2 in the driest year (2013) to 221.4 g C m−2 in the wettest year (2011), but the cumulative growing season fluxes varied by a factor of five between sites. Our results suggest that increased or more intense precipitation in the future has the potential to increase alpine tundra R S, although this effect will be buffered to some degree by compensatory responses from dry, mesic, and wet alpine tundra.