Impacts of an anomalously warm year on soil CO2 efflux in experimentally manipulated tallgrass prairie ecosystems

摘要

Modeling analyses suggest that an increase in growth rate of atmospheric CO2 concentrations during an anomalously warm year may be caused by a decrease in net ecosystem production (NEP) in response to increased heterotrophic respiration (R-h). To test this hypothesis, 12 intact soil monoliths were excavated from a tallgrass prairie site near Purcell, Oklahoma, USA and divided among four large dynamic flux chambers (Ecologically Controlled Enclosed Lysimeter Laboratories (EcoCELLs)). During the first year, all four EcoCELLs were subjected to Oklahoma air temperatures. During the second year, air temperature in two EcoCELLs was increased by 4 degrees C throughout the year to simulate anomalously warm conditions. This paper reports on the effect of warming on soil CO2 efflux, representing the sum of autotrophic respiration (R-a) and R-h.During the pretreatment year, weekly average soil CO2 efflux was similar in all EcoCELLs. During the late spring, summer and early fall of the treatment year, however, soil CO2 efflux was significantly lower in the warmed EcoCELLs. In general, soil CO2 efflux was correlated with soil temperature and to a lesser extent with moisture. A combined temperature and moisture regression explained 64% of the observed variation in soil CO2 efflux. Soil CO2 efflux correlated well with a net primary production (NPP) weighted greenness index derived from digital photographs. Although separate relationships for control and warmed EcoCELLs showed better correlations, one single relationship explained close to 70% of the variation in soil CO2 efflux across treatments and years. A strong correlation between soil CO2 efflux and canopy development and the lack of initial response to warming indicate that soil CO2 efflux is dominated by R-a. This study showed that a decrease in soil CO2 efflux in response to a warm year was most likely dominated by a decrease in R-a instead of an increase in R-h.