干旱区生态环境知识资源中心

Understanding the temporal dynamics and influencing factors of ecosystem respiration ( Reco) in semi-arid shrublands is critical for predicting how their carbon balance may respond to climate change. Using the eddycovariance technique, we quantified the net ecosystem CO2 exchange (NEE) in a semi-arid shrubland of northern China from July 2011 to December 2016, and partitioned NEE into Reco and gross primary productivity (GPP). Annual Reco varied from 300 g C m(-2) yr(-1) in 2014 to 426 g C m(-2) yr(-1) in 2012, and GPP ranged from 277 g C m(-2) yr(-1) in 2014 to 503 g Cm-2 yr(-1) in 2012. The relationship between half-hourly nighttime Reco and air temperature (Ta) was well-described by the Lloyd & Taylor model. Indicators of the seasonal temperature sensitivity (E-0 and Q(10)) of Reco increased with both the annual integral and seasonal amplitude of GPP. However, when averaged into 1 degrees C T-a bins, nighttime Reco increased with T-a up to an optimal temperature of similar to 20 degrees C, abovewhich it decreased with increasing Ta. Periods of low soil moisture in spring and summer markedly depressed Reco, contributing to its seasonal and interannual variations. In addition, low soil moisture had little effect on nighttime Reco when T-a was below 15 degrees C, but substantially reduced nighttime Reco when T-a was above 15 degrees C. Ecosystem respiration increased linearly with GPP at both seasonal and interannual scales, with the slope being 0.50 and 0.55, respectively. Our results have important implications for predicting Reco under climate change, considering continuous warming and increases in the frequency and intensity of extreme events (e.g., heatwaves, droughts). Moreover, our results suggest that process-based carbon models should adequately represent the effects of substrate supply (e.g., by GPP) on Reco and its temperature sensitivity. (C) 2019 Elsevier B.V. All rights reserved.