Global patterns of soil autotrophic respiration and its relation to climate, soil and vegetation characteristics

Soil autotrophic respiration (RA) is one of the key components of carbon and nutrient cycling in terrestrial ecosystems. Despite conflicting findings on the dominant climatic, edaphic and vegetation controls of RA from localized studies, little is known about global RA patterns and their potential drivers. To fill the knowledge gap, we evaluated the direct and indirect effects of climatic, edaphic and vegetation variables on global RA using regression analysis and a structural equation model (SEM) based on the most updated Global Soil Respiration Dataset with 499 observations. Results showed that mean (+/- standard deviation) RA was 358 +/- 283 g C m(-2) yr(-1) across global terrestrial ecosystems and differed significantly among forest ecosystems, croplands and grasslands (p < 0.01), highlighting the importance of vegetation on RA. Mean RA was the highest in evergreen broad-leaf forests (633 +/- 397 g C m(-2) yr(-1)) and the lowest in deciduous needle-leaf forests (226 +/- 170 g C m(-2) yr(-1)). RA was positively correlated with mean annual temperature and precipitation, gross primary production (GPP), ecosystem respiration (ER) and total belowground carbon allocation (TBCA), whereas it was negatively correlated with silt and clay contents, and a unimodal relation was observed between RA and sand content with a maximum at 68%. These findings implied that the amount of carbon respired from roots back into the atmosphere may increase in the context of long-term climate change, mediated by soil and vegetation. SEM analysis indicated that TBCA was the most important factor controlling RA, accounting for 50% of the variation in RA at the global scale. Globally, RA ratios to GPP and ER were 20% and 23%, respectively, which, combined with developed global GPP and ER products, provide a new concept to estimate the spatiotemporal patterns of RA at regional or global scales. However, to increase the reliability of global RA estimates using these two ratios, more simultaneous observations of RA, ER or GPP from different ecosystems are strongly recommended.