Global annual soil respiration in relation to climate, soil properties and vegetation characteristics: Summary of available data

Soil respiration (Rs) is a key process in the terrestrial carbon cycle. Measurement and simulation of Rs has received much attention recently. We collected annual Rs field datasets to examine key controls of temporal and spatial variability in annual Rs at the global scale. Published studies that reported annual field Rs, mean annual temperature (MAT), annual precipitation (AP), soil (0-20 cm) properties and vegetation characteristics were compiled. MAT, AP and soil organic carbon (SOC) were the three most important variables in these datasets, together being responsible for 50% of the variance in annual Rs in a global model (MAT&AP&SOC-model). Combining other site soil properties (e.g. pH) and vegetation variables, such as tree age (TA), tree height (TH), litter fall biomass (LF) and leaf area index (LAI), into the MAT&AP&SOC-model improved model performance. The site characteristic that explained the most variation in Rs was AP followed by MAT, SOC, net primary productivity (NPP), pH, TA, TH, LF, LAI, elevation (EL) and diameter at breast height (DBH). Among the simulated models, the model based on MAT, AP, SOC and pH had the best fit for annual Rs variance. There was a highly significant logarithmic relationship between Rr/Rs (the contribution of root respiration to Rs) and AP. The AP value of 0.4 m was a threshold for Rr/Rs, corresponding to Rr/Rs of 0.4 which reflects water limitation of root growth and plant productivity.