Aim: Root production and turnover play a key role in regulating carbon (C) flow in terrestrial ecosystems. However, a general pattern reflecting the responses of roots to increasing nitrogen (N) input has yet to be described. 

　　Location: Global terrestrial ecosystems. 

　　Methods: We conducted a meta-analysis to assess the central tendencies of root production, turnover rate and standing crop with respect to the experimental addition of N. We evaluated the effect of the form of N, root diameter and climatic (mean annual temperature, MAT; mean annual precipitation, MAP), biotic (ecosystem type, plant type and forest stand age) and forcing factors (experimental duration, N addition rate and cumulative amount of N) on the variations in root response. 

　　Results: Globally, the addition of N significantly decreased root production and turnover rate but had only a minor impact on root standing crop. In different ecosystems, the three root variables exhibited heterogeneous responses to N enrichment. Additionally, root production and turnover rate responded distinctly to diverse forms of N. The responses of root production and turnover rate to the addition of N were generally positively correlated with MAT and MAP but negatively related to forest stand age and experimental duration. The response pattern of root standing crop was negatively affected by MAT, MAP and forest stand age. However, none of the three root parameters had any obvious correlations with N addition rate or cumulative amount of N. 

　　Main conclusions: Our results demonstrate that, on aggregate, the addition of N decreased root production and turnover rate at the global scale. These root response patterns and the regulatory factors can be incorporated into earth system models to improve the prediction of belowground C dynamics.