Ecosystem respiration in rangelands is dominated by contributions from soil organic matter (SOM) decomposition and root/rhizosphere respiration. These processes may have similar or different responses to land use or climate change, so a predictive understanding of carbon (C) cycling requires separation of these processes. For example, increasing temperatures and atmospheric carbon dioxide (CO2) concentrations may both stimulate decomposition rates more than root respiration, potentially providing a positive feedback to global warming. Stable isotopes and manipulative experiments can be applied to partition the sources of respiration in ecosystems over a range of spatial and temporal scales. In systems where an isotopic label has been applied, such as in many elevated CO2 experiments, differences in the C stable isotopic composition between newly fixed C and older SOM may allow tracing of inputs as well as efflux of new C. For example, decomposition rates were stimulated by 20-80% under elevated compared to ambient CO2 treatments in shortgrass steppe, whereas root respiration rates were not altered. In systems without an added label, natural variations in oxygen isotopes or manipulations such as root exclusion may provide insight into respiration sources. Examples from grassland and sagebrush steppe ecosystems will be presented demonstrating different approaches to partitioning.