Despite heavy study in past decades, responses of the major components of soil respiration (root-derived vs.soil microbe derived) to environmental factors is still not well-understood.This is mainly due to the inaccessibility of major soil organisms for experimental manipulation.Based on data collected in the 2009 growing season, we used a field trenching method for soil respiration partitioning in a mixed-grass rangelands with two different grazing histories.Total soil respiration was modeled as a function of soil temperature, soil water content and leaf area index, while microbial respiration (from trenched plots) was modeled as a function of soil temperature and soil water content.The models explained 62% to 67% of the variations in data.Over 111 days, from May 14 to September 1, 2009, the average fraction of root respiration (out of the total respiration) was estimated to be 54% and 67% under moderate grazing and heavy grazing, respectively.These percentages include root respiration, rhizosphere respiration, and the microbial respiration of decaying plant roots.Both modeled and measured root respiration rates showed higher values during the rapid growth period from May to early July.Root respiration mainly responded to large rain events and was relatively insensitive to temperature pulses.Microbial respiration responded to both the rain and temperature pulses.Respiration cost for leaf production (and presumably also for forage production) tended to be lower under moderate grazing than under heavy grazing.Further study involving multi-year data and alternative models is needed to minimize the underestimation seen in this paper for predicting soil respiration under field conditions.