Contemporary forest management practices are increasingly designed to optimize novel objectives, such as maximizing biomass feedstocks and/or maintaining ecological legacies, but many uncertainties exist regarding how these practices influence forest carbon (C) cycling. We examined the responses of soil respiration (R_s) to biomass harvesting and green-tree retention in an effort to empirically assess their impacts on C cycling. We measured R_s and soil microclimatic variables over four growing seasons following implementation of these management practices using a fully replicated, operational-scale experiment in aspen-dominated forests in northern Minnesota. Treatments included three levels of biomass removal within harvested areas: whole-tree harvest (no slash deliberately retained), 20% slash retained, and stem-only harvest (all slash retained), and two levels of green-tree retention: 0.1 ha aggregate or none. The relative amount of biomass removed had a negligible effect on R_s in harvested areas, but treatment effects were probably obscured by heterogeneous slash configurations and rapid post-harvest regeneration of aspen in all of the treatments. Discrete measurements of R_s and soil temperature within green-tree aggregates were not discernible from surrounding harvested areas or unharvested control stands until the fourth year following harvest, when R_s was higher in unharvested controls than in aggregates and harvested stands. Growing season estimates of Rs showed that unharvested control stands had higher R_s than both harvested stands and aggregates in the first and third years following harvest. Our results suggest that retention of larger forest aggregates may be necessary to maintain ecosystem-level responses similar to those in unharvested stands. Moreover, they highlight the innate complexity of operational-scale research and suggest that the initial impacts of biomass harvest on R_s may be indiscernible from traditional harvest in systems where incidental breakage is high.