Background and aims The knowledge of individual tree species impacts on soil respiration based on rigorous experimental designs is limited, but is crucial to help guide selection of species for reforestation and carbon (C) management purposes.

Methods We assessed monthly soil respiration and its components, litterfall input, fine root production and mortality under 19-year-old native coniferous Cunninghamia lanceolata and broadleaved Mytilaria laosensis plantations in sub-tropical China.

Results Total soil respiration from October 2011 to March 2013 was significantly lower under the C. lanceolata than the M. laosensis plantation. The difference in respiration rates derived from fine roots and the litter layer explained much of the variation of total soil respiration between the two tree species. We used an exponential equation and base temperature (10 °C) to normalize soil respiration rate and its components (R10) and determined the correlation between R10 and soil moisture. Although soil moisture had a positive relationship with R10 derived from roots or litter under both C. lanceolata and M. laosensis forests, these positive correlations were masked by negative relationships between soil moisture and R10 derived from root-freesoil, which resulted in a neutral correlation between total R10 and soil moisture under C. lanceolata forests. Monthly litterfall input was associated with variation in concurrent total soil respiration rate under the M. laosensis plantation and respiration rate lagging 3 months behind under the C. lanceolata plantation, which may suggest that litterfall input from M. laosensis can more rapidly produce C substrates for microbial respiration than litterfall from C. lanceolata. Conclusions This study highlighted that tree speciesinduced variation in the quality and quantity of fine roots and litterfall can impact not only the soil respiration rate but also the seasonal variation model of forest soil respiration.