The present study was performed in 16 plots spanning a forest age gradient from 20 to 120 years. We investigated the effects of the successional stage on soil respiration (Rs), autotrophic respiration (Ra) and heterotrophic respiration (Rh) in an evergreen subtropical broad-leaved forest. Rs showed significant differences among three successive stages, and it was the highest in the young stage. Ra was higher in the young stage than in the medium stage. Cumulative annual Rs and Ra peaked in the young and old stages, respectively. Cumulative annual Rh and respiration measured from soil organic matter (RSOM) decreased, whereas RL (respiration of above ground litter) increased with forest age. The SEM revealed that cumulative annual Rs was influenced by fine root biomass and soil volumetric water content (SVWC). Our results indicated that the dominant force regulating Rs on a seasonal scale is soil temperature (ST); however, on a successional scale, belowground carbon emerges as the dominant influential factor.

Rs was measured using a closed chamber connected to a portable infrared gas 144 analyzer from April 2009 to March 2012. In each plot, six shallow polyvinyl chloride (PVC) collars (20 cm interior diameter and 8 cm height) were permanently installed along a transect with 1 m intervals between collars, to depth of 3 cm, thus protruding 5 cm above the ground. Six deep collars (20 cm interior diameter and 70 cm height) were installed to a soil depth of 65 cm to exclude plant roots and organic matter input from decomposing roots (> 90% of the belowground biomass is distributed in the top 30 cm of soil) . The shallow and deep collars were used to measure Rs and Rh, respectively. The difference between Rs and Rh was used to estimate Ra. Litter was removed from half of the respiration collars to separate Rs into litter and mineral soil components, and the collars were covered with nylon mesh with a 2 mm pore size to exclude plant litterfall. These three collars were used to measure soil respiration without litter 155 (RNL). The difference between Rs and RNL was used to estimate RL (respiration from aboveground litter), and the difference between Rh and RL was used to estimate RSOM (respiration from soil organic matter). Both Rs and Rh were measured manually twice with an interval of 1 hour every two months. Soil temperature at a depth of 5 cm (ST) was recorded simultaneously with the Rs 162 using a soil temperature probe connected to the portable soil respiration system. From April 2009 to August 2010, the SVWC of the upper 10 cm was measured gravimetrically. In August 2010, we added two EM50 data loggers with an air temperature and humidity Sensor and an ECH2O soil temperature and moisture probe in the middle of each successional stage. The air sensor and ECH2O were used to measure air temperature and humidity at 1.50 m aboveground and ST and SVWC, respectively, at 30 min intervals.

CSP01 CSP02 CSP04 CSP05 CSP08 CSP09 CSP12 CSP13 CSP14 CSP16 CSP17 CSP18 CSP22 CSP23 CSP25 CSP26

Ordinary linear regression was used to investigate the relationship between forest age and ST, SVWC, Rh, RSOM and RL, and the relationship between FRB and Rs, Rh and Ra. The impact of successional stage on Rs and environment parameters was examined by one-way analysis of variance (ANOVA), and a least square deviation (LSD) test was used to compare the mean differences of Rs and environment parameters among successional stages.