ALBERT

Description: Background: Although numerous studies have been carried out in recent decades, soil respiration remains one of the less understood elements in global carbon budget research. Tropical forests store a considerable amount of carbon, and a well-established knowledge of the patterns, components, and controls of soil respiration in these forests will be crucial in global change research. Methods: Soil respiration was separated into two components using the trenching method. Each component was measured at multiple temporal scales and in different microhabitats. A commercial soil efflux system (Li8100/8150) was used to accomplish soil respiration monitoring. Four commonly used models were compared that described the temperature dependence of soil heterotrophic respiration using nonlinear statistics. Results and Conclusions: Trenching has a limited effect on soil temperature but considerably affects soil water content due to the exclusion of water loss via tree transpiration. Soil respiration decreased gradually from 8 to 4 μmol·m−2·s−1 6 days after trenching. Soil autotrophic (Ra) and heterotrophic respiration (Rh) have contrasting diel patterns and different responses to temperature. Rh was negatively correlated with temperature but positively correlated with relative humidity. Both Ra and Rh varied dramatically among microhabitats. The Q10 value of Rh derived using the Q10 model was 2.54. The Kirschbaum–O’Connell model, which implied a strong decrease of Q10 with temperature, worked best in describing temperature dependence of Rh. Heterotrophic respiration accounted for nearly half of the total soil efflux. We found an unexpected diurnal pattern in soil heterotrophic respiration which might be related to diurnal moisture dynamics. Temperature, but not soil moisture, was the major controller of seasonal variation of soil respiration in both autotrophic and heterotrophic components. From a statistical perspective, the best model to describe the temperature sensitivity of soil respiration was the Kirschbaum–O’Connell model. Soil respiration varied strongly among the microhabitats and played a crucial role in stand-level ecosystem carbon balance assessment.