Forest ecosystems in northeastern China play an important role in both local and national carbon budgets because of their large area extent and huge amount of carbon storage. The spatial and temporal changes in soil surface CO₂ flux (R_S), the major CO₂ source to the atmosphere from terrestrial ecosystems, directly influence the local and regional carbon budgets. However, few data on R_S were available for this region. In this study, we used an infrared gas exchange analyzer (LI_COR 6400) to measure the R_S and related biophysical factors, and examined soil temperature and moisture effects on soil respiration for six secondary temperate forest ecosystem types: Mongolian oak (dominated by Quercus mongolica), poplar_birch (dominated by Populus davidiana and Betula platyphylla), mixed_wood (no dominant tree species), hard_wood forests (dominated by Fraxinus mandshurica, Juglans mandshurica and Phellodendron amurense), Korean pine (Pinus koraiensis) and Dahurian larch (Larix gmelinii) plantations. Our specific objectives were to: 1) compare the soil temperature, soil moisture, R_S, and Q₁₀ (temperature coefficient) of the six forest types; 2) quantify the seasonality of R_S and related environmental factors; and 3) determine the environmental factors affecting the R_S, and construct models of R_S against the related environmental factors.

Soil temperature, soil moisture and their interactions significantly (p < 0.01) influenced the R_S, but their effects depended on forest type and soil depth. These factors could explain 67.5%-90.6% of the variations in the R_S data. During the growing season, the soil temperature at 10 cm depth in the different forest types did not differ significantly but soil moisture did. The R_S for the oak, pine, larch, hardwood, mixed_wood, and poplar_birch stands varied from 1.89-5.23, 1.09-4.66, 0.95-3.52, 1.13-5.97, 1.05-6.58, and 1.11-5.76 μmol CO ₂·m^-2·s^-1, respectively; the Q₁₀ values for those stands were 2.32, 2.76, 2.57, 2.94, 3.55 and 3.54, correspondingly. The seasonality of R_S was driven mainly by soil temperature and moisture, and was roughly consistent with that of soil temperature. The broad_leaved forests had a higher soil respiration rate than those of coniferous forests probably because of a more suitable soil thermal and moisture regimes and other biological factors.

The temperature sensitivity coefficient of soil respiration (Q₁₀) showed a convex_type curve along a soil moisture gradient. The Q₁₀ tended to increase when soil moisture increased from 30.19 to 40.7, and then declined probably because the extremely high soil moisture content in the hardwood forest may impede activities of soil microbes and plant roots, and thus decrease decomposition rates and soil CO₂ emission. Our study strongly recommended that estimation of soil surface CO₂ flux from forest ecosystems should consider the comprehensive effects of both soil temperature and moisture on soil respiration so as to reduce uncertainties of ecosystem carbon budget studies in this region.