资源环境科技发展态势分析平台

To determine the effectiveness of forests in sequestering atmospheric carbon (C), we must know the amount of fixed carbon dioxide (CO2) that is subsequently lost due to heterotrophic microbial activity in the soil. Furthermore, the heterotrophic proportion of total soil respiration (R-s) must be quantified as it changes between different physiographic regions, seasons, and silvicultural treatments. This research quantified heterotrophic contributions to R-s in loblolly pine (Pins taeda) plantations in the Piedmont (n = 3) and Upper Coastal Plain (n = 3) of the Southeastern USA under control, fertilized, and herbicide treatments over an annual cycle. Heterotrophic respiration (R-h) was separated in the field from autotrophic root respiration (R-a) using metal root-excluding collars. The R-h proportion of R-s was not significantly different between regions or treatments, but demonstrated some seasonal variance. The average R-h proportion across the study was found to be similar to 73 +/- 2% but ranged from similar to 70% in winter, spring, and summer to 82% in the fall. Statistical models using microbial biomass, temperature, moisture, and other soil characteristics explained 82% and 75% of R-s and R-h variability, respectively. In contrast, the process based DAYCENT model, parameterized for each site to model R-s, R-h, and R-h proportion compared poorly to field measurements. Model predicted mean seasonal R-h proportions also extended beyond the range of those measured (65-88%) from 61 +/- 1.3% to 94 +/- 0.4%. DAYCENT performed slightly better (i.e., lower root mean square error) for Piedmont than Coastal Plain sites. DAYCENT does not simulate CO2 fluxes below 20 cm and may be missing substantial fluxes from deeper roots and microbial activity. The results from this study suggest that statistical models such as multiple regression may provide more accurate estimates of R-h proportion for regional extrapolation than the current formulation of the process based DAYCENT model. It is unclear, however, if either approach captures seasonal variation in R-h or how strongly R-h varies with season. Finally, the empirical field data suggest the use of fertilizer and herbicides in these ecosystems increases ecosystem productivity without increasing R-h, which results in an increase in net ecosystem productivity that may lead to greater rates of C sequestration.