1] Environmental Sciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA [2] Climate Change Science Institute, Oak Ridge National Laboratory, Oak Ridge, TN, USA.
1] Climate Change Science Institute, Oak Ridge National Laboratory, Oak Ridge, TN, USA [2] Biosciences Division, Oak Ridge National Laboratory, Oak Ridge, TN, USA.
ISME J. 2015 Jan;9(1):226-37. doi: 10.1038/ismej.2014.120. Epub 2014 Jul 11.
Climate feedbacks from soils can result from environmental change followed by response of plant and microbial communities, and/or associated changes in nutrient cycling. Explicit consideration of microbial life-history traits and functions may be necessary to predict climate feedbacks owing to changes in the physiology and community composition of microbes and their associated effect on carbon cycling. Here we developed the microbial enzyme-mediated decomposition (MEND) model by incorporating microbial dormancy and the ability to track multiple isotopes of carbon. We tested two versions of MEND, that is, MEND with dormancy (MEND) and MEND without dormancy (MEND_wod), against long-term (270 days) carbon decomposition data from laboratory incubations of four soils with isotopically labeled substrates. MEND_wod adequately fitted multiple observations (total C-CO2 and (14)C-CO2 respiration, and dissolved organic carbon), but at the cost of significantly underestimating the total microbial biomass. MEND improved estimates of microbial biomass by 20-71% over MEND_wod. We also quantified uncertainties in parameters and model simulations using the Critical Objective Function Index method, which is based on a global stochastic optimization algorithm, as well as model complexity and observational data availability. Together our model extrapolations of the incubation study show that long-term soil incubations with experimental data for multiple carbon pools are conducive to estimate both decomposition and microbial parameters. These efforts should provide essential support to future field- and global-scale simulations, and enable more confident predictions of feedbacks between environmental change and carbon cycling.
土壤的气候反馈可能源于环境变化，继而引起植物和微生物群落的响应，和/或与养分循环的变化相关。由于微生物的生理和群落组成及其对碳循环的相关影响发生变化，明确考虑微生物的生活史特征和功能，可能对于预测气候反馈是必要的。在这里，我们通过纳入微生物休眠和追踪碳的多种同位素的能力，开发了微生物酶介导分解（MEND）模型。我们针对四种具有同位素标记底物的土壤进行了长达 270 天的实验室培养，用两种版本的 MEND（有休眠的 MEND 和无休眠的 MEND，即 MEND_wod）对其进行了测试，这两种版本的 MEND 都能对长期碳分解数据进行很好的拟合（总 C-CO2 和 (14)C-CO2 呼吸以及溶解有机碳），但代价是显著低估了总微生物生物量。MEND 通过将微生物生物量的估计值提高了 20-71%，从而改善了 MEND_wod 的估计值。我们还使用基于全局随机优化算法的关键目标函数指数方法，以及模型复杂度和观测数据的可用性，来量化参数和模型模拟的不确定性。综合我们对培养实验的模型外推结果，长期土壤培养实验与多种碳库的实验数据有利于估计分解和微生物参数。这些努力应该为未来的野外和全球尺度模拟提供重要支持，并使环境变化和碳循环之间的反馈预测更加有信心。
