Soil carbon diversity can be an important property for the stability of soil carbon. A problem is the lack of techniques for measuring this diversity. I suggest here the use of a combination of a general statistical principle, MAXimum ENTropy (MaxEnt), and a mechanistic model of organic matter decomposition, the Q model. The Q model provides the temporal development of the average carbon quality of litter and amount of soil organic C, which can be applied in a MaxEnt calculation to obtain a distribution of soil C over qualities. This distribution may not be the actual distribution but it is the most probable one. This distribution can be used to calculate aggregate properties for the total of soil C. I will use this distribution to calculate the temporal development of the variance in C quality as an expression of C diversity. The general tendency is that the variance declines with time of decomposition. Six long-term bare fallow (LTBF) from different climatic and management conditions were used to investigate which system properties are most important for the temporal development of the variance. The initial quality of the litter forming soil C is the dominant property. Chemical shifts in NMR spectra were tested as a possible way of measuring the variance in C quality.

土壤碳多样性可能是土壤碳稳定性的重要属性。问题是缺乏测量这种多样性的技术。我在这里建议结合使用一般的统计原理，最大熵（Maxentum Entropy，MaxEnt）和有机物分解的机械模型Q模型。Q模型提供了凋落物的平均碳质量和土壤有机碳含量的时间变化，可以将其应用于MaxEnt计算中，以获得土壤C在品质上的分布。这种分布可能不是实际的分布，但它是最可能的分布。此分布可用于计算土壤C总量的聚集特性。我将使用此分布来计算C品质方差随C多样性表达的时间发展。通常的趋势是方差随分解时间而降低。使用六个来自不同气候和管理条件的长期裸休（LTBF）来研究哪些系统属性对于方差的时间发展最重要。凋落物形成土壤C的初始质量是主要属性。测试了NMR光谱中的化学位移，作为测量C质量变化的一种可能方式。