Einfluss des Wassergehalts, des Eintrags von Wurzeln und gelöster organischer Substanz und räumlicher Unzugänglichkeit auf den C-Umsatz & Bestimmung der räumlichen Variabilität von Bodeneigenschaften

新鲜有机质供应减少、有机质与矿质相的相互作用以及空间不可达性影响了地下土壤中碳储量。然而，为了更好地了解每一种不同过程对地下土壤中碳固存的贡献和改进地下土壤碳模型所需的定量信息很少。地下土壤有机质分解速率的主要控制因素也是如此。此外，关于底土不同性质的空间变异性的信息很少。现有的少数研究将近红外光谱和中红外光谱（NIRS/MIRS）与地质统计学方法结合起来，表明有可能创建空间地图，显示土壤剖面中生物活动增加的热点及其对C含量分布的影响。目标是(i)通过结合14C测量应用分馏程序来确定不同馏分中底土C的平均停留时间；（ii）研究水分含量、13c标记根和溶解有机质的输入以及空间不可达性对自动微观系统中碳转换的影响；（iii）利用近红外光谱（NIRS）和MIRS确定土壤的一般性质和土壤的生物和化学特征，以及（iv）利用不同的空间插值技术将测量和估计的土壤性质外推到垂直剖面。对于NIRS/MIRS应用，将优化样品预处理（风干与冻干样品）和校准程序（改进的偏最小二乘（MPLS）方法与结合MPLS或PLS的遗传算法）。我们假设，结合化学分馏与14C测量和培养实验的结果，将得到被动、中间、不稳定和极不稳定C和N的池大小以及不稳定和极不稳定C和N的平均停留时间。这些结果将使PC子项目开发的新定量模型初始化成为可能。此外，我们假设样品预处理“冷冻干燥”将比空气干燥对土壤生物学特性的估计更有用。预计GA-MPLS和GA-PLS方法比MPLS和PLS方法能更好地估计土壤特性。不同底土特征的空间分布图与温度和含水量的空间分布图相结合，可能有助于解释C含量的空间异质性。

It is well established that reduced supply of fresh organic matter, interactions of organic matter with mineral phases and spatial inaccessibility affect C stocks in subsoils. However, quantitative information required for a better understanding of the contribution of each of the different processes to C sequestration in subsoils and for improvements of subsoil C models is scarce. The same is true for the main controlling factors of the decomposition rates of soil organic matter in subsoils. Moreover, information on spatial variabilities of different properties in the subsoil is rare. The few studies available which couple near and middle infrared spectroscopy (NIRS/MIRS) with geostatistical approaches indicate a potential for the creation of spatial maps which may show hot spots with increased biological activities in the soil profile and their effects on the distribution of C contents. Objectives are (i) to determine the mean residence time of subsoil C in different fractions by applying fractionation procedures in combination with 14C measurements; (ii) to study the effects of water content, input of 13C-labelled roots and dissolved organic matter and spatial inaccessibility on C turnover in an automatic microcosm system; (iii) to determine general soil properties and soil biological and chemical characteristics using NIRS and MIRS, and (iv) to extrapolate the measured and estimated soil properties to the vertical profiles by using different spatial interpolation techniques. For the NIRS/MIRS applications, sample pretreatment (air-dried vs. freeze-dried samples) and calibration procedures (a modified partial least square (MPLS) approach vs. a genetic algorithm coupled with MPLS or PLS) will be optimized. We hypothesize that the combined application of chemical fractionation in combination with 14C measurements and the results of the incubation experiments will give the pool sizes of passive, intermediate, labile and very labile C and N and the mean residence times of labile and very labile C and N. These results will make it possible to initialize the new quantitative model to be developed by subproject PC. Additionally, we hypothesize that the sample pretreatment “freeze-drying” will be more useful for the estimation of soil biological characteristics than air-drying. The GA-MPLS and GA-PLS approaches are expected to give better estimates of the soil characteristics than the MPLS and PLS approaches. The spatial maps for the different subsoil characteristics in combination with the spatial maps of temperature and water contents will presumably enable us to explain the spatial heterogeneity of C contents.