【作者】 林慧龙；

【导师】 任继周；

【作者基本信息】 兰州大学 ， 草业科学， 2007， 博士

【摘要】 践踏是放牧家畜对草地的三大影响因素之一，也是导致草原全面退化的最主要的原因。为了深入探索草地退化机理和草地恢复机制，研究家畜对草地的践踏作用是非常必要的，但是由于技术和方法的限制，相关的研究报道较少，极大地削弱了我们对草地农业生态系统的主体—草畜界面生态动力机制的整体认识。基于上述原因，在我国干旱半干旱的黄土高原典型草原区——兰州大学环县草畜生产系统野外实验点，通过滩羊夏季轮牧试验践踏强度的跟群观测、践踏同质性试验、模拟降水与试验践踏的双因子试验进行了为期2年的野外试验系列研究。首先建立了践踏强度指标和严格的践踏试验方法，然后围绕践踏强度这一主线重点揭示土（土壤理化性质，土壤侵蚀）—草（地下生物量）—畜（践踏）在不同降水条件下的互作机制，对家畜践踏这一不受关注的“小”问题，应用数学生态学的原理和方法开展综合性、系统性的研究，探讨家畜践踏对草地的作用。主要结果如下：1、提出践踏强度是在特定的放牧方式（自由放牧或轮牧）下，单一种或组合种畜群单位时间内（通常指一个放牧季内）对单位草地面积所施加的践踏重力或践踏面积，它应有三种表示方法：重力表示法、面积表示法、频率表示法。首次提出了践踏单位和践踏当量概念，以明确践踏强度的计算标准和折算标准，有助于使当前不统一的践踏强度表示方法规范化。2、通过在环县典型草原不同放牧强度下轮牧试验的跟群观测，获得了家畜践踏强度依放牧强度的幂函数式增大的定量结论，揭示了放牧强度与践踏强度的关系，该定量结论为调控践踏提供了实践途径，为制定科学的放牧制度提供了又一新的理论与实践的基础。3、研制了模拟践踏器，通过对试验践踏样区与传统的滩羊夏季轮牧样区的土壤和植被的6项指标的对比研究表明，模拟践踏与放牧滩羊的践踏具有同质性，为本研究提供了将践踏与采食和排泄的影响区分开来的试验手段，使能对践踏进行精确的研究。4、根系是植物体受放牧家畜采食影响较小，而受放牧活动机械干扰（践踏）较大的部位，但践踏对草地地下生物量影响的专项研究较少。对以0-10cm地下生物量为目标函数，以践踏强度、牧草生长期降水量为决策变量的践踏、降水耦合模型，以地下生物量最大化为准则，进行解析和寻优分析，确定出环县典型草原适宜的草地放牧强度是3.84-5.09羊单位／hm²。5、对模拟降水与试验践踏的双因子野外控制试验的48个试验处理组合的5项土壤化学指标进行分析，结果表明在践踏强度梯度上碱解氮（mg／kg）和全磷（％）差异显著；在模拟降水量梯度上，碱解氮（mg／kg）、全磷（％）、速效磷（mg／kg）和有机质（％）差异显著，表明土壤化学性状对模拟降水更为敏感。践踏强度与模拟降水量交互效应对碱解氮（mg／kg）、全磷（％）、有机质（％）影响显著，说明践踏和模拟降水均是引起土壤化学性质变化的驱动力因素。6、通过模拟降水与试验践踏的野外控制试验，对不同试验组合样区表土（0-15cm）质地的机械组成，利用分形模型计算了土壤粒径分形维数，计算结果表明：分维数值均表现为重度践踏区＜中度践踏区＜轻度践踏区＜对照区，说明分维数指标对践踏强度呈一定的负相关，即践踏强度越大土壤的分维数值越小。但分形维数递减幅度与模拟降水处理水平相关，在模拟降水处理水平依次为干旱、自然降水、平水、丰水时，粒径分形维数递减幅度依次为4.33％、2.61％、1.34％和1.23％。降水与践踏的匹配关系，对形成确定的分形维数具有决定性作用，其回归模型是：F=2.427+2.0982×10^-4W-1.3333x10^-4T。土壤侵蚀过程在土壤颗粒分形维数上亦可表现为分形维数降低，因此，在此种意义上来说，土壤分维数的测算对土壤侵蚀程度的评估具有一定的指示意义，对进一步研究放牧草地土壤侵蚀机理可能具有推动作用。7、降水和践踏通过各种不同强度的侵蚀外营力组合方式，改变土壤可蚀性。以牧草生长期单位面积累计践踏量和模拟降水量为自变量的土壤可蚀性K值的ANN（Artificial Neural Networks）关系模型具有较好的拟合结果和预测能力，说明直接从输入到草地生态系统的外侵蚀营力着手，跨越系统内土壤可蚀性变化的内在的复杂的隐含过程，建立的输出端—土壤可蚀性K值与土壤侵蚀外营力的ANN（Artificial Neural Networks）关系模型是准确确定土壤可蚀性K值的一次全新的成功尝试。8、模拟降水和试验践踏通过各种不同的组合方式改变土壤抗蚀力，从而起到对草地土壤侵蚀的增减作用。在牧草生长期，中、高强度践踏对土壤的扰动，加剧了干旱状况下土壤风蚀的风险，降水具有双向调节作用，即在践踏强度升高时具有降低土壤侵蚀的作用，而在践踏强度降低时又具有升高土壤侵蚀的作用，但显然模拟降水对土壤侵蚀的限制作用小于践踏对侵蚀的增加作用。构建的以试验期土壤侵蚀模数为因变量，以践踏、模拟降水累计量为自变量的机理模型，可为黄土高原放牧地土壤侵蚀产沙量提供计算方法。据此，以土壤侵蚀最小化为目标，对于不同的牧草生育期降水量，解析得到的最适放牧强度显然是对正确确定载畜量的一次新的尝试。9、在可持续发展前提下，仅仅根据草畜的供求关系管理放牧系统远远不够，科学管理放牧系统还需要兼顾“地下生物量”和“土壤侵蚀”状况，为此提出：适宜的放牧率=min（理论载畜量，以地下生物量最大化为准则确定的适宜放牧率，以土壤侵蚀最小化为目标确定的最适放牧率）。由此得出环县典型草原在牧草生长期降水量为224.9mm时，最适宜的践踏强度为：39次羊践踏／m²·期，对应的放牧强度为：3.8羊单位／hm²。更多还原

【Abstract】 An animal exerts three main influences on pasture - it treads, removes leaves （defoliates）, and excretes upon it. Continued overgrazing and erosion cause range degradation by excessive animal trampling. Published studies that have examined the effects of grazing often do not distinguish between effects arising from these separate activities. Little is known about the result of trampling-induced feedback among the soil environment, grass and their interdependency. In addition, more information is needed on the effects of animal trampling on the managed grazing ecosystems. In typical steppe of loess plateau regions, where the management objective is maintaining grassland ecosystem structure and function, the impacts of grazing trampling can be a serious problem. This was one of the reasons for conducting this study on trampling effects. The object of the study was to evaluate separately the effect of trampling in typical steppe of Huanxian County in eastern Gansu Province, China. A series of field experiments were conducted there, which lasted for two years. They were involved of trampling intensity observation of tailing up Tan-sheep rotational grazing trail in summer season, trampling homogeneity trial, double factors trail between simulated precipitation and experimental trampling. Firstly, the index of trampling intensity and the strict trial method of experimental trampling were set up. Techniques of Mathematical Ecology were used to describe the variation among the response variables for the use levels （trampling intensity and simulated precipitation）. This paper quantifies the impact of controlled experimental trampling on belowground plant biomass, soil physical and chemical properties and soil erosion under different precipitation conditions with tramping intensity as its core. The results were as follows.1. The trampling intensity （TI） proposed in the paper is defined under free or rotational grazing, trampling area or trampling pressure of grazing single or multiple livestock per hectare during a unit time. It has three expressive methods: trampling pressure method, trampling area method, trampling frequency method. The trampling unit and Trampling Equivalent Unit （TEU） were put forward for the first time in order to define the calculating standard and converting standard of TI.2. By tailing up and observing the flocks on typical steppe of Huanxian County in Eastern Gansu, China, a relation equation by regression analysis showed that livestock trampling intensity is increased by the power function of grazing intensity. This fully demonstrated that adjusting and controlling the trampling effects of grazing livestock is an important indication of grazing management. The trampling effects of grazing livestock played a key role in preventing the degeneration and maintaining the wholesomeness of the grassland.3. Manufacture of trampler was imitated for mechanic characteristics of Tan-sheep trampling. By a simulation homogeneity experiment, we compared experimental trampling and traditional Tan-sheep rotational grazing in summer season. Statistical analyses （t test） indicated that there were no significant differences between modelling Tan-sheep trampling and traditional rotational grazing in the height of leaf layer, species number of plant, the water content of top layer （0-15cm）, the soil porosity, bulk density and soil compaction. There is homogeneity between experimental trampling and traditional grazing in the effect of grassland trampled. This lay the foundation for future research in animal trampling effects for more effectively.4. Early research on grassland root systems focused on qualitative descriptions of root system structure. But the response effects of roots for different trampling intensity are not well-understood. Taking trampling intensity and rainfall as independent variables, the coupling model of underground plant biomass of 0-10cm layer was scanned optimization with underground plant biomass maximum as a target. The most suitable grazing intensity for the typical steppe of Huanxian County was decided to be 3.84-5.09 sheep unit/hm² which is converted by the relation equation between trampling intensity and grazing intensity. The models offered a new pattern and quantitative tool for the grazing system management.5. There are 48 experimental treatment combinations on five soil chemical indicators in field controlling experiment of two factors: experimental trampling and simulated precipitation. The analysis results by multivariate and multiple analyses of variance indicate that there is significant difference between hydro N （mg/kg） and total P （%） in trampling intensity gradient. However, there are significant differences among hydro N （mg/kg）, total P （%）, available P （mg/kg） and the content of soil organic matter （%） in simulated precipitation gradient. This shows that soil chemical properties are more sensitive to simulated precipitation. The effects of interaction between experimental trampling and simulated precipitation have significant influences on hydro N （mg/kg）, total P （%） and organic content. Trampling and simulated precipitation factors play a role in facilitating ecosystem state change in the typical steppe of Huanxian County in Eastern Gansu, China.6. We analyzed the soil mechanical composition （0-15cm） of different trial sample plots in field controlling experiment of two factors: experimental trampling and simulated precipitation. Fractal dimension can be calculated by using fractal model. The result shows that fractal dimension characterized by heavy trampling plot<moderate trampling plot<light trampling plot<comparison plot. A negative correlation was found between fractal dimension and trampling intensity. That is to say, the smaller fractal dimension, the higher space heterogeneity will be. However, the decrease extent of fractal dimension was relation to simulated precipitation level. When simulated precipitation level was dry, natural precipitation, moderate rainfall, and abundant rainfall in turn, the decline extent of fractal dimensions value was 4.33%, 2.61%, 1.34% and 1.23% accordingly. The matching relation between precipitation and trampling had significance with determinate fractal dimension values. Two variable regression patterns expressed: F = 2.427 + 2.0982×10^-4 W -1.3333×10^-4 T. The soil eroding process also can be characterized fractal dimension values decrease. So, in this sense, fractal dimension measurement has certain indication significance with evaluation of soil erosion. This pushes forward future research in soil erosion mechanism in grazing grassland.7. The sensitivity of soil erosion to trampling-caused was changed with various different combinations of experimental trampling and simulated precipitation. Compared with traditional regression model, the artificial neural network model by Error Back Propagation （BP） can complete the news treatment of the whole network by means of a mutual function between the neural units, and has a lot of merits of self-learning, self-adjusting and fault tolerance such like. Therefore it is feasible that the soil erodibility factor K value is calculated by the neural network. Soil erodibility is complicated, which is affected by many factors such as trampling and precipitation. The artificial neural network model skipped the mechanism of soil erodibility, conducted research on soil erodibility by both trampling and rainfall factors. Obviously, this was an innovative attempt to precisely evaluate soil erodibility for understanding soil erosion regularity.8. Trampling directly facilitates erosion by hoof. Trampling of the soil is the most important factor contributing to erosion. Study results indicated that in grass growth period, moderate and heavy trampling aggravated soil erosion risk under arid circumstances. Precipitation has the effects of two-side regulating soil erosion. With increasing trampling intensities, precipitation can lower soil erosion modulus, but soil erosion modulus will increase with decreasing trampling intensities. However, the restriction effects of simulated precipitation on soil erosion were less than the trampling effects on soil erosion. The mechanism model, which took soil erosion modulus as dependent variable and takes experimental trampling and simulated precipitation as independent variable, was validated by experimental data. This can provide a new method for the evaluation of soil erosion in arid and semi-arid grassland regions. With soil erosion minimum as a target, the most suitable grazing intensity can be obtained by analyzing precipitation amount in different forage growth period. Obviously, this is an innovative attempt for ascertaining grazing capacity rationally. 9. The ecological restoration of degraded grassland ecosystems has become the key task for the future sustainable development. So an integrated method for calculating a rational stocking density should be adopted. It gives attention to two things: underground plant biomass and soil erosion, besides the balance between forage yield and livestock utilization. The most suitable grazing intensity calculated by the integrated method can be denoted as minimum （"determine livestock carrying capacity according to grass", the suitable grazing intensity with underground plant biomass maximum as a target, the suitable grazing intensity with soil erosion minimum as a target）. As a result, the most suitable trampling intensity for the typical steppe grassland of Huanxian County is decided to be 39 UST /hm², in which the precipitation amount in forage growth period is 224.9mm, and the corresponding most suitable grazing intensity converting by the relation equation between trampling intensity and grazing intensity is 3.8 sheep unit/hm².更多还原