【作者】 邹松兵；

【导师】 冯兆东；

【作者基本信息】 兰州大学 ， 自然地理， 2006， 博士

【摘要】 人类对自然环境的改造深刻地影响着地球系统的生物地球化学过程,该过程通过生态过程、水文过程和气候过程之间的高度耦合作用,使支撑人类社会经济的自然生态环境系统发生着人类不可预知的变化。为了社会经济的可持续发展,需要理解人类活动对于生态环境系统的影响程度,其基础性工作就是关于潜在自然植被区域分布的探索。为探索中国黄土高原生态环境建设之路,构建本区域可持续发展格局,需要理解以植被为主导的生态系统的原本面貌。从中国黄土高原生态—水文—气候过程的相互耦合联系的角度,本研究进行了气候植被尺度和群系物种尺度上潜在植被分布的规律性探索。 本研究进行了四个方面的工作:1、构建用于进行潜在植被类型分布模拟的气候要素的分布式模型。气候要素主要包括:降水、气温、地面温度、最低气温、最高气温、空气湿度、蒸发及潜在太阳辐射。首先通过对气候要素详细的空间结构模式的探索,提出了中国黄土高原地区的气候分布结构函数;之后对剔除结构性分量后的残差进行详细的空间内插分析和模型比较,最后形成本区域空间分布式气候要素模型。2、采用柯本模型、吉良模型与HOLDRIDGE模型进行本区域的气候植被分类模拟,并对HOLDRIDGE模型进一步分类,形成HOLDRIDGE主要气候植被模型和HOLDRIDGE过渡性气候植被综合模型模型,其中HOLDRIDGE主要气候植被模型与柯本和吉良模型具有可对比性,且模拟效果要好,包含过渡性植被类型的综合模型能模拟出植被型尺度上的中国黄土高原潜在植被分布,模型结果基本上体现了本区域气候控制下的潜在植被类型。3、利用预设预测规则的遗传算法(GARP)、BIOCLIM算法、DOMAIN算法及生态位因子分析算法(ENFA)四种算法对本地区主要的12种地带性植被型中17种典型群系物种进行了潜在空间分布模拟。其中GARP能模拟出物种潜在分布的空间格局;DOMAIN算法与BIOCLIM算法模拟的结果具有相似的结构特征,但DOMAIN算法适合于连续大片分布的物种模拟,BIOCLIM算法在小尺度上具有对潜在物种模拟的良好效果;对于具有明显垂直地带性分布特征的物种,ENFA算法具有其他算法不可比拟的优势。4、根据模拟的结果初步分析了中国黄土高原潜在植被分布的地理空间特征,根据不同类型生态系统典型物种的空间分布态势,提出人类活动干预在本地区植被物种潜在分布具有重要影响的观点,即现代气候条件下,人类活动能导致植被物种地理分布的转型,黄土高原大部具备森林和森林草原存在的气候条件,其生态环境的恢复和治理需要人类活动基于生态规律的积极干预。更多还原

【Abstract】 Human modifications of the natural environment profoundly affect biogeochemical processes of the Earth’s climatic systems. These processes may bring about unpredictable changes of the natural eco-environment that supports our human socio-economic systems. To promote the sustainability of socio-economic developments, the need to assess the degree to which human activities can affect ecosystems is pressing. In other words, we need to know how much of the natural ecosystems have been altered. Or to simply put, what are the potential natural ecosystems. While nurturing the Chinese civilization, the Chinese Loess Plateau (CLP) has been radically changed in its visage of natural environments, as expressed by environmental issues such as ecological deterioration and soil degradation. After a long period of abusive economic over-exploration, the consensus has been recently reached by the government and the general public that the disrupted ecosystems need to be restored under scientific guidelines of potential ecology with a full consideration of current socio-economic status. To explore scientifically sounding ways of ecological restoration and to plan for a long-term socio-economic sustainability, we need to understand the maximum potentials of the widely-advocated ecological restoration in the Chinese Loess Plateau. This thesis is designed to explore the spatial patterns of the maximum potentials of various vegetation types both in the climate-vegetation scale and in the community-species scale by fully investigating the interacting and coupling relationships among ecological, hydrological and climatic processes in the plateau.Four aspects of work are involved in this thesis. First, distributed statistical models were constructed to spatially interpolate all of key environmental factors that contribute to the formation of ecological patterns. These key environmental factors include the following: precipitation, air temperature, surface temperature, minimal air temperature, maximal air temperature, air humidity, evaporation and potential solar radiation. These models first explore in detail the spatial structural functions and address spatially-distributed climate structure functions. Then, we conducted spatial analyses and model comparisons of the residues after removals of the structural components. Finally, the spatially-distributed climate factor models for the study area were constructed. Second, the Koppen, Kira, and HOLDRIDGE models were used to simulate the spatial distributions of climate-vegetation types in the study area. The normal HOLDRIDGE model was particularly used for further classifying HOLDRIDGE-defined major sub-climate-vegetation types, and a HOLDRIDGE transitional model was then used for delineating the spatial patterns of transitional climate-vegetation types. The normal HOLDRIDGE modeling results appear to be well acceptable and can be well compared with the results of the KSppen and Kira modeling. The latter model (HOLDRIDGE transitional model) seems to be quite effective in simulating the maximum potentials of vegetation distributions in the CLP, especially regarding the influences of regional climate conditions. Third, four algorithms (i.e. GARP, BIOCLIM, DOMAIN and the ENFA algorithms) were employed to simulate the potential distributions of 17 typical community-species out of 12 zonal vegetation types that dominate the study area. Among these simulations, GARP seems to well depict the spatial pattern of potential distributions of community-species. Although DOMAIN and BIOCLIM algorithms share some similar structural characteristics, the former depicts large-patch distributions of certain specieswhile the latter yields better results in simulating small-patch distributions of certain species. For species with characteristics of conspicuous vertical zonality, ENFA algorithm has unparalleled advantages over the other three algorithms. Finally, we propose that interventions of human activities can play a significant role in modifying the maximum potential distribution of vegetation-species in the study area according to the spatial trends of typical species distributions in different ecosystems. Human activities may results in reorganizations in the spatial distribution of vegetation-species in the context of modern climatic conditions. The results show that most parts of the Chinese Loess Plateau has the potential climatic conditions favoring forest development (forest and forest-steppes), but restoration and management of the eco-environment in the CLP require visionary and well-thought human interventions.更多还原