基于植被动态的黄土高原生态地理分区

1 引言

自然区划作为地理学的核心议题和重要工作,在认识地理分异规律、社会生产以及区域规划活动中都发挥着极为重要的作用。自然区划着重从区域角度出发,通过观察和研究地表自然综合体以揭示地域分异规律,并探讨自然地域单元的划分与合并^{[ 1]}。基于不同的研究目标与不同发展阶段,自然区划产生了诸多分支,相关概念也十分丰富,如：综合自然区划^{[ 2]}、自然地理区划^{[ 3]}、生态地理区划^{[ 4]}、生物（态）气候区划^{[ 5]}等。其中,生态地理区划是结合了生态学思想,在自然区划的原理和方法上发展起来的新分支,是基于自然地理学的整体性和地域分异规律,从生态学的角度揭示自然生态区域的相似性和差异性,从而对该区域进行系统的整合和分区^{[ 6, 7]}。其中,特别关注类型分区与地域分区,生态地理分区研究是进行生态地理区划的前期阶段与基础工作^{[ 8]}。相对于自然区划,生态地理区划本质上更加关注具有相似生产潜力、受到相似生态胁迫和具备相似结构特征的生态单元^{[ 4]}。因此,通过生态地理区划所得到的内部生态特征与生产潜力高度均一性的地块,成为国家及地区生态恢复与重建、自然资源合理开发以及区域可持续发展的基本空间单元。

生态地理区划与分区因其综合性和整体性优势,一直以来都是国内外地理学者研究的热点领域之一。在国际上,1976年Bailey对美国生态系统进行了等级划分,第一个具有真正生态地理意义的区划方案自此诞生^{[ 9]}。此后,Rowe等学者对生态地理分区的原则、指标选择进行了进一步探索^{[ 10- 12]};Bailey于1989年绘制的全球生态地理区划图又进一步明确了应将气候、地形、植被要素作为生态地理分区的基础^{[ 13]}。近年来,更多的学者开始将大尺度遥感数据作为生态地理分区的可靠依据^{[ 14]},Mackey等基于MODIS数据对澳大利亚进行了生物地理分区,并提出一个适用于全部陆地区域的系统化通用框架^{[ 15]};Cruz等则指出使用RBFs神经网络与遥感数据结合的方法可大大提升分区效率,克服传统实地考察法中高时间成本与地域限制的缺点^{[ 16]}。与此同时,自然植被与农作物作为敏感的气候变化指标成为了辅助区域划分的最佳选择^{[ 17]},Gao等借助地统计学思想,基于NDVI将半方差分析与分形维数应用于青藏高原生态地理分区的边界验证^{[ 18]};Liu等则对生态地理分区的传统指标体系进行了优化,提出了有机质含量等5个新的土壤指标以替代单纯的土壤类型^{[ 19]}。在国内,黄秉维等在综合自然区划方面的成果^{[ 2, 20]}成为中国生态地理区划的雏形;1988年《中国自然生态区划与大农业发展战略》^{[ 21]}方案的发表则标志着生态地理区划研究在中国的正式兴起。2000年以来,郑度等采用传统三级划分法、专家集成与模型定量法实现了中国的生态地理区划^{[ 7, 22]};欧阳志云等着重关注生态敏感性区划在国家层面的应用^{[ 23]}。近年来,生态地理区划及其相关研究工作仍然十分活跃,但总体而言,现有研究大多注重对单一分区方案的解释与描述,对于不同分区方案之间的对比与优选缺乏定量化途径,优选结果往往依赖研究者的目视比较或主观判定。此外,现有的生态分区多基于传统的专家集成法^{[ 24]},使用自上而下的三级演绎途径予以划分^{[ 7, 25]},区划结果相对于自下而上的综合集成法更容易受到人为影响,并且容易忽略中小尺度的地带性分异规律。

随着计算机技术在自然区划研究中的应用愈发丰富,人工神经网络、遗传算法、模拟退火算法等新技术被广为利用^{[ 24]},大大提高了区划工作的效率和客观性。其中,利用自组织特征映射（self-organizing feature maps,SOFM）网络来实现自下而上的地理学及生态学分区研究已经有了部分文献与案例支持。张学儒等构建了地形综合指数、温暖指数、湿润指数、地被指数和水文指数等5个综合指标,采用神经网络与GIS技术对青藏高原东部进行了基于土地类型的区划研究^{[ 26]};韩忆楠等以内蒙古自治区农牧交错带为例,从大尺度景观空间分异的角度构建评价指标,运用SOFM网络将该地划分为6个景观分区^{[ 27]};Peng等重点关注快速城市化地区的土地利用分区,将不透水表面指数与SOFM技术相结合,划分深圳市土地利用区^{[ 28]};郝成元等基于云南省南部30个气象台站多年实测数据,利用SOFM网络识别出哀牢山对北来冷空气南下的屏障作用^{[ 29]},是SOFM网络在小样本区划工作中的重要实例。

黄土高原作为世界上黄土分布面积最大、黄土地貌最发育的地理单元,生态环境问题异常突出,成为中国乃至世界上水土流失最严重、生态环境最脆弱的地区之一^{[ 30, 31]},受到国内外众多地理学者的广泛重视。为缓解当地恶劣的自然环境与人地关系,自20世纪50年代起,中国诸多重点生态建设项目均在此实施,如1986-1990年“黄土高原综合治理”项目^{[ 32]},1999年启动的退耕还林（草）项目^{[ 33]},针对大型天然林种植与保护的“天保”工程^{[ 34]}及“三北”防护林工程^{[ 35]},缓解水土流失的小流域治理工程等。总体而言,得益于多个生态建设项目的顺利实施,黄土高原地区植被覆盖在近年来整体呈现增长趋势^{[ 36]}。对于自然条件严苛的黄土高原来说,植被恢复直接影响着区域水土流失强度与土地生产能力,因此植被生长状况可以作为当地生境质量改变的“指示剂”,选择其作为黄土高原区划工作的依据并进一步参与分区方案优选,所得到的黄土高原生态地理分区不仅具有传统自然区划的科学内涵,而且能体现研究区特殊的生态现状和政策背景;而通过详细了解各生态小区的植被恢复空间分异特征,也将对退耕还林等综合治理项目在黄土高原的进一步实施产生促进作用,对于因地制宜地进行生态恢复与规划管理也有着十分重要的意义。

因此,根据黄土高原的自然基底及土地覆被特征,选取气温、积温日数、太阳辐射、降水、干燥度、NDVI、海拔及植被覆盖等生态地理指标,基于自组织映射网络（SOFM）与GIS的空间分析及制图功能,对黄土高原进行生态地理分区;基于植被动态一致性准则对多种分区方案进行对比,采用NPP判定最优分区方案;并将分区结果与郑度等的相关区划方案^{[ 7, 22]}以及黄土高原等温线、等降水量线、山脉水系图等对照,以期探索基于植被动态一致性准则的生态地理分区方案优选方法,并为黄土高原生态环境地域分异研究提供基础数据。

2 研究区概况与数据来源

2.1 研究区概况

黄土高原地处半干旱半湿润气候带,水土流失严重,生态环境脆弱,是黄河泥沙的主要来源,也是中国水土保持重点区域^{[ 37]}。黄土高原位于中国地势第二级阶梯,地处中国中部偏北,形状略呈矩形,约处于北纬34°~41°N,98°~114°E。包括宁夏与山西全境、陕西中北部,甘肃的陇中和陇东地区,青海东北部以及内蒙古的河套平原和鄂尔多斯高平原,总面积达64.40×10⁴ km²（图1）。

显示原图| 下载原图ZIP| 生成PPT

图1 黄土高原地理位置及地形

Fig. 1 The location and topography of the Loess Plateau

黄土高原海拔1200~1600 m,地势南高北低且沟壑纵横,绝大多数地表被深厚的黄土层覆盖,具有塬、梁、峁等典型黄土地貌,土层最厚可达150~180 m。黄土高原的气候特征具有明显的地带性^{[ 38]},气温大致随着纬度升高与地势抬升逐渐降低,降水的分布也依循这一趋势,且具有强烈的季节变化和年际变化,6-9月降水占全年的70%~80%。总体来说,黄土高原属于中国东部湿润区与西北干旱半干旱区的过渡地带,由于区内地形、季风以及距海洋的远近不同,其内部在生物气候上存在着复杂的地域分异。

对于生境恶劣的黄土高原来说,植被覆盖对其可持续发展具有至关重要的作用,在1999年退耕还林（草）政策大规模实施以前,黄土高原植被覆盖以小幅波动为主,此后NDVI显著增强^{[ 39]};并且,通过耕地向林地、草地等其他用地类型的转化,黄土高原景观生态结构均得以优化^{[ 40]}。同时,植被恢复直接使得黄土高原地区水土流失、土地退化等问题得到显著缓解。因此,基于植被多年变化的动态一致性准则进行生态地理分区方案优选,有助于进一步了解黄土高原特殊的生态基底与政策背景。

2.2 数据来源

主要数据来源及处理方法：① 气象数据来自中国气象科学数据共享服务网（http://cdc.cma.gov.cn/）提供的全国756个气象站点1982-2010年的气温和降水量数据,为保证数据精度,选取黄土高原及周边共170个站点的数据;根据林忠辉等对气象要素插值方法的对比结果^{[ 41]},采用ArcMap 10.2中的Kriging法进行插值;② 全国范围1 km分辨率DEM数据来源于“黑河计划数据管理中心”（http://westdc.westgis.ac.cn）;③ 1982-2010年逐月NDVI数据来自寒区旱区科学数据中心（http://westdc.westgis.ac.cn/data/1cad1a63-ca8d-431a-b2b2-45d9916d860d）的NOAA/AVHRR-GIMMS数据集;④ 全国1 km分辨率植被类型数据来自SPOT4卫星GLC2000数据集（http://westdc.westgis.ac.cn/data/f1aaacad-9f42-474e-8aa4-d37f37d6482f）,中国植被共分为22类,黄土高原地区共有17类;⑤ 土壤类型数据来自地球系统科学数据共享平台（http://www.geodata.cn/Portal/metadata/viewMetadata.jsp?id=100101-10641）的全国1 $\begin{matrix} ∶ \end{matrix}$ 400万土壤类型分布图,分类方法依据中国1990年土壤系统分类方案。所用栅格数据统一具有GCS-WGS-84地理坐标系下的经纬度信息。

3 研究方法

3.1 生态地理分区指标体系

生态地理分区由自然分区演化而来,指标体系在原有基础上有所继承和发展。生态地貌特征作为分区的重要基础^{[ 42]},生物与土壤特征是其在地表的直观表达,这又在根本上决定于自然基底的水热条件,因而生态地理区的初级单位划分通常选择温度、水分、地形等作为划分依据;在高级单位的划分时更多参考植被等生态指标。综合考虑指标的合理性以及数据的时效性,选择四大类9个指标作为黄土高原生态地理分区的基础指标。

（1）热量类指标。具体包括日均温≥10℃的年积累日数、多年1月平均气温、多年7月平均气温、年均太阳辐射等4个指标。热量类指标主要用于黄土高原温度带的划分,其核心指标为日均温≥10℃的年积累日数。因为从生物学的角度出发,温度≥10℃是大多数植物和农作物正常生长的保障,可将≥10℃的天数作为植被的生长期。此外,为反映各地极端温度对黄土高原近30年来植被生长的影响,选择最冷月（1月）和最暖月（7月）平均气温作为辅助指标^{[ 22]}。太阳辐射指标对植被生长和干物质的积累有着重要作用,并已有学者证实植被NPP与太阳辐射呈显著正相关^{[ 43]}。

（2）水分类指标。具体包括年均降水量、年均干燥度等两个指标。水分与热量条件共同决定自然地表区域性差异。在水分类指数中,年降水量能够反映区域水分的补给状况,200 mm、400 mm、800 mm等降水量线也被认为是划分干旱地区、半干旱地区、半湿润地区、湿润地区的分界线。而干燥度作为划分研究区干湿区的核心因子,反映了区域内水分收支状况,并间接反映了除蒸发以外可用于植物生长的水资源量,一般使用潜在蒸散和年降水量的比值得到^{[ 44]}。在具体计算时,潜在蒸散ET₀使用联合国粮农组织（FAO）于1998年修正的Penman-Monteith公式计算^{[ 45]},具体公式如下所示：

式中：净辐射量R_n（MJ/m²）是ET₀重要的计算基础;D是饱和水汽压曲线斜率（kPa/℃）;G为土壤热通量（MJ/m² $\begin{matrix} \cdot \end{matrix}$ d）;g为干湿常数（kPa/℃）;T为平均气温（℃）;U₂为校正后的2 m高处风速（m/s）;e_s为某时段内平均饱和水汽压（kPa）;e_a为实际水汽压（kPa）。

（3）地形类指标。主要是DEM表征的海拔。地形是生态地理系统中的重要组成部分,黄土高原广泛分布有塬、梁、峁等典型黄土景观,地势起伏悬殊,由DEM数据所表征的海拔不仅直接影响着温度和水分的地域分异特征,而且在一定程度上控制着植被、土壤的水平和垂直分布,进而在不同尺度上决定着当地的植被类型与长势,因此在进行黄土高原自然分区时必须将海拔高度作为主要的地形类指标。

（4）地表覆被类指标。具体包括年均NDVI和土壤类型两个指标。土壤作为连接大气圈、岩石圈、水圈与生物圈的关键角色,其质地直接决定了上覆植被的类型;并且退耕还林政策的顺利实施也必须考虑所选育树种、草种对于当地土壤的适宜性。而NDVI数据是黄土高原30年来植被演替的最直观表征,因此,天然植被作为黄土高原生态地理分区的重要指标,直接指示了不同下垫面与水热条件下生态恢复的波动状况^{[ 42]}。

以上四类指数分别指代不同的自然地表特征,能够全面反映1982-2010年间黄土高原地区生态与气候条件的总体状况,将观测数据不全的站点剔除相应的年份,以各指标的多年平均值作为最终的生态地理分区依据。此外,除以上9种指标以外,选取GLC2000植被类型数据集对黄土高原各分区的植被类型进行统计,选取所占面积比重最大的植被作为该分区的主导类型,进而对各分区结果进行命名。

3.2 生态地理分区方法

在分区方法的选择方面,中国现有生态地理分区研究大都选择自上而下的演绎途径从高到低予以划分^{[ 2, 42, 46, 47]},而传统分区方法大都有分区过程复杂、耗时、主观性强的缺点。本文使用SOFM神经网络模型实现自下而上的分区,该方法所得到的分区单元界线更加精确并且可靠性较高^{[ 4]}。

SOFM神经网络又称自组织特征映射网络,属于人工神经网络的一种^{[ 48]}。包括竞争层和输入层两部分^{[ 49]},其中输入层用于接收样本,竞争层对样本进行分类并输出分类结果。SOFM神经网络可以直接对未知的输入数据进行学习或模拟,而并不需要进行先验学习^{[ 50]};因此与传统的三级分区方案相比,SOFM所形成的分类中心能保证原数据的拓扑不变性,是一种非监督型的模式识别技术,可以克服传统专家经验集成的区划范式下主观性强的缺陷。基于计算机强大的计算能力,该方法也能够快速把握区域的分异特点。SOFM编程及数据处理在专业数学软件MATLAB R2014a中完成,根据四类指标构建9×9297输入层数据,初始权值为[0,1]的随机数,基本学习速率为0.1,最大循环次数1000次,其他参数采用默认值。

为提取足够的样本作为SOFM网络的输入数据,借助ArcMap 10.2的Random Point工具基于9组数据集生成随机点（以3 km作为半径共提取10000个随机点,排除无效值后共9297个）,并提取随机点位的相应数值。由于上述9项指标具有不同的量纲,选用极差标准化方法对数据进行归一化处理。处理后的数据平均值为0,且所有数据的数值在（-1,1）之间,其计算公式如下：

式中： $\begin{matrix} x_{i}^{*} \end{matrix}$ 表示标准化后的数据; 表示平均值。

由于黄土高原地区已有分区方案^{[ 5, 7]}的总体分区均在5个以上,考虑到区域空间分异表征准确性及其刻画细致性的需要,依据植被动态一致性准则对SOFM网络的11种分区方案（分别划分出5~15类生态地理区）进行筛选。植被动态一致性是指所识别的各个分区之间即使在气温、降水、地形、下垫面甚至植被类型等方面均有差异,但经过黄土高原长时间的植被演替以及1999年退耕还林之后,各分区的植被生长是以相同趋势发展的。植被动态一致性准则从根本上体现了黄土高原近30年来的植被生长特征,而植被的净初级生产力（net primary productivity,NPP）又是NDVI与气温、降水等自然要素综合作用的结果^{[ 51]},因此,将其作为多分区方案优选的关键指标。NPP采用基于光能利用率的CASA模型计算得到,数据源为1982-2010年GIMMS-NDVI数据和气象观测数据。

具体来说,多分区方案的优选主要分为两步进行：第一步,利用不同分区方案下各分区的NPP平均值动态进行初步筛选。以时间为横坐标分别绘制各分区NPP平均值的多年变化趋势图,若各分区曲线相交一次则记为一次异常,即该分区方案并不能够客观而清晰地解释当年的植被地域分异规律。总体来看,29年间异常次数越少分区准确性越强,并且折线在多年间应以相同趋势增或减,以此方法选出相交次数较少的3~4个方案;第二步,针对初选出来的分区方案,利用各分区逐栅格多年平均NPP变异系数（coefficient of variation,CV）进行再筛选：CV用于表征各分区内NPP数值的离散程度,离散程度越小则说明分区内部相似性越高,分区之间差异越大;所有分区离散系数的平均值最小,所得到的分区方案越合理。由于各方案的具体分区布局不同,各区CV值并不可直接对比;因此对各方案离散系数的算术平均值（ $\begin{matrix} \bar{CV} \end{matrix}$ ）及按像元个数加权的平均值（ $\begin{matrix} C V^{*} \end{matrix}$ ）进行了计算, $\begin{matrix} \bar{CV} \end{matrix}$ 和 $\begin{matrix} C V^{*} \end{matrix}$ 分别代表了黄土高原全局尺度及像元尺度的多年平均NPP的离散程度。其中, $\begin{matrix} C V^{*} \end{matrix}$ 的计算公式如下：

式中：n代表分区个数;CV_i代表该分区的CV值;N_i代表该分区的像元个数。

4 结果分析

4.1 黄土高原生态地理要素空间分异

为更加直观地了解黄土高原水热环境与生态资源空间配置现状,基于研究区1982-2010年的多期遥感数据和气象站点观测数据,利用ArcMap 10.2软件的空间分析与制图功能对3.1中所选择的9种生态地理要素进行空间制图（图2）,从中可发现各指标均呈现十分显著的空间分异,并且热量类指数、水分类指数、地形指数与地表覆被指数之间格局差异较大。

显示原图| 下载原图ZIP| 生成PPT

图2 黄土高原生态地理指标的空间分异

Fig. 2 The eco-geographical indexes in the Loess Plateau

从热量类指数看（图2a~ 图2d）,黄土高原东南部地区总体较为温暖且太阳辐射充足,而西北部地区则与之相反。并且,黄土高原气温的季相差异明显,冬季（图2b）最低温一般出现在青海东部、内蒙古呼和浩特市与巴彦淖尔市境内,1月最低温仅有-14℃左右,陕北榆林和甘肃兰州、固原等地冬季气温也相对较低（-8℃~-6℃）;而关中地区与河南省西北部的冬季气温则相对最高,最低温在0℃以上。夏季（图2c）较低气温出现在甘肃南部和青海东部,夏季均温仅为12℃~15℃左右,最高温仍分布于西安市等南部地区,夏季高温可达30℃。黄土高原的气温产生季相差异的主要原因是不同地区的地表覆被与地形差异显著,由此影响了冬夏季的地表比热与局地小环境。值得注意的是积温日数指标,由于当日均温稳定升至10℃以上时,喜凉作物开始迅速生长,喜温作物也开始播种,因此其在农业资源评价中有重要作用^{[ 7]}。由图2a中可以看出,黄土高原东南大部分地区的日均温 $\begin{matrix} \geq \end{matrix}$ 10℃的天数可达200天以上,这些地区的作物可达两年三熟或一年两熟,属于粮食产量较高的地区;而黄土高原西部积温日数则普遍较低,最低地区每年日均温 $\begin{matrix} \geq \end{matrix}$ 10℃的天数仅占全年的20%左右,这些地区的作物多为一年一熟。

干燥度（图2e）和年均降水（图2f）作为水分类指数,在分区中主要用干湿地区命名。黄土高原地区以其水土流失和干旱少雨广为人知,全区多年平均降雨量约为466 mm,并呈现从东南向西北递减的趋势,极低值多出现在内蒙古乌兰布和沙漠及毛乌素沙地,极高值出现在河南洛阳及关中地区。对于干燥度指数来说,其本质上代表着某一地区水分支出与收入的比值,数值越大代表该地越干旱。一般来说,当干燥度 $\begin{matrix} < \end{matrix}$ 1.0时为湿润地区,[1.0,1.7)和[1.7,3.0)分别为半湿润及半干旱地区, $\begin{matrix} \geq \end{matrix}$ 3.0则为干旱地区^{[ 44]};从图2e中可知,黄土高原干燥度在1.46~6.95之间,因此跨越了半湿润、半干旱、干旱三个干湿区,总体呈现东西方向的条带状分布,从具体数值来看,干燥度<1.7的地区多位于黄土高原东南部,而西北地区则最为干旱,干燥度大多 $\begin{matrix} \geq \end{matrix}$ 3.0。对干燥度和降水量的划分对黄土高原最终的分区命名起到了重要的作用。

地形和地表覆被决定了黄土高原生态系统自然景观的差异性,其中土壤（图2g）和地形（图2i）直接决定了所生长植被的种类与长势。总体来说,黄土高原的土壤以壤质普通黄绵土为主,广泛分布于山西太原至甘肃定西一带;该土种的主要特征是剖面发育不明显,土壤侵蚀较为严重,但可耕作性好,适宜长期耕种^{[ 52]}。黄土高原总体呈西高东低的地势特征,黄河流域沿线海拔最低,总体分为陇中高原、陇东陕北高原、山西高原及渭河平原。由此可知,黄土高原无论在下垫面条件抑或是气候水热条件方面都有着极大的区域差异,不同的土壤和坡度也导致了退耕还林过程中因地制宜的植被恢复模式。NDVI作为植被生长状况的最直接表征,在水热充足、地势平缓的东南部呈现出较高的植被覆盖度,延安的子午岭山区也有着大片密集的森林资源;植被覆盖较低的地区主要分布在黄土高原北部的干旱沙地,而河套平原则因相对充足的水资源而成为了北部的“绿岛”。

4.2 黄土高原生态地理分区方案优选

基于SOFM网络将极差标准化后的9项评价指标划分为5~15类,初步筛选所绘制的各分区NPP均值多年变化趋势如图3所示。

显示原图| 下载原图ZIP| 生成PPT

图3 黄土高原生态地理分区方案对比：不同分区NPP平均值年际变化 (a)~(c),(e)~(l)分别为5~15分区方案,(d)为修正6分区方案

Fig. 3 The interannual variation of NPP among the multi-regionalization plan of the Loess Plateau (a)~(c), (e)~(l) represent for 5 regions plan to 15 regions plan respectively, (d) represents for the modified 6 regions plan

根据统计结果可知,随着分区个数的增加,相交折线和交点数目显著增多,相应的折线重合度升高。其中5~7分区方案的交点个数最少,分别为10个、8个与4个,其余分区方案的折线交点数均在10个以上,13~15分区中甚至达到了50个以上（图3j~ 图3l）,并且各分区均表现出了较强的不稳定性,曲线重叠严重、变化趋势杂乱。通过观察图3a和图3c可知5分区和7分区方案仅有两条折线出现相交现象（分别为4区和5区、6区和7区）,其他分区均表现出较好的独立性;而6分区方案虽然仅有8个交点,但1区和2区的折线变化趋势并不十分一致。综合以上情况,按照交点个数最少原则可知在黄土高原的11个分区方案中的5~7分区方案都是比较合理的（图4a~ 图4c）。其中,7分区方案中的6区与7区折线的重合度极高,说明这两个分区并不能独立表现植被的动态变化特征,可以予以合并,进而得到更为优化的分区方案（图4d）,将其命名为“修正6分区方案”。通过分析新分区方案的NPP均值多年变化状况（图3d）可知,6条折线变化趋势相近且并无相交现象,说明各分区独立性较高且多年植被生产力也以相同趋势增加与衰减。

显示原图| 下载原图ZIP| 生成PPT

图4 黄土高原生态地理最优分区方案初选结果

Fig. 4 The first step of comparison and screening of the optimal regionalization plan of the Loess Plateau

在初步筛选后需要借助各分区NPP的离散程度对图4所示的四种方案进行再次筛选,对比各方案的变异系数（图5）可知,5分区方案的第5区与其他3种方案的第6区多年NPP离散度最大,该分区正好位于地表覆被差异极大的西北部,常年飞沙的毛乌素沙地与水土丰泽的河套平原共居其中,因此直接造成了NPP的显著离散。并且,在四种分区方案中 $\begin{matrix} \bar{CV} \end{matrix}$ 均呈现“修正6分区<6分区<7分区<5分区”的特征,而 $\begin{matrix} C V^{*} \end{matrix}$ 则略有不同,“修正6分区<7分区<5分区<6分区”;通过对两个指标的比较均可发现,7分区合并之后的新方案NPP离散程度最小。由于离散系数能够客观地表达一组数据距离其平均值的波动程度,数值越大说明数据集越分散,内部统一性越差;而数值越低则说明数据的团聚性越强,各分区内多年平均NPP越相似。综上所述,可以认为,修正6分区方案是基于植被动态的理想黄土高原生态地理分区方案。

显示原图| 下载原图ZIP| 生成PPT

图5 黄土高原生态地理分区初选方案年均NPP变异系数CV值对比

Fig. 5 The comparison of the CV value of NPP in each region between 5-7 regions and modified 6 regions

4.3 黄土高原生态地理分区方案

利用非监督性质的SOFM神经网络将黄土高原生态系统分为六大区域,参照中国传统的气候区划方法^{[ 7, 53]}使用三级划分法进行命名。需要注意的是,传统的分区方法是“逐级推进、层层嵌套”的模式,即首先采用温度指标命名温度带,在温度带的内部用水分类指标划分干湿地区,再在干湿区内部进一步划分自然区^{[ 7, 54]}。在分区时采用SOFM神经网络进行各指标的统一划分,仅使用三级体系进行最终命名。

在分区结果的基础上,分别以日均温 $\begin{matrix} \geq \end{matrix}$ 10℃的年积累日数、年均干燥度、植被类型数据为命名温度带、干湿区与自然区的核心因子,其他因子作为辅助指标。各指标的阈值主要参考《中国生态地理区域系统研究》^{[ 7]}与《中国自然地理系列专著：中国气候》^{[ 53]}。最终分区方案及命名结果如表1和图6所示：黄土高原六大生态地理分区中共有两个分区属于中温带,在空间上位于黄土高原西北部;有3个分区属于暖温带,主要位于黄土高原中部及东南部;而高原温带地区位于西部的青海和甘肃境内。从干湿区角度来看,研究区全区仅有东南端的分区IIA1属于半湿润气候,与此相对,西北端沙漠广布的IC1区则是仅有的干旱地区,中部的其他4个分区均属于半干旱气候。根据GLC2000数据对各分区主导植被进行统计可知,干旱少雨的IC1区仅分布有少量荒漠草原,而中部干旱程度稍弱的半干旱地区则以草原为主要植被类型,根据当地坡度、海拔等差异分别有着不同的优势草种：如从西北向东南分别生长有平原草原、坡面草原、高山亚高山草甸,雨水最为丰沛的半湿润IIA1区则以相对优越的自然条件滋养了落叶阔叶林等高大乔木。此外,除天然植被以外,在暖温带的3个分区中分布有较大面积的耕地。总体来说,黄土高原干旱少雨、水土流失较为严重的生态背景更加适宜草类植被的生长,从另一个方面可以推知草类较之林木有更强的耐旱能力,适宜种植的范围更广。

表1 黄土高原生态地理分区命名体系

Tab. 1 Naming system for eco-geographical regionalization of the Loess Plateau

温度带		干湿地区		自然区
I	中温带	B	半干旱	1	平原草原区
I	中温带	C	干旱	1	荒漠及荒漠草原区
II	暖温带	A	半湿润	1	落叶阔叶林耕作区
		B	半干旱	1	平原草原耕作区
		B	半干旱	2	坡面草原耕作区
III	高原温带	B	半干旱	1	高山、亚高山草甸区

新窗口打开

显示原图| 下载原图ZIP| 生成PPT

图6 黄土高原生态地理分区方案

Fig. 6 Eco-geographical regionalization of the Loess Plateau

为进一步探究各生态指数在不同分区中的数量分布状况,本文基于ArcMap 10.2的分区统计工具对各分区的评价指标平均值进行了统计（表2）。一方面对于热量类指数而言,低纬地区且海拔较低的分区IIA1虽然所接收的太阳辐射量并不高,但全年 $\begin{matrix} \geq \end{matrix}$ 10℃的积温日数可以达到200天以上,且最冷、最暖月的平均温度都为6个分区中最高;相反,位于黄土高原最西端的IIIB1分区热量类指数都相对较低,结合地形指数分析可知,该地区处于黄土高原海拔最高的地区,因而气温会随着海拔升高发生了梯度性下降。另一方面对于水分类指数而言,干燥度与降水量相辅相成,共同决定了一个地区的水分收支平衡;黄土高原六大分区从西北向东南整体呈现干燥度降低、降水量升高的现象。将各分区水热条件与海拔、地形等要素结合可知,低海拔、气候湿润温暖的平原地区相应的植被覆盖也最良好^{[ 5]},因此黄土高原的NDVI高值区和主要的林区都分布于IIA1区内。

表2 黄土高原生态地理分区评价指标平均值统计

Tab. 2 Average values of the indexes in each eco-geographical zone of the Loess Plateau

			生态地理分区
指数类型	具体指标	IC1	IB1	IIB1	IIB2	IIA1	IIIB1
热量指数	≥10℃积温日数	185	179	183	179	209	146
	太阳辐射（MJ）	5890.61	5675.22	5444.72	5291.93	4717.69	5631.00
	1月均温（℃）	-8.61	-8.63	-8.52	-7.36	-2.41	-7.31
	7月均温（℃）	23.84	23.12	23.36	21.86	24.56	17.26
水分指数	干燥度	5.09	3.46	2.78	2.32	1.8	1.81
水分指数	年均降水（mm）	239.68	325.68	385.96	423.92	546.82	459.61
地表指数	NDVI	0.21	0.23	0.29	0.39	0.55	0.45
地表指数	主导土壤类型	普通灰钙土	砂壤质普通黄绵土	壤质普通黄绵土	壤质普通黄绵土	普通塿土与壤质普通黄绵土符合土	普通寒毡土
地形指数	DEM（m）	1420.55	1510.89	1356.43	1758.92	988.76	2857.23
植被指数	主导植被类型	荒漠草原	平原草原	平原草原及耕地	坡面草原及耕地	落叶阔叶林及耕地	高山亚高山草甸

新窗口打开

（1）IC1：中温带干旱荒漠及荒漠草原区。该区包括河套平原广大地区,黄河上游流经于此,区内地形指数较低,年最大温差可达32℃左右,平均降水量仅有239.68 mm,属于6个分区中最低,因此该分区整体气候干热;该地广泛分布有沙地,著名的毛乌素沙地即位于此,干旱少雨的气候条件和下垫面决定了其稀疏的地表覆盖,荒漠草地成为该地的优势草种,但是在水分丰沛的河套平原地区则分布有大片耕地。该区的主要土种为普通灰钙土,与其他区域有较大区别。

（2）IB1：中温带半干旱平原草原区。该分区在黄土高原6大分区中所占面积最小,由东北向西南贯穿鄂尔多斯、陕北榆林、宁夏及甘肃兰州。该区气候较为寒冷干燥,1月均温为6个分区中最低,干燥度仅次于干旱区的IC1地区,年均降水量325.68 mm;平均海拔相对较高,因此主要植被类型多为典型平原草原以及少量草甸和坡面草地,在该区南部海拔较高地区分布着一部分耕地,由于其生态系统植被类型以草类为主,因此该区的NDVI并不高,仅仅略优于沙漠广布的分区IC1。

（3）IIB1：暖温带半干旱平原草原耕作区。本区位于内蒙古呼和浩特、陕北榆林及延安、甘肃庆阳及兰州的大部分地区;热量类指数与相邻的IB1分区较为相近,但气候更加湿润。该分区位于黄土高原的中央位置,地势平坦、海拔较低,植被覆盖类型以平原草原为主,在甘肃庆阳北部生长有密集的坡面草地,在吕梁山山脉以西分布有大量的耕地,主要原因是该地跨越黄河干流及多条支流,水分供应充足。整体植被覆盖状况优于分区IB1,全区的多年NDVI平均值可达0.3左右。

（4）IIB2：暖温带半干旱坡面草原耕作区。该区东西跨度最大,横跨6个省市自治区,最主要面积位于山西境内的吕梁山山脉。气温及所接受的太阳辐射都比IIB1区略低,降水却比IIB1有较明显增长,因为包含吕梁山和乌鞘岭两大山脉,故其平均海拔仅次于IIIB1区。起伏度较大的地形决定了坡面草地成为其主导的植被类型,在该区的中东部也成为黄土高原主要的农业用地与粮食产区。

（5）IIA1：暖温带半湿润落叶阔叶林耕作区。本区位于黄土高原东南部,在6个分区中面积最大。除黄河下游外,渭河、汾河、泾河以及洛河均流经于此,地表水极为丰富,年均降水量也为全区最高（546.82 mm）。气候暖温,年均无霜期在200天以上,冬季气候也比较温和;海拔较低,地形大都为平原,适宜作物种植。故该地为黄土高原最大的耕地与针阔叶林分布区,其暖湿气候及优势地形下的高植被覆盖（年均NDVI约为0.55）使其成为黄土高原生物气候最优的地区。

（6）IIIB1：高原温带半干旱高山亚高山草甸区。该分区在6个分区中地势最高,平均海拔可达2857.23 m;主要位于甘南藏族自治州、海北藏族自治州以及祁连山南麓,在研究区内被IIB2分割为两个小区。该区气温处于6大分区的最低水平,每年适宜植被生长的无霜期只有近150天,全年最暖月的平均温度不足20 ℃,全年温差较小;然而该区水分类指数较高,年均降水可达459.61 mm,总体属于寒冷湿润的气候。在地表覆被方面,该分区植被类型较为单一,土壤类型以普通寒毡土为主,全区广泛均分布着高山亚高山典型耐寒草类品种,因此该区多年NDVI平均值仅次于IIA1区。

5 讨论

5.1 黄土高原生态地理分区方案对比

为讨论基于植被动态一致性准则的黄土高原生态地理分区的合理性,将区划结果与郑度等^{[ 7]}2008年公布的中国生态地理区域系统黄土高原部分研究结果进行对比（图7）,并对分区面积的一致性进行统计（表3）。从图7中可以看出,在郑度等的区划方案中,黄土高原共分为三个温度带（中温带、暖温带、高原温带）、三个干湿区（半湿润区、半干旱区、干旱区）和该分类系统下的7个自然小区,总体呈现由东北向西南的条带状分布,本文分区的空间分布与这一特征极为符合;其次,本文对中温带和暖温带边界的识别比较合理,与郑度等的研究结果基本相似,尤其是本研究中温带干旱荒漠及荒漠草原区（IC1）与中温带半干旱平原草原区（IB1）的边界与其对于中温带干旱地区鄂尔多斯及内蒙古高原西部荒漠草原区（IID1）与中温带半干旱地区内蒙古东部草原区（IIC3）的交界线非常接近,面积重合度达78.92%。

表3 黄土高原分区面积一致性统计(%)

Tab. 3 Statistics of consistency ratio in area of the Loess Plateau regionalization(%)

		黄土高原生态地理分区
		IB1	IC1	IIA1	IIB1	IIB2	IIIB1
生态地理区域系统^{[ 7]}	IIC3	65.16	2.66	0	34.80	10.64	0
	IID1	4.07	78.92	0	0	0	0
	IID2	3.08	6.50	0	2.80	2.23	0
	IIIB3	0	0	27.33	0	6.91	0
	IIIB4	0	0	72.67	0	6.23	13.60
	IIIC1	27.69	11.92	0	62.40	55.92	21.57
	HIIC1	0	0	0	0	18.07	64.83

新窗口打开

显示原图| 下载原图ZIP| 生成PPT

图7 黄土高原生态地理分区与郑度等的分区方案对比

Fig. 7 The eco-geographical regionalization overlay with Zheng's of the Loess Plateau

另外,中温带干旱荒漠及荒漠草原区（IC1）实际包含了生态地理区域中的中温带干旱地区阿拉善及河西走廊荒漠区（IID2）和鄂尔多斯及内蒙古高原西部荒漠草原区（IID1）两部分,IC1区与IID2、IID1的重合度达到85.42%。与此相近,IIA1分区实际也包含了生态地理区域系统中的IIIB3和IIIB4分区,这两个区域同属暖温带半湿润地区,与IIA1相同（IIA1区的范围完全落于IIIB3和IIIB4之内）,都以落叶阔叶林为主导植被类型。对于横跨东西、形状狭长的暖温带半干旱坡面草原耕作区（IIB2）与生态地理区域系统中的6个分区都有重叠,不确定性相对较高,但其主体面积仍主要与IIIC1重合,都属于暖温带半干旱地区,植被类型均以草地为主导。通过表3可知,本文所划6大分区与各自重合度最高的已有分区结果均吻合,所产生差异的原因主要是区划尺度的不同,生态地理区域系统的划分是着眼于全国的,而本文则基于中尺度的黄土高原,划分结果不可避免会有所不同,但是总体空间分异格局都是基本吻合的。

除与已有分区方案进行对比以外,将分区结果与黄土高原年均温等值线图、年降水量等值线图以及山脉水系图等对照^{[ 55- 58]},结果发现：本文分区方案对于水分类指数与热量类指数的分区效果较好,在各区域间体现出了明显的梯度变化性;并且将六大分区在温度带上总体划分为暖温带、中温带与高原温带的分类方法较为合理,符合黄土高原由东南至西北的气候变化特征。结合黄土高原年降水量等值线图发现：划分的三类干湿区与该地的400 mm、200 mm等降水量线由东南向西北延伸的趋势相符。总体而言,通过与已有黄土高原自然要素图件对比,本文的分区方案基本反映了黄土高原的自然地理界线。

5.2 基于植被动态的生态地理分区方法适用性

生态地理分区及区划是地理学综合研究的核心主题,长期以来受到地理学者广泛关注。区划或分区要素在区域之间差异最大、区域内部差异最小,这是生态地理分区或区划有效性的基本要求。在中国早期的生态区划和自然区划工作中,黄秉维等^{[ 2, 59- 61]}多根据已有生态地理要素图件及研究者对于地理环境的直观认识与经验积累进行;近年来,随着计算机技术的发展,SOFM神经网络成为生态地理分区的重要定量方法^{[ 24, 29]}。而由于SOFM网络自身的特性分类数目的改变会显著影响分区结果,生态地理分区方案的对比与优选因而成为基于SOFM神经网络的生态地理分区关键环节。但是,既有的生态分区、地理区划及土地利用分区等相关研究,大多依据研究区空间分异的主观感知,基于不同方案下的分区图直观判定;仅有部分学者尝试采用聚集度等景观格局指数定量对比不同类型分区的空间聚集程度,基于整体聚集度高低筛选最优分区方案^{[ 28]};但这种最适分区判定仅仅考虑了空间形态,未能关涉分区核心要素在不同分区方案之间的对比。因此,本研究提出基于植被动态一致性准则的生态地理最适分区判定方法,以植被动态作为生态地理要素空间分异的地表景观综合判定依据,有助于提升生态地理分区的客观性与准确性。

不过,需要指出的是,由于这一方法假定植被动态一致的样点具有相似的生态地理要素特征,亦即是说植被动态主要受生态地理因子制约。因此,人类活动等非生态地理因素对研究区植被动态的影响不予考虑;而要达到这一要求,或者人类活动对植被动态影响极为微弱,或者人类活动影响在研究区没有显著的空间差异。对于黄土高原来说,该地区近十年间最大的生态演变特征便是退耕还林还草等生态建设作用下的全域植被恢复,区域内部植被动态差异基本反映了生态地理要素的空间分异。同时,这一基于植被动态的生态地理分区优选方法,其适用性也受研究区空间异质性的本底特征影响。在表现出典型干湿地带性的黄土高原,该方法表现出对中等程度异质性的良好表征效果。而对于高度异质性的复杂地域而言,考虑到其下垫面、气候条件及植被覆盖的多样性及破碎化格局,在使用本文所提出的多方案优选方法时,更细化的分区方案对真实的生态环境将有更好的反映。对于环境构成更为单调的区域,因其自然本底特征与覆被状况的均质性,从生态地理分区或区划的视角来看其本身在某种程度上已成为不可再分的整体,分区效果并不一定更优。

6 结论

以黄土高原多年来的植被生长与恢复为重点,基于植被动态一致性准则讨论了生态地理分区多方案优选的新方法。在定量统计四大类9个综合自然指数的基础上,将其作为分区指标体系输入SOFM神经网络模型,对黄土高原用地类型进行了自下而上的合并;并借助NPP数据对多种分区方案进行讨论,进行最优分区方案筛选,最终将黄土高原划分为六大生态地理分区。研究表明：

（1）研究区可以划分为：中温带干旱荒漠及荒漠草原区（IC1）、中温带半干旱平原草原区（IB1）、暖温带半干旱平原草原耕作区（IIB1）、暖温带半干旱坡面草原耕作区（IIB2）、暖温带半湿润落叶阔叶林耕作区（IIA1）以及高原温带半干旱高山亚高山草甸区（IIIB1）六大生态地理区。根据分区特点能够显示出黄土高原地区的气候要素及植被覆盖的条带性分布,较好地揭示出该地区的自然地域分异特征。

（2）通过将SOFM神经网络分类结果与郑度等^{[ 7]}的全国区划及黄土高原年均温等值线图、年降水量等值线图以及山脉水系图等对照,以此对SOFM网络的分类结果进行验证。对比结果显示,SOFM神经网络聚类方法所得的自然分区结果与黄土高原自然生态系统的真实分异特征具有较好的一致性,且划分层次明确,区域分割清晰,结果可信。

（3）对于典型的植被恢复带黄土高原而言,本文的生态地理分区抓住了该地独特的生态变化及政策背景,并且有针对性地对多种分区方案进行了筛选及讨论,使所得结果更具实用性。分区结果能够达到分区内部植被多年变化差异最小,分区之间特异性最大的目标,对于黄土高原未来的退耕还林以及生态管护都有着较大的借鉴意义。

论文仍然存在一些不足。基于SOFM网络的区划技术虽然具有诸多优点,但是其作为一种非监督型的识别技术,在具体的运算过程中具有对操作过程以及结果的不可控性。根据研究者输入的特定区划类别,SOFM网络均能得出相应的分区方案,因而容易忽略生态系统的真实区划状况。这也从另一方面反映了多方案优选的重要性。此外,在SOFM网络构建时,维数大小的组成方式、网络拓扑结构等的选择和设置,都一定程度上受到研究者的主观干扰,如何避免该方面的影响则需要在后续的工作中进行进一步的探究。

The authors have declared that no competing interests exist.

参考文献

原文顺序

文献年度倒序

文中引用次数倒序

被引期刊影响因子

[1]	郑度, 葛全胜, 张雪芹, 等. 中国区划工作的回顾与展望 . 地理研究, 2005, 24(3): 330-344. https://doi.org/10.3321/j.issn:1000-0585.2005.03.002 URL [本文引用: 1] 摘要在回顾国内外区划工作基础上,总结了我国区划工作各个发展阶段的特点,简单概述了我国区划工作中的方法论,讨论了区划工作中存在的若干问题.20世纪后半叶起,我国的区划研究进入系统研究和全面发展期.这期间我国区划工作的主要特点可归纳为: (1)众多区划方案的提出都有其深刻的历史背景,既是 [Zheng Du, Ge Quansheng, Zhang Xueqin, et al. Regionalization in China: retrospect and prospect. Geographical Research, 2005, 24(3): 330-344.] https://doi.org/10.3321/j.issn:1000-0585.2005.03.002 URL [本文引用: 1] 摘要在回顾国内外区划工作基础上,总结了我国区划工作各个发展阶段的特点,简单概述了我国区划工作中的方法论,讨论了区划工作中存在的若干问题.20世纪后半叶起,我国的区划研究进入系统研究和全面发展期.这期间我国区划工作的主要特点可归纳为: (1)众多区划方案的提出都有其深刻的历史背景,既是
[2]	黄秉维. 中国综合自然区划的初步草案 . 地理学报, 1958, 24(4): 348-365. URL [本文引用: 4] 摘要正综合自然区划是根据地表自然综合体(即自然的总情况)的相似性与差异性将地域加以划分,并进而按区划单位来认识自然综合体的发生、发展与分布的规律。认识自然综合体的目的是:(1)阐明自然资源与自然条件对于生产与建设的有利方面与不利方面;(2) [Huang Bingwei. A preliminary scheme of integrated physicogeographical zoning in China. Acta Geographica Sinica, 1958, 24(4): 348-365.] URL [本文引用: 4] 摘要正综合自然区划是根据地表自然综合体(即自然的总情况)的相似性与差异性将地域加以划分,并进而按区划单位来认识自然综合体的发生、发展与分布的规律。认识自然综合体的目的是:(1)阐明自然资源与自然条件对于生产与建设的有利方面与不利方面;(2)
[3]	高江波, 黄姣, 李双成, 等. 中国自然地理区划研究的新进展与发展趋势 . 地理科学进展, 2010, 29(11): 1400-1407. https://doi.org/10.11820/dlkxjz.2010.11.032 URL [本文引用: 1] 摘要 As is well known,there exist similarities and differences among geographical regions,and zonality is a universal theorem governing geographical patterns.Physio-geographical regional system is a hierarchical system,which is formed by division or combination of natural features based on geographic relativity and geographic zonality.Researches on physio-geographical regional system aim at observing and studying natural complex of the earth's surface,revealing rules of regional differentiations and exploring formation,development,divi-sion,combination,relativity and demarcation of natural regions at different scales,from a regional point of view. As a radical method to obtain geographical information,physio-geographical regionalization has always been studied and widely used by geographers.Studies on physio-geographical regionalization in China have made great progress in principle,methodology,and theory since the 1930s.It could provide a scientific basis for social and economic development and planning,and conservation,improvement and rational utilization of natural environment.Based on brief review of physio-geographical regionalization at home and abroad,the paper summarizes the characteristics of researches in recent years:various types of elements;distinct ecologicalization in comprehensive natural regionalization;a complete series of spatial scales;scarce creativity in technology and methods;emphasizing regionalization methodology;enhancing application of regionalization schemes.Then,the paper analyzes academic problems in the present researches:lack of epistemological study;weak methodological study.Finally,the paper explores the prospects of physio-geographical regionalization:continuous increase in requirement at the national level;strengthening the standardization of regionalization,and raises several trends in geographical regionalization. [Gao Jiangbo, Huang Jiao, Li Shuangcheng, et al. The new progresses and development trends in the research of physio-geographical regionalization in China. Progress in Geography, 2010, 29(11): 1400-1407.] https://doi.org/10.11820/dlkxjz.2010.11.032 URL [本文引用: 1] 摘要 As is well known,there exist similarities and differences among geographical regions,and zonality is a universal theorem governing geographical patterns.Physio-geographical regional system is a hierarchical system,which is formed by division or combination of natural features based on geographic relativity and geographic zonality.Researches on physio-geographical regional system aim at observing and studying natural complex of the earth's surface,revealing rules of regional differentiations and exploring formation,development,divi-sion,combination,relativity and demarcation of natural regions at different scales,from a regional point of view. As a radical method to obtain geographical information,physio-geographical regionalization has always been studied and widely used by geographers.Studies on physio-geographical regionalization in China have made great progress in principle,methodology,and theory since the 1930s.It could provide a scientific basis for social and economic development and planning,and conservation,improvement and rational utilization of natural environment.Based on brief review of physio-geographical regionalization at home and abroad,the paper summarizes the characteristics of researches in recent years:various types of elements;distinct ecologicalization in comprehensive natural regionalization;a complete series of spatial scales;scarce creativity in technology and methods;emphasizing regionalization methodology;enhancing application of regionalization schemes.Then,the paper analyzes academic problems in the present researches:lack of epistemological study;weak methodological study.Finally,the paper explores the prospects of physio-geographical regionalization:continuous increase in requirement at the national level;strengthening the standardization of regionalization,and raises several trends in geographical regionalization.
[4]	程叶青, 张平宇. 生态地理区划研究进展 . 生态学报, 2006, 26(10): 3424-3433. https://doi.org/10.3321/j.issn:1000-0933.2006.10.035 URL [本文引用: 3] 摘要生态地理区划是国内外地理学和宏观生态学研究的热点问题之一.国外学者对生态地理区划概念、原则和指标体系和方法等进行了大量的研究和讨论,并就全球、国家和区域尺度的生态地域系统进行划分.其研究经历了从单生态因子分异的规律性认识、生态系统聚合、人类作为生态系统一部分的单要素划分到层次约束、相互包容与不包容的等级斑块思维的整体性功能、过程和结构性划分的发展路线.我国生态地理区划在借鉴自然区划的基础上发展较为迅速,并在区划的原则、方法、理论等方面取得重要进展,就全国或地区的生态地理综合区划及要素区划等进行了深入的探讨.但仍然处于探索阶段,理论方法和技术体系不完善.中国未来生态地理区划应重点展开基础理论研究,构建严密的区划体系和理论体系,更加关注人文要素在区划中的重要影响,注重新方法和技术的应用,并加强对典型生态地域系统区划研究.为区域经济发展与生态环境保护、改良和合理利用提供科学依据. [Cheng Yeqing, Zhang Pingyu. Progrest on eco-geographical regionalization researches. Acta Ecologica Sinica, 2006, 26(10): 3424-3433.] https://doi.org/10.3321/j.issn:1000-0933.2006.10.035 URL [本文引用: 3] 摘要生态地理区划是国内外地理学和宏观生态学研究的热点问题之一.国外学者对生态地理区划概念、原则和指标体系和方法等进行了大量的研究和讨论,并就全球、国家和区域尺度的生态地域系统进行划分.其研究经历了从单生态因子分异的规律性认识、生态系统聚合、人类作为生态系统一部分的单要素划分到层次约束、相互包容与不包容的等级斑块思维的整体性功能、过程和结构性划分的发展路线.我国生态地理区划在借鉴自然区划的基础上发展较为迅速,并在区划的原则、方法、理论等方面取得重要进展,就全国或地区的生态地理综合区划及要素区划等进行了深入的探讨.但仍然处于探索阶段,理论方法和技术体系不完善.中国未来生态地理区划应重点展开基础理论研究,构建严密的区划体系和理论体系,更加关注人文要素在区划中的重要影响,注重新方法和技术的应用,并加强对典型生态地域系统区划研究.为区域经济发展与生态环境保护、改良和合理利用提供科学依据.
[5]	张厚华, 黄占斌. 黄土高原生物气候分区与该区生态系统的恢复 . 干旱区资源与环境, 2001, 15(1): 64-71. https://doi.org/10.3969/j.issn.1003-7578.2001.01.011 URL [本文引用: 3] 摘要根据不同的生物气候状况 ,可将黄土高原分成五个生物气候区 :暖温带湿润半湿润森林区、暖温带半湿润半干旱森林草原区、中温带半干旱典型草原区、中温带干旱半干旱荒漠草原区以及中温带干旱草原化荒漠区。紧密结合上述生物气候区域特征 ,将生态学的有关原理引入到该区的治理与生态恢复中 ,提出了治理黄土高原的新模式 :正确实施生物措施 ,人工模拟天然生态系统。该区的生态恢复应从系统的角度考虑 ,并严格遵守生态学的有关规律在不同的生物气候区构建与该区大环境一致的不同的生态系统 ,改变过去那种在黄土高原全面造林而又成效不佳的做法 [Zhang Houhua, Huang Zhanbin. Bio-climatic division and restoration of the degraded ecosystem on the Loess Plateau. Journal of Arid Land Resources and Environment, 2001, 15(1): 64-71.] https://doi.org/10.3969/j.issn.1003-7578.2001.01.011 URL [本文引用: 3] 摘要根据不同的生物气候状况 ,可将黄土高原分成五个生物气候区 :暖温带湿润半湿润森林区、暖温带半湿润半干旱森林草原区、中温带半干旱典型草原区、中温带干旱半干旱荒漠草原区以及中温带干旱草原化荒漠区。紧密结合上述生物气候区域特征 ,将生态学的有关原理引入到该区的治理与生态恢复中 ,提出了治理黄土高原的新模式 :正确实施生物措施 ,人工模拟天然生态系统。该区的生态恢复应从系统的角度考虑 ,并严格遵守生态学的有关规律在不同的生物气候区构建与该区大环境一致的不同的生态系统 ,改变过去那种在黄土高原全面造林而又成效不佳的做法
[6]	Liu G H, Fu B J. The principle and characteristics of ecological regionalization. Advances in Environmental Science, 1998, 6(6): 67-72. [本文引用: 1]
[7]	郑度. 中国生态地理区域系统研究. 北京: 商务印书馆, 2008. [本文引用: 12] [Zheng Du.Chinese Ecological Geographic Area System Research. Beijing: The Commercial Press, 2008.] [本文引用: 12]
[8]	彭建, 王军. 基于Kohonen神经网络的中国土地资源综合分区 . 资源科学, 2006, 28(1): 43-50. https://doi.org/10.3321/j.issn:1007-7588.2006.01.008 URL [本文引用: 1] 摘要本文在综述我国土地资源综合分区研究进展的基础上,建立了土地资源综合分区的原则、依据和等级体系.在前人相关工作的基础上,采用自上而下的演绎途径,根据水热气候指标与地势差异划分出我国的¨个土地资源区,即东北山地平原农林用地区、华北平原旱作农业与工矿建设用地区、黄土高原旱作农业与林牧用地区、四川盆地及秦岭山地农林用地区、长江中下游平原农渔与工矿建设用地区、江南丘陵山地农林用地区、云贵高原农林用地区、华南、滇南热带亚热带农林渔果与工矿建设用地区、内蒙古高原及长城沿线旱作农业与牧业用地区、西北内陆干旱荒漠与绿洲农牧用地区、青藏高原高寒荒漠与林牧用地区;在各土地资源区内部,采用自下而上归并的途径,依据区域土地资源利用结构与社会经济属性指标,以县为基本单元,应用Kohonen网络进行非监督分类,在全国续分出41个土地资源亚区. [Peng Jian, Wang Jun. Land use zoning in China based on Kohonen network. Resources Science, 2006, 28(1): 43-50.] https://doi.org/10.3321/j.issn:1007-7588.2006.01.008 URL [本文引用: 1] 摘要本文在综述我国土地资源综合分区研究进展的基础上,建立了土地资源综合分区的原则、依据和等级体系.在前人相关工作的基础上,采用自上而下的演绎途径,根据水热气候指标与地势差异划分出我国的¨个土地资源区,即东北山地平原农林用地区、华北平原旱作农业与工矿建设用地区、黄土高原旱作农业与林牧用地区、四川盆地及秦岭山地农林用地区、长江中下游平原农渔与工矿建设用地区、江南丘陵山地农林用地区、云贵高原农林用地区、华南、滇南热带亚热带农林渔果与工矿建设用地区、内蒙古高原及长城沿线旱作农业与牧业用地区、西北内陆干旱荒漠与绿洲农牧用地区、青藏高原高寒荒漠与林牧用地区;在各土地资源区内部,采用自下而上归并的途径,依据区域土地资源利用结构与社会经济属性指标,以县为基本单元,应用Kohonen网络进行非监督分类,在全国续分出41个土地资源亚区.
[9]	Bailey R G. Ecoregions of the United States 1:7500000 . Washington: US Government Printing Office, 1976. URL [本文引用: 1]
[10]	Rowe J S, Sheard J W. Ecological land classification: A survey approach. Journal of Environmental Management, 1981, 5(5): 451-464. https://doi.org/10.1007/BF01866822 URL [本文引用: 1] 摘要 A landscape approach to ecological land mapping, as illustrated in this article, proceeds by pattern recognition based on ecological theory. The unit areas delineated are hypotheses that arise from a knowledge of what is ecologically important in the land. Units formed by the mapper are likely to be inefficient or irrelevant for ecological purposes unless he possesses a sound rationale as to the interactions and controlling influences of the structural components of ecosystems. Here is the central problem with what have been called “objective” multivariate approaches to mapping based on grid units and the sometimes arbitrary attributes thereof; they tend to conceal the importance of ecological theory and the necessity for theory-based supervision of pattern recognition. Multivariate techniques are best used iteratively to verify and refine map units initially recognized and delineated by theoretical considerations. These ideas are illustrated by an example of a reconnaissance survey in the Northwest Territories of Canada.
[11]	Wiken E B. Terrestrial Ecozones of Canada: Ecological Land Classification Series NO.19. Quebec city: Environment Canada, 1982. URL
[12]	Bailey R G, Zoltai S C, Wiken E B. Ecological regionalization in Canada and the United States. Geoforum, 1985, 16(3): 265-275. URL [本文引用: 1]
[13]	Bailey R G. Explanatory supplement to ecoregions map of the continents. Environmental Conservation, 1989, 16(4): 307-309. https://doi.org/10.1017/S0376892900009711 URL [本文引用: 1] 摘要 This explanatory note is intended to accompany the map included in this issue. The map is on a scale of 1:30,000,000 (1 cm = 300 km) and shows regional-scale 'ecosystem' units, or ecoregions, differentiated according to a scheme modified from J.M. Crowley (1967), and using climate and vegetation as indicators of the extent of each unit. The units are similar in concept to ‘ecobiomes’ as proposed by Polunin (1984). The map was developed following a proposal by Bailey & Hogg (1986) to supplement the Udvardy (1975) system of biogeographical provinces with a treatment of higher resolution. Three levels or categories of this hierarchy are shown. Of these the broadest, domains, and within them divisions, are based largely on the broad ecological climate zones following the Koppen system as modified by Trewartha (1968), and summarized in our Tables I and II. Climate is emphasized at the broadest levels because of its overriding effect on the composition and productivity of ecosystems etc. from region to region.
[14]	Pan T, Wu S H, He D M, et al. Effects of longitudinal range-gorge terrain on the eco-geographical pattern in Southwest China. Journal of Geographical Sciences, 2012, 22(5): 825-842. https://doi.org/10.1007/s11442-012-0966-6 URL [本文引用: 1] 摘要 China’s southwestern special terrain pattern as parallel arrangement between longitudinal towering mountains and deep
[15]	Mackey B G, Berry S L, Brown T. Reconciling approaches to biogeographical regionalization: A systematic and generic framework examined with a case study of the Australian continent. Journal of Biogeography, 2008, 35(2): 213-229. https://doi.org/10.1111/j.1365-2699.2007.01822.x URL [本文引用: 1] 摘要 Aim68 To develop a systematic and generic framework for biogeographical regionalizations that can assist in reconciling different approaches and advance their application as a research tool. Location68 The Australian continent is used as a case study. Methods68 A review of approaches to biogeographical regionalization revealed two basic methodologies: the integrated survey method and the parametric approach. To help reconcile these different approaches, we propose a simple, four-step, flexible and generic framework. (1) Identification of the thematic foci from the three main themes (composition and evolutionary legacy; ecosystem drivers; ecosystem responses). (2) Proposal of a theory defining the purpose. (3) Application of a numeric agglomerative classification procedure that requires the user to make explicit assumptions about attributes, the number of classification groups, the spatial unit of analysis, and the metric for measuring the similarity of these units based on their attribute values. (4) Acquisition of spatial estimates of the required input attribute data. For this case study, an agglomerative classification strategy was applied using the functions withinpatn3.03, a software package facilitating large-scale, multivariate pattern analysis. The input data to the classifications were continental coverages of 11 environmental variables and three indices of gross primary productivity stored at a grid cell resolution of c. 25065m. The spatial units of analysis were surface hydrological units (SHU), which were derived from a continental digital elevation model based on the contributing areas to stream segments or the area draining into a local sink where there is no organized drainage. The Minkowski series (Euclidean distance) was selected as the association measure to allow weightings to be applied to the variables. Results68 Two new biogeographical regionalizations of the Australian continent were generated. The first was an environmental domain classification, based on 11 climatic, terrain and soil attributes. This regionalization can be used to address hypotheses about the relationship between environmental distance and evolutionary processes. The classification produced 151 environmental groups. The second was a classification of primary productivity regimes based on estimates of the gross primary productivity of the vegetation cover calculated from moderate resolution imaging spectroradiometer (MODIS) normalized difference vegetation index (NDVI) data and estimates of radiation. This classification produced 50 groups, and can be used to examine hypotheses concerning productivity regimes and animal life-history strategies. The productivity classification does not capture all the properties related to biological carrying capacity, process rates and differences in the characteristic biodiversity of ecosystems. Some of these ecologically significant properties are captured by the environmental domain classification. Main conclusions68 Our framework can be applied to all terrestrial regions, and the necessary data for the analyses presented here are now available at global scales. As the spatial predictions generated by the classifications can be tested by comparison with independent data, the approach facilitates exploratory analysis and further hypothesis generation. Integration of the three themes in our framework will contribute to a more comprehensive approach to biogeography.
[16]	Cruz M, Espínola M, Ayala R, et al. How can neural networks speed up ecological regionalization friendly? Replacement of field studies by satellite data using RBFs . In: Proceedings of the International Conference on Fuzzy Computation and International Conference on Neural Computation. Trier, Computer Science Bibliography, 2010: 295-300. URL [本文引用: 1]
[17]	Su Y, Li Z, Wang Y. Study on climatic regionalization for mango planting in Guangxi based on "3S" technology . In: Proceedings of 2011 International Conference on Agricultural and Biosystems Engineering (ICABE 2011). Hong Kong, 2011: 230-234. URL [本文引用: 1]
[18]	Gao J B, Li S C, Zhao Z Q. Validating the demarcation of eco-geographical regions: A geostatistical application. Environmental Earth Sciences, 2010, 59(6): 1327-1336. https://doi.org/10.1007/s12665-009-0120-7 URL [本文引用: 1] 摘要 Eco-geographical regional system is important for the study of global environmental changes and sustainable development, and it serves as a scientific basis for rationally managing and sustainably utilizing ecosystems and natural resources, such as constructing healthy eco-environments and making policies of environmental management. This paper explained the necessity of validation in the demarcation of eco-geographical regions, which is difficult and may be achieved with some assumptions and presumptions because of the existence of transition zones. Also in this paper, we explored the use of geostatistics in validating regions and boundaries using a case study in Qinghai-Tibet Plateau on the basis of normalized difference vegetation index (NDVI) data. The results show that: (1) eco-geographical regions have different spatial complexity and spatial heterogeneity [i.e. different characteristic values (nugget/sill, fractal) of NDVI], and regions with similar patterns of temperature and moisture have similar mean NDVI values and spatial characteristics [i.e. similar spatial characteristic values (nugget/sill, fractal) of NDVI]. Thus, based on the similarity of spatial heterogeneity or spatial patterns of distribution, demarcations of eco-geographical regions with similar conditions of temperature and moisture, such as IID1 (Ngari montane desert-steppe and desert zone), IID2 (Qaidam montane desert zone), and IID3 (Northern slopes of Kunlun montane desert zone), meet the regional validation requirement. (2) Based on the comparison of spatial heterogeneity or spatial patterns of distribution, the boundary between IIA/B1 (Western Sichuan鈥揺astern Xizang montane coniferous forest zone) and IB1 (Golog
[19]	Liu Y, Wu S H, Zheng D, et al. Soil indicators for eco-geographic regionalization: A case study in mid-temperate zone of eastern China. Journal of Geographical Sciences, 2009, 19(2): 200-212. https://doi.org/10.1007/s11442-009-0200-3 URL [本文引用: 1] 摘要 Eco-geographic regional system is formed by division or combination of natural features based on geographic relativity and comparison of major ecosystem factors (including biological and non-biological) and geographic zonality. In previous studies, soil types were often taken as a basis for soil regionalization. However, the quantitative characteristics of soil indicators are fitter than the qualitative ones of soil types for modern regionalization researches. Based on the second China' national soil survey data and the provincial soil resource information, by principal analysis and discriminant analysis, this paper discusses the appropriate soil indicators as the complement of eco-geographic region indicator systems and the relationships between these soil indicators and soil types in regionalization. The results show that five indicators are used in eco-geographic zonality in mid-temperate zone of eastern China which are organic matter content, cation exchange capacity, pH, clay content and bulk density in topsoils. With a regression-kriging approach, the maps of soil indicators in mid-temperate zone of eastern China are compiled with a resolution of 1 km in every grid and the indicative meanings of these soil indicators are discussed. By cluster analysis it is proved that these soil indicators are better than the soil types and soil regionalization in delineating eco-geographic regions.
[20]	任美锷, 包浩生. 中国自然区域及开发整治. 北京: 科学出版社, 1992. [本文引用: 1] [Ren Mei'e, Bao Haosheng. China's Natural Regionalization and Development Regulation. Beijing: Science Press, 1992.] [本文引用: 1]
[21]	侯学煜. 中国自然生态区划与大农业发展战略. 北京: 科学出版社, 1988. URL [本文引用: 1] [Hou Xueyu.China Natural Regionalization and Agricultural Development Strategy. Beijing: Science Press, 1988.] URL [本文引用: 1]
[22]	傅伯杰, 刘国华, 陈利顶, 等. 中国生态区划方案 . 生态学报, 2001, 21(1): 1-6. URL [本文引用: 3] 摘要 Ecological regionalization is a base for rational management and sustainable utilization of ecosystems and natural resources It can provide scientific basis for constructing healthy ecological environments and making policies of environmental management In this paper, based on synthetical analysis of the characteristics of ecological environments of China, the principles of ecological regionalization are discussed, and indices and nomenclature of ecological regionalization are proposed, The ecoregions in national scale are divided The results show that there are 3 domains, 13 ecoregions and 57 ecodistricts [Fu Bojie, Liu Guohua, Chen Liding, et al. Scheme of ecological regionalization in China. Acta Ecologica Sinica, 2001, 21(1): 1-6.] URL [本文引用: 3] 摘要 Ecological regionalization is a base for rational management and sustainable utilization of ecosystems and natural resources It can provide scientific basis for constructing healthy ecological environments and making policies of environmental management In this paper, based on synthetical analysis of the characteristics of ecological environments of China, the principles of ecological regionalization are discussed, and indices and nomenclature of ecological regionalization are proposed, The ecoregions in national scale are divided The results show that there are 3 domains, 13 ecoregions and 57 ecodistricts
[23]	欧阳志云, 王效科, 苗鸿. 中国生态环境敏感性及其区域差异规律研究 . 生态学报, 2000, 20(1): 10-13. URL [本文引用: 1] 摘要 The eco environmental degradation has been attracted more and more attention in China. It is becoming a challenge for scientists to find effective measures to tackle the eco environmental problems. In this study, we reviewed the spatial pattern and correlationship of major eco environmental problems: soil erosion, desertification, salinization, and acid rain. The concept of eco environment sensitivity was proposed, and the factors which influence the eco environment sensitivity in China were analyzed, especially climatic factors were focused. A regional classification of eco environment sensitivity in China was developed, and the characteristics of each sub region of eco environment sensitivity was described. [Ouyang Zhiyun, Wang Xiaoke, Miao Hong. China's eco-environmental sensitivity and its spatial heterogeneity. Acta Ecologica Sinica, 2000, 20(1): 10-13.] URL [本文引用: 1] 摘要 The eco environmental degradation has been attracted more and more attention in China. It is becoming a challenge for scientists to find effective measures to tackle the eco environmental problems. In this study, we reviewed the spatial pattern and correlationship of major eco environmental problems: soil erosion, desertification, salinization, and acid rain. The concept of eco environment sensitivity was proposed, and the factors which influence the eco environment sensitivity in China were analyzed, especially climatic factors were focused. A regional classification of eco environment sensitivity in China was developed, and the characteristics of each sub region of eco environment sensitivity was described.
[24]	黄姣, 高阳, 赵志强, 等. 基于GIS与SOFM网络的中国综合自然区划 . 地理研究, 2011, 30(9): 1648-1659. https://doi.org/10.3724/SP.J.1011.2011.00415 URL [本文引用: 3] 摘要 Comprehensive physiographic regionalization has long been a core issue of physical geography in China.A great number of regionalization themes have been developed and applied as guidelines for regional development and geography teaching.However,these themes mainly use the traditional expertise-experiences-based regionalization methodology,which probably make themselves unreliable due to certain prejudices and different knowledge backgrounds of each individual.In order to overcome this obstacle,and to enrich regionalization research theoretically and methodologically,this paper tries to apply SOFM neural network to the regionalization.Supported by GIS technology and following the traditional three-level-strategy,we construct and operate SOFM neural networks at each level,using temperature factors,moisture factors and supplement factors respectively.Finally,we divide Chinese mainland into 8 temperature zones,17 moisture regions and 43 natural sub-regions,then compare this scheme with those based on traditional methods.The result shows that based on GIS platform,applying SOFM neural network into comprehensive physiographic regionalization has significant advantages,which is an important supplement and development to traditional regionalization paradigm. [Huang Jiao, Gao Yang, Zhao Zhiqiang, et al. Comprehensive physiographic regionalization of China using GIS and SOFM neural network. Geographical Research, 2011, 30(9): 1648-1659.] https://doi.org/10.3724/SP.J.1011.2011.00415 URL [本文引用: 3] 摘要 Comprehensive physiographic regionalization has long been a core issue of physical geography in China.A great number of regionalization themes have been developed and applied as guidelines for regional development and geography teaching.However,these themes mainly use the traditional expertise-experiences-based regionalization methodology,which probably make themselves unreliable due to certain prejudices and different knowledge backgrounds of each individual.In order to overcome this obstacle,and to enrich regionalization research theoretically and methodologically,this paper tries to apply SOFM neural network to the regionalization.Supported by GIS technology and following the traditional three-level-strategy,we construct and operate SOFM neural networks at each level,using temperature factors,moisture factors and supplement factors respectively.Finally,we divide Chinese mainland into 8 temperature zones,17 moisture regions and 43 natural sub-regions,then compare this scheme with those based on traditional methods.The result shows that based on GIS platform,applying SOFM neural network into comprehensive physiographic regionalization has significant advantages,which is an important supplement and development to traditional regionalization paradigm.
[25]	刘闯. 中尺度对地观测系统支持下中国综合自然地理区划新方法论研究 . 地理科学进展, 2004, 23(6): 1-9. https://doi.org/10.3969/j.issn.1007-6301.2004.06.001 URL [本文引用: 1] 摘要 One of the most outstanding achievements in Chinese history of geography studies during the 20th century is that from the comprehensive physical geographical regionalization. There are seven specific themes of the Comprehensive Physical Geographical Regionalization of China at least during the last half century, while the very beginning one was published by Professor LIN Chao from Peking University in 1954. China is a huge country with great pressure on agriculture, natural resources and environment. How to make the regional development more efficiently, the way of Comprehensive Physical Geographical Regionalization could be useful method to help scientists and decision makers to identify the differential features in the regional scale. From 1958~1983, five different themes had been developed for supporting agriculture development in China. Prof. HUANG Bingwei from Institute of Geography, Chinese Academy of Sciences led the activities, Prof. REN Mei'er from Nanking University, Prof. HOU Xueyu, ZHAO Songqiao from Chinese Academy of Sciences, Prof. XI Chengfan from Ministry of Agriculture of China proposed themes with revisions from different opinions. Even so, the hierarchical approach and the top-down methodology make them more successful in geography teaching, agriculture development and food support. With the economic development and the population increased rapidly during the last 20 years in China, more and more pressures come on ecosystems and environment. It is recognized that China has to deal with the challenges from the regional development and adaptation of the global environmental change. At the same time period, the earth observation systems and related moderate scale data has been accumulated, these are valuable information for monitoring envi ronment change. About 280TB data from EOS are archived in China, and only a very small part of them are used in the Comprehensive Physical Geographical Regionalization of China. A new methodology supported by the EOS data in moderate scale for the Comprehensive Physical Geographical Regionalization of China is discussed and propose in this paper, including the methods of data capturing, data mining with computational models and the data sharing issues ad well. A new theme of the Comprehensive Physical Geographical Regionalization of China for the natural resources, environment and sustainable development with the methods from both of top-down and bottom-up ways has been incubated. This paper has been presented in the Anural Conference of China National Committee, for IGBP, 2004, Beijing, and I would like to present it to Professor LIN Chao also for celebrating the 50th anniversary of the first theme of Comprehensive Physical Geographical Regionalization of China, which was created by him. [Liu Chuang. A new methodology for comprehensive physical regionalization of China supported by EOS in the moderate scale. Progress in Geography, 2004, 23(6): 1-9.] https://doi.org/10.3969/j.issn.1007-6301.2004.06.001 URL [本文引用: 1] 摘要 One of the most outstanding achievements in Chinese history of geography studies during the 20th century is that from the comprehensive physical geographical regionalization. There are seven specific themes of the Comprehensive Physical Geographical Regionalization of China at least during the last half century, while the very beginning one was published by Professor LIN Chao from Peking University in 1954. China is a huge country with great pressure on agriculture, natural resources and environment. How to make the regional development more efficiently, the way of Comprehensive Physical Geographical Regionalization could be useful method to help scientists and decision makers to identify the differential features in the regional scale. From 1958~1983, five different themes had been developed for supporting agriculture development in China. Prof. HUANG Bingwei from Institute of Geography, Chinese Academy of Sciences led the activities, Prof. REN Mei'er from Nanking University, Prof. HOU Xueyu, ZHAO Songqiao from Chinese Academy of Sciences, Prof. XI Chengfan from Ministry of Agriculture of China proposed themes with revisions from different opinions. Even so, the hierarchical approach and the top-down methodology make them more successful in geography teaching, agriculture development and food support. With the economic development and the population increased rapidly during the last 20 years in China, more and more pressures come on ecosystems and environment. It is recognized that China has to deal with the challenges from the regional development and adaptation of the global environmental change. At the same time period, the earth observation systems and related moderate scale data has been accumulated, these are valuable information for monitoring envi ronment change. About 280TB data from EOS are archived in China, and only a very small part of them are used in the Comprehensive Physical Geographical Regionalization of China. A new methodology supported by the EOS data in moderate scale for the Comprehensive Physical Geographical Regionalization of China is discussed and propose in this paper, including the methods of data capturing, data mining with computational models and the data sharing issues ad well. A new theme of the Comprehensive Physical Geographical Regionalization of China for the natural resources, environment and sustainable development with the methods from both of top-down and bottom-up ways has been incubated. This paper has been presented in the Anural Conference of China National Committee, for IGBP, 2004, Beijing, and I would like to present it to Professor LIN Chao also for celebrating the 50th anniversary of the first theme of Comprehensive Physical Geographical Regionalization of China, which was created by him.
[26]	张学儒, 张镱锂, 刘林山, 等. 基于SOFM神经网络模型的土地类型分区尝试: 以青藏高原东部样带为例 . 地理研究, 2013, 32(5): 839-847. https://doi.org/10.11821/yj2013050007 URL [本文引用: 1] 摘要 The bottom-up physical regionalization based on land type units can establish more clear boundaries,which is dramatic breakthroughs of research in physical regionalization.The study area is located in the mountainous areas of eastern Tibetan Plateau.The research on bottom-up physical regionalization based on land type units is implemented by means of a combined method of SOFM network models and GIS.Topographic index,warmth index,humidity index,LCI index and hydrological index can be seen as input layer variables,and land type units and natural zones are regarded as background and regionalizing target.The results show that:(1) the land types are aggregated into five natural regions,namely alpine sparse vegetation plateau,alpine meadow plateau,alpine shrub-meadow plateau,alpine shrub-meadow in mountain-valleys,and coniferous forestland in mountain-valleys.(2) The boundaries of this zoning are close to scheme of ecological regionalization in China.Therefore,the results of network classification show that there is high concentration at spatial scale,which represents natural geographical characteristics in the mountainous areas of eastern Tibetan Plateau.This study can provide new ideas and methods for bottom-up physical regionalization based on land type units. [Zhang Xueru, Zhang Yili, Liu Linshan, et al. Zoning by land types based on SOFM network: A case study on transect of eastern Tibetan Plateau. Geographical Research, 2013, 32(5): 839-847.] https://doi.org/10.11821/yj2013050007 URL [本文引用: 1] 摘要 The bottom-up physical regionalization based on land type units can establish more clear boundaries,which is dramatic breakthroughs of research in physical regionalization.The study area is located in the mountainous areas of eastern Tibetan Plateau.The research on bottom-up physical regionalization based on land type units is implemented by means of a combined method of SOFM network models and GIS.Topographic index,warmth index,humidity index,LCI index and hydrological index can be seen as input layer variables,and land type units and natural zones are regarded as background and regionalizing target.The results show that:(1) the land types are aggregated into five natural regions,namely alpine sparse vegetation plateau,alpine meadow plateau,alpine shrub-meadow plateau,alpine shrub-meadow in mountain-valleys,and coniferous forestland in mountain-valleys.(2) The boundaries of this zoning are close to scheme of ecological regionalization in China.Therefore,the results of network classification show that there is high concentration at spatial scale,which represents natural geographical characteristics in the mountainous areas of eastern Tibetan Plateau.This study can provide new ideas and methods for bottom-up physical regionalization based on land type units.
[27]	韩忆楠, 彭建, 王仰麟. 基于自组织特征映射(SOFM)网络的农牧交错带景观分区: 以内蒙古自治区为例 . 应用生态学报, 2013, 24(5): 1224-1230. URL [本文引用: 1] [Han Yinan, Peng Jian, Wang Yanglin. Landscape regionalization of agro-pastoral transitional zone using self-organizing feature maps (SOFM) network: A case study of Inner Mongolia Autonomous Region, China. Chinese Journal of Applied Ecology, 2013, 24(5): 1224-1230.] URL [本文引用: 1]
[28]	Peng J, Ma J, Yuan Y. Integrated urban land-use zoning and associated spatial development: Case study in Shenzhen, China. Journal of Urban Planning and Development, 2014. doi: 10.1061/(ASCE)UP.1943-5444.0000245. URL [本文引用: 2]
[29]	郝成元, 吴绍洪, 李双成. 基于SOFM的区域界线划分方法 . 地理科学进展, 2008, 27(5): 121-127. https://doi.org/10.11820/dlkxjz.2008.05.016 URL [本文引用: 2] 摘要区域分异研究是人们对地理环境认知深度和自然地理研究水平的重要标志之一,划定分区界线就成为一项迫切而意义重大的工作,尤其是在气候复杂、地貌多样的我国西南高原、山地组合区。云南省南部地区由于多季风系统和大地形作用的影响,气候复杂多样。雨季,温暖湿润的西南夏季风给研究区西部带来大量降水,东部雨量少;干季,整个研究区主要在西风南支急流控制之下,天气晴朗、少雨,同时也使得植被种类及盖度差别较大。基于研究区30个气象台站的海拔高度、多年平均气温和降水、风速、活动积温、潜在蒸散以及MODIS-EVI等数据,利用神经网络技术构建了非线性分类器,即自组织特征映射模型(SOFM),对所有气象台站进行了聚类研究。结果显示,哀牢山成为阻挡北来冷空气进入西南山地的屏障,是我国冬季东北风和夏季西南风的分界线,因此也成为研究区东、西两类气候的分界线。SOFM网络应用于地形复杂、地貌多样的生态地理区域分异研究,基本能反映不同区域之间界线两侧的相似性和差异性,能够揭示一个由量变到质量过程的连续性,不失为一种较好的综合自然地理区划方法。 [Hao Chengyuan, Wu Shaohong, Li Shuangcheng. Study on the method of areal differentiation based on SOFM. Progress in Geography, 2008, 27(5): 121-127.] https://doi.org/10.11820/dlkxjz.2008.05.016 URL [本文引用: 2] 摘要区域分异研究是人们对地理环境认知深度和自然地理研究水平的重要标志之一,划定分区界线就成为一项迫切而意义重大的工作,尤其是在气候复杂、地貌多样的我国西南高原、山地组合区。云南省南部地区由于多季风系统和大地形作用的影响,气候复杂多样。雨季,温暖湿润的西南夏季风给研究区西部带来大量降水,东部雨量少;干季,整个研究区主要在西风南支急流控制之下,天气晴朗、少雨,同时也使得植被种类及盖度差别较大。基于研究区30个气象台站的海拔高度、多年平均气温和降水、风速、活动积温、潜在蒸散以及MODIS-EVI等数据,利用神经网络技术构建了非线性分类器,即自组织特征映射模型(SOFM),对所有气象台站进行了聚类研究。结果显示,哀牢山成为阻挡北来冷空气进入西南山地的屏障,是我国冬季东北风和夏季西南风的分界线,因此也成为研究区东、西两类气候的分界线。SOFM网络应用于地形复杂、地貌多样的生态地理区域分异研究,基本能反映不同区域之间界线两侧的相似性和差异性,能够揭示一个由量变到质量过程的连续性,不失为一种较好的综合自然地理区划方法。
[30]	董锁成, 吴玉萍, 王海英. 黄土高原生态脆弱贫困区生态经济发展模式研究: 以甘肃省定西地区为例 . 地理研究, 2003, 22(5): 590-600. URL [本文引用: 1] [Dong Suocheng, Wu Yuping, Wang Haiying. A study on the eco-economic development model in the eco-environmental vulnerable and needy region on Loess Plateau: The case of Dingxi prefecture in Gansu province. Geographical Research, 2003, 22(5): 590-600.] URL [本文引用: 1]
[31]	汪文霞, 周建斌, 严德翼, 等. 黄土区不同类型土壤微生物量碳、氮和可溶性有机碳、氮的含量及其关系 . 水土保持学报, 2006, 20(6): 103-106. https://doi.org/10.3321/j.issn:1009-2242.2006.06.025 URL [本文引用: 1] 摘要研究了黄土高原南部地区不同土壤类型及不同利用方式下土壤微生物量碳、氮和可溶性有机碳、氮的含量.结果表明:不同土地利用方式下,土壤微生物量碳、氮和可溶性有机碳、氮含量均为林地＞农田,其中林地枯枝落叶层＞林地0～20 cm土层.农田土壤微生物量碳、氮的含量均为红油土＞黑垆土＞淋溶褐土;农田土壤中可溶性有机碳含量为淋溶褐土＞红油土＞黑垆土,而可溶性有机氮含量则为黑垆土＞红油土＞淋溶褐土.方差分析表明,不同土壤类型土壤微生物量氮含量之间的差异达显著水平,而不同土壤类型间土壤微生物量碳、可溶性有机碳、氮含量之间的差异未达显著水平.土壤微生物量碳、氮占土壤有机碳和全氮的比例明显高于可溶性有机碳、氮占土壤有机碳和全氮的比例.相关分析发现,土壤微生物量碳与可溶性有机碳之间以及土壤微生物量氮与可溶性有机氮之间的相关性达显著或极显著水平,说明土壤微生物量碳、氮和土壤可溶性有机碳、氮之间有密切联系. [Wang Wenxia, Zhou Jianbin, Yan Deyi, et al. Contents of soil microbial biomass C, N and K₂SO₄-extractable organic C, N and their relations in different soil types on Loess Plateau of China. Journal of Soil and Water Conservation, 2006, 20(6): 103-106.] https://doi.org/10.3321/j.issn:1009-2242.2006.06.025 URL [本文引用: 1] 摘要研究了黄土高原南部地区不同土壤类型及不同利用方式下土壤微生物量碳、氮和可溶性有机碳、氮的含量.结果表明:不同土地利用方式下,土壤微生物量碳、氮和可溶性有机碳、氮含量均为林地＞农田,其中林地枯枝落叶层＞林地0～20 cm土层.农田土壤微生物量碳、氮的含量均为红油土＞黑垆土＞淋溶褐土;农田土壤中可溶性有机碳含量为淋溶褐土＞红油土＞黑垆土,而可溶性有机氮含量则为黑垆土＞红油土＞淋溶褐土.方差分析表明,不同土壤类型土壤微生物量氮含量之间的差异达显著水平,而不同土壤类型间土壤微生物量碳、可溶性有机碳、氮含量之间的差异未达显著水平.土壤微生物量碳、氮占土壤有机碳和全氮的比例明显高于可溶性有机碳、氮占土壤有机碳和全氮的比例.相关分析发现,土壤微生物量碳与可溶性有机碳之间以及土壤微生物量氮与可溶性有机氮之间的相关性达显著或极显著水平,说明土壤微生物量碳、氮和土壤可溶性有机碳、氮之间有密切联系.
[32]	朱显谟, 孙林夫, 杨文治, 等. 黄土高原综合治理分区 . 中国科学院西北水土保持研究所集刊: 黄土高原综合治理专集, 1985, (1): 2-66. URL [本文引用: 1] 摘要本文在介绍了黄土高原的基本概况与建设成就之后,详细剖析了黄土高原地区存在的问题及其产生的原因,总结了其中的经验教训.文中还着重分析了该地区的自然资源及农林牧副各业生产潜力, 提出了资源合理利用与综合经营的具体意见。在此基础上,制定了5个地带25个区域的分区综合治理方案.为了保证这个方案的实施,本文在最后一部分提出了 10条必需的保证措施. [Zhu Xianmo, Sun Linfu, Yang Wenzhi, et al. Regionalization of comprehensive control in Loess Plateau. Memoir of NISWC, Academia Sinica, 1985, (1): 2-66.] URL [本文引用: 1] 摘要本文在介绍了黄土高原的基本概况与建设成就之后,详细剖析了黄土高原地区存在的问题及其产生的原因,总结了其中的经验教训.文中还着重分析了该地区的自然资源及农林牧副各业生产潜力, 提出了资源合理利用与综合经营的具体意见。在此基础上,制定了5个地带25个区域的分区综合治理方案.为了保证这个方案的实施,本文在最后一部分提出了 10条必需的保证措施.
[33]	焦峰, 温仲明, 李锐. 黄土高原退耕还林(草)环境效应分析 . 水土保持研究, 2005, 12(1): 26-29. https://doi.org/10.3969/j.issn.1005-3409.2005.01.008 URL [本文引用: 1] 摘要退耕还林(草)是黄土高原恢复植被、改善生态的重要措施之一.随着退耕还林(草)工程的深入,其产生的经济、社会和生态效应也越来越广泛地受到人们的关注.对黄土高原退耕还林(草)影响下的土地利用格局演变、生态和社会效应进行了分析,提出了黄土高原尚待或进一步深化研究的几个问题. [Jiao Feng, Wen Zhongming, Li Rui. Analysis on environment effect of the returning farm land into forest and grassland on the Loess Plateau. Research of Soil and Water Conservation, 2005, 12(1): 26-29.] https://doi.org/10.3969/j.issn.1005-3409.2005.01.008 URL [本文引用: 1] 摘要退耕还林(草)是黄土高原恢复植被、改善生态的重要措施之一.随着退耕还林(草)工程的深入,其产生的经济、社会和生态效应也越来越广泛地受到人们的关注.对黄土高原退耕还林(草)影响下的土地利用格局演变、生态和社会效应进行了分析,提出了黄土高原尚待或进一步深化研究的几个问题.
[34]	国家林业局. 长江上游、黄河上中游地区天然林资源保护工程实施方案., 2015-04-27. URL [本文引用: 1] [State Forestry Administration. Implementation of natural forest resources protection project in the upper middle reaches of the Yangtze River and upper middle reaches of the Yellow River. , 2015-04-27.] URL [本文引用: 1]
[35]	国家林业局. 关于在西北、华北、东北风沙危害和水土流失重点地区建设大型防护林的规划., 2015-04-27. URL [本文引用: 1] [State Forestry Administration. Planning of large scale shelter forest in the northwest, North China, northeast sandstorm hazard and soil erosion. , 2015-04-27.] URL [本文引用: 1]
[36]	Feng X M, Fu B J, Lu N, et al. How ecological restoration alters ecosystem services: An analysis of carbon sequestration in China's Loess Plateau. Scientific Reports, 2013, 3. doi: 10.1038/srep02846. URL PMID: 24088871 [本文引用: 1] 摘要 Restoring disturbed and over-exploited ecosystems is important to mitigate human pressures on natural ecosystems. China has launched an ambitious national ecosystem restoration program called Grain to Green Program (GTGP) over the last decade. By using remote sensing techniques and ecosystem modelling, we quantitatively evaluated the changes in ecosystem carbon sequestration since China's GTGP program during period of 2000-2008. It was found the NPP and NEP in this region had steadily increased after the initiative of the GTGP program, and a total of 96.1Tg of additional carbon had been sequestered during that period. Changes in soil carbon storage were lagged behind and thus insignificant over the period, but was expected to follow in the coming decades. As a result, the Loess Plateau ecosystem had shifted from a net carbon source in 2000 to a net carbon sink in 2008. The carbon sequestration efficiency was constrained by precipitation, and appropriate choices of restoration types (trees, shrubs, and grasses) in accordance to local climate are critical for achieving the best benefit/cost efficiency.
[37]	信忠保, 许炯心, 郑伟. 气候变化和人类活动对黄土高原植被覆盖变化的影响 . 中国科学D辑: 地球科学, 2007, 37(11): 1504-1514. https://doi.org/10.3321/j.issn:1006-9267.2007.11.009 URL [本文引用: 1] 摘要利用GIMMS和SPOT VGT两种归一化植被指数(NDVI)数据对黄土高原地区1981～2006年期间植被覆盖的时空变化进行了研究, 并从气候变化和人类活动的角度分析了植被覆盖变化的原因.黄土高原地区植被覆盖经历了以下4个阶段: ① 1981～1989年植被覆盖持续增加时期; ② 1990～1998年以小幅波动为特征的相对稳定时期; ③ 1999～2001年植被覆盖迅速下降时期; ④ 2002～2006年植被覆盖进入迅速上升时期.黄土高原地区植被覆盖变化存在显著的空间差异, 内蒙古和宁夏沿黄农业灌溉区和鄂尔多斯退耕还林还草生态恢复区的植被覆盖明显提高, 而黄土丘陵沟壑区和六盘山、秦岭北坡等山地森林区的植被覆盖明显退化.从不同的植被类型来看, 沙地、草地和耕地的NDVI上升趋势显著, 而森林植被的NDVI呈明显的下降趋势.研究表明: 植被覆盖变化是气候变化和人类活动共同作用的结果.黄土高原地区气候变暖在加剧土壤干燥化抑制夏季植被生长的同时, 提高了春、秋季节植被生长活性, 延长了植被生长期.黄土高原地区植被覆盖和降水关系密切, 降水变化是植被覆盖变化的重要原因.农业生产水平的提高致使农业区NDVI在不断上升, 同时, 正在黄土高原大规模进行的退耕还林还草工程建设, 其生态效应也正在呈现. [Xin Zhongbao, Xu Jiongxin, Zheng Wei. Response of vegetation cover change to climate change and human activities in Loess Plateau. Science in China Series D: Earth Sciences, 2007, 37(11): 1504-1514.] https://doi.org/10.3321/j.issn:1006-9267.2007.11.009 URL [本文引用: 1] 摘要利用GIMMS和SPOT VGT两种归一化植被指数(NDVI)数据对黄土高原地区1981～2006年期间植被覆盖的时空变化进行了研究, 并从气候变化和人类活动的角度分析了植被覆盖变化的原因.黄土高原地区植被覆盖经历了以下4个阶段: ① 1981～1989年植被覆盖持续增加时期; ② 1990～1998年以小幅波动为特征的相对稳定时期; ③ 1999～2001年植被覆盖迅速下降时期; ④ 2002～2006年植被覆盖进入迅速上升时期.黄土高原地区植被覆盖变化存在显著的空间差异, 内蒙古和宁夏沿黄农业灌溉区和鄂尔多斯退耕还林还草生态恢复区的植被覆盖明显提高, 而黄土丘陵沟壑区和六盘山、秦岭北坡等山地森林区的植被覆盖明显退化.从不同的植被类型来看, 沙地、草地和耕地的NDVI上升趋势显著, 而森林植被的NDVI呈明显的下降趋势.研究表明: 植被覆盖变化是气候变化和人类活动共同作用的结果.黄土高原地区气候变暖在加剧土壤干燥化抑制夏季植被生长的同时, 提高了春、秋季节植被生长活性, 延长了植被生长期.黄土高原地区植被覆盖和降水关系密切, 降水变化是植被覆盖变化的重要原因.农业生产水平的提高致使农业区NDVI在不断上升, 同时, 正在黄土高原大规模进行的退耕还林还草工程建设, 其生态效应也正在呈现.
[38]	孟庆香. 基于遥感、GIS和模型的黄土高原生态环境质量综合评价 . 杨凌: 西北农林科技大学博士学位论文, 2006. URL [本文引用: 1] 摘要本研究针对黄土高原生态环境主要问题和国家生态环境建设的需要,总结已有研究,初步提出了生态环境质量综合评价思路框架,采用遥感和地理信息系统等手段, 充分利用已有的调查和研究成果,在黄土高原生态环境单因子进行动态分析基础上,建立了黄土高原生态环境数据库。在确定生态环境质量综合评价指标体系和方法基础上,建立了基于主成分分析的复合评价模型,以县为单位,进行了黄土高原生态环境质量动态分析及其预测。由于针对黄土高原大范围、多因子、多时期的生态环境质量评价在我国鲜有研究,因而该研究对深化黄土高原生态环境的科学认识,制订科学的生态环境建设规划提供了理论参考;同时也为进一步研究黄土高原生态环境提供了先进的技术手段和科学的数据基础。本研究取得以下主要结论和成果: 1.提出了生态环境质量综合评价的基本思路。在可持续发展理论、环境容载力理论和景观生态学理论指导下,分别从生态环境承载能力、生态环境发展能力、生态环境持续能力3个方面进行生态环境质量评价。根据综合评价模型,计算生态环境质量综合指数,主要由生态环境承载能力、生态环境发展能力和生态环境持续能力构成。 2.建立黄土高原生态环境质量综合评价指标体系。根据评价的基本思路,通过对生态环境系统特征因子的变量分析,提出了比较完整的、多级别的评价指标体系, 包括生态环境质量综合指数(目标层)、生态环境承载能力、生态环境发展能力和生态环境持续能力(准则层)和24个指标(指标层)。 3.建立了较为完备的生态环境数据库并进行了系列制图。深入细致地研究黄土高原地形、土地利用、土壤侵蚀、植被覆盖、气候插值、土壤可蚀性因子、土地适宜性和限制性等自然因子以及社会经济要素,以100m栅格为单元,进行黄土高原生态环境单因子分析。在此基础上建立了黄土高原20世纪80年代中期和90年代中后期2期生态环境数据库,并进行了系列制图与分析,得到一些有意义的结论,如景观多样性指数增大,生态环境弹性度提高,水土流失强度指数下降,归一化植被指数上升等。生态环境单因子分析为深刻理解黄土高原及其生态环境建设提供了理论参考,数据库建设为进一步研究黄土高原生态环境提供了科学的数据基础。 4.采用基于主成分分析的复合评价模型,以县为评价单元,分别计算了20世纪80年代中期和90年代中后期生态环境承载能力指数、生态环境发展能力指数、生态环 [Meng Qingxiang. Integrative assessment of eco-environmental quality on the Loess Plateau based on remote sensing, GIS and models. Yangling: Doctoral Dissertation of Northwest Agriculture and Forestry University, 2006.] URL [本文引用: 1] 摘要本研究针对黄土高原生态环境主要问题和国家生态环境建设的需要,总结已有研究,初步提出了生态环境质量综合评价思路框架,采用遥感和地理信息系统等手段, 充分利用已有的调查和研究成果,在黄土高原生态环境单因子进行动态分析基础上,建立了黄土高原生态环境数据库。在确定生态环境质量综合评价指标体系和方法基础上,建立了基于主成分分析的复合评价模型,以县为单位,进行了黄土高原生态环境质量动态分析及其预测。由于针对黄土高原大范围、多因子、多时期的生态环境质量评价在我国鲜有研究,因而该研究对深化黄土高原生态环境的科学认识,制订科学的生态环境建设规划提供了理论参考;同时也为进一步研究黄土高原生态环境提供了先进的技术手段和科学的数据基础。本研究取得以下主要结论和成果: 1.提出了生态环境质量综合评价的基本思路。在可持续发展理论、环境容载力理论和景观生态学理论指导下,分别从生态环境承载能力、生态环境发展能力、生态环境持续能力3个方面进行生态环境质量评价。根据综合评价模型,计算生态环境质量综合指数,主要由生态环境承载能力、生态环境发展能力和生态环境持续能力构成。 2.建立黄土高原生态环境质量综合评价指标体系。根据评价的基本思路,通过对生态环境系统特征因子的变量分析,提出了比较完整的、多级别的评价指标体系, 包括生态环境质量综合指数(目标层)、生态环境承载能力、生态环境发展能力和生态环境持续能力(准则层)和24个指标(指标层)。 3.建立了较为完备的生态环境数据库并进行了系列制图。深入细致地研究黄土高原地形、土地利用、土壤侵蚀、植被覆盖、气候插值、土壤可蚀性因子、土地适宜性和限制性等自然因子以及社会经济要素,以100m栅格为单元,进行黄土高原生态环境单因子分析。在此基础上建立了黄土高原20世纪80年代中期和90年代中后期2期生态环境数据库,并进行了系列制图与分析,得到一些有意义的结论,如景观多样性指数增大,生态环境弹性度提高,水土流失强度指数下降,归一化植被指数上升等。生态环境单因子分析为深刻理解黄土高原及其生态环境建设提供了理论参考,数据库建设为进一步研究黄土高原生态环境提供了科学的数据基础。 4.采用基于主成分分析的复合评价模型,以县为评价单元,分别计算了20世纪80年代中期和90年代中后期生态环境承载能力指数、生态环境发展能力指数、生态环
[39]	张宝庆, 吴普特, 赵西宁. 近30a黄土高原植被覆盖时空演变监测与分析 . 农业工程学报, 2011, 27(4): 287-293. https://doi.org/10.3969/j.issn.1002-6819.2011.04.051 URL [本文引用: 1] 摘要 Time-series GIMMS and SPOT VGT NDVI datasets, which served as an evaluation index of the vegetation cover, were used in this paper to detect and analyze the spatial and temporal variation of vegetation cover in the Loess Plateau from 1982 to 2009. Results showed that before the large-scale implementations of vegetation construction (1982-1998), the annual mean NDVI of the Loess Plateau fluctuated in a narrow range yearly; most places indicated no significant changes in NDVI except for some improvement in a few places. However, since the large-scale implement of vegetation construction in 1999, the annual mean NDVI of the Loess Plateau has increased significantly. Most regions (covering 66.12% of the total area) on the Loess Plateau showed significant positive correlation between NDVI and time over the past decade, especially in the part traversed by the Loess Hilly-gully Region, where the achievements of vegetation construction were quite noticeable. Moreover, the vegetation cover in 15°～25° and 6°～15° slope areas experienced great improvement from 1999 to 2009, which may be helpful to control soil and water losses. The large-scale implementations of vegetation construction accelerated the vegetation restoration in the Loess Plateau. Nevertheless, regions with sparse vegetation on the Loess Plateau still covered a great part by 2009, so it is necessary to further strengthen the ecological environment construction in the future. [Zhang Baoqing, Wu Pute, Zhao Xining. Detecting and analysis of spatial and temporal variation of vegetation cover in the Loess Plateau during 1982-2009. Transactions of the CSAE, 2011, 27(4): 287-293.] https://doi.org/10.3969/j.issn.1002-6819.2011.04.051 URL [本文引用: 1] 摘要 Time-series GIMMS and SPOT VGT NDVI datasets, which served as an evaluation index of the vegetation cover, were used in this paper to detect and analyze the spatial and temporal variation of vegetation cover in the Loess Plateau from 1982 to 2009. Results showed that before the large-scale implementations of vegetation construction (1982-1998), the annual mean NDVI of the Loess Plateau fluctuated in a narrow range yearly; most places indicated no significant changes in NDVI except for some improvement in a few places. However, since the large-scale implement of vegetation construction in 1999, the annual mean NDVI of the Loess Plateau has increased significantly. Most regions (covering 66.12% of the total area) on the Loess Plateau showed significant positive correlation between NDVI and time over the past decade, especially in the part traversed by the Loess Hilly-gully Region, where the achievements of vegetation construction were quite noticeable. Moreover, the vegetation cover in 15°～25° and 6°～15° slope areas experienced great improvement from 1999 to 2009, which may be helpful to control soil and water losses. The large-scale implementations of vegetation construction accelerated the vegetation restoration in the Loess Plateau. Nevertheless, regions with sparse vegetation on the Loess Plateau still covered a great part by 2009, so it is necessary to further strengthen the ecological environment construction in the future.
[40]	易扬, 信忠保, 覃云斌, 等. 生态植被建设对黄土高原农林复合流域景观格局的影响 . 生态学报, 2013, 33(19): 6277-6286. https://doi.org/10.5846/stxb201306071389 URL [本文引用: 1] 摘要 Due to loose soil property and sparse vegetation coverage,the Loess Plateau is a famous soil erosion region in China or even the world. Although widespread ecological restoration efforts have been devoted in past decades,little progress could be seen in the loess region as a result of its unique natural environments. The ecological environment has not been significantly restored until 1999 with the introduction of the Grain for Green Project. As the largest program in history to restore ecological environment,this project has attracted the participation of huge number of farmers. Furthermore,it has involved the most extensive area and the largest amount of domestic investment and as a result,the degraded ecological environment has been gradually restored. With the implementation of the project,it is expected the land use / landscape has been changed accordingly. Prior studies have widely investigated the spatial-temporal dynamics and their driving forces of landscape patterns in medium-to-large watersheds by analyzing Landsat TM remote sensing data. In contrast,studies on landscape patterns in small watersheds through high-resolution remote sensing images are lacking,in particular on the land use dynamics resulted from various ecological restoration projects. Up to now,the Grain for Green projects has been implemented for more than 10 years,and it is now in the benefit consolidation phase. Therefore,an urgent need at present is to quantitatively evaluate the impacts of the ecological restoration project on land use change and landscape patterns through remote sensing and GIS techniques,with which to provide guidelines for future policies and decision making. Based on the Luoyugou watershed in Tianshui(Gansu Province),this study evaluated the dynamic changes of land use and landscape patterns between 2002 and 2008 by analyzing Spot5 remote sensing images. The results show that its land use and landscape patterns have been substantially changed due to the implementation of the Grain for Green project,with a total change of 36.77%. Major changed land use categories were cropland,orchards,young forest plantation,and forestland. In particular,28.95% of the cropland had been converted to other land use patterns during the study period,of which,48. 83% had been converted to orchards and 44.69% to young forest plantation. The changes of landscape patterns showed an obvious spatial distribution,with most of the gentle slope regions being converted to orchards and the medium-to-high altitudes mainly to young forest plantation regions. After the implementation of the Grain for Green project,the ecological landscape in the Luoyugou watershed has been greatly improved,indicating a good sign for future development. [Yi Yang, Xin Zhongbao, Qin Yunbin, et al. Impact of ecological vegetation construction on the landscape pattern of a Loess Plateau watershed. Acta Ecologica Sinica, 2013, 33(19): 6277-6286.] https://doi.org/10.5846/stxb201306071389 URL [本文引用: 1] 摘要 Due to loose soil property and sparse vegetation coverage,the Loess Plateau is a famous soil erosion region in China or even the world. Although widespread ecological restoration efforts have been devoted in past decades,little progress could be seen in the loess region as a result of its unique natural environments. The ecological environment has not been significantly restored until 1999 with the introduction of the Grain for Green Project. As the largest program in history to restore ecological environment,this project has attracted the participation of huge number of farmers. Furthermore,it has involved the most extensive area and the largest amount of domestic investment and as a result,the degraded ecological environment has been gradually restored. With the implementation of the project,it is expected the land use / landscape has been changed accordingly. Prior studies have widely investigated the spatial-temporal dynamics and their driving forces of landscape patterns in medium-to-large watersheds by analyzing Landsat TM remote sensing data. In contrast,studies on landscape patterns in small watersheds through high-resolution remote sensing images are lacking,in particular on the land use dynamics resulted from various ecological restoration projects. Up to now,the Grain for Green projects has been implemented for more than 10 years,and it is now in the benefit consolidation phase. Therefore,an urgent need at present is to quantitatively evaluate the impacts of the ecological restoration project on land use change and landscape patterns through remote sensing and GIS techniques,with which to provide guidelines for future policies and decision making. Based on the Luoyugou watershed in Tianshui(Gansu Province),this study evaluated the dynamic changes of land use and landscape patterns between 2002 and 2008 by analyzing Spot5 remote sensing images. The results show that its land use and landscape patterns have been substantially changed due to the implementation of the Grain for Green project,with a total change of 36.77%. Major changed land use categories were cropland,orchards,young forest plantation,and forestland. In particular,28.95% of the cropland had been converted to other land use patterns during the study period,of which,48. 83% had been converted to orchards and 44.69% to young forest plantation. The changes of landscape patterns showed an obvious spatial distribution,with most of the gentle slope regions being converted to orchards and the medium-to-high altitudes mainly to young forest plantation regions. After the implementation of the Grain for Green project,the ecological landscape in the Luoyugou watershed has been greatly improved,indicating a good sign for future development.
[41]	林忠辉, 莫兴国, 李宏轩, 等. 中国陆地区域气象要素的空间插值 . 地理学报, 2002, 57(1): 47-56. https://doi.org/10.3321/j.issn:0375-5444.2002.01.006 URL [本文引用: 1] 摘要在区域农田生态系统生产力模拟模型研究中,空间插值可以提供每个计算栅格的气象要素资料.然而,在众多的气象要素空间插值方法中,并没有一种适合每一个气象要素的普适的最佳插值方法.本文以全国725站1951～1990年整编资料中的旬平均温度和计算得来的675站的月平均光合有效辐射日总量(PAR)为数据源,选用了距离平方反比法(IDS)、梯度距离平方反比法(GIDS)和普通克立格法 (OK)等3种插值方法,进行了方法选取的探讨.交叉验证结果表明:3种方法中,温度插值的平均绝对误差(MAE)的排序为IDS＞OK＞GIDS,其值分别为2.15℃、1.90℃和1.32℃;在作物生长季节(4～10月),MAE分别2.0℃、1.9℃和1.2℃,表明GIDS法在温度插值方面更具实用价值;对于PAR,MAE的排序为OK＞GIDS＞DS,其值分别为+.83MJ/m2、071MJ/m2和0.46MJ/m2,说明复杂的方法并不必然具有更好的效果.对这2个气象要素的空问分布特征分析表明:温度和PAR的经、纬向梯度和高度梯度均具有明显的季节性变化特征;温度的纬向梯度有近似正弦曲线的较强的季节变化,表现为夏季高,而冬、春季低;温度的高度梯度年内变化范围为一0.0033～0.0048℃/m,GIDS法能较细致地反映温度随海拔高度的变化;PAR也具有明显的纬向梯度,其季节变化特征与温度类似.最后,对中国陆地区域以1'×1'进行插值,生成了中国陆地区域的温度和 PAR的空间分布栅格图. [Lin Zhonghui, Mo Xingguo, Li Hongxuan, et al. Comparison of three spatial interpolation mehtods for climate variables in China. Acta Geographica Sinica, 2002, 57(1): 47-56.] https://doi.org/10.3321/j.issn:0375-5444.2002.01.006 URL [本文引用: 1] 摘要在区域农田生态系统生产力模拟模型研究中,空间插值可以提供每个计算栅格的气象要素资料.然而,在众多的气象要素空间插值方法中,并没有一种适合每一个气象要素的普适的最佳插值方法.本文以全国725站1951～1990年整编资料中的旬平均温度和计算得来的675站的月平均光合有效辐射日总量(PAR)为数据源,选用了距离平方反比法(IDS)、梯度距离平方反比法(GIDS)和普通克立格法 (OK)等3种插值方法,进行了方法选取的探讨.交叉验证结果表明:3种方法中,温度插值的平均绝对误差(MAE)的排序为IDS＞OK＞GIDS,其值分别为2.15℃、1.90℃和1.32℃;在作物生长季节(4～10月),MAE分别2.0℃、1.9℃和1.2℃,表明GIDS法在温度插值方面更具实用价值;对于PAR,MAE的排序为OK＞GIDS＞DS,其值分别为+.83MJ/m2、071MJ/m2和0.46MJ/m2,说明复杂的方法并不必然具有更好的效果.对这2个气象要素的空问分布特征分析表明:温度和PAR的经、纬向梯度和高度梯度均具有明显的季节性变化特征;温度的纬向梯度有近似正弦曲线的较强的季节变化,表现为夏季高,而冬、春季低;温度的高度梯度年内变化范围为一0.0033～0.0048℃/m,GIDS法能较细致地反映温度随海拔高度的变化;PAR也具有明显的纬向梯度,其季节变化特征与温度类似.最后,对中国陆地区域以1'×1'进行插值,生成了中国陆地区域的温度和 PAR的空间分布栅格图.
[42]	杨勤业, 郑度, 吴绍洪. 中国的生态地域系统研究 . 自然科学进展, 2002, 12(3): 287-291. URL [本文引用: 3] [Yang Qinye, Zheng Du, Wu Shaohong. Ecological region system research of China. Progress in Natural Science, 2002, 12(3): 287-291.] URL [本文引用: 3]
[43]	毛德华, 王宗明, 韩佶兴, 等. 1982-2010年中国东北地区植被NPP时空格局及驱动因子分析 . 地理科学, 2012, 32(9): 1106-1111. URL [本文引用: 1] 摘要 AVHRR NDVI at 8 km spatial resolution and MODIS NDVI at 1 km spatial resolution were integrated using per pixel linear regression model method to construct NDVI dataset from 1982 to 2010 at 8 km spatial resolution and covering Northeast China. Based on the constructed NDVI dataset and meteorological data, monthly net primary productivity in Northeast China in1982-2010 were estimated using the CASA model. Simulation accuracy of net primary productivity was above 75% and through accuracy and consistency check. Annual total vegetation net primary productivity in Northeast China is 6.5×10⁸ t/a (carbon), and the distribution of net primary productivity varied from vegetation types, climatic and topography differences. There are obvious difference and spatial heterogeneity for NPP in Northeast China. For the vegetation in different topography areas, net primary productivity decreased in the order of mountains, plains, plateaus. For different vegetation types, net primary productivity decreased in the order of broadleaved deciduous forest and evergreen coniferous forest, mixed coniferous and broadleaved forest and deciduous coniferous forest, crop, grass. Trend analysis at pixel extent showed that the pixels with decreased trends mainly distributed in the transitional zone between typical forest zone and plain, Hulun Buir grassland and Chifeng semi-aird region. Result of spatial variation analysis showed that the NPP of grassland vegetation had the highest changes compared with other vegetation types in the past 29 years. But net primary productivity of grassland have an obvious increase at a whole, this was maybe resulted from the conservative grazing and grassland protection. Net primary productivity exhibited obviously increasing trend in 1982-2010. Consequently, vegetation carbon fixation had enhanced as a whole acquired from the decade analysis. Climatic changes for different parameters are very important affecting factors on spatial pattern and annual dynamic of NPP. Land cover changes in Northeast China during the past 29 years influenced the spatial pattern and annual dynamic for net primary productivity of different vegetation types. Economic development in this region still makes a pressure for vegetation net primary productivity, especially the forest. [Mao Dehua, Wang Zongming, Han Jixing, et al. Spatio-temporal pattern of net primary productivity and its driven factors in Northeast China in 1982-2010. Scientia Geographica Sinica, 2012, 32(9): 1106-1111.] URL [本文引用: 1] 摘要 AVHRR NDVI at 8 km spatial resolution and MODIS NDVI at 1 km spatial resolution were integrated using per pixel linear regression model method to construct NDVI dataset from 1982 to 2010 at 8 km spatial resolution and covering Northeast China. Based on the constructed NDVI dataset and meteorological data, monthly net primary productivity in Northeast China in1982-2010 were estimated using the CASA model. Simulation accuracy of net primary productivity was above 75% and through accuracy and consistency check. Annual total vegetation net primary productivity in Northeast China is 6.5×10⁸ t/a (carbon), and the distribution of net primary productivity varied from vegetation types, climatic and topography differences. There are obvious difference and spatial heterogeneity for NPP in Northeast China. For the vegetation in different topography areas, net primary productivity decreased in the order of mountains, plains, plateaus. For different vegetation types, net primary productivity decreased in the order of broadleaved deciduous forest and evergreen coniferous forest, mixed coniferous and broadleaved forest and deciduous coniferous forest, crop, grass. Trend analysis at pixel extent showed that the pixels with decreased trends mainly distributed in the transitional zone between typical forest zone and plain, Hulun Buir grassland and Chifeng semi-aird region. Result of spatial variation analysis showed that the NPP of grassland vegetation had the highest changes compared with other vegetation types in the past 29 years. But net primary productivity of grassland have an obvious increase at a whole, this was maybe resulted from the conservative grazing and grassland protection. Net primary productivity exhibited obviously increasing trend in 1982-2010. Consequently, vegetation carbon fixation had enhanced as a whole acquired from the decade analysis. Climatic changes for different parameters are very important affecting factors on spatial pattern and annual dynamic of NPP. Land cover changes in Northeast China during the past 29 years influenced the spatial pattern and annual dynamic for net primary productivity of different vegetation types. Economic development in this region still makes a pressure for vegetation net primary productivity, especially the forest.
[44]	孟猛, 倪健, 张治国. 地理生态学的干燥度指数及其应用评述 . 植物生态学报, 2004, 28(6): 853-861. URL [本文引用: 2] 摘要 The aridity Index (AI, here defined as the climatic aridity) is an index tha t de scribes the dry and wet conditions of a site. Historically, the AI has been ofte n used in long-term studies of geography and ecology. Recently it has became one of the most frequently used climatic factors in studies of global change, espec ially in studies of climate change, ariditification and desertification. A total of 22 AIs used worldwide were briefly introduced in this paper, and of thes e ei ght of the more commonly used AIs were evaluated. We discussed their princip les, c alculation methods, their applications in ecological and geographical studies, a nd the advantages and shortcomings of each indice were analyzed based on their a pplications and practice in China. Our analyses indicated that three AIs, th e mo dified Selianinov AI, the de Martonne AI and the Holdridge potential eva potranspiration ratio (another AI to some extent) have clear applicability and significa nce for physics and ecology. These AIs are suitable for characterizing the phys ical environment of China and can be used in future studies of climate change, a ridification and desertification in China. [Meng Meng, Ni Jian, Zhang Zhiguo. Aridity index and its applications in geo-ecological study. Acta Phytoecologica Sinica, 2004, 28(6): 853-861.] URL [本文引用: 2] 摘要 The aridity Index (AI, here defined as the climatic aridity) is an index tha t de scribes the dry and wet conditions of a site. Historically, the AI has been ofte n used in long-term studies of geography and ecology. Recently it has became one of the most frequently used climatic factors in studies of global change, espec ially in studies of climate change, ariditification and desertification. A total of 22 AIs used worldwide were briefly introduced in this paper, and of thes e ei ght of the more commonly used AIs were evaluated. We discussed their princip les, c alculation methods, their applications in ecological and geographical studies, a nd the advantages and shortcomings of each indice were analyzed based on their a pplications and practice in China. Our analyses indicated that three AIs, th e mo dified Selianinov AI, the de Martonne AI and the Holdridge potential eva potranspiration ratio (another AI to some extent) have clear applicability and significa nce for physics and ecology. These AIs are suitable for characterizing the phys ical environment of China and can be used in future studies of climate change, a ridification and desertification in China.
[45]	Raes D. The ETo Calculator Reference Manual Version 3.2: Evapotranspiration from A Reference Surface. Québec: Food and Agriculture Organization of the United Nations, 2009. [本文引用: 1]
[46]	黄秉维. 中国综合自然区划图. 见: 科学出版社. 中国自然保护图集. 北京: 科学出版社, 1989: 20-21. URL [本文引用: 1] [Huang Bingwei.Comprehensive natural zoning map. In: Science Press. Atlas of Natural Protection of China. Beijing: Science Press, 1989: 20-21.] URL [本文引用: 1]
[47]	黄秉维. 中国气候区划与自然地理区划的回顾与展望 . 地理集刊, 1989, 21: 1-9. URL [本文引用: 1] [Huang Bingwei. Retrospect and prospect of climatic regionalization and natural geographical regionalization of China. Collection of Geographical Papers, 1989, 21: 1-9.] URL [本文引用: 1]
[48]	Kohonen T. Engineering applications of SOFM. IEEE Transactions on Neural Network, 1996, 84: 1358-1383. [本文引用: 1]
[49]	李双成. 中国可持续发展水平区域差异的人工神经网络判定 . 经济地理, 2001, 21(5): 523-526. https://doi.org/10.3969/j.issn.1000-8462.2001.05.003 URL [本文引用: 1] 摘要本文在对目前区域可持续发展水平度量方法进行分析的基础上 ,提出以人工神经网络 (ANN)作为区分中国可持续发展水平区域差异的模型工具。构建了ANN模型分析中较为先进的自组织特征映射网络 (Self-Organiz ingMaps,SOFM) ,并以中国 31个省市 (自治区 ) 1996年的社会、经济、资源和环境状况作为待分样本 ,用SOFM进行了可持续发展区域差异的判定。网络运行结果表明 ,1996年中国可持续发展水平的区域差异可分成 5类 ,SOFM分类结果与专家的判断基本近似。可见 ,人工神经网络是一条具有发展和应用前景的途径 [Li Shuangcheng. The analysis on regional differentiations of sustainable development by using artificial neural networks. Economic Geography, 2001, 21(5): 523-526.] https://doi.org/10.3969/j.issn.1000-8462.2001.05.003 URL [本文引用: 1] 摘要本文在对目前区域可持续发展水平度量方法进行分析的基础上 ,提出以人工神经网络 (ANN)作为区分中国可持续发展水平区域差异的模型工具。构建了ANN模型分析中较为先进的自组织特征映射网络 (Self-Organiz ingMaps,SOFM) ,并以中国 31个省市 (自治区 ) 1996年的社会、经济、资源和环境状况作为待分样本 ,用SOFM进行了可持续发展区域差异的判定。网络运行结果表明 ,1996年中国可持续发展水平的区域差异可分成 5类 ,SOFM分类结果与专家的判断基本近似。可见 ,人工神经网络是一条具有发展和应用前景的途径
[50]	Stankiewicz A, Kosiba P. Advances in ecological modelling of soil properties by self-organizing feature maps of natural environment of Lower Silesia [Poland]. Acta Societatis Botanicorum Poloniae, 2009, 78(2): 167-174. URL [本文引用: 1]
[51]	于强, 谢贤群, 孙菽芬, 等. 植物光合生产力与冠层蒸散模拟研究进展 . 生态学报, 1999, 19(5): 156-165. https://doi.org/10.1088/0256-307X/16/12/025 URL [本文引用: 1] 摘要 From the beginning of the 90's,the simulation of plant physiological ecology came into a stage of process-based simulation from empirical methods We generalize and evaluate the basic processes of physiological ecology at leaf and canopy level,and then analyze some famous models and their methods in simplifying and parameterizing The leaf photosynthesis model is based on biochemical model and stomatal conductance model,such as Ball\\|Berry model Canopy model is based on Ross's solar radiation transfer theory and ecological model scaled up from the leaf Because of the change of feedback mechanisms between physiological processes and environment,search for the balance between simplicity for the convenience in usage and complexity in mechanisms is the focus in physiological ecology simulation in the future [Yu Qiang, Xie Xianqun, Sun Shufen, et al. Andances in simulation of plant photosynthetic productivity and canopy evapotranspiration. Acta Ecologica Sinica, 1999, 19(5): 156-165.] https://doi.org/10.1088/0256-307X/16/12/025 URL [本文引用: 1] 摘要 From the beginning of the 90's,the simulation of plant physiological ecology came into a stage of process-based simulation from empirical methods We generalize and evaluate the basic processes of physiological ecology at leaf and canopy level,and then analyze some famous models and their methods in simplifying and parameterizing The leaf photosynthesis model is based on biochemical model and stomatal conductance model,such as Ball\\|Berry model Canopy model is based on Ross's solar radiation transfer theory and ecological model scaled up from the leaf Because of the change of feedback mechanisms between physiological processes and environment,search for the balance between simplicity for the convenience in usage and complexity in mechanisms is the focus in physiological ecology simulation in the future
[52]	黄自立. 陕北地区黄绵土分类的研究 . 土壤学报, 1987, 24(3): 266-271. URL [本文引用: 1] 摘要 Yellow loessial soils are infant soils developed on lossial material, without distinct development of genetic horizons and zonal characteristics of profile. The soils are similar to their parent material. The profile pattern is mainly composed of horizons of Ap—C or A—C. Based the diegnostic horizon and profile, the soils are divided into two subgroups, i.e. primary yellow cultivated loessial soils and lightly yellow cultivated lossial soil. Based on the hydrothermal regimes of the soil in crop growing season (April to Sept.), the soils are subdivided into such soil families as warm-moist, warm-irrigated, warm-wet, warm-dry, cool-moist yellow cultivated lossial soils etc. Based on the texture and O.M. content, the soils are further subdivided into 28 soil species. The soil are mainly distributed in northern Shaanxi with an altitude of lower then 1200 m. [Huang Zili. Discussion of classification of yellow cultivated loessial soils in Shaanxi. Acta Pedoligica Sinica, 1987, 24(3): 266-271.] URL [本文引用: 1] 摘要 Yellow loessial soils are infant soils developed on lossial material, without distinct development of genetic horizons and zonal characteristics of profile. The soils are similar to their parent material. The profile pattern is mainly composed of horizons of Ap—C or A—C. Based the diegnostic horizon and profile, the soils are divided into two subgroups, i.e. primary yellow cultivated loessial soils and lightly yellow cultivated lossial soil. Based on the hydrothermal regimes of the soil in crop growing season (April to Sept.), the soils are subdivided into such soil families as warm-moist, warm-irrigated, warm-wet, warm-dry, cool-moist yellow cultivated lossial soils etc. Based on the texture and O.M. content, the soils are further subdivided into 28 soil species. The soil are mainly distributed in northern Shaanxi with an altitude of lower then 1200 m.
[53]	中国科学院中国自然地理编辑委员会. 中国自然地理系列专著: 中国气候. 北京: 科学出版社, 2013. [本文引用: 2] [China natural Geographic Editorial Committee of the Chinese Academy of Sciences. China natural Geography Series of Monographs: The Climate of China. Beijing: Science Press, 2013.] [本文引用: 2]
[54]	吴绍洪, 杨勤业, 郑度. 生态地理区域系统的比较研究 . 地理学报, 2003, 58(5): 686-694. https://doi.org/10.3321/j.issn:0375-5444.2003.05.006 URL [本文引用: 1] 摘要 Eco-geographical region is a major ecosystem in geographic zonality. A hierarchical system, which is formed by division or combination of natural features based on geographic relativity and comparison of major ecosystem factors and geographic zonality, is called eco-geographic regional system. A comparison of the studies of this topic between Chinese and overseas scientists will enrich ideology of this field, facilitate academic exchange with understandable scientific language and introduce China's progress to the world. The authors compared both China's and overseas eco-geographic regional systems in development process, hierarchical units, delineating criteria, mapping procedures and regions. The results indicate that there are main common points in the subject, such as objectives, research contents and service objects connecting with environment, ecology and global change. Of the existing eco-regional systems, Robert G Bailey's systems of the United States, North America and the continent are comparatively close to China's system in hierarchical units, mapping procedure and regions. [Wu Shaohong, Yang Qinye, Zheng Du. Comparative study on eco-geographic regional systems between China and USA. Acta Geographica Sinica, 2003, 58(5): 686-694.] https://doi.org/10.3321/j.issn:0375-5444.2003.05.006 URL [本文引用: 1] 摘要 Eco-geographical region is a major ecosystem in geographic zonality. A hierarchical system, which is formed by division or combination of natural features based on geographic relativity and comparison of major ecosystem factors and geographic zonality, is called eco-geographic regional system. A comparison of the studies of this topic between Chinese and overseas scientists will enrich ideology of this field, facilitate academic exchange with understandable scientific language and introduce China's progress to the world. The authors compared both China's and overseas eco-geographic regional systems in development process, hierarchical units, delineating criteria, mapping procedures and regions. The results indicate that there are main common points in the subject, such as objectives, research contents and service objects connecting with environment, ecology and global change. Of the existing eco-regional systems, Robert G Bailey's systems of the United States, North America and the continent are comparatively close to China's system in hierarchical units, mapping procedure and regions.
[55]	中国科学院黄土高原综合科学考察队. 黄土高原地区自然环境及其演变. 北京: 科学出版社, 1991. [本文引用: 1] [The Scientific Expedition Team of the Loess Plateau of the Chinese Academy of Sciences. Natural Environment and Its Evolution in the Loess Plateau. Beijing: Science Press, 1991.] [本文引用: 1]
[56]	中国科学院黄土高原综合科学考察队. 黄土高原地区植被资源及其合理利用. 北京: 中国科学技术出版社, 1991. [The Scientific Expedition Team of the Loess Plateau of the Chinese Academy of Sciences. Vegetation Resources and Its Reasonable Utilization in the Loess Plateau. Beijing: China Science and Technology Press, 1991.]
[57]	吴钦孝, 杨文治. 黄土高原植被建设与持续发展. 北京: 科学出版社, 1998. URL 摘要本书在对现有林草植被资源及其效益试验分析的基础上，全面总结了我国40年来黄土高原植被建设的实践和经验，介绍水土流失区强化植被建设的技术和措施，包括低产林改造、经济林建设和高效生产技术、草地生产’持续发展等，提出了植被建设分区、最佳森林覆盖率和可能恢复程度，为该地区林草植被的持续发展，以及政府有关部门治理和开发黄土高原提供科学依据。可供从事林业、草地、水土保持、环境科学等方面的科研、教学人员和管理工 [Wu Qinxiao, Yang Wenzhi.Vegetation Construction and Sustainable Development of Loess Plateau. Beijing: Science Press, 1998.] URL 摘要本书在对现有林草植被资源及其效益试验分析的基础上，全面总结了我国40年来黄土高原植被建设的实践和经验，介绍水土流失区强化植被建设的技术和措施，包括低产林改造、经济林建设和高效生产技术、草地生产’持续发展等，提出了植被建设分区、最佳森林覆盖率和可能恢复程度，为该地区林草植被的持续发展，以及政府有关部门治理和开发黄土高原提供科学依据。可供从事林业、草地、水土保持、环境科学等方面的科研、教学人员和管理工
[58]	彭祥林, 贾恒义. 黄土高原草地土壤生态. 西安: 世界图书出版西安公司, 1997. [本文引用: 1] [Peng Xianglin, Jia Hengyi.Grassland Soil Ecology in the Loess Plateau. Xi'an: The World Book Publishing Company of Xi'an, 1997.] [本文引用: 1]
[59]	侯学煜, 姜恕, 陈昌笃,等. 对于中国各自然区的农、林、牧、副、渔业发展方向的意见 . 科学通报, 1963, (9): 8-26. URL [本文引用: 1] [Hou Xueyu, Jiang Shu, Chen Changdu, et al. Proposals about how to develop agriculture, forestry, animal husbandry, subsidiary business and fishery to each natural areas in China. Chinese Science Bulletin, 1963, (9): 8-26.] URL [本文引用: 1]
[60]	赵松乔. 中国综合自然地理区划的一个新方案 . 地理学报, 1983, 38(1): 1-10. URL 摘要 A new scheme for comprehensive physical regionalization in China is developed and adopted in the newly written textbook "Physical Geography of China" (both in Chinese and in English). Three natural realms (Eastern Monsoon China, Northwest Arid China, Tibetan Frigid Plateau), seven natural divisions (Temperate humid subhumid Northeast China, Warm-temperate humid subhumid North China, subtropical humid Central South China, tropic humid South China, temperate grassland of Inner Mongolia, temperate warm-temperate desert of Northwest China, Tibetan Plateau), and 33 natural regions are demarcated. They are listed in table 2 and shown in map 1. The classification of lower-level regional units (natural sub-regions and natural areas) is also briefly discussed. [Zhao Songqiao. A new scheme for comprehensive physical regionalization in China. Acta Geographica Sinica, 1983, 38(1): 1-10.] URL 摘要 A new scheme for comprehensive physical regionalization in China is developed and adopted in the newly written textbook "Physical Geography of China" (both in Chinese and in English). Three natural realms (Eastern Monsoon China, Northwest Arid China, Tibetan Frigid Plateau), seven natural divisions (Temperate humid subhumid Northeast China, Warm-temperate humid subhumid North China, subtropical humid Central South China, tropic humid South China, temperate grassland of Inner Mongolia, temperate warm-temperate desert of Northwest China, Tibetan Plateau), and 33 natural regions are demarcated. They are listed in table 2 and shown in map 1. The classification of lower-level regional units (natural sub-regions and natural areas) is also briefly discussed.
[61]	席承藩, 张俊民, 丘宝剑. 中国自然区划概要. 北京: 科学出版社, 1984. URL [本文引用: 1] [Xi Chengfan, Zhang Junmin, Qiu Baojian.Outline of Physical Regionalization of China. Beijing: Science Press, 1984.] URL [本文引用: 1]

中国区划工作的回顾与展望

2005

... 自然区划作为地理学的核心议题和重要工作,在认识地理分异规律、社会生产以及区域规划活动中都发挥着极为重要的作用.自然区划着重从区域角度出发,通过观察和研究地表自然综合体以揭示地域分异规律,并探讨自然地域单元的划分与合并^[1].基于不同的研究目标与不同发展阶段,自然区划产生了诸多分支,相关概念也十分丰富,如：综合自然区划^[2]、自然地理区划^[3]、生态地理区划^[4]、生物（态）气候区划^[5]等.其中,生态地理区划是结合了生态学思想,在自然区划的原理和方法上发展起来的新分支,是基于自然地理学的整体性和地域分异规律,从生态学的角度揭示自然生态区域的相似性和差异性,从而对该区域进行系统的整合和分区^[6,7].其中,特别关注类型分区与地域分区,生态地理分区研究是进行生态地理区划的前期阶段与基础工作^[8].相对于自然区划,生态地理区划本质上更加关注具有相似生产潜力、受到相似生态胁迫和具备相似结构特征的生态单元^[4].因此,通过生态地理区划所得到的内部生态特征与生产潜力高度均一性的地块,成为国家及地区生态恢复与重建、自然资源合理开发以及区域可持续发展的基本空间单元. ...

中国区划工作的回顾与展望

2005

中国综合自然区划的初步草案

1958

... 生态地理区划与分区因其综合性和整体性优势,一直以来都是国内外地理学者研究的热点领域之一.在国际上,1976年Bailey对美国生态系统进行了等级划分,第一个具有真正生态地理意义的区划方案自此诞生^[9].此后,Rowe等学者对生态地理分区的原则、指标选择进行了进一步探索^[10-12];Bailey于1989年绘制的全球生态地理区划图又进一步明确了应将气候、地形、植被要素作为生态地理分区的基础^[13].近年来,更多的学者开始将大尺度遥感数据作为生态地理分区的可靠依据^[14],Mackey等基于MODIS数据对澳大利亚进行了生物地理分区,并提出一个适用于全部陆地区域的系统化通用框架^[15];Cruz等则指出使用RBFs神经网络与遥感数据结合的方法可大大提升分区效率,克服传统实地考察法中高时间成本与地域限制的缺点^[16].与此同时,自然植被与农作物作为敏感的气候变化指标成为了辅助区域划分的最佳选择^[17],Gao等借助地统计学思想,基于NDVI将半方差分析与分形维数应用于青藏高原生态地理分区的边界验证^[18];Liu等则对生态地理分区的传统指标体系进行了优化,提出了有机质含量等5个新的土壤指标以替代单纯的土壤类型^[19].在国内,黄秉维等在综合自然区划方面的成果^[2,20]成为中国生态地理区划的雏形;1988年《中国自然生态区划与大农业发展战略》^[21]方案的发表则标志着生态地理区划研究在中国的正式兴起.2000年以来,郑度等采用传统三级划分法、专家集成与模型定量法实现了中国的生态地理区划^[7,22];欧阳志云等着重关注生态敏感性区划在国家层面的应用^[23].近年来,生态地理区划及其相关研究工作仍然十分活跃,但总体而言,现有研究大多注重对单一分区方案的解释与描述,对于不同分区方案之间的对比与优选缺乏定量化途径,优选结果往往依赖研究者的目视比较或主观判定.此外,现有的生态分区多基于传统的专家集成法^[24],使用自上而下的三级演绎途径予以划分^[7,25],区划结果相对于自下而上的综合集成法更容易受到人为影响,并且容易忽略中小尺度的地带性分异规律. ...

... 在分区方法的选择方面,中国现有生态地理分区研究大都选择自上而下的演绎途径从高到低予以划分^[2,42,46,47],而传统分区方法大都有分区过程复杂、耗时、主观性强的缺点.本文使用SOFM神经网络模型实现自下而上的分区,该方法所得到的分区单元界线更加精确并且可靠性较高^[4]. ...

... 生态地理分区及区划是地理学综合研究的核心主题,长期以来受到地理学者广泛关注.区划或分区要素在区域之间差异最大、区域内部差异最小,这是生态地理分区或区划有效性的基本要求.在中国早期的生态区划和自然区划工作中,黄秉维等^[2,59-61]多根据已有生态地理要素图件及研究者对于地理环境的直观认识与经验积累进行;近年来,随着计算机技术的发展,SOFM神经网络成为生态地理分区的重要定量方法^[24,29].而由于SOFM网络自身的特性分类数目的改变会显著影响分区结果,生态地理分区方案的对比与优选因而成为基于SOFM神经网络的生态地理分区关键环节.但是,既有的生态分区、地理区划及土地利用分区等相关研究,大多依据研究区空间分异的主观感知,基于不同方案下的分区图直观判定;仅有部分学者尝试采用聚集度等景观格局指数定量对比不同类型分区的空间聚集程度,基于整体聚集度高低筛选最优分区方案^[28];但这种最适分区判定仅仅考虑了空间形态,未能关涉分区核心要素在不同分区方案之间的对比.因此,本研究提出基于植被动态一致性准则的生态地理最适分区判定方法,以植被动态作为生态地理要素空间分异的地表景观综合判定依据,有助于提升生态地理分区的客观性与准确性. ...

中国综合自然区划的初步草案

1958

中国自然地理区划研究的新进展与发展趋势

2010

中国自然地理区划研究的新进展与发展趋势

2010

生态地理区划研究进展

2006

... [4].因此,通过生态地理区划所得到的内部生态特征与生产潜力高度均一性的地块,成为国家及地区生态恢复与重建、自然资源合理开发以及区域可持续发展的基本空间单元. ...

生态地理区划研究进展

2006

黄土高原生物气候分区与该区生态系统的恢复

2001

... 由于黄土高原地区已有分区方案^[5,7]的总体分区均在5个以上,考虑到区域空间分异表征准确性及其刻画细致性的需要,依据植被动态一致性准则对SOFM网络的11种分区方案（分别划分出5~15类生态地理区）进行筛选.植被动态一致性是指所识别的各个分区之间即使在气温、降水、地形、下垫面甚至植被类型等方面均有差异,但经过黄土高原长时间的植被演替以及1999年退耕还林之后,各分区的植被生长是以相同趋势发展的.植被动态一致性准则从根本上体现了黄土高原近30年来的植被生长特征,而植被的净初级生产力（net primary productivity,NPP）又是NDVI与气温、降水等自然要素综合作用的结果^[51],因此,将其作为多分区方案优选的关键指标.NPP采用基于光能利用率的CASA模型计算得到,数据源为1982-2010年GIMMS-NDVI数据和气象观测数据. ...

... 为进一步探究各生态指数在不同分区中的数量分布状况,本文基于ArcMap 10.2的分区统计工具对各分区的评价指标平均值进行了统计（表2）.一方面对于热量类指数而言,低纬地区且海拔较低的分区IIA1虽然所接收的太阳辐射量并不高,但全年

\begin{matrix} \geq \end{matrix}

10℃的积温日数可以达到200天以上,且最冷、最暖月的平均温度都为6个分区中最高;相反,位于黄土高原最西端的IIIB1分区热量类指数都相对较低,结合地形指数分析可知,该地区处于黄土高原海拔最高的地区,因而气温会随着海拔升高发生了梯度性下降.另一方面对于水分类指数而言,干燥度与降水量相辅相成,共同决定了一个地区的水分收支平衡;黄土高原六大分区从西北向东南整体呈现干燥度降低、降水量升高的现象.将各分区水热条件与海拔、地形等要素结合可知,低海拔、气候湿润温暖的平原地区相应的植被覆盖也最良好^[5],因此黄土高原的NDVI高值区和主要的林区都分布于IIA1区内. ...

黄土高原生物气候分区与该区生态系统的恢复

2001

\begin{matrix} \geq \end{matrix}

The principle and characteristics of ecological regionalization.

1998

2008

... [7,25],区划结果相对于自下而上的综合集成法更容易受到人为影响,并且容易忽略中小尺度的地带性分异规律. ...

... 因此,根据黄土高原的自然基底及土地覆被特征,选取气温、积温日数、太阳辐射、降水、干燥度、NDVI、海拔及植被覆盖等生态地理指标,基于自组织映射网络（SOFM）与GIS的空间分析及制图功能,对黄土高原进行生态地理分区;基于植被动态一致性准则对多种分区方案进行对比,采用NPP判定最优分区方案;并将分区结果与郑度等的相关区划方案^[7,22]以及黄土高原等温线、等降水量线、山脉水系图等对照,以期探索基于植被动态一致性准则的生态地理分区方案优选方法,并为黄土高原生态环境地域分异研究提供基础数据. ...

... 从热量类指数看（图2a~图2d）,黄土高原东南部地区总体较为温暖且太阳辐射充足,而西北部地区则与之相反.并且,黄土高原气温的季相差异明显,冬季（图2b）最低温一般出现在青海东部、内蒙古呼和浩特市与巴彦淖尔市境内,1月最低温仅有-14℃左右,陕北榆林和甘肃兰州、固原等地冬季气温也相对较低（-8℃~-6℃）;而关中地区与河南省西北部的冬季气温则相对最高,最低温在0℃以上.夏季（图2c）较低气温出现在甘肃南部和青海东部,夏季均温仅为12℃~15℃左右,最高温仍分布于西安市等南部地区,夏季高温可达30℃.黄土高原的气温产生季相差异的主要原因是不同地区的地表覆被与地形差异显著,由此影响了冬夏季的地表比热与局地小环境.值得注意的是积温日数指标,由于当日均温稳定升至10℃以上时,喜凉作物开始迅速生长,喜温作物也开始播种,因此其在农业资源评价中有重要作用^[7].由图2a中可以看出,黄土高原东南大部分地区的日均温

\begin{matrix} \geq \end{matrix}

10℃的天数可达200天以上,这些地区的作物可达两年三熟或一年两熟,属于粮食产量较高的地区;而黄土高原西部积温日数则普遍较低,最低地区每年日均温

\begin{matrix} \geq \end{matrix}

10℃的天数仅占全年的20%左右,这些地区的作物多为一年一熟. ...

... 利用非监督性质的SOFM神经网络将黄土高原生态系统分为六大区域,参照中国传统的气候区划方法^[7,53]使用三级划分法进行命名.需要注意的是,传统的分区方法是“逐级推进、层层嵌套”的模式,即首先采用温度指标命名温度带,在温度带的内部用水分类指标划分干湿地区,再在干湿区内部进一步划分自然区^[7,54].在分区时采用SOFM神经网络进行各指标的统一划分,仅使用三级体系进行最终命名. ...

... [7,54].在分区时采用SOFM神经网络进行各指标的统一划分,仅使用三级体系进行最终命名. ...

... 在分区结果的基础上,分别以日均温

\begin{matrix} \geq \end{matrix}

10℃的年积累日数、年均干燥度、植被类型数据为命名温度带、干湿区与自然区的核心因子,其他因子作为辅助指标.各指标的阈值主要参考《中国生态地理区域系统研究》^[7]与《中国自然地理系列专著：中国气候》^[53].最终分区方案及命名结果如表1和图6所示：黄土高原六大生态地理分区中共有两个分区属于中温带,在空间上位于黄土高原西北部;有3个分区属于暖温带,主要位于黄土高原中部及东南部;而高原温带地区位于西部的青海和甘肃境内.从干湿区角度来看,研究区全区仅有东南端的分区IIA1属于半湿润气候,与此相对,西北端沙漠广布的IC1区则是仅有的干旱地区,中部的其他4个分区均属于半干旱气候.根据GLC2000数据对各分区主导植被进行统计可知,干旱少雨的IC1区仅分布有少量荒漠草原,而中部干旱程度稍弱的半干旱地区则以草原为主要植被类型,根据当地坡度、海拔等差异分别有着不同的优势草种：如从西北向东南分别生长有平原草原、坡面草原、高山亚高山草甸,雨水最为丰沛的半湿润IIA1区则以相对优越的自然条件滋养了落叶阔叶林等高大乔木.此外,除天然植被以外,在暖温带的3个分区中分布有较大面积的耕地.总体来说,黄土高原干旱少雨、水土流失较为严重的生态背景更加适宜草类植被的生长,从另一个方面可以推知草类较之林木有更强的耐旱能力,适宜种植的范围更广. ...

... 为讨论基于植被动态一致性准则的黄土高原生态地理分区的合理性,将区划结果与郑度等^[7]2008年公布的中国生态地理区域系统黄土高原部分研究结果进行对比（图7）,并对分区面积的一致性进行统计（表3）.从图7中可以看出,在郑度等的区划方案中,黄土高原共分为三个温度带（中温带、暖温带、高原温带）、三个干湿区（半湿润区、半干旱区、干旱区）和该分类系统下的7个自然小区,总体呈现由东北向西南的条带状分布,本文分区的空间分布与这一特征极为符合;其次,本文对中温带和暖温带边界的识别比较合理,与郑度等的研究结果基本相似,尤其是本研究中温带干旱荒漠及荒漠草原区（IC1）与中温带半干旱平原草原区（IB1）的边界与其对于中温带干旱地区鄂尔多斯及内蒙古高原西部荒漠草原区（IID1）与中温带半干旱地区内蒙古东部草原区（IIC3）的交界线非常接近,面积重合度达78.92%. ...

... Statistics of consistency ratio in area of the Loess Plateau regionalization(%)

		黄土高原生态地理分区
		IB1	IC1	IIA1	IIB1	IIB2	IIIB1
生态地理区域系统^[7]	IIC3	65.16	2.66	0	34.80	10.64	0
	IID1	4.07	78.92	0	0	0	0
	IID2	3.08	6.50	0	2.80	2.23	0
	IIIB3	0	0	27.33	0	6.91	0
	IIIB4	0	0	72.67	0	6.23	13.60
	IIIC1	27.69	11.92	0	62.40	55.92	21.57
	HIIC1	0	0	0	0	18.07	64.83

图7

黄土高原生态地理分区与郑度等的分区方案对比 ...

... （2）通过将SOFM神经网络分类结果与郑度等^[7]的全国区划及黄土高原年均温等值线图、年降水量等值线图以及山脉水系图等对照,以此对SOFM网络的分类结果进行验证.对比结果显示,SOFM神经网络聚类方法所得的自然分区结果与黄土高原自然生态系统的真实分异特征具有较好的一致性,且划分层次明确,区域分割清晰,结果可信. ...

2008

... [7,25],区划结果相对于自下而上的综合集成法更容易受到人为影响,并且容易忽略中小尺度的地带性分异规律. ...

\begin{matrix} \geq \end{matrix}

10℃的天数可达200天以上,这些地区的作物可达两年三熟或一年两熟,属于粮食产量较高的地区;而黄土高原西部积温日数则普遍较低,最低地区每年日均温

\begin{matrix} \geq \end{matrix}

10℃的天数仅占全年的20%左右,这些地区的作物多为一年一熟. ...

... [7,54].在分区时采用SOFM神经网络进行各指标的统一划分,仅使用三级体系进行最终命名. ...

... 在分区结果的基础上,分别以日均温

\begin{matrix} \geq \end{matrix}

... Statistics of consistency ratio in area of the Loess Plateau regionalization(%)

		黄土高原生态地理分区
		IB1	IC1	IIA1	IIB1	IIB2	IIIB1
生态地理区域系统^[7]	IIC3	65.16	2.66	0	34.80	10.64	0
	IID1	4.07	78.92	0	0	0	0
	IID2	3.08	6.50	0	2.80	2.23	0
	IIIB3	0	0	27.33	0	6.91	0
	IIIB4	0	0	72.67	0	6.23	13.60
	IIIC1	27.69	11.92	0	62.40	55.92	21.57
	HIIC1	0	0	0	0	18.07	64.83

图7

黄土高原生态地理分区与郑度等的分区方案对比 ...

基于Kohonen神经网络的中国土地资源综合分区

2006

基于Kohonen神经网络的中国土地资源综合分区

2006

Ecoregions of the United States 1:7500000

1976

Ecological land classification: A survey approach.

1981

Terrestrial Ecozones of Canada: Ecological Land Classification Series NO.19.

1982

Ecological regionalization in Canada and the United States.

1985

Explanatory supplement to ecoregions map of the continents.

1989

Effects of longitudinal range-gorge terrain on the eco-geographical pattern in Southwest China.

2012

Reconciling approaches to biogeographical regionalization: A systematic and generic framework examined with a case study of the Australian continent.

2008

How can neural networks speed up ecological regionalization friendly? Replacement of field studies by satellite data using RBFs

2010

Study on climatic regionalization for mango planting in Guangxi based on "3S" technology

2011

Validating the demarcation of eco-geographical regions: A geostatistical application.

2010

Soil indicators for eco-geographic regionalization: A case study in mid-temperate zone of eastern China.

2009

1992

1988

中国生态区划方案

2001

... （1）热量类指标.具体包括日均温≥10℃的年积累日数、多年1月平均气温、多年7月平均气温、年均太阳辐射等4个指标.热量类指标主要用于黄土高原温度带的划分,其核心指标为日均温≥10℃的年积累日数.因为从生物学的角度出发,温度≥10℃是大多数植物和农作物正常生长的保障,可将≥10℃的天数作为植被的生长期.此外,为反映各地极端温度对黄土高原近30年来植被生长的影响,选择最冷月（1月）和最暖月（7月）平均气温作为辅助指标^[22].太阳辐射指标对植被生长和干物质的积累有着重要作用,并已有学者证实植被NPP与太阳辐射呈显著正相关^[43]. ...

中国生态区划方案

2001

中国生态环境敏感性及其区域差异规律研究

2000

中国生态环境敏感性及其区域差异规律研究

2000

基于GIS与SOFM网络的中国综合自然区划

2011

... 随着计算机技术在自然区划研究中的应用愈发丰富,人工神经网络、遗传算法、模拟退火算法等新技术被广为利用^[24],大大提高了区划工作的效率和客观性.其中,利用自组织特征映射（self-organizing feature maps,SOFM）网络来实现自下而上的地理学及生态学分区研究已经有了部分文献与案例支持.张学儒等构建了地形综合指数、温暖指数、湿润指数、地被指数和水文指数等5个综合指标,采用神经网络与GIS技术对青藏高原东部进行了基于土地类型的区划研究^[26];韩忆楠等以内蒙古自治区农牧交错带为例,从大尺度景观空间分异的角度构建评价指标,运用SOFM网络将该地划分为6个景观分区^[27];Peng等重点关注快速城市化地区的土地利用分区,将不透水表面指数与SOFM技术相结合,划分深圳市土地利用区^[28];郝成元等基于云南省南部30个气象台站多年实测数据,利用SOFM网络识别出哀牢山对北来冷空气南下的屏障作用^[29],是SOFM网络在小样本区划工作中的重要实例. ...

基于GIS与SOFM网络的中国综合自然区划

2011

中尺度对地观测系统支持下中国综合自然地理区划新方法论研究

2004

中尺度对地观测系统支持下中国综合自然地理区划新方法论研究

2004

基于SOFM神经网络模型的土地类型分区尝试: 以青藏高原东部样带为例

2013

基于SOFM神经网络模型的土地类型分区尝试: 以青藏高原东部样带为例

2013

基于自组织特征映射(SOFM)网络的农牧交错带景观分区: 以内蒙古自治区为例

2013

基于自组织特征映射(SOFM)网络的农牧交错带景观分区: 以内蒙古自治区为例

2013

Integrated urban land-use zoning and associated spatial development: Case study in Shenzhen, China.

2014

基于SOFM的区域界线划分方法

2008

基于SOFM的区域界线划分方法

2008

黄土高原生态脆弱贫困区生态经济发展模式研究: 以甘肃省定西地区为例

2003

... 黄土高原作为世界上黄土分布面积最大、黄土地貌最发育的地理单元,生态环境问题异常突出,成为中国乃至世界上水土流失最严重、生态环境最脆弱的地区之一^[30,31],受到国内外众多地理学者的广泛重视.为缓解当地恶劣的自然环境与人地关系,自20世纪50年代起,中国诸多重点生态建设项目均在此实施,如1986-1990年“黄土高原综合治理”项目^[32],1999年启动的退耕还林（草）项目^[33],针对大型天然林种植与保护的“天保”工程^[34]及“三北”防护林工程^[35],缓解水土流失的小流域治理工程等.总体而言,得益于多个生态建设项目的顺利实施,黄土高原地区植被覆盖在近年来整体呈现增长趋势^[36].对于自然条件严苛的黄土高原来说,植被恢复直接影响着区域水土流失强度与土地生产能力,因此植被生长状况可以作为当地生境质量改变的“指示剂”,选择其作为黄土高原区划工作的依据并进一步参与分区方案优选,所得到的黄土高原生态地理分区不仅具有传统自然区划的科学内涵,而且能体现研究区特殊的生态现状和政策背景;而通过详细了解各生态小区的植被恢复空间分异特征,也将对退耕还林等综合治理项目在黄土高原的进一步实施产生促进作用,对于因地制宜地进行生态恢复与规划管理也有着十分重要的意义. ...

黄土高原生态脆弱贫困区生态经济发展模式研究: 以甘肃省定西地区为例

2003

黄土区不同类型土壤微生物量碳、氮和可溶性有机碳、氮的含量及其关系

2006

黄土区不同类型土壤微生物量碳、氮和可溶性有机碳、氮的含量及其关系

2006

黄土高原综合治理分区

1985

黄土高原综合治理分区

1985

黄土高原退耕还林(草)环境效应分析

2005

黄土高原退耕还林(草)环境效应分析

2005

2015

2015-04-27

How ecological restoration alters ecosystem services: An analysis of carbon sequestration in China's Loess Plateau.

2013

气候变化和人类活动对黄土高原植被覆盖变化的影响

2007

... 黄土高原地处半干旱半湿润气候带,水土流失严重,生态环境脆弱,是黄河泥沙的主要来源,也是中国水土保持重点区域^[37].黄土高原位于中国地势第二级阶梯,地处中国中部偏北,形状略呈矩形,约处于北纬34°~41°N,98°~114°E.包括宁夏与山西全境、陕西中北部,甘肃的陇中和陇东地区,青海东北部以及内蒙古的河套平原和鄂尔多斯高平原,总面积达64.40×10⁴ km²（图1）. ...

气候变化和人类活动对黄土高原植被覆盖变化的影响

2007

基于遥感、GIS和模型的黄土高原生态环境质量综合评价

2006

... 黄土高原海拔1200~1600 m,地势南高北低且沟壑纵横,绝大多数地表被深厚的黄土层覆盖,具有塬、梁、峁等典型黄土地貌,土层最厚可达150~180 m.黄土高原的气候特征具有明显的地带性^[38],气温大致随着纬度升高与地势抬升逐渐降低,降水的分布也依循这一趋势,且具有强烈的季节变化和年际变化,6-9月降水占全年的70%~80%.总体来说,黄土高原属于中国东部湿润区与西北干旱半干旱区的过渡地带,由于区内地形、季风以及距海洋的远近不同,其内部在生物气候上存在着复杂的地域分异. ...

基于遥感、GIS和模型的黄土高原生态环境质量综合评价

2006

近30a黄土高原植被覆盖时空演变监测与分析

2011

... 对于生境恶劣的黄土高原来说,植被覆盖对其可持续发展具有至关重要的作用,在1999年退耕还林（草）政策大规模实施以前,黄土高原植被覆盖以小幅波动为主,此后NDVI显著增强^[39];并且,通过耕地向林地、草地等其他用地类型的转化,黄土高原景观生态结构均得以优化^[40].同时,植被恢复直接使得黄土高原地区水土流失、土地退化等问题得到显著缓解.因此,基于植被多年变化的动态一致性准则进行生态地理分区方案优选,有助于进一步了解黄土高原特殊的生态基底与政策背景. ...

近30a黄土高原植被覆盖时空演变监测与分析

2011

生态植被建设对黄土高原农林复合流域景观格局的影响

2013

生态植被建设对黄土高原农林复合流域景观格局的影响

2013

中国陆地区域气象要素的空间插值

2002

... 主要数据来源及处理方法：① 气象数据来自中国气象科学数据共享服务网（http://cdc.cma.gov.cn/）提供的全国756个气象站点1982-2010年的气温和降水量数据,为保证数据精度,选取黄土高原及周边共170个站点的数据;根据林忠辉等对气象要素插值方法的对比结果^[41],采用ArcMap 10.2中的Kriging法进行插值;② 全国范围1 km分辨率DEM数据来源于“黑河计划数据管理中心”（http://westdc.westgis.ac.cn）;③ 1982-2010年逐月NDVI数据来自寒区旱区科学数据中心（http://westdc.westgis.ac.cn/data/1cad1a63-ca8d-431a-b2b2-45d9916d860d）的NOAA/AVHRR-GIMMS数据集;④ 全国1 km分辨率植被类型数据来自SPOT4卫星GLC2000数据集（http://westdc.westgis.ac.cn/data/f1aaacad-9f42-474e-8aa4-d37f37d6482f）,中国植被共分为22类,黄土高原地区共有17类;⑤ 土壤类型数据来自地球系统科学数据共享平台（http://www.geodata.cn/Portal/metadata/viewMetadata.jsp?id=100101-10641）的全国1

\begin{matrix} ∶ \end{matrix}

400万土壤类型分布图,分类方法依据中国1990年土壤系统分类方案.所用栅格数据统一具有GCS-WGS-84地理坐标系下的经纬度信息. ...

中国陆地区域气象要素的空间插值

2002

\begin{matrix} ∶ \end{matrix}

400万土壤类型分布图,分类方法依据中国1990年土壤系统分类方案.所用栅格数据统一具有GCS-WGS-84地理坐标系下的经纬度信息. ...

中国的生态地域系统研究

2002

... 生态地理分区由自然分区演化而来,指标体系在原有基础上有所继承和发展.生态地貌特征作为分区的重要基础^[42],生物与土壤特征是其在地表的直观表达,这又在根本上决定于自然基底的水热条件,因而生态地理区的初级单位划分通常选择温度、水分、地形等作为划分依据;在高级单位的划分时更多参考植被等生态指标.综合考虑指标的合理性以及数据的时效性,选择四大类9个指标作为黄土高原生态地理分区的基础指标. ...

... （4）地表覆被类指标.具体包括年均NDVI和土壤类型两个指标.土壤作为连接大气圈、岩石圈、水圈与生物圈的关键角色,其质地直接决定了上覆植被的类型;并且退耕还林政策的顺利实施也必须考虑所选育树种、草种对于当地土壤的适宜性.而NDVI数据是黄土高原30年来植被演替的最直观表征,因此,天然植被作为黄土高原生态地理分区的重要指标,直接指示了不同下垫面与水热条件下生态恢复的波动状况^[42]. ...

中国的生态地域系统研究

2002

1982-2010年中国东北地区植被NPP时空格局及驱动因子分析

2012

1982-2010年中国东北地区植被NPP时空格局及驱动因子分析

2012

地理生态学的干燥度指数及其应用评述

2004

... （2）水分类指标.具体包括年均降水量、年均干燥度等两个指标.水分与热量条件共同决定自然地表区域性差异.在水分类指数中,年降水量能够反映区域水分的补给状况,200 mm、400 mm、800 mm等降水量线也被认为是划分干旱地区、半干旱地区、半湿润地区、湿润地区的分界线.而干燥度作为划分研究区干湿区的核心因子,反映了区域内水分收支状况,并间接反映了除蒸发以外可用于植物生长的水资源量,一般使用潜在蒸散和年降水量的比值得到^[44].在具体计算时,潜在蒸散ET₀使用联合国粮农组织（FAO）于1998年修正的Penman-Monteith公式计算^[45],具体公式如下所示： ...

... 干燥度（图2e）和年均降水（图2f）作为水分类指数,在分区中主要用干湿地区命名.黄土高原地区以其水土流失和干旱少雨广为人知,全区多年平均降雨量约为466 mm,并呈现从东南向西北递减的趋势,极低值多出现在内蒙古乌兰布和沙漠及毛乌素沙地,极高值出现在河南洛阳及关中地区.对于干燥度指数来说,其本质上代表着某一地区水分支出与收入的比值,数值越大代表该地越干旱.一般来说,当干燥度

\begin{matrix} < \end{matrix}

1.0时为湿润地区,[1.0,1.7)和[1.7,3.0)分别为半湿润及半干旱地区,

\begin{matrix} \geq \end{matrix}

3.0则为干旱地区^[44];从图2e中可知,黄土高原干燥度在1.46~6.95之间,因此跨越了半湿润、半干旱、干旱三个干湿区,总体呈现东西方向的条带状分布,从具体数值来看,干燥度<1.7的地区多位于黄土高原东南部,而西北地区则最为干旱,干燥度大多

\begin{matrix} \geq \end{matrix}

3.0.对干燥度和降水量的划分对黄土高原最终的分区命名起到了重要的作用. ...

地理生态学的干燥度指数及其应用评述

2004

\begin{matrix} < \end{matrix}

1.0时为湿润地区,[1.0,1.7)和[1.7,3.0)分别为半湿润及半干旱地区,

\begin{matrix} \geq \end{matrix}

\begin{matrix} \geq \end{matrix}

3.0.对干燥度和降水量的划分对黄土高原最终的分区命名起到了重要的作用. ...

The ETo Calculator Reference Manual Version 3.2: Evapotranspiration from A Reference Surface.

2009

1989

中国气候区划与自然地理区划的回顾与展望

1989

中国气候区划与自然地理区划的回顾与展望

1989

Engineering applications of SOFM.

1996

... SOFM神经网络又称自组织特征映射网络,属于人工神经网络的一种^[48].包括竞争层和输入层两部分^[49],其中输入层用于接收样本,竞争层对样本进行分类并输出分类结果.SOFM神经网络可以直接对未知的输入数据进行学习或模拟,而并不需要进行先验学习^[50];因此与传统的三级分区方案相比,SOFM所形成的分类中心能保证原数据的拓扑不变性,是一种非监督型的模式识别技术,可以克服传统专家经验集成的区划范式下主观性强的缺陷.基于计算机强大的计算能力,该方法也能够快速把握区域的分异特点.SOFM编程及数据处理在专业数学软件MATLAB R2014a中完成,根据四类指标构建9×9297输入层数据,初始权值为[0,1]的随机数,基本学习速率为0.1,最大循环次数1000次,其他参数采用默认值. ...

中国可持续发展水平区域差异的人工神经网络判定

2001

中国可持续发展水平区域差异的人工神经网络判定

2001

Advances in ecological modelling of soil properties by self-organizing feature maps of natural environment of Lower Silesia [Poland].

2009

植物光合生产力与冠层蒸散模拟研究进展

1999

植物光合生产力与冠层蒸散模拟研究进展

1999

陕北地区黄绵土分类的研究

1987

... 地形和地表覆被决定了黄土高原生态系统自然景观的差异性,其中土壤（图2g）和地形（图2i）直接决定了所生长植被的种类与长势.总体来说,黄土高原的土壤以壤质普通黄绵土为主,广泛分布于山西太原至甘肃定西一带;该土种的主要特征是剖面发育不明显,土壤侵蚀较为严重,但可耕作性好,适宜长期耕种^[52].黄土高原总体呈西高东低的地势特征,黄河流域沿线海拔最低,总体分为陇中高原、陇东陕北高原、山西高原及渭河平原.由此可知,黄土高原无论在下垫面条件抑或是气候水热条件方面都有着极大的区域差异,不同的土壤和坡度也导致了退耕还林过程中因地制宜的植被恢复模式.NDVI作为植被生长状况的最直接表征,在水热充足、地势平缓的东南部呈现出较高的植被覆盖度,延安的子午岭山区也有着大片密集的森林资源;植被覆盖较低的地区主要分布在黄土高原北部的干旱沙地,而河套平原则因相对充足的水资源而成为了北部的“绿岛”. ...

陕北地区黄绵土分类的研究

1987

2013

... 在分区结果的基础上,分别以日均温

\begin{matrix} \geq \end{matrix}

2013

... 在分区结果的基础上,分别以日均温

\begin{matrix} \geq \end{matrix}

生态地理区域系统的比较研究

2003

生态地理区域系统的比较研究

2003

1991

... 除与已有分区方案进行对比以外,将分区结果与黄土高原年均温等值线图、年降水量等值线图以及山脉水系图等对照^[55-58],结果发现：本文分区方案对于水分类指数与热量类指数的分区效果较好,在各区域间体现出了明显的梯度变化性;并且将六大分区在温度带上总体划分为暖温带、中温带与高原温带的分类方法较为合理,符合黄土高原由东南至西北的气候变化特征.结合黄土高原年降水量等值线图发现：划分的三类干湿区与该地的400 mm、200 mm等降水量线由东南向西北延伸的趋势相符.总体而言,通过与已有黄土高原自然要素图件对比,本文的分区方案基本反映了黄土高原的自然地理界线. ...

1991

1998

1997

对于中国各自然区的农、林、牧、副、渔业发展方向的意见

1963

对于中国各自然区的农、林、牧、副、渔业发展方向的意见

1963

中国综合自然地理区划的一个新方案

1983

中国综合自然地理区划的一个新方案

1983

1984

基于植被动态的黄土高原生态地理分区

Eco-geographical regionalization in Loess Plateau based on the dynamic consistency of vegetation

1 引言

2 研究区概况与数据来源

2.1 研究区概况

2.2 数据来源

3 研究方法

3.1 生态地理分区指标体系

3.2 生态地理分区方法

4 结果分析

4.1 黄土高原生态地理要素空间分异

4.2 黄土高原生态地理分区方案优选

4.3 黄土高原生态地理分区方案

5 讨论

5.1 黄土高原生态地理分区方案对比

5.2 基于植被动态的生态地理分区方法适用性

6 结论

参考文献 文献选项 原文顺序 文献年度倒序 文中引用次数倒序 被引期刊影响因子

参考文献

原文顺序

文献年度倒序

文中引用次数倒序

被引期刊影响因子