【作者】 石淑芹；

【导师】 陈佑启；

【作者基本信息】 中国农业科学院 ， 农业遥感， 2009， 博士

【摘要】 随着空间信息获取及处理技术、计算机技术的发展、数理统计理论及方法的应用及深入,在面向全球变化、自然资源、生态、环境及人文等领域的对地观测系统数据丰富发展的情况下,将观测数据转化为区划相关的知识,将经典区划方法论与计算科学有机结合起来;在继承经典区划方法论的基础上,创造以计算机技术、空间信息技术等为代表的具有信息时代特点的区划新方法论,已经成为一种必然趋势和客观要求。本文以吉林省玉米为例,将空间分析技术、多源数据、地统计学理论、数理统计模型、作物生产潜力模型以及传统区划理论与方法等多种理论、技术及方法相结合,在地理信息系统和气象、土壤、农业统计等基础数据支持下,对研究区域的作物生产力区划展开研究。与国内外研究相比较,本文主要在以下两个方面有所创新:⑴在利用现代空间信息技术、基于定量分析的地统计理论、数理统计方法、作物模型与传统区划理论与方法相结合方面具有创新性;⑵在基于多源数据的区划要素空间化上进行了探索并取得了新进展,区划要素数据在获取、空间化及精度方面有所提高。主要结论有以下几点:⑴玉米生产力区划指标体系根据作物生产力区划的基本理论及目的,遵循作物生产力区划的原则,参考学术界已有的研究成果,结合作物生产力的形成要素及过程,从基础地理要素类指标、作物各级生产潜力、作物现实生产力、自然灾害类指标、利用与管理类指标及作物品质类指标等方面构建指标体系。⑵基于多源数据的区划要素获取及空间化①气象要素数据空间化在对气象要素与经纬度、海拔、坡度、坡向、地形遮蔽度等地理要素进行的相关性分析的基础上,从克里金插值方法、协克里金插值方法以及多远回归+残差内插相结合的三种方法出发,分别采用三种方法进行气象要素的空间化研究,并且对其精度进行对比及验证,最后选择精度较高的方法作为应用于区划的1km×1km分辨率栅格的气象要素数据。②土壤养分数据空间化利用ArcGIS软件的地统计模块,结合GIS的空间数据管理功能和分析功能,研究吉林省土壤中pH值、有机质、速效钾、碱解氮和速效磷元素的空间变异状况,并且在对这几种元素与土壤中其他土壤性质如有效硼(B),有效钙(Ca),有效镁(Mg),有效铜(Cu)和有效锌(Zn)等,以及与经纬度、海拔、坡度、坡向、地形遮蔽度、剖面曲率和平面曲率等地理要素进行相关性分析的基础上,通过克里金插值方法、基于土壤类型和微量元素的协克里金插值方法的比较,最终选择基于土壤类型和微量元素的协克里金插值方法绘制了1km×1km的五种土壤养分元素空间分布图。③统计单产数据空间化结合吉林省农业综合分区,将研究区域划分为Ⅰ区(吉林省)、Ⅱ区(中部平原农业区和西部平原农牧区)、Ⅲ区(东部半山农林区和长白山地林农区);利用GIS软件,应用多元回归方法分别对Ⅰ区、Ⅱ区和Ⅲ区这三个不同区域中建立农业统计数据的玉米单产和这些环境要素之间的多元回归分析模型,应用多元回归方法进行吉林省粮食单产数据的空间化,生成1km×1km分辨率栅格的粮食单产,并采用统计单产的数据对其进行了修正。④玉米生产潜力基于GIS平台以及由前面分析得到的气象要素、土壤要素方面的数据,对作物生产力逐步衰减模型中对温度、降水和土壤有效系数订正,并与ArcGIS空间分析中GRID建模相结合建立了作物生产潜力模型,在利用已建立的模型计算出吉林省作物各级生产潜力和资源有效系数的基础上,对吉林省作物生产潜力进行了分级。⑶玉米生产力区划研究按照作物生产力区划的目的、原则、指标体系及方法,结合了作物生长机理和生态学理论中物质迁移和能量流动的特点、根据区域形成的自然地理过程和条件以及社会经济发展的差异,利用已有的自然、地形地貌、农业综合分区、气候等区划方案均作为本次分区过程中的辅助和校正材料,以地理信息系统、经验判别以及自上而下逐级划分和自下而上组合相结合的方法,将吉林省划分为6大区域:Ⅰ区(吉林省中部生产力最高区)、Ⅱ区(吉林省中部和东部生产力较高区)、Ⅲ区(吉林省西部和中部生产力中等区)、Ⅳ区(吉林省西部和中部生产力较差区)、Ⅴ区(吉林省东部半山农林区和长白山地林农区生产力较差区)、Ⅵ区(吉林省长白山地林农区生产力最低区)。相对潜力较高的地区主要分布在Ⅰ区和Ⅱ区,相对潜力较差的地区主要分布在其他地区;各级潜力系数的空间分布都比较分散。对不同类型区域,应选择其相应的适度开发对策,最终使吉林省玉米生产实现经济与生态的协调发展。更多还原

【Abstract】 Along with the promotion and development of spatial information acquisition and data processing, computation technology and statistical theory, how to transform these space-based observation data, from various fields such as global change, natural resources, ecology, environment and humanity, etc. into knowledge or rule for regionalization, combine the classical theory and methodology of regionalization with the computation science to develop the integrated methodology of regionalization, have already become a kind of inexorable trend and current demand.Selected Jilin Province maize as a case study, this paper conducted the study on crop productivity regionalization, by integrating with the GIS technique, multiple-source based data acquisition, geostatistics theory, mathematical statistics model, crop potential productivity model and traditional regionalization theory and method. Compared with the related literature both at home and abroad, this paper mainly innovated in two respects:⑴The paper integrated the spatial information technology, geostatistical theory, statistical method, crop potential productivity model and traditional regionalization theory and method based on quantitative analysis method innovatively;⑵The paper explored and advanced the spatialization technology of the elements in regionalization based on multiple source data, the precision of the elements in regionalization have been improved in aspects of data acquisition and spatialization. The main conclusions of this paper are briefly described as below:⑴Indicator system of regionalization of maize productivityAs the first step toward our final goal, we first considered to develop an indicator system of maize regionalization. According to basic theories, purpose and principle of crop productivity regionalization, and the background of out quantitative knowledge of the environmental factors of maize productivity, the paper established an indicator system of regionalization of maize productivity, which included geographical environmental factors, crop potential productivity, current crop productivity, natural hazards index, utilization and the management index and crop type index, etc.⑵Data acquisition and spatialization of the regionalization element①meteorological conditionsThrough the correlation analysis among the environmental variables, such as meteorological variables, longitude, latitude, elevation, slope, aspect and hillshade, etc, the paper try to characterize the spatial distribution of the meteorological element. By contrast of the precision and validation of the result from three interpolation methods, including kriging, cokriging and multivariate regression with residues kriging, the paper chose the spatial interpolation method with highest precision to map spatial distribution of the meteorological element with a spatial resolution of 1km.②soil nutrientsUsing GIS and geostatistics technique, the paper quantified the spatial variability characteristics of soil pH,soil organic matter, available K, available P and available NH4 in the soil in study area,at the same time , the correlation among these main soil nutrients and other soil properties were analyzed, such as available boron(B) ,available calcium(Ca), available magnesium(Mg), available copper(Cu), available zinc (Zn), etc, and the correlation with the geographical elements was also calculated, such as longitude, latitude, elevation, slope, aspect, plane curvature and profile curvature, etc. Then from the contrast of the results derived from the method of kriging and the cokriging applied to the soil type and microelements, the paper determined the cokriging as spatial interpolation method for the spatial distribution of five types of soil nutrient elements with each grid being 1km by 1km.③maize productivityAccording to the previous agricultural regionalization of Jilin Province, the paper first construct a framework of regionalization with the districtⅠ(Jilin province), districtⅡ(middle plain farming region and western plain farming and pastoral region), districtⅢ(agricultural forest zone of Mid-levels in the east and forest agricultural zone of Changbai Mountain). Then on a GIS platform, the multivariate regression model was developed to simulate the relationship between the statistics-based maize yield and the geographical environmental elements, while the correlation analysis were conducted to validate the result of modeling. Finally, the spatial distribution of maize yield was mapped with a spatial resolution of 1km .④Maize potential productivityUsing spatial analysis technique embedded in GIS platform, combined the variables of meteorological conditions and soil nutrient status, we developed a crop potential productivity model by an combination of the traditional crop productivity attenuating model and the GRID model in GIS, and to quantify the spatial distribution of the maize potential productivity.⑶Maize productivity regionalizationOn the understanding basis of the purpose, principle, the above indicator system and method of crop productivity regionalization, fully considering the mechanism of crop growth and ecological theory on material movement and energy flow, spatial heterogeneity of physical geographic processes, condition and status of social economic development, the advancement of current natural, topographical, agricultural and climatic regionalization, the paper developed the methodology for maize productivity regionalization using GIS technique, experience criterion, in conjunction with the general rules of aggregation (from bottom to top with the partition and from top to bottom). According to the steps mentioned above, six different regions were divided in Jilin Province: DistrictⅠ(highest productivity of the middle and eastern region of Jilin Province), districtⅡ(Second highest productivity of middle and east region of Jilin Province), districtⅢ(moderate productivity of west and middle region of Jilin Province), districtⅣ(low-moderate productivity of west and middle region of Jilin Province), districtⅤ(lower productivity of agricultural forest zone of Mid-levels in the east and forest agricultural zone of Changbai Mountain in Jilin Province), districtⅥ(lowest productivity of forest agricultural zone of Changbai Mountain in Jilin Province). The higher relative potential productivity distributed mainly in the regions of districtⅠand districtⅡ,and the lowest relative potential productivity distributed in the other areas extensively; The spatial or geographic property of the coefficient of potentiality was scattered over the Jilin province. As a general conclusion, the corresponding appropriate countermeasure of agricultural development should be adopted according to the characteristic of different region, and some scientific-based support should be provided for maintaining and promotion the sustainable development of regional economic and ecological system.更多还原