基于网格的耕地质量分等成果省级汇总方法

Cultivated land quality grading results integration method at provincial level based on grid

耕地质量分等成果由县、省、国家3级成果组成,省级成果由县级成果汇总获取,是农用地分等工作中必不可少的环节。由于传统的汇总方法对汇总底图的过分依赖性,导致汇总效率低下,汇总后的质量完全取决于汇总底图的质量。基于该缺陷,该文提出利用网格取代汇总底图,将县级成果按照"权属-地类-等别"进行分类,根据面积占优原则确定对应网格所属类别,并通过空间最邻近方法寻找距离县级单元最近的网格单元,建立"网格-县级分等单元"的追溯关系。该文提出了以县行政代码、网格所在行列码、等别、地类组成21位省级汇总网格的编码方法。最后利用平均中心和标准差椭圆检验汇总前后耕地空间分布的相似性。以北京市大兴区为研究区进行了实例验证,大兴区根据"地类-自然等-利用等-经济等"组合共分为8类,根据大兴区耕地的分布,最终确定688个有效耕地网格。利用该文方法实现了省级汇总。经过检验表明,汇总前后耕地分布的平均中心和分布方向仅相差405.6 m和0.34°,说明该汇总方法保证了汇总前后耕地空间分布的一致性。与依赖汇总底图的方法相比较,省去了制作底图的工作量,提高了工作效率,并且将省级耕地定位在每个网格中,通过网格编码能够快速查找到对应位置的耕地质量,方便管理和应用。

The results of cultivated land quality gradation are made of the results of county, provincial, national levels. The results at the provincial level are integrated by the results at the county level. Provincial integration is an essential link in cultivated land quality grading work. The traditional provincial integration methods excessively depend on the provincial integration base map, which makes the integration efficiency low. Moreover, the quality of integration work completely depends on the quality of integration base map. Based on these defects, this paper comes up with a method that uses the grid instead of the integration base map. In provincial integration work, we need to build traceability relation between county units and provincial units. It requires all county units can find provincial units and all provincial units have their corresponding county units. Moreover, there is a principle we have to abide by： the ownership, the land type and the grade should be coincident between the county units and corresponding provincial units. That is ＂ownership-land type-grade＂ coincident principle. The purpose of this method is to make spatial distribution of provincial cultivated land quality is similar to that of the county when abiding by this principle. First we need to classify the county results according to ＂ownership-land type-grade＂. Then we determine the corresponding grid category according to the area dominant principle. Through the spatial nearest neighboring method, we find the grid unit closest to the unit at the county level to build ＂Grid-County grading unit＂ relations. But maybe there are some categories which can＇t be found in the provincial grids. Based this situation, the barycenter of the missing category was firstly calculated, then an uncultivated area grid closest to the barycenter was searched out, and lastly this grid was marked as the missing category. The grid coding method proposed in this paper stipulates that grid code is formed by county＇s administrative code, the grid rank, quality gradation and land class code. The length of the code is 21. At last, this paper uses the mean center and standard deviational ellipse to test the similarity of the spatial distribution of cultivated land before and after integration. This paper takes Daxing district, Bejing as a study area. By ＂ownership-land type-grade＂, there are 8 categories in Daxing. According to the distribution of cultivated land in Daxing district, 688 valid cultivated land grids were finally determined. Using the method in this article, we completed the integration at the provincial level for Daxing. After test, it showed that the difference of mean center between provincial grid results and county results is only 405.6 m, and the difference of distribution direction is only 0.34°, which means that this method can ensure the consistency of spatial distribution of cultivated land at different levels. Compared with the method which depends on the integration base map, the method based on grid doesn＇t make base map, reduces workload and improves work efficiency. This method orientates the provincial cultivated land into each grid and codes for these grids. Through grid code, we can quickly search cultivated land quality in the corresponding position, which is very convenient for both management and application.