基于地理回归的农作物播种面积统计数据空间化方法

农作物空间格局反映了农作物种植结构和特征,是了解生产资源利用状况及进行农业结构调整的重要依据。研究旨在探索农作物播种面积统计数据空间化的方法,实现对历史农业统计数据的空间化表达。将传统的农业统计调查与先进的遥感技术、空间地理信息技术相结合,通过多元Logistic回归分析农作物格局与自然地理因素和社会经济因素之间的关系,构建农作物空间适宜性分布概率,在此基础上将农作物播种面积统计数据利用空间迭代分配方法,实现统计数据空间化的研究。论文选取中国东北三省作为方法研究区,实现了对该地区2000—2010年农作物播种面积空间化表达,经检验,该方法对东北三省水稻空间化精度为0.76,能够较好地完成农作物时空播种面积统计数据空间化工作。该方法是农作物调查和遥感时空格局解译研究的有效补充,为丰富农作物空间数据提供了技术手段。

Spatialization of Statistical Crop Planting Area Based on Geographical Regression

The sptial pattern of crops reflects the planting structure and characteristics of crops, which is an important basis for understanding agricultural resource utilization and adjusting crop planting structure. This study aims to explore the method for specializing statistical data of crop planting area, and thus spatially express historial agricultural statistics data. This study used the traditional agricultural statistical survey data and remote sensing imagery data with geographic information technologies. The spatial probability distributions of suitabilities of crops are estimated using the Binary Logistic regression analysis that characterizes the relationships between the crop planting structure and the geographical factors as well as social-economic factors. Based on the spatial probability distribution, the statistical data of crop planting area were spatially distributed by using spatial iteratative allocation. Northeast China was taken as the study area and the spatial expression of sown area in this area during 2000-2010 was completed. The spatial accuracy of 0.76 was achieved by using this multi-scale and multi-resolution analysis method, which demonstrated it is superior in spatially expressing statistical data of crop planting. The method can be taken as an effective complement for crop field survey and remote sensing-based crop interpretation, and thus provides novel technical means for enriching crop spatial data.