关键词: 物候信息图谱, 种植模式, 种植模式谱, 贝叶斯网络, 江汉平原

Abstract:

【Objective】 Cropping patterns are a summary of crop sequence which reflects the use patterns and efficiency of cropland resources. Through analyzing the phenological differences of different cropping patterns, the cropland phenology information atlas and cropping pattern spectrum for croplands on the Jianghan Plain were constructed, and the major cropping patterns in this area were mapped. 【Method】 The cropland phenology information atlas including different cropland use patterns was formed through expressing graphically the spatial difference between vegetation index states and cropland use patterns, according to the prior knowledge of crop planting and the phenological differences of different cropping patterns under the framework of geo-information atlas. Taking the major cropping patterns on the Jianghan Plain as the study cases, the vegetation index states in the key phenological periods were arranged and combined to establish the information remapping rule from the vegetation index states to the cropping patterns, their phenological characteristics were explored, and the cropping pattern spectrum was constructed. Then the data during the key phenological periods and phenological knowledge were integrated to map cropping patterns on the Jianghan Plain by using the Naive Bayes Networks. The vegetation index states of the key phenological periods were quantitatively expressed by using the knowledge probability coding method.【Result】The cropping pattern spectrum on the Jianghan Plain was constructed, and it was found that the cropping pattern spectrum on the Jianghan Plain was composed of eight cropping patterns: Spring single-cropping, Summer single-cropping, Spring-and-Summer double-cropping, Summer-and-Autumn double-cropping, double-cropping paddy-rice, cash crops, aquaculture ponds, trees or abandoned croplands. The results showed that the proposed cropping pattern spectrum and the method of mapping cropping patterns based on the key phenological periods and the Naive Bayesian Networks could extract all cropping patterns accurately, at the same time, which had good performance and suitability. There was a significant trend of the expansion of Summer-and-Autumn double-cropping and the shrink of Spring-and-Summer double-cropping and the Summer single-cropping on the Jianghan Plain during the study period.【Conclusion】The cropping pattern spectrum gave a picture of the overall situation of intensive utilization of croplands on the Jianghan Plain, for improving the accuracy of monitoring the use of croplands and enriching the connotation of the use of cropland resources.

Key words: phenology information atlas, cropping pattern, cropping pattern spectrum, Bayesian Network, the Jianghan Plain