| 海河低平原典型县种植制度与农田景观格局变化遥感监测 |
| |
| 引用本文： | 王小慧,姜雨林,傅漫琪,尹小刚,陈阜.海河低平原典型县种植制度与农田景观格局变化遥感监测[J].农业工程学报,2022,38(1):297-304. |
| |
| 作者姓名： | 王小慧 姜雨林 傅漫琪 尹小刚 陈阜 |
| |
| 作者单位： | 中国农业大学农学院/农业农村部农作制度重点实验室，北京 100193 |
| |
| 基金项目： | 国家重点研发计划课题（课题编号：2016YFD0300201） |
| |
| 摘 要： | 作物是影响农田景观格局和生态功能的最主要因素，种植制度与农田景观优化对集约化农业的可持续发展影响显著。该研究主要利用2013－2019年Landsat 8 OLI影像，基于主成分分析法-随机森林分类法（Principal Component Analysis - Random Forest, PCA-RF）解译遥感影像，结合转移矩阵和景观指数法，以河北吴桥为典型案例研究了海河低平原复种指数、种植模式的空间分布特征及其变化动态，在此基础上分析了农田景观破碎度和多样性的变化特征。结果发现，PCA-RF法解译研究区种植模式总体精度达90%以上，Kappa系数高于0.84，效果较好；研究区复种指数从163%上升至174%，一熟转变为两熟区的面积是两熟转变为一熟区面积的1.64倍；麦玉两熟面积保持稳定，棉花一熟面积缩减了80.93%，而粮林复合种植模式的面积增长了64.54%；棉花一熟是转变为麦玉两熟的主体模式，占麦玉两熟转入面积的81.15%；同时，麦玉两熟和棉花一熟是转变为玉米一熟的主体模式，分别占玉米一熟转入面积的46.43%和41.43%；研究区香农多样性指数上升8.57%，麦玉两熟和棉花一熟的分离度指数显著增大。结果表明，7年来，棉花一熟转变为麦玉两熟导致研究区复种指数提升；棉花一熟缩减且转变为其他模式和种植模式增多导致农田景观多样性增加；同时，麦玉两熟和棉花一熟的破碎度增大导致农田景观破碎度增大，不利于主体模式的进一步规模化生产。
|
| 关 键 词： | 遥感 作物 监测 海河低平原 农田景观 复种指数 种植模式 |
| 收稿时间： | 2021/9/23 0:00:00 |
| 修稿时间： | 2021/12/16 0:00:00 |
Cropping patterns and farmland landscape at the county level using remote sensing in Haihe Lowland Plain |
| |
| Wang Xiaohui,Jiang Yulin,Fu Manqi,Yin Xiaogang,Chen Fu.Cropping patterns and farmland landscape at the county level using remote sensing in Haihe Lowland Plain[J].Transactions of the Chinese Society of Agricultural Engineering,2022,38(1):297-304. |
| |
| Authors: | Wang Xiaohui Jiang Yulin Fu Manqi Yin Xiaogang Chen Fu |
| |
| Institution: | College of Agronomy and Biotechnology, China Agricultural University/Key Laboratory of Farming System, Ministry of Agriculture and Rural Affairs, Beijing 100193, China; 1. College of Computer Science and Technology, Henan Polytechnic University, Jiaozuo 454000, China; 2. School of Surveying and mapping and Land Information Engineering, Henan Polytechnic University, Jiaozuo 454000, China; |
| |
| Abstract: | Crop is one of the most important indicators to determine the farmland landscape pattern, crop productivity, and ecological benefits. It is necessary to clarify the changes in cropping patterns, thereby optimizing the farmland landscape during the sustainable intensification development of agriculture. Taking a typical county in Haihe Lowland Plain as the study area, this study aims to conduct the spatial distribution, variables of multiple cropping index, and cropping patterns at the county level from 2013 to 2019. A quadratic difference set was firstly used to collect the multiple cropping index, according to the smooth NDVI time-series curves by Harmonic Analysis of Time Series (HANTS). A Principal Component Analysis - Random Forest (PCA-RF) was then selected to classify the cropping patterns via the image interpretation using multi-temporal Landsat 8 OLI and Sentinel 2 images. General cropping patterns were interpreted including the cotton, pepper, peanut, maize, and tree crop single cropping, as well as the wheat-maize and tree crop double cropping. Finally, the transfer matrix was operated to quantify the transformations between every two cropping patterns, and then landscape metrics (e.g. Splitting Index (SPLIT), Shannon''s Diversity Index (SHDI) and Shannon''s Evenness Index (SHEI)) were calculated by Fragstats 4.2, further to describe the landscape fragmentation and diversity of the farmland. The overall accuracy and Kappa coefficient exceeded 90% and 0.84, respectively, after interpreting multiple cropping indexes using PCA-RF. Overall, the multiple cropping index increased from 163% to 174%, and the area of single cropping transformed to double cropping was 1.64 times that of the double cropping transformed to single cropping. The area of wheat-maize pattern remained stable, and the area of cotton single cropping decreased by 80.93%, with the area reduction of tree crop planting pattern increased by 64.54%. Cotton single cropping was the major pattern that transformed into wheat-maize double cropping, accounting for 81.15% of the transformed area. At the same time, the wheat-maize double and cotton single cropping were the main cropping patterns that transformed into the maize single cropping, accounting for 46.43% and 41.43% of the transformed area, respectively. The SHDI of the cropping pattern increased by 8.57%, and the SPLIT of wheat-maize double cropping and cotton single cropping increased significantly. In conclusion, the major variation in the farmland landscape was that: 1) The increase of the multiple indexes was mainly caused by transformation from cotton single cropping to wheat-maize double cropping. 2) The diversity of farmland landscape increased in two ways. One is that the cotton single cropping was transformed to other patterns (e.g. wheat-maize double cropping, tree crop double cropping, and cotton single cropping). The other is that more cropping patterns (such as peanut single cropping and maize single cropping) appeared, with the increased area of those cropping patterns in smaller areas. 3) Larger fragmentation of wheat-maize double cropping and cotton single cropping resulted in the more fragmented farmland landscape, in turn further preventing the large-scale crop production of the main cropping patterns. |
| |
| Keywords: | remote sensing crops monitoring Haihe lowland plain farmland landscape multiple cropping index cropping pattern |
|
| 点击此处可从《农业工程学报》浏览原始摘要信息 |
| 点击此处可从《农业工程学报》下载免费的PDF全文 |
|
