Knowledge Resource Center for Ecological Environment in Arid Area
DOI | 10.1016/j.rse.2018.09.008 |
Tracking annual cropland changes from 1984 to 2016 using time-series Landsat images with a change-detection and post-classification approach: Experiments from three sites in Africa | |
Xu, Yidi1; Yu, Le1,2; Zhao, Feng R.3; Cai, Xueliang4; Zhao, Jiyao1; Lu, Hui1,2; Gong, Peng1,2 | |
通讯作者 | Yu, Le |
来源期刊 | REMOTE SENSING OF ENVIRONMENT
![]() |
ISSN | 0034-4257 |
EISSN | 1879-0704 |
出版年 | 2018 |
卷号 | 218页码:13-31 |
英文摘要 | Ensuring food security has been the top priority of many regions, particularly in developing countries in Africa. In recent decades, increasing population, together with growing food demands, have put great pressure on the world’s food production. Long-term, up-to-date, annual cropland mapping at high resolution (i.e., at tens-of-metre levels) is in urgent demand for tracking spatial and temporal patterns of cropland change. However, because of the difficulty of capturing seasonality and flexible cropping systems, few studies have focused on understanding the dynamics of cropland using Landsat data in Africa. Here, we propose a new method of updating annual cropland mapping using a change-detection approach and post-classification to improve on traditional bi-temporal change vector analysis. Three Landsat footprints in Africa were selected (Egypt, Ethiopia and South Africa) as our study areas based on their different cropping systems and field sizes. The potential annual change areas were detected by employing multiple indices and thresholds in reference and long-term annual composite Landsat images. Next, map updates were conducted in the potential change pixels using random forest-based classification. Different training sample metrics were used (seasonal and annual samples) and compared in the classification step. The long-term cropland mapping accuracies for these three sites ranged from 88.04% to 94.30% (Egypt), 76.28% to 82.88% (Ethiopia) and 56.52% to 67.53% (South Africa). The results showed improvements in the accuracy and consistency of updating the annual cropland information using change-detection approaches, accounting for accuracy increases of 2.40%, 10.62% and 0.55% compared with a yearly cropland mapping approach in our previous research. The best results using annual samples extracted from the same season with the classified images supported the use of annual and growing samples in long-term annual mapping. Overall, a common trend of cropland expansion in all three sites was revealed, with an increase rate of 10.06, 3.73 and 1.35 kha/year in Egypt, Ethiopia and South Africa, respectively. The results indicated a rapid increasing pattern from bare land (desert) to irrigated systems (Egyptian site) but smaller and stable cropland changes in smallholder and farming-pastoral ecotones (Ethiopian and South African site), where limited land was still available for an expansion of agricultural area. This study highlights the potential application of time-series Landsat data in documenting and contributing missing cropland distribution information required for assessing and solving food security in Africa. |
英文关键词 | Cropland Change detection Africa Time-series Landsat |
类型 | Article |
语种 | 英语 |
国家 | Peoples R China ; USA ; Netherlands |
收录类别 | SCI-E |
WOS记录号 | WOS:000449449800002 |
WOS关键词 | CLIMATE-CHANGE IMPACTS ; FOREST COVER LOSS ; FROM-GLC ; MODIS ; DATABASE ; DYNAMICS ; CHINA ; AREA ; AGRICULTURE ; DATASET |
WOS类目 | Environmental Sciences ; Remote Sensing ; Imaging Science & Photographic Technology |
WOS研究方向 | Environmental Sciences & Ecology ; Remote Sensing ; Imaging Science & Photographic Technology |
来源机构 | 清华大学 |
资源类型 | 期刊论文 |
条目标识符 | http://119.78.100.177/qdio/handle/2XILL650/212702 |
作者单位 | 1.Tsinghua Univ, Dept Earth Syst Sci, Minist Educ, Key Lab Earth Syst Modeling, Beijing 100084, Peoples R China; 2.Joint Ctr Global Change Studies, Beijing 100875, Peoples R China; 3.Univ Maryland, Dept Geog Sci, College Pk, MD 20742 USA; 4.UNESCO IHE, Westvest 7, NL-2611 AX Delft, Netherlands |
推荐引用方式 GB/T 7714 | Xu, Yidi,Yu, Le,Zhao, Feng R.,et al. Tracking annual cropland changes from 1984 to 2016 using time-series Landsat images with a change-detection and post-classification approach: Experiments from three sites in Africa[J]. 清华大学,2018,218:13-31. |
APA | Xu, Yidi.,Yu, Le.,Zhao, Feng R..,Cai, Xueliang.,Zhao, Jiyao.,...&Gong, Peng.(2018).Tracking annual cropland changes from 1984 to 2016 using time-series Landsat images with a change-detection and post-classification approach: Experiments from three sites in Africa.REMOTE SENSING OF ENVIRONMENT,218,13-31. |
MLA | Xu, Yidi,et al."Tracking annual cropland changes from 1984 to 2016 using time-series Landsat images with a change-detection and post-classification approach: Experiments from three sites in Africa".REMOTE SENSING OF ENVIRONMENT 218(2018):13-31. |
条目包含的文件 | 条目无相关文件。 |
除非特别说明,本系统中所有内容都受版权保护,并保留所有权利。