干旱区生态环境知识资源中心(Arid): Study on mapping method of irrigated cultivated land-taking Nebraska as an example

Arid

DOI	10.1016/j.ecolind.2024.112271
	Study on mapping method of irrigated cultivated land-taking Nebraska as an example
	Zhu, Xiufang; Liu, Ying; Xu, Kun
通讯作者	Liu, Y
来源期刊	ECOLOGICAL INDICATORS
ISSN	1470-160X
EISSN	1872-7034
出版年	2024
卷号	166
英文摘要	The distribution information of irrigated cultivated land is important to many studies, including grain yield estimation, water resources management, drought risk assessment, and so on. This study developed a feature variable, called the irrigation probability index (IPI), based on the principle that irrigation can slow down or inhibit the evolution from meteorological drought to agricultural drought. Nebraska, which has a good irrigation information database, was selected as the study area. Combining IPI with the irrigation statistical data and other selected feature variables, irrigated cultivated land in Nebraska was mapped in 2017 using both the spatialization method and remote sensing classification methods (random forest and support vector machine). We verified the effectiveness of IPI in identifying irrigated cultivated land, analyzed the importance of different feature variables for irrigated cultivated land identification, and compared the accuracy of different irrigation mapping methods. The results show that the IPI is significantly correlated with the actual irrigation area, the area of irrigation facilities, and the number of active irrigation wells. It has a better indicative ability of irrigation in areas with a suitable climate or arid areas, than in areas with a humid climate. The crop water stress index (CWSI), IPI, vegetation index (EVI and NDVI), and land surface temperature difference between day and night are the most important feature variables that are used to distinguish irrigated cultivated land from rain-fed cultivated land. There are differences between the recognition accuracies of the three mapping methods in different climatic regions. Overall, the spatialization method had the highest accuracy (overall accuracy = 86.92 %, kappa = 0.74) and the highest similarity with the existing high-resolution irrigation maps. This study provides a direct and useful reference for other researchers to select feature variables and mapping methods in irrigated cultivated land mapping research.
英文关键词	Feature selection Cultivated land Random forest Support vector machine Irrigation mapping
类型	Article
语种	英语
收录类别	SCI-E
WOS记录号	WOS:001259564700001
WOS关键词	MAP ; TEMPERATURE ; AREAS ; MODIS
WOS类目	Biodiversity Conservation ; Environmental Sciences
WOS研究方向	Biodiversity & Conservation ; Environmental Sciences & Ecology
资源类型	期刊论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/403415
推荐引用方式 GB/T 7714	Zhu, Xiufang,Liu, Ying,Xu, Kun. Study on mapping method of irrigated cultivated land-taking Nebraska as an example[J],2024,166.
APA	Zhu, Xiufang,Liu, Ying,&Xu, Kun.(2024).Study on mapping method of irrigated cultivated land-taking Nebraska as an example.ECOLOGICAL INDICATORS,166.
MLA	Zhu, Xiufang,et al."Study on mapping method of irrigated cultivated land-taking Nebraska as an example".ECOLOGICAL INDICATORS 166(2024).