干旱区生态环境知识资源中心(Arid): Towards a global soil taxonomy and classification tool for predicting multi-level soil hierarchy

Arid

DOI	10.1007/s40808-021-01191-8
	Towards a global soil taxonomy and classification tool for predicting multi-level soil hierarchy
	Mallah, Sina; Bagheri-Bodaghabadi, Mohsen
通讯作者	Mallah, S (corresponding author), Agr Res Educ & Extens Org AREEO, Soil & Water Res Inst, Dept Soil Phys & Irrigat, Karaj, Iran.
来源期刊	MODELING EARTH SYSTEMS AND ENVIRONMENT
ISSN	2363-6203
EISSN	2363-6211
出版年	2021
英文摘要	It is essential to update previous classifications of the soil pedons in reports and databases as keys to soil taxonomy and classification editions getting updated. The conventional soil inventory procedure is manually time- and cost-consuming. Meanwhile, it usually depends on users' expert opinion and is mostly prone to human error as well as personalization. Therefore, a user friendly tool that simulate the soil classification automatically can be a great solution for this purpose. Hence, this paper represents a mobile-based tool to classify the soil pedons according to USDA soil taxonomy classification at multiple hierarchical levels. The framework is a computational basis in two dimensions, including horizontal and vertical ranking in order of priority and soil taxonomy criteria, respectively. The outputs were validated using pedons of Iran National Soil Dataset (INSD) and National Cooperative Soil Survey (NCSS) which were selected randomly at order, sub-order and great group levels. The NCSS results showed that the model could achieve 63.88% overall accuracy with a certain (100%) prediction performance in orders of gelisols, histosols, spodosols, oxisols, vertisols, aridisols and entisols and moderate accuracy (66.6%) for inceptisols. The overall accuracy of the classification prediction in the hierarchical level of order and suborder obtained 100% and 58% for NCSS, while 87.5% and 92.85% for INSD, respectively. These findings support the helpfulness of smart-phone-based technology for soil identification exclusively in soils of arid and semi-arid regions. Concluding that, there is still more attempts required to develop and achieve reliable predictions for automatic soil classification.
英文关键词	Soil survey Numerical taxonomy Software Automatic soil classification
类型	Review ; Early Access
语种	英语
收录类别	ESCI
WOS记录号	WOS:000657549200001
WOS类目	Environmental Sciences
WOS研究方向	Environmental Sciences & Ecology
资源类型	期刊论文
条目标识符	http://119.78.100.177/qdio/handle/2XILL650/353031
作者单位	[Mallah, Sina] Agr Res Educ & Extens Org AREEO, Soil & Water Res Inst, Dept Soil Phys & Irrigat, Karaj, Iran; [Bagheri-Bodaghabadi, Mohsen] Agr Res Educ & Extens Org AREEO, Soil & Water Res Inst, Dept Soil Genesis Classificat & Cartog, Karaj, Iran
推荐引用方式 GB/T 7714	Mallah, Sina,Bagheri-Bodaghabadi, Mohsen. Towards a global soil taxonomy and classification tool for predicting multi-level soil hierarchy[J],2021.
APA	Mallah, Sina,&Bagheri-Bodaghabadi, Mohsen.(2021).Towards a global soil taxonomy and classification tool for predicting multi-level soil hierarchy.MODELING EARTH SYSTEMS AND ENVIRONMENT.
MLA	Mallah, Sina,et al."Towards a global soil taxonomy and classification tool for predicting multi-level soil hierarchy".MODELING EARTH SYSTEMS AND ENVIRONMENT (2021).