摘要
传统土壤调查和制图建立在调查者经验思维基础上,目前对高精度土壤信息的大量需求对传统技巧性土壤调查方式提出了挑战,量化的土壤———景观模型日益受到土壤科学家的重视。近十年内世界各国开展了大量研究,试图将数字地形分析,GIS技术和土壤调查技术相结合,通过对景观信息的分析预测土壤信息。本文简要阐述土壤景观模型的基本原理和定义,对线性回归、回归树、判别分析、模糊聚类、地统计学、模糊推理和规则网络等模型的研究进展进行总结。在此基础上讨论了模型的发展方向和应用前景。
Traditional soil survey and mapping are based on the conceptual models that are constructed from the experience of the soil surveyors. The ever-increasing demands of high-resolution soil information have challenged this empirical method. Quantified soil-landscape models have received the recognition of more and more soil scientists. In the last decade, many works have been done a of over the world to integrate digital terrain analysis, GIS and soil survey techniques for soil information prediction. This paper reviewed the theory and definition of soil-landscape models, and briefly described the development of the modeling technique such as multiple linear regression, regression tree, discriminant analysis, geostatistic, fuzzy cluster, fuzzy logic reference and rule network. Finally, the research direction and possible applications of the models were discussed.
出处
《土壤通报》
CAS
CSCD
北大核心
2004年第3期339-346,共8页
Chinese Journal of Soil Science
基金
国家自然科学基金重点项目(49831004)资助
