Pedotransfer functions for estimating soil water retention properties of northern China agricultural soils: Development and needs*
文献类型： 外文期刊
作者： Xu, Xu 1 ; Li, Huawei 2 ; Sun, Chen 3 ; Ramos, Tiago B. 4 ; Darouich, Hanaa 5 ; Xiong, Yunwu 1 ; Qu, Zhongyi 6 ; Huang, G 1 ;
作者机构： 1.China Agr Univ, Chinese Israeli Int Ctr Res & Training Agr, Beijing 100083, Peoples R China
2.Henan Tech Coll Construct, Dept Engn Management, Zhengzhou, Peoples R China
3.Chinese Acad Agr Sci, Inst Environm & Sustainable Dev Agr, Beijing, Peoples R China
4.Univ Lisbon, Inst Super Tecn, MARETEC, Lisbon, Portugal
5.Univ Lisbon, Inst Super Agron, Ctr Invest Agron Alimentos Ambiente & Paisagem LE, Lisbon, Portugal
6.Inner Mongolia Agr Univ, Coll Water Conservancy & Civil Engn, Hohhot, Peoples R China
关键词： neural network; non-linear fitting; pedotransfer functions; soil hydraulic parameters; soil water model; textural data conversion
期刊名称：IRRIGATION AND DRAINAGE （ 影响因子：1.328； 五年影响因子：1.435 ）
ISSN： 1531-0353
年卷期：
页码：
收录情况： SCI
摘要： Many agro-environmental studies focusing on the efficient management of soils and water resources make use of soil water simulation models. Reliable soil hydraulic properties are critical for ensuring the accuracy of model simulations. However, soil hydraulic parameters in northern China are generally derived using external pedotransfer functions (PTFs) that do not take into account the specificities of local edaphoclimatic conditions due to the lack of a better alternative. Therefore, the main objective of this paper was to develop PTFs for estimating the soil water retention curve (SWRC) in northern China agricultural soils (named PTF-ANC). A total of 440 soil horizons were collected from the existing literature. A flexible soil-textural conversion program was first developed to harmonize soil particle-size data into the United States Department of Agriculture (USDA) classification system. The SWRC parameters of the van Genuchten model were also generated by curve fitting. Then, the PTF-ANC were developed using artificial neural networks, with soil texture and bulk density being used as input data and with a basic three-layer back-propagation neural network. The PTF-ANC showed an acceptable accuracy when predicting the SWRC, indicating a strong application potential for northern China soils. Comparison of estimates with two widely used external PTFs also showed that these were not suitable for characterizing the SWRC of northern China agricultural soils. This is due to the fact that the main soil textures (silt and silty loam textures) found in northern China soils were misrepresented in those external soil databases. Overall, this paper presented the absolute necessity of developing specific PTFs for northern China agricultural soils.
- 相关文献
