共 50 条
Quantification of the effect of soil erosion factors on soil nutrients at a small watershed in the Loess Plateau, Northwest China
被引:14
|作者:
Wang, Ziguan
[1
,2
]
Wang, Guangcai
[3
,4
]
Zhang, Yongguang
[5
]
Wang, Ruixue
[2
]
机构:
[1] Beijing Normal Univ, Fac Geog Sci, Beijing 100875, Peoples R China
[2] Kunming Univ Sci & Technol, Fac Land Resource Engn, Kunming 650093, Yunnan, Peoples R China
[3] China Univ Geosci, State Key Lab Biogeol & Environm Geol, Beijing 100083, Peoples R China
[4] China Univ Geosci, MOE Key Lab Groundwater Circulat & Environm Evolu, Beijing 100083, Peoples R China
[5] Capital Normal Univ, Informat Engn Coll, Beijing 100048, Peoples R China
基金:
中国国家自然科学基金;
关键词:
Gradient boosting decision tree;
Quantitative assessment;
Soil erosion factors;
Soil nutrients;
ECO-GEOLOGICAL ENVIRONMENT;
SPATIAL-DISTRIBUTION;
ENRICHMENT RATIO;
RAINDROP IMPACT;
ORGANIC-CARBON;
TOTAL NITROGEN;
RAINFALL;
SLOPE;
REGION;
ERODIBILITY;
D O I:
10.1007/s11368-019-02458-5
中图分类号:
X [环境科学、安全科学];
学科分类号:
08 ;
0830 ;
摘要:
Purpose This research aimed to quantitatively assess the effect of soil erosion factors on soil nutrients under similar rainfall conditions at a small watershed in the Loess Plateau, China, with emphasis on quantification of each factor's contribution to soil nutrients. Materials and methods Forty-five soil samples were collected and the soil texture and nutrients were analyzed. The integrated fertility index (IFI) was quantified with laboratory-measured available phosphorus, available potassium, total nitrogen, and soil organic matter based on principal component analysis and the soil erosion factors were computed. The feature importance of soil erosion factors that affect soil nutrients was estimated based on the gradient-boosting decision tree (GBDT) method. Results and discussion The GBDT optimized model revealed that the feature importance of soil erosion factors soil erodibility (K), slope length (L), slope steepness (S), vegetation cover (C), and mechanical erosion control (P) was 0.41, 0.11, 0.03, 0.39, and 0.06, respectively. The K and C factors occupied 80% of the importance and were the two predominant soil erosion factors affecting the soil comprehensive nutrients. The Pearson correlation coefficient between K, L, S, C, or P, and IFI was - 0.387, - 0.23, - 0.25, - 0.479, and - 0.532, respectively. The different order of factors between the Pearson correlation and feature importance may indicate that linear relation cannot be used to indicate the significance of each factor. Conclusions The soil erodibility and the vegetation cover could basically determine the quality of soil nutrients in the small watershed under similar rainfall conditions. The soil property itself and associated effects on the amount and rate of runoff under a certain rainfall erosivity had a dominant impact on the soil nutrients. Factor C greatly contributed to runoff resistance and protection of nutrients from erosion. The L, S, and P factors had weak effects on the IFI because of the generally low rainfall intensity, short rainfall duration, and other factors that may counteract the effects in the area.
引用
收藏
页码:745 / 755
页数:11
相关论文