集成学习支持的复杂地貌类型区土壤全钾含量自适应曲面建模
|
摘要
土壤属性空间分布受复杂地学环境要素影响,空间分异特征十分明显,用单一全局插值模型进行土壤属性模拟,难以实现高精度模拟。对于空间不连续、全局插值模型精度有限及适应性差的特点,本文提出了一种集成学习支持的、融合地学环境变量的土壤属性自适应曲面建模方法(ASM-SP)。利用2013年采集的110个样点数据,使用回归克里金(RK)、贝叶斯克里金(BK)、普通克里金插值法(OK)、反距离加权法(IDW)、ASM-SP,分别对青海湖复杂地貌类型区进行土壤全钾含量的插补。本文采用逐点交叉验证(LOOCV)插值方法模拟精度。结果表明,ASM-SP不仅考虑了地学环境变量与土壤属性的非线性关系,而且融合了多个模型的适应性优势,是实现复杂地貌类型区土壤全钾含量的高精度模拟的一种新方法。
The spatial distribution of soil attributes is influenced by the elements of complex geoscience environment,and the spatial differentiation characteristics are very obvious.It is difficult to simulate soil attributes with a single global interpolation model.For the characteristics of spatial discontinuity,limited precision of global interpolation model and poor adaptability,a soil attribute adaptive surface modeling method(ASM-SP) based on integrated learning support and integrated geoscience environment variables was proposed.Based on 110 samples collected in 2013,regression Kriging(RK),Bayesian Kriging(BK),ordinary Kriging(OK) and inverse distance weighting method(IDW),ASM-SP were used to interpolate the soil total potassium content in the complex geomorphological area of Qinghai Lake,respectively.In this paper,the simulation accuracy of different interpolation methods was evaluated by Leave One Out Validation.ASM-SP not only considers the nonlinear relationship between geo-environmental variables and soil attributes,but also combines the adaptive advantages of multiple models.It is a new method to simulate soil total potassium content in complex geomorphological regions with high precision.
The spatial distribution of soil attributes is influenced by the elements of complex geoscience environment,and the spatial differentiation characteristics are very obvious.It is difficult to simulate soil attributes with a single global interpolation model.For the characteristics of spatial discontinuity,limited precision of global interpolation model and poor adaptability,a soil attribute adaptive surface modeling method(ASM-SP) based on integrated learning support and integrated geoscience environment variables was proposed.Based on 110 samples collected in 2013,regression Kriging(RK),Bayesian Kriging(BK),ordinary Kriging(OK) and inverse distance weighting method(IDW),ASM-SP were used to interpolate the soil total potassium content in the complex geomorphological area of Qinghai Lake,respectively.In this paper,the simulation accuracy of different interpolation methods was evaluated by Leave One Out Validation.ASM-SP not only considers the nonlinear relationship between geo-environmental variables and soil attributes,but also combines the adaptive advantages of multiple models.It is a new method to simulate soil total potassium content in complex geomorphological regions with high precision.
引文
[1]陈鸽,汤春纯,李祖胜,等.不同施肥措施对洞庭湖平原区旱地肥力及作物产量的影响[J].中国生态农业学报,2017,25(5):689-697.
[2]董洪芳,于君宝,孙志高,等.黄河口滨岸潮滩湿地植物-土壤系统有机碳空间分布特征[J].环境科学,2010,31(6):1594-1599.
[3]OBALUM S E,OPPONG J,IGWE C A,et al.Spatial variability of uncultivated soils in derived savanna[J].International Agrophysics,2013,27(1):57-67.
[4]ROSEMARY F,VITHARANA U W A,INDRARATNES P,et al.Exploring the spatial variability of soil properties in an Alfisol soil catena[J].Catena,2017,150:53-61.
[5]赵明松,张甘霖,王德彩,等.徐淮黄泛平原土壤有机质空间变异特征及主控因素分析[J].土壤学报,2013,50(1):1-11.
[6]龙军,张黎明,沈金泉,等.复杂地貌类型区耕地土壤有机质空间插值方法研究[J].土壤学报,2014,51(6):1270-1281.
[7]史文娇,岳天祥,石晓丽,等.土壤连续属性空间插值方法及其精度的研究进展[J].自然资源学报,2012,27(1):163-175.
[8]马成霞,丁建丽,王璐,等.绿洲土壤表层含盐量空间变异分析的插值方法研究[J].水土保持研究,2014,21(4):317-320.
[9]赵巧丽,郑国清,冯晓,等.河南省安阳县三种土壤全氮含量空间插值方法的比较分析[J].土壤通报,2012,43(5):1162-1166.
[10]谢云峰,陈同斌,雷梅,等.空间插值模型对土壤Cd污染评价结果的影响[J].环境科学学报,2010,30(4):847-854.
[11]KURIAKOSE S L,DEVKOTA S,ROSSITER D G,et al.Prediction of soil depth using environmental variables in an anthropogenic landscape:a case study in the Western Ghats of Kerala,India[J].Catena,2009,79(1):27-38.
[12]江贵荣.干旱区不同尺度土壤盐分空间变异特征及不确定性分析[D].北京:中国地质大学,2012.
[13]吴春生,黄翀,刘高焕,等.黄河三角洲土壤含盐量空间预测方法研究[J].资源科学,2016,38(4):704-713.
[14]杨琳,朱阿兴,张淑杰,等.土壤制图中多等级代表性采样与分层随机采样的对比研究,土壤学报,2015,52(1):28-37.
[15]王胜利,刘伟,张连蓬,等.地学环境变量支持的土壤全钾含量自适应曲面建模---以青海湖流域典型地区为例[J].水土保持研究,2018,25(1):132-138.
[16]黄文忠.宜宾市土壤钾素的空间变异特征及影响因素研究[D].雅安:四川农业大学,2010.
[17]徐爱萍,圣文顺,舒红.时空积和模型的数据插值与交叉验证[J].武汉大学学报(信息科学版),2012,37(7):766-769.
[18]李佳,段平,吕海洋,等.基于改进的逐点交叉验证的RBF形态参数优化方法及其空间插值实验[J].地理与地理信息科学,2016,32(3):39-42.
[19]顾春雷,杨漾,朱志春.几种建立DEM模型插值方法精度的交叉验证[J].测绘与空间地理信息,2011,34(5):99-102.
[2]董洪芳,于君宝,孙志高,等.黄河口滨岸潮滩湿地植物-土壤系统有机碳空间分布特征[J].环境科学,2010,31(6):1594-1599.
[3]OBALUM S E,OPPONG J,IGWE C A,et al.Spatial variability of uncultivated soils in derived savanna[J].International Agrophysics,2013,27(1):57-67.
[4]ROSEMARY F,VITHARANA U W A,INDRARATNES P,et al.Exploring the spatial variability of soil properties in an Alfisol soil catena[J].Catena,2017,150:53-61.
[5]赵明松,张甘霖,王德彩,等.徐淮黄泛平原土壤有机质空间变异特征及主控因素分析[J].土壤学报,2013,50(1):1-11.
[6]龙军,张黎明,沈金泉,等.复杂地貌类型区耕地土壤有机质空间插值方法研究[J].土壤学报,2014,51(6):1270-1281.
[7]史文娇,岳天祥,石晓丽,等.土壤连续属性空间插值方法及其精度的研究进展[J].自然资源学报,2012,27(1):163-175.
[8]马成霞,丁建丽,王璐,等.绿洲土壤表层含盐量空间变异分析的插值方法研究[J].水土保持研究,2014,21(4):317-320.
[9]赵巧丽,郑国清,冯晓,等.河南省安阳县三种土壤全氮含量空间插值方法的比较分析[J].土壤通报,2012,43(5):1162-1166.
[10]谢云峰,陈同斌,雷梅,等.空间插值模型对土壤Cd污染评价结果的影响[J].环境科学学报,2010,30(4):847-854.
[11]KURIAKOSE S L,DEVKOTA S,ROSSITER D G,et al.Prediction of soil depth using environmental variables in an anthropogenic landscape:a case study in the Western Ghats of Kerala,India[J].Catena,2009,79(1):27-38.
[12]江贵荣.干旱区不同尺度土壤盐分空间变异特征及不确定性分析[D].北京:中国地质大学,2012.
[13]吴春生,黄翀,刘高焕,等.黄河三角洲土壤含盐量空间预测方法研究[J].资源科学,2016,38(4):704-713.
[14]杨琳,朱阿兴,张淑杰,等.土壤制图中多等级代表性采样与分层随机采样的对比研究,土壤学报,2015,52(1):28-37.
[15]王胜利,刘伟,张连蓬,等.地学环境变量支持的土壤全钾含量自适应曲面建模---以青海湖流域典型地区为例[J].水土保持研究,2018,25(1):132-138.
[16]黄文忠.宜宾市土壤钾素的空间变异特征及影响因素研究[D].雅安:四川农业大学,2010.
[17]徐爱萍,圣文顺,舒红.时空积和模型的数据插值与交叉验证[J].武汉大学学报(信息科学版),2012,37(7):766-769.
[18]李佳,段平,吕海洋,等.基于改进的逐点交叉验证的RBF形态参数优化方法及其空间插值实验[J].地理与地理信息科学,2016,32(3):39-42.
[19]顾春雷,杨漾,朱志春.几种建立DEM模型插值方法精度的交叉验证[J].测绘与空间地理信息,2011,34(5):99-102.