[1]张宏帅,朱高龙,吴家煜,等.基于BP神经网络与Kriging结合的土壤有机质空间分布模拟——以福建省华安县为例[J].亚热带农业研究,2021,17(01):40-47.[doi:10.13321/j.cnki.subtrop.agric.res.2021.01.008]
 ZHANG Hongshuai,ZHU Gaolong,WU Jiayu,et al.Simulation of spatial distribution simulation of soil organic matter based on BP neural network and Kriging interpolation—Taking Hua’an County, Fujian Province as an example[J].,2021,17(01):40-47.[doi:10.13321/j.cnki.subtrop.agric.res.2021.01.008]
点击复制

基于BP神经网络与Kriging结合的土壤有机质空间分布模拟——以福建省华安县为例()
分享到:

《亚热带农业研究》[ISSN:1006-6977/CN:61-1281/TN]

卷:
17
期数:
2021年01期
页码:
40-47
栏目:
出版日期:
2021-05-15

文章信息/Info

Title:
Simulation of spatial distribution simulation of soil organic matter based on BP neural network and Kriging interpolation—Taking Hua’an County, Fujian Province as an example
作者:
张宏帅1 朱高龙2 吴家煜1 吴锡麟2
1. 福州大学空间数据挖掘与信息共享教育部重点实验室, 福建 福州 350108;
2. 闽江学院海洋学院, 福建 福州 350108
Author(s):
ZHANG Hongshuai1 ZHU Gaolong2 WU Jiayu1 WU Xilin2
1. Key Lab of Spatial Data Mining and Information Sharing of China Ministry of Education, Fuzhou University, Fuzhou, Fujian 350108, China;
2. College of Oceanography, Minjiang University, Fuzhou, Fujian 350108, China
关键词:
土壤有机质神经网络虚拟变量空间插值华安县
Keywords:
soil organic matterneural networkvirtual variablespatial interpolationHua’an county
分类号:
S159
DOI:
10.13321/j.cnki.subtrop.agric.res.2021.01.008
摘要:
[目的] 提高县域尺度耕地土壤有机质空间插值精度。[方法] 基于福建省漳州市华安县215个土壤有机质野外采样数据,将样地土壤类型、土地利用方式两种定性因素转化为虚拟变量,结合土壤质地、海拔高度、坡度等定量因素,构建了BP神经网络与克里金插值(Kriging)相结合的非线性拟合法(BP_OK),并与回归克里金插值法(RK)、普通克里金插值法(OK)进行对比。[结果] 利用30个验证样点计算BP_OK、RK、OK法的均方根误差分别为3.55、3.73、4.92 g·kg-1,相关系数分别为0.72、0.68、0.35。[结论] 结合土壤有机质采样数据和外界辅助因素的BP_OK法、RK法插值精度明显优于仅考虑土壤有机质采样数据空间自相关性的OK法,其中采用非线性拟合的BP_OK法预测精度最高,证明了BP_OK法可以有效改善县域尺度下耕地土壤有机质空间分布模拟精度。
Abstract:
[Purpose] To improve the accuracy of spatial interpolation of soil organic matter in cultivated land at county scale. [Method] Based on the field sampling data of 215 soil organic matter in Hua’an County, Zhangzhou City, Fujian Province, the two qualitative factors of soil type and land use mode of the sample plots were converted into virtual variables, and combined with quantitative factors such as soil texture, altitude and slope to construct a BP neural network. A non-linear fitting method (BP_OK) was constructed to combine the network with Kriging interpolation, and compare to regression Kriging interpolation (RK) and ordinary Kriging interpolation (OK). [Result] The root mean square errors of the BP_OK, RK, and OK methods calculated using 30 verification samples were 3.55, 3.73, 4.92 g·kg-1, respectively, and the correlation coefficients were 0.72, 0.68, and 0.35, respectively. [Conclusion] The interpolation accuracy of the BP_OK and RK methods, which combines with soil organic matter sampling data and external auxiliary factors, is significantly better than the OK method, which only considers the spatial autocorrelation of soil organic matter sampling data. Among these methods, the BP_OK method with nonlinear fitting has the highest prediction accuracy. When it combines with the network and Kriging interpolation, the BP_OK method can effectively improve the simulation accuracy of the spatial distribution of soil organic matter in cultivated land at the county scale.

参考文献/References:

[1] 李启权,王昌全,岳天祥,等.基于RBF神经网络的土壤有机质空间变异研究方法[J].农业工程学报,2010,26(1):87-93.
[2] 杨奇勇,杨劲松,姚荣江,等.耕地土壤有机质空间变异性的随机模拟[J].农业工程学报,2010,26(12):324-329.
[3] 李启权.基于RBF神经网络的土壤属性信息空间插值方法研究[D].雅安:四川农业大学,2006.
[4] MOHAMED E S, ABU-HASHIM M, ABDELRAHMAN M A, et al. Evaluating the effects of human activity over the last decades on the soil organic carbon pool using satellite imagery and GIS techniques in the Nile Delta area, Egypt[J]. Sustainability, 2019,11(9):2644.
[5] 鲍丽然,周皎,李瑜,等.渝西北土壤有机质空间变异及影响因素分析[J].西南农业学报,2017,30(11):2541-2547.
[6] 刘永红,倪中应,谢国雄,等.浙西北丘陵区农田土壤微量元素空间变异特征及影响因子[J].植物营养与肥料学报,2016,22(6):1710-1718.
[7] FUSARO C, SARRIA-GUZMáN Y, CHáVEZ-ROMERO Y A, et al. Land use is the main driver of soil organic carbon spatial distribution in a high mountain ecosystem[J]. Peer J, 2019,7(5):e7897.
[8] 江叶枫,郭熙,叶英聪,等.基于辅助变量和神经网络模型的土壤有机质空间分布模拟[J].长江流域资源与环境,2017,26(8):1150-1158.
[9] 聂哲,李秀芬,吕家欣,等.东北典型黑土区表层土壤有机质含量高光谱反演研究[J].土壤通报,2019,50(6):1285-1293.
[10] 陈琳,任春颖,王宗明,等.基于克里金插值的耕地表层土壤有机质空间预测[J].干旱区研究,2017,34(4):798-805.
[11] 姚聪.基于卷积神经网络的耕层土壤有机质含量估测模型研究[D].泰安:山东农业大学,2020.
[12] 江叶枫,郭熙,叶英聪,等.应用集成BP神经网络模型预测土壤有机质空间分布[J].江苏农业学报,2017,33(5):1044-1050.
[13] 刘二永,刘健,余坤勇,等.基于环境因子和R-STPS的林地土壤有机质预测模型[J].农业机械学报,2015,46(1):133-137.
[14] GEORGE J, KUMAR S, ARYA RAJ R. Soil organic carbon prediction using visible-near infrared reflectance spectroscopy employing artificial neural network modelling[J]. Current Science, 2020,119(2):377-381.
[15] 任丽,杨联安,王辉,等.基于随机森林的苹果区土壤有机质空间预测[J].干旱区资源与环境,2018,32(8):141-146.
[16] 吴鹏飞.川东平行岭谷区土壤有机质时空变异特征研究:以达县为例[D].重庆:西南大学,2011.
[17] 李可,杨勇,刘亚军,等.基于回归克里格的河北省土壤有机质时空变化特征研究[J].中国土壤与肥料,2020(3):1-7.
[18] 田海涛.带有序多分类解释变量的线性回归模型估计方法及其应用研究[D].成都:西南交通大学,2019.
[19] 戴金辉.虚拟变量回归及其应用[J].统计与决策,2019,35(5):77-80.
[20] 王政权.地统计学及在生态学中的应用[M].北京:科学出版社,1999:74-96.
[21] 宋娜.多元Logistic分布及其参数估计[D].北京:北京工业大学,2007.
[22] 叶勤,姜雪芹,李西灿,等.基于高光谱数据的土壤有机质含量反演模型比较[J].农业机械学报,2017,48(3):164-172.
[23] AGHAJANI M, JALALIAN A, BESALATPOUR A A. Soil particulate organic matter (POM) prediction in a mountainous watershed using artificial neural networks[J]. Communications in Soil Science and Plant Analysis, 2015,46(8):925-938.
[24] 张森,卢霞,聂格格,等.SVM和BP检测滨海湿地土壤有机质[J].光谱学与光谱分析,2020,40(2):556-561.
[25] CAMBARDELLA C A, MOORMAN T B, NOVAK J M, et al. Field-scale variability of soil properties in central Iowa soils[J]. Soil Science Society of America Journal, 1994,58(5):1501-1511.
[26] 陈桂香,高灯州,曾从盛,等.福州市农田土壤养分空间变异特征[J].地球信息科学学报,2017,19(2):216-224.
[27] 余笑眉,吕晓男,王美琴,等.南方丘陵耕地土壤有机质空间插值精度比较:以衢江区为例[J].浙江农业学报,2010,22(5):639-643.
[28] 姜赛平,张怀志,张认连,等.基于三种空间预测模型的海南岛土壤有机质空间分布研究[J].土壤学报,2018,55(4):1007-1017.
[29] PICCINI C, MARCHETTI A, FRANCAVIGLIA R. Estimation of soil organic matter by geostatistical methods:use of auxiliary information in agricultural and environmental assessment[J]. Ecological Indicators, 2014,36:301-314.
[30] 杨顺华,张海涛,郭龙,等.基于回归和地理加权回归Kriging的土壤有机质空间插值[J].应用生态学报,2015,26(6):1649-1656.
[31] 尹群,郭纪敏,张世文,等.基于辅助变量的县域土壤有机质预测[J].江苏农业科学,2020,48(24):267-273.

备注/Memo

备注/Memo:
收稿日期:2021-02-02。
基金项目:国家自然科学基金项目(41871233);福建省科技厅农业引导性(重点)项目(2018N0024)。
作者简介:张宏帅(1996-),男,硕士研究生。研究方向:耕地质量评价。Email:2476402452@qq.com。
通讯作者:朱高龙(1974-),男,教授。研究方向:植被遥感与耕地质量评价。Email:zhugaolong@163.com。
更新日期/Last Update: 1900-01-01