基于BP神经网络的作物Cd含量预测及安全种植分区

doi:10.18402/resci.2018.12.09

摘要/Abstract

摘要：

为科学预测作物重金属含量,实现重金属污染农用地的安全利用,本研究利用地理探测器选取对作物Cd含量影响较大的土壤Cd含量、土壤pH值、与交通主干线的距离等10个因素为输入因子,农作物可食部分Cd含量作为输出因子,分别建立小麦、水稻、油菜籽、蔬菜可食部分Cd含量的BP神经网络预测模型,对作物种植污染情况进行预测分析,划分作物安全利用种植区。研究结果表明：① BP神经网络模型预测精度明显优于多元回归预测模型;② 对小麦、水稻和油菜籽的可食部分Cd含量预测结果进行评价,得到作物污染可能的空间分布及特征;③ 依据评价结果,对3种作物进行配置,划分得到4种作物适宜种植区,并提出管控策略。研究可为污染农用地的安全利用及作物种植调整提供思路和依据,兼具理论和现实意义。

关键词: 作物Cd含量, BP神经网络, 预测模型, 重金属污染, 农用地安全利用, 小麦, 水稻, 油菜籽, 蔬菜, 作物适宜种植区

Abstract:

In order to predict the heavy metal content of crops and realize the safe utilization of heavy metal contaminated agricultural land, 10 factors including soil Cd content, soil pH value, soil organic matter, cation exchange capacity, distance to mining land, and distance to traffic lines etc. which affect Cd content in crops were selected by GeoDetector model, being used as input factors. The Cd content in edible parts of crops was used as output factor. The BP neural network prediction models of Cd content in edible parts of wheat, rice, rapeseed, and vegetables were established respectively. The pollution of crop planting was predicted and analyzed. The safety zones of planting was categorized. The results illustrate that: (1) The BP neural network exhibit a good predictive effect, which has a nice applicability to the prediction of Cd content of edible parts of crops. The prediction accuracy of BP neural network models was better than that of the multiple regression prediction models according to the accuracy evaluation indexes. (2) Due to the low prediction accuracy of vegetables prediction model, the support for prediction and management partitions was limited. Only the prediction results of Cd content in edible parts of wheat, rice, and rapeseed were evaluated. The crop pollution prediction distribution map and distribution characteristics were obtained. (3) With the goal of safe use of agricultural land, the three crops in the study area were re-allocated based on the evaluation results, which was divided into four crop suitable planting safety zones. The findings of this study can provide ideas and basis for safe utilization of contaminated agricultural land and crop planting adjustment, which is of both theoretical and practical significance.

Key words: Cd content of crops, BP neural network, prediction models, heavy metal pollution, safety use of agricultural land, wheat, rice, rapeseed, vegetables, zoning of suitable planting for crops

侯艺璇, 赵华甫, 吴克宁, 李凯. 基于BP神经网络的作物Cd含量预测及安全种植分区[J]. 资源科学, 2018, 40(12): 2414-2424.

Yixuan HOU, Huafu ZHAO, Kening WU, Kai LI. Prediction of crop Cd content and zoning of safety planting based on BP neural network[J]. Resources Science, 2018, 40(12): 2414-2424.

图/表 14

图1

图2

表1

表2

图3

表3

表4

图4

表5

图5

表6

表7

图6

图7

参考文献 28

[9]	周开利, 康耀红. 神经网络模型及其MATLAB仿真程序设计[M]. 北京:清华大学出版社, 2005.
	[Zhou K L, Kang Y H.Neural Network Model and Its MATLAB Simulation Program Design [M]. Beijing: Tsinghua University Press, 2005.]
[10]	周政. BP神经网络的发展现状综述[J]. 山西电子技术, 2008, (2):90-92.
	[Zhou Z.Survey of current progress in BP neural network[J]. Shanxi Electronic Technology, 2008, (2):90-92.]
[11]	樊霆, 叶文玲, 陈海燕, 等.农田土壤重金属污染状况及修复技术研究[J].生态环境学报, 2013, (10):1727-1736.
	[Fan T, Ye W L, Chen H Y, et al. Review on contamination and remediation technology of heavy metal in agricultural soil[J]. Ecology and Environmental Sciences, 2013, (10):1727-1736.]
[12]	张倩, 赵华甫, 吴克宁, 等. 重金属超标农用地安全利用的地类转换模式研究[J]. 土壤通报, 2016, 47(5):1071-1076.
	[Zhang Q, Zhao H F, Wu K N, et al. Land type transfer mode for the safety utilization of heavy metals over standard agriculture land[J]. Chinese Journal of Soil Science, 2016, 47(5):1071-1076.]
[13]	朱静, 黄标, 孙维侠,等. 农田土壤有效态微量元素的时空变化及其影响因素研究[J]. 南京大学学报(自然科学), 2007, 43(1):1-12.
	[Zhu J, Huang B, Sun W X, et al. Temporal-Spatial Variability and Its Influencing Factors of Available Trace Metals in Soils[J]. Journal of Nanjing University (Natural Sciences), 2007, 43(1):1-12.]
[14]	王振宇. 保护性耕作对冀西北高原土壤物理性质和养分的影响[D]. 保定: 河北农业大学, 2010.
	[Wang Z Y.Effects of Conservation Tillage on Soil Physical Properties and Nutrient in Northwest Plateau of Hebei Province[D]. Baoding: Hebei Agricultural University, 2010.]
[15]	Dayton E A, Basta N T, Payton M E, et al. Evaluating the contribution of soil properties to modifying lead phytoavailability and phytotoxicity[J]. Environmental Toxicology & Chemistry, 2006, 25(3):719-725.
[16]	王劲峰, 徐成东. 地理探测器:原理与展望[J]. 地理学报, 2017, 72(1):116-134.
	[Wang J F, Xu C D.Geodetector: principle and prospective[J]. Acta Geographica Sinica, 2017, 72(1):116-134.]
[17]	张雨浓, 杨逸文, 李巍. 神经网络权值直接确定法[M]. 广州: 中山大学出版社, 2010.
	[Zhang Y N, Yang Y W, Li W.Neural Network Weight Direct Determination Method[M]. Guangzhou: Sun Yat-sen University Press, 2010. ]
[18]	邹志红, 王学良. BP模型在河流水质预测中的误差分析[J]. 环境科学学报, 2007, 27(6):1038-1042.
[1]	Legind C N, Trapp S.Comparison of prediction methods for the uptake of As, Cd and Pb in carrot and lettuce[J]. Sar&Qsar in Environmental Research, 2010, 21(5-6):513-525.
[2]	Pandey J, Pandey U.Atmospheric deposition and heavy metal contamination in an organic farming system in a seasonally dry tropical region of India[J]. Journal of Sustainable Agriculture, 2009, 33(4):361-378.
[18]	[Zou Z H, Wang X L. The errors analysis for river water quality prediction based on BP-modeling[J]. Act a Scientiae Circumstantiae, 2007, 27(6):1038-1042.]
[19]	中华人民共和国国家卫生和计划生育委员会, 国家食品药品监督管理总局. 食品安全国家标准食品污染物限量(GB 2762-2017)[EB/OL](2017-04-14)[2018-09-08]..
[3]	王梦梦, 何梦媛, 苏德纯. 稻田土壤性质与稻米镉含量的定量关系[J]. 环境科学, 2018, (4):1-11.
	[ Wang M M, He M Y, Su D C.Quantitative relationship between paddy soil properties and cadmium content in rice grains[J]. Environmental Science, 2018, (4):1-11.]
[19]	[National Health and Family Planning Commission of People’s Republic of China, China Food and Drug Administration. National Food Safety Standard for Limit of Contaminants in Food(GB 2762-2017)[EB/OL] (2017-04-14)[2018-09-08]..]
[20]	中华人民共和国环境保护部. 关于印发《全国土壤污染状况评价技术规定》的通知(环发[2008]39号)[EB/OL]. (2012-01-12) [2018-09-08]..
[4]	Bešter P K, Lobnik F, Eržen I, et al. Prediction of cadmium concentration in selected home-produced vegetables[J]. Ecotoxicology & Environmental Safety, 2013, 96(6):182-190.
[5]	Hough R L, Young S D, Crout N M J. Modelling of Cd, Cu, Ni, Pb and Zn uptake, by winter wheat and forage maize, from a sewage disposal farm[J]. Soil Use & Management, 2010, 19(1):19-27.
[20]	[Ministry of Environmental Protection of People's Republic of China. Notice on Printing and Distributing the Technical Regulations on the Evaluation of Soil Pollution Status in the Whole Country (Huanfa [2008]No. 39) [EB/OL] (2012-01-12)[2018-09-08].. ]
[21]	陈博, 欧阳竹. 基于BP神经网络的冬小麦耗水预测[J]. 农业工程学报, 2010, 26(4):81-86.
[6]	施亚星, 吴绍华, 周生路,等. 土壤-作物系统中重金属元素吸收、迁移和积累过程模拟[J].环境科学,2016,37(10):3996-4003.
	[Shi Y X, Wu S H, Zhou S L, et al. Simulation of the absorption, migration and accumulation process of heavy metal elements in soil-crop system[J]. Environmental Science, 2016, 37(10):3996-4003.]
[21]	[Chen B, Ouyang Z.Prediction of winter wheat evapotranspiration based on BP neural networks[J]. Transactions of the CSAE, 2010, 26(4): 81-86.]
[22]	夏运生, 王凯荣, 张格丽. 土壤镉生物毒性的影响因素研究进展[J]. 农业环境科学学报, 2002, 21(3):272-275.
	[Xia Y S, Wang K R, Zhang G L.Research advances in influence factors of phytotoxicity of cadmium in soil[J]. Journal of Agro-Environment Science, 2002, 21(3):272-275.]
[23]	牛雅典. 不同油菜品种对Cd耐性差异及其初步机理研究[D].武汉:华中农业大学, 2012.
[7]	Chen H, Yuan X, Li T, et al. Characteristics of heavy metal transfer and their influencing factors in different soil-crop systems of the industrialization region, China[J]. Ecotoxicology & Environmental Safety, 2016, 126(2):193-201.
[8]	Novotná M, Mikeš O, Komprdová K.Development and comparison of regression models for the uptake of metals into various field crops[J]. Environmental Pollution, 2015, 207(3):357-364.
[23]	[Niu Y D.Differences of Cd Tolerance among Different Rapeseed Species and Their Preliminary Mechanism[D]. Wuhan: Huazhong Agricultural University, 2012.]
[24]	张景阳, 潘光友. 多元线性回归与BP神经网络预测模型对比与运用研究[J].昆明理工大学学报(自然科学版), 2013, (6):61-67.
	[Zhang J Y, Pan G Y.Comparison and application of multiple regression and BP neural network prediction model[J]. Journal of Kunming University of Science and Technology (Natural Science Edition), 2013, (6):61-67.]
[25]	冯金飞. 高速公路沿线农田土壤和作物的重金属污染特征及规律[D]. 南京: 南京农业大学, 2010.
	[Feng J F.Heavy Metal Pollution Characteristics and Laws of Farmland Soils and Crops along Expressway[D]. Nanjing: Nanjing Agricultural University, 2010.]
[26]	汤丽玲. 作物吸收Cd的影响因素分析及籽实Cd含量的预测[J]. 农业环境科学学报, 2007, 26(2):699-703.
	[Tang L L.Effects of soil properties on crop Cd uptake and prediction of Cd concentration in grains[J]. Journal of Agro-Environment Science, 2007, 26(2):699-703.]
[27]	严连香, 黄标, 邵学新, 等. 不同工业企业周围土壤-作物系统重金属Pb、Cd的空间变异及其迁移规律[J]. 土壤学报, 2009, 46(1):52-62.
	[Yan L X, Huang B, Shao X X, et al. Spatial variability and transfer of Pb and Cd in soil-crop system around different types of factories[J]. Acta Pedologica Sinica, 2009, 46(1):52-62.]
[28]	Sharma R K, Agrawal M, Marshall F M.Atmospheric deposition of heavy metals (Cu, Zn, Cd and Pb) in Varanasi City, India[J]. Environmental Monitoring & Assessment, 2008, 142(1-3):269-278.

变量	因子	单位	变量	因子	单位
x₁	土壤Cd含量	mg/kg	x₆	土壤Zn含量	mg/kg
x₂	土壤pH值	-	x₇	与交通主干道距离	m
x₃	阳离子交换量（CEC）	cmol/kg	x₈	与建成区距离	m
x₄	土壤有机质含量（SOM）	%	x₉	与采矿用地距离	m
x₅	土壤粘粒含量	%	x₁₀	与灌溉水源距离	m

变量	x₂	x₃	x₄	x₅	x₆	x₇	x₈	x₉	x₁₀
小麦	0.810	0.818	0.839	0.879	0.853	0.754	0.897	0.932	0.758
水稻	0.649	0.849	0.912	0.669	0.896	0.593	0.852	0.890	0.663
油菜籽	0.591	0.783	0.409	0.496	0.824	0.786	0.810	0.716	0.684
蔬菜	0.966	0.962	0.531	0.964	0.958	0.365	0.369	0.535	0.523

超标等级	E	超标程度
Ⅰ	E ≤ 1	未超标
Ⅱ	1< E ≤ 2	轻微超标
Ⅲ	2< E ≤ 3	轻度超标
Ⅳ	3< E ≤ 5	中度超标
Ⅴ	E > 5	重度超标

	R²		RMSE		MAE		MRE
	训练集	检验集	训练集	检验集	训练集	检验集	训练集/%	检验集/%
小麦	0.998**	0.979**	0.010	0.132	0.008	0.087	12.231	27.084
水稻	0.993**	0.969**	0.021	0.129	0.008	0.082	8.995	22.998
油菜籽	0.969**	0.916**	0.009	0.038	0.007	0.029	19.413	27.517
蔬菜	0.961**	0.546*	0.331	0.429	0.240	0.258	134.060	83.530

	预测方程
小麦	lgCd_（小麦）=1.000lg x₁-2.267lg x₂-0.875lg x₃-0.427lg x₄-0.169lgx₅+0.123lg x₆+0.047lg x₇+0.169lg x₈-0.223lg x₉+0.041lg x₁₀+3.491
水稻	lgCd_（水稻）=1.059lg x₁-2.539lg x₂-0.663lg x₃-1.816lg x₄-0.730lg x₅-0.815lg x₆+0.200lg x₇-0.174lg x₈-0.153lg x₉+0.312lg x₁₀+7.699
油菜籽	lgCd_{（油菜籽）}=0.410lg x₁-0.174lg x₂-1.374lg x₃+0.761lg x₄+0.629lg x₅-0.604lg x₆-0.154lg x₇+0.050lg x₈-0.137lg x₉+0.013lg x₁₀+2.304
蔬菜	lgCd_（蔬菜）=0.541lg x₁+2.282lg x₂-1.598lg x₃-0.094lg x₄-1.199lg x₅-0.563lg x₆-0.165lg x₇-0.121lg x₈-0.080lg x₉-0.535lg x₁₀+6.429