欢迎访问《中国农学通报》,

中国农学通报 ›› 2021, Vol. 37 ›› Issue (2): 108-115.doi: 10.11924/j.issn.1000-6850.casb20200300217

所属专题: 玉米 烟草种植与生产

• 资源·环境·生态·土壤·气象 • 上一篇    下一篇

基于环境因子的春玉米产量结构模型分析研究

许艺馨1(), 冯磊1, 苏永秀2(), 莫申萍3, 刘军4   

  1. 1广西物流职业技术学院,广西贵港537100
    2广西壮族自治区气象科学研究所,南宁 530022
    3贵港市气象局,广西贵港 5371003
    4贵港市农业技术推广中心,广西贵港 537100
  • 收稿日期:2020-03-18 修回日期:2020-06-28 出版日期:2021-01-15 发布日期:2021-01-14
  • 通讯作者: 苏永秀
  • 作者简介:许艺馨,女,1989年出生,广西贵港人,工程师,硕士,主要研究方向为农业气象。通信地址:537100 广西壮族自治区贵港市金港大道580号 贵港市气象局,E-mail:261786751@qq.com
  • 基金资助:
    公益性行业(气象)科研专项“我国东南部地区主要热带果树的关键气象保障技术研究”(GYHY201406027);广西科技计划项目“广西新兴热带果树寒冻害风险预警与精细区划技术研究”(桂科AB16380260)

The Yield Structure Model of Spring Corn Based on Environmental Factors

Xu Yixin1(), Feng Lei1, Su Yongxiu2(), Mo Shenping3, Liu Jun4   

  1. 1Guangxi Logistics Vocational and Technical College, Guigang Guangxi 537100
    2Guangxi Meteorological Disaster Mitigation Institute, Nanning Guangxi 530022
    3Guigang Meteorologic Bureau, Guigang Guangxi 537100
    4Guigang Agricultural Technology Extension Center, Guigang Guangxi 537100
  • Received:2020-03-18 Revised:2020-06-28 Online:2021-01-15 Published:2021-01-14
  • Contact: Su Yongxiu

摘要:

环境因子对作物产量的影响是现代农业气象研究的重要内容之一,建立春玉米产量结构模型可为春玉米的科学生产提供依据。本研究分析贵港春玉米不同生育阶段的环境因子与产量结构的相关性,并建立全因子、显著因子的多元线性回归模型和BP神经网络模型。结果表明,对春玉米产量结构影响最大的生育期为拔节—抽雄期,10~40 cm的土壤水分体积含水率与产量结构最为密切;四种产量结构预测模型优度(R2)比较,全因子模型(AF)优于显著因子模型(SF),多元线性回归(MLR)模型优于BP神经网络(BPNN)模型。试报检验模型发现MLR模型的泛化能力不及BPNN模型,其中BPNN_AF模型对理论产量、果穗粗的预测最为精准。BPNN全因子模型(BPNN_AF)可作为春玉米产量结构预测的最优模型,能较好捕捉作物产量结构与环境因子之间的非线性影响规律,预测结果较为合理准确。

关键词: 气象, 土壤, 产量结构, 相关分析, 模型

Abstract:

The impact of environmental factors on crop yield is one of the important contents of modern agrometeorological research. Establishing a spring corn yield structure model can provide a basis for the scientific production of spring corn. The correlation between environmental factors and yield structure in different growth stages of spring corn in Guigang was analyzed. And multiple linear regression models and BP neural network models with full factors and significant factors were established. The results showed that the growth period that has the greatest impact on the yield structure of spring corn is from jointing to tasseling. The soil water volume and moisture content of 10 cm to 40 cm were the most closely related to the yield structure. The four models of yield structure prediction (R2) comparison showed that the factor model (AF) was superior to the significant factor model (SF).The multiple linear regression (MLR) model was superior to the BP neural network (BPNN) model. The test reported model found that the generalization ability of the MLR model was not as good as that of the BPNN model. Among them, the BPNN_AF model was the most accurate predictor of theoretical yield and ear diameter. The BPNN full factor model (BPNN_AF) can be used as the optimal model for predicting the yield structure of spring corn, it can better capture the non-linear influence law between the crop yield structure and environmental factors, and the prediction results are more reasonable and accurate.

Key words: meteorology, soil, yield structure, correlation analysis, model

中图分类号: