湖北农业科学 ›› 2023, Vol. 62 ›› Issue (6): 46-50.doi: 10.14088/j.cnki.issn0439-8114.2023.06.009

• 资源·环境 • 上一篇    下一篇

湖北省绿茶气候品质综合指数模型研究

邓环环1,2,3, 秦鹏程1,2,3, 万素琴1,2,3, 邓爱娟1,2,3, 汤阳1,2,3   

  1. 1.武汉区域气候中心,武汉 430074;
    2.三峡国家气候观象台,湖北 宜昌 443002;
    3.中国气象局流域强降水重点开放实验室,武汉 430205
  • 收稿日期:2022-05-07 出版日期:2023-06-25 发布日期:2023-07-18
  • 通讯作者: 万素琴,正高级工程师,主要从事农业气象研究,(电子信箱)11662106@qq.com。
  • 作者简介:邓环环(1980-),女,湖北钟祥人,高级工程师,主要从事生态与农业气象研究,(电话)18971382090(电子信箱)11662106@qq.com。
  • 基金资助:
    湖北省科技基金项目(2019Y05)

Study on climatological quality evaluation model with comprehensive indexes for green tea in Hubei Province

DENG Huan-huan1,2,3, QIN Peng-cheng1,2,3, WAN Su-qin1,2,3, DENG Ai-juan1,2,3, TANG Yang1,2,3   

  1. 1. Wuhan Regional Climate Center, Wuhan 430074, China;
    2. Three Gorges National Climatological Observatory,Yichang 443002,Hubei,China;
    3. Key Laboratory of Basin Heavy Rainfall,CMA,Wuhan 430205,China
  • Received:2022-05-07 Online:2023-06-25 Published:2023-07-18

摘要: 利用湖北省典型茶区2018—2020年绿茶不同开采期品质数据及开采前15 d气象观测数据,首先通过决策树和随机森林等机器学习方法分析了影响绿茶品质的关键气象因子及其响应关系,然后基于模糊数学理论,构建了气温、日照时数、风速和相对湿度单因子隶属函数模型,并采用综合加权得到绿茶气候品质综合指数模型,最后基于遗传算法对模型参数进行优化求解,并确定了等级评价标准。结果表明,符合实际等级的样本占样本总数的67.2%,相差一个等级的样本占样本总数的32.8%,说明绿茶气候品质综合指数模型能够反映不同气候条件下绿茶品质的差异。

关键词: 绿茶, 气候品质, 评价模型, 机器学习, 湖北省

Abstract: Quality analysis data of green tea within different green tea-plucking periods and observation data of meteorological elements at 15 days before spring tea-plucking in typical tea-producing regions of Hubei Province from 2018 to 2020 was used. Firstly, the key meteorological elements affecting green tea quality and its response relationship were analyzed by using machine learning methods including decision tree and random forest model. Secondly, the single-factor subordinate function model for temperature, sunshine, wind speed and relative humidity was built based on fuzzy mathematics theory. In the meanwhile, a comprehensive index model for climatological quality of green tea was built by using the comprehensive weighted method. Finally, parameters in this model were optimized based on the genetic algorithm. Additionally, grade evaluation standard was also determined. The results showed that the number of samples corresponding and existing one grade difference to the actual grade of samples accounted for 67.2% and 32.8% of the total samples, respectively. The results indicated that this model for green tea could reflect its quality difference under different climate situations.

Key words: green tea, climatological quality, evaluation model, machine learning, Hubei Province

中图分类号: