龙井茶适制茶树品种多性状综合评价——基于多元统计分析

确定适宜加工龙井茶的茶树品种的多性状评价指标,建立龙井茶适制品种农艺性状的综合评价体系,为鉴定筛选优良龙井茶适制品种或资源提供技术支撑。对浙江省15份主要栽培茶树品种（系）的芽叶长宽比、百芽重、展叶角度、亮度值(L)、红绿属性(a)、黄蓝属性(b)、色彩饱和度(C)、氨基酸、茶多酚、水浸出物、咖啡碱、酚氨比、外形、汤色、香气、滋味和叶底17项农艺性状进行综合比较,并联合使用主成分分析、聚类分析和层次分析对茶树品种（系）的龙井茶适制性进行综合评价。17项农艺性状的平均变异系数为12.05%,百芽重的变异系数最大(28.15%),叶底的变异系数最小(0.86%),适制龙井茶农艺性状变量间既相对独立又密切相关,应用主成分分析法,将17个评价指标压缩成6个综合指标,经过分析6个主成分函数式中的17项农艺性状系数,可以将17项农艺性状分为外形因子、内质因子、感官因子3个主要指标,这3个主要指标可以较准确地评价龙井茶适制品种,其中芽叶长宽比、a值、茶多酚、酚氨比、外形、香气和汤色是主要性状。采用聚类分析,可以将15个品种分为2个大的类群,采用层次分析可以将茶树品种采制龙井茶的适宜性进行综合排序。采用多元统计分析中的主成分分析和聚类分析对15份茶叶品种的17项农艺性状进行龙井茶适制性综合评价是可行的,结合层次分析可以得到更加准确的排序,可以从不同角度进行较全面客观评价,为龙井茶适制品种的选育提供参考。

Comprehensive Evaluation on Characters of Tea Varieties Suitable for Longjing Tea: Based on Multiple Statistic

The paper aims to determine the multi-trait evaluation index of tea varieties suitable for processing Longjing tea, and to establish a comprehensive evaluation system for agronomic traits of Longjing tea varieties, so as to provide technical support for identifying and screening suitable varieties or resources of Longjing tea. A total of 17 agronomic traits including leaf length-width ratio, 100-bud weight, leaf unfolding angle, L, a, b, C, amino acids, tea polyphenols, water extract, caffeine, phenol-ammonia ratio, appearance, soup color, aroma, taste and leaf bottom of 15 main tea cultivars (lines) in Zhejiang were compared. The suitability of Longjing tea varieties (lines) was evaluated by principal component analysis, cluster analysis and hierarchical analysis. The average coefficient of variation of 17 agronomic traits was 12.05%, the coefficient of variation of 100-bud weight was the largest (28.15%), and the coefficient of variation of leaf bottom was the smallest (0.86%). The variables of agronomic traits of Longjing tea were relatively independent and closely related. By analyzing 17 agronomic trait coefficients of 6 principal component function formulas, 17 agronomic traits could be divided into 3 main indexes: shape factor, endoplasmic factor and sensory factor. These 3 main indexes could accurately evaluate the suitable varieties of Longjing tea, among which the ratio of bud length to leaf width, a value, tea polyphenols, phenol to ammonia ratio, shape, aroma and soup color were main characters. By cluster analysis, 15 varieties could be divided into 2 large groups, and the suitability of Longjing tea could be sorted by AHP. It is feasible to comprehensively evaluate 17 agronomic traits of Longjing tea by using principal component analysis and cluster analysis in multivariate statistical analysis. Combining with AHP, more accurate ranking can be obtained, and comprehensive and objective evaluation can be made from different angles, which could provide references for tea plant breeding suitable for Longjing tea.