Resumen
Rock zones have an important influence on the yield and quality of Wuyi rock tea. In this study, OPLS-DA combined with machine learning was used to analyze the effects of different rock zones and tea tree varieties on the physicochemical properties of rhizosphere soil, the growth of the tea tree and the quality of the tea leaves using tea trees in different rock zones. The results showed that rock zones had significant effects on rhizosphere soil physicochemical indexes, soil enzyme activities, tea tree growth and tea quality indexes, while there was little difference between different tea tree varieties. The interaction analysis showed that the physicochemical indexes of rhizosphere soil in different rock zones significantly affected tea quality, while also affecting growth indexes. The main indexes affecting tea yield and caffeine content were soil pH, available nitrogen, total phosphorus, total nitrogen and available phosphorus, while the main indexes affecting tea quality were available potassium, organic matter, total potassium, protease, polyphenol oxidase and urease. Analyses of PCA, OPLS-DA models and KNN and ANN machine learning showed that different rock zones could be effectively distinguished from each other with 100% accuracy, while different tea varieties had little difference and could not be distinguished. TOPSIS analysis found that the physicochemical indexes most affected by rock zone were available nitrogen, available potassium and sucrose, and the quality indexes most affected by rock zone were tea polyphenols and theanine. The growth index most affected by rock zone was tea yield. It was evident that the key difference between tea trees in different rock zones was yield and quality, with high yields in continent zones, and good quality in semi-rock zones and rock zones. This study provides a crucial foundation for tea-plantation management, the artificial regulation of tea yield and the quality of different rock zones of Wuyi rock tea.