Field-Based Spectral and Metabolomic Analysis of Tea Geometrid (Ectropis grisescens) Feeding Stress

Abstract

Tea is one of the most widely consumed non-alcoholic beverages globally, yet its yield and quality are significantly impacted by herbivory from tea geometrids. To accurately detect herbivory stress in tea leaves, this study integrated metabolomics with visible-near-infrared spectroscopy (VIS-NIRS) to explore its in situ capabilities and underlying mechanisms. The results demonstrated that metabolomic data, combined with PCA-based linear dimensionality reduction, could effectively distinguish between tea leaves subjected to herbivory by different densities of tea geometrids. VIS-NIRS successfully identified herbivore-damaged leaves, achieving an optimal average classification accuracy of 0.857. Furthermore, VIS-NIRS was able to differentiate leaves subjected to herbivory on different days. The application of appropriate preprocessing techniques significantly enhanced temporal classification, achieving the highest average classification accuracy of 0.773. By integrating metabolomics and spectral band analysis, the spectral range of 800–2500 nm was found to more accurately identify leaves exposed to herbivory for a prolonged period. Compared to using the full spectrum, the model built within this wavelength range improved classification accuracy by 10%. In conclusion, this study provides a solid theoretical foundation for the in situ, rapid detection of tea geometrid herbivory stress in the field using VIS-NIRS, offering key technical support for future applications.

Plant phenotyping relevance

VIS-NIRSを用いて茶葉の食害ストレスを直接推定し、前処理・波長選択・分類精度を評価する手法研究であり、植物状態の取得方法が中心的です。

Code and data availability

The supplied blocks describe field VIS-NIRS spectral collection, metabolomics, and PLSDA/RF modeling in Python, but contain no data availability statement, public dataset deposit, or author code repository URL. No paper-specific public asset is identifiable.

No evidence-backed public reproduction asset is currently recorded.