The green tea leafhopper, Matsumurasca onukii, is an important pest that feeds on new tea shoots, causing substantial damage and significantly reducing plucking yield. However, tolerable injury levels and control thresholds for efficient pest management are not clearly defined. Here, we conducted three-year controlled experiments with different insecticide treatments in 'Yabukita' tea fields in Shizuoka Prefecture, and analyzed the relationships among M. onukii population density (assessed by beating method), shoot damage, and plucking yield of second-flush tea. Results showed significant negative correlations between pre-plucking insect density and damage severity with 100-bud weight (yield index). Yield losses exceeding 50% were observed under high infestation conditions, demonstrating the serious impact of this pest. Using regression equations derived from these relationships, we estimated tolerable injury levels based on 100-bud weight and corresponding control thresholds. When the tolerance level was set at 5% yield reduction, the percentage of damaged shoots and degree of damage (scale: 0-1) were 8.0% and 0.06, respectively. The control threshold at the optimal timing (early leaf opening stage) was calculated as 0.35 insects per beating point. For 10% yield loss tolerance, these values were 15.9%, 0.12 and 0.69 insects per beating point, respectively. These thresholds were prominently low, suggesting that pest control is necessary in most tea fields during the second flush season. Effective management of M. onukii requires an integrated pest management approach, combining appropriate insecticide selection with cultural control practices.

In the present study, we evaluated the effectiveness of tunnel-type covering cultivation using shading materials as a countermeasure against high summer temperatures in newly transplanted tea seedlings. Tunnel coverings with varying shade levels (30–85%) were applied during the summers of 2021–2023 and 2025, and both aboveground and belowground growth responses were assessed.

Shade treatments of 30–60% maintained a light environment sufficient to meet the light saturation point while suppressing increases in leaf temperature, thereby improving transpirational conditions through a reduction in the leaf-to-air vapour pressure deficit (VPD_leaf). Aboveground growth, assessed using the Plant Growth Index (PGI), and belowground growth, evaluated by image analysis of excavated root systems, were significantly enhanced under shaded conditions compared with the unshaded control.

However, under the 85% shade treatment, although aboveground growth improved relative to that of the control, belowground growth remained comparable to that observed in the unshaded treatment. These findings indicate that 30–60% shading is effective in promoting both aboveground and belowground growth in transplanted tea seedlings under high summer temperatures.

In addition, a controlled environmental experiment examining the effects of root-zone temperature and exposure duration revealed that root respiratory activity declined prior to the appearance of visible symptoms in the aboveground parts. This finding suggests that under high-temperature conditions, belowground impairment in tea seedlings may precede aboveground responses.

This study aimed to investigate the potential of using Unmanned Aerial Vehicles (UAVs) for diagnosing the condition of tea branches after pruning. The results revealed significant correlations between the spectral characteristics of tea branches approximately one month after pruning and those of leaves at the same locations approximately two months after pruning.

Multispectral images with four bands (green, red, red-edge, and near-infrared) were acquired at a tea plantation in Nichinan City, Miyazaki Prefecture, using a DJI Mavic 3 Multispectral drone at approximately one and two months after pruning. Reflectance and NDVI images were generated using Pix4Dmapper, and branch, leaf, and shadow classes were classified using a Support Vector Machine (SVM) algorithm in ArcGIS Pro. The classification accuracy for the branch class reached 93 %, supporting the reliability of the extracted branch data.

Correlation analysis was conducted using 76 analysis areas, comparing the mean spectral values of branch pixels at one month with those of leaf pixels at two months. Strong correlations were observed, with absolute correlation coefficients exceeding 0.7 for indices and bands such as NDVI, near-infrared (NIR), and red. Notably, the correlation between branch NIR and leaf NDVI suggests that the internal structural properties of branches may influence subsequent leaf development. In addition, the characteristics of the red band in branches indicate the possible presence of chlorophyll within the bark.

This study provides fundamental insights toward the development of diagnostic indicators for tea plant growth following medium pruning, highlighting the potential of UAV-based spectral analysis as a non-destructive tool for early growth assessment in tea cultivation.

As tea farmers manage larger areas in Shizuoka Prefecture, the demand for simplified methods to evaluate leaf maturity is growing. This study developed a technique to estimate key indicators of leaf maturity—total nitrogen (TN) and neutral detergent fiber (NDF)—using deep learning models applied to images of fresh tea leaves captured with low-cost devices. Using a model trained on the first flush tea of ‘Yabukita’ cultivar, we confirmed that TN estimation is possible. The effects of image sampling frequency, leaf quantity, and leaf withering on prediction accuracy were also examined, providing practical insights for field application. While the model showed strong potential for the first flush tea of ‘Yabukita’ cultivar, limitations were observed for other cultivars and withered leaves, underscoring the need for further model development and validation under diverse field conditions.

We examined factors affecting profitability using data on unit price, cultivation management, and tea composition from the Kagoshima tea market. Unit price was negatively associated with yield. Gross profit was mainly influenced by bud number and harvest date, while unit price and yield were affected by deep trimming height. A neural network model using fertilization data accurately predicted bud number and harvest date. Both variables were also influenced by tea variety and final harvest timing. Total nitrogen content was strongly affected by shading period, whereas the effect of applied nitrogen fertilizer was limited. These results indicate the potential for data-driven tea production based on field data.