Author：Kang Le, Assistant Researcher at Taiwan Agricultural Research Institute
Chen Chun-Chung, Associate Researcher at Taiwan Agricultural Research Institute

The demand for tomatoes as a vegetable remains stable in Taiwan. Many tomato farmers have invested in facilities to mitigate the risks associated with outdoor farming. With the increasing demand in sales channels and markets, tomato farmers are striving to improve the cultivation management technology to meet the contractual cultivation conditions and supply sufficient compliant produce within the scheduled periods. This has become the critical objective for tomato production and cultivation management. The priority of ensuring stable production of compliant tomatoes often surpasses the goal of increasing yields or producing premium-grade products. On the other hand, how sales channels and vendors effectively cope with sudden surges in supply or unavoidable periods of undersupply from the production end is also a practical challenge that must be addressed.

To achieve the above objectives, adopting a crop model to understand the potential growth variations of crops in different environments is essential to adjust the cultivation environment and control the growth of crops. Moreover, utilizing the future weather forecast information to predict the crop growth conditions and proactively respond to them in advance may serve as a potential solution.

TOMGRO, a dynamic model for tomato growth and yield, was published by James W. Jones et al. in 1991. This model employs a series of mathematical equations to depict the quantitative changes in various organs of tomatoes, such as stems, leaves, and fruits, as well as harvesting under greenhouse cultivation from the perspective of carbon balance and source-sink balance. It quantifies the effects of temperature, carbon dioxide, and light on tomato growth, thereby providing greenhouse cultivation managers with the means to develop and evaluate optimal strategies.

This model is based on the fundamental crop mechanisms, such as photosynthesis and respiration, to depict the formulas for growth and development. The crop model of this type is generally considered to produce more consistent simulation results across different environments.

Therefore, we validated and calibrated the accuracy of TOMGRO model for application in Taiwan by incorporating actual temperature and light intensity measurements, using commonly grown large-fruited tomato varieties. The findings revealed that many of the photosynthesis-related formulas had parameters consistent with the testing results found in literatures. However, due to differences in pruning and leaf removal practices commonly used in Taiwan compared to other regions, we introduced additional formulas to depict the cultivation models and adjusted coefficients related to development, leaf growth, and source-sink balance that may vary among different varieties. Subsequently, we re-simulated the growth and yield of large-fruited tomatoes in Taiwan. The results indicated significant improvements in the growth of stem nodes, an increase in leaf area index, and yield simulation. Moreover, parameters and formulas established from two-year survey data were used to simulate the third-year data, and the results obtained were satisfactory. With the changes in stem nodes which serve as the basis for growth and development calculations in the model, and the yield per unit area, both simulated and measured R2 values exceeded 0.98. This indicates that the modified TOMGRO model, when provided with temperature and light intensity data and aligned with the cultivation practices in Taiwan, can accurately simulate and predict the growth and yield of large-fruited tomatoes under conditions of well-managed plant nutrient supply and pest control.

These findings offer us the chance to convert environmental information into crop growth information through crop model. This enables the greenhouse managers to shift their decision-making emphasis from environmental control to crop growth management when handling greenhouse environmental control facilities, thereby regulating the crop growth to better match fluctuations in yield with contract supply planning.

On the other hand, since the TOMGRO model assumes optimal conditions for tomato nutrient supply and pest control, if continuous discrepancies between actual growth conditions and simulation results are presented, it may indicate potential challenges with nutrient supply or pest infestation. This serves as a reminder for farmers to inspect the greenhouse related systems and implement the cultivation management adjustments accordingly.

For tomato farmers with limited environmental control capabilities or those lacking relevant control measures, there is still an opportunity to utilize the TOMGRO model to forecast tomato growth and yield based on weather forecast data. This enables farmers to arrange manpower in advance and provides sufficient time for logistics and sales teams to respond with appropriate supply and marketing strategies.

The crop models may offer assistance in production, logistics, or sales management. Take TOMGRO for example. Despite being in development for over three decades, TOMGRO has yet to become widely applied in the tomato industry. This could be attributed to the complexity of the equations within the crop model, making it difficult to evaluate their benefits. However, advancements in computer technology, such as improved access to environmental monitoring data, enhanced accuracy and accessibility of weather forecasts, and the widespread availability of cloud and mobile device functionalities, have made the application of the crop model more convenient. We anticipate that with the assistance of agricultural digital service industries, TOMGRO can be simplified and made more accessible, providing relevant decision-making support for the tomato industry and contributing to greater stability in tomato production and distribution in the near future.