Using a Real-Time Photosynthesis and Transpiration Monitoring System to Develop Random Forests Models for Predicting Cherry Tomato Yield in a Commercial Greenhouse

• DOI: 10.2525/ecb.62.29
• Publication Date: 2024-05-09T22:14:42Z
• Authors: Naomichi FUJIUCHI; Kazue INABA; Shinchu OH; Sayaka OKAJIMA; Yuichiro ASAI; Hiroshige NISHINA; Kotaro TAKAYAMA

Predicting the yield of horticultural crops is crucial to meet expectations retailers and consumers. In this study, we developed random forests (RF) based on measured amounts whole-plant photosynthesis transpiration predict cherry tomato fruit yields in a commercial greenhouse Japan. Whole-plant daily net (Photo) (Trans) were by using real-time monitoring system. Variables environmental conditions (Env), including solar irradiation, air temperature, atmospheric water vapor deficit, also greenhouse. Data with different 7 variable combinations (Env, Photo, Trans, Env+Photo, Env+Trans, Photo+Trans, Env+Photo+Trans) 21 timeframes (from 1 6 consecutive weeks past weeks) used train models for predicting subsequent week. RF 3 until 2 before date prediction (3W2) 4W2 Photo+Trans had relatively low normalized root mean square error (RMSE%; 9.8-10.3%). The model that timeframe combination Photo best accuracy (RMSE% = 9.8%). These indicate are good predictors yield.