Analyzing the effect of light on lettuce Fv/Fm and growth by machine learning

Abstract

Varied plant growth cultivations require different light conditions. To explore them with lettuce (Lactuca sativa L. cv. ‘Hong Kong glass’), five photosynthetic photon flux density (PPFD, 50, 150, 300, 450 and 600 μmol·m−2·s−1) levels were set in completely controlled circumstances. Using ordinary least squares (OLS), the appropriate PPFD was determined depending on the dark fluorescence characteristics. Following 14 h of treatment, a new PPFD of 200 μmol·m−2·s−1 was employed to monitor the adaptation process obtained using the genetic algorithm-support vector regression (GA-SVR) model. The results showed that Fv/Fm declined as PPFDs increased except for L50, and OLS with R2 and RMSE of 0.9710 and 0.0416 suggested 140.80 μmol·m−2·s−1 was the ideal PPFD. Little changed in the new PPFD adaptation procedure for, L50, L150, and L300. During 42 h of adaptation, the discretized curves of Fv/Fm using the GA-SVR model at various PPFDs revealed that Fv/Fm increasingly approached and eventually reached the same level. The quick change was in [290, 480] μmol·m−2·s−1 depending on the probability density distributions. The difference in change rates between the two PPFD treatments decreased, indicating that lettuce's light tolerance improved after the two phases. The largest sum Fv/Fm in L150 indicated a maintained growth state. Full-cycle cultivation revealed that 150 and 250 μmol·m−2·s−1 had the highest fresh biomass. In conclusion, in [290, 480] μmol·m−2·s−1, short-term high light could improve plant light adaptability in variable light. In the constant light of the plant factory with artificial lighting (PFAL), it was recommended to utilize the light intensity of 140.80 μmol·m−2·s−1, which can both maintain growth and save light energy. A more suitable solution might be chosen depending on the cultivation requirements.