The Implementation of Response Surface Methodology and Artificial Neural Networks to Find the Best Germination Conditions for <i>Lycopersicon esculetum</i> Based on Its Phenological Development in a Greenhouse

Abstract

The incorporation of biodegraded substrates during the germination of horticultural crops has shown favorable responses in different crops; however, most of these studies evaluate their effect only in the first days of seedling life, and do not follow up on the production process under greenhouse or open field conditions. The objective of this study was to evaluate the phenological development of Lycopersicon esculetum (tomato) seedlings in greenhouses that were germinated with biodegraded substrate mixed with peat moss. To find the best plant performance condition and determine whether the biodegraded substrate allows tomato plants to be obtained with the conditions for their production, the response surface methodology (RSM) and artificial neural network (ANN) were used. Three response surface models and three neural network models were developed to analyze the plant growth, the leaf length and the leaf width. The results obtained show that plant height during the first days presented statistically significant differences among the different treatments, with an initial average height of 5.3 cm. The length of the leaves at transplantation was statistically different, maintaining a length of 2.4, and the width of the leaves at transplantation measured 1.8 cm. The RSM and ANN models allowed the estimation of the optimal value of the adequate amount of degraded substrate to germinate Lycopersicon esculetum and reduce the use of peat moss. The coefficient of determination (r2) indicates that the ANNs presented a better data fit (r2 > 0.99) to predict the experimental conditions that maximize the study variables; in this sense, the plants obtained with 100% biodegraded substrate showed a better development, which suggests its use as an alternative substrate in the germination process and to reduce the use of peat moss.

Keywords

• biodegraded substrate
• peat moss
• artificial neural networks (ANN)
• response surface methodology (RSM)
• organic agriculture
• soils