Establishment and verification of prediction models for evaluating the physical and chemical properties of soilless substrates

In soilless culture, a suitable mixed substrate that provides a balanced and stable rhizosphere environment is vital for promoting plant growth. The present study was undertaken to establish seven prediction models of physical and chemical properties, including bulk density (DB), total porosity (TP), water-holding porosity (WHP), air porosity (AP), WHP/AP, electrical conductivity (EC) and cation exchange capacity (CEC) of mixed substrate based on regression equations of measured values from 76 substrate combinations. These seven models were verified using the measured values of 12 mixed substrates, and the average relative prediction errors (REs) were all less than 10%. A comprehensive property prediction model was established by weighted summation of the seven models of physical and chemical properties. According to the set values of DB, TP, WHP, AP, WHP/AP, EC and CEC, the comprehensive property model predicted the 6 mixture proportions of mixed-substrate, as verified using the measured values. This study is the first to establish prediction models of the physical and chemical properties of mixed substrates. The comprehensive property model could be used to evaluate the physical and chemical properties of commercial mixed substrates, and to provide the optimal mixture substrate formulations according to the setting property value of production requirement. Keywords: prediction model, mixed substrate, physical and chemical properties, multiple regression, genetic algorithm DOI: 10.25165/j.ijabe.20211402.5815 Citation: Gong B B, Wang N, Zhang T J, Li S, Wu X L, Tian J, et al. Establishment and verification of prediction models for evaluating the physical and chemical properties of soilless substrates. Int J Agric & Biol Eng, 2021; 14(2): 9–18.