Modeling relations of tomato yield and fruit quality with water deficit at different growth stages under greenhouse condition

Nowadays more and more attentions are paid to fruit quality in the production of tomato. In order to better understand the effects of deficit irrigation on tomato yield and fruit quality, four years of deficit irrigation experiments were investigated to simulate water-yield and water-fruit quality relationships of tomato in greenhouses. The yield and fruit quality parameters like total soluble solids (TSS), reducing sugars (RS), organic acids (OA), sugar/acid content ratio (SAR), vitamin C (VC), firmness (Fn), color index (CI) were correlated with seasonal evapotranspiration (ET) and ET deficit at flowering and fruit development stage (Stage II) and fruit ripening stage (Stage III) using linear regression. Three water-yield models (Jensen, Stewart, Minhas) and three water-fruit quality models (multiplicative, additive, exponential) were applied to simulate the relationships of tomato yield and fruit quality parameters with water deficit at various growth stages. The water deficit sensitivity indexes (λ/Ky/δ or γ/Kq/ψ) of the models were calculated with the method of multiply linear regression. The performance and sensitivity analysis of the models were evaluated. Results showed that the relative yield decreased linearly with the drop of relative seasonal ET, mainly due to the yield depression by ET deficit at Stage II and Stage III; the relative values of fruit quality parameters increased with the drop of relative seasonal ET, mostly because of the enhancement by ET deficit at Stage III. The calculated water deficit sensitivity indexes indicated that both the yield and fruit quality were hardly sensitive to water deficit at Stage I, but sensitive to water deficit at Stage II and that at Stage III; TSS, RS, SAR and VC were much more sensitive to water deficit at Stage III than that at Stage II; RS, SAR and VC were more sensitive to water deficit than TSS, OA, Fn and CI. The Minhas model with its water deficit sensitivity indexes was recommended to simulate water-yield relations of greenhouse tomato in the study area; multiplicative model and additive model were, respectively, recommended to simulate the relationships of fruit quality parameters like TSS, RS, SAR, Fn and fruit quality parameters like OA, VC, CI with water deficit at various growth stages. The water-yield and water-fruit quality models would be helpful to optimally allocate irrigation water during the growth season, thus achieving efficient production of tomato in greenhouses in consideration of the compromise between tomato yield and fruit quality.