Variations in tomato yield and quality in relation to soil properties and evapotranspiration under greenhouse condition

Soil properties and crop evapotranspiration (ET) are two important factors influencing variations in tomato yield and fruit quality. Quantifying the effect of soil properties and ET on variations in yield and fruit quality is critical for adopting precision irrigation and fertilization management practices in a greenhouse. The experiments were conducted to investigate the variations in drip-irrigated tomato yield and quality parameters and their relationships with spatial variability of soil properties from 0 to 20cm depth and ET from 0 to 60cm depth in a greenhouse of Northwest China during March to July 2012 (the 1st season) and October 2012 to April 2013 (the 2nd season). The results showed that tomato yield was more variable than any quality parameters. The coefficients of variation for yield, quality parameters and ET were 18.2, 8.6–14.5 and 8.2–20.0% in the 1st season, and 11.8, 6.3–9.8 and 9.1–11.5% in the 2nd season, respectively. In both seasons, yield was significantly positively correlated with soil moisture content (θv), nitrate nitrogen content and ET, while significantly negatively related to sand content, bulk density and ammonium nitrogen content in soil. The contents of total soluble solids (TSS), vitamin C (VC) and soluble sugar (SS) in fruit and fruit firmness (Fn) were significantly negatively related to θv and ET. The principal component analysis (PCA) was conducted for combination of the selected soil properties and ET, and the retained principal components (PCs) were used in stepwise regression with yield and quality parameters. PCA results showed that a maximum of 83% of the total variability in soil properties and ET was explained by the first four PCs. The stepwise regression analysis indicated that soil properties and ET could explain 81, 65, 66, 66, 64 and 60%, and 69, 47, 35, 52, 33 and 28% of total variance in tomato yield, TSS, VC, SS, organic acid and Fn in the 1st and 2nd season, respectively. The selected soil properties and ET were more important in estimating yield than quality parameters which were difficult to accurately predict in the greenhouse.