Response surface optimization of cultivation conditions for yield of tomato (Lycopersicon esculentum Mill.) in greenhouses

There is a great potential for greenhouse tomato fruit yield improvement in China for the low yield per hectare. We evaluated the effects of multi-factors (plant density, nitrogen (N) and K₂O fertilizer) on fruit yield of tomato (Lycopersicon esculentum Mill. cv. Jinfan 4) by response surface methodology with a 5-level-3-factor central composite design. A multivariate quadratic regression model of fruit yield was established. Results showed that N fertilizer was the most significant for fruit yield, followed by K₂O fertilizer and plant density. Fruit yield showed a parabolic trend with increasing fertilizer levels or plant density. There was a significant interaction effect between plant density and fertilizer levels. Optimal conditions were obtained: 4.83 × 10⁴ plants ha^?1 for density, 262 kg ha^?1 for N and 513 kg ha^?1 for K₂O. Under these conditions, the predicted fruit yield was 119,381 kg ha^?1, while the actual fruit yield from verification test was 121,005 kg ha^?1.