Objectives Excessive phosphorous fertilization is common in facility vegetable production. In this study, the effect of reducing phosphorus application was investigated from the points of crop nutrition and soil available P accumulation, in order to provide theoretical basis for phosphorus fertilizer reduction in facilities.

Methods A field experiment was conducted inside a plastic greenhouse, using tomato as tested materials. The conventional phosphorus application rate of 891 kg/hm² was taken as the control (CK), two phosphorus reduction treatments were set, including 50% less (P1) and 70% less (P2) in 2015, and extra treatment of no phosphorus (P0) in 2016. The dry weight of roots, stems, leaves and fruits were investigated, and the N, P and K contents in the organs were measured in 2016 during fruit expanding and full-fruiting stages. The available phosphorus content in 0–60 cm soil layer was measured before planting, during full-fruiting and during uprooting stage.

Results In the soil with high available P (about 220 mg/kg), reducing 70% of P input for two consecutive years or no P input for one year did not affect tomato yield. Tomato yields in 2015 and 2016 were 53.9–55.1 t/hm² and 50.2–52.7 t/hm², respectively. Compared with CK, all other treatments significantly increased the dry matter weight and the allocation rate of NPK in fruits, and decreased those in leaves at full fruit bearing stage. During fruit-expanding period, 62.8%–65.7% of the dry matter was distributed in the leaves of tomato plants, and the total N, P and K uptake by the plants were 83.2–89.9 kg/hm², 10.3–11.1 kg/hm², 75.0–85.9 kg/hm², respectively. At this time, leaves and stems of tomato were the main accumulative parts of nutrients. The distribution rates of N, P and K in stems and leaves were 84.4%–86.4%, 79.4%–83.4%, 76.9%–82.3%, and the absorption ratio of N, P and K in tomatoes was 1∶0.12∶0.84–0.96. During the full fruiting period, 43.0%–44.6% and 37.0%–44.6% of dry matter were distributed in fruits and leaves. At this time, tomato fruits and leaves were the main accumulative parts of nutrients. The sum of N, P and K distribution rates in fruits and leaves were 84.6%–86.7%, 78.5%–82.7%, 81.4%–83.9%. The N, P and K uptake of the plants were 197–226 kg/hm², 33–37 kg/hm² and 200–247 kg/hm², respectively. The total K content and uptake of tomato leaves and fruits treated with CK were significantly higher than those treated with P reduction, which showed that reducing P fertilizer could reduce the luxury absorption of K in tomato. The accumulation of available P in surface soil decreased significantly after two years of P reduction, but P migrated downward in all soil profiles. The soil available P content at fruit-expanding stage in 0–20 cm soil layer decreased by 27.0–60.9 mg/kg, compared with that before planting, and the increment of available P in 20–60 cm soil layer was 11.8–50.1 mg/kg. The increment of available P in 20–60 cm soil layer was significantly reduced by P reduction treatments.

Conclusions In the soil with high available P, reducing 70% of P will not affect tomato yield, but reduce the luxury absorption of K, and slow down the accumulation of soil available P. However, the content of soil available P is still at a high level, indicating that the potential of P reduction in greenhouse vegetable production is great in the studied area.