Objectives The diverse types of vegetables have varying nutrient demand characteristics. We constructed a tomato Nutrient Expert system based on yield response and agronomic efficiency for fertilizer recommendation, aiming to provide a scientific, simplified, and efficient fertilizer recommendation method for tomato production in China.

Methods We collected data from 314 tomato field experiments conducted between 2000 and 2023, calculated parameters such as yield response, relative yield, agronomic efficiency, and internal nutrient efficiency, and used the QUEFTS model to estimate the optimal nutrient demand parameters for tomatoes at different target yields. Based on this, we established a tomato Nutrient Expert system (NE) that incorporates yield response and agronomic efficiency. In 2023, field validation experiments were conducted at seven locations in Tianjin and Xuzhou of Jiangsu Province, with six treatments: farmer practice (FP), soil testing-based recommendation (ST), tomato Nutrient Expert system recommendation (NE), and three additional treatments based on NE but without N, P, or K application, respectively. Tomato yield, economic benefits, nutrient uptake by plants (N, P, K), apparent nutrient balance, and nutrient use efficiency were investigated.

Results The average results from our existing experiments in China were: tomato yields 73.34 t/hm², harvest index 0.54 kg/kg, the aboveground of plant uptake 218.60 kg/hm² for N, 52.66 kg/hm² for P, and 370.14 kg/hm² for K, with respective nutrient harvest indices of 0.51, 0.52, and 0.52 kg/kg, and internal nutrient efficiencies of 352.75, 1445.45, and 209.64 kg/kg for N, P, and K. The QUEFTS model results indicated that within the range of achieving 70% of potential yield, the amounts of N, P, and K required for 1 t of tomato fruit yield were constant, at 2.6, 0.7, and 4.6 kg, respectively, with a nutrient demand ratio of approximately 3.8:1.0:6.8 for N:P:K. There was a highly significant negative linear correlation between tomato yield response to N, P, and K fertilizers and relative yield (P<0.01); and a highly significant quadratic relationship between yield response and agronomic efficiency (P<0.01). Field validation results showed that compared with FP, NE reduced N, P, and K fertilizer application rates by 31.3%, 59.2%, and 22.7%, respectively, while increasing tomato yield and net economic benefits by 2.0% and 4.2%, respectively. Compared with ST, NE reduced N and P fertilizer application but increased K fertilizer application. The apparent nutrient balance results showed that FP had large surpluses of N, P, and K; ST had higher surpluses of N and P but a deficit of K; and NE had nutrient surpluses closest to balance for N, P, and K. The recovery rates of N, P, and K fertilizers with NE were higher than those of FP with 13.1, 7.9 and 13.6 percentage points, respectively, and higher than those of ST with 6.8, 6.7 and 1.9 percentage points, respectively. The partial factor productivities of N and P with NE were higher than those with FP and ST, while the partial factor productivity of K with NE was higher than that with FP but lower than that with ST.

Conclusion The tomato Nutrient Expert system, built based on yield response and agronomic efficiency, further optimizes fertilizer application rates and proportions compared with conventional recommendation methods under different production conditions across China. It enhances tomato yield and fertilizer use efficiency while reducing soil nutrient surpluses. Therefore, it is a recommended fertilizer application method suitable for smallholder tomato cultivation in various regions of China.