Objectives In order to explore long-term effect of different fertilization treatments on crop yields and evolution of soil fertility, a reasonable integrated evaluation method of soil fertility was studied in this paper.
Methods The study was based on a 7-years’ localized fertilization experiment in the purple soil in the southwest of China. The treatments included non-fertilization (CK), optimal fertilization (OP), organic N replace 50% chemical N (MF) and organic N replace 100% N (OM). The crop yields were investigated, and the soil physic-chemical properties were measured. The methods of Nemoro index, correlation coefficient and factor analysis were respectively used to calculate values of the integrated soil fertility index (IFI) under each treatment, the accuracy of the three methods were compared.
Results The average yields of vegetables and corn of the MF were the highest among the fertilization treatments. In the vegetable season, the MF treatment significantly increased the yield. Compared with the OP and OM treatments, the mustard yields were increased by 27.8% and 16.0%, respectively, and the Chinese cabbage yields were increased by 28.6% and 8.7%, respectively. In the maize season, the maximum yield of the MF treatment was 9907 kg/hm2, and increased by 1199 kg/hm2 and 1273 kg/hm2 compared with the OP and OM treatments, respectively. By analyzing the correlations between individual soil fertility index and crop yields, there were significant or extremely significant positive correlations between soil organic matter, total nitrogen, available phosphorus and crop yields, indicating that soil organic matter, total nitrogen and available phosphorus contributed significantly to soil fertility in purple soil and were important indicators to characterize soil fertility in purple soil. The IFI was in the order of OM > MF > OP > CK under three evaluated ways, and the IFI value of the application organic manure was significantly higher than that of the OP treatment. The IFI values of the three evaluation methods showed significant or extremely significant relationship with crop yields, but the mean of the correlation coefficients of the three evaluation methods were in the order of the correlation coefficient method (0.5292) > the Nemoro index method (0.5252) > the factor analysis method (0.5130), indicating that the correlation coefficient method was more reliable than the Nemoro index method and the factor analysis, and was preferred in practice.
Conclusions In vegetable and maize rotation system, the replace part of chemical nitrogen with organic nitrogen can increase crop yields and improve the integrated soil fertility level. From the accuracy of the three evaluation methods, the correlation coefficient method is more reliable than the Nemoro index and factor analysis method, and is preferred in practice.