Optimization Fertilizer Application of Filling Type Flue-cured Tobacco Based on Partial Least-squares Regression

This study was conducted to provide a quantitative fertilization model for producing high quality filling flue-cured tobacco leaves and reducing environmental pollution and the quality decline of tobacco leaf problem resulted from over-fertilization. A mathematic model for the relations of flue-cured tobacco production and principal chemical components was established by the Partial Least-squares Regression (PLSR) fertilization function model through massive field experiments. The results showed that there was significant regression relationship between N, P, K fertilizers and flue-cured tobacco production and chemical components. first order term modulus showed that the primary and secondary relationship between N, P, K fertilizers affected flue-cured tobacco; reciprocation term and quadratic term modulus showed N, P, K fertilizers reasonable combination had a critical point, when combination to achieve reasonable N, P, K fertilizers existed the synergistic effects, The antagonistic effects were shown in the higher range. To obtain the max economic returns, the optimum fertilization was N:P:K=1:1.6:2.1. To improve flue-cured tobacco aroma, inflammability and safety, the optimum fertilization was N:P:K=1:1.3:2.9.