Study On Establishment And Application Of The Fertilizer Effect Function Based On Field Experiment Of "3414"

According to the whole region "3414"exeperment, using the common model of the fertilizer effect function, the second success rate of the fertilizer effect function were statistically analyzed and studied on different models of selection of the fertilizer effect function in the main crop; research on the effects of coefficient and fertilizer rate in the building process of the fertilizer effect functions after adding blank output (N0P0K0).As well as studies on methods to establish of the regional fertilizer effect function .The results showed that: In the P<0.05 function in the secondary effect, one- factor model highest success rate ,in the P<0.05 and obeying the reward descending rule, soybean, wheat, tomatoes have the same change laws,namely:one-factor model>two-factor model> three-factor model, basically ,corn,millet is one-factor model> three-factor model> two- factor model. The optimal model of the fertilizer effect is the square rool for corn,millet and tomatoes. For wheat, the best models are linear plus plateau for nitrogen fertilizer and phosphorus fertilizer,potassium fertilizer is the square rool.For soybean ,the best model is linear plus plateau for nitrogen fertilizer, but the optimal model is one-factor for phosphorus fertilizer and potassium fertilizer. The nature of the original function has not changed after adding a blank yield, they have the same change Laws: the constant term b0 (the basic yield) have decreased, coefficient b1 and quadratic coefficient b2 have increased. When the same of output, fertilizer rate is decreased for the model of adding blank output. Calculation of regression coefficient, economic optimum fertilization and the highest quantities of fertilizers application are almost same using average method of regression coefficient and original data, by unlike the decimal point median only choice is different. By contrast, the regression coefficient is better than the original data ,because it can be used to reject abnormal data,the average method of original data were not the case.