Analysis of identifying important ecological factors influencing winter wheat protein content based on artifical neural networks

Temperature, rainfall, illumination time and soil nutrients are major ecological factors to influence protein content forming of winter wheat. This study focused on the evaluation of the relative weighting of those factors on winter wheat grain quality (protein) based on the wheat planting, soil and weather data in Beijing, China. artificial neural network (ANN) analysis is employed in this study. The result indicated that the 10 factors have significant impact on the formation of wheat protein. The most important factor is illumination time from 6th June to 10th June, followered by the number of days which the temperature above 32℃, available nitrogen content of soil, average temperature from 1st May to 10th June, average temperature from 26 May to 30 May, accumulated temperature from 20th May to 10th June, average temperature from 1st June to 5th June, range of temperature from 20th May to 10th June, rainfall from 20th May to 10th June, and organic matter in soil respectively. Then, the response curves for key factors are generated by the ANN models in order to reflect the wheat protein variant trend according to the different ecological factors. The results of this study can probably be used for provided the reference basis for the winter wheat quality regionalization of Beijing area