PJB-2018-523
PREDICTION MODEL OF CH4 EMISSIONS AT THE GROWTH STAGEOF RICE FIELDS IN COLD REGION BASED ON BP NEURAL NETWORK
yu lihong
Abstract
Abstract: In view of the methane emission from rice vulnerable to various factors such as soil properties, temperature, fertilization and water management, a prediction method for methane emissions at growth stage of rice fields in cold region is proposed. Based on the statistical analysis of the field measurement data of CH4 emissions from rice in cold region of Heilongjiang, China, the soil pH value, the redox potential, the soil temperature at 10cm, the N application rate and the depth of the water layer are selected as the input, and the CH4 emissions as the output, a BP neural network prediction model, with five nodes in hidden layer and three levels of network, for CH4 emissions at the growth stage of rice fields in cold region is established with Matlab. The correlation coefficient reaches 0.998, the average error is 7.86%, and the prediction accuracy is 92.14%.
To Cite this article:
Download
