A dynamic knowledge model on appropriate nutrient indexes in rice

Rice (Oryza sativa L.) is the most important food crop in the world and China. Studying the dynamics of nitrogen, phosphorus, and potassium nutrient accumulation and analyzing nutrient content in above ground plant is very important, which could help farmers to make rational decision on fertilizer application. Based on the analysis and refinement of rice raising rationale and latest technical research, a general knowledge model for nutrient index dynamics in rice was developed through the quantitative description of dynamic relationship of N, P and K accumulation and concentration in the above-ground part of rice to varieties, ecological environment factors and technique levels. The knowledge model was established using Visual C++ as programming language, and encapsulated in the form of automation with the standard of COM (Component Object Model) on operational platform as Windows 2000 in Chinese version, which could be used to predict the time-course dynamics of above-ground plant nutrient contents and accumulation for the purpose of nutrient (diagnosis) and growth regulation under different target yields with different rice varieties, environmental condition and production levels. Case studies were carried out using the daily meteorological data and research results in normal climatic year of Nanjing and Changde in China. Results showed that RMSE of nutrient accumulation and concentration trend were 8.97 kg/hm2 and 0.32%, respectively, indicating that the time-course dynamics of nutrient index designed by the knowledge model for two eco-sites and different rice cultivars were overall consistent with the practical growth patterns (under) specific production systems. Therefore, the quantitative knowledge model would overcome the weakness of traditional crop expert system with poor spatial and temporal adaptation, such as specific site limitation, massive knowledge base and low quantification, and lay a foundation for the digital decision support on rice management.