Securing China's Rice Harvest: Unveiling Dominant Factors in Production Using Multi- Source Data and Hybrid Machine Learning Models

Abstract

Abstract Ensuring the security of China's rice harvest is imperative for sustainable food production. This study addresses this critical need by employing a comprehensive approach that integrates multi-source data, including climate, remote sensing, soil properties and statistical information. The research evaluates various single and hybrid machine learning models to predict rice production across China, particularly focusing on the main rice cultivation areas. The investigation identifies the hybrid models have performed better than single models, the best scenario was recorded in scenarios 8 (soil variables + SA) and 11 (All variables) based RF-XGB by decreasing the RMSE by 38% and 31% respectively in comparison with the single model (RF). Moreover, the soil properties contribute as the predominant factors influencing rice production, exerting an 87% and 53% impact in east and southeast China, respectively. Additionally, the study explores the implications of temperature and precipitation changes on rice production. Notably, it observes a yearly increase of 0.16°C and 0.19°C in maximum and minimum temperatures, coupled with a 20 mm/year decrease in precipitation. These climatic shifts contribute to a 2.2% annual reduction in rice production on average in southeast China. This research provides valuable insights into the dynamic interplay of environmental factors affecting China's rice yield, informing strategic measures to enhance food security in the face of evolving climatic conditions.