<p dir="ltr">This dataset contains the spatial distribution of total CH₄ emissions from global rice paddies during 2001–2020. The data was generated using machine learning methods and supports the analysis presented in the manuscript "Rising Greenhouse Gas Emissions from Global Rice Paddies and Mitigation Strategies." The dataset is at a global scale with a spatial resolution of 0.5° × 0.5°, providing detailed insights into methane emissions trends and their geographical variability. Unit is Tg CH4 yr-1.</p><p dir="ltr">This Excel file contains experimental data from irrigation and non-continuous irrigation (AWD and related practices) field trials used for meta-analysis. The dataset includes site-level information (site name, latitude, longitude, experimental years), treatment and control values for CH₄ and N₂O emissions, and associated management practices (irrigation regime, fertilizer input, residue management, biochar application, and tillage). These data were used to support meta-analysis and model evaluation of irrigation effects on soil GHG emissions in rice systems.</p><p dir="ltr">These two code files provide workflows for applying the trained Random Forest (RF) models to generate regional predictions of soil GHG emissions (CH<sub>4</sub>, N<sub>2</sub>O) and crop yield under varying environmental and management conditions. The codes use the previously trained RF models to process input datasets and produce spatially explicit simulation outputs for regional analysis.</p><p dir="ltr"><br></p><p dir="ltr"><br></p>