Abstract
Paddy rice, one of the world's primary cereals, plays a critical role in food security and environmental sustainability. Existing global rice distribution datasets provide valuable insights but are often constrained by time and coverage. In this study, we developed the GloRice (I), a long-term global gridded dataset mapping rice harvested and physical areas at a 5-arcminute resolution for 1961-2021. The dataset consists of three components: (1) GloRice-2000hvst: four types of rice harvested area maps for the year 2000, using multiple established datasets (M3, SPAM, GAEZ); (2) GloRice-hvst: annual harvested area maps (1961-2021) integrating national and subnational statistics; and (3) GloRice-phsc: annual physical area maps (1961-2021) derived from the rice multiple cropping index (MCI) data. Validation demonstrated strong agreement at global and regional levels, with consistent accuracy in subnational comparisons for major rice-producing regions, including China, South Asia, and Southeast Asia. The GloRice (I) provides essential data input for regional and global modelling and assessments, with broad applicability in agriculture, climate, and environmental research.