Full Dataset for Geo-ROCKET: Climate-based Estimation of Multi-cropping Rice Planting Dates

Overview This dataset contains processed multi-source agricultural and environmental data for rice yield prediction in the Monsoon Asia region. The data spans the period 2019-2020 and includes climate variables, rice phenology calendar information, crop production statistics, and spatial mask files for the Asian monsoon region. Data Sources and Description Data Type Source Resolution Coverage Content/Variables Data Quality/Methodology Access/References Climate Data (AgERA5) Agrometeorological indicators from 1979 to present derived from reanalysis, based on hourly ECMWF ERA5 data at surface level 0.1° spatial, daily temporal 2019-2020 • Daily mean 2m air temperature (°C) • Daily precipitation flux (mm/day) • Daily surface solar radiation flux (W/m²) Data aggregated to daily time steps at local time zone and corrected towards finer topography using grid and variable-specific regression equations trained on ECMWF's operational high-resolution atmospheric model (HRES) Available from Copernicus Climate Data Store (CDS) Rice Phenology Calendar (MARC) Monsoon Asia Rice Calendar derived from Sentinel-1 and Sentinel-2 imagery 0.5° × 0.5° 2019-2020 Transplant and harvest timing, number of cropping seasons Minimal bias (4 days early for transplant, 6 days late for harvest), reduced errors compared to prevailing phenology charts • Paper: DOI: 10.5194/essd-16-3893-2024 • Data: DOI: 10.17595/20230728.001 Crop Production Data (SPAM 2020) IFPRI's Spatial Production Allocation Model (SPAM) 2020 5-minute grid 2020 Rice physical area data (all technologies together - irrigated and rainfed portions) Cross-entropy method to generate plausible estimations of crop distribution within disaggregated units DOI: 10.7910/DVN/SWPENT, Harvard Dataverse, V4 Spatial Mask Data Administrative boundary data from Resource and Environmental Science Data Platform, Chinese Academy of Sciences 0.5° grid Monsoon Asia region Administrative boundaries for Asian monsoon region Used to define the 0.5° grid and enable grid-based sampling of other datasets Available from Resource and Environmental Science Data Platform Data Processing and Production The data has been processed using the 0.5° grid of the Asian monsoon region (defined by the administrative boundary data mentioned above) to perform grid-based sampling of climate, calendar, and SPAM2020 data. Usage Instructions Data Setup 1. Download the data from this Zenodo repository2. Extract the `data.zip` file3. Place the extracted `data` directory in the root directory of the cloned code repository4. Follow the instructions in the main code repository's README.md to run the code