Geostatistical Dataset for Topsoil Organic Carbon Modelling in India Using CMIP6 Climate Projections and LUH2 Land Use Data

This dataset provides a geostatistical feature space in ESRI Shapefile format representing present and future geo-environmental variables that regulate the spatio-temporal variability of Soil Organic Carbon (SOC) stock across India. The dataset is spatially constrained to the national territory of India and is structured to support modelling, simulation, and exploratory analysis of SOC dynamics under both current and projected environmental conditions. The feature space is organized into two main components: static environmental features and dynamic forcings. Static features represent relatively stable geo-environmental characteristics influencing SOC formation and redistribution, including soil properties, geology, terrain attributes, and spatial location factors consistent with the SCORPAN conceptual framework (soil, climate, organisms, relief, parent material, age, and spatial position). The dataset also includes a dynamic feature space composed of climate and land-use forcings that capture temporal variability affecting SOC processes. Climate variables were derived from the NASA Earth Exchange Global Daily Downscaled Projections (NEX-GDDP) dataset, representing statistically downscaled CMIP6 climate simulations for historical (1979–2014) and future periods (2015–2100) under multiple Shared Socioeconomic Pathway (SSP) scenarios: SSP126, SSP245, SSP370, and SSP585. Climate data from multiple CMIP6 models were aggregated into monthly variables including precipitation totals and mean air temperature, with units standardized and time series re-indexed to a continuous temporal axis. To reduce regional biases typical of downscaled climate products, an India-specific bias-correction procedure based on Quantile Delta Mapping (QDM) was applied using observational climate references from the India Meteorological Department (IMD). Bias correction was performed at grid-cell level across the entire country and subsequently combined into a multi-model ensemble (MME) to represent robust climate forcing conditions. Land-use and land-management variables were obtained from the Land Use Harmonization v2 (LUH2) database associated with CMIP6/LUMIP experiments, providing annually resolved historical and scenario-based information on land states, irrigation practices, and fertilizer management. Fractional land-use variables were converted to physically interpretable area-based units to ensure consistent comparison across historical and projected periods. All climate and land-use variables were organized into annually indexed shapefile layers aligned to the 0.25° × 0.25° NEX-GDDP grid, enabling seamless integration with SOC modelling frameworks. In addition to geospatial feature space for SOC modelling, the dataset includes a ready-to-process modelled SOC stock time-series of India, that can be used for wavelet decomposition and long-term trend analysis. This regionally harmonized dataset as such provides a comprehensive dataset for studying the interactions between climate change, land-use transitions, and SOC dynamics across India.