Topsoil Organic Carbon Stocks and Uncertainty in Florida Grazing Lands Derived from Quantile Regression Forest (30 m Resolution)

This dataset provides spatially explicit predictions of topsoil soil organic carbon (SOC) stocks and associated prediction uncertainty for grazing lands across Florida, USA. The data were generated using an uncertainty-aware digital soil mapping framework based on Quantile Regression Forest (QRF) modeling and the integration of contemporary and legacy soil observations. Version 2.0 update: This version contains revised SOC prediction and uncertainty products generated following revisions to the associated manuscript during peer review. Version 2.0 corresponds to the analysis and results presented in the manuscript accepted for publication in Geoderma and supersedes version 1.0 for use with the final peer-reviewed study. Version 1.0 remains available through the Zenodo version history to preserve the provenance and reproducibility of the earlier preprint-associated release. Dataset contents The dataset consists of five raster layers in GeoTIFF format at 30 m spatial resolution: Grazing land mask (binary): Spatial extent of Florida grazing lands used for modeling (1 = grazing land; 0 = non-grazing land). Mean SOC stock map (t ha⁻¹): Predicted mean topsoil SOC stock. 5th percentile map (Q5; t ha⁻¹): Lower bound of the 90% prediction interval. 95th percentile map (Q95; t ha⁻¹): Upper bound of the 90% prediction interval. 90% prediction interval width (t ha⁻¹): Difference between Q95 and Q5, representing the spatial magnitude of predictive uncertainty. Together, these layers support both deterministic and probabilistic interpretation of SOC spatial variability across Florida grazing lands. The grazing land mask defines the spatial domain of the study and can be used to subset the SOC prediction and uncertainty layers. Data characteristics Spatial resolution: 30 m Spatial extent: Grazing lands across Florida, USA Coordinate reference system: EPSG:4326 File format: GeoTIFF (.tif) Units: t ha⁻¹ for SOC stock predictions, percentile bounds, and prediction interval width Important notes Version 2.0 is the recommended version for use with the final peer-reviewed study accepted for publication in Geoderma. The SOC layers represent model-based estimates rather than direct measurements and should be interpreted accordingly. Prediction uncertainty is derived from the QRF predictive distribution and is represented by the 5th and 95th percentiles and their corresponding 90% prediction interval. The prediction intervals quantify model-based predictive uncertainty but do not explicitly account for all potential sources of uncertainty, including sampling bias and measurement error. Users are encouraged to consider the mean SOC predictions together with the associated uncertainty bounds when using these data for analysis, carbon accounting, grazing land management, or decision-making. Citation If you use this dataset, please cite both the dataset and the associated peer-reviewed publication: Zhao, C., Song, J., Dubeux, J., Grunwald, S., Bretas, I. L., Liao, H.-Y., Tziolas, N., Harley, J. B., Zare, A., Babaeian, E., Garcia, L., Queiroz, L., & Mendes, C. T. E. (2026). Topsoil Organic Carbon Stocks and Uncertainty in Florida Grazing Lands Derived from Quantile Regression Forest (30 m Resolution) (Version 2.0) [Dataset]. Zenodo. https://doi.org/10.5281/zenodo.22000721 Zhao, C., Song, J., Dubeux, J., Grunwald, S., Bretas, I. L., Liao, H.-Y., Tziolas, N., Harley, J. B., Zare, A., Babaeian, E., Garcia, L., Queiroz, L., & Mendes, C. T. E. (2026). Spatiotemporal controls on soil organic carbon stocks in subtropical grazing lands: An uncertainty-aware digital soil mapping approach. Geoderma. Associated resources Interactive Web GIS application:An interactive web-based GIS application for exploring and visualizing the spatial distribution of topsoil SOC stocks and associated uncertainty is available at:https://es-geoai.rc.ufl.edu/agroes-grazing-soc/ Source code:https://github.com/Ecosystem-Services-GeoAI/florida-grazing-soc-qrf Funding This research was supported by the 2022–2023 Florida State Legislative Budget AI-HARVEST program; Florida Milk Checkoff; Florida Cattle Enhancement Board (P0326003); USDA-NIFA Research Capacity Hatch Funds (FLA-AGR-006393); Florida Agricultural Experiment Station UF/IFAS Archer Early Career Seed Grant (P00133052); and startup funds from the Florida Agricultural Experiment Station, UF/IFAS, University of Florida. The Florida Soil Carbon and Pedon Database (FSCPD) was supported by USDA-CSREES-NRI grant award 2007-35107-18368 (PI: S. Grunwald).