A standardised and reusable dataset of soil data to be used in Monitoring Reporting and Verification systems for carbon credits certification, in Excel and Soilwise Geopackage format and described by standard vocabularies: A Case Study in the Po Plain (Northern Italy)

The dataset presented here originates from field sampling campaigns (2023-2025) carried out within the Horizon MRV4SOC project in the surroundings of the CREA-FL experimental farm (Mezzi), in the municipality of Casale Monferrato (Piemonte, Northern Italy). Set across an area defined by two primary soil types and their corresponding physiographic classes (active floodplains and fluvial terraces) the dataset is designed to support the development, calibration, and testing of standardized Monitoring, Reporting, and Verification (MRV) frameworks. The dataset integrates depth-stratified laboratory-analysed physical and chemical soil properties. Samples were collected across fields with diverse land-use histories (>15 years), including conventional arables, poplar plantations, agroforestry systems (i.e., poplar intercropped with arable crops), set-aside, and unmanaged patches with spontaneous woody vegetation. Sampling followed a rigorous design: for each field, three geo-referenced monitoring points (A, B, C) were established to capture local environmental variability of the different land-uses within every sampled field. The samples were taken at three fixed depth intervals (0–30 cm, 30–60 cm, and 60–90 cm). The samples taken at the 0–30 and 30–60 cm were analysed in laboratory for the following soil properties: soil organic carbon content (dag/kg), bulk density (g/cm3), coarse fragments content (daL/m3, > 2 mm diameter), soil fine earth textural classes (USDA standard for thresholds between classes) content (dag/kg). Spectral measurements and preprocessing: Air-dried soil samples sieved to 2 mm were scanned using an ASD FieldSpec 3 spectroradiometer (Malvern Panalytical). Spectra were pre-processed by splice correction, conversion to absorbance, Savitzky–Golay smoothing (11-point window, second-order polynomial), removal of low signal-to-noise regions (<500 nm and 2450–2500 nm), resampling at 10 nm intervals, and standard normal variate (SNV) transformation. Spectral alignment using the Lucky Bay sand reference, following the International Soil Standard (ISS) protocol, was performed according to Ben-Dor et al. (2015). Soil properties at 0–30 and 30–60 cm were measured, whereas those at 60–90 cm were estimated using Cubist machine-learning models based on spectral data. The following datasets are included: a dataset in Excel format which includes data itself and metadata sheets. In the metadata sheets a detailed description is given for each term used and for each soil property, including unit of measure and the procedures adopted. The terms and the codes used are chosen from controlled vocabularies, with permanent URI (Uniform Resource Identifier) taken from GLOSIS (Global Soil Information System), AGROVOC and SOILVOC. In case the analytical procedure was not available in standard vocabularies the DOI of the reference scientific publication is given. a dataset in Soilwise Geopackage format (Lachi et al., 2026), which is the relational (SQLite based) container adopted by the SoilWise project (Horizon Europe – Mission Soil) to enable exchange, storage, and GIS native use of soil data, with the explicit goal of making them FAIR and reusable across European policies, research, and land management workflows. To correctly open the GeoPackage and browse it through the custom interface developed for the project, it is recommended to follow the instructions provided in the "Open Geopackage" guide: https://soilwise-he.github.io/Geopackage-so/qgis/open_geopackage.html The complete SoilWise GeoPackage documentation is also available at: https://soilwise-he.github.io/Geopackage-so/ The documentation is organized into several sections describing: the use of the GeoPackage in QGIS through the project's customized interface: https://soilwise-he.github.io/Geopackage-so/qgis/ the use of QField for data collection during field survey campaigns: https://soilwise-he.github.io/Geopackage-so/qfield/ the detailed description of the database structure, including tables, fields, relationships, and layers that compose the data model: https://soilwise-he.github.io/Geopackage-so/tables/ The documentation also includes a search function that allows users to quickly locate tables, fields, layers, or specific topics without having to manually browse through the different sections. https://soilwise-he.github.io/Geopackage-so/qgis/open_geopackage.html