SCARFACE: a harmonized spatio-temporal dataset integrating socio-economic, environmental, and agricultural indicators for the Po Valley (Italy), 2011--2024

The SCARFACE research initiative SCARFACE (Sequestering CARbon through Forests, AgriCulture, and land usE - https://www.paolomaranzano.net/scarface) is a research initiative funded by the University of Milano-Bicocca (UniMiB), Italy. The project blends complementary and interdisciplinary research experiences from the statistical, data science and environmental and atmospheric chemistry backgrounds at UniMiB. Along with researchers from UniMiB, the project involves researchers from the Council for Agricultural Research and Analysis of Agricultural Economics - Agricultural Policies and Bioeconomy (CREA-PB), Italy, and the School of Mathematics and Statistics of the University of Glasgow (Scotland, UK). The SCARFACE dataset The project assembles a harmonized spatio-temporal dataset that integrates several domain, such as climate, air quality, pollution emissions, land cover, soil properties, agro-industry dynamics and socio-economic indicators, to jointly investigate interconnected processes linking agricultural systems, atmospheric dynamics, emissions, and socioeconomic conditions in the Po Valley (Northern Italy), an area characterized by strong interactions among agricultural systems, environmental processes, and human activities. The spatial reference unit adopted in SCARFACE is the Agrarian Sub-Region (ASR), a territorial classification defined by the Italian National Statistics Office (ISTAT). ASRs represent groups of contiguous municipalities that are considered relatively homogeneous with respect to natural conditions, agronomic characteristics, and agricultural production systems. The Po Valley can be partitioned into m=256 ASRs with different sizes and shapes. The SCARFACE dataset integrates information for the period from 2011 to 2024 (i.e., T=14 time stamps), with the initial and final temporal coverage depending on the availability of the individual data sources. Therefore, the final database adopts an annual panel structure defined over ASR spatial units and composed of a total number of spatio-temporal observations equal to N=mxT=256x14=3584 for each variable. Overall, SCARFACE comprises a set of p=2748 variables (plus three unique identifiers, that is, year, ASR and geometry) that include administrative records, gridded environmental products, satellite-derived land information, and survey-based socio-economic indicators sourced from national and international public institutions, covering a wide range of thematic domains. Farm activity and agro-economic indicators are derived from the Farm Accountancy Data Network (FADN) survey coordinated by the Italian Council for Agricultural Research and Agricultural Economics Analysis (CREA). Emissions data are obtained from the EDGAR inventories developed by the European Commission, while air quality information is sourced from both the European Environment Agency (EEA) and the Copernicus Atmosphere Monitoring Service (CAMS). Meteorological variables are retrieved from the ERA5-Land reanalysis produced by the European Centre for Medium-Range Weather Forecasts (ECMWF), and extreme weather indicators are provided by the European Drought Observatory (EDO). Land cover information is based on the CORINE Land Cover dataset from Copernicus and the Global Dynamic Land Cover (GDLC) dataset. Livestock data are obtained from the Italian National Livestock Registry (BDN) managed by the Italian Ministry of Health, while socio-economic indicators are produced by ISTAT. Finally, geographical features and administrative metadata are derived from a combination of Amazon Web Service (AWS), ISTAT and Eurostat. The dataset is designed for broad reuse across a range of analytical settings, including panel data analysis at moderate spatial resolution, spatio-temporal modeling with heterogeneous covariates, spatial clustering and the identification of regional typologies, and policy-oriented applications related to agricultural transitions, air quality management, and climate variability. The building process of the dataset is detailed in the companion paper.