Dataset: The Agrivoltaic Paradox: Why Product Carbon Labelling May Penalise Integrated Food-Energy Systems

This dataset contains the primary and processed data supporting the life cycle assessment (LCA) presented in the manuscript "The Agrivoltaic Paradox: Why Product Carbon Labelling May Penalise Integrated Food-Energy Systems" (under review at Nature Sustainability). The data were collected during the 2023 growing season from two adjacent 1‑hectare experimental vineyards in Apulia, southeastern Italy (40°21′ N, 17°35′ E). The study compares the carbon footprint of wine produced in a conventional organic system versus an agrivoltaic system where grapevines are grown underneath bifacial photovoltaic panels. The dataset includes:Primary inventory data for both the organic and agrivoltaic systems, detailing:Vineyard characteristics (planting density, row spacing, vine age).Annual agricultural inputs (fertilisers, pesticides, biostimulants).Fuel consumption (diesel for tillage, spraying, harvesting, and maintenance).Grape yields and transport distances.Packaging materials (glass bottle weight, cork, PVC capsule).Technical specifications of the agrivoltaic infrastructure (PV panel type, array dimensions, support structures, electrical components).Emission factor database compiled from peer‑reviewed literature and authoritative sources (IPCC, IEA, Ecoinvent, Agri‑footprint, and others), used to calculate greenhouse gas emissions.Calculated carbon footprints per functional unit (0.75 L bottle for organic wine; 0.5 L bottle for agrivoltaic wine) and per hectare, including detailed breakdowns by life cycle stage.Results of alternative allocation methods (economic, land‑area, energy‑content, and system expansion) for the agrivoltaic system.One‑at‑a‑time sensitivity analysis outputs showing the influence of key parameters (prices, PV lifetime, yield, etc.) on the agrivoltaic wine carbon footprint.Monte Carlo simulation results (10,000 iterations), including input distributions and output distributions for both systems, providing 95% confidence intervals.The data are provided in multiple formats (Excel spreadsheets and CSV files) to facilitate reuse. All calculations follow ISO 14040/14044 and ISO 14067 guidelines.Data collection methods: On‑farm inputs were recorded in logbooks by vineyard staff; fuel consumption was measured from refuelling records; yields were weighed at harvest; PV infrastructure data were supplied by the system installer. Emission factors were sourced from the most recent available databases and literature, prioritising Italian‑specific data where possible.Usage notes: The data are organised into separate sheets/tables corresponding to the supplementary information in the related manuscript. Users may reproduce the LCA results by applying the provided emission factors to the inventory data. The accompanying Python code (available in a separate GitHub repository, linked will be available below) implements the full calculation framework, including all allocation methods and uncertainty analysis.Related materials:Manuscript: [The Agrivoltaic Paradox: Why Product Carbon Labelling May Penalise Integrated Food-Energy Systems], under review at Nature Sustainability.Python code: will be available shortly