An open-access dataset of global experimental yields in organic, diversified and conventional agricultural systems.

The productive performance of organic and diversified agricultural systems remains the subject of intense debate. Despite a large body of empirical studies and several meta-analyses, an up-to-date, centralized dataset assembling medium- to long-term experimental evidence with annual yield time series is still lacking. Here, we present an open-access global database of published field experiments comparing yields under organic or diversified management with conventional reference systems. The database compiles 102 peer-reviewed articles based on 73 experimental sites, representing 322 experimental units (i.e., unique yield comparisons). It provides standardized yield time series together with harmonized metadata describing crop species, geographic locations, experimental design, and key management practices (e.g., irrigation, fertilization, tillage, and diversification practices). This dataset is designed to support reproducible syntheses of yield levels and yield variability across farming systems and crop types and to facilitate the integration of future long-term experimental results. The dataset is intended to support meta-analyses, model calibration, and scenario analyses of yield levels and yield stability across agricultural management systems.