Critical analysis of life cycle inventory datasets for organic crop production systems

Abstract

Abstract Purpose Organic agriculture (OA) has gained widespread popularity due to its view as a more sustainable method of farming. Yet OA and conventional agriculture (CA) can be found to have similar or varying environmental performance using tools such as life cycle assessment (LCA). However, the current state of LCA does not accurately reflect the effects of OA; thus the aim of the present study was to identify gaps in the inventory stage and suggest improvements. Methods This article presents for the first time a critical analysis of the life cycle inventory (LCI) of state-of-the-art organic crop LCIs from current and recommended LCA databases ecoinvent and AGRIBALYSE®. The effects of these limitations on LCA results were analyzed and detailed ways to improve upon them were proposed. Results and discussion Through this analysis, unrepresentative plant protection product (PPP) manufacturing and organic fertilizer treatment inventories were found to be the main limitations in background processes, due to either the lack of available usage statistics, exclusion from the study, or use of unrepresentative proxies. Many organic crop LCIs used synthetic pesticide or mineral fertilizer proxies, which may indirectly contain OA prohibited chemicals. The effect of using these proxies can contribute between 4–78% to resource and energy-related impact categories. In a foreground analysis, the fertilizer and PPP emission models utilized by ecoinvent and AGRIBALYSE® were not well adapted to organic-authorized inputs and used simplified modeling assumptions. These critical aspects can be transferred to respective LCAs that use this data, potentially yielding unrepresentative results for relevant categories. To improve accuracy and to contribute novel data to the scientific community, new manufacturing LCIs were created for a few of the missing PPPs, as well as recommendations for fertilizer treatment LCIs and more precise emission models for PPPs and fertilizers. Conclusions The findings in the present article add much needed transparency regarding the limitations of available OA LCIs, offers guidance on how to make OA LCIs more representative, allow for more accurate comparisons between conventional and OA, and help practitioners to better adapt LCA methodology to OA systems.