Mapping life cycle assessment (LCA) scientific research in agriculture: what do we still have to do?

Scientific literature documents and dictates the direction of knowledge and to recognize it we rely on techniques that compile information to point out the next paths in science. One of the promising techniques is bibliometric analysis, especially when applied to the recent and under developing approach, such as Life Cycle Assessment (LCA). Considering that the methodology still has limitations and can generate simplified results, we sought to map its characteristics in agricultural use, characterizing the evolution and trends of scientific research when using LCA in the agricultural sector over a period of twenty years, from 2002 to 2022. We used bibliometric analysis for a systematic review of the literature, using the Scopus repository as a database. To organize and process the data, we used the VOSviewer software. It was found 4943 articles, of which the publication peaks occurred in the last ten years, published predominantly by journals related to environmental sciences. This is due to the study characteristics being mainly focused on methodology improving. They have been produced mainly in the USA, Europe, and China, without emphasis on regions, such as Latin America, Africa, and Asia, although they largely concentrate global agricultural production and with relevant sustainability challenges. Through bibliometric analysis, three main clusters were formed: Impact Assessment (which deals with methodological development); Types of products (about raw materials for alternative energy production); and Substances and activities (particularly substances related to greenhouse gas emissions and activities related to animal and feed farming). We note that the methodology still neglects certain types of impacts, such as biodiversity loss, soil degradation, and partly the (eco) toxicity of pesticides. The methodology is still insufficient for decision-making, with a limited understanding of agricultural functions and difficulties in modeling certain persistent effects, although it continues to be strategically improved. Graphical abstract