Soil seedbank: Old methods for new challenges in agroecology?

The weed soil seedbank is of interest in agroecosystems as a major source of weed infestation in fields and as a reservoir of plant and seed-feeder diversity. A seedbank is a characteristic of annual plants and has been the focus of numerous studies, as it reflects the past aboveground vegetation and is the reservoir of the future vegetation. Therefore, it potentially can be used to evaluate the past, present and future annual weed communities. The goal of this paper was to provide guidelines to help researchers to do a weed seedbank survey. Through a qualitative review of 60 weed seedbank articles, we investigate why and how the seedbank has been studied in agronomy. It shows that seedbank studies have been performed to address the following four major objectives: (a) the assessment of weed management practices on weed communities; (b) the relationship between seedbank and aboveground vegetation; (c) the study of composition and diversity of seedbank in a given area; and, (d) the quantification of seedbank as a food resource for wildlife. Because the analysis highlighted a wide range of methodologies to estimate the seedbank, we critically reviewed them. We show that the selected methodology strongly affects the seedbank estimate. Nevertheless, in our sample of research articles, the analysis revealed that the choice of the methodology was not always justified in terms of achieving a particular scientific goal, but was often determined by the resources available for the experiment (e.g., workload). While studying the soil seedbank remains of interest for scientists (proved by the amount of recent publications), it is time consuming and requires considerable botanical skill. Innovative methods of estimation are scarce and novel methodological developments are needed to increase the quality and reliability of the data obtained.