Modelling early growth under different sowing conditions: A tool to predict variations in intercrop early stages

Early growth is a critical phase of the crop cycle, which lasts from emergence to the beginning of competition between plants and is sensitive to sowing conditions and species characteristics. Providing tools to improve the management of this critical phase in intercrops is a challenge for agroecology as these cropping systems are the subject of renewed interest for their ability to maintain yields while requiring fewer inputs. The aim of the present study was to investigate variations in early growth under different sowing conditions in different species with contrasted seed and seedling characteristics (seed mass, hypogeal or epigeal emergence, and legume and non-legume species), especially species used as intercrops. Experiments were carried out in glasshouses using different sowing depths and levels of mineral nutrition, first for each species separately, then in mixed sowings. In the first set of experiments, biomass at emergence and relative growth rate after emergence were measured and then modelled as a function of seed mass, mineral nutrition, and time to emergence. Predictive equations were tested by comparing simulations with biomass measured in the second set of experiments, for two intercrops grown under varied sowing conditions. Finally, simulations were run to analyse variations in the early growth of two intercrops (durum wheat/pea or alfalfa) under a wider range of sowing conditions (seed mass, sowing depth, and with or without mineral nutrition). Biomass at emergence was positively correlated with seed mass, and in epigeal species, was also negatively impacted by time to emergence. Relative growth rate was highly stimulated by mineral nutrition whereas its response to time to emergence varied among species. The amount of seminal reserves at emergence (in hypogeal species) and the cotyledon specific mass (in epigeal species) were correlated with the establishment of the relative growth rate. When evaluated, the model was shown to satisfactorily predict the early growth of two intercrops. Used as a simulation tool, the model indicated that all the sowing techniques tested can have a major influence on total biomass and on the proportion of each component species when competition begins. This model can thus contribute to the management of sowing techniques for sole as well as combined crops whose effects are difficult to predict and are also difficult to analyse from field experiments alone because of the number of possible combinations and interactions. (C) 2013 Elsevier B.V. All rights reserved.