Estudo do crescimento de plantas daninhas do gênero Amaranthus por modelos não lineares ; Study in growth of weeds of the Amaranthus genus by nonlinear models

The concept of weeds emerged with the beginning of agriculture, as it was from that moment onwards that human beings began to select plants. In the world there are several known species of weeds, such as those of crucial importance of the genus Amaranthus that are present throughout Brazil and in several areas. They have high seed production, easy dissemination, and rapid growth. Besides, they can promote competition for water, light and nutrients, reducing the yield of the crops in which they compete. The study of how this type of plants grows makes it possible to analyze its interference in the environment, and in this way, they can contribute to the development of management systems. The objective of this work was to study the growth and development of five weed species of the genus Amaranthus, under non-competitive conditions, based on the accumulation of total dry mass, roots and reproductive structures, with data obtained by the transversal method, using the non-linear Logistic, Gompertz and von Bertalanffy models, the first to fourth order derivatives were also explored, analyzing the critical points of the growth curves. In addition, it was evaluated which is the most suitable model to analyze the growth of the species. Data were extracted from Carvalho, López-Ovejero and Christoffoleti (2008); the experiment was carried out in a greenhouse at the Department of Plant Production of the Higher School of Agriculture “Luiz de Queiroz” – ESALQ/USP, between September and December 2005, and consisted of eight evaluations of growth, spaced 10 days apart, except for the first one which was carried out at 20 days after sowing and then every 10 days until the total of 90 days. The non-linear Logistic, Gompertz and von Bertalanffy models were adjusted using the open access software R, considering the AR(1) autoregressive error structure and heteroscedasticity of variances, when necessary. To compare the adjustments, the following evaluators were used: coefficient of determination R², residual standard deviation ...