Description of weed germination curves at different temperatures using nonlinear models

The spread of weeds can occur in various ways, both through seeds that can detach from plants already present in the area and be carried by machinery, wind, animals, and humans. The use of statistical models can help in a better understanding of the cumulative seed germination process and aid in the management system. The objective of this study was to use the non-linear Logistic and Gompertz models, evaluating which one is more suitable to describe the cumulative germination curve of three weed species, as well as exploring the first to fourthorder derivatives to analyze the critical points of the curves. The following conditions were used: photoperiod (8 hours of light/16 hours of darkness) with temperature (alternating or constant), and the percentage germination of seeds was evaluated at 2, 4, 6, 8, 10, 12, and 14 days after sowing. The model parameters were estimated using the method of least squares with the iterative Gauss-Newton method. The quality of fit was assessed using the coefficient of determination, residual standard deviation, and Akaike information criterion, through R software. The Gompertz model was found to be the most suitable for describing the data. Under both germination conditions, the species A. viridis showed the lowest and highest values of beta and k, respectively, followed by A. spinosus and A. deflexus. Regarding critical points, A. viridis showed the earliest germination, but A. spinosus had the highest germination percentage.