Analysis of potato canopy coverage as assessed through digital imagery by nonlinear mixed effects models

Abstract

The characterization of the dynamics of canopy coverage represents a relevant matter of study in the field of crop physiology. The objective of this work was to calibrate a model able to simulate potato canopy coverage as a function of thermal time, but including the error structure in such model. This was accomplished by using a mixed effects modelling approach where random effects were added to the average response model. By applying this modelling approach, the structure of the data was taken into account. Calibration data for the model were obtained from canopy coverage estimates derived from image processing analysis. Digital images were taken periodically within 11 potato fields located in the Mantaro Valley (Peru) during the 2005–2006 growing season. This model gave a better fit in comparison with the traditional fixed parameters model. An additional uncertainty analysis with the objective of estimating the confidence region for the predictions of the mixed effects model was carried out. By exploring the data structure, a more comprehensive overview of the potato canopy coverage was achieved with the mixed effects model.