Abstract

A nonlinear regression model was developed in order to explain the shade generated by individual trees of Cordia alliodora depending on its age, in agroforestry systems with coffee in Lebanon, Tolima, Colombia. Trunk diameter at chest height and crown area were measured, occlusion of tree top was estimated with hemispherical photography from 50 individual trees (1-19 years old). 14 generic nonlinear regression models were studied to develop the model, the hyperbolic one was selected as the best fit due to its lower mean square error and the Akaike and Bayesian criteria. A practical tool was developed for coffee producers, with which they can estimate the density of C. alliodora that provides the desired shade level. This tool was built based on the estimates of individual tree shade from the growth model and the necessary shade for the crop management. The tree top of C. alliodora presented occlusion averages of 0.79 and crown area of 60.4 m².tree^-1. Shade growth is higher in young trees than in adult trees (5.5 vs 3.0 m².treel^-1.year^-1 for trees under 5 years and about 20 years, respectively). Similarly, young C. alliodora trees had an average of 20 cm² of basal area per m² of shade compared to 62.7 cm².m^-2 in the adults. Research results and the tool generated from this study are key for the design and management of the shade by coffee producers since it allows defining the density of trees needed.

Keywords: Shade area; shade increment; design and management; non-lineal regression; occlusion