Epsilon Archive for Student Projects

Abstract

The research objective is to construct a model for crown radius (CR) based on DBH and significant quantifiers for wild pear (Pyrus pyraster), using regression analyses, to develop guidelines for silvicultural practices for wild pear. To establish the model empirical data of a wild pear population in Western Germany was collected. Site measurements include prevailing vegetation, soil characteristics, location and neighbouring trees. Tree measurements include DBH, height, crown radius and quality characteristics: living branches, forking, bending and spiral grain. 52 trees have been measured of which 41 have been used for modelling purposes. Selection criteria for the model are residual plots and coefficient of determination (R²). In order to find a linear relationship, (semi-) logarithmic transformations of the dependent and independent variables have taken place. A backward variable elimination from a full model was used to find significant predictors. The logarithmic model is selected best as residual plots show a homoscedastic distribution and follow a normal distribution around the zero-line. The R² of 0.6101 gives an indication of a significant relationship. The final relation can be formulated as:CR=e^(0.61388∗ln(DBH)−0.75029), where CR in m and DBH in cm. No other independent variables than DBH have been significant. From the derived model, tree spacing and number of stems per ha can be calculated at any DBH. Based on collected data and the model silvicultural guidelines for a pure and two mixed wild pear stands including planting density and layout, pruning, thinning and regeneration cutting are established.