Author(s): Coloma, Eduardo Collado
Date: 2016
Persistent ID: http://hdl.handle.net/10400.5/12939
Origin: Repositório da UTL
Subject(s): total height distribution; Johnson's function; skewness; kurtosis; probability density function
Author(s): Coloma, Eduardo Collado
Date: 2016
Persistent ID: http://hdl.handle.net/10400.5/12939
Origin: Repositório da UTL
Subject(s): total height distribution; Johnson's function; skewness; kurtosis; probability density function
Mestrado Mediterranean Forestry and Natural Resources Management (MEDfOR) - Instituto Superior de Agronomia - UL
Cork oak (Quercus suber L.) represents a crucial role on montado ecosystem. This work will contribute for the improvement of the SUBER model, a growth and yield model developed for cork oak stands in Portugal. In order for the SUBER model to allow the simulation of new stands, it requires the simulation of the model tree state variables. During the first years of tree growth the state variable used by the model is the tree total height. The purpose of this thesis is the modelling of the total heights distributions of young cork oaks plantations. For that, it uses tree measurements taken in 42 plots distributed throughout Portugal with ages between 6 and 22 years old. Following partial objectives were fulfilled: 1) selecting the best probability density function (pdf) to simulate total height distribution; 2) modelling the parameters recovery of the pdf selected, using stand and edaphoclimatic variables as input; 3) validation of the proposed framework, to estimate the total height distributions using the Kolmogorov Smirnov statistic. The results show: 1) the Johnson’s distribution was the best pdf; 2) the mean height, the standard deviation, the skewness and kurtosis were the moments and function of moments used in the parameter recovery; 3) the use of stand and edaphoclimatic variables in the proposed models allows to predict total mean height, standard deviation, skewness and kurtosis with model efficiency values of 0.91, 0.61, 0.35 and 0.21, respectively; 4) 66% of the simulated height distributions followed the real distribution according to the Kolmogorov – Smirnov test. The general behavior of the simulated Johnson’s distribution is acceptable to predict the total height distributions of new stands in SUBER model.