Predicting individual tree mortality in northern hardwood stands under uneven-aged management in southern Québec, Canada

Abstract

-This study proposes a generalized linear mixed model to predict individual tree mortality in northern hardwood stands under uneven-aged management. The model is based on a complementary log-log (CLL) link function, and was calibrated using permanent-plot data. Tree vigor, stem product, diameter at breast height and stand basal area were tested as explanatory variables. A plot and an interval random effect were specified to account for spatial correlations. When compared with the traditional logit link function, the CLL facilitates the inclusion of the time factor. In this case study, there was an important variability of mortality predictions between the plots and the intervals for a given plot. The interval random effect is thought to be associated with catastrophic mortality. Since both tree vigor and stem product proved to be significant mortality predictors, we recommend that these variables be evaluated to increase the accuracy of mortality models. mortality / generalized linear mixed model / random effects / hardwoods / single-tree selection cutting Résumé -Prévision de la mortalité des tiges individuelles dans les peuplements de feuillus nobles sous aménagement inéquienne dans le sud du Québec, Canada. Cette étude présente un modèle linéaire mixte généralisé pour la prévision de la mortalité dans les peuplements de feuillus nobles sous aménagement. Le modèle utilise une fonction de lien log-log complémentaire (LLC) et a été étalonné à l'aide de données de placettes permanentes. La vigueur de l'arbre, le produit, le diamètre à hauteur de poitrine et la surface terrière ont été testés comme variables explicatives. Des effets aléatoires de placette et d'intervalle ont été spécifiés dans le modèle afin de tenir compte des corrélations spatiales. Comparée au traditionnel logit, la fonction de lien LLC facilite l'inclusion du facteur temps. Dans ce cas d'étude, il existait une importante variabilité des prévisions entre les placettes et les intervalles de temps d'une même placette. On présume que l'effet aléatoire d'intervalle représente la mortalité catastrophique. Puisque la vigueur et le produit des tiges se sont avérés être des variables significatives, il est recommandé de les évaluer afin d'améliorer la précision des modèles de mortalité. mortalité / modèle linéaire mixte généralisé / effets aléatoires / feuillus nobles / jardinage par pied d'arbre

Mortality of silver fir and Norway Spruce in the Western Alps — a semi-parametric approach combining size-dependent and growth-dependent mortality

“…Other authors have developed parametric regressions (such as logistic regression) using permanent plot data which included both growth (either directly or indirectly through competition indexes) and size to obtain synthetic mortality models including regular and irregular mortality (Eid and Tuhus, 2001;Fortin et al, 2008;Moore et al, 2004;Uriarte et al, 2004;Wunder et al, 2007;Yao et al, 2001). Nevertheless, it has been demonstrated that due to unbalanced data-sets from permanent plots and to the highly skewed shape of mortality, parametric models assuming a strict model shape may lead to biased mortality estimates and wrong interpretations concerning species 305p7 Ann.…”

“…When permanent plot data are available, competition indexes (or growth) and size are often combined in a parametric regression, such as the logistic regression, to determine mortality estimates (Eid and Tuhus, 2001;Fortin et al, 2008;Moore et al, 2004;Uriarte et al, 2004;Wunder et al, 2007;Yao et al, 2001). Parametric functions have two disadvantages when trying to calibrate mortality models.…”

Contacts and foot and mouth disease transmission from wild to domestic bovines in Africa

“…It was necessary to account in statistical models for this variation in the length of the time interval over which serological transitions were observed. To do so, a complementary log-log link function was used and the log of the length of the time interval was included as an offset in all the models presented below (Fortin et al 2008). Goodness of fit was assessed through Pearson overdispersion test (Bolker et al 2009).…”

Étude comparative des taux de mortalité entre les arbres dominants et les arbres indicateurs moyens de la station

“…Indeed, trees in the same plot are likely to be correlated, trees observed in the same period for a plot and observations made on the same tree in a plot-method as well. A cloglog link was used and an offset variable corresponding to the log of time interval between two measurements was included in the model and its coefficient was set to 1 (Fortin et al, 2008;Rose et al, 2006). The general model considered for the five species of interest was:…”