Estimating mortality in multi-species forests in Durango, Mexico

Mortality models are important elements for predicting future forest conditions. The objective of the present study was the estimation of survival probability at tree level using the logistic function. The model was constructed using representative data from two inventories of a trial network involving 58 permanent plots established by the Unidad de Conservacion y Desarrollo Forestal de El Salto, Mexico (Tab. I, Fig. 1). The results show that the classic logistic function approach proved to be helpful for estimating mortality in multi-species and uneven-aged stands in this forest region. For both pine and oak species groups, the competition index C66 and the quotient of the diameter at breast height (bhd(-1)) showed a significantly greater contribution to the model than the BAL- and the iterative Hegyi-indices (Tab. 2). The tree number decline is inversely proportional to the diameter at breast height and directly proportional to the competition index C66 (Figs. 2, 3 and 4). The logit mortality model confirmed that for the studied tree species groups there were no significant differences among the estimated and observed tree mortalities (Tab. 3, Figs. 2 and 3). A negative relationship between the P-LOGIT value and the probability of tree mortality estimated by the three-parameter logistic model was found (Tab. 4, Tab. 5, Fig. 5).