COMPETITION INDICES AND STAND MODELS FOR INDIGENOUS HARDWOOD SPECIES IN AKURE FOREST RESERVE, ONDO STATE, NIGERIA

Abstract

Competition between trees exists when resource availability falls below the sum requirement of stand population for optimal growth. The competitive pressure individual tree undergo during this competition is estimated by calculating its Competition Indices, as this represent how individual tree is affected by its local neighbour. Therefore, this study aims to estimate the competitive stress of individual tree using selected Competition Indices and to also develop models for determining the effectiveness of the selected competition indices in predicting the growth of individual tree in Akure Forest Reserve. Data were collected from the four permanent sample plots (PSP) located in the Strict Nature Reserve within Akure forest reserve, each plot covering an area of 0.25ha. The Dbh, Diameter at the base, Diameter at the middle, Diameter at the top, total height, and distance from the subject tree to neigbouring trees were measured and recorded to estimate tree volumes and the competition indices. The data were sorted according to species to obtain the four selected families for this study. The selected families are Annonaceae, Meliceae, Sterculiaceae, and Ulmaceae. Three Distance-Independent Competition Indices and Seven Distance-Dependent were estimated and involved for model generation. Correlation coefficients between each of the Competition Index and tree growth variables were evaluated for each family. Also, individual trees were treated as subject trees in calculating the competition index for each tree. Descriptive statistics were carried out for each of these selected four families to observe the differences in their growth variables. Volume and competition models were developed according to species and families. Assessment and validation of the models were carried out with several statistical criteria to test for their suitability and flexibility for prediction. Linear and non-linear regression models were adopted. The results indicated a total of 266 trees (67 trees/ha) in the four permanent sample plots but the study made use of the 171 trees that fell to the four families that were selected for this study (22 trees in Annonaceae, 17 in Meliceae, 87 in Sterculiaceae and 45 in Ulmaceae families). The result showed that CI4, CI5, and CI10 are mostly correlated with tree growth variables of each family except Ulmaceae family where none of the index was strongly correlated with its tree growth variables. The Multiple Linear-polynomial Model was recommended for predicting the growth variables of the trees in each family. The result showed that CI4, was the only consistent significant growth predictor among the CIs for each of the selected family except in Annonaceae where CI10 gave the best model for diameter and volume and CI10 and CI5 for height and volume in Meliaceae. Also, local indicators of spatial association to detect “hot spots,” or clusters of similar-sized trees, and “cold spots,” or clusters of trees of varying sizes within sample plots was also used in this study. This study concluded that the distance dependent CIs had higher capacity and considerable improvement in predicting individual tree growth than the non-distance dependent ones. The result of the volume model revealed that Ratkowsky model was outstanding in describing the relationship between the diameter at breast height and volume for the entire forest because it gives the lowest AIC and standard error values of 29.56 and 1.08 respectively.