Projection of stand development over time relies on accurate height-diameter functions. In this study, we evaluated the capability of 43 nonlinear models to estimate hornbeam heights in a portion Rezaeian experimental forest in Gorgan, Golestan province. We applied a systematic random sampling method to collect field data within a 150×200 meter network (3.33% intensity). It resulted in 200 circular plots with 17.84 m (0.1 ha) radius. In each plot tree species, height and diameter at breast height (DBH) of all trees with DBH>7.5 cm were measured, resulting in 2144 pairs of measured heightdiameter. From the available dataset, we included 70% in the model development and the remaining 30% to validate the models. The relationship between height (dependent variable) and DBH (independent variable) was analyzed using 43 non-linear regression models. The results showed no significant difference between the applied model diagnostics, and the applied t-test showed non-significant mean stand height estimation using all models and actual height at 99% confidence level. In addition, the results of peal reed, prodan, Morgan-Merser-Florin, Logestic modifield and Double-Sigmoid models with R2 of 0.826, 0.825, 0.825, 0.825 and 0.825 and RMSE% of 7.74%, 7.662%, 7.67%, 7.683% and 7.76%, respectively were almost similar in that they were better predictors of hornbeam height. Based on the results, we conclude that these models can be used for predicting hornbeam height in similar broadleaved stands of northern Iran, provided that comparative studies are conducted elsewhere to approve the results obtained here.
Type of Study:
Research |
Subject:
Special Received: 2019/01/14 | Accepted: 2019/02/20