Abstract:
Volume estimation is an importantparameter for forest inventory, improve productivity andfor decision making and sustainable management. Height (H), Diameter at breast height(DBH) and form factor are the most essential variables toestimate stem volume. However,few studies have been available in Ethiopiaand no studies onPinus patula species inAmhara region. Therefore, there is a need to determine the form factor and develop sitespecific height and stem volume estimation models for predicting treevolume. The study wasaimed at determine form factors and develop yield estimation model for Pinus patula trees inSouth Gondar Zone, Amhara Region, Ethiopia. To achieve this objective a total of 84individual trees were selected for harvesting and 14 representative trees were selectedrandomly from each 6 diameter class.From the total data, 75% (63 trees) data was used formodel fitness and 25 % (21 trees) data were used for model validation. Model fitness andvalidated data were selected from total sampling data through simple random samplingwithout replacement.Microsoft Excel, descriptive statistics was used to analyze the formfactor and Regression analysis using SPSS software version 21 was used to analyze treeheight and volume prediction model.The performance of the fitted models was assessed usingseveral statistics such as Coefficient determination (R²), Root Mean Square Error (RMSE),Mean Absolute Error (MAE) and Mean Absolute Percent Error (MAPE) and paired sample t-test were used for modelvalidation. In the study, the mean real form factor value of the Pinuspatula was 0.505 (±0.056). The form factor consistently decreased as diameter size increaseswith significant variation. The form factor of Cupressus lusitanica (0.560) that has beenusedby Amhara Forest Enterprise wassignificanthigherwith the value of the present study. Thelog-transformed data shows that DBH explained 73.1% of the variation in the height withMAPE of 8.1% and the validation succeeded and then model 2(lnH=lnbo+b1lnDBH)was thebest estimator of tree height. The DBH explained 97.8% variability in the stem volume withMAPE of 9.7%, whereas adding H improved the model performance with MAPE of 6.6%andthus models with DBH and Hmodel 3(lnV =lnbo+b1lnDBH+b2lnH) and model5(lnV=lnbo+b1lnDBH +b2lnDBH*H)were found to be the best models for the accurate estimationof volume stocks of Pinus patula species or plantation stand in the study area
