Forest resource monitoring of different forest types is of great importance in sustainable forest management and climate change mitigation. Monitoring the productivity of forest resources could be achieved by modeling the basic tree parameters necessary for forest growth and yield. This study was conducted to develop a height-diameter at breast height (h-d) model necessary for tree height (h) estimation since h measurement is difficult in the field, especially in dense forests, and to estimate the forest productivity of the Diguyo limestone forest within the Northern Sierra Madre Natural Park (NSMNP). The diameter at breast height (d) and h of 124 trees were measured in seven 400-m2 plots as the basis for the model development. The h-d model was developed using different non-linear models such as the Chapman-Richards (CR), exponential (EX), Korf/Lundqvist (KL), modified logistic (ML), Schnute (SC), and Weibull (WE) models. The models were evaluated using the adjusted coefficient of determination (R2 adjusted), Akaike information criterion (AIC), Bayesian information criterion (BIC), mean absolute error (MAE), root mean square error (RMSE), percentage root mean squared error (PRMSE), and root mean squared percentage error (RMSPE). The performance of the species-specific allometric models and the generic models were compared for the biomass productivity of the limestone forest. Results showed that the CR h-d model performed best with MAE, RMSE, PRMSE, RMSPE, R2 adjusted, AIC, and BIC values of 1.47 m, 1.74 m, 19.31%, 28.71%, 0.79, 32.46, and 36.00, respectively. The highest average predicted tree biomass and carbon stock of the Diguyo limestone forest was 112.52 ± 97.65 t/ha and 50.64 ± 43.94 tC/ha, respectively, which is lower than other karst forests in Asia. The low forest resource productivity is due to the physical condition of the forest aggravated by natural and anthropogenic disturbances, thereby needing immediate attention to achieve forest sustainability.
Read full abstract