A nonlinear mixed-effects tree height prediction model: Application to Pinus pinaster Ait in Northwest Spain

Abstract

This study introduced a new height prediction model based on the modification of power function distribution. The model was fitted to the Pinus pinaster Ait (Maritime pine) data set comprised of 14339 trees measured from 155 permanent sample plots in northwest Spain. Nonlinear mixed-effect was to refit the model and calibrated with the random effects predicted from one to four sample trees per plot using validation data set (3621 trees from 29 plots). The models were evaluated based on different indices including root mean square error (RMSE), critical error (Ecrit) and adjusted coefficient of determination (R¯2). The results showed that the performance of the model was improved by the inclusion of random parameter with RMSE, Ecrit and R¯2of 1.143, 0.700 and 0.910, respectively. The calibration response of the mixed-effect model involved the selection of four trees per sample plot – each close to the 25th, 50th, 75th quartiles and maximum diameter. It resulted in about 40% reduction of the RMSE compared to other alternatives.

Keywords

• Power function distribution
• height-diameter
• mixed-effect
• model calibration
• Pinus pinaster