Table 2 Assessment of the dominant/codominant tree height models for the Norway spruce in the Carpathian Mountains. The abbreviations are: Giurgiu Giurgiu–Draghiciu, Pars. Parsimonious, Var variance

From: Nonlinear parsimonious forest modeling assuming normal distribution of residuals

Site index Model Bias (%) MAE (%) RMSE (%) Correlation coefficient (%)
Mean Var Mean Var Mean Var
21.9 Giurgiu −2.1 84 12.9 84 8.7 84 91.7
Bayesian −2.0 82 13.3 82 8.8 83 91.4
Pars. 10 terms −1.9 12 13.3 50 8.8 57 91.4
Pars. 30 terms −2.0 12 13.3 50 8.8 57 91.4
26.9 Giurgiu 0.01 95 6.1 95 7.3 95 97.4
Bayesian −0.05 94 6.3 94 7.7 94 97.1
Pars. 10 terms −0.04 −9.7 6.3 29 7.7 72 97.1
Pars. 30 terms −0.04 −9.7 6.3 29 7.7 72 97.1
31.8 Giurgiu 3.4 96.3 4.8 96.3 5.6 96.3 98.1
Bayesian −0.5 95.3 5.1 95.3 6.0 95.3 97.7
Pars. 10 terms −0.5 −25.9 5.1 72.3 6.0 84.1 97.7
Pars. 30 terms −0.5 −25.9 5.1 72.3 6.0 84.1 97.7
36.9 Giurgiu −0.08 96.8 4.3 96.8 5.2 96.8 98.3
Bayesian −0.5 94.9 4.9 94.9 6.1 94.9 97.4
Pars. 10 terms −0.5 −38.7 4.9 80.4 6.1 94.6 97.4
Pars. 30 terms −0.5 −38.7 4.9 80.4 6.1 94.6 97.4