Analysis and comparison of nonlinear tree height prediction strategies for Douglas-fir forests

摘要: Using an extensive Douglas-fir data set from southwest Oregon, we examined the (1) performance and suitabil- ity of selected prediction strategies, (2) contribution relative position stand-density measures in improving tree height (h) values, (3) effect different subsampling designs to fill missing h values a new stand using regional nonlinear model. Nonlinear mixed-effects models (NMEM) substantially improved accuracy precision over conventional fixed-effects model (NFEM) that assumes observations are inde- pendent, particularly when few trees subsampled for height. The predictive correction factor on NFEM with was comparable NMEM four or more were When two heights randomly subsampled, efficiently explained differences height-diameter relationship because variations density without having incorporate them into only one selecting largest diameter would result lower predicted root mean square error (RMSE) than height, re- gardless form fitting strategy used.