A Random-Parameter Height-Dbh Model for Cherrybark Oak

摘要

A random-parameter model was used to relate total height to diameter at breast height (dbh) for cherrybark oak (Quercus pagoda Raf,), Data were obtained from 561 trees located in 50 stands occurring on bottomland hardwood sites in East Texas, near thewestern extent of the cherrybark oak natural range. Mixed-model estimation techniques were used to fit fixed-effects parameters to the height-dbh relationship for cherrybark oak, with random-effects parameters representing sample stands from which treedata were obtained. The fixed-effect parameter estimates can be used to predict average cherrybark oak height for a given dbh in the region from which the data were obtained. Because random parameters associated with stands were used in the data-fittingprocess, the models can be calibrated to fit new stands by obtaining measurements to fit appropriate random parameters for that stand. This calibration improves height predictions for individual stands while requiring less data than would the developmentof a completely new height prediction model for that stand.