Bayesian and classical biomass allometries for open grown valonian oaks ( Q. ithaburensis subs. macrolepis L.) in a silvopastoral system

Allometric models predicting aboveground woody biomass for open grown valonian oak ( Q. ithaburensis subs. macrolepis L.) trees growing in a Mediterranean silvopastoral system were built based on Bayesian and classical statistical techniques. The simple power model M = aD b was used for predicting aboveground woody biomass ( M ), stem ( M S ) and branch ( M B ) biomass through tree diameter ( D ). An informative Bayesian approach (IB) based on prior information about a and b and increasing variance of predicted values in relation to D was applied on 25 destructively sampled trees for estimating M . Non-informative Bayesian (NB), log-linear regression (LR) and non-linear regression were also built for M , M S and M B . Quite similar M distribution was derived from LR and NB across the D range, totally different from IB predictions which provided biologically sound estimates. Tree height, stem length and crown length did not substantially improve predictions for M , M S and M B . Comparisons to oak trees growing in closed stands indicated that open-grown oaks sustain much less stem biomass but maintain larger branch biomass than forest-grown counterparts. Comparisons to published values for open-grown green ash trees supported the hypothesis that open grown broadleaved specimens may sustain similar M values, irrespectively of species, growth conditions and tree size. On the contrary, allocation pattern of organic matter to stem and branches seems to vary by species and/or site conditions. Finally, predictions for b = 2.67 derived from a theoretical model was not supported by this dataset.