木材密度包括基本密度、气干密度等，在12%含水率条件下的气干密度(D12)较常用，因此有必要将木材气干密度换算为基本密度(Db)。目前利用木材气干密度计算基本密度的模型有Reyes、Chave、Simpson和Vieilledent模型等，然而这些模型预测结果不完全一致。利用中国林业科学研究院木材工业研究所(Research Institute of Wood Industry，Chinese Academy of Forestry，CRIWI)和法国农业国际合作研究发展中心(French Agricultural Research Centre for International Development，CIRAD)的木材D12和Db数据，首先基于CRIWI的木材密度数据建立D12与Db的关系模型，然后将CRIWI和CIRAD的D12数据分别代入Reyes模型、Chave模型、Simpson模型、Vieilledent模型和新建模型，获得每个树种木材Db的预测值，并根据Db预测值和实测值计算残差绝对值均值。不同模型残差绝对值均值比较结果表明:Reyes模型在利用CRIWI和CIRAD的木材密度数据时预测Db的准确性都比较高，适用性最广;Simpson模型、新建模型在D12高于1。0 g/cm3时预测Db的准确性降低。
Comparison of Correlation Models Between Wood Basic Density and Air-Dry Density
Wood density includes basic density,air-dry density,etc.The air-dry density(D12)measured at the moisture content of 12%is more commonly used,so it is necessary to convert wood air-dry density to basic density(Db).Currently,the models that use air-dry density to calculate basic density,include Reyes,Chave,Simpson,and Vieilledent models.However,these models do not offer completely consistent predictions.This study used data of wood basic density and air-dry density from Research Institute of Wood Industry,Chinese Academy of Forestry(CRIWI)and French Agricultural Research Centre for International Development(CIRAD)to evaluate the prediction precision difference among the models mentioned above.Firstly,a correlation model-between D12 and Db was established based on CRIWI wood density database.Then,the air-dry density at 12%moisture content of individual wood species from the CRIWI and the CIRAD wood density database was applied to the Reyes model,Chave model,Simpson model,Vieilledent model,and the newly fitted model respectively to predict basic densities of each wood species.Then the mean absolute value of the error(MAE)was calculated from the predicted and the measured value of the wood basic density.By comparing the MAE of different models,the results indicate that only the Reyes model can appropriately predict basic density corresponding to either the CRIWI or CIRAD wood density databases.For both the Simpson model and the newly fitted model,the accuracy in predicting basic density significantly decreases when the air-dry density at 12%moisture content exceeds 1.0 g/cm3.
国家科技期刊平台
NSTL
万方数据
