ABSTRACT A method for simulation of distributions of structural timber properties based on geographical data and stand-level forest inventory data is presented. The method is based on Monte Carlo simulation of randomly selected sites from volume distributions of standing timber, in our case from the Norwegian National Forest Inventory. Simulation models of density, modulus of elasticity and bending strength were estimated from data from 2369 boards of Norway spruce from 27 sites in Norway. The models were estimated as mixed models including covariate models with fixed effects of latitude, altitude, site index and stand age, random site effects and residual variances, representing the within-site variances. Correlations between the simulated properties were obtained by transformation of the random effects in the simulations. The transformations were based on Cholesky decomposition of the correlation matrices of the random site-effects and the residuals of the simulation models. The simulations provide distributions for timber from each selected site and from all selected sites within defined areas, with different limitations on geographical data and stand-level forest inventory data. Simulations of Norway spruce from Eastern Norway are presented for demonstration, comparing timber from maturity classes, altitudes and site indices.

中文翻译：

---

基于地理数据和林分森林清单数据的结构木材特性模拟

摘要提出了一种基于地理数据和林分森林清单数据模拟结构木材特性分布的方法。该方法基于从立木体积分布中随机选择的地点的蒙特卡罗模拟，在我们的案例中来自挪威国家森林清单。密度、弹性模量和弯曲强度的模拟模型是根据来自挪威 27 个地点的 2369 块挪威云杉板的数据估算的。这些模型被估计为混合模型，包括具有纬度、海拔、场地指数和林龄的固定效应、随机场地效应和残差方差的协变量模型，代表场地内方差。模拟特性之间的相关性是通过模拟中随机效应的转换获得的。转换基于随机站点效应的相关矩阵和模拟模型的残差的 Cholesky 分解。模拟提供了来自每个选定地点和定义区域内所有选定地点的木材分布，对地理数据和林分级森林清单数据有不同的限制。展示了来自挪威东部的挪威云杉的模拟以供演示，比较来自成熟度等级、海拔高度和场地指数的木材。