Abstract Stand structure develops with stand age. Old‐growth forests with well‐developed stand structure support many species. However, development rates of stand structure likely vary with climate and topography. We modeled structural development of 4 key stand variables and a composite old‐growth index as functions of climatic and topographic covariates. We used a hierarchical Bayesian method for analysis of extensive snap‐shot National Forest Inventory (NFI) data in Japan (n = 9244) to account for differences in stand age. Development rates of structural variables and the old‐growth index exhibited curvilinear responses to environmental covariates. Flat sites were characterized by high rates of structural development. Approximately 150 years were generally required to attain high values (approximately 0.8) of the old‐growth index. However, the predicted age to achieve specific values varied depending on environmental conditions. Spatial predictions highlighted regional variation in potential structural development rates. For example, sometimes there were differences of >100 years among sites, even in the same catchment, in attainment of a medium index value (0.5) after timber harvesting. The NFI data suggested that natural forests, especially old natural forests (>150 years), remain generally on unproductive ridges, steep slopes, or areas with low temperature and deep snow, where many structural variables show slow development rates. We suggest that maintenance and restoration of old natural forests on flat sites should be prioritized for conservation due to the likely rapid development of stand structure, although remaining natural forests on low‐productivity sites are still important and should be protected.

中文翻译：

---

基于气候和地形的天然林结构发展过程的空间显性经验模型

摘要 林分结构随着林分年龄的增长而发展。具有发育良好林分结构的古老森林支持许多物种。然而，林分结构的发展速度可能因气候和地形而异。我们模拟了 4 个关键林分变量的结构发展和作为气候和地形协变量的函数的复合老生长指数。我们使用分层贝叶斯方法分析日本（n = 9244）的大量快照国家森林清单 (NFI) 数据，以解释林龄差异。结构变量的发展速度和旧增长指数对环境协变量表现出曲线响应。平坦场地的特点是结构发展速度快。通常需要大约 150 年的时间才能达到旧增长指数的高值（大约 0.8）。然而，达到特定值的预测年龄因环境条件而异。空间预测突出了潜在结构发展速度的区域差异。例如，有时即使在同一流域，不同地点之间在木材采伐后达到中等指数值 (0.5) 时也存在超过 100 年的差异。NFI 数据表明，天然林，尤其是古老的天然林（>150 年），普遍停留在非生产性山脊、陡坡或低温和积雪较深的地区，许多结构变量显示出缓慢的发展速度。由于林分结构可能快速发展，我们建议优先保护和恢复平坦场地上的古老天然林，