Data from: Modelling the probability of microhabitat formation on trees using cross-sectional data

1. Context: Tree-related microhabitats (TreMs), such as trunk cavities, peeled bark, cracks or sporophores of lignicolous fungi, are essential to support forest biodiversity because they are used as substrate, foraging, roosting or breeding places by bryophytes, fungi, invertebrates and vertebrates. Biodiversity conservation requires the continuous presence of TreMs in a forest. However, little is known about their dynamics. Moreover, we usually have only cross-sectional TreM data (observations of many trees at a single time), making it difficult to estimate TreM formation rates. 2. Method: This study adapted the methods of survival and reliability analysis to model the rate of TreM formation per unit of diameter increment as a function of tree diameter at breast height (DBH). We tested three variants of this model: the TreM formation rate independent of, proportional to or increasing non-linearly with DBH. We calculated the likelihood of the models, considering cross-sectional observations either of TreM presence/absence or TreM number on trees of different sizes. We calibrated the models in six sub-natural montane forests dominated by European beech (Fagus sylvatica) and silver fir (Abies alba) – in the French Pyrenees. Assuming an annual DBH increment value, the annual formation rate of TreMs was predicted both at the level of the tree and at the level of the forest stand. 3. Results: This method provided a coherent framework to model the probability that a TreM forms on a tree during a unit growth step and produces realistic predictions of TreM accumulation on trees. TreM formation accelerated as trees grew for A. alba but not for F. sylvatica. The TreM formation rate was twice as fast on F. sylvatica as on A. alba. We estimated a formation of 0.82–1.28 TreMs/ha per year and 0.5–0.9 TreM bearing trees/ha per year in the sub-natural forests studied. 4. Synthesis and applications: This method makes rigorous modelling of the formation of TreMs possible during the growth of trees and forest stands. The quantitative evaluation of TreM fluxes will help to design forest biodiversity conservation strategies favouring the development and temporal continuity of TreMs.

1. 研究背景：树木相关微生境（Tree-related microhabitats, TreMs），如树干树洞、剥落树皮、裂隙或木质真菌子实体，是支撑森林生物多样性的核心要素，因为苔藓植物、真菌、无脊椎动物和脊椎动物可将其用作基质、觅食场、栖息处或繁殖场所。生物多样性保护要求森林中持续存在树木相关微生境，但目前对其动态变化的认知仍较为匮乏。此外，现有数据通常仅为横断面的TreM观测数据（即单次时间点对大量树木的观测），这使得难以准确估算TreM的形成速率。
2. 研究方法：本研究借鉴生存分析与可靠性分析的方法，以胸径（diameter at breast height, DBH）为自变量，构建单位直径生长增量对应的TreM形成速率模型。我们测试了该模型的三种变体：TreM形成速率与胸径无关、与胸径成比例，或随胸径呈非线性增长。针对不同胸径级树木的TreM存在/缺失状况以及树木上TreM数量的横断面观测数据，我们计算了各模型的似然值。研究在法国比利牛斯山脉的6处以欧洲山毛榉（Fagus sylvatica）和银冷杉（Abies alba）为优势种的近天然山地森林中完成模型校准。在设定年胸径生长增量的前提下，我们分别在单棵树木和林分尺度上预测了TreM的年形成速率。
3. 研究结果：本方法提供了一套严谨的分析框架，可用于模拟单位生长周期内树木形成TreM的概率，并能对树木上TreM的累积情况生成符合实际的预测结果。银冷杉的TreM形成速率随树木生长逐渐加快，但欧洲山毛榉未呈现该规律。欧洲山毛榉上的TreM形成速率是银冷杉的两倍。在所研究的近天然森林中，我们估算得到年TreM形成量为0.82–1.28个/公顷，年携带TreM的树木数量为0.5–0.9株/公顷。
4. 综合与应用：本方法可实现树木与林分生长过程中TreM形成的严谨建模。对TreM通量的定量评估将有助于制定有利于TreM发育与时间连续性的森林生物多样性保护策略。