Long-term studies of vegetation recovery following post-mining restoration in low-productivity, high-stress areas are limited, but essential for understanding underlying ecological processes and evaluating management practices. This study's goal was to describe temporal patterns of recovery (up to 37 years) in vegetation structure, floristic diversity, and composition following post-mining restoration at two sites, and identify potential drivers of restoration outcomes, in the nutrient-poor, seasonally dry, species-rich, fire-prone kwongan vegetation of southwest Western Australia. Vegetation development is progressing and restoration measures are within range of native reference values, but there is large variation in both patterns observed and restoration outcomes. Several patterns described share similarities with post-fire recovery of kwongan, and post-disturbance recovery of other low-productivity, high-stress, fire-prone systems. However, differences in some patterns between sites indicate differences in the underlying mechanisms of recovery. Many management and environmental variables emerge as significant drivers of restoration outcomes but age, fire, and the planting of seedlings account for the largest amount of variation. Adaptive management at both sites appears to be facilitating improved restoration outcomes over time. Using a combination of space-for-time, plot level, and linear mixed effects modeling perspectives to examine patterns provides greater insights into restoration recovery than a chronosequence perspective alone. This study will inform restoration practices and outcomes not only in kwongan but in other comparable systems also.

中文翻译：

---

采矿后恢复的物种丰富的灌木林的结构、多样性和组成的模式和驱动因素

对低生产力、高压力地区采矿后恢复后植被恢复的长期研究是有限的，但对于了解潜在的生态过程和评估管理实践至关重要。本研究的目标是描述两个地点采矿后恢复后植被结构、植物群多样性和组成的恢复时间模式（长达 37 年），并确定恢复结果的潜在驱动因素，在营养贫乏、季节性干旱的地区，西澳大利亚西南部物种丰富、易着火的广安植被。植被正在发展，恢复措施在本地参考值范围内，但观察到的模式和恢复结果存在很大差异。所描述的几种模式与 kwong 的火灾后恢复以及其他低生产力、高压力、易发生火灾的系统的扰动后恢复有相似之处。但是，站点之间某些模式的差异表明恢复的潜在机制存在差异。许多管理和环境变量是恢复结果的重要驱动因素，但年龄、火灾和幼苗的种植造成的变化量最大。随着时间的推移，这两个站点的适应性管理似乎有助于改善恢复结果。使用空间换时间、情节级别和线性混合效应建模视角的组合来检查模式，比单独使用时间顺序视角更深入地了解恢复恢复。