For centuries, forest management has provided great benefits for local and global populations, especially in terms of wood production (Wagner et al. 2004, Puettmann et al. 2009). However, recent events have begun to undermine the confidence that our management practices are adequate to ensure the continued provision of desired ecosystem services (Figure 1). For example, large-scale insect infestations in western Canada and western Europe and large, high-intensity fires in Australia and California have encouraged discussions, with many people suggesting that we will have to deal with such “extreme events” (defined as rare, but high-impact events that lead to irreversible, unacceptable outcomes, see Box 1Box 1) more frequently in the future (Dale et al. 2001, Sheehan et al. 2015, Seidl et al. 2018). A closer look at statistical and analytical concepts suggests that extreme events are an inherent part of forest ecosystems, and that we may be well advised to acknowledge this fact in research, education, and planning and implementing forest management in the face of a novel and highly uncertain future. One of the reasons that these events have not gotten sufficient attention is that because of their rarity, extreme events are hard to study, and thus often are easily and—understandably—ignored in silvicultural and management decisions. For example, we have paid limited attention to the role of fires in the moist, western forests in the Pacific Northwest with a fire return interval of 300 years or longer (Agee 1993), even though describing the disturbance regime by its fire return interval acknowledges that we understand the high impact that fires can have on the landscape. Furthermore, typically used statistical approaches and assumptions may not apply to extreme events (see Box 1Box 1). Thus, acknowledging extreme events suggest the need for a critical review of our silviculture practices and any assessment should use a solid scientific basis.

中文翻译：

---

极端事件：在预料之外的情况下管理森林

几个世纪以来，森林管理为当地和全球人口带来了巨大好处，尤其是在木材生产方面（Wagner 等人，2004 年，Puettmann 等人，2009 年）。然而，最近的事件开始削弱我们的管理实践足以确保持续提供所需生态系统服务的信心（图 1）。例如，加拿大西部和西欧的大规模昆虫侵扰以及澳大利亚和加利福尼亚州的大规模高强度火灾引发了讨论，许多人建议我们将不得不应对此类“极端事件”（定义为罕见、但导致不可逆、不可接受结果的高影响事件，见框 1框 1）在未来更频繁（Dale et al. 2001, Sheehan et al. 2015, Seidl et al. 2018）。对统计和分析概念的仔细研究表明，极端事件是森林生态系统的固有组成部分，我们可能会被建议在研究、教育、规划和实施森林管理中，面对新的高度不确定的未来。这些事件没有得到足够重视的原因之一是，由于它们的罕见性，极端事件很难研究，因此在造林和管理决策中往往很容易并且可以理解地被忽略。例如，我们很少关注火灾在太平洋西北部潮湿的西部森林中的作用，其火灾重现间隔为 300 年或更长时间（Agee 1993），尽管通过火灾重现间隔来描述扰动状况承认我们了解火灾可能对景观产生的重大影响。此外，通常使用的统计方法和假设可能不适用于极端事件（见方框 1 方框 1）。因此，承认极端事件表明需要对我们的造林实践进行批判性审查，任何评估都应使用坚实的科学基础。