Combining spatial, genetic, and environmental risk data to define and prioritize in situ conservation units

In situ management aims to preserve species and their genetic integrity within their natural habitat. To achieve this, conservation strategies must strike a balance between safeguarding genetic diversity, mitigating environmental risks, and addressing practical management constraints. Here, we present a clear and reproducible framework that addresses these goals. We applied this framework to the Nightcap reserves in the Gondwanan Rainforests of Australia, a UNESCO World Heritage site impacted by the 2019/20 Black Summer fires. We analysed the genetic diversity of 12 rainforest tree species, including three endangered species—Eidothea hardeniana, Uromyrtus australis, and Elaeocarpus sedentarius—and examined how fire risk, influenced by the presence of fire-dependent species such as eucalypts, impacts genetic diversity. To guide specific in situ management for the endangered species, we developed a flexible framework that uses clustering algorithms (DBSCAN and k-means) to define spatial management units while considering resource limitations (e.g., maximum perimeter or area). Our framework also incorporates a composite genetic value metric (combining Essential Biodiversity Variables heterozygosity, allelic richness, and genetic differentiation) and evaluates future fire risk based on vegetation flammability. This approach allowed us to identify priority management areas while adhering to resource constraints. We provide some reproducible examples of how the proposed framework can be applied, either partially or fully, to optimize in situ conservation efforts. Its flexibility allows for adjustments to fit different habitat types, species, and environmental threats, making it a valuable tool for enhancing conservation management across diverse conservation contexts.

原位保护管理（in situ management）旨在在物种的自然栖息地内留存物种及其遗传完整性（genetic integrity）。为实现这一目标，保护策略需在保护遗传多样性（genetic diversity）、降低环境风险与应对实际管理约束三者之间寻求合理平衡。本研究提出了一套清晰且可复现的研究框架以达成上述目标。
我们将该框架应用于澳大利亚冈瓦纳雨林（Gondwanan Rainforests）内的奈特卡普保护区——这是一处受2019-2020年“黑色夏季”山火（Black Summer fires）影响的联合国教科文组织世界遗产地（UNESCO World Heritage site）。我们分析了12种雨林乔木的遗传多样性，其中包含3种濒危物种（endangered species）：Eidothea hardeniana、Uromyrtus australis及Elaeocarpus sedentarius，并探究了受桉属植物（eucalypts）等火依赖型物种（fire-dependent species）影响的火险如何作用于遗传多样性。
为指导濒危物种的针对性原位保护管理，我们开发了一套灵活框架：该框架借助聚类算法（clustering algorithms）中的密度聚类算法（DBSCAN）与K均值聚类算法（k-means）划定空间管理单元（spatial management units），同时兼顾资源限制条件（如最大周长或面积限制）。本框架还整合了一项综合遗传价值指标（composite genetic value metric）——该指标结合了生物多样性核心变量（Essential Biodiversity Variables）中的杂合度（heterozygosity）、等位基因丰富度（allelic richness）与遗传分化（genetic differentiation）——并基于植被可燃性（vegetation flammability）评估未来火险等级。该方法使我们能够在遵循资源限制的前提下，识别出优先管理区域。
我们提供了若干可复现的示例，展示如何部分或完整应用所提出的框架以优化原位保护工作。该框架具备灵活性，可针对不同栖息地类型、物种与环境威胁进行调整，因此可作为提升各类保护场景下保护管理水平的实用工具。