1 INTRODUCTION

Biodiversity is facing the strongest crisis since the last mass extinction, with extinction rates estimated to be 100 times higher than background rates (Ceballos et al., 2015). The current biodiversity crisis is caused by multiple processes such as habitat modifications, climate change, overexploitation, spread of diseases, anthropogenic nitrogen deposition, and introduction of invasive species (Stuart et al., 2008). However, these drivers do not influence populations independently, but can act together or agonistically at multiple levels, complicating the identification of key processes affecting populations and species survival (Didham et al., 2007; Falaschi et al., 2019).

Amphibians are an exemplary case of the current biodiversity crisis, with >40% of species recognised as threatened by the International Union for Nature Conservation (Hoffmann et al., 2010). Land-use change and alien vertebrates and crustaceans are listed among their main threats (Hof et al., 2011). Habitat loss is the threatening factor affecting the largest number of amphibian species; therefore, we might expect it to be the strongest driver of populations trends. However, the intensity of land-use change can vary through space and time with areas of the world that have suffered strong habitat loss through the past centuries (Falcucci et al., 2007; Goldewijk et al., 2011; Hansen et al., 2013). By contrast, invasive alien species are a growing issue (Falaschi et al., 2020). The number of invasive alien species is exponentially increasing at the global scale, exerting heavy impact even in areas with well-conserved habitat (Denoël et al., 2019; Seebens et al., 2017).

Population dynamics of amphibians can result from processes acting at different scales. Both the local characteristics of breeding sites and the features of the landscape surrounding them can strongly affect the temporal dynamics of amphibian populations (Dalpasso et al., 2022; Falaschi et al., 2021; Lowe & Bolger, 2002). Additionally, amphibians often exploit discrete patches of breeding habitat connected by dispersing individuals, called spatially structured populations (SSP; Revilla & Wiegand, 2008). Hence, factors influencing the connectivity among patches can also be crucial in determining population dynamics (Manenti et al., 2020). Appropriate management of these features, such as the manipulation of water regime and wetlands density across the landscape, can mitigate the negative impact of other stressors and halt biodiversity loss (Mathwin et al., 2020; Rannap et al., 2009).

Studying population dynamics is the most straightforward way to gather useful information for ecology and biodiversity conservation, and information on abundance changes is pivotal to evaluate the conservation status of species (IUCN, 2001). However, the potential drivers of abundance are often assessed in snapshot correlative studies, in which population abundance at a given time is related to the spatial variation of candidate environmental predictors. Such snapshot studies are often unable to identify the main factors determining temporal dynamics, and similar studies can even yield strongly contrasting results. For instance, Ficetola et al. (2011) found a negative correlation between the abundance of larval amphibians and the presence of alien crayfish, while similar analyses performed in a different area did not detect clear negative relationships (Bélouard et al., 2019). The study of abundance dynamics can be challenging because it requires planning long-term sampling and analysing the collected data with appropriate models. For instance, population growth at a given time can be density-dependent (Cayuela et al., 2019). Additionally, the detection probability of individuals can be low, requiring several surveys at each site within each sampling season to obtain reliable measures of abundance (Falaschi, 2021; Ficetola et al., 2018; Kellner & Swihart, 2014).

These issues strongly limited quantitative assessments of abundance dynamics for species for which exhaustive census data are available. Despite possible challenges, measuring demographic trends, and studying the factors determining temporal changes is essential should we want to assess the conservation status of species, and ensure their long-term persistence (IUCN, 2001; Redford et al., 2013). Recently developed demographic models integrating analyses of detection probability are extremely promising, as they can quantify species trends and identify their drivers (Manenti et al., 2020; Zhao et al., 2019). However, they have been rarely used so far to measure trends at a broad (e.g. regional) scale.

We performed multiple surveys between 1996 and 2020 and used state-space demographic models to assess the importance of multiple factors in determining regional-scale abundance dynamics of a particularly vulnerable group of amphibians: newts, which form complex networks of populations linked to both freshwater and terrestrial environments (Beebee, 2014). We evaluated the effect of candidate drivers at different scales, and we considered: (1) the area, hydroperiod, presence of fish, and the presence of an invasive crayfish as characteristics of breeding sites; (2) terrestrial habitat availability as a landscape feature; and (3) incidence of crayfish and incidence of the focal species in the surrounding landscape as factors acting on the connectivity among wetlands. A previous study assessed the drivers of amphibian occupancy, including newts, in the same study region (Falaschi et al., 2021), but a lack of quantitative estimations of abundance hampered a complete evaluation of demographic changes. Here, we exploit the power of N-mixture models to estimate trends of abundance at the regional scale and evaluate the drivers of abundance through long-term monitoring of representative sites.

中文翻译：

---

解释意大利北部蝾螈丰度的下降

1 简介

生物多样性正面临自上次大灭绝以来最严重的危机，灭绝率估计是背景率的 100 倍（Ceballos 等人，  2015 年）。当前的生物多样性危机是由多种过程引起的，例如栖息地改变、气候变化、过度开发、疾病传播、人为氮沉降和入侵物种的引入（Stuart et al.,  2008）。然而，这些驱动因素不会独立影响种群，而是可以在多个层面共同或相互竞争，从而使影响种群和物种生存的关键过程的识别变得复杂（Didham 等人，  2007 年；Falaschi 等人，  2019 年）。

两栖动物是当前生物多样性危机的典型案例，超过 40% 的物种受到国际自然保护联盟的威胁（Hoffmann 等人，  2010 年）。土地利用变化和外来脊椎动物和甲壳类动物被列为主要威胁之一（Hof 等人，  2011 年）。栖息地丧失是影响两栖动物数量最多的威胁因素；因此，我们可能期望它成为人口趋势的最强大驱动力。然而，土地利用变化的强度会随着空间和时间的变化而变化，在过去的几个世纪中，世界上遭受严重栖息地丧失的地区（Falcucci 等人，  2007 年；Goldewijk 等人，  2011 年；Hansen 等人，  2013）。相比之下，外来入侵物种是一个日益严重的问题（Falaschi 等人，  2020 年）。外来入侵物种的数量在全球范围内呈指数级增长，即使在栖息地保存完好的地区也会产生重大影响（Denoël 等人，  2019 年；Seebens 等人，  2017 年）。

两栖动物的种群动态可能来自不同尺度的过程。繁殖地的当地特征和它们周围的景观特征都会强烈影响两栖动物种群的时间动态（Dalpasso et al.,  2022 ; Falaschi et al.,  2021 ; Lowe & Bolger,  2002）。此外，两栖动物经常利用分散的个体连接的离散的繁殖栖息地，称为空间结构种群（SSP；Revilla 和 Wiegand，  2008 年）。因此，影响斑块之间连通性的因素对于确定种群动态也至关重要（Manenti et al.,  2020）。对这些特征进行适当管理，例如控制整个景观的水情和湿地密度，可以减轻其他压力源的负面影响并阻止生物多样性丧失（Mathwin 等人，  2020 年；Rannap 等人，  2009 年）。

研究种群动态是为生态和生物多样性保护收集有用信息的最直接方法，而丰度变化信息对于评估物种保护状况至关重要（IUCN，  2001）。然而，丰富的潜在驱动因素通常在快照相关研究中进行评估，其中给定时间的人口丰度与候选环境预测因子的空间变化有关。这种快照研究通常无法确定决定时间动态的主要因素，类似的研究甚至可以产生强烈对比的结果。例如，Ficetola 等人。( 2011) 发现两栖动物幼虫的数量与外来小龙虾的存在呈负相关，而在不同地区进行的类似分析并未发现明显的负相关关系（Bélouard 等人，  2019 年）。丰度动态研究可能具有挑战性，因为它需要规划长期采样并使用适当的模型分析收集的数据。例如，给定时间的人口增长可能与密度有关（Cayuela 等人，  2019 年）。此外，个体的检测概率可能很低，需要在每个采样季节在每个地点进行多次调查以获得可靠的丰度测量值（Falaschi，  2021；Ficetola 等，  2018；Kellner & Swihart，  2014）。

这些问题极大地限制了对可获得详尽普查数据的物种的丰度动态的定量评估。尽管可能存在挑战，但如果我们想要评估物种的保护状况并确保它们的长期存在，测量人口趋势和研究决定时间变化的因素是必不可少的（IUCN，  2001；Redford 等，  2013）。最近开发的集成了检测概率分析的人口统计模型非常有前景，因为它们可以量化物种趋势并确定其驱动因素（Manenti 等人，  2020 年；Zhao 等人，  2019 年）。然而，迄今为止，它们很少用于衡量广泛（例如区域）范围内的趋势。

我们在 1996 年至 2020 年期间进行了多项调查，并使用状态空间人口模型来评估多种因素在确定特别脆弱的两栖动物群体的区域规模丰度动态方面的重要性：蝾螈，它们形成了与淡水和淡水相关的复杂种群网络陆地环境（Beebee，  2014）。我们评估了不同尺度的候选驱动因素的影响，我们考虑了：（1）区域、水周期、鱼类的存在和入侵小龙虾的存在作为繁殖地点的特征；(2) 作为景观特征的陆地栖息地可用性；(3) 小龙虾的发生率和周边景观中重点物种的发生率是影响湿地之间连通性的因素。先前的一项研究评估了同一研究区域内两栖动物（包括蝾螈）居住的驱动因素（Falaschi 等人，  2021 年），但缺乏对丰度的定量估计阻碍了对人口变化的完整评估。在这里，我们利用N的力量- 混合模型来估计区域尺度的丰度趋势，并通过对代表性地点的长期监测来评估丰度的驱动因素。