Aliens in the twilight zone: using robots to study and manage invasive lionfish on Caribbean mesophotic coral ecosystems

入侵的狮子鱼（Pterois spp.）在西大西洋广泛传播，对当地海洋生物多样性构成了重大威胁。狮子鱼不受限制地捕食本地鱼类种群，与狮子鱼在印度-太平洋地区的本土范围不同，这些鱼类没有进化出的防御能力。目前的管理很大程度上依赖于娱乐性的水肺潜水员的机会性捕杀，但这种方法只能提供局部的短期喘息机会。这是因为休闲潜水员倾向于停留在30米以上，然而狮子鱼在这个深度以下的数量非常多，因此错过了。此外，最近的证据表明，这些生活在中孔珊瑚生态系统（MCEs: 30-150米）中的深层狮子鱼比生活在浅水区的狮子鱼更大，繁殖能力更强。对mce的研究很少，因为它们对传统的水肺潜水来说太深了，但又不够深，不足以证明昂贵的研究潜水器是合理的。Wallacea行动与美国的慈善机构机器人环境服务组织（RSE）进行了新的合作，该组织开发了一种小型的、具有成本效益的机器人，能够进入mce并捕获狮子鱼。在2019年的试点测试中，该机器人成功地在85米的深度下运行，并被设计为超过200米的深度。研究问题：这些深海狮子鱼种群可能在很大程度上促进了浅层珊瑚礁的繁殖，通过允许在扑杀后快速重新定居，破坏了管理工作。这种潜力的实现程度，以及它对管理工作的影响程度，需要对MCE狮子鱼有更深入的生态了解。我们的目标是利用这些机器人令人兴奋的新功能来(1)确定MCE狮子鱼在物种种群动态中的作用，(2)量化入侵狮子鱼对更广泛的MCE鱼类群落的影响，以及(3)使用空间结构的人口统计学模型来制定MCE狮子鱼种群的最佳管理实践。实地工作将主要在洪都拉斯的Wallacea行动的长期海洋研究站点进行（自2005年以来一直活跃；自2016年以来有23篇同行评议的科学论文）。使用外部资助的、安装有摄像头的定制原型机器人，将收集狮子鱼及其栖息生态系统的标准化数据，特别是关注mce。具体来说，视频调查将用于调查狮子鱼在其全深度范围内的分布，并确定猎物的可用性。机器人还将捕捉狮子鱼并将它们放回水面，以便分析肠道内容和跨深度的人口变化。利用这些数据，将开发一个空间结构的人口统计模型来了解狮子鱼的种群动态。利用这个模型，我们将探索不同的筛选情景，并为最优策略提出假设。这些策略将通过涉及机器人深度筛选的现场实验进行测试。基于结果，种群模型将被改进以应用于其他地区，并为全球狮子鱼管理提供信息。第一年：强化培训/入职培训。文献综述。视频观察（洪都拉斯）。丰度分布、种群结构、捕食行为和更广泛的MCE鱼类群落分析。二年级：肠道含量，繁殖力和身体状况分析。开发空间结构的种群模型，为筛选实验提供信息。第三年：实地实验，以确定最佳的剔除策略和验证/改进模型。制定和实施入侵狮子鱼管理的最佳实践指南。

Research contextInvasive lionfish (Pterois spp.) have spread widely across the western Atlantic and are a substantial threat to native marine biodiversity. Lionfish predate unchecked on native fish populations which, unlike in the lionfish's native Indo-Pacific range, have no evolved defence. Current management relies largely on opportunistic culling by recreational SCUBA divers yet this method only provides local short-term respite. This is because recreational divers tend to stay above 30 m, yet lionfish are at high abundance below this depth and are consequently missed. Also, recent evidence shows these deeper lionfish - in mesophotic coral ecosystems (MCEs: 30-150 m) - are larger and more fecund than those found in the shallows. MCEs are poorly studied because they are too deep for conventional SCUBA diving but not deep enough to justify expensive research submersibles. Operation Wallacea has a new collaboration with US-based charity Robots in Service of the Environment (RSE), which has developed a small, cost-effective robot capable of accessing MCEs and capturing lionfish. The robot has successfully operated at 85m depth during 2019 pilot tests and is designed to reach beyond 200m. Research questionsThese deep populations of lionfish may contribute substantially to recruitment in shallow reefs, undermining management efforts by allowing rapid recolonisation after culls. How far this potential is realised, and thus how much it compromises management efforts, requires greater ecological understanding of MCE lionfish.We aim to exploit these robots exciting new capability to (1) determine the role of MCE lionfish in the species' population dynamics, (2) quantify the impacts of invasive lionfish on the wider MCE fish community, and (3) use spatially-structured demographic models to develop best management practice for culling MCE lionfish populations.MethodsFieldwork will primarily be at Operation Wallacea's long-term marine research sites in Honduras (active since 2005; 23 peer-reviewed scientific papers since 2016). Using externally funded, bespoke prototype robots mounted with cameras, standardised data will be collected on lionfish and the ecosystems they inhabit, particularly focusing on MCEs. Specifically, video surveys will be used to survey the distribution of lionfish at their full depth range and determine prey availability. The robots will also catch lionfish and return them to the surface to enable gut content analysis and demographic changes across depth. Using these data, a spatially-structured demographic model will be developed to understand lionfish population dynamics. With this model, we will explore different culling scenarios and develop hypotheses for optimal strategies. These strategies will be tested via field experiments involving robot-enabled deep culling. Based on the results, the population model will be refined for application to other locations, and to inform lionfish management globally.TimelineYear 1: intensive training/orientation. Literature review. Video observations (Honduras). Analysis of abundance distribution, population structure, predatory behaviour and wider MCE fish community. Year 2: Gut content, fecundity and body condition analyses. Develop spatially-structured population model to inform culling experiment.Year 3: Field experiments to determine optimal culling strategy and validate/refine model. Develop and implement best-practice guidelines for invasive lionfish management.