Salmon on the lam: Drivers of escaped farmed fish abundance in rivers

Abstract: The production of Atlantic salmon in aquaculture has grown substantially over the last 40 years. The unintentional release of domesticated salmon poses a significant risk to the long-term persistence of wild Atlantic salmon populations through ecological interactions and genetic introgression. Our ability to link aquaculture production to farmed escaped salmon in rivers is still limited and hinders identifying the appropriate production capacity of salmon aquaculture to reduce unwanted interactions between wild and escaped Atlantic salmon. Here, we use a 14-year dataset of farmed escapee abundance in rivers to model how the a priori selected covariables of wild salmon abundance, aquaculture intensity, river discharge, hydropower, and fjord placement of the river affects escapee abundance across 54 rivers in western Norway. Then, we evaluate the predictive strength of the model to provide context for its use to minimize escapees. We found that the abundance of farmed escaped Atlantic salmon in rivers is correlated to aquaculture intensity. Furthermore, the abundance of wild Atlantic salmon, mean yearly discharge, and the interaction between fjord placement and wild salmon abundance were important predictors of escapee abundance in rivers. The model was 40% accurate when predicting the abundance of farmed escaped salmon in rivers. However, the accuracy improved to 75% when using risk categories derived from modeled intrusion rates that induced long-term genetic changes to the wild population (low &lt; 4%, medium 4 – 10%, and high &gt; 10% escaped farmed salmon). <em>Synthesis and applications: </em>This study links aquaculture production, at relevant spatiotemporal scales (75 km from rivers), to the abundance of escaped farmed Atlantic salmon in rivers, and provides governmental agencies with a tool to help regulate domesticated salmon production based on the carrying capacity of the system to buffer against introgression between conspecifics. Furthermore, understanding this relationship will be beneficial when establishing new aquaculture sites in pristine ecosystems where they would overlap with wild Atlantic salmon. Finally, future mitigation efforts should continue to focus on new technologies (e.g., triploid females) that can eliminate the risk of introgression without limiting aquaculture production.摘要：过去40年来，水产养殖（aquaculture）大西洋鲑（Atlantic salmon）的产量大幅增长。养殖鲑的意外逃逸会通过生态交互与遗传渐渗（genetic introgression），对野生大西洋鲑种群的长期存续构成显著风险。当前我们将水产养殖产量与河流中养殖逃逸鲑关联起来的能力仍存在局限，这阻碍了我们确定鲑鱼水产养殖的合理产能，以减少野生大西洋鲑与逃逸养殖鲑之间的不良交互。 本研究基于挪威西部54条河流为期14年的养殖逃逸鲑丰度数据集，构建模型，探究野生鲑丰度、水产养殖强度、河流径流量、水电开发以及河流峡湾位置这些先验选定的协变量（covariables），如何影响研究区域内54条河流中的逃逸鲑丰度。随后，我们评估了该模型的预测能力，为其在最小化逃逸鲑风险中的应用提供依据。 研究发现，河流中养殖逃逸大西洋鲑的丰度与水产养殖强度存在相关性。此外，野生大西洋鲑丰度、年平均径流量，以及峡湾位置与野生鲑丰度的交互项，均是预测河流中逃逸鲑丰度的重要因子。该模型对河流中养殖逃逸鲑丰度的预测准确率为40%。不过，当采用基于模型推导的入侵风险等级（可对野生种群造成长期遗传改变：低风险<4%、中风险4%–10%、高风险>10%的养殖逃逸鲑占比）时，预测准确率提升至75%。 **合成与应用（Synthesis and applications）**：本研究将相关时空尺度（spatiotemporal scales，距河流75公里范围内）的水产养殖产量，与河流中养殖逃逸大西洋鲑的丰度建立关联，可为政府管理机构提供一套工具，帮助其基于生态系统承载能力规范养殖鲑生产，以规避种内（conspecifics）遗传渐渗带来的风险。此外，在与野生大西洋鲑栖息地重叠的原生生态系统中新建水产养殖场址时，明晰这一关联将具有重要参考价值。最后，未来的防控措施应持续聚焦于可消除遗传渐渗风险且不限制水产养殖产能的新技术（例如三倍体雌鱼（triploid females））。