Preventing biodiversity loss in the face of global change is a major challenge in ecology and conservation (Folke et al., 2004;Scheffer et al., 2012). As global change accelerates (Hoegh-Guldberg et al., 2018), species-and the services that they provide (Pecl et al., 2017)-are being lost at an unprecedented rate (Barnosky et al., 2012;Pimm et al., 2014). Still, some species can persist or even increase their abundance despite the increasingly frequent and intense disturbance events, as a consequence of global change (Antão et al., 2020;Blowes et al., 2019;van Klink et al., 2020). Such an ability to persist after a disturbance depends, to a large extent, on the species' inherent ability to resist and recover from such events, their resilience (Capdevila, Stott, et al., 2020;Hodgson et al., 2015). Therefore, understanding what makes some species more/less resilient than others is crucial to developing effective management and conservation plans (Pressey et al., 2007). Yet, the lack of data regarding species' natural population's responses to disturbances and robust methods to quantify resilience have hampered

Setting appropriate conservation strategies in a multi-threat world is a challenging goal, especially because of natural complexity and budget limitations that prevent effective management of all ecosystems. Safeguarding the most threatened ecosystems requires accurate and integrative quantification of their vulnerability and their functioning, particularly the potential loss of species trait diversity which imperils their functioning. However, the magnitude of threats and associated biological responses both have high uncertainties. Additionally, a major difficulty is the recurrent lack of reference conditions for a fair and operational measurement of vulnerability. Here, we present a functional vulnerability framework that incorporates uncertainty and reference conditions into a generalizable tool. Through in silico simulations of disturbances, our framework allows us to quantify the vulnerability of communities to a wide range of threats. We demonstrate the relevance and operationality of our framework, and its global, scalable and quantitative comparability, through three case studies on marine fishes and mammals. We show that functional vulnerability has marked geographic and temporal patterns. We underline contrasting contributions of species richness and functional redundancy to the level of vulnerability among case studies, indicating that our integrative assessment can also identify the drivers of vulnerability in a world where uncertainty is omnipresent.

The development ofReimaanlok, a national framework for the planning and establishment of community-based conservation areas in the Marshall Islands, is outlined. A team composed of international experts and local resource management professionals selected and modified an ecoregional planning approach, defined key concepts, selected conservation features and targets, compiled biogeographical information from scientific and local knowledge and carried out a national-level ecological gap assessment. Past development of community-based fisheries and conservation plans was reviewed and the lessons learned informed the development of a robust community-based planning process for the design and establishment of conservation areas on individual atolls, integrating ecosystem based management (EBM) theory, traditional knowledge and management, and the particular socio-economic needs of island communities. While specific geographic, historical, cultural and economic characteristics of the Marshall Islands have created a framework that is unique, several aspects of this process offer ideas for national strategic conservation planning in other Small Island Developing States where there is a paucity of scientific data, significant and increasing threats, and where decision-making about the use of natural resources occurs primarily at the local level.