Introduction
Human activities are causing a global decline in marine biodiversity (Butchart et al. 2010). Local anthropogenic impacts on ecosystems, such as overfishing or pollution (Cinner et al. 2018), combined with global impacts including ocean acidification and climate change (Zhao et al. 2021) degrade ecosystems (De’ath et al. 2012). Coral reefs support most of the diversity of marine life on Earth (Hughes et al. 2002), which translates directly into ecosystem services upon which several billions of people depend in coastal areas (Teh et al. 2013). Fishes represent the main actors of the provision of ecosystem services (Holmlund and Hammer 1999) contributing to biomass production, food security and nutrient cycles, and generating cultural value at the core of activities such as ecotourism (Heyman et al. 2010). The decline of fish threatens tropical reef services (Hughes et al. 2003) and urges scientists, stakeholders, and industries to better monitor the change of fish diversity on tropical reefs to help in conservation and restoration decisions (Obura et al. 2019).
Environmental governance suffers from a long delay between detecting biodiversity decline and implementing conservation measures (Wetzel et al. 2015), a delay that can be shortened by emergent monitoring technology (Polanco Fernández et al. 2021). In coastal marine ecosystems, Underwater Visual Census (UVC) is traditionally used for fish diversity assessments but are time consuming to perform (Colton and Swearer 2010). Additionally, UVCs are likely to miss the most elusive species in need of monitoring for conservation (Boussarie et al. 2018). Genetic technologies, such as eDNA metabarcoding, are rapidly developing and can now identify species assemblages from water samples containing trace DNA from organisms in the environment (Pedersen et al. 2015). When combined with a genetic reference database, eDNA metabarcoding provides an inventory of species composition in aquatic systems that often better recovers elusive and cryptic species of monitoring focus (Deiner et al. 2015; Harrison et al. 2019, Polanco Fernández et al. 2021). Studies of eDNA on coral reefs have shown a strong ability for biodiversity detection showing capacity to match inventories from traditional surveys (Sigsgaard et al. 2019; West et al. 2021; Polanco Fernández et al. 2021). Beyond inventories, eDNA could allow rapid quantification of biodiversity and ecosystem quality indices which, in combination with functional or phylogenetic information, may help monitor shifts in ecosystem processes and states (Holman et al. 2019, Marques et al. 2021).
As eDNA monitoring is sensitive to detect biodiversity responses to environmental gradients, such tools could be deployed to quantify marine biodiversity and deliver overall ecosystem indices to better monitor, manage and conserve ecosystems (Cristescu et al. 2018). Marine eDNA metabarcoding has been shown to discriminate species composition along biogeographic clines (Holman 2021, Polanco Fernández et al. 2021, West et al. 2021), or between different habitats in very localized signals (Jeunen et al. 2019). This method should thus be further able to discriminate assemblage properties in response to anthropogenic stresses (DiBattista et al. 2020). The massive amount of DNA sequence data from eDNA metabarcoding could be compounded into ecological indices, where the cumulated species-specific responses translate into measures of environmental quality (Cordier 2020). Furthermore, by combining with functional traits (e.g. including body size, trophic level) or phylogenetic information (Keck et al. 2018, Marques et al. 2021), eDNA could generate proxies of ecosystem structure and functioning more informative than those from taxonomic lists alone (D’Alessandro and Mariani 2021). The use of functional or phylogenetic indices should be first evaluated along contemporary gradients of anthropogenic pressures before future application in monitoring of assemblages (Carvalho et al. 2020).
Among bioregions with high cover of coral reefs, the Caribbean Sea harbors reefs that are degrading rapidly with a loss of ~ 50% in just four decades as a result of anthropogenic factors (Wilkinson 2000, O’Dea et al. 2020). Coral decline is associated with a marked decrease in biodiversity and shifts in fish composition (Bellwood et al. 2004). If the present trend continues, at least 60% of Caribbean coral reefs could be lost over the next 30 years, motivating data-driven actions for improved monitoring and management (Pittman et al. 2018, Camacho et al. 2020). The decline of coral reefs has been associated with a cumulative set of anthropogenic factors, including poorer water quality from runoff and pollution, damage from tourism overuse, unsustainable fishing and climate change (Duran et al. 2018). With few exceptions (Lester et al. 2020) the lack of monitoring has limited our understanding of the relative effects of those stressors, and this gap could be filled with eDNA monitoring. Curaçao, an Island of the Lesser Antilles, has been known to support a large stretch of among the least degraded coral reefs in the Caribbean (Jackson et al. 2014). However, the decline in reef cover has increased recently because of poorer water quality, the overexploitation of fish populations, unsustainable coastal development, as well as industrial waste issues (Jackson et al. 2014). Along the coast of Curaçao, wide differences in the levels of anthropogenic pressures are nonetheless observed (Waitt institute 2017, de Bakker et al. 2016), which should be associated with contrasted fish assemblage composition either in proximity to dense human settlements or more isolated from human activities.
Here, we investigated the variation in fish taxonomic composition, as well as functional and phylogenetic indices recovered from eDNA metabarcoding along the coast of Curaçao. In particular, we compared two coastal areas with contrasted environmental and anthropogenic conditions: the first a coastal stretch in proximity to the capital, Willemstad, a dense area with nutrient rich water; and a second stretch, more isolated and generally less accessible. In each of these two coastal areas, we collected eDNA samples in 2020, which we further compared with UVCs conducted in 2015. From this collection of data, we asked the following questions: (i) Are there differences in taxonomic ecological and phylogenetic indices between the two areas associated with contrasted environmental conditions? (ii) Do we observe similar species and assemblage composition responses across the two coastal areas recovered from eDNA and UVC? (iii) Do we observe distinct occupancy responses of species in proximity or away from densely populated areas and does it vary across fish families?