AxIOM: Amphipod crustaceans from insular Posidonia oceanica seagrass meadows

Abstract Background The Neptune grass, Posidonia oceanica (L.) Delile, 1813, is the most widespread seagrass of the Mediterranean Sea. This foundation species forms large meadows that, through habitat and trophic services, act as biodiversity hotspots. In Neptune grass meadows, amphipod crustaceans are one of the dominant groups of vagile invertebrates, forming an abundant and diverse taxocenosis. They are key ecological components of the complex, pivotal, yet critically endangered Neptune grass ecosystems. Nevertheless, comprehensive qualitative and quantitative data about amphipod fauna found in Mediterranean Neptune grass meadows remain scarce, especially in insular locations. New information Here, we provide in-depth metadata about AxIOM, a sample-based dataset published on the GBIF portal. AxIOM is based on an extensive and spatially hierarchized sampling design with multiple years, seasons, day periods, and methods. Samples were taken along the coasts of Calvi Bay (Corsica, France) and of the Tavolara-Punta Coda Cavallo Marine Protected Area (Sardinia, Italy). In total, AxIOM contains 187 samples documenting occurrence (1775 records) and abundance (10720 specimens) of amphipod crustaceans belonging to 72 species spanning 29 families. The dataset is available at http://ipt.biodiversity.be/resource?r=axiom.


Introduction
The Neptune grass, Posidonia oceanica (L.) Delile, 1813, is the most widespread seagrass of the Mediterranean Sea. This foundation species forms large meadows that are of crucial ecological and economic importance. Their complex, multi-layered structure offers a suitable habitat to hundreds of animal and plant species, as well as micro-organisms (Buia et al. 2000). In addition, through the epiphytes that grow on all parts of the plants, its dead and decaying tissues and, to a lesser extent, its living tissues, P. oceanica supports elaborate food webs (Vizzini 2009. Thanks to these habitat and trophic services, Neptune grass meadows, which cover up to 50000 km , are biodiversity hotspots in the Mediterranean Sea. Although these meadows are legally protected and included in numerous marine protected areas (MPAs), they are nevertheless threatened by direct and indirect impacts of multiple anthropogenic activities (Giakoumi et al. 2015).
In P. oceanica meadows, amphipods are one of the dominant groups of vagile invertebrates, forming an abundant and diverse taxocenosis (Gambi et al. 1992). They mostly feed on seagrass epiphytes with species-specific dietary preferences (Michel et al. 2015b). Through their feeding activity, they act as ecosystem engineers, as they exert selective top-down control on epiphytic assemblages and modulate nutrient availability for their seagrass host (Michel et al. 2015a). Overall, amphipod crustaceans can be considered key ecological components of the complex, pivotal, yet critically endangered Neptune grass ecosystems. Despite their ecological importance, quantitative and widely available data about amphipod fauna of Mediterranean Neptune grass meadows remain scarce. This is especially true for amphipod assemblages from meadows situated along the coasts of Mediterranean islands, whose structure has recently been showed to differ from their mainland counterparts (Bellisario et al. 2015). In this context, the aim of the AxIOM dataset is to make data collected in the framework of ecological studies freely available on the Global Biodiversity Information Facility (GBIF) portal.

General description
Purpose: AxIOM is a sample-based dataset (n = 187 samples) documenting occurrences of amphipod crustaceans associated to Posidonia oceanica seagrass meadows from Mediterranean Islands (Corsica, Sardinia). In total, it contains 1775 records, documenting occurrence and abundance of 10720 amphipod specimens belonging to 72 species spanning 29 families. Samples were collected over different periods 3 consecutive years, both during the day and during the night. A nested hierarchical sampling design was set up, and multiple sampling methods were combined to ensure a holistic view of the taxocenosis. The dataset package is composed of two data files: one describing sampling events, and the other reporting occurrence data of amphipod crustaceans.

Project description
Title: Multidisciplinary study of trophic diversity and functional role of amphipod crustaceans associated to Posidonia oceanica meadows AND Multiscale variability of amphipod assemblages in Posidonia oceanica meadows: A comparison between different protection levels

Personnel: Loïc N. Michel, Nicolas Sturaro and Gilles Lepoint
Design description: The AxIOM dataset was generated during two doctoral research programmes that took place at University of Liège, Belgium. The first one focused on ecology of amphipod crustaceans from Posidonia oceanica meadows, on their place in the food web and on their role in the ecosystem (Michel 2011). The second one focused on the multiscale variability patterns of amphipod assemblages associated to P. oceanica meadows, and their potential responses among different protection levels (Sturaro 2012).

Sampling methods
Study extent: AxIOM contains 187 sampling events, spanning 3 consecutive years. Sampling took place in two regions: Corsica (Calvi Bay) and Sardinia (Tavolara-Punta Coda Cavallo Marine Protected Area, TMPA). Samples were taken during different periods of the year (November, March, June, July, August) to acknowledge seasonal variation of communities (Gambi et al. 1992). Since amphipod assemblages from Posidonia oceanica meadows also exhibit diel variations (Sánchez-Jerez et al. 1999), samples were taken during both day and night. To ensure efficient and representative sampling of the amphipod taxocenosis, 4 complementary methods were used: hand-towed net, litter collection, air-lift and 2 slightly different types of light traps. In both investigated regions, a nested hierarchical sampling design was setup to fully capture the spatial variability of amphipod community structure over multiple scales spanning four orders of magnitude (1 to 1000 metres; Sturaro et al. 2015). Sampling stations encompass various levels of environmental protection and anthropogenic pressure, including integral reserve (TMPA zone A), partial reserve (TMPA zone B), general reserve (TMPA zone C), unprotected pristine zone (Calvi Bay) and heavily impacted zone (Gulf of Olbia).
Sampling description: Most samples were taken following a nested hierarchical sampling design that focused on variability at 4 spatial scales, ranging from 1 to 1000 metres. In each sampling region (Corsica and Sardinia), zones separated by > 1000 m were chosen. In each zone, 2 sites (separated by ~ 100 m) were picked. Inside each site, 2 or 4 sectors (separated by ~10 m) were randomly selected within each site. Each sector was delimited by a permanent frame circumscribing an area of 9 m . Depending on the method used, sampling events either covered a full sector or were taken randomly inside a sector and separated by ~1 m were collected. Details of the sampling design are given in Sturaro et al. (2015), Sturaro et al. (2014) Levels of this design are documented in the "event.txt" file of the dataset using matching hierarchized parent event IDs.
All sampling was performed by SCUBA diving at depths ranging from 10.4 to 15 metres. Detailed methodology for the hand-towed net (labelled "Net" in the "samplingProtocol" column of the "event.txt" file of the dataset), the air-lift ("Airlift") and the first type of light traps ("Trap1") can be found in Michel et al. (2010).
Litter collection ("Litter") consisted in hand-picking of litter fragments. A 25 x 40 cm quadrate was randomly thrown in the meadow, to estimate sampling area, and all litter present among this meadow patch was handpicked by fistfuls, and quickly placed in a container. By doing so, vagile organisms associated to litter fragments were also collected. This procedure was repeated until a standardized container of 2 litres was filled.
The second type of light traps ("Trap2", Fig. 1) were made of two nested 1 litre translucent plastic containers. The top container was pierced with vertical rectangular slits (1 cm wide x 12 cm long), and was then inserted in the bottom one. Traps were anchored using metal stakes (∅: 3 mm) that were directly stuck in the matte. Each trap presented vertical rectangular slits (1 cm wide x 15 cm long) in its upper part. A diving emergency light stick was fixed in the bottom part of each trap. These sticks emit light for >12 hours, and the vagile invertebrates, attracted by the light, entered the trap through the slits. They gathered in the bottom part, the presence of a bottleneck in the middle of the trap limiting their potential escape. Traps were placed at twilight and recovered the next morning.
Quality control: Sampling protocols were standardized to avoid biases. Amphipods were and identified using primarily the keys of the Mediterranean amphipod fauna of Bellan-Santini et al. (1982), Bellan-Santini et al. (1989), Bellan-Santini et al. (1993), Bellan-Santini et al. (1998) and the interactive key of Myers et al. (2001). In some cases, more recent diagnoses and redescriptions of species were also used. This was notably the case for the genera Apherusa (Krapp-Schickel and Sorbe 2006) and Caprella (Guerra-García and Takeuchi 2002, Krapp-Schickel and Takeuchi 2005, Krapp-Schickel and Vader 1998. After identification, specimens were randomly selected to be re-examined by either first or last author in order to check identification accuracy. Species names were matched against the authoritative, expert-driven World Register of Marine Species (WoRMS).
Step description: After collection, all samples were sieved on 400 µm nylon mesh to eliminate sediment and fine particulate organic matter. They were subsequently fixed for >24 hours in a formaldehyde solution (4% in 0.22 µm-filtered seawater). Samples were then sorted to isolate amphipods and transfer them to a preservation solution consisting of 70% ethanol in distilled water to which 1% glycerine was added to prevent evaporation. After identification, specimens were stored in this preservation solution in airtight vials. Calvi Bay lies in the Ligurian Sea (western Mediterranean), on the north-western coast of Corsica (France; 42°35'N, 8°45'E). It is bound by Punta Revellata Cape in the West, and by Punta Spanu Cape in the East. Temperature of water is typically minimal in February (12°C ) and maximal in August (26°), with a notable vertical thermal stratification from May to September. Salinity of the water of Calvi Bay is around 38 and shows no major seasonal variation. Calvi Bay is an oligotrophic area and shows low inorganic nutrient and particulate organic matter concentrations (Lepoint et al. 2004).
In Calvi Bay, Posidonia oceanica meadows cover 4.94 km i.e. about 50% of the area of the bay. They are found at depths ranging from 3 to 38 m. Meadows mostly grow on soft bottoms and show, in most places, a continuous extension, but local erosion ("intermattes") occurs . Meadows of Calvi Bay are relatively dense, and show an important foliar biomass and production despite the oligotrophic character of the area (Gobert et al. 2003). Overall, the coastal areas surrounding the bay are weakly urbanised and the ecological status of seawater of Calvi Bay is considered as good ).
The Tavolara Zone A (5.29 km ) is an integral reserve and no-take/no-access zones. Access of zone A is restricted to scientists, reserve staff and police authorities. Zone B (31.13 km ) is a partial reserve where access is permitted, but only professional fishermen inhabiting the nearby coastal villages are allowed to fish. Zone C (117.15 km ) is a general reserve where access as well as professional and recreational fishing are allowed under restricted conditions defined by the MPA management consortium.
In TMPA, temperature of water is nearly the same as Calvi Bay, with variation between 14°C and 26°C. Salinity is around 38 and is constant the whole year. P. oceanica meadows cover a total surface of 4415 Ha and are found at depths ranging from 0.5 to 41 m (Tavolara-Punta Coda Cavallo Marine Protected Area management consortium pers. comm.). At sampling depth, shoot density, leaf and epiphyte biomasses do not show differences among protection levels (Sturaro et al. 2014