A total of 39 stations were sampled at eight sites of transitional environments (sensu McLusky and Elliott 2007) in the eastern English Channel along French coasts (Fig. 1). Five sites were harbours and were considered as water bodies heavily modified by human activities (WFD 2000/60/EC), i.e. Caen-Ouistreham (CO), Le Havre (LH), Calais (CL), Boulogne-sur-Mer (BL) and Dunkerque (DK). The three other sites were considered as less impacted sites moderately influenced by human activities (Nasseh and Texier 2000; Poirier et al. 2006; Caplat et al. 2006), i.e. the Bay of Veys (BV), the Orne River (O) and the Authie Estuary (AE). For all sites, the sampling date and a brief description of each station is reported in Table 1. All stations from harbours were sampled in shallow subtidal environment (maximum depth: 18 m) except for all stations in Boulogne-sur-Mer and the station LH1 in Le Havre that were sampled in intertidal environment. All stations of sites considered as less impacted outside harbours were sampled in intertidal environment.

Figure 1. 

Location of the sampling sites (red outline) and position of the sampling stations (black outline) on French coasts of the English Channel. White crosses and circles respectively represent intertidal and subtidal sites. Names of sites are abbreviated such as BV: Bay of Veys, CO (Caen-Ouistreham harbour) O (Orne estuary), LH (Le Havre harbour), AE (Authie estuary), BL (Boulogne-sur-Mer harbour), CL (Calais harbour), DK (Dunkerque harbour).

Environmental parameters were assessed from four replicated sediment cores at each station. Three replicates were dedicated to measurement of Total Organic Carbone (TOC) and one to grain size characterisation. The three replicates for TOC analysis were first frozen, freeze-dried, crushed and pre-acidified. Then, TOC content was determined by high-temperature combustion of dry samples (60 °C, 48 h). Measurements of CO₂ were finally done by thermal conductometry using an elemental analyser (FlashEA, Thermo Electron Corporation). Sediment grain size was obtained by diffraction and diffusion of a monochromatic laser beam on suspended particles (Trentesaux et al. 2001).

Foraminiferal community compositions were assessed from three replicate cores, except in Boulogne-sur-Mer where only one replicate was sampled. The surface sediment (0–1 cm) was sampled from three different deployments with a Reineck corer (160 cm²) for subtidal stations or a handcorer (56 cm²) for intertidal stations. Sediment samples were preserved in ethanol and Rose Bengal solution (2 g L^-1). A total of 107 samples from the 39 stations were used for foraminiferal analysis in this study.

In the laboratory, samples were sieved through a 63µm-mesh and the fraction >63µm was dried at 50 °C. Foraminifera were then concentrated by flotation using trichloroethylene (density = 1.46). For each sample, at least 300 living (stained) benthic foraminiferal individuals were collected and identified to the species level when possible, for both statistical validity (Fatela and Taborda 2002) and representativity of foraminiferal species assemblages (Schönfeld et al. 2012). Among these 300 individuals, Ammonia species were identified morphologically under a stereomicroscope following Pavard et al. (2021). Specifically, Ammonia aberdoveyensis, A. confertitesta and A. veneta were morphologically discriminated following a dichotomous procedure of discrimination based on the pore diameter and the elevation of sutures on the central part of the spiral side (Richirt et al. 2019; Richirt et al. 2021).

For each station, relative abundances (mean ± standard deviation) and absolute abundances (ind 50cm^-2, mean ± standard deviation) were calculated (excluding for the site of Boulogne-sur-Mer, where only one sample was taken). Relative abundances were also calculated at the scale of sites (containing several stations). As number of stations sampled was not constant between sites (e.g. four in the Authie Estuary and seven in Dunkerque harbour) and sampling was not done at the same time (e.g. Boulogne-sur-Mer and Dunkerque harbours), relative abundances data were used instead of absolute abundances to have comparable data. Count, normalised and relative abundance data are available as Suppl. materials 1–3.

The Indicator Value index, IndVal, allows to determine if a species is indicative of a specific habitat (Dufrêne and Legendre 1997). We applied the calculation of IndVal on the normalised abundance dataset to assess if the three Ammonia species investigated here could be typical of stations in harbours or of stations located in less impacted sites outside harbours. IndVal is calculated as:

IndVal_ij = A_ij × B_ij × 100

where IndVal is the Indicator Value (%) of a species i at stations of group j (i.e. either harbour or less impacted), with A_ij = Nindividuals_ij /Nindividuals_i, and B_ij = Nstations_ij / Nstations_j. A_ij is a measure of specificity of species i where Nindividuals_ij is the mean number of individuals of species i across sites of group j. Nindividuals_i is the sum of mean numbers of individuals of species i over all groups. B_ij is a measure of fidelity. Nstations_ij is the number of stations of j group where species i is present and Nstations_j is the total number of stations in that group j. Normalised numbers of individuals (for 50 cm²) were used to calculate A_ij. The IndVal index calculation (iterations: n = 999; package labdsv 2.0-1, Roberts and Roberts 2016) was performed using the software R 4.2-1 (R Core Team, 2022).

To investigate the causal link between the presence of commercial harbours and the occurrence of Ammonia confertitesta, we compiled datasets on its distribution from this study and from previously published distribution of this species in Europe (Holzmann 2000; Langer and Leppig 2000; Hayward et al. 2004; Ertan et al. 2004; Schweizer et al. 2011; Saad and Wade 2016; Chronopoulou et al. 2019; Richirt et al. 2019, 2021; Bird et al. 2020; Francescangeli et al. 2021). We also indicated the distance between the species occurrences and close commercial harbours and their equivalent volume of tonnage of freight (Suppl. material 4).

Total Organic Carbon values ranged from 0.81 ± 0.09% (BV3) to 10.69 ± 0.40% (D1; Table 2). Stations in the Bay of Veys exhibited the lowest values of TOC and ranged between 0.81 ± 0.09% (BV3) and 0.79 ± 0.04% (BV2). Conversely, stations in Dunkerque showed the highest TOC values, ranging from 1.45 ± 0.18% (DK4) to 10.69 ± 0.40% (DK1). On average, TOC values were higher in Dunkerque, Calais, Boulogne-sur-Mer and Le Havre harbours, indicating an enrichment in organic matter (Table 2). It is more contrasted in Caen-Ouistreham harbour where only two stations over six showed values of TOC > 2.0%. In these harbours, the poorest stations in TOC were CO1, CL5 and CL6, all situated at the most opened parts (seaward). Symmetrically, highest concentrations were measured in the most enclosed parts of the harbours, most of the time separated from the open sea by a sluice (e.g. DK1, CL3 and CO5). Sediment samples from harbour sites (i.e. Dunkerque, Calais, Le Havre) generally had a smaller grain size than samples in less impacted ones (Authie, Orne, Bay of Veys; Table 2). Silt content in Dunkerque harbour ranged from 84.3% (DK2) to 95.0 (DK7) and from 93.6% (LH2) to 95.6% (LH3) in Le Havre harbour. Conversely, stations of the Bay of Veys, the Orne and the Authie estuaries exhibited in general a higher sand content, often reaching from 20% to more than 40% in BV4 for example (Table 2). It is more contrasted in stations more exposed to open sea (i.e. CO1 and CL5) which are less muddy than other stations from their sampling sites. Finally, highest content in sand were measured in the Boulogne-sur-Mer harbour with values ranging from 66.4% to 81.3% except for the most external station BL5, which showed a substantially lower value (13.8%).

Among Ammonia species, A. aberdoveyensis dominated in less impacted sites (Fig. 2A), especially in both the Bay of Veys (56.2 ± 27.9%) and the Orne Estuary (68.7 ± 12.0%). In contrast, A. confertitesta dominated Ammonia assemblages in almost all harbours, especially in Le Havre and Caen-Ouistreham where it dominated all stations (Fig. 2). It was more contrasted in Dunkerque, where only three out of seven stations were dominated by A. confertitesta (i.e. DK3, DK4, DK5) and A. aberdoveyensis dominated elsewhere (Fig. 2B). Only few Ammonia individuals were found in DK6 and DK7, and as such those results should be considered with caution (i.e. < 20 individuals per sample; Suppl. material 1). Ammonia veneta was barely found in harbours, with relative abundances ranging from 1.4 ± 2.4% in Dunkerque to 9.4 ± 13.7% in Caen-Ouistreham but showed greater proportions in the less impacted sites of the Bay of Veys (28.5 ± 36.2%) and the Orne Estuary (27.5 ± 11.6%). No Ammonia individuals were found in the Authie Estuary. Finally, over the 35 stations of sites investigated where Ammonia was present, only four stations (i.e. 11.4%) did not contain any Ammonia individuals, three (8.6%) contained only one of the three species, two species co-occurred in four stations (11.4%) and the three Ammonia species co-occurred in 24 stations (68.6%).

Figure 2. 

Maps representing pie charts of proportions of Ammonia veneta (red), Ammonia aberdoveyensis (green) and Ammonia confertitesta (blue) in: A sampled sites B Dunkerque harbour (DK) C Calais harbour (CL) D Boulogne-sur-Mer harbour (BL) E Le Havre harbour (LH) F Caen-Ouistreham harbour (CO) and Orne river (O) and G Bay of Veys (BV). AE: Authie estuary. White crosses represent sites and stations sampled in intertidal environments. White circles represent those which were sampled in subtidal environments.

Only A. confertitesta in harbours (i.e. highly impacted habitats) showed IndVal value greater than 60.0% (86.6%, p < 0.001; Table 3). Concerning the other two species, none of the IndVal values were significant for any habitat.

In the present work, living individuals of the three Ammonia species were present at all sites, except in the Authie Estuary. More specifically, the three species of Ammonia co-occurred in most stations. These results are sharply contrasting with molecular studies that found at most one or two Ammonia species co-occurring in a same station (Hayward et al. 2004; Saad and Wade 2016; Bird et al. 2020; Richirt et al. 2021). The three Ammonia species were recorded occurring together once at Ouistreham (Richirt et al. 2021). Noticeably, when A. confertitesta occurred, it barely co-occurred with other Ammonia species (five out of 29 locations; Bird et al. 2020), which sharply contrasts with our observations. These differences may be related to differences in the method of species identification, i.e. molecular and morphometrical (Scanning Electron Microscope, SEM) vs. morphological (stereomicroscope). Indeed, single cell molecular identification is often based on few individuals, conversely to morphological methods that allow to identify a much larger number of individuals, i.e. tens to thousands. Moreover, the stereomicroscope method, compared to the SEM one, is faster to identify Ammonia spp. individuals and then, could more quickly generate data. In the context of globalisation, where worldwide exchanges increase and ecosystems are rapidly altered, the investigation of more sites and more individuals quickly allows to build distribution maps way easier (i.e. cheaper and less-time consuming) to respectively track NIS and potential global changes associated to anthropogenic impact. However, note that if the proportion of Ammonia species in an assemblage is represented by low count data (i.e. 0 to 20 individuals), the user should be aware of the inherent error rate of the morphological method (Pavard et al. 2021).Therefore, results for DK1, DK6, DK7, CL1, CL2, CL5, CL6, BL5 and BV2 should be taken with caution.

The distribution pattern of A. veneta, which is present at almost all sites over the French coasts of the eastern part of the English Channel, is congruent with its cosmopolitan distribution characteristic (Holzmann and Pawlowski 2000; Hayward et al. 2021). It should nevertheless be stressed that this species was generally present in low abundances and proportions compared to others Ammonia species, with the exception of the intertidal stations in both the Orne River and the Bay of Veys. This intertidal preference is consistent with previous observations reporting this species in intertidal ecosystems both in mud and muddy sand soft-sediments (Saad and Wade 2016).

To date, A. aberdoveyensis has only been encountered in the North Atlantic and the Mediterranean Sea (see review in Hayward et al. 2021). In the present study, A. aberdoveyensis was the major species in the less impacted habitats of the Orne Estuary and the Bay of Veys, but also in a few stations in harbours (DK1, DK2, DK6, CL1, CL2 and LH1). Results in DK6, CL1 and CL2 should, however, be taken with care as this species was represented by only one individual at each of those stations and could be the result of events of passive transport, i.e. wave action and turbulence (Alve 1999). This species was also observed at some subtidal stations, such as in a core sampled at 34 m depth in the Grevelingen lake (Richirt et al. 2020). However, Bird et al. (2020) described A. aberdoveyensis as a species restricted to intertidal environment. In the harbour of Le Havre, while it is dominant in the intertidal station LH1 of the recreational part of the harbour, it was also recorded in the subtidal stations in the international shipping area, suggesting that the species can dwell in both intertidal and subtidal zones.

Ammonia confertitesta is known to have a disjunct geographical distribution between Asia and Europe (Hayward et al. 2004, 2021). In this study, it was the dominant Ammonia species in the subtidal studied harbours. It was only present at low abundances in the intertidal stations of the Bay of Veys and in the Orne Estuary (O1) but with high abundances in all stations of Boulogne-sur-Mer and in the sole intertidal station of Le Havre harbour (LH1). This observation contrasts with previous studies that considered A. confertitesta as an intertidal species (Saad and Wade 2016; Bird et al. 2020), but is consistent with observations of this species from both subtidal and intertidal brackish waters (Bird et al. 2020; Holzmann 2000; Pavard et al. in press; Saad and Wade 2016; Schweizer et al. 2011) to marine subtidal waters (Petersen et al. 2016; Bird et al. 2020; Richirt et al. 2020), confirming the statement of the species being an euryhaline species (Bird et al. 2020; Hayward et al. 2021). This tolerance to a large spectrum of salinity may offer an advantage to colonise different sites as shown in this study. Considering that A. confertitesta is a NIS that may have the potential to become invasive species and replace its congeneric (see Richirt et al. 2022; Pavard et al. in press and the discussion below), A. aberdoveyensis could have found a refuge higher on shores or in the inner part of an estuary, such as in the Orne Estuary, where A. confertitesta is not established yet or do not favour these kinds of environments (Richirt et al. 2021). A similar pattern was previously reported in the Brancaster Staithe high marsh (England; Richirt et al. 2021), in the Verse Meer (The Netherlands; Richirt et al. 2021) and in the Gironde Estuary (France; Pavard et al. in press) where A. aberdoveyensis was the main Ammonia species in higher parts of these systems instead of A. confertitesta, which was more abundant in lower parts. In the Elbe Estuary, A. confertitesta was reported as the dominant foraminiferal species in living assemblages, though it co-occurred with A. veneta instead of A. aberdoveyensis, which was only present in dead assemblages (Francescangeli et al. 2021). In the same study, the authors showed that A. confertitesta replaced the species Elphidium selseyense, in the living assemblages over the last 40 years. This could demonstrate the ability of A. confertitesta to replace other species at a decadal scale.

The latest observations of A. confertitesta in Europe tend to confirm that it is a NIS (reported in Suppl. material 4). Its distribution pattern and its dominance over its congeneric species in international commercial harbours compared to less impacted sites highly suggests that harbours may constitute the main gateway of the introduction of this species in Europe, as previously suggested by Pawlowski and Holzmann (2008). Conversely, in Asia, A. confertitesta preferably occurs in natural sites in Japan (e.g. the Ramsar site of Lake Nakaumi, Toyofuku et al. 2005) and in highly anthropised site in China (e.g. Quingdao Bay, Hayward et al. 2021). Specifically, A. confertitesta was the major Ammonia species in Le Havre, Dunkerque, Calais and Caen-Ouistreham harbours, which are respectively 1^st, 3^rd, 4^th and 13^th French harbours in terms of tonnage of freight in 2019 (https://www.statistiques.developpement-durable.gouv.fr/publicationweb/326?type, https://www.caen.port.fr/le-port-de-caen-1.html) and exhibit an intense international worldwide trade. In addition, most observations of this species in France (Fig. 3, Suppl. material 4), i.e. in the Authie Estuary, the Seine Estuary, the Bay of Aiguillon (Richirt et al. 2021), the Gironde Estuary (Pavard et al. in press) and the Rhone delta (Richirt et al. 2019) were made no more than 35 km away from the international harbours of HAROPA Port, La Rochelle, Bordeaux or Marseille harbours. Likewise, in Great Britain, most observations of this species (Saad and Wade 2016; Bird et al. 2020; Richirt et al. 2021) were in the vicinity of international commercial harbours (Fig. 3). At the continental scale, from the Netherlands to Germany, Denmark or even Sweden, the distribution of A. confertitesta is geographically associated with harbours, which supports its NIS status in Europe. The species has likely been introduced through ballast waters or via hull fouling as already shown for other species of benthic foraminifera (Deldicq 2019; McGann et al. 2019; Bouchet et al. 2023).

Figure 3. 

Map showing occurrences of Ammonia confertitesta in Europe (blue dots: morphological identification; orange squares: molecular identification) and nearest commercial harbours (black crosses) from occurrences. Black circles represent freight of each harbour (in Millions of tons).

Noticeably, some smaller harbours (lower freight tonnage, i.e. Boulogne-sur- Mer, Le Tréport in France, Den Helder in The Netherlands, Husum in Germany, Ahus in Sweden) and/or A. confertitesta occurrences are located rather far from international commercial harbours. These smaller harbours are part of national to regional trade networks and are often connected to international harbours through shipping or smaller boats. The hypothesis of a secondary spread of a NIS (i.e. from international to national/regional harbours) has been well studied (Simkanin et al. 2009; Mineur et al. 2010; Zabin 2014; Costello et al. 2022) and could explain the occurrences of A. confertitesta relatively far from harbours dedicated to international trade. Over time, A. confertitesta could then have spread further from its arrival point following this hypothesis of a secondary spread. In this context, the restricted occurrence of A. confertitesta at the mouth of the Orne River (O) may been surprising given its vicinity next to the Caen-Ouistreham harbour (CO). These observations may, however, correspond to several scenarios. First, this could reflect the ongoing invasion of this species in the Orne River. Another hypothesis relates to the ability of the species to disperse according to both anthropogenic and natural structures of sites (i.e. topography and/or artificial dams). Specifically, stations from these two sites are separated by a sluice which could limit their connectivity. It is also the case in the Grevelingen lake where it is separated from the sea by a sluice (e.g Richirt et al. 2020). Moreover, the tortuosity of the estuary compared to the canal could add another obstacle to the spread of A. confertitesta.

Finally, large proportions of A. confertitesta in stations (up to 87.5%) among autochthonous Ammonia species in assemblages clearly question its invasiveness ability. In the Grevelingen lake, the recent analysis of a sediment core covering the period 1972–2012 indicated that A. confertitesta was recorded from 1986 onwards and was progressively replacing its two congeneric A. veneta and especially A. aberdoveyensis over time (Richirt et al. 2022). In addition, a monthly sampling survey sampling along a transect of the Gironde Estuary has demonstrated the dominance of A. confertitesta among congeneric species over both time and space (Pavard et al. in press), further confirming the hypothesis of its invasiveness. As done for other foraminiferal NIS (McGann et al. 2012; Deldicq 2019; Richirt et al. 2022), the sampling of long sediment cores in sampled sites of this study would allow us to determine when A. confertitesta appeared in species assemblages and eventually when it outcompeted indigenous congeneric species in stations where it was the dominant species. Nonetheless, it would be relatively difficult to sample long-cores in some dynamic systems such as estuaries (e.g. Orne). Finally, a regular monitoring of sampled sites of this study should be done to see if A. confertitesta outcompetes its congenerics over space and time and track its possible ongoing invasion, notably in the less impacted sites, outside harbours.