Introduction

The Wadden Sea stretches from the Netherlands to Denmark and represents, with 4700 km2 emerging tidal flat area during low tide, the largest coherent dynamic intertidal system in the world (Wehrmann 2016). Due to its young post-glacial evolution and complex structure (barrier islands, tidal flats, estuaries, salt marshes), it provides numerous, so far unoccupied niches (Wolff 1999) for species immigrating by expansion or shift of their biogeographic range. Although species dispersal is an ongoing natural process increasing the species richness (Beukema and Dekker 2011), rapid changes in community compositions can have a drastic influence on the balance within this ecosystem. Increasing globalisation, maritime transport and expanded aquaculture favour the unintended introduction of non-indigenous species (NIS) and despite monitoring and management programmes, the rate at which new introductions are observed is rising (Reise et al. 1999; Büttger et al. 2022).

Annually, about two additional NIS are introduced into the German North Sea and a total of 92 macrobiota species were officially registered in this area until today (Büttger et al. 2022). Nevertheless, a complete local extinction of native species was not documented yet and most NIS remain neutral and additive to the ecosystem (Gollasch and Nehring 2006; Buschbaum et al. 2012; Reise et al. 2023). The manifold burdens and benefits and their role in biodiversity changes and ecosystem services are intensely discussed (Schlaepfer et al. 2011; Boltovskoy et al. 2022; Kourantidou et al. 2022). For example, the Pacific oyster Magallana gigas (Thunberg, 1793), which was introduced to the German North Sea in 1991 (Reise 1998; Wehrmann et al. 2000), was suspected to outcompete the blue mussel Mytilus edulis Linnaeus, 1758. It was revealed that although the condition indices of both species were reduced, neither their growth rates nor mortality were affected by their coexistence (Joyce et al. 2021). However, a potential negative impact of NIS cannot be ruled out. Invasive species can alter the biotic community, habitats, and trophic structure (Crooks 1998; Kolar and Lodge 2001; Markert et al. 2010; Büttger et al. 2022); provide a basis for subsequent invasions (Markert et al. 2014); or cause economic damages like the Chinese mitten crab Eriocheir sinensis H. Milne Edwards, 1853 by undermining coastal protection dikes of the estuaries and infesting fish nets and traps (Gollasch 2011). Especially when facing future species distribution shifts and ecosystem changes in the context of the climate crisis and resulting socio-economic changes, monitoring of NIS is inevitable (Essl et al. 2020; Simões et al. 2021).

The majority of NIS in the Wadden Sea originates from the Western Atlantic or Pacific (Wolff 2005; Büttger et al. 2022). This is also the case for the dwarf surf clam Mulinia lateralis (Say, 1822) examined in this study. Its distribution range is originally in the Western North Atlantic from the Gulf of St. Lawrence to the Gulf of Mexico (Walker and Tenore 1984; Montagna and Kalke 1995; Brunel et al. 1998; Turgeon et al. 2009). Since August 2017, it was detected outside its native range along the Dutch coast by Craeymeersch et al. (2019) and Klunder et al. (2019). However, it was initially (mis-)identified as Spisula subtruncata (da Costa, 1778), as this species is native to the North Sea. M. lateralis was able to establish rapidly high densities of up to 5872.4 ind./m2 in the Dutch Voordelta (Craeymeersch et al. 2019) and was recorded annually from 2017 to 2021 at multiple locations (Wood et al. 2022). In Belgium, M. lateralis has been detected at several sites from Knokke to de Panne (Kerckhof 2019; Walles et al. 2020; Wood et al. 2022). Since 2017, M. lateralis has also been recorded in the German part of the Ems Dollard estuary (Klunder et al. 2019; G. Scheiffarth pers. comm.). Additional isolated records were reported further east at the JadeWeserPort (IfAÖ 2020), near Tossens, and in the inter- and subtidal around the Island of Sylt (K. Reise pers. comm.; U. Schückel pers. comm.).

The aim of the present study is (i) to test the hypothesis of a stepwise spread of the Mulinia lateralis larvae by anti-clockwise residual tidal currents as known from previous invasions (Brandt et al. 2008; Markert et al. 2014) by (ii) documenting the initial distribution based on population parameters (i.e., abundances, length-frequency distributions, biomasses) from a systematic survey along the coast which allows (iii) to define the present status of this non-native species in the central Wadden Sea.

Material and methods

To document the present occurrence status of Mulinia lateralis along the central Wadden Sea coast, we developed a survey design, which allows easy access to the potential intertidal habitats. As known from some other non-native species (e.g. Magallana gigas, Hemigrapsus takanoi Asakura & Watanabe, 2005), invasion into the central Wadden Sea is often driven by an eastward drift of larvae due to residual tidal currents (Brandt et al. 2008; Markert et al. 2014). Therefore, sampling was carried out between February and May 2022 along 10 shore-normal transects (T1 to T10) in the Lower Saxon Wadden Sea National Park distributed regularly from the outer Ems River estuary in the west to the outer Elbe River estuary in the east (Fig. 1; Table 1).

Fig. 1
figure 1

Location of the surveyed transects in the central Wadden Sea and position of previous findings (asterisks) of Mulinia lateralis since August 2017 (Craeymeersch et al. 2019; Klunder et al. 2019; IfAÖ 2020; G. Scheiffarth pers. comm.). Greenish: above mean high-water; yellowish: intertidal; bluish: subtidal

Full size image
Table 1 List of transects sampled during initial monitoring of Mulinia lateralis. Start of each transect is close to the local high-water line whereas the end of each transect is approximate to the respective low-water line
Full size table

Salinity in the study area covers a broad range from 13 to 22 in the outer Ems River estuary (T1 and T2) over 28 to 31 in the central tidal basins (T3), 30 to 34 in the Jade channel (T7 and T8) to 16 to 22 in the outer Elbe River estuary (T10) (Reineck and Flemming 1990; Becker 1998; Kaiser and Niemeyer 1999; Glorius et al. 2021). The tidal range of the semidiurnal tide varies between 2.9 m in the outer Ems River estuary (T1), 2.5 m in the backbarrier tidal flats (T3), 3.8 m in the Jade (T8), and 2.9 m in Cuxhaven (T10) (BSH 2022).

Each transect was 1 km in length covering 40 sampling stations of 1 m2, each 25 m apart from the other, starting close to the mean high-water line of the mainland coast following the topographic gradient down close to the low-water line. Bathymetric data along the transects were taken from the 2016 digital terrain model (10 m grid length) of the EasyGSH-DB platform (https://mdi-dienste.baw.de/geoserver/EasyGSH_Bathymetrie/wms; https://doi.org/10.48437/02.2020.K2.7000.0002). Bathymetric data refer to the standard elevation datum NHN. The top sediment layer (uppermost 3–4 cm) inside the 1 × 1 m-frame was sieved through a 1mm-sieve in the field. All specimens of M. lateralis were collected and stored for 1–5 days at 4 °C until further identification and measurements in the lab. Shell length was measured with a digital caliper rule (accuracy: 0.01 mm) from the posterior margin to the anterior margin. Live-wet weight (i.e. shell, flesh and mantle water) was measured with a KERN EW 220-3NM scale (accuracy: 1 mg). The specimens were identified using a Leica M205 C stereomicroscope with the identification key provided by Craeymeersch et al. (2019). All specimens were preserved in 96% denatured Ethanol. The occurrence data were summarised and uploaded as a new dataset to the open source platform EurOBIS (Gismann et al. 2023, https://doi.org/10.14284/602).

Statistical analyses were calculated in RStudio (R Core Team 2022, version 4.2.1; RStudio Team 2022) using the package ggplot2 (Wickham 2016).

Condition indices were calculated separately for all sampled specimens of M. lateralis (n = 897; Fig. 5), with the formula (Jakob et al. 1996; Marzec et al. 2010; Smith et al. 2000):

$$a=\frac{W}{{L}^{b}},$$
(1)

with W as the measured live-wet weight and L as the shell length. The exponent b is the slope calculated in the power relationship between live-wet weight and shell length (Fig. 5). To test the condition indices between the transects, the Kruskal-Wallis test (1952) and post hoc Dunn test (1964) were performed, using the Benjamini-Hochberg correction (1995).

The bimodality of the length distribution was analysed in two ways, using the function is.bimodal() available in the package LaplacesDemon (Statisticat 2021) and the function bimodality_coefficient() of the package mousetrap (Wulff et al. 2021).

In addition to the characteristic morphological features, we confirmed the species determination using DNA-based analyses of three specimens from each of the transects T1, T2 and T4 to T9 and two specimens from T3. The targeted COI gene region was amplified by polymerase chain reactions (PCRs) in Mastercycler pro S thermocycler (Eppendorf, Hamburg, Germany) with a final reaction volume of 20 μL, comprising 10 μL of Accustart II PCR SuperMix, 0.5 μL of each species-specific primers Mul2L (5′-TTATTCGAATGGAGTTAACATC-′3) and Mul1R (5′-GAACCTCTTTCCGCATAGGT-′3; Hare et al. 2000), 8 μL of ultrapure water, and 1 μL DNA extract. The PCR reactions were performed by an initial denaturation at 94 °C for 3 min, followed by 38 cycles consisting of 30 s of denaturation at 94 °C, 45 s of annealing at 45 °C, and 1 min of elongation at 72 °C, and a final extension for 3 min at 72 °C. PCR products were Sanger sequenced by Macrogen Europe (Amsterdam, Netherlands) in both directions. Resulting sequences were processed in Geneious™ (version R7.1.9.) and quality checked using the BLAST tool of NCBI (Altschul et al. 1997).

Results

Morphological and genetic identification of Mulinia lateralis

Following the identification key provided by Craeymeersch et al. (2019), we could exclude other species of the family Mactridae occurring in the NE Atlantic Ocean with similar morphological characteristics (most similar to Spisula subtruncata (da Costa, 1778)). The following morphological features were decisive for the identification of M. lateralis (Fig. 2): (1) the triangular shell outline; (2) the distinct radial ridge along the posterior end of the valves; (3) the ligament was exclusively internal; (4) accessory lamella well developed; (5) anterior lateral teeth in the right valve of different sizes, the ventral one longer; two posterior lateral teeth similar in size; (6) shell colour whitish to cream; (7) shell surface smooth; (8) shell distinctly convex and (9) the cardial area between beaks was broad in large specimens.

Fig. 2
figure 2

Morphological identification characteristics of Mulinia lateralis; a Inside left valve with posterior lateral tooth (PLT), anterior lateral tooth (ALT), cardinal teeth (CT), and accessory lamella (AL); b Inside right valve with internal ligament (IL). Scale bar for a and b 5 mm; c Inside view with pallial sinus and pallial line; d Outside view with posterior radial ridge (RR). Scale bar for c and d 10 mm

Full size image

All sequenced specimens matched 100% with prior uploaded sequences from Klunder et al. (2019) (MN207099.1; MN207097.1) and with sequences submitted by the Smithsonian Environmental Research Center (KT959410.1; KT959381.1). Sequence vouchers of specimens investigated in this study are publicly accessible through GenBank (Accession No. OP575829 to OP575848).

Abundance and spatial distribution

In the sampling area of 392 m2, 897 specimens of M. lateralis were found at 9 of 10 transects (Table 1; Fig. 3). No specimens of M. lateralis were found in the easternmost transect T10. The majority of individuals (86%) was found in three transects (Fig. 3), namely T1 (n = 343), T5 (n = 144), and T6 (n = 284). The highest mean abundance was recorded in T1 (10.72 ± 9.78 ind./m2), followed by T6 (7.10 ± 5.96 ind./m2) and T5 (3.60 ± 4.29 ind./m2). Recorded abundances in T2, T4, and T8 were lower, ranging between 0.63 ± 1.27 ind./m2 and 1.48 ± 1.50 ind./m2. Few individuals were found in T3, T7, and T9, resulting in a mean abundance of 0.1 ind./m2 or less (Table 1; Fig. 3).

Fig. 3
figure 3

Abundance of Mulinia lateralis along the elevation gradient per transect (T1 station 1–32; T2–T10 station 1–40), n is the number of individuals per transect

Full size image

M. lateralis is mainly distributed in an elevation range from + 0.35 m to  − 0.40 m NHN (Fig. 4). Specimens in transects T1, T4, T5, and T8 are distributed in a depth range from  + 0.25 m to − 0.43 m NHN, with the highest abundances at a depth below 0 m NHN (Fig. 3). In transect T1, the abundance increases with depth, with an average abundance of 30.25 ± 4.03 ind./m2 between  − 0.33 m (station 29) and  − 0.39 m NHN (station 32). In T2 and T6, the individuals are distributed in shallower depth ranges of  + 0.52 m to  − 0.24 m NHN (Fig. 3). All individuals found in transect T9 (n = 4) and two individuals in T7 (n = 3) are distributed in an elevation range from  + 0.44 m to  + 0.26 m NHN. The third individual in T7 was found at a higher elevation of  + 0.94 m NHN.

Fig. 4
figure 4

Histogram of absolute frequency of Mulinia lateralis per elevation range

Full size image

Population structure

In total, the specimens of M. lateralis ranged in size between 3.98 and 23.55 mm. 3.57% (n = 32) of the individuals found in this study exceed the previously known maximum size of 21.2 mm (Craeymeersch et al. 2019). Live-wet weight ranged between 0.016 and 3.115 g. No pattern in length distribution along the elevation gradient was observed (Fig. 1 supplementary material).

Shell length and live-wet weight show a power relationship (Fig. 5). The log–log relationship is linear (linear model, R-squared: 0.95, p-value: < 2.2e-16) with log(intercept) =  − 7.874 and slope = 2.846. This results in the formula:

$$W= {e}^{-7.874}*{L}^{2.846},$$
(2)

with W as live-wet weight and L as shell length.

Fig. 5
figure 5

Correlation of shell length (L) and live-wet weight (W) of Mulinia lateralis. Black dots depict specimens of transects T2 to T9, red dots depict specimens of transect T1

Full size image

The condition indices calculated with the Formula (1) range from 0.51 × 10−4 in T6 to 7.18 × 10−4 in T8 (Fig. 6). The Kruskal-Wallis and the following post hoc Dunn test of the condition index per transect confirm a significantly lower condition index in T1 (\(\widetilde{x}\) = 3.52 × 10−4) than every other transect tested (Figs. 5 and 6). Transect T8 (\(\widetilde{x}\) = 4.16 × 10−4) shows a significant higher condition index than T1, T5, and T6 (Fig. 6).

Fig. 6
figure 6

Boxplots showing the condition indices (a) for transects with n ≥ 25 individuals (T1, T2, T4–T6, and T8). T1 differs significantly from every other transect. T5 and T6 differ significantly from T8

Full size image

The probability density function on shell length for all sampled individuals is bimodal, suggesting the presence of two size classes. Bimodality is supported by the function is.bimodal(), but did not reach the threshold criterium for the bimodality coefficient according to Pfister et al. (2013).

Separate testing of bimodality in T1, T2, T4, T5, T6, and T8, with n ≥ 25, reveals a bimodal shell length distribution fulfilling the threshold criterion of at least one bimodality test at every transect except T4. The density curves of transects T1, T2, T6, and T8 are similar, with a first increase in density ranging from 8 to 14 mm shell length and a second increase at a shell length of 16 to 20 mm. The bimodal distribution in transect T5 is shifted towards higher shell lengths, with a first peak at a shell length of 12.5 to 17.5 mm and a second peak at 20 to 22.5 mm (Fig. 7). The density curve of T4 shows a single peak at a shell length of 15 mm to 20 mm (Fig. 7).

Fig. 7
figure 7

Shell length density functions of Mulinia lateralis per transect, with n ≥ 25 individuals (T1, T2, T4–T6, and T8)

Full size image

Discussion

Mulinia lateralis occurs in high densities in its native range along the North American East Coast (Flint and Younk 1983; 74022 ind./m2 in the Hillsborough Bay, USA (Santos and Simon 1980); 63168 ind./m2 in the intertidal area of the estuarine Wassaw Sound Bay, USA (Walker and Tenore 1984)), as well as in the non-native European coastal waters (5872 ind./m2 in the sublittoral of the Dutch Voordelta (Craeymeersch et al. 2019)). Craeymeersch et al. (2019) documented lower densities in the intertidal, ranging from 2.4 to 9.8 ind./m2, with exception of the intertidal in the Dutch Westerschelde (820.0 ind./m2).

This study confirms the occurrence of M. lateralis from the outer Ems River estuary up to the Dorumer Watt at the outer Weser River estuary. The maximum abundance of 36 ind./m2 was recorded at station 31 (− 0.37 m NHN) in transect T1 in the outer Ems River estuary, exceeding the maximum abundance of 9.8 ind./m2 known so far for this area (Craeymeersch et al. 2019) and the German North Sea.

The maximum shell length of 23.55 mm recorded in this study exceeds previous reports. Say (1822) first described a maximum shell length of 12.7 mm. Others state a maximum length of 20 mm (Calabrese 1969; Zettler and Alf 2021). Craeymeersch et al. (2019) reported a maximum shell length of 21.2 mm within the Dutch Wadden Sea area. In our study, 3.57% (n = 32) of all individuals investigated exceed the currently known maximum size of 21.2 mm (Craeymeersch et al. 2019).

Based on a life expectancy of 2 years known for M. lateralis (Calabrese, 1969), it can be assumed that the bimodal distribution of the length frequency (Fig. 7) shows the presence of two cohorts. This leads to the assumption that M. lateralis is able to survive and successfully reproduce.

Differences in condition indices between transects may be influenced by the time of sampling (Marzec et al. 2010). Over the past 50 years, there has been a consistent rise in monthly mean temperatures during all seasons in the Wadden Sea (Beukema et al. 2009; van Aken 2008). In cold-blooded animals, metabolism is regulated by outside temperatures. During a mild winter, metabolic rates are high and food availability low as the main food source, unicellular phytoplankton, is dependent on light. This results in a negative energy balance and greater weight loss (Beukema 1992). Therefore, the significantly lower condition index in T1 could be due to greater weight loss during winter. Sampling started in February (T1) and March (T2, T4, T6, T7), and lasted until April (T3, T5) and May (T8–T10). Following previous studies, the growing season coincides with the phytoplankton cycle from April to September (Dörjes 1992; Ramón 2003). Since the other transects (T2–T10) were sampled later in the year when food availability is greater, animals in these transects had time to regain the weight lost during the winter, resulting in a higher condition index compared to T1. This is supported by the finding of the highest condition index sampled in transect T8 (Table 1; Fig. 6). The sampling time may influence the shift in shell length distribution between the transect (Fig. 7).

Spatial differences in abundance may be dependent on hydrodynamics and sediment composition (Rosenberg 1995; Kröncke 2006; Schückel et al. 2013). Strong currents prevent larvae and small juveniles from settling and thus prevent a successful recruitment. Higher abundances of M. lateralis are reported in sandy mud and mud compared to coarser sediment (Walker and Tenore 1984; Klunder et al. 2019). Since no data about the sediment structure and hydrodynamics was collected, no statement about its influence on the recorded abundances can be made.

M. lateralis tolerates a wide range of salinity (Parker 1975) and occurs in mixohaline waters (Brunel et al. 1998; McKeon et al. 2015). In the Dutch Wadden Sea area, no distinct effect of salinity on the distribution of M. lateralis has been reported (Klunder et al. 2019). This can be supported by our study where the salinities of the transects in which M. lateralis was found are in a broad range from brackish (salinity 13) to normal marine (salinity 34; Reineck and Flemming 1990; Becker 1998; Kaiser and Niemeyer 1999; Glorius et al. 2021).

As known from some other NIS (e.g. Magallana gigas, Hemigrapsus takanoi), the spreading into the central Wadden Sea is often driven by an eastward drift of larvae due to anti-clockwise residual tidal currents (Brandt et al. 2008; Markert et al. 2014). Decreasing abundance from T1 (outer Ems River estuary; 10.72 ± 9.78 ind./m2) to T9 (outer Weser River estuary; < 1 ind./m2) and the absence of M. lateralis in transect T10 (outer Elbe River estuary) clearly indicate an eastward spread.

It cannot be excluded that the low abundances in T7, T9, and T10 are due to the sampled elevation range (+ 1.75 to + 0.21 m NHN) not covering the entire main distribution range of M. lateralis (+ 0.35 to − 0.40 m NHN; Fig. 3).

High densities and high fecundity (Santos and Simon 1980; Flint and Younk 1983; Walker and Tenore 1984; Lu et al. 1996; Craeymeersch et al. 2019), the ability to reproduce several times a year (Calabrese 1970), and the ability to survive high environmental adversity (Parker 1975) indicate a high potential for M. lateralis to become invasive (Craeymeersch et al. 2019; Gittenberger et al. 2019; Klunder et al. 2019), meaning that by definition their introduction threatens the biological diversity of the ecosystem (Büttger et al. 2022). M. lateralis is sensitive to competition (Parker 1975; Klunder et al. 2019) but known to increase rapidly in abundance in areas that have recently been affected by a disturbance event (e.g. dredging (Kaplan et al. 1974)). Construction work in the Ems River estuary led to tidal amplification, increased fine sediment transport (van Maren et al. 2015) and, thus, may have allowed M. lateralis to quickly multiply. Flint and Younk (1983) observed a similar rapid recolonisation of M. lateralis after a disturbance event in the Corpus Christi Bay (Texas). M. lateralis was able to recolonize the disturbed area faster than the competing species. With the recurrence of the competing species, the abundance of M. lateralis decreased. As M. lateralis is similar to the native cockle Cerastoderma edule (Linnaeus, 1758) with respect to size, habitat preference, feeding mode, and food source, a competition is most likely. However, Craeymeersch et al. (2019) noted a disadvantage of M. lateralis in prolonged phases of starvation, making M. lateralis a weak competitor.

The assumption of IfAÖ (2020), Zettler and Alf (2021), and Wood et al. (2022), that M. lateralis is already or will become established in the German Wadden Sea, can be confirmed by this study. Whether the species will become invasive cannot be assessed at this early stage of bioinvasion.

Conclusion

After the first findings of the non-native bivalve Mulinia lateralis in the westernmost part of the central Wadden Sea in 2017, we expected an eastward directed dispersal due to larval drift by residual tidal currents. Therefore, we designed a survey of 10 shore-normal transects distributed in tidal flats of the central Wadden Sea between the outer Ems River estuary in the west and the outer Elbe River estuary in the east to document the actual invasion status.

The survey took place from February to May 2022. In total, we sampled 897 living specimens of M. lateralis from 392 sampling stations (each 1 × 1 m2) which reflects a mean abundance of 2.3 ind./m2. Highest abundance was observed at a station in the most western part with 36 ind./m2. There is also a clear trend of high abundances in the mid-tidal level (+ 0.35 m and  − 0.40 m NHN; Fig. 4). The shell length varies between 3.98 and 23.55 mm, which exceeds the so far known maximum shell length of 20–21 mm. M. lateralis was absent in the most eastern transect. Bimodality of shell length distribution indicates the presence of at minimum two cohorts within the population.

Besides the characteristic morphological features, the presence of M. lateralis was also confirmed by DNA-based analyses where all sequenced specimens matched 100% with prior uploaded sequences.

Except for the most eastern transect, M. lateralis was found at each transect of the central Wadden Sea which supports the hypothesis of an eastward directed dispersal by larval drift. Due to the maximum lifespan of 2 years based on earlier studies, the evidence of a high number of small individuals, and the continuous distribution, reproduction of M. lateralis must have taken place in the German North Sea. As some findings were recently also reported from the northern part of the Wadden Sea, the present status of this non-indigenous species can be classified as established.

Further studies should focus on detailed population and reproduction dynamics, the genetic diversity of the newly established population, and the competitive traits to native species like Cerastoderma edule with respect to space and food.