The benthic mussel stages can be dispersed over long distances, while attached to ship hulls or floating substrata. Alternatively, larvae can be transported by water currents and in the ballast and bilge water of ships and vessels. To gain insight into dispersal potential and habitat suitability, survival of the benthic stages of two invasive dreissenid species (chosen based on their occurrence in salinity gradients) were tested for their tolerance to salinity. They were exposed to various salinities in outdoor mesocosms during three long-term experiments (Van der Gaag et al. 2016, 2017, 2018). Mussel survival was studied without prior acclimation, reflecting conditions they experienced when they are attached to ship hulls while travelling along a salinity gradient from fresh or brackish water to seawater, or vice versa. Initially, mussels react to salinity shock by temporarily closing their valves, suspending ventilation and feeding. This cannot be maintained for long periods and adaptation to higher salinity must eventually occur. In our tests, bivalve survival was monitored till the last specimen of a test cohort died. The survival periods were compared with the duration of actual ship voyages to estimate the real-world survival potentials of species dependent of salinity changes, travel distances and durations. Salinity shocks during the trip will be survived within the favourable salinity range but these species tolerate only for a few weeks the unfavourable salinity range. Variability in salinity of dispersal corridors may function as a barrier for the spread of species. At faster and more frequent shipping in the future, salinity can become less important as a dispersal barrier (Van der Gaag et al. 2016).

Genetic information should be considered with respect to environmental tolerances such as salinity. Recently, it became clear that with respect to the COI gene of the invasive populations of M. leucophaeata in Europe, only one haplotype was involved. This haplotype is the same for the populations along the temperate east Atlantic coast of North America (Heiler et al. 2010; Fernandes et al. 2018, 2022). This relatively cold-tolerant lineage may be presenting a distinct species that originated in the mid-eastern or north-eastern USA and deviating because of past glaciations or the formation of the Labrador Current. This haplotype has colonised Eurasia via harbours by the shipping trade, which function as the hubs for further dispersal (Fernandes et al. 2018). The haplotype can be considered a subspecies or a species separate from M. leucophaeata for which in both cases the name cochleata is available as described by Kickx in Nyst (1835). So, we established the salinity-temperature tolerance of this (sub) species.

Populations of D. polymorpha in the Netherlands resembled genetically most those of invasive populations in other Eurasian countries such as Poland, Spain, and Russia (Volga River). Its sister group consists of populations from the Danube River, Hudson River, and the Erie Canal (Stepien et al. 2014).

Mytilopsis leucophaeata cochleata and D. polymorpha show a different occurrence with respect to salinity. To understand and predict their colonization, dispersal, population development and survival, it is important to know their salinity tolerances. To get information about their salinity limits in the field in relation to other environmental factors several approaches can be used varying from short to long-term observations and measurements in the field, transplants in cages and settlement plates at various locations. To get a better understanding controlled outdoor and indoor mesocosms and aquarium experiments were carried out at varying combinations of conditions (Van der Gaag et al. 2016 and literature therein).

Data on species occurrence in salinity gradients in estuaries or canals connecting seaways to freshwater harbours can also provide valuable information on their salinity tolerance (Wolff 1969; Strayer and Smith 1993). Therefore, the distribution of two invasive dreissenids (M. leucophaeata cochleata and D. polymorpha) has been studied in a salinity gradient within the North Sea canal (Noordzeekanaal, the Netherlands).

The present study aims to test the reliability of the experimentally derived salinity-tolerance limits for these species (Van der Gaag et al. 2016), using available data on their distribution in the salinity gradients of the North Sea canal and elsewhere in the Netherlands to get insight in the environmental conditions for co-existence and sympatry, niche segregation and invasiveness of these mussels and to predict their occurrence (Van der Velde et al. 1998).

The distribution related to salinity of M. leucophaeata cochleata and D. polymorpha was studied in the North Sea canal (Province of North-Holland) in the Netherlands. This canal connects the Amsterdam harbours with the North Sea (Figs 2, 3).

The North Sea canal opened in 1876 for shipping after 11 years of implementation. The construction of this waterway started with 7 km digging through the coastal dunes near Velsen. At the east end the Oranjesluizen (Oranje sluices) were built in 1872 and 17 km of dykes constructed in the lakes ‘Wyker meer’ and ‘IJ meer’. This resulted in several land reclaims and a canal with a depth of 7.5 m and a bottom width of 27 m (Van Haaren and Tempelman 2006). After several enlarging and deepening works, the North Sea canal is today 15–18 m deep and 170 m wide (Zindler et al. 2004). The distance along the canal from the sluices at IJmuiden to the city of Amsterdam is indicated as kilometres (km). The 21 km western part of the canal (km 0–21) is called the North Sea canal and the 7 km eastern part (km 21–28) is the river IJ. The IJ in Amsterdam is 10–12 m deep and about 100 m wide (Van Haaren and Tempelman 2006). The banks along the canal mainly consist of Belgian and German rock armour, sheet piles (harbours) and reed beds (some side canals).

The main function of the canal is transport by shipping from the harbours of Amsterdam to the North Sea and vice versa. Large seagoing ships can use all harbours along the canal up to the Mercurius harbour that is located at 21.5 km from the sea and 3 km from the city centre of Amsterdam. In the canal yearly 45,000 ship movements were recorded (Rijkswaterstaat 2023 https://www.rijkswaterstaat.nl/water/waterbeheer/bescherming-tegen-het-water/waterkeringen/dammen-sluizen-en-stuwen/sluizencomplex-ijmuiden#feiten-en-cijfers). Other functions of the canal are the discharge of water and sediment and use of cooling water by power stations and industries. The yearly water discharge through the sluices is 4.6 billion m³ (Rijkswaterstaat 2023 https://www.rijkswaterstaat.nl/water/waterbeheer/bescherming-tegen-het-water/waterkeringen/dammen-sluizen-en-stuwen/sluizencomplex-ijmuiden#feiten-en-cijfers).

The sampling sites of M. leucophaeata cochleata and D. polymorpha in the North Sea canal, the river IJ and the harbour of IJmuiden are situated in the littoral zones along the main channel of the canal (Fig. 3, Table 2). Sampling sites 1 and 2 are situated outside the sluices (seawater) and, sampling sites 10 and 11 are in the IJ (Amsterdam). All the other sampling sites are situated along the borders of the North Sea canal.

The harbour of IJmuiden is located West of the sea sluices, where seawater (normally with a salinity of 30–35) is mixed with brackish water from the canal resulting in a salinity of 17 measured at low tide. The freshwater lake IJmeer is situated east of the Oranje sluices (Fig. 3). The North Sea canal receives freshwater from the River Rhine through the Amsterdam-Rhine canal. The surface water of the North Sea canal is clearly a mixture of seawater from the North Sea and discharged freshwater originating from the River Rhine and surrounding polder areas.

The North Sea canal is a water system with salt stratification. When ships pass the sluices, seawater penetrates the canal. This results in a salt wedge, which moves over the bottom in an eastern direction. The water with a lower salinity (brackish water) and lower density flows over the salt wedge in the direction of the sea. The salt wedge reaches until the threshold of the emergency barrier separating the Amsterdam-Rijnkanaal from Het IJ/North Sea canal. Both water layers poorly mix so that a salt jump develops at a certain depth.

By maintaining a minimal flow rate of 10 m³.s^-1 in the Amsterdam-Rijnkanaal the salt wedge of the Amsterdam-Rijnkanaal is banned. The result of the combination of a salt wedge flowing to the east and the brackish upper layer flowing to the west is a salt gradient in horizontal as well as vertical direction. The mean residence time of the salt wedge is 40 days and for the freshwater layer 20 days. The flow rate is < 0.2 m.s^-1 and the mean water drainage 85 m³.s^-1 (Van Beersum et al. 1993). Water from the Amsterdam-Rijnkanaal, lake IJmeer and the polder waters are directed towards the North Sea. By these flow rates a relatively brackish, well mixed upper layer (epilimnion) develops with a thickness of 5–6 m with a decreasing thickness from west to east. An intermediary layer with a thickness of 8 m develops, with a linearly increasing salinity (metalimnion) and a basic layer of 1–2 m with a constant salinity, of which the thickness increases and the salinity decreases from west to east (hypolimnion) (Karelse and Van Gils 1991). At each lock sea water enters the canal reinforcing the salt wedge (ten times per day for the Noordersluis). Now and then the salt wedge enters the Amsterdam-Rijnkanaal under the influence of water flow and wind.

In October 1989, we collected one to three basalt stones in the littoral zones of the North Sea canal at seven sampling sites (km 0.4, 2.5, 6.5, 10, 15, 19.5, 24) alongside the canal for the sampling of M. leucophaeata cochleata and D. polymorpha (Fig. 3). Mytilus edulis could be collected only from the basalt stones at the seaside of the sluices in IJmuiden at low tide (Van der Velde et al. 1998). In April, June and September 1990, four to five sampling sites alongside the canal have been sampled and in June 1991 only km 24. All sampled stones were transported to the laboratory in plastic bags, provided with the number of the sampling location and were stored at 4 ⁰C until analyses. Subsequently, all mussels present on the stones were removed and counted. The sampling area was measured in cm². The shell length and height of each individual of M. leucophaeata cochleata and D. polymorpha were measured with an accuracy of 0.05 mm. For practical reasons the minimum length of the specimens used for size measurements was 1 mm. The density of each species and the ratio between the numbers of individuals of both species were calculated.

Salinity of the upper water layer was measured with an YSI salinity meter in the North Sea canal and water temperature was measured with a mercury thermometer at each sampling date and site. To study the seasonal variability of salinity and water temperature of the North Sea canal, monthly measurements were carried out near the Municipality of Velsen (km 2.5) for four years. All other salinity values of the North Sea canal used were derived from chloride data of Van Couwelaar and Van Dijk (1989). Salinity data of the canal through Voorne data were derived from chloride values mentioned in Janssen and Janssen-Kruit (1967). Salinity was calculated by multiplication of chloride content values (g.l^-1) with 1.80655 (Salinity = 1.80655 * Cl) (Wooster et al. 1969).

SigmaPlot for Windows (Version 11.0 (Build 11.0.0.77) Copyright © 2008. Systat Software Inc., 1735 Technology Drive Suite 430 San José, CA95110, U.S.A.) was used for statistical analyses and graphical representation of data.

A large population of M. leucophaeata cochleata is still present in the North Sea canal and the connected river IJ (Figs 3, 4). The ships in the canal produce waves and currents resulting in a supply of suspended organic matter to mussels in the littoral zone. Power plants and other industries use water of the canal for cooling purposes and discharge cooling water with temperatures above ambient conditions. The higher water temperatures in mixing zones of these thermal discharges are beneficial for the reproduction and growth of mussels. In cooling water systems with water intake from the canal, the higher temperature and the continuous flow of food stimulate the growth of the mussels that settled in the condenser tubes causing blockage of the water flow (Jenner et al. 1998).

At km 0.4, 2.5, 4.5, 6.5, 10, 15, 19.5, and 24 M. leucophaeata cochleata was found and, at km 19.5 and km 24 this species was found to co-exist with D polymorpha (Tables 2, 3). At km 19.5 one specimen of D. polymorpha was found in October 1989 but none in 1990. At km 28 only D. polymorpha was found. Dreissena polymorpha co-existed with M. leucophaeata cochleata at the farthest end of the canal with most Rhine water influence (Fig. 5). The occurrence of D. polymorpha was restricted to a salinity below 4 (Fig. 6). At km -1.5 and km -0.1, outside of the sea sluices, no dreissenid mussels were observed (see below Discussion Salinity tolerance).

The salinity measurements in the North Sea canal showed a gradual decrease from km 0–km 25 (Fig. 5). Table 4 summarizes the yearly mean, maximum, minimum and median water temperature of sampling site km 2.5 (municipality of Velsen).

Monthly, salinity and water temperature in the period 1989–1992 were measured in the North Sea canal at km 2.5 (Velsen). Salinity fluctuated between 3.0 and 9.5 (Fig. 7), while the temperature course is predictable with a maximum in June–October (20–24 ⁰C) and a minimum in December–February (4–8 ⁰C).

Mytilopsis leucophaeata cochleata was present over nearly the whole length of the North Sea canal with the highest densities close to the sea sluices where also the highest salinities and water temperatures were measured. The densities in the canal decreased gradually at larger distances from the sea (Fig. 5) where salinity was lower (Fig. 6).

The occurrence of co-existing D. polymorpha and M. leucophaeata cochleata was studied at km 24 in the period October 1989–June 1991. The density of M. leucophaeata cochleata was on average 2.4 times higher than that of D. polymorpha (Table 5). Further eastward at ‘freshwater’ sites only D. polymorpha was found.

From 1895 until 1932, M. leucophaeata cochleata expanded its range in the Netherlands through the brackish water systems in the province of North-Holland and from the coastal areas near Hellevoetsluis to Groningen (Van der Gaag 2021). Most of the populations in the province of North-Holland disappeared due to a decline in water salinity after the closure of the Zuiderzee, leading to the development of the freshwater lake IJsselmeer in 1932. In 1932, in the Oosterdock (part of the harbour of Amsterdam) Vorstman performed a study with slide glasses, to get insight in the life cycle of M. leucophaeata cochleata (Vorstman 1933a, b, 1934) demonstrating that this area was still a brackish water system.

The delta works in the south-western part of the Netherlands resulted in the closure of estuaries by dams and other storm surge barriers. These flood defence structures changed brackish estuaries into freshwater lakes causing the local extinction of several populations of M. leucophaeata cochleata in the province of Zuid-Holland (e.g., the population in the canal through Voorne by the closure of the Haringvliet estuary in 1971) (Fig. 4). Most of the populations in the provinces of Friesland and Groningen also disappeared before the 21^th century due to a strong decline of the brackish water systems by increased inlet of fresh water for agricultural purposes. Although considered an alien species the high local extinction rate led to putting M. leucophaeata cochleata on the red list of vulnerable species (Gmelig Meyling and De Bruyne 2022).

An established population of M. leucophaeata cochleata is still present in the canal from Ghent to Terneuzen in the south of the Netherlands (Kuijper 2000). A small population occurs in the Eems canal and the Oosterhornhaven in Weiwerd near Delfzijl (Kuijper 2002) in Groningen. The species also occurs in the Van Harinxma canal at Harlingen in the north, in the eastern part of the Western Scheldt near Rilland in the south and in a water body near Den Helder in North-Holland (De Bruyne et al. 2013) (Figs 2–4).

Janssen and Janssen-Kruit (1967) qualitatively collected dreissenid mussels at seven locations with different salinities in the littoral zone of the canal through Voorne.

The canal through Voorne between the river Brielse Maas and Haringvliet estuary was digged out by hand and finished in 1829 (Figs 2, 3). The Haringvliet was at that time in open connection with the North Sea. The canal has a width of 30 m, a sluice width of 10 m and a depth of 5.3–5.6 m, which was large enough for most ships at that time (Benschop 2023). The canal through Voorne had connections by the sluice with the Brielse Maas and with the Haringvliet via a sluice at Hellevoetsluis. By the construction of this canal the sailing time for ships from the North Sea to the Rotterdam harbour was since 1829 shortened half a day. The canal lost its function after the construction of the Nieuwe Waterweg between the North Sea and the Port of Rotterdam in 1872. The Haringvliet was at that time in open connection with the North Sea. In 1960 the canal through Voorne was dammed of in the north (Wikipedia 2022). The Haringvlietdam was finished in 1971 changing the Haringvliet into a freshwater body. The desalination of the canal through Voorne resulted in the disappearance of M. leucophaeata cochleata after 1971 (Fig. 4). A salinity gradient was present from the northern sluice (km 0), which connected the canal to the freshwater of the Brielse Maas, and the southern sluice (km 9), which connected the canal to the seawater of the Haringvliet. Salinity data of the first nine months of 1965 were delivered as chloride content by Rijkswaterstaat. Their qualitative sampling approach allowed to calculate the ratio between the number of individuals of both species. In the low salinity part of the canal, a few M. leucophaeata cochleata were found and many D. polymorpha. In the high salinity part of the canal the opposite was found. In the middle of the canal both species were present in equal amounts as salinity values of the canal through Voorne were low (Table 2).

In 1988, Stichting Ecotest commissioned by the department Water Management of Rijkswaterstaat Directorate Noord-Holland investigated the littoral fauna of the North Sea canal, side channels and harbours (Van Couwelaar and Van Dijk 1989). They found populations with a density higher than 200 mussels.m^-2 for M. leucophaeata cochleata between km 5–km 25. For D. polymorpha they found a density of more than 200 mussels.m^-2 at km 25–km 28, low numbers at km 22 and very low numbers from km 7.5–km 13.5 (Tables 2, 3). In the side channel C, near the municipality of Spaarndam (Fig. 3), D. polymorpha had high numbers too, because of fresh water discharged by the sluice of the Spaarne (Tables 2, 3). In the period 1989–1990, we found high numbers for M. leucophaeata cochleata from km 0–km 25, high numbers for D. polymorpha at km 25 – km 28 and low numbers at km 19.5 (Table 3). In the western part of the North Sea canal (the deep part with the big harbours) the salinity (3–8) is too high for the settlement of a permanent D. polymorpha population with high density. Mytilopsis leucophaeata cochleata is more tolerant to salinity (2–8) and showed high densities from km 0–km 25. In the most eastern part of the canal, near the Oranje sluices, the salinity (1–2) is too low for a viable population of M. leucophaeata cochleata. In 1989 and in 1990, both species had well established populations. They co-existed at km 25 at the north side of the canal just opposite of the central station of Amsterdam as also found by Van Couwelaar and Van Dijk (1989). De Bruyne and Neckheim (2001) mentioned a record of co-existence of both species near the quay De Ruyterkade in Amsterdam in 1951 at a salinity of 2.4. Walton (1996) mentioned coexistence of both species in the Hudson River, New York where D. polymorpha occurs at salinities of 0–3 (none found at salinities of 5–9). All these data are in agreement with each other and fall within the favourable ranges found by our mesocosms study (look under the heading Salinity tolerance below).

Mytilopsis leucophaeata cochleata was also found in fresh water outside the brackish water areas in the large rivers Rhine, Meuse and Scheldt (Fig. 4). Jaeckel (1962) mentioned it even stream upwards, from the inner Rhine-Ruhr harbour of Duisburg (Germany). Because their shells often carry growth of brackish water species such as bryozoans and tube worms which live in the North Sea canal area but lack in fresh water, they must have been transported by ships most likely through the Amsterdam-Rhine canal to the Rhine and there became detached from the ship hulls. Adult M. leucophaeata cochleata can survive in fresh water for a long time (120–150 days in mesocosm and transplantation experiments), but reproduction was not observed under such conditions (Verhofstad et al. 2013; Van der Gaag et al. 2016).

In the past, M. leucophaeata cochleata occurred on piles and wooden and stony bank protection in the docks of Amsterdam, as these harbour areas were still connected with the Zuyder Sea (Zuiderzee) and sea water penetrated into the canals of Amsterdam and other waters around Amsterdam. In 1932 the Zuiderzee was closed by a dam, and changed very soon in a freshwater lake, called lake IJsselmeer. This led to the disappearance of M. leucophaeata cochleata in the period 1935–1940 by freshening of the water in the canals of Amsterdam, except the IJ. In the Zuiderzee, M. leucophaeata cochleata was lacking and this species did not invade the newly created lake IJsselmeer in spite of a short brackish phase after the closing. Dreissena polymorpha invaded the whole lake (Van Benthem Jutting 1943) profiting from freshening of the IJsselmeer and availability of shells as a substrate due to mass mortality of marine bivalves such as Mya arenaria L., 1758 (Smit et al. 1993) (Figs 2, 4). Elsewhere, freshening of the water caused by dams in river estuaries by the Delta works and the influence of the enclosed freshwater IJsselmeer on salinity of surrounding water bodies led to the disappearance of M. leucophaeata cochleata at many localities where it was recorded before (Fig. 4, blue squares) in contrast to D. polymorpha. Mytilopsis leucophaeata cochleata was found in the Netherlands in brackish water but not in the so-called shock habitats where the salinity changes from sea to fresh water and daily fluctuates with the tidal regime (Wolff 1969). This species occurs where seasonal fluctuation of salinity is predictable allowing to fulfil its life cycle. Therefore, it favours brackish canals and harbours.

Salinity tolerance of Mytilopsis leucophaeata cochleata and D. polymorpha was studied in long-term studies in outdoor mesocosms during 1992 – 1995 by which favourable and unfavourable salinity ranges for the survival of full-grown mussels could be distinguished for the mussel species occurring in the North Sea canal (Van der Gaag et al. 2016). The results of the outdoor salinity tolerance experiments allowed us to distinguish favourable (f.: high tolerance) and unfavourable (u.: no or low tolerance) salinity ranges for each species, viz. for D. polymorpha 0.2–6.0 (f.), 7.0–30.0 (u.), for M. leucophaeata cochleata 0.2–17.5 (f.), 20.0–30.0 (u.). At the unfavourable salinities, those mussels died within 14 days of initial exposure. The maximum duration of survival of single specimens of D. polymorpha was 318 days at a salinity of 3.2, and 781 days for M. leucophaeata cochleata at a salinity of 15.0 (Van der Gaag et al. 2016).

Favourable salinity ranges measured by adult survival time in the mesocosms, appeared to be 0.2–17.5 for M. leucophaeata cochleata, and 0.2–6.0 for D. polymorpha (Van der Gaag et al. 2016). The salinity range based on the distribution of M. leucophaeata cochleata in the North Sea canal was 1.5–9.2, 1.5–3.3 for D. polymorpha, and 13.0–16.2 for Mytilus edulis. Van Couwelaar and Van Dijk (1989) found in the North Sea canal for M. leucophaeata cochleata 1.0–4.2 and D. polymorpha 1.0–3.5. Janssen and Janssen-Kruit (1967) found for M. leucophaeata cochleata in Canal through Voorne 0.2–2.8, and for D. polymorpha 0.2–2.8. Van Couwelaar and Van Dijk (1989) found this sympatric occurrence in the North Sea canal at a salinity of 1.0–3.5. In our study this was recorded for a salinity range of 1.5–3.3 (Tables 2, 3). All these salinity values are within the favourable ranges of both species, such as determined in the long-term survival experiments in the mesocosms.

New invading mussels in the North Sea canal and the Netherlands are the brackish water Hooked mussel or Bent mussel Ischadium recurvum (Rafinesque, 1820) (Mytilidae) since 2012 (first record in Europe), originating from the Gulf of Mexico (Goud et al. 2019) and the freshwater Quagga mussel (Dreissena rostriformis bugensis (Andrusov, 1897) (Dreissenidae) since 2006, originating from the Ponto-Caspian area (Bij de Vaate 2006).

Ischadium recurvum shows increased mortality below a salinity of 6 and needs a minimum salinity of 4.5 (Allen 1960). The species tolerates a maximum salinity of 36 (Gulf of Mexico salinity) (Goud et al. 2019). Maximum salinity for occurrence of Dreissena rostriformis bugensis is 4 (Mills et al. 1996). Maximum salinity for reproduction is 2 (Wright et al. 1996).

The occurrence of these newcomers means that competition between four mussel species is likely when salinity tolerances are similar but excluded when salinity tolerances are different. It is already observed that D. rostriformis bugensis outcompetes D. polymorpha under freshwater conditions (Bij de Vaate et al. 2014; Matthews et al. 2014; D’Hont et al. 2018). Besides new species in the North Sea canal also new constructions were made leading to opportunities for alien and native species. In 1993–1994 the Seaport Marina of IJmuiden was constructed and is with the harbour of Scheveningen nowadays one of the few seawater harbours along the Dutch coast. The Seaport Marina is constructed at the southern part of the old inner pier of IJmuiden outside the sluices. Other opportunities for marine species are provided by the harbour near the steel fabric Tata Steel at the northern side of the canal. Both can act as a hub for the dispersal of newly introduced species. Since January 2022, at opening of the new lock (Zeesluis IJmuiden), the biggest lock in the world, each time 10,000 ton sea water is entering the North Sea canal unlike the old lock (Noordersluis) with 6,000 ton sea water. Construction of a dam (2022–2024) with at the bottom a gate in the inner blow down channel (Binnenspuikanaal) will made it possible to drain the salt sea water to the North Sea. In the case that the salinity is rising too high when the salt wedge is moving too far inland, and salt water is entering the Amsterdam-Rhine canal the old lock will be used (Rijkswaterstaat 2023). The biological consequences of these hydrological changes have to be studied as this situation creates opportunities for establishment of new alien mussel species and other biofouling species. Studies on the interactions between these new invaders and existing populations in the North Sea canal are urgent.