Biological invasions are considered one of the main threats to the marine environment and are a key driver of biodiversity change (Ruiz et al. 1999; Streftaris et al. 2005; Zorita et al. 2013; Jimenez et al. 2018). Non-indigenous species (NIS) have become ubiquitous across coastal ecosystems, such as estuaries, bays, and coastal lagoons (Ruiz et al. 2009), because these sheltered zones are subject to high levels of anthropogenic pressure and activity (Ruiz et al. 2000). The adverse impacts of NIS are multiple and cross-disciplinary, spanning economic sectors—such as fisheries, aquaculture, and tourism (Galil et al. 2014)—and ecological dimensions, including single-species interactions, population declines, local extinctions, changes in community composition, and effects on entire ecosystem processes and functions (Katsanevakis et al. 2014; Anton et al. 2019).

Despite the growing knowledge of NIS impacts, this information is not yet adequately addressed in European Union legal frameworks for the protection and management of marine and freshwater ecosystems. Although the temporal occurrence, abundance, and spatial distribution of NIS are indicators (descriptor D2) used to assess the Environmental Status (ES) of marine waters under the Marine Strategy Framework Directive (MSFD 2008), no specific consensus-based tools for measuring the magnitude of bioinvasion impacts have been formally adopted. Likewise, the indices proposed to assess the Ecological Quality Status (EQS) of transitional and coastal waters under the Water Framework Directive (WFD 2000) do not address the impacts of NIS (Zaiko and Daunys 2015).

One of the potential ecological hazards posed by the introduction of NIS that should be considered in assessment tools and by decision-makers is the displacement of native species (Truhlar et al. 2014). This can occur for numerous reasons, such as restricted mobility during competition for available space (Shinen and Morgan 2009), direct predation (Battini et al. 2021), ecosystem engineering by NIS (Rilov et al. 2024), parasitism and diseases (Stiger-Pouvreau and Thouzeau 2015), or a high potential for food competition between native species and NIS (Comeau et al. 2015; Nielsen et al. 2017; Sonier 2017; Rodrigues et al. 2023). Clearance rate (CR), defined as the volume of water cleared of particles per unit of time, may provide insight into potential trophic competition between suspension-feeding bivalves. An extensive body of literature on CR spanning several decades (Møhlenberg and Riisgård 1978; Iglesias et al. 1998; Troost et al. 2009; Yu et al. 2017) has revealed the sensitivity of CR values to several factors, such as temperature (Pernet et al. 2008; Cranford et al. 2011), salinity (Casas et al. 2018), distinct experimental setups (Riisgård 2001), food quantity and quality (Navarro and Iglesias 1993), and contamination levels at sampling sites (Widdows 2001). To date, these experiments have been conducted primarily under laboratory conditions (Møhlenberg and Riisgård 1978, 1979; Denis et al. 1999; Nakamura 2001; Haure et al. 2003; Troost et al. 2009; Joyce et al. 2019) or using artificial diets (Navarro and Widdows 1997; Petersen et al. 2004; Pascoe et al. 2009; Nagasoe et al. 2011; Tamayo et al. 2013). Nevertheless, laboratory CR results may deviate from in situ measurements (Cranford 2001; Petersen et al. 2004; Clavier and Chauvaud 2010; Rosa et al. 2018). Therefore, different species should be compared within the same trophic competition study (Troost et al. 2009), with the animals maintained under the same experimental conditions and preferably grazing on the full diversity of seston naturally available in seawater.

Despite the extensive research on CR, several terms (e.g., filtration rate, retention efficiency, pumping rate, and capture efficiency) have been used interchangeably (Rosa et al. 2018; Steeves et al. 2022), leading to confusion. The term retention efficiency is commonly used to refer to the proportion of particles that encounter the gill filaments and are cleared from the water, and it must be determined before measuring a bivalve’s CR (Cranford et al. 2016). Following the recommendation of Rosa et al. (2018), “retention efficiency” is hereafter referred to as capture efficiency (CE). This is because the number of particles retained by the gills can be accurately determined only using direct in vivo approaches (Shimeta and Jumars 1991; Ward et al. 1998; Rosa et al. 2018).

One of the most exploited native bivalves in Europe is the common cockle, Cerastoderma edule (Magalhães et al. 2018). This species occurs in coastal lagoons and estuaries along the Atlantic Ocean, from the Barents Sea to Mauritania (Freitas et al. 2014), and is known to be crucial for ecosystem functioning, providing a wide variety of services, namely carbon sequestration, food provision, and biodiversity enhancement (Carss et al. 2020).

The Portuguese oyster, Magallana angulata, was previously considered native to the northeastern Atlantic Ocean; however, in recent decades, several studies have suggested that this species is either a close phylogenetic relative of or identical to the Pacific oyster, Magallana gigas (Foighil et al. 1998; Yu et al. 2026). One proposed hypothesis is that the species was unintentionally introduced into Portugal during the 16^th century through shipping routes originating from Asia (Buroker et al. 1979). Despite this, M. angulata is still managed as a native aquatic living resource by the competent Portuguese authorities (e.g., Decree-Law No. 92/2019 on the prevention and management of the introduction and spread of invasive alien species), resulting in a contentious and ambiguous scenario for the management of this species. As oyster taxonomy was not the primary focus of this study, the Portuguese oyster is considered a non-indigenous species in Portugal, taking into account the most recent scientific findings.

The non-indigenous Manila clam, Ruditapes philippinarum, is a bivalve native to the Indo-Pacific Ocean that was first reported in Portugal in 1984 after being introduced for aquaculture purposes (Ruano and Sobral 2000). Its presence across European countries has been reported to affect biogeochemical cycles and macroalgal growth (Bartoli et al. 2001) and to cause reductions in both the density and spatial extent of native bivalves (Pranovi et al. 2006; ICES 2008). Furthermore, Maia et al. (2025) suggested that R. philippinarum holds a competitive advantage over native sympatric species, not only because of its rapid growth and early sexual maturity but also because of its broad diet and flexibility in feeding on multiple food sources according to their availability in the water column.

The purpose of the present study was to assess the potential trophic competition between native and non-indigenous bivalves feeding on natural particles. The specific objectives were to (i) compare the CE of one native species (the cockle, C. edule) and two non-indigenous species (the Portuguese oyster, M. angulata, and the Manila clam, R. philippinarum), (ii) estimate their CR under field and laboratory conditions, and (iii) provide data to competent authorities and regulators for potential management measures regarding the aforementioned NIS.

Despite the generally similar mean shell lengths across experiments within each species, the mean AFDW of the body tissue was higher in the field experiments, regardless of species (Table 1). Considering this outcome, no direct comparisons between the field and laboratory experiments were performed.

At the onset of the experiments, the average concentration of particles present in the seawater was estimated for each particle size (Fig. 4). Particle concentrations in the field were, on average, 6–11 times higher than those in the laboratory. The maximum average number of particles per mL was recorded for the 4–5 μm size class (14,853 ± 5,229 particles mL⁻¹, n = 64, and 2,184 ± 967 particles mL⁻¹, n = 108, for the field and laboratory experiments, respectively), whereas the minimum was observed for particles within the 13–14 μm size class (2,017 ± 760 particles mL⁻¹, n = 64 and 188 ± 118 particles mL⁻¹, n = 108, for field and laboratory, respectively).

In the field, particle size had a significant effect on CE regardless of species (see Suppl. material 5). CE was positively related to particle size for all three species (Fig. 5A), with smaller particles (4–8 μm) being associated with lower CE values than larger particles (8–14 μm). Specifically, particles belonging to the 4–5, 6–7, and 7–8 μm size classes showed significant differences compared with particles from the 8–13 μm size classes (see Suppl. material 6).

CE results under laboratory conditions (Fig. 5B) were consistent with those obtained in the field (see Suppl. material 5): CE was unaffected by species under laboratory conditions. Moreover, as found in the field, particle size significantly affected CE, with the 4–5 and 11–12 μm particle-size classes showing significant differences relative to most of the remaining size classes, indicating that their CE values diverged more broadly across the comparisons (see Suppl. material 7). Overall, for both native species and NIS, CE tended to increase with particle size.

No significant differences (p = 0.47 and p = 0.59 for the field and laboratory experiments, respectively) were found in the average CR_ind among species (Fig. 6A).

Taking the AFDW (g) of each specimen’s body tissue into account, C. edule exhibited a significantly higher CR_AFDW (p < 0.05) than M. angulata (C. edule: 0.97 L h⁻¹ g⁻¹ and M. angulata: 0.33 L h⁻¹ g⁻¹), whereas R. philippinarum (CR_AFDW: 0.49 L h⁻¹ g⁻¹) showed no significant differences compared with either of the other species in the field experiments (Fig. 6B). In the laboratory experiments, C. edule once again exhibited the highest CR_AFDW value (2.50 L h⁻¹ g⁻¹), showing statistically significant differences (p < 0.001 and p < 0.01) relative to both M. angulata (0.79 L h⁻¹ g⁻¹) and R. philippinarum (1.16 L h⁻¹ g⁻¹).

No statistically significant relationships were observed between CR and CI measured in either the field or laboratory experiments for any species (see Suppl. material 8).

Field and laboratory experiments were conducted using the natural seston present in seawater and therefore provide a good demonstration of which particle sizes are naturally available in the coastal environment. Despite the particle supply before the start of the experiments being higher in the field than in the laboratory, the number of particles available in seawater was highest for the smallest size class measured (4–5 μm) and decreased exponentially with increasing particle size, as reported for other locations (Strohmeier et al. 2012).

Previous studies on various bivalve species have reported that CE generally increases in a nonlinear, plateau-like fashion with the diameter of the available particles, particularly those equal to or larger than 4 μm (Stuart and Klumpp 1984; Jørgensen 1989; Langdon and Newell 1990). Despite the difference in particle concentrations between experiments, CE was dependent on particle size regardless of the experimental conditions.

Although previous research attempted to replicate in the laboratory the feeding conditions present in the natural environment by adding cultured microalgae and sediment particles, this approach did not provide an accurate depiction of field conditions (Kiørboe and Møhlenberg 1981; Bricelj and Malouf 1984; Bayne et al. 1989; Navarro et al. 1992). Stenton-Dozey and Brown (1992) and Stenton-Dozey and Brown (1994) went further using natural food and incorporating the respective temporal variability into their experiments, although their trials were still conducted in the laboratory. In the present study, the results revealed that the CE responses of Cerastoderma edule, Magallana angulata, and Ruditapes philippinarum were similar under field and laboratory conditions: the three species tended to capture particles larger than approximately 8 μm more efficiently. This outcome is consistent with the findings of Iglesias et al. (1992), who tested the selective preference of C. edule for particles within the range of 2–12 μm. Likewise, Arakawa et al. (1998) reported a capture rate of 60% or higher for particles between 2 and 25 μm for R. philippinarum, although this study did not rely solely on natural water. Additionally, previous research conducted with Ruditapes decussatus, a species that is very closely related phylogenetically, concluded that this species efficiently retained particles from 6 to 9.6 μm, with 9.6 μm being the maximum dimension assessed (Sobral and Widdows 2000). To the best of knowledge, there are no previous studies focusing on the CE or retention efficiency of M. angulata. In the present study, M. angulata had the highest CE for larger particles (i.e., above 8 μm), which is consistent with the trend observed for both Magallana gigas (Marks et al. 2024) and Crassostrea virginica (Haven and Morales-Alamo 1970).

Higher CE does not necessarily imply higher CR because the relationship between CE and pumping rate depends on several factors, such as particle size or gill structure (Steeves et al. 2022). The average CR_ind observed per individual was very similar across experiments and among species, with the non-indigenous M. angulata recording a slightly higher value (0.74 L h⁻¹ ind⁻¹ and 0.95 L h⁻¹ ind⁻¹ for the field and laboratory experiments, respectively). No previous research targeting the individual CR of this species was found; however, Casas et al. (2018) reported CRs of 1.82 L h⁻¹ ind⁻¹ and 4.95 L h⁻¹ ind⁻¹ during winter and summer, respectively, for C. virginica. Even though the approach was similar to that used in the current study (static method), the experiments described above were conducted using a mixture of artificial seawater and six microalgal species, which may explain the discrepancy in the results. C. edule presented an average CR_ind of 0.62 L h⁻¹ ind⁻¹ and 0.79 L h⁻¹ ind⁻¹ for the field and laboratory experiments, respectively, which is lower than other values ranging from 1.00 to 4.00 L h⁻¹ ind⁻¹ reported in the literature (Møhlenberg and Riisgård 1979; Urrutia et al. 1997; Widdows and Navarro 2007). However, the latter studies used either a different methodology to measure CR (i.e., flow-through) or cultured microalgae added to the bivalves’ diet. Lastly, the non-indigenous R. philippinarum recorded marginally lower CR_ind values (field: 0.52 L h⁻¹ ind⁻¹ and laboratory: 0.72 L h⁻¹ ind⁻¹). As with the oyster, no previous research reporting individual CR values was found for the Manila clam, although Sobral and Widdows (2000) reported a higher value (1.30 L h⁻¹ ind⁻¹) for R. decussatus clearing microalgae added to seawater. It is important to note that the use of cultured algae leads to higher CRs than seston-based diets (Cranford et al. 2011).

Determining individual CRs is crucial for estimating the ecological carrying capacity of a specific ecosystem (Cranford 2019), because abundance data are generally the most commonly available type of data collected in ecological studies. Furthermore, even when biomass data are available for the ecosystem, they take into consideration the total weight of the organisms, including the shell, whereas CR per gram is frequently calculated using the ash-free dry weight of the organism’s body tissue. This may lead to an overestimation of the clearance time (CT; the time needed for the bivalves to clear all the water in the ecosystem) and, consequently, an overestimation of the clearance ratio, which is the ratio between CT and the water residence time (RT). Such a situation could have severe implications when establishing aquaculture production limits because, with a clearance ratio (CT/RT) < 1, bivalves are considered to exert pressure on pelagic communities through grazing because the water is cleared faster than it is renewed (Smaal and van Duren 2019).

On the other hand, when the aim is to identify or quantify the potential impacts of NIS on native species by directly comparing particular species, individual CRs are not very useful, especially for species with such divergent shell lengths, such as oysters and clams. In this case, CR_AFDW is the most appropriate measure. The native C. edule presented the highest CR_AFDW (field: 0.97 L h⁻¹ g⁻¹ and lab: 2.50 L h⁻¹ g⁻¹) regardless of the experiment. These values are consistent with findings in the literature for this species, which report CR_AFDW values ranging from 1.00 to 3.58 L h⁻¹ g⁻¹ (Smaal et al. 1997; Karlsson et al. 2003; Cranford et al. 2011; Nilin et al. 2012; Ong et al. 2017). M. angulata, in turn, showed the lowest values of CR_AFDW (field: 0.33 L h⁻¹ g⁻¹ and laboratory: 0.79 L h⁻¹ g⁻¹). The limited research conducted on this species and the large variation in the results, ranging from 0.009 L h⁻¹ g⁻¹ to 7.68 L h⁻¹ g⁻¹ (Mathers 1974; Haure et al. 2003; Yu et al. 2017), hamper a straightforward comparison, particularly because none of the studies were conducted under conditions similar to those of the present study. Likewise, low CR_AFDW values were also observed for R. philippinarum (field: 0.49 L h⁻¹ g⁻¹ and laboratory: 1.16 L h⁻¹ g⁻¹), which are within the range of values reported in the literature for this species (Goulletquer et al. 1989; Nakamura 2001; Koo and Seo 2020; Jiang et al. 2023). It is also noteworthy that the interspecies CR_AFDW patterns were similar under both experimental settings.

The high variability in the values of both CR_ind and CR_AFDW reported in the literature can be explained by differences in experimental setup and methodology (Cranford et al. 2011), as well as other factors known to influence bivalve filtration activity, such as temperature (Goulletquer et al. 1989; Tamayo et al. 2013), salinity (Casas et al. 2018; Barr et al. 2023), current velocity (Sobral and Widdows 2000; Karlsson et al. 2003; Fernandes et al. 2007; Widdows and Navarro 2007), contaminants (Nilin et al. 2012), the type and amount of food supplied (Prins et al. 1991; Iglesias et al. 1996; Barillé et al. 1997), acidification (Ong et al. 2017), and hypoxia or anoxia (Sobral and Widdows 1997). In fact, particle concentration is known to affect CR, with multiple studies showing that higher seston concentrations lead to lower CRs (Iglesias et al. 1992, 1996; Navarro and Widdows 1997). Seston diversity in coastal seawater is known to shift across seasons, which would suggest that different bivalve species might benefit more at particular times of the year. In this study, however, such season-specific advantages were not evident. Seasonal variability was inherently incorporated into the experimental design (field experiments were conducted in summer, whereas laboratory experiments were performed in winter), and although the absolute magnitude of CR_ind and CR_AFDW differed between experiments, the relative pattern among species remained consistent. This indicates that species-specific feeding performance was robust to the seasonal changes in seston composition captured. It should be noted that the exclusion of inactive bivalves from the estimates may result in values that reflect the maximum potential CR. This approach was applied deliberately to ensure that the calculations excluded individuals with closed valves as a response to handling stress, although it might also have excluded specimens undergoing natural periods of inactivity as part of the regulation of feeding activity (Riisgård et al. 2003). Despite the fact that bivalves must maintain a minimum valve opening to allow particle entry and gas exchange (O₂–CO₂), this intuitive expectation of a positive relationship between gape and feeding activity is rarely supported empirically. It is often assumed that wider valve opening reflects higher pumping activity; however, studies have shown that gaping rhythms correlate only weakly with physiological processes. Maire et al. (2007), for example, reported very low correspondence between valve gape and pumping rate in Mytilus galloprovincialis, particularly under starvation. Their work further demonstrated that the surface area of the exhalant siphon opening is a far more reliable predictor of pumping rate, a conclusion consistent with observations by Newell et al. (2001) for M. edulis. These findings indicate that physiological processes such as pumping depend more on the activation of ciliary beating than on the extent of valve opening itself. Consequently, establishing a direct, generalizable relationship between gaping behavior and feeding physiology remains challenging, and any meaningful interpretation must account for the species-specific and context-dependent nature of these interactions.

Conducting experiments under laboratory conditions to study CR provides essential observations on the physiology of bivalves by controlling each variable and evaluating specific responses (Cranford et al. 2011). Nonetheless, the effects of temporal changes due to tidal influence, seasons, and other variables on food quantity and quality can be replicated only in field studies (Cranford et al. 2011). In the absence of field data for a certain region, models can estimate carrying capacity using CR values reported in the literature from laboratory trials, despite the associated limitations (Gray et al. 2022). Results from different regions may vary not only because of geographic differences in the species-specific shell lengths of bivalves or site-specific food resources but also because individual CRs tend to be higher than the average CR in a bed due to the potential refiltration of seawater in the environment (Troost et al. 2009). Thus, direct comparisons between native species and NIS should be performed within the same study and under similar conditions, as demonstrated by several studies conducted in recent years (Troost et al. 2009; Galimany et al. 2017; Joyce et al. 2019; Rodrigues et al. 2023).

The fact that the non-indigenous M. angulata and R. philippinarum did not exhibit greater efficiency in selecting and retaining particles within the range of 4–14 μm compared with the sympatric native species (C. edule) suggests that other biological traits may explain their establishment success. In fact, Cardeccia et al. (2018) assessed the traits of the 68 most widespread non-indigenous species in European seas, including R. philippinarum, and determined that dispersal ability and reproductive rates were among the most relevant traits.

Although the native C. edule was the most efficient species in clearing particles from seawater (i.e., on the basis of tissue weight), NIS nonetheless remains a threat because their collective abundance may exert notable pressure on the ecosystem’s carrying capacity. Indeed, filter-feeding bivalves are recognized as impacting the availability of food resources to other species and directly affecting phytoplankton communities by reducing their abundance and altering their species composition (Prins et al. 1991; Rahman et al. 2020). Monitoring studies of bivalve communities conducted in the last decade have shown that R. philippinarum and M. angulata are dominant species in the Tagus and Sado estuaries in terms of both abundance and biomass (Cabral et al. 2020; NIPOGES 2022). This dominance is consistent with the historical role of M. angulata as the main species targeted by the aquaculture industry in the Sado Estuary (Biguino 2020). Considering that the aquaculture production area in Portugal increased by nearly 20% between 2020 and 2021 (INE 2023) and that the recently approved National Plan for Aquaculture in Transitional Waters allows a fourfold expansion of the areas designated for aquaculture (RCM76 2022), a resource-limited scenario might be forthcoming, in which both M. angulata and R. philippinarum may contribute to the potential collapse of the pelagic food web (Brito et al. 2023). Moreover, it has already been demonstrated that R. philippinarum competes with native species such as Ruditapes decussatus and Cerastoderma glaucum for food resources in the Tagus Estuary because the three species rely on the same particulate sources (Dias et al. 2019). Pranovi et al. (2006) also suggested that the decreasing density of both R. decussatus and C. glaucum in the Venice Lagoon was a result of competition with the introduced R. philippinarum, which has caused the displacement of several bivalve species, as well as substantial changes in ecosystem functioning in this region. These alterations are even more alarming, taking into consideration that NIS are unable to effectively fulfill the ecosystem functions provided by native species after their replacement (Steger et al. 2022).

Consequently, governing entities and regulators should be cautious when establishing new areas for bivalve aquaculture production and ensure that data on the CE and CR of both native species and NIS are taken into account because their ecological niches overlap. In addition, considering that the individuals used in these experiments were representative of the exploitable populations because their shell lengths were all above the minimum marketing size, this information should also be considered when managing NIS and native species stocks.

In conclusion, the results indicate that C. edule, M. angulata, and R. philippinarum have similar capture efficiencies for the particle-size classes measured in this study (ranging from 4 to 14 µm). Moreover, M. angulata and R. philippinarum did not exhibit a competitive advantage in clearing suspended particles compared with the native C. edule, with the latter being the most efficient species on a tissue-weight basis. Therefore, the invasion success of NIS may be explained by other biological characteristics or advantages in ecological interactions, such as reproductive rates, competition for space, overlapping niche spaces, or trophic competition related to diet composition. It is highly recommended that future studies focus on these aspects to determine the full extent of the potential impacts of NIS on native species. Nevertheless, NIS may still jeopardize the entire ecosystem by exerting additional pressure on its carrying capacity. Finally, this study provides an important baseline contribution to the management of NIS by determining capture efficiencies and clearance rates exclusively using natural seawater, thereby providing a more accurate representation of the processes that occur in the environment.