Prioritising non-native fish species for management actions in three Polish rivers using the newly developed tool—dispersal-origin-status-impact scheme

Introduction

Non-native species actively or passively translocated by human actions in regions they have no evolutionary history with (Soto et al., 2024), are recognised among the major threats to global biodiversity, affecting all aspects of ecosystems (Simberloff et al., 2013; Cepic, Bechtold & Wilfing, 2022). These impacts are modulated and often magnified by synergistic interactions with other drivers such as habitat loss, which is considered ‘immense, insidious and usually irreversible’ (Strayer, 2010; Caffrey et al., 2014). Freshwater ecosystems are, among all ecosystems, the most vulnerable to being affected by external drivers such as climate change, pollution, and biological invasions (Havel et al., 2015; Haubrock et al., 2021; Cuthbert et al., 2023). Moreover, in the last three decades, biodiversity declined faster in freshwater ecosystems than in marine and terrestrial ecosystems (Collier, Probert & Jeffries, 2016; Reid et al., 2019, but see Van Klink et al., 2020), with non-native species introductions being among the main extinction drivers (Blackburn et al., 2014). The intrinsic connectivity of freshwater ecosystems due to e.g., the canalization of large rivers, facilitates the spread of non-native species and ultimately increases the homogenisation of ecosystems (Marr et al., 2013). Consequently, mitigation of the effects of non-native species has become one of the most pressing problems ecologists, decision makers, and stakeholders face (Simberloff, 2015).

Considering the growing distribution of countless non-native species and the increasing evidence of their staggering negative effects on recipient ecosystems (Roy et al., 2023) that are increasingly difficult to monitor and manage (Moon, Blackman & Brewer, 2015; Crowley, Hinchliffe & McDonald, 2017), there is a rising need for reliable, accessible, and robust tools to assess the potential threat different populations of these non-native species present. Within the last two decades, several protocols (including both ‘risk assessment protocols’ (Hawkins et al., 2015) and ‘risk screening’ (a.k.a. ‘risk identification’) (Vilizzi et al., 2022); please also see Srebaliene et al. (2019) and Hill et al. (2020) for a comparison of impact and risk assessment methods) have been developed and implemented worldwide, targeting various taxonomic groups and evaluating current and potential impacts of non-native species. Most of the available assessment protocols share a common feature: they enable the classification of non-native species based on the level of risk they do or may present to a specific assessment area. They, however, differ in complexity (e.g., number of assessed aspects of the species), the underlying scoring system, and the range of impacts assessed. However, although they are designed and tested by scientific experts, a recently conducted—yet critized—comprehensive consistency analysis revealed considerable inconsistency among taxonomic groups, scoring systems, expertise of assessors, and impact evaluated (environmental only or with socio-economic; González-Moreno et al., 2019). One pressing issue is that most of these protocols are employed at the national (Tarkan et al., 2017) or continental scale (Haubrock et al., 2021; Vilizzi et al., 2021), which is valuable for national information systems or larger political entities like the European Union but lacks granularity considering the variability of non-native species populations (Haubrock et al., 2024). These generalised approaches can lead to underestimating or overestimating impacts at particular sites by assuming that local effects can be generalised at the species level and be superimposed across regions and ecosystems with similar conditions. Vice versa, an assessment at the national scale, even when informed by local risk screenings, may still underestimate the threat a non-native species presents at specific sites, as generalisations across regions and ecosystems with similar conditions may overlook critical local variations. Another important issue is that several of the currently available protocols consider only environmental impacts (González-Moreno et al., 2019) as socio-economic impacts are usually difficult to quantify due to the lack of information, despite it being widely accepted that the economic consequences of biological invasions prerequisite an efficient allocation of financial resources e.g., management actions (Lodge et al., 2016; Bang et al., 2022; Soto et al., 2023; Tarkan et al., 2024). This further underlines the urgency to easily differentiate and prioritise non-native species for management interventions, resulting in more efficient actions (Lodge et al., 2016).

In Poland, over 60% (17 out of 28) of non-native freshwater fish species were introduced more than three decades ago and now form self-sustained populations in the wild or do not breed in natural condition but are keep in aquaculture and used for stocking several water bodies (Grabowska, Kotusz & Witkowski, 2010). One of the most important pathways aiding the range extension of non-native aquatic species in inland waters of Poland is the European central invasion corridor (Jażdżewski, 1980; Bij de Vaate et al., 2002; Fig. 1). This route was used by several non-native fish species to spread in Polish inland waters (Grabowska, Pietraszewski & Ondračková, 2008; Semenchenko et al., 2011). In response to changing temporal invasion dynamics of non-native species in Polish freshwater ecosystems, alongside recent European Union regulations, the national project run by the government institution The General Directorate for Environmental Protection was completed in 2018. It aimed to determine the degree of invasiveness of non-native species in Poland and identify species that pose the greatest threat to invaded ecosystems. To achieve that goal the Harmonia⁺ protocol was implemented in Poland and named Harmonia^+PL (Tokarska-Guzik et al., 2019). Among the non-native fish species considered in this recent national project (General Directorate for Environmental Protection, 2024), species recently established in Poland include four species of Ponto-Caspian gobies (round, monkey, western tubenose and racer goby; Neogobius melanostomus, N. fluviatilis, Proterorhinus semilunaris and Babka gymnotrachelus, respectively), the Chinese sleeper Perccottus glenii, and the topmouth gudgeon Pseudorasbora parva, but also one species present in European inland waters (including Poland) since the end of the 18^th Century, namely the brown bullhead Ameiurus nebulosus. The last species included was pirapitinga Piaractus brachypomus that is very occasionally recorded as single individuals released by aquarists (Grabowska, Kotusz & Witkowski, 2010).

All those species analysed via Harmonia^+PL protocol, despite their wide distribution across the country (except pirapitinga), were categorised as a low priority in the case of gobies and as a medium priority in the case of Chinese sleeper and topmouth gudgeon. This resulted in the removal of all four goby species from the list of harmful non-native species considered a national Polish concern following the implementation of EU regulations (1143/2014). Furthermore, these changes translate directly into the management of gobies. Although it is still forbidden to introduce them or move them within the environment, it is now allowed to keep them (e.g., in the aquarium or private pond), stock, sell, or exchange them. This can, in practice, result in e.g., intentional introductions via anglers using gobies as live baits (Drake & Mandrak, 2014). Although there is limited evidence of monkey, western tubenose, and racer goby negatively affecting ecosystems they are introduced to Grabowska et al. (2023), this is not the case for the round goby (Cerwenka et al., 2023). Thus, the only fish species among the non-native species currently occurring in Polish waters that remained on the lists of Union or Polish concern are the Chinese sleeper, the topmouth gudgeon, the pumpkinseed (Lepomis gibbosus; which was not assessed by Harmonia^+PL), and the brown bullhead (EU regulations 1143/2014 and its implementation at the national level in Dz. U. 2021 poz. 1718).

However, there is growing recognition that biological invasions are context-specific, with considerable variations in the potential of individuals to spread and exert impacts among populations influenced by diverse environmental and biological factors (Soto et al., 2024; Haubrock et al., 2024). Consequently, there is a need for accurate and standardised assessment protocols that consider the varied effects (both presence and impact) within populations of the same species. The first steps have already been made by Soto et al. (2024), who sorted out the confusion in biological invasion nomenclature and proposed a new assessment scheme—The Dispersal-Origin-Status-Impact (DOSI). The advantage of this approach stems from its thorough yet adaptable framework, which can be applied to specific populations or at broader regional or ecosystem scales in precise and scientific communication. Therefore, some populations might be identified at different scales of prioritisation and can change over time due to the inherent temporal dynamics of an invasion (e.g., population expanding or higher impacts). DOSI improves upon previous management practices by assisting stakeholders and managers, who often face resource constraints (Adelino et al., 2021) in selecting non-native species populations for management actions, thereby enabling them to assess and prioritise non-native species.

To test the relevance and applicability of DOSI, we applied it to non-native species, in three Polish rivers: Bzura, Pilica, and Skrwa Prawa, tributaries of the Vistula River, i.e., the Polish section of the European central corridor of invasions aiming to assess different populations of non-native fish species in rivers of different size. For this, monitoring studies were conducted at least twice on each river, allowing to document several non-native species by examining the entire length of the rivers (i.e., from their sources to their mouths), enabling to obtain an understanding of ongoing changes in the distribution and abundance of these species. The DOSI scheme implementation should provide insight into the threat of non-native species at the population level, enable comparisons with results from the previously conducted Harmonia^+PL to identify potential discrepancies and thereby direct future management efforts to particular localities. The DOSI application may also reveal variability in the level of risk that different populations of the same non-native species may pose in different water bodies, as the population level is usually overlooked by more general metrics (e.g., Harmonia^+PL).

Materials and Methods

The Dispersal-Origin-Status-Impact assessment scheme

The DOSI assessment scheme (Fig. 2) exclusively focuses on negative impacts, emphasizing that these potential threats are significantly more important and distinct than any potential benefits (Carneiro et al., 2024). However, DOSI’s objective is to prioritise non-native populations for management interventions by considering local risks only, without considering the feasibility or availability of appropriate methods, or the species’ potential to spread beyond their current locations. The focus on the population level distinguishes DOSI from other assessment tools, like the Harmonia^+PL protocol, that are commonly applied at varying regional scales (i.e., assessment regions) without a strict focus on the population level. The Harmonia^+PL protocol looks at non-native species at the national level and consists of 30 questions divided into the two main modules “invasion process” and “impact” and a final score calculated based on combined results obtained for both modules.

DOSI prioritisation is structured around a hierarchy of primary dispersal mechanisms, distinguishing between non-native populations that can (a) spread independently and invade areas beyond the introduction site, (b) rely mainly on human assistance and the presence of pathways and vectors, (c) have the capability for both assisted and independent spread (i.e., evaluated for both a and b), and (d) the populations’ status, which defines the state of a population within the target site and the local impact it exerts. This means, that populations that can spread independently and with assistance, and those showing changes in abundance and range, are ranked higher than those with only one type of dependency. This is because the former scenarios indicate a greater and more harmful invasion potential. Similarly, populations with one static and one expanding dependency are also ranked higher. Conversely, if a population is determined to have no known local impact, it is lowered in the priority ranking and thus requires a different response (Fig. 3).

To test the DOSI assessment scheme for each river, we considered all non-native fish species identified. We assessed each identified non-native fish species in the Pilica River (seven non-native species), the Bzura River (seven non-native species), and the Skrwa Prawa River (four non-native species; Table 1) using DOSI to provide an objective overview for the prioritisation of each rivers’ non-native species populations (Fig. 3). Information on changes in abundance growth or range extension were not always precise based on the field samplings, thus we filled information gaps based on our expert knowledge of the study sites and the respective non-native species invasion histories. Consequently, we discussed the DOSI assessment outcomes for the assessed species with the previous screening based on Harmonia^+PL to identify discrepancies and ultimately test if the population level considered by DOSI provides relevant variability.

Table 1:

Summary of non-native fish species occurrence found in each river (Pilica, Bzura and Skrwa Prawa).

Species	Common name	Pilica	Bzura	Skrwa Prawa
Babka gymotrachelus	racer goby	+	+	+
Neogobius fluviatilis	monkey goby	+	+	−
Proterorhinus semilunaris	western tubenose goby	+	+	+
Percottuss glenii	Chinese sleeper	+	+	+
Ameiurus nebulosus	brown bullhead	+	−	−
Carassius gibelio	gibel carp	+	+	+
Pseudorasbora parva	topmouth gudgeon	+	+	−
Cyprinus carpio	common carp	−	+	−

DOI: 10.7717/peerj.18300/table-1

Results

Within the three tested rivers (i.e., Pilica, Bzura and Skrwa Prawa rivers), eight non-native fish species were identified, three of which (i.e., the monkey goby, racer goby, and western tubenose goby) were of Ponto-Caspian origin, another three (i.e., the topmouth gudgeon, Chinese sleeper, and gibel carp) originated from Eastern Asia, while one (brown bullhead) originated from North America, and another one (common carp) from the Danube catchment (Tables S1–S3). All goby species as well as Chinese sleeper and topmouth gudgeon in each evaluated river (Pilica, Bzura, Skrwa Prawa) were classified as independently dispersing, whereas brown bullhead, gibel carp, and common carp as spreading depending on human assistance.

The DOSI ranking was not consistent among species and rivers highlighting the context-dependency of invasions (Fig. 4). The monkey goby was designated as Highest Priority in the Pilica and Bzura Rivers (the species was absent in the Skrwa Prawa and could not be evaluated there) due to its increasing range and abundance leading to competitive pressure on native species (Błońska et al., 2016; Błońska et al. under review). The monkey goby ranking was constant across the analysed sites (Highest). The second one was the western tubenose goby, which also received the status High Priority in all rivers. Although the species is also continually extending its range and abundance, no negative impact has been observed (yet). The third was the topmouth gudgeon, which was ranked as Medium Priority based on static range and abundance in both Pilica and Bzura.

Both the racer goby and the gibel carp were ranked as Highest Priority in the Skrwa Prawa and Pilica River, respectively. In other sites, both species were ranked as High or Medium Priority. These discrepancies result from the inconsistent dynamics of both species. Besides range extension and abundance increase, they displayed a negative effect on native biota at one site while having no influence at another (e.g., gibel carp in the Pilica vs. Bzura River). The Chinese sleeper was scored as Medium Priority in both the Pilica and Bzura Rivers and High Priority in the Skrwa Prawa, where its abundance was increasing rather than static. The only species designated with Low Priority was the brown bullhead, whose decreasing range and abundance are probably due to the less suitable riverine habitat for this species compared to more stagnant waters such as oxbow lakes.

The DOSI ranking differentiated among populations (ranging from High to Low Priority) and was not complementary with the Harmonia^+PL results, which differentiated the six previously assessed species into moderately invasive and potentially invasive (Table 2).

Discussion

In the current study, we evaluated the risk posed by non-native species in three temperate lowland rivers in Poland (the Pilica, Bzura and Skrwa Prawa River) by applying the DOSI scheme (as assessment protocol) and comparing our results to the Harmonia^+PL screening protocol outcomes. Although DOSI and Harmonia^+PL are not directly comparable for this reason, as DOSI focuses on population-level prioritization while Harmonia^+PL identify species risk based on broader ecological implications of species introduction, this distinction allows DOSI to provide more granular, site-specific risk rankings that may vary across locations. This variability was reflected in our findings, where species were not consistently assigned the same rank across the three rivers, demonstrating how DOSI can capture localized differences in species impact, even within similar ecological contexts. Across the three rivers, species were not always designated with the same rank. High and Medium Priority ranks dominated (six times each) with four Highest Priority and only one Low Priority, underlining the DOSI’s ability to prioritise non-native species at the population level.

Conclusion

The application of the DOSI scheme in evaluating the risk posed by non-native species in three temperate lowland rivers in Poland demonstrates that ranking non-native species is both feasible and effective. The study highlights substantial differences between DOSI’s population-level assessments, and the species-level risk screening provided by Harmonia^+PL. These differences underscore the importance of localized and population-specific evaluations in understanding and managing non-native species. DOSI’s ability to assess the risk at the population level provides nuanced insights that are critical for effective management. By identifying the specific threats and prioritising non-native species based on their local impact and spread, DOSI enables more targeted and relevant management decisions. This approach helps in determining the most appropriate management strategies, whether it involves population management for independently spreading species or pathway management for those spreading through human assistance. Applying DOSI in combination with monitoring surveys could enhance the accuracy and timeliness of risk assessments, allowing for more proactive intervention strategies. Expanding the application of DOSI to other geographical regions and aquatic environments may reveal further insights into its effectiveness in addressing varying ecological contexts.

Supplemental Information