Global warming scenarios for the Eastern Adriatic Sea indicate a higher risk of invasiveness of non-native marine organisms relative to current climate conditions

Introduction

Marine bioinvasions represent one of the major sources of pressure on the environment. Because of their obvious negative consequences (Katsanevakis et al. 2014), non-native invasive species are therefore widely recognised as a threat that may lead to profound changes in the structure and function of the invaded ecosystems (Costello et al. 2010). The abundance, proportion, and biomass of non-native species are particularly pronounced in the Mediterranean Sea, which is one of the world's most invaded marine regions (Edelist et al. 2013). As a result, the biodiversity of the Mediterranean Sea has been profoundly influenced by the introduction of non-native (invasive) species over the last 150 years, not only because of the opening of the Suez Canal but also due to other human-related pathways and vectors including shipping, aquaculture, and the aquarium trade (Slišković et al. 2021; Tarkan et al. 2021; Zenetos et al. 2012).

The Mediterranean Sea is one of the two most prominent climate change hotspots in the world (Giorgi 2006), implying that the impacts of climate change on its environment could be especially severe by affecting precipitation, air, and sea water temperatures. For the Adriatic Sea, a positive trend in both sea surface and air temperatures have been recorded recently (Bonacci and Vrsalović 2022; Pisano et al. 2020). According to International Panel for Climate Change (https://www.ipcc.ch/) scenarios, the sea surface temperature and salinity of the Adriatic Sea are projected to increase in the future, together with changes in precipitation regime. The projected future climate for the EAS is characterised by a warming of up to 5 ℃ by the end of the twenty-first century (Branković et al. 2013), with rainfalls expected to increase in winter and decrease in summer by 20% (Zampieri et al. 2012). It has been pointed that the long-term sea surface temperature warming trend in the Adriatic Sea has already impacted on its marine ecosystems and biodiversity (Bonacci et al. 2021). Additionally, a recent study has demonstrated a decreasing trend in habitat suitability for many target species (almost 50%) with a decrease in landings from 13.5 to 86.9% (depending on scenarios) and their likely replacement in the short-to-medium term with southern or tropical species (Cavraro et al. 2023). Thus, the future climatic conditions for the Adriatic Sea under tested scenarios indicate that it could become a suitable environment for tropical species. To this end, the introduction of thermophilic non-native species is expected to cause competition with local species and alter the biological cycles of acclimatised marine species to temperate cold climates (Cavraro et al. 2023).

The Adriatic Sea is an 800 km-long, semi-enclosed basin of the Mediterranean Sea whose ecosystems are influenced by the regular exchange of seawater from the eastern part of the basin (Gačić et al. 2010). This incoming nutrient-rich and warmer seawater, especially under current conditions of global warming, affects the function of the local ecosystems in terms of primary and secondary production and has caused several warm-water species to move towards higher latitudes (Grbec et al. 1998; Orlić et al. 1992; Zore-Armanda 1972). The arrival into the Adriatic Sea of non-native and thermophilic 'neo-native' marine species (i.e., range-expanding species previously not recorded in the Adriatic Sea but present in other parts of the Mediterranean Sea) is therefore expected to increase because of sea warming (Dulčić and Grbec 2000), although other factors including ballast water and aquaculture escapees are likely to play an important role (Glamuzina et al. 2021a).

Like other Mediterranean countries, those facing the Eastern Adriatic Sea (EAS), namely Slovenia, Croatia, Bosnia-Herzegovina, Montenegro, and Albania, have been facing increasing problems from the introduction during the last decades of non-native marine species (Dragičević and Dulčić 2010; Dulčić and Dragičević 2011; Lipej et al. 2012; Petović and Mačić, 2017; Pešić et al. 2020). According to a recent review (Pećarević et al. 2013), as of 2013 in the EAS, there were 113 established non-native species (15 phytoplankton, 16 macroalgae, 16 zooplankton, 44 zoobenthos, and 22 fish species), and of these, some are invasive and known to exert impacts on marine ecosystem services and biodiversity (Dulčić et al. 2011b; Mancinelli et al. 2016). Since then, several new species that may alternative community composition, but whose invasiveness is still unknown, have been recorded in the EAS (Marić et al. 2017; Petrocelli et al. 2019; Slišković et al. 2021; Tsiamis et al. 2019). For this reason, preventative actions and policy measures for the early detection of non-native species and the determination of their risk of invasiveness are needed (Azzurro et al. 2019; Chan et al. 2019; Ojaveer et al. 2018). However, preventative actions and related policy are currently lacking in the countries that share the EAS coastal area (Dulčić et al. 2018). At the same time, post-introduction actions, such as eradication, control, and containment, are known to be difficult to implement and unlikely to be successful (Williams and Grosholz 2008), and especially in the marine environment (Werschkun et al. 2014). There is therefore a need to implement defensible strategies relying on both monitoring and risk assessment for the identification, control, and management of non-native (invasive) species.

Overall, marine bioinvasions require the integration of existing monitoring tools with complementary early detection strategies for more efficient management measures. However, to be successful, these strategies must rely on defensible evidence based on risk analysis, which consists of three sequential components: risk screening (or hazard identification), risk assessment, and risk communication and management (e.g., Copp et al. 2016a; Vilizzi and Piria 2022). Apart from a recent risk screening study for an estuarine area of Croatia and Bosnia-Herzegovina that included 12 non-native marine species (Glamuzina et al. 2021b) and another study concerned with nine non-native marine species screened for the coastline of Montenegro (Tomanić et al. 2022), neither a comprehensive risk screening for non-native marine species nor related management actions have been implemented in any of the EAS countries.

The aim of this study was to determine the risk of invasiveness of both extant and 'horizon' (i.e., not yet present) non-native marine species under both current and predicted climate conditions for the EAS coasts of Slovenia, Croatia, and Montenegro—the risk assessment area. The specific objectives were twofold: (1) identify the non-native extant and horizon marine species introduced into the EAS that are likely to pose the highest threat to native biodiversity; and (2) evaluate the risk of invasiveness of the selected species under both current and predicted climate conditions for the risk assessment area. It is anticipated that the outcomes of this study will be used by national authorities in all three countries to propose measures for preventing the expansion and colonisation of new areas by those species likely to pose the highest threat to the biodiversity of the EAS. Of note, as this study focuses on the screening component of risk analysis applied to non-native species, no account will be made of the (follow-up) risk assessment of those species identified as high risk, which would involve detailed species-specific evaluation of their impacts and responses to climate change conditions in terms of spread and distribution. This is an important distinction that is often overlooked in risk analysis studies of invasion science (e.g., González-Moreno et al. 2019; Marcot et al. 2019) and has led to misinterpretations (Hill et al. 2020).

Materials and methods

Species selection

Selection of the species for risk screening was performed according to three main criteria (after Zenetos et al. 2010; Supplementary files: Table 1): (1) extant established species (EE), i.e., species recorded at least twice (invertebrates and plants) or three times (fishes) in the risk assessment area; (2) extant not established species (ENE), i.e., species recorded only once (invertebrates and plants) or twice (fishes) in the risk assessment area; (3) horizon species (H), i.e. species likely to enter the risk assessment area in the near future as defined from a horizon-scanning exercise based on consultation of the Centre for Agriculture and Bioscience International Invasive Species Compendium Horizon Scanning tool (www.cabi.org/HorizonScanningTool) refined by expert knowledge.

The lists of EE and ENE species were prepared based on: (1) species identified by Pećarević et al. (2013); (2) search of published papers with records on non-native fishes (Chordata—Actinopterygii), invertebrates (Arthropoda, Bryozoa, Ctenophora, Mollusca, Porifera), and plants (macroalgae, vascular plants) in the risk assessment area; and (3) European Alien Species Information Network (https://easin.jrc.ec.europa.eu/easin) database searches for Croatia and Slovenia (with a first search yielding 100 items for Croatia and 38 for Slovenia in total from which fishes, invertebrates, and plants were extracted). The list of H species was prepared based on a literature search of the non-native species recorded in the near proximity of the risk assessment area (primarily with a focus on the Western Adriatic and Ionian Seas) using the following criteria: (1) invasive species; (2) species with documented harmful effect on humans; (3) species with importance for fishery in their native or introduced area; or (4) species valuable in the ornamental industry. In all searches, cryptogenic, north-expanding species of indigenous origin as well as 'questionable' species were excluded.

Risk screening

Risk screening was undertaken with the Aquatic Species Invasiveness Screening Kit (AS-ISK) v2.3.3 (Copp et al. 2016b, 2021; www.cefas.co.uk/nns/tools/). This taxon-generic, decision-support tool consists of 55 questions of which 49 comprise the Basic Risk Assessment (BRA) and six the Climate Change Assessment (CCA). BRA address the biogeography/invasion history and biology/ecology of the species; and the CCA require the assessor to predict how future predicted climatic conditions are likely to affect the BRA with respect to risks of introduction, establishment, dispersal, and impact. Screenings were conducted separately by 19 assessors, with each assessor screening from two to 13 species, and with 56 species being evaluated by a single assessor, 19 by two independent assessors, and nine by three independent assessors.

To achieve a valid screening, the standard protocol described in Vilizzi et al. (2022a) and Vilizzi and Piria (2022) was followed. In detail, the assessor provided for each question a response, confidence level, and justification based on literature sources, which result in two score outcomes (BRA and BRA + CCA). The outcomes were a BRA score, which ranges from − 20 to 68, and a (composite) BRA + CCA score, which ranges from − 32 to 80 (i.e., after adding or subtracting up to 12 points to the BRA score or leaving it unchanged in case of a CCA score equal to 0). Scores < 1 suggest a 'low risk' of the species being or becoming invasive in the risk assessment area, whereas scores ≥ 1 indicate a 'medium risk' or a 'high risk'. Distinction between medium and high risk was defined using a calibrated threshold that was obtained, whenever possible (see below), by Receiver-Operating Characteristic (ROC) curve analysis. In this study, for the fishes and invertebrates classified as high risk, an additional 'very high risk' category was distinguished using an ad hoc threshold. As part of the risk analysis process, identification of the very high-risk species may help prioritise allocation of resources for comprehensive risk assessment. This examined in detail the risks of: (1) introduction (entry); (2) establishment (of one or more self-sustaining populations); (3) dispersal (more widely within the RA area, i.e., so-called secondary spread or introductions); and (4) impacts (to native biodiversity, ecosystem function and services, and the introduction and transmission of diseases).

Confidence levels in the responses to questions in the AS-ISK were ranked using a 1–4 scale (1 = low; 2 = medium; 3 = high; 4 = very high) as per the Intergovernmental Panel on Climate Change (Copp et al. 2016b). Based on the confidence level (CL) allocated to each response, a confidence factor (CF) was obtained as

where CL_Qi is the CL for Qi, 4 is the maximum achievable value for confidence (i.e., very high: see above) and 55 is the total number of questions comprising the AS-ISK questionnaire. The CF ranges from a minimum of 0.25 (i.e., all 55 questions with confidence level equal to 1) to a maximum of 1 (i.e., all 55 questions with confidence level equal to 4). Based on all 55 Qs of the AS-ISK questionnaire, the 49 Qs comprising the BRA and the six Qs comprising the CCA: for the CL, the CL_Total, CL_BRA and CL_CCA, and for the CF, the CF_Total, CF_BRA and CF_CCA were computed.

The a priori categorisation of species required for ROC curve analysis was implemented as per the standard protocol (Vilizzi et al. 2022a; Supplementary files: Table 1). An ROC curve is a graph of sensitivity vs 1—specificity for each threshold value, where in the present context, sensitivity and specificity are the proportion of a priori invasive and non-invasive species, respectively, that are correctly identified by the AS-ISK. This requires at least 15–20 species to achieve a successful, statistically robust calibrated threshold score with which to distinguish between medium-risk and high-risk species; these species must consist of both a priori non-invasive and invasive in a 'relatively balanced' proportion. Unlike fishes and invertebrates, in this study, such requirements were not met for plants, so the generalised threshold equal to 32 for marine plants was used (after Vilizzi et al. 2021). Implementation of ROC curve analysis also followed the standard protocol (Vilizzi et al. 2022a), with the true/false–positive/negative outcome distinction not applied to medium-risk species, because their further evaluation in a comprehensive risk assessment depends on policy/management priorities and/or the availability of financial resources. A measure of the accuracy of the calibration analysis is the area under the curve (AUC). If the AUC is equal to 1, then the test is 100% accurate, because both sensitivity and specificity are 1, and there are neither 'false positives' (a priori non-invasive species classified as high risk, hence invasive) nor 'false negatives' (a priori invasive species classified as low risk, hence non-invasive). If the AUC is equal to 0.5, then the test is 0% accurate as it cannot discriminate between 'true positives' (a priori invasive species classified as high risk, hence invasive) and 'true negatives' (a priori non-invasive species classified as low risk, hence non-invasive). Values for the AUC were interpreted as: 0.7 ≤ AUC < 0.8 = acceptable discriminatory power, 0.8 ≤ AUC < 0.9 = excellent, and 0.9 ≤ AUC = outstanding. Following ROC analysis, the best threshold value that maximises the true-positive rate and minimises the false-positive rate was determined using Youden's J statistic; whereas the 'default' threshold of 1 was set to distinguish between low-risk and medium-risk species. Notably, the true/false positive/negative outcome distinction was not applied to the medium-risk species, as they can be either included or not into a full (comprehensive) RA (see "Risk screening") depending on priority and/or availability of financial resources.

Fitting of the ROC curve for fishes and invertebrates was with pROC (Robin et al. 2011) for R × 64 v4.0.5 (R Core Team 2022). Permutational ANOVA with normalisation of the data was used to test for differences in confidence factor (CF) between components (BRA and BRA + CCA) and organism groups (fishes, invertebrates, and plants); this used a Bray–Curtis dissimilarity measure, 9999 unrestricted permutations of the raw data, and with statistical effects evaluated at α = 0.05.

Results

A risk screening report (Online Resource1) was generated for each of the 116 screenings for the 84 species in total evaluated for the EAS (Fig. 1).

Figure  .  Map of the Adriatic Sea including the coasts of Slovenia, Croatia, and Montenegro representing the risk assessment area for which 84 non-native marine species (see Supplementary files: Table 1) were screened for their risk of invasiveness

DownLoad: Full-Size Img PowerPoint

Fishes

Of the 37 species, 28 were categorised a priori as non-invasive and nine as invasive (Supplementary files: Table 1). The ROC curve analysis provided an AUC of 0.7004 (0.5050–0.8598 95% CI) and a calibrated threshold of 15.25.

Based on the BRA scores (Supplementary files: Table 2, Fig. 2A), 17 (45.9%) species were ranked as high risk, 16 (43.2%) as medium risk, and 4 (10.8%) as low risk. Of the 28 a priori non-invasive species, four were true negatives (bluespotted seabass Cephalopholis taeniops, nakedband gaper Champsodon nudivittis, sapphire devil Chrysiptera cyanea, and narrow-barred Spanish mackerel Scomberomorus commerson) and ten were false positives (smallscale codlet Bregmaceros nectabanus, white grouper Epinephelus aeneus, black-barred halfbeak Hemiramphus far, Oceanic puffer Lagocephalus lagocephalus, barred knifejaw Oplegnathus fasciatus, red seabream Pagrus major, soi-uy mullet Planiliza haematocheilus, common lionfish Pterois miles, yellowstripe barracuda Sphyraena chrysotaenia, and burrowing goby Trypauchen vagina). Of the nine a priori invasive species, seven were true positives (bluespotted cornetfish Fistularia commersonii, silver-cheeked toadfish Lagocephalus sceleratus, striped eel catfish Plotosus lineatus, redcoat Sargocentron rubrum, brushtooth lizardfish Saurida lessepsianus, dusky spinefoot Siganus luridus, and marbled spinefoot Siganus rivulatus). Of the 16 medium-risk species, 14 were a priori non-invasive and two invasive.

Figure  2.  Aquatic Species Invasiveness Screening Kit (AS-ISK) scores (± SE, where applicable) and risk ranks for fishes. A BRA (basic risk assessment) scores for the first 19 species. B BRA scores for the last 18 species; C BRA + CCA (BRA + climate change assessment) scores for the first 19 species. D BRA + CCA scores for the last 18 species. Orange bars = very high-risk species; Green bars = high-risk species; Purple bars = medium-risk species; white bars = low-risk species. Solid line = very high-risk threshold; hatched line = high-risk threshold; dotted line = medium-risk threshold (thresholds as per Supplementary files: Table 2)

DownLoad: Full-Size Img PowerPoint

Based on the BRA + CCA scores (Supplementary Table 2, Fig. 2B), 23 (62.2%) species were ranked as high risk, 11 (29.7%) as medium risk, and 3 (8.1%) as low risk. Of the a priori non-invasive species, three were true negatives (same species as for BRA except for Scomberomorus commerson) and 15 false positives (same species as for BRA plus Indian scad Decapterus russelli, fangtooth moray Enchelycore anatina, orange-spotted grouper Epinephelus coioides, Golani round herring Etrumeus golanii, and longfin yellowtail Seriola rivoliana). Of the a priori invasive species, eight were true positives (same species as for BRA plus goldband goatfish Upeneus moluccensis). Of the eleven medium-risk species, ten were a priori non-invasive and one invasive.

All 17 high-risk species for the BRA were also ranked as high risk after accounting for climate change predictions (cf. BRA + CCA), which resulted in the inclusion of six additional species. Based on an ad hoc very high-risk threshold ≥ 35, the highest-scoring species were Epinephelus aeneus and Sphyraena chrysotaenia for both the BRA and BRA + CCA, and Bregmaceros nectabanus, Fistularia commersonii, Hemiramphus far, Lagocephalus lagocephalus, Lagocephalus sceleratus, Oplegnathus fasciatus, Plotosus lineatus, Pterois miles, Saurida lessepsianus, Siganus luridus, and Siganus rivulatus for the BRA + CCA only (Fig. 2). The CCA resulted in an increase in the BRA score (cf. CCA) for 31 species (of which 14 achieved the maximum increase of 12 points), in no change for four, and in a decrease for two (Supplementary Table 2).

Invertebrates

Of the 38 species, 14 were categorised a priori as non-invasive and 24 as invasive (Supplementary files: Table 1). The ROC curve analysis provided an AUC of 0.7278 (0.5607–0.8948 95% CI) and a calibrated threshold of 23.5.

Based on the BRA scores (Supplementary Table 2; Fig. 3A), 15 (39.5%) species were ranked as high risk and 23 (60.5%) as medium risk. Of the 14 a priori non-invasive species, one was a false positive (Japanese bubble snail Haloa japonica). Of the 24 a priori invasive species, 14 were true positives (transverse ark Anadara transversa, Asian date mussel Arcuatula senhousia, Brachidontes pharaonis, Atlantic blue crab Callinectes sapidus, Say mud crab Dyspanopeus sayi, Pacific oyster Magallana gigas, sea walnut Mnemiopsis leidyi, Paraleucilla magna, rayed pearl oyster Pinctada radiata, blue swimming crab Portunus pelagicus, African blue swimming crab Portunus segnis, purple whelk Rapana venosa, Manila clam Ruditapes philippinarum, and pleated sea squirt Styela plicata). Of the 23 medium-risk species, 13 were a priori non-invasive and ten invasive.

Figure  3.  AS-ISK scores (± SE, where applicable) and risk ranks for invertebrates. A BRA scores for the first 19 species. B BRA scores for the last 19 species. C BRA + CCA scores for the first 19 species. D BRA + CCA scores for the last 19 species. Notation and formatting as per Fig. 2

DownLoad: Full-Size Img PowerPoint

Based on the BRA + CCA scores (Supplementary Table 2; Fig. 3B), 18 (47.4%) species were ranked as high risk and 20 (52.6%) as medium risk. Of the a priori non-invasive species, six were false positives (same species as for BRA plus northern brown shrimp Penaeus aztecus, Penaeus Hathor, and striped false limpet Siphonaria pectinata). Of the a priori invasive species, 14 were true positives (same species as for BRA except for Styela plicata and plus spiny oyster Spondylus spinosus). Of the 20 medium-risk species, ten were a priori non-invasive and ten invasive.

Except for Styela plicata, all other 15 high-risk species for the BRA were also ranked as high risk after accounting for climate change predictions (cf. BRA + CCA), which resulted in the inclusion of three additional species. Based on an ad hoc very high-risk threshold ≥ 40, the highest-scoring species were Arcuatula senhousia, Callinectes sapidus, Mnemiopsis leidyi and Rapana venosa for both the BRA and BRA + CCA, and Brachidontes pharaonis, Dyspanopeus sayi, Magallana gigas and Pinctada radiata for the BRA + CCA only (Fig. 2). The CCA resulted in an increase in the BRA score (cf. CCA) for 25 species (of which six achieved the maximum increase of 12 points), in no change for eight, and in a decrease for five (Supplementary Table 2).

Plants

Of the nine species, one was categorised a priori as non-invasive and eight as invasive (Supplementary files: Table 1), and the generalised threshold of 32 was used.

Based on the BRA scores (Supplementary Table 2; Fig. 4A), 1 (11.1%) species was ranked as high risk and 8 (88.9%) as medium risk. Of the eight a priori invasive species, one was a true positive (sponge seaweed Codium fragile). Of the three medium-risk species, one was a priori non-invasive and two invasive.

Figure  4.  AS-ISK scores (± SE, where applicable) and risk ranks for plants. A BRA scores. B BRA + CCA scores. Notation and formatting as per Fig. 2

DownLoad: Full-Size Img PowerPoint

Based on the BRA + CCA scores (Supplementary Table 2; Fig. 4B), 6 (66.7%) species were ranked as high risk and 3 (33.3%) as medium risk. Of the a priori invasive species, six were true positives (same species as for BRA plus Acrothamnion preissii, Red Sea plume Asparagopsis taxiformis, Caulerpa cylindracea, killer algae Caulerpa taxifolia, and broadleaf seagrass Halophila stipulacea). Of the three medium-risk species, one was a priori non-invasive and two invasive.

Only Codium fragile was ranked high-risk for both the BRA and BRA + CCA, whereas accounting for climate change predictions (cf. BRA + CCA) resulted in the inclusion of five additional species (Fig. 2). The CCA resulted in an increase in the BRA score (cf. CCA) for seven species (of which three achieved the maximum increase of 12 points), in no change for one, and in a decrease for another one (Supplementary Table 2).

Confidence

The mean CF_Total was 0.674 ± 0.012, the mean CF_BRA 0.687 ± 0.013, and the mean CF_CCA 0.563 ± 0.015. For fishes: CF_Total = 0.618 ± 0.014, CF_BRA = 0.630 ± 0.015, CF_CCA = 0.518 ± 0.017; for invertebrates: CF_Total = 0.688 ± 0.019, CF_BRA = 0.701 ± 0.019, CF_CCA = 0.576 ± 0.025; for plants: CF_Total = 0.821 ± 0.029, CF_BRA = 0.838 ± 0.028, CF_CCA = 0.678 ± 0.049. In all cases, the level of confidence was medium. Overall, CF_BRA was higher than the CF_CCA (F_{1, 226}^# = 38.73, P < 0.001; # = permutation value), and the CF for plants was higher than the CF for invertebrates, which was higher than the CF for fishes (F_{2, 226}^# = 21.22, P < 0.001; plants vs invertebrates: t^# = 3.67, P < 0.001; plants vs fishes: t^# = 7.11, P < 0.001; invertebrates vs fishes: t^# = 3.36, P < 0.001;). There was no Component × Group interaction, indicating that differences between the CF_BRA and CF_CCA were unrelated to the organism groups.

Discussion

This study has provided the first extensive risk screening of extant and horizon non-native marine species for the EAS coasts of Slovenia, Croatia, and Montenegro, and it is the first extensive screening study for the Adriatic Sea. With respect to the 84 species screened, there was an increase in those ranked as high risk from 33 (39.3%) under current climate conditions to 47 (56.0%) under global warming scenarios, and for those ranked as very high risk, the increase was from 6 (7.1%) to 21 (25.0%). This outcome emphasises the need to implement measures aimed at the mitigation of the potential impacts caused by the (higher risk) extant non-native species and the prevention of entry of the (higher risk) horizon species.

The calibrated threshold of 15.25 identified in this study for the marine fishes of the EAS is comparable to the global thresholds (both generalised and climate-specific: 12.75 and 19.5, respectively) estimated for this group of aquatic organisms (Vilizzi et al. 2021). And the same is true relative to the calibrated thresholds for marine fishes in the south-western coasts of Anatolia (18.5: Bilge et al. 2019) and in the eastern Mediterranean Sea (15.75: Tarkan et al. 2021), including the threshold for the five pufferfishes in the Eastern Mediterranean Sea (Filiz et al. 2017), which was based on the calibrated one by Bilge et al. (2019). Whereas a comparison cannot currently be made with the higher calibrated threshold for the marine fishes screened for the Mediterranean Sea (27.5: Yapici 2021), which should be subject to revision (Vilizzi and Piria 2022). Conversely, the calibrated threshold of 23.5 identified in this study for the marine invertebrates of the EAS is higher than the global thresholds (both generalised and climate-specific: 15.1 in both cases) estimated for this group of aquatic organisms (Vilizzi et al. 2021). And the same is true relative to the calibrated thresholds for marine invertebrates (jellyfish: Killi et al. 2020; crustacean decapods and barnacles: Stasolla et al. 2021) in the Mediterranean Sea as a whole. This is possibly an outcome of the elevated invasive potential of the screened invertebrate species in the present study for the risk assessment area under investigation due to the different climate types of the coastal areas found in its northern and southern parts. A discussion of the higher risk species for each group of organisms as identified in this study is provided in Supplementary Information 2.

The Mediterranean Sea is a marine biodiversity hotspot (Coll et al. 2010) with nearly half (49%) of its species recorded in the Adriatic Sea, and this rich diversity is threatened by the non-native species introduced since the 1970s (Zenetos et al. 2017; Servello et al. 2019; Tsiamis et al. 2019). This study has identified several high-risk species most of which have reached the risk assessment area by expanding their distributional range as migrants from the Suez Canal (Dulčić et al. 2011a, 2011b; Dulčić and Dragičević 2014; Joksimović et al. 2009; Pallaoro and Dulčić, 2001) likely due to changes in the ecosystems of the Adriatic Sea as a result of global warming (Bonacci et al. 2021). On the contrary, most of the macroinvertebrates and plants have been introduced into the risk assessment area by shipping (Slišković et al. 2021), with ports and marinas acting as hotspots for their first introduction (Petrocelli et al. 2019). Since the coast of the EAS is characterised by hard substrata with highly diverse habitats separated by estuarine ecosystems, such habitats may be potentially beneficial for the further spread and establishment of these non-native species (Glamuzina et al. 2021a; Slišković et al. 2021).

Several species identified in this study as high risk and already established in the EAS represent a threat to traditional aquaculture or fishing, causing damage to economy and tourism. One such example is Magallana gigas, which represents a threat to local shellfish farming and especially for the cultivation of the European flat oyster Ostrea edulis (Ezgeta-Balić, et al. 2019; Stagličić et al. 2020; Šegvić-Bubić et al. 2016). Magallana gigas has been successful in establishing self-sustaining populations in the Mediterranean Sea (Troost 2010). Consequently, the native O. edulis has been listed as a 'Threatened and declining species', with only small relic populations present in European waters (Stechele et al. 2022). However, M. gigas has not yet been recorded in the central and southern parts of the EAS, which include the coastal areas of Croatia, Bosnia-Herzegovina, and Montenegro (Ezgeta-Balić, et al. 2019; Hansen et al. 2022), where the most valuable farms for O. edulis are located. Should M. gigas become established also in these areas, eradication is expected to be costly and extremely difficult to implement (Hansen et al. 2022), with resulting considerable damage to the remnant local oyster populations (Troost 2010) and related food industry.

A second example is the already established Callinectes sapidus with abundant populations along the EAS coasts, including the vulnerable ecosystems of the River Neretva Estuary (Glamuzina et al. 2021a, 2021b). The high abundance of this species may cause the decline of commercially important native species thereby impacting on the ecosystem services of the region (Laugen et al. 2015). In this regard, the traditional fyke net traps used in the area are not effective for the successful capture of C. sapidus, although by changing the design of traps (e.g., into American wire traps), their effectiveness may become higher (Glamuzina et al. 2021b). Whilst C. sapidus may be of interest for introduction as a new gastronomical 'delicacy' as occurred in several countries (Piria et al. 2021), such practice is to be discouraged in the risk assessment area as it would replace the traditional values and identity.

Future climate change models for the Adriatic Sea project changes in both precipitation regime and sea surface temperature and salinity towards the end of the twenty-first century (Branković et al. 2013). Specifically, rainfall is expected to increase in winter and decrease in summer by 20% (Zampieri et al. 2012), and sea surface temperature may increase up to + 5 ℃ (Bonacci et al. 2021) coupled with an increase in salinity (Paklar et al. 2020). These abiotic factors are likely to affect the future distribution of most species in the marine environment (Azzurro et al. 2013; Vitelletti et al. 2023) by forcing native marine species to adapt to the new conditions by finding suitable ecological niches, with the concomitant risk of being replaced by southern or tropical species (Cavraro et al. 2023; Marras et al. 2015). Future climate scenarios for the Adriatic Sea also show a decreasing trend in habitat suitability for many native commercial fish species (Cavraro et al. 2023), thereby favouring Suez Canal migrants such as Siganus rivulatus (Marras et al. 2015), which competes for food and space with native herbivorous fish such as salema Sarpa salpa (Bariche et al. 2004). Furthermore, the distribution of coralligenous outcrops, which are important biodiversity hotspots (Piazzi et al. 2019), is likely to be affected by changes in salinity, temperature, and nitrate concentration (Vitelletti et al. 2023). Overall, the environmental variations projected under climate change conditions are expected to favour the spread of opportunistic organisms that are more tolerant of stressful conditions at the expense of the more vulnerable species. This will result in a shift in the distribution of the available habitats, with a consequent potential loss of biodiversity in the Adriatic Sea (Vitelletti et al. 2023).

Overall, the above examples point to the importance of improving management efforts for the mitigation and control of non-native species entry and consequent impacts on the EAS coasts (Slišković et al. 2021). However, legislation and research efforts vary between the countries sharing the EAS coastline, including a lack of networking initiatives (Rak et al. 2019). Despite the Adriatic Sea coastline being shared by six countries (i.e., Albania, Bosnia-Herzegovina, Croatia, Italy, Montenegro, and Slovenia), a common policy following the Marine Strategy Framework Directive Descriptor criterion D2C1 (EU 2017) is still not developed. This affects the development and implementation of preventative measures for the management and control of non-native species introduction routes and pathways. This should be a priority approach (Tsiamis et al. 2019), especially given that non-native marine species, once established, are extremely difficult to control or eradicate (Simberloff 2021). Additionally, to achieve these goals, scientific efforts will have to be supported by extensive education of local stakeholders from different fishery and public sectors and by the promotion of internet networks and mobile applications at the international level for the immediate report and fast identification of invasive species.

The EU Invasive Alien Species Regulation 1143/2014 is the core legislation for invasive alien species management in Europe, and its importance is highlighted in the EU Biodiversity Strategy for 2030. However, invasive marine species are under-represented despite their posing a major socioeconomic and environmental threat (Kleitou et al. 2021). In the third update of the Union List entered into force on 2 August 2022, marine species remain overall under-represented and none of the extant species ranked as high risk in this study for the EAS are included in the list. These species should therefore be considered as candidates to complement the list of non-native species of Union concern due their potential impacts on biodiversity and economy, as demonstrated in this study.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1007/s42995-023-00196-9.

Acknowledgements

Marina Piria was funded by an ERASMUS+ (EPPKA2-Cooperation for innovation and the exchange of good practices, CBHE-JP-Capacity building in higher education) within the project "Educational capacity strengthening for risk management of non-native aquatic species in Western Balkans (Albania, Bosnia and Herzegovina and Montenegro)-RiskMan" (project reference: 619384-EPP-1-2020-1-TR-EPPKA2-CBHEJP). The authors from Slovenia acknowledge the national research project "Monitoring the species diversity and abundance of non-native species in the Slovenian Sea" and the Slovenian Research Agency (research core funding No. P1-0237). Ante Žuljević, Ivan Cvitković, and Marija Despalatović acknowledge the Croatian Science Foundation for the support under Project No. HRZZ-IP-2019-04-6702 (Benthic NIS).

Author contributions

BG and PT designed and conceived the study with the assistance of MP and LV. The above authors contributed to the write-up of the manuscript. Data and literature collection was by BG, PT, AŽ, IC, MD, and SMS. Preparation of data for analysis was by PS and ZV, and analysis of data by LV. ABC, AF, AP, AŽ, BM, BD, BG, DT, IC, JT, LL, LG, MP, MBŠ, MP, PT, SG, TD, and VB conducted the screenings.

Data availability

The data supporting the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Conflict of interest

The authors declare no competing interests. The authors alone are responsible for the content and writing of the article.

Animal and human rights statement

This article does not contain any studies with human participants or animals performed by the authors.