This study is based on data accumulated in the Information system on Aquatic Non-Indigenous and Cryptogenic Species–AquaNIS, where all geographic information is arranged in a hierarchical order ranging from oceans, ocean sub-regions, LMEs, sub-regions of LMEs to smaller entities, such as ports (Olenin et al., 2014; AquaNIS, 2016). All countries are linked to relevant LMEs or LME sub-regions. This provides database search combinations “country + LME” or “country + LME sub-region” for different coasts and for a country that borders different seas, e.g.,: “Germany within the LME 23 Baltic Sea,” “Italy within the Adriatic Sea, a sub-region of LME 26 Mediterranean.” Such data may also be aggregated at different geographical scales and in different combinations, e.g., “LME 22 North Sea + LME 23 Baltic Sea,” or “Germany within both the North Sea and Baltic Sea coasts,” which would be needed to define the level of primary introduction.

The basic data entry in AquaNIS is an introduction event record, documenting a species introduction into a recipient region, defined as a country or a country sub-area within an LME or LME sub-region. Registration of an introduction event includes the date of the first record when a species was noticed in a recipient region as well as pathways and vectors of introduction according to different levels of certainty. In addition, AquaNIS gathers and disseminates information on species biological traits, environmental tolerance limits, availability of molecular data for identification, habitats, etc. Moreover, the information system is equipped with a structured “search” function that allows for retrieving and organizing data by multiple and complex search criteria (for details see Olenin et al., 2014).

nNIS is the number of new NIS in an assessment unit, which were recorded and compared with the initial or periodic assessment. In this study, the assessment unit is equal to a recipient region as it is in AquaNIS. To illustrate such a calculation we selected a range of assessment units from different marine environments as examples, where all entered data has been checked for quality. The areas selected were the Baltic Sea and coastal waters of Italy and the records may be examined on-line (AquaNIS, 2016).

In the Baltic Sea, there are 10 recipient regions: eight bordering countries and the two separate regions of the Russian Federation, the Sankt-Petersburg area in the Gulf of Finland (RU_S) and the Kaliningrad area in the south-eastern Baltic (RU_K). In Italy, there are three recipient regions: the Adriatic Sea, Western Mediterranean (the western coast of the Italian mainland and north coast of Sicily) and Eastern Mediterranean (the south coast of the Italian mainland and south coast of Sicily).

The initial assessment is the first inventory of all NIS present in a recipient region. For example, most EU Member States will have performed an initial environmental status assessments under the MSFD and reported the cumulative number of NIS in the waters under their jurisdiction recorded by 2010. In the present study, all new NIS, arriving after this date, were counted. A periodic assessment is a record of new NIS arrived to a recipient region since the first inventory. The periodicity of the assessment may be defined by the management needs, for example, it will be 6 years for MSFD and in maximum 5 years for granting ballast water management exemptions under BWMC (Olenin et al., 2016).

A primary introduction is the first arrival of a NIS to a particular assessment unit, while the secondary spread is its further dispersal to other locations. The level of a primary introduction can be assessed at different geographical scales, from a recipient region to an LME or a larger biogeographical area. From the environmental policy point of view, more important are those primary introductions, which are new not only for a coast of a particular country (recipient region), but for an entire LME or, even for a larger biogeographical region, for example for two or more neighboring LMEs or LME sub-regions.

The levels of primary introduction should be defined for each case study separately, depending on the availability of data for larger geographical scales. In the Baltic case study, the lowest level of primary introduction (L1) is one of the 10 recipient regions, the next level (L2) is the entire LME (Baltic Sea), and the highest level (L3) is two neighboring LMEs (Baltic Sea and North Sea). Thus, nNISL1 shows how many new NIS were recorded in a particular country since the previous assessment, nNISL2 shows how many of them were new for the Baltic Sea, and nNISL3 indicates the number of NIS new for both the Baltic and the North seas. In the Italian case study, the lowest level (L1) is one of the three recipient regions, while the next level (L2) is all Italian marine areas together. The highest level here would be the entire Mediterranean Sea, the data for which currently is under development.

AquaNIS offers an opportunity to extract the value of nNISL1 directly, using the built-in “Search” function for the recipient region and year, from which the new arrivals should be calculated. The system can retrieve the number of species (i.e., nNISL1) and the number of introduction events. Data extraction for nNISL2 and L3 values involves several steps, using a combination of “Search” and “Comparison of search results” functions (Table 1).

The calculations of all nNIS values presented here are based on data that has accumulated in AquaNIS to July 28, 2016 (AquaNIS, 2016). All entries for cryptogenic species were not considered in our calculations.

According to the AquaNIS (2016) definition, an introduction event should be ascribed to a pathway/vector with the defined level of certainty (Table 2).

In total, 26 NIS involving 36 introduction events were recorded in 10 recipient regions within the Baltic Sea since 2010. Of these, 12 NIS are new to the Baltic (Table 3), while 14 were involved in secondary spread within the Baltic, i.e., previously were known from at least one of the 10 recipient regions. However, it is important to distinguish between the NIS, which were known in the Baltic Sea before and after the previous assessment. In this case, one species, the brackish water clam Rangia cuneata, first in the Baltic was recorded in 2010 in the Russian part of the south-eastern Baltic (RU_K) and during the assessment period have spread to three other recipient regions: Poland, Lithuania, and Estonia (Table 3). The sedentary polychaete Hypania invalida was found simultaneously in a water body shared between Germany and Poland (Szczecin Lagoon); therefore, the primary introduction is ascribed to 2 countries.

The difference between nNIS L1 and L2 for a recipient region indicates the number of species, which this particular region received during the assessment period due to secondary spread from other parts of the Sea. For example, since the initial assessment 12 NIS were recorded for Germany, i.e., nNIS L1_Germany = 12 (Table 3). Of these, seven were primary introductions to the Baltic Sea (nNIS L2_Germany = 7), and one of these six (the amphipod Echinogammarus trichiatus) is new at the level of the larger biogeographical region, comprising both the Baltic Sea and the North Sea LME (nNIS L3_Germany = 1). In Poland, seven new species were recorded (nNIS L1_Poland = 7). Two of them were new for the Baltic (nNIS L2_Poland = 2): H. invalida and the tubificid oligochaete Limnodrilus cervix, while R. rangia which first appeared in 2011 was not counted at L2 as it was earlier recorded in the neighboring region of Russia. Finally, L. cervix was new at the scale L3, i.e., this species entered the Baltic Sea and North Sea via the Polish coast. In Sweden six new NIS appeared, however only one of them was new at the LME level, i.e., nNIS L2_Sweden = 1.

The pathway analysis at the level of the Baltic Sea LME (L2) reveals that “Vessels” were responsible for 10 primary inoculations, involving ballast water, ballast tank sediments, hull fouling, etc. (AquaNIS, 2016; Table 3), with levels of certainty ranging from “Direct evidence” (the sea anemone Diadumene lineata found on ship hull) to “Very likely” for three and six “Possible” primary introductions. The pathway “Natural spread from neighboring regions,” indicating secondary spread of NIS was ascribed for four primary introductions. The pathway “Culture activities” involving the vectors aquaculture equipment, stock movements and releases and escapees, was ascribed for two primary introductions.

At the scale of a larger biogeographical region, covering two neighboring LMEs (L3), four species were found to be new for both seas, while 18 were known earlier from the North Sea and may have spread from there to the Baltic Sea.

A total of 33 NIS were registered in the three Italian recipient regions since 2010, including 24 species previously not registered in Italy, and a further 9 NIS, introduced before 2010, which have spread further to adjacent coastal regions (AquaNIS, 2016). New arrivals have been recorded in similar numbers in all three coastal regions: 12, 14, and 15 NIS for the Adriatic Sea, Italian Eastern and Italian Western Mediterranean, respectively (Table 4). Some of these nNIS appeared in multiple regions along the Italian coast, having spread rapidly.

The number of likely pathways responsible for the new introductions is higher in Italy than in the Baltic Sea region: besides vessels and culture activities, Italian coasts have been receiving a worrying high number of species that have likely entered the Mediterranean through the Suez Canal, including species of high concern for human health (e.g., the stinging jellyfish Rhopilema nomadica and the toxic pufferfish Lagocephalus sceleratus), as well as species possibly associated with aquarium trade.

In a recent review, Marchini et al. (2015), highlighted that inventories listing cumulative numbers of marine alien species in the Mediterranean Sea present several types of uncertainty, unfortunately resulting in a confused picture of the phenomenon. Problematic species identifications, doubtful records and unknown native origin affect large portions of such inventories, thus preventing a comprehensive assessment of marine bioinvasions (e.g., Katsanevakis et al., 2016). However, modern taxonomy and molecular tools (Zaiko et al., 2015; Bucklin et al., 2016; Raupach et al., 2016; Viard et al., 2016), combined with a growing scientific concern for marine bioinvasions, are contributing to improve the quality of the current species records. In other words, while it is extremely challenging to reliably reconstruct the past history of redistribution of species due to human intervention (Carlton, 2009; Clavero, 2014), we now have more effective and accurate tools to measure the changes currently occurring. Modern records of new arrivals are often delivered with more in-depth analysis of possible vectors and more detailed knowledge on the NIS ecology (e.g., Reusch et al., 2010), and can therefore provide higher-quality knowledge to support scientifically based advice on management decisions. Therefore, an indicator based on new arrivals (nNIS), despite its inherent time-limitation, could offer a more reliable picture of the problem of bioinvasions and supply information essential for early warning initiatives, horizon-scanning programs (sensu Roy et al., 2014) and identification of an environmental target for the MSFD Good Environmental Status (GES) descriptor 2 “Non-indigenous species” (European Commission, 2008). Further, such an indicator enables an assessment how effective implemented vector and pathway management measures are.

Defining the environmental target for the nNIS indicator, the following considerations should be taken into account:

nNIS, as such, has little information value without further breakdown to pathways and vectors involved in transmission of new species, therefore an environmental target should be related to a particular pathway/vector;

Thus, in general, the environmental target for nNIS may be formulated for a country as “No new primary introductions of NIS by a particular pathway/vector per assessment period via the territory of that country.” For example, it may sound like “No new primary introductions of NIS by ship's ballast water to the Baltic Sea via territory of Lithuania for the assessment period,” i.e., the environmental target is: “nNISL2_Lithuania (by ballast water) = 0.” In this case, the environmental target will be achieved if during the assessment period no NIS, new for the entire Baltic Sea, entered the marine area under jurisdiction of Lithuania by ships ballast water. Thus, only primary introductions at the level of entire LME (L2) are counted, i.e., secondary spread is excluded.

The environmental target should be harmonized at the level of LME or a larger region, including several neighboring LMEs, where secondary dispersal of NIS may take place with currents and other natural means. For example, the Baltic Sea Action Plan (HELCOM, 2007) sets the environmental management objective “No introductions of alien species from ships.” In fact, nNIS indicates the success or failure of the preventing measures and its highest target value could be set as “No new human-mediated primary introductions on the level of the European regional seas,” although this seems to be impossible with the management options we have today.

The size of the assessment unit for the nNIS indicator may vary depending on the practical considerations, management needs, and, naturally, data availability. In this study, the assessment unit was equal to a country marine area within an LME or LME sub-region, i.e., to a recipient region as it is defined in AquaNIS (2016). Such subdivision is determined by practical needs, because the management decisions on preventive measures are taken at the level of national authorities.

The smallest possible level is a port and/or its vicinities, where it is practical to perform a NIS survey. Such biological surveys in ports are obligatory, for example, for taking decision on granting exemptions under BWMC (David et al., 2013; David and Gollasch, 2015; Olenin et al., 2016).

The largest level for the nNIS assessment is a marine region under a regional convention, e.g., North-East Atlantic (OSPAR) or Baltic Sea (HELCOM), where measures to prevent NIS introduction may be practically coordinated. The higher geographical scale to calculate nNIS including all regional seas of an entire continent so far is not achievable, because such datasets do not exist, or, at least are not publically available.

The technical precondition for the calculation of nNIS is the availability of verified and continuously updated source of information on introduction events, e.g., a NIS database. Depending of the size of the assessment unit, the geographical coverage of such a database may range from national to regional or interregional. Ideally, such information source should be global or, at least, continent-wide. For example, AquaNIS, used in this study, is being regularly updated by the ICES Working Group on Introductions and Transfers of Marine Organisms and contains information from other world regions as well.

Our results show that the average number of new arrivals having occurred in the past few years in different European sea regions is high, and therefore a reliable database can be achieved only by a continuous and scrupulous work of data checking and update. However, a scientific community needs to achieve the long-term maintenance and reliability of such databases, because they require frequent updating and corrections (Costello et al., 2014). Without continuous maintenance, update and data quality control, the usefulness of the database diminishes over time and its users may be hampered by outdated and therefore misleading information (Olenin et al., 2014) as it was shown in the Mediterranean Sea case (Marchini et al., 2015).

Ideally, all newly published records of NIS introduction events in journals, such as Biological Invasions, Aquatic Invasions, BioInvasion Records, Marine Biodiversity Records, Mediterranean Marine Science, shall be standardized and immediately entered in a global online NIS information system. That would speed up the transfer of knowledge on biological invasions and aid the analysis of new arrivals in relation to all other relevant data stored in the database.

Another technical precondition is that nNIS are scored for recipient regions or LMEs, where regular NIS surveys or, at least, a well-established long-term biological monitoring is in place (Olenin et al., 2011; Lehtiniemi et al., 2015). Taking into account that there are several international instruments (BWMC, ICES Code of Practice, MSFD, etc.) which include NIS monitoring for management purposes and for measuring progress toward achieving their goals, it would be feasible to coordinate NIS surveys regionally. For example, rapid assessment surveys focused on target species (e.g., Minchin et al., 2009, 2016) may be arranged simultaneously by several countries within an LME, in the same way as it is done for fishery surveys (ICES, 2014).

It should be taken into account, also that in some cases our ability to distinguish between the primary introduction and secondary spread may be limited. This is because, multiple introductions of a NIS from outside an LME area also possible as in the case of the American comb jelly Mnemiopsis leidyi “very-likely” spread via ballast waters from two distinct source populations from the western Atlantic to the Black Sea and the North Sea (Reusch et al., 2010). Multiple introductions make distinctions between primary introduction and secondary spread within an LME or larger biogeographical region more difficult. Development of eDNA techniques could assist in determining origin in the future (Rius et al., 2015).

It is difficult to predict those nNIS that may become invasive and cause problems to the environment, economy and/or human health in a recipient region. No control or eradication of invasive alien species without affecting other components of the ecosystem is feasible after an invasion process is underway and a species has become established within an ecosystem. Given the severity of problems that can be caused by biological invasions, it is mandatory for policy and management to focus on the pathways and vectors with the aim to prevent further introductions (e.g., European Commission, 2014).

The nNIS indicator evaluates the effectiveness of prevention measures where these can be employed, be it the ballast water management, precautionary approach in aquaculture or life food trade regulation. For example, strict ballast water management rules applied in recent decades in US and Canada resulted in no new fresh water introductions (i.e., nNIS L3 = 0) in the Laurentian Great Lakes Region since 2006 (Scriven et al., 2015). In contrast, the nNIS value for Italy, obtained in the present study, clearly shows a high exposure arising from a large geographical dispersion of introduction events and several pathways. For example, the relatively high number of newly arrived NIS that can be associated with aquarium releases (3 out of 24) indicates a requirement for a greater implementation of a code of practice in aquarium trade (Scalera et al., 2012), as well as a greater general public awareness. There is a need to review management for those NIS that might be prevented from becoming established where direct anthropogenic introductions can be regulated.

It is important to manage primary introductions at the scale L2 and L3, because the secondary spread, which can be inevitable, may seriously compromise the ability for any practical regulation. This is because a further spread may involve dispersal by the same pathway or by multiple pathways that might act in relay. What is not possible to manage are the natural processes involving tidal movements, alongshore drift, waterfowl, and other aquatic biota as a carrier of NIS either within an LME or from neighboring LME.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.