Transylvania is the largest region of Romania, covering an area of 102,226 km² (43% of Romania). It is bordered by the Carpathian Mountains in the north, east, and south, and by the Pannonian Plain in the west (see Fig. 1). It encompasses three biogeographical regions: Alpine (29.7%), Continental (55.3%), and Pannonian (15.0%). The majority of the larger rivers in the region originate from the Carpathians and flow towards the west (Pannonian Plain). The Mureș is the longest river of Transylvania, stretching over 766 km (718 km in Romania) with a discharge/outflow of 165 m³/s at the Romanian-Hungarian border. Other large rivers include the Someș River (Someșul Mic and Someșul Mare form the united Someș), which has a length of 345 km in Romania (Transylvania) and a discharge of 118 m³/s, the Criș Rivers (Crișul Repede with 148 km, Crișul Negru with 144 km, and Crișul Alb with 238 km), Bega River (244 km), and the Timiș River (359 km). The lowest sampling site from our study is located at 78 m above sea level on the Timiș River, while the highest sampling site is situated at 1356 m above sea level on the Someșul Rece River.

Figure 1. 

Position of Transylvania, Romania, in Europe, showcasing the main rivers, river basins, Natura 2000 sites (SCIs – Site of Community Interest), and the sampling sites surveyed in the study.

In the past two centuries, river regulations have affected Transylvanian rivers, especially floodplains and marshes. The largest marsh, the Ecsed Moor, situated on the boundary between Romania and Hungary, was drained in the 19^th century. Another important lowland floodplain, the Ier River valley, was also drained. Most of this land has been converted into agricultural land. After accession to the European Union in 2007, agriculture intensified significantly, with monocultures taking priority over small parcels of land, which probably have an effect on the fish communities of the rivers. Although there are relatively few large cities in the region, the numerous villages may pose an important source of pollution. The five largest cities in Transylvania are Cluj Napoca (with 286,598 inhabitants), Timișoara (250,849), Brașov (237,589), Oradea (183,105), and Arad (145,078) (2021 population census).

Data were collected between 31 March 2007 and 29 October 2022 from a total of 679 sampling sites, including all rivers and major tributaries in Transylvania (Fig. 1). Our survey was focused on rivers, still we occasionally sampled backwaters, ponds, and drainage channels with stagnant water to gather data on species that inhabit these waters and are threatened by habitat loss. Standing waters represent 6% of the total sampling sites. Fishponds and artificial lakes were not surveyed at all. We chose the location of sampling sites to ensure relatively uniform coverage of each river. Although we invested higher sampling efforts in some areas, we attempted to achieve representative coverage of all rivers (Fig. 1). Fishing was carried out with a 12V battery-powered electrofishing device (Samus 725 or Samus 1000), and data collection was performed by wading (Sály et al. 2009). This method consists of moving slowly upstream in the shallow waters and fishing on one side of the stream in a single pass. Due to current legislative constraints, we were unable to fish from inflatable boats, which is preferable on larger rivers and lakes. Therefore, the fishing method used in our study may underestimate the presence and abundance of species inhabiting deep and/or large (stagnant) water bodies. The length of the sampling sites was set to 150 m in small and medium-sized rivers and 200–300 m in larger rivers, although occasionally, the length of the sampling sites had to be adjusted according to the local field conditions (e.g., shortened if the site was inaccessible). After capture, identification, and taking occasionally morphometric measurements, all individuals were released in good condition. The raw survey data are stored in the OpenFishMaps database, an open-source database available at https://openfishmaps.ro/, and are available in Suppl. material 1.

We compiled the distribution maps of species using the data from our survey and information provided by anglers, angling associations, fish stocking projects, or the Facebook page “Ichthyology of Romania” (https://www.facebook.com/groups/ichthyologyofromania). From the sources other than our own capture data we only used data that were supported by documentary photographs so that the species could be accurately identified. Information obtained from these sources and additional personal occurrence data for four species (Cottus gobio, Eudontomyzon mariae, Sabanejewia romanica, Umbra krameri) is not included in the raw data of distribution (Suppl. material 1) but is indicated separately in the distribution maps (Suppl. material 2). We excluded all hybrids from the dataset, particularly the Barbus barbus × Barbus biharicus, Romanogobio uranoscopus × Gobio gobio sensu lato, and the Barbus petenyi × Barbus barbus. The two Sabanejewia species previously belonging to S. aurata (S. balcanica and S. bulgarica) were treated as Sabanejewia sp. because the identification of these species in some rivers was uncertain.

The survey data stored in the OpenFishMaps database were exported to the R statistical environment (v. 4.2.2; R Core Team 2022) using ESRI shape files. As the database contains data from all regions of Romania, we first selected only the data points within the boundary of Transylvania, then we applied descriptive statistics. For the graphical visualization of the data, to better showcase the distributional patterns of species and different groups of species within Transylvania, we assigned to all sampling sites the cell codes of the overlapping 50 × 50 km ETRS grid, and counted the total number of species, number of native species, number of non-native species, and the number of Natura 2000 species (found in Annex II of the EU’s Habitat Directive) for each grid cell. Detailed distribution maps for each species separately are provided in Suppl. material 2. The filtering of the spatial data was performed with the “sf” R package (Pebesma 2018), while data operations were performed with the “dplyr” R package (Wickham et al. 2022). Data visualization was performed using “tmap” R package (Tennekes 2018) and Quantum GIS (version 3.22; QGIS Development Team 2022).

Out of the 77 identified species recorded during our survey, 19 (24.7%) are introduced, while 24 (31.2%) species belong to Natura 2000 species. Of the total of 129,212 captures, 6,553 individuals (5.1%) belong to non-native species, and 46,497 individuals (36%) belong to Natura 2000 species. Overall, the abundance of non-native species in our study region can be considered relatively low. For example, in Hungarian waters, 28.8% of identified species and 18.3% of total captures are non-native (Takács et al. 2017). In various parts of the Mediterranean Basin 25% of fish species are non-native, and in the Iberian Peninsula, where the majority of Europe’s threatened fish populations can be found, the proportion of alien species reaches 50% (Leprieur et al. 2008; Clavero et al. 2010; Maceda-Veiga 2013). The distribution of both native and Natura 2000 species throughout Transylvania (Figs 3, 4) and the small number and low abundance of non-native species (Fig. 5) demonstrates that the ichthyofauna in the rivers of Transylvania is much closer to a natural state. Although anecdotal evidence suggests that stocking is still low, irresponsibly repopulating river sectors and lakes could potentially exert significant pressure on the river ecosystems from this region. For instance, Salvelinus alpinus is reported for the first time in Transylvanian natural waters. A few individuals were caught by fisherman in the Someșul Cald River upstream of the Fântânele reservoir, where it is presumed the species was introduced without authorization.

Comparing our data with those collected during the last comprehensive survey by Bănărescu (1964, 1969), major changes can be observed in the distribution of several species. We present the status of these species below.

Carassius carassius was prevalent in most floodplains in the past (i.e., before 1964) but has now vanished from most of its former habitats (Suppl. material 2: map S17). On the other hand, Carassius gibelio, which was present in only a few habitats before 1964, has now expanded its distribution over the main rivers of Transylvania, excluding mountainous habitats (Suppl. material 2: map S18).

Our data indicates that the distribution range of Zingel zingel has decreased, as the species has disappeared from the Someșul Mare, Someșul Mic, Crișul Repede, Olt Rivers and the middle part of the Mureș River. We found viable populations of the species in the Someș, Crișul Negru, Crișul Alb, Mureș, and Timiș Rivers, and a very fragile population in the Bega River (Suppl. material 2: map S77). Zingel streber has apparently disappeared from the Tur, Someșul Mic, Crasna, Barcău, Arieș, and Bega Rivers (Suppl. material 2: map S76). Our observation supports the findings of Brinker et al. (2018), who also noted a reduction in the historical range of the species in the upper Danube basin due to population fragmentation and habitat loss. It has to be mentioned though that our fishing method is not proper for evaluating populations of Zingel species (Szalóky et al. 2021), therefore the occurrence and abundance of these two Zingel species might be underestimated. Our results suggest that the species still maintains significant populations in the Mureș, Crișul Negru, Crișul Alb, and Nera Rivers. The last recorded sighting of the species in the Someș River was in 1964 (Bănărescu 1964), although several studies have been conducted on the ichthyofauna of the Someș River since then (Bănărescu et al. 1999; Năstase and Oţel 2016). Our survey found the species at five sampling sites along the Someș River and at one site along the Someșul Mare River. When studying the ichthyofauna of the Mureș River, Nalbant (1995) only found a few individuals of this species in the fishermen’s catch at the Gura Arieșului locality. Our survey found viable populations in the lower and upper-middle part of the Mureș River (the species was present at 24 sampling stations). The species is also present in the Timiș, Olt, and Târnava Rivers, but in much smaller numbers (Suppl. material 2: map S76).

We have observed a drastic reduction in the distribution of Gymnocephalus schraetser, as this species was not identified during our surveys, except a few records from other verified sources, although it was found in several rivers (Mureș, Crișul Repede, Crișul Negru, Crișul Alb Rivers) in the 1990s (Nalbant 1995; Bănărescu et al. 1999) and later in the Timiș River (Bănăduc et al. 2013). The species is still present in the Someș, Timiș, and Crișul Negru Rivers (single individuals were observed by local anglers; Suppl. material 2: map S35). Harka and Csipkés (2009) observed a similar drastic contraction of distribution in the Hungarian Bodrog River. Further surveys are needed to map the remaining populations of this species.

Umbra krameri has disappeared from most of its known habitats, particularly from the Ier River valley and from the Carei Plane in north-west Transylvania. In a survey, the species was found only in two out of 13 sites where the species was formerly recorded (Wilhelm 2008a), but three new populations were discovered in the upper valley of the Ier River and one new, but fragile population in the Homorod River of the Crasna River basin. A new population was also found in the Timiș River basin by Covaciu-Marcov et al. (2018) and confirmed by the present study (Suppl. material 2: map S74). These new findings are likely not due to a range expansion of the species in recent decades, but rather because this region of the Romania is understudied.

The presence of Romanogobio vladykovi has increased as a result of human activities in the Tisa River basin (Telcean and Bănărescu 2002) and our study confirms former findings (Suppl. material 2: map S59). Another species, Leuciscus leuciscus, which had only one confirmed occurrence in the 1990s, has shown significant recovery and was detected in the catchment area of 7 rivers (Suppl. material 2: map S43).

Hucho hucho has returned to the upper Mureș River basin due to stocking (Cengher 2007) after a few decades of absence, but the construction of the Răstolița dam may affect the survival of the species. We found viable population of the species in the Tisa and upper Mureș River basins. (Suppl. material 2: map S36).

Although the method used in our study was moderately suitable for assessing Cyprinus carpio populations, our data indicates a massive decline of the species (Suppl. material 2: map S26), especially of the wild form. This change is possibly due to hybridization and river regulation (Freyhof 2010).

Eudontomyzon mariae is reported for the first time in Transylvanian waters and is present in the Olt River basin (Suppl. material 2: map S29). A new, large population of Cobitis elongata was found in the Caraș River, in addition to the previously known population in the Nera River basin (Suppl. material 2: map S20).

Three species, Babka gymnotrachelus, Neogobius fluviatilis, and Perccottus glenii, have recently appeared in Transylvania (Covaciu-Marcov et al. 2011; Cocan et al. 2014, 2016). Our observations indicate that these species have expanded their range of distribution. Babka gymnotrachelus was found in the Timiș River basin (Suppl. material 2: map S7), Neogobius fluviatilis was found in the Someș, Timiș, and Olt River basins (Suppl. material 2: map S46), and Perccottus glenii was found in the Tur, Crișul Repede, and Bega Rivers (Suppl. material 2: map S50). The rapid spread of Perccottus glenii in Europe and its impact on native fish fauna (Koščo et al. 2003; Reshetnikov 2003, 2013; Reshetnikov and Ficetola 2011; Horvatić et al. 2022) raise concerns, as it is already present in three Transylvanian rivers and its further spread is expected, posing a significant threat to the native fish fauna, especially to the vulnerable Umbra krameri (Grabowska et al. 2019).

Pseudorasbora parva, a non-native species, was not present in Transylvania before 1964, but we found it in almost all river basins and at 19.9% of the sampling sites (Suppl. material 2: map S54). Lepomis gibbosus was present only in the western part of the region before 1964, but we found it in most of the river basins surveyed (Suppl. material 2: map S39). Ameiurus nebulosus was the dominant Ameiurus species in Transylvania’s waters until the 2000s (Wilhelm 2013), but it has now been almost completely replaced by Ameiurus melas (Suppl. material 2: maps S5, S6). Only one specimen of A. nebulosus was identified in the Tur River. This replacement of A. nebulosus is similar to what has been observed in Hungarian waters (Takács et al. 2017) and confirms the observation of Jaćimović et al. (2019) regarding the invasive potential of this species.

Many of the Natura 2000 sites from Transylvania have been designated predominantly in mountainous areas to enhance the protection of Natura 2000 fish species, although only a few of these species occur there (as seen in E530N255 and E535N255). However, important river sectors in hilly and lowland areas, which have a high number of Natura 2000 fish and lamprey species, remain unprotected (such as parts of the Crișul Alb River from E520N265, the Bega and Bega Veche River from E515N255, the lower part of the Niraj River from E540N270, and important sectors of the Someș River and the middle and lower part of the Lăpuș River at E530N280). These hilly and lowland river sectors require protection as they are vulnerable to anthropogenic disturbances, particularly due to river regulation. These areas are home to most of the native and Natura 2000 species (Figs 3, 4). Furthermore, the presence of rare or endangered species alone is a sufficient reason to protect an entire aquatic habitat (such as the Umbra krameri in the Timișul Mort River or Homorodul Vechi River). Therefore, we propose that in future designation of protected areas, not only the number of Natura 2000 species should be considered (Fig. 3), but also species occurrence maps. This is because, in some cases, the presence of rare or endangered species is a compelling argument for the establishment and designation of a protected area. Based on our survey, we suggest several river sectors for inclusion in the Natura 2000 network to ensure the protection of the most valuable and diverse river stretches (as listed in Table 2).

Considering that Romania does not currently have an officially adopted Red List for fish and lamprey species, our results can contribute to the creation of such a list. Some species, although not listed as Natura 2000 species, are of prime conservation concern due to drastic reductions in their distribution (e.g., Carassius carassius, Leucaspius delineatus, Lota lota, Tinca tinca) or diminished abundance (Thymallus thymallus) in recent decades. It is crucial to assess their current conservation status to ensure their long-term survival. Many environmental impact assessments are hindered by a lack of up-to-date data on fish fauna, and often rely on assessments that are not appropriate for studying fish communities. Our data can provide valuable information for these conservation studies.

The species pool and distribution of some species in the study area is likely greater than what our survey shows due to several reasons. Firstly, we mainly sampled rivers and several species that were recorded in the past or are present in neighboring countries are expected to occur in the area, especially in stagnant or enclosed water bodies. Although we did sample a few backwaters to gather data on species that inhabit stagnant waters and are threatened by habitat loss (e.g., Carassius carassius, Leucaspius delineatus, Umbra krameri), a comprehensive survey of these habitats was not conducted. Additionally, the fish fauna of thermal springs and lakes was excluded from the study, despite of some of these habitats are known to host exotic fish populations (Bănărescu et al. 1997). These waters are important sources and dispersal hotspots for some aquaristic cultivated fish species (Takács et al. 2015; Weiperth et al. 2015; Kordás and Juhász 2020) and may also serve as starting points for invasive species. Further sampling of these water bodies is likely to increase the number of introduced species in the Transylvanian fish fauna. Finally, the absence of certain species from our survey (e.g., Pelecus cultratus) and apparent gaps in the distribution of others, such as Abramis brama, Leuciscus aspius or Sander lucioperca, can be attributed to the limitations of our sampling method.

No conflict of interest was declared.

No ethical statement was reported.

No funding was reported.

The study was designed by AAN and PLP. AAN, II collected data from the field investigations, NE, GB and AF analyzed the data. AAN and PLP wrote the manuscript with critical feedback from all co-authors. All authors gave final approval for publication and agreed to be accountable for the aspects of work that they conducted.

Nándor Erős https://orcid.org/0000-0001-8921-482X

Attila Fülöp https://orcid.org/0000-0001-5337-336X

Péter László Pap https://orcid.org/0000-0002-3659-7684

All of the data that support the findings of this study are available in the main text or Supplementary Information.