Introduction

Freshwater eels, genus Anguilla Schrank, 1978, are one of the most unique groups of Anguilliformes and consist of over 800 species (Arai 2020). Freshwater eels are currently recognized as 19 species and subspecies (Ege 1939; Watanabe et al. 2004; Arai 2016, 2020, 2021) and are split into tropical eels (13 species/subspecies) and temperate eels (6 species/subspecies) based on their biogeography (Arai 2020) (Table 1, Fig. 1). Although a large number of research outcomes have been accumulated for centuries, vital biological and ecological aspects of freshwater eels have not been unveiled. Spawning grounds of freshwater eels are located in the open ocean, and hence, no spawning has ever been seen in the realm of nature. Freshwater eels display catadromous life patterns as they migrate between continental growth habitats and offshore spawning grounds. In addition to biological and ecological interests, freshwater eels are an economically important fish species and are widely consumed, especially in Asian and European countries, as delicacy food resources. Populations of freshwater eels are extremely vulnerable to exploitation and global climate change. Drastic declines in glass eel recruitments in the temperate eels and tropical eels have been reported worldwide (Dekker 2003; Doole 2005; Arai 2014a, b, 2021; Jacoby et al. 2015; Jellyman 2021).

Table 1 Nineteen species and subspecies of freshwater eels of the genus Anguilla and their geographical distribution
Full size table
Fig. 1
figure 1

Biogeography of freshwater eels of the genus Anguilla. The detailed geographical distribution in each species is summarized in Table 1. Blue: temperate eels, red: tropical eels. The base map was downloaded from the OpenStreetMap and OpenStreetMap Foundation (open access) at https://www.openstreetmap.org/#map=2/7.2/-23.9 (Data©OpenStreetMap contributors)

Full size image

The life history of freshwater eels involves five developmental stages: the leptocephalus (larva), glass eel (juvenile), elver (pigmented glass eel, young eel), yellow eel (immature eel) and silver eel (mature eel) stages (Tesch 1977). Freshwater eels can be distinguished into three migratory phases throughout their lives: oceanic migration to continental habitats during the leptocephalus and glass eel stages, continental migration during the glass eel, elver, yellow eel and silver eel stages and oceanic migration to spawning grounds during the silver eel stage (Fig. 2). During oceanic migration to continental habitats, the metamorphosis from leptocephalus to glass eel is the most remarkable event throughout their lives. The timing of metamorphosis and the length of the leptocephalus stage play a vital factor in determining the overall geographical ranges in freshwater eels (Arai et al. 2001, 2020). A longer larval period during the oceanic stage is believed to establish the global geographical distribution and the consequent speciation in anguillid eels. The leptocephalus periods prolonged from several months to nearly one year in temperate eels A. anguilla (Lecomte-Finiger 1992; Arai et al. 2000a; Wang and Tzeng 2000), A. japonica (Cheng and Tzeng 1996; Arai et al. 1997; Leander et al. 2012), A. rostrata (Wang and Tzeng 1998, 2000; Arai et al. 2000a), A. australis australis (Arai et al. 1999a; Shiao et al. 2001, 2002), A. australis schmidtii (Arai et al. 1999a; Shiao et al. 2001) and A. dieffenbachii (Marui et al. 2001) and tropical eels A. celebesensis (Arai et al. 1999b, 2001, 2003a; Marui et al. 2001), A. marmorata (Arai et al. 1999b, 2001, 2002a, b; Marui et al. 2001; Robinet et al. 2003, 2008; Réveillac et al. 2008; Leander et al. 2012; Hewavitharane et al. 2019), A. bicolor bicolor (Arai et al. 1999b; Robinet et al. 2003, 2008), A. bicolor pacifica (Arai et al. 1999c, 2001; Marui et al. 2001), A. reinhardtii (Shiao et al. 2002), A. mossambica (Robinet et al. 2003, 2008; Réveillac et al. 2009), A. bengalensis labiata (Robinet et al. 2003), A. luzonensis (Han et al. 2016), A. megastoma (Hewavitharane et al. 2019) and A. obscura (Hewavitharane et al. 2019). Recruitment to the coast in tropical glass eels occurs throughout the year (Arai et al. 1999d; Sugeha et al. 2001a; Robinet et al. 2003; Shen and Tzeng 2007; Leander et al. 2012; Hewavitharane et al. 2018), and it is considered to be due to year-round spawning (Arai et al. 2001, 2016; Shiao et al. 2002; Arai and Abdul Kadir 2017a) and a constant larval period (Arai et al. 2001). On the other hand, seasonal recruitment in temperate glass eels is considered to be due to the seasonal spawning (Shiao et al. 2001; Leander et al. 2012).

Fig. 2
figure 2

Migration of freshwater eels throughout their whole lives. The migration of freshwater eels can be distinguished into three migratory phases: oceanic migration to continental habitats, continental migration and oceanic migration to spawning grounds. The scales of two oceanic migration phases can display short (local), middle and large (global) scales. During continental migration, freshwater eels have four migration patterns: sea eels (freshwater eels live in marine water), estuarine eels (freshwater eels live in brackish water), river eels (freshwater eels live in fresh water) and habitat shifters (freshwater eels shift habitats in various saline waters). S: spawning grounds

Full size image

During the continental growth phase after recruitment, glass eels and elvers grow as yellow eels and live in coastal waters, estuaries, rivers, ponds and lakes. The immature yellow eels undergo another metamorphosis into silver eels accompanied by the onset of gonad maturation mainly in the autumn and winter in temperate eels, and thereafter, they initiate oceanic migration to the spawning ground (Tesch 1977). In tropical eels, however, the period of downstream migrations extends throughout the year in certain species and regions (Arai et al. 2016; Arai and Abdul Kadir 2017a). During the continental migration phase, otolith microchemistry studies have revealed that some yellow and silver eels never migrate into freshwater but spend their entire life history in the ocean. Studies on the incorporation of strontium (Sr) into otoliths of freshwater eels found that the otolith Sr:calcium (Ca) ratio increased linearly with salinity but was less influenced by other factors such as water temperature, food and physiologies in fish (Tzeng 1996; Lin et al. 2007; Arai and Chino 2017). Thus, otolith microchemistry can trace whether each fish truly enters freshwater after oceanic migration to continental growth habitats and lives in freshwater, brackish water or marine water during the continental growth period just before the initiation of oceanic migration to spawning grounds or whether fish shift habitats and live in separate saline areas. The application of otolith microchemistry to reconstruct the migratory history has also elucidated diverse migration to those of marine residents (sea eels) and freshwater residents (river eels) in temperate eels A. anguilla (Tzeng et al. 1997, 2000; Arai et al. 2006, 2009, 2019; Shiao et al. 2006), A. japonica (Tsukamoto and Arai 2001; Arai et al. 2003b, c, 2008, 2009; Kotake et al. 2003, 2005; Shiao et al. 2003; Tzeng et al. 2003; Chino and Arai 2009), A. rostrata (Jessop et al. 2002, 2008; Lamson et al. 2006), A. australis schmidtii and A. dieffenbachii (Arai et al. 2004) and tropical eels A. marmorata (Shiao et al. 2003; Chino and Arai 2010a, b; Lin et al. 2012; Arai et al. 2013; Arai and Chino 2018), A. celebesensis (Chino and Arai 2010b), A. bicolor bicolor (Chino and Arai 2010b, c; Arai and Chino 2019; Arai et al. 2020), A. bicolor pacifica (Briones et al. 2007; Arai et al. 2013; Arai and Chino 2019) and A. bengalensis bengalensis (Arai et al. 2020). These studies clearly found brackish water residents (estuarine eels) and habitat shifts between separate salinity regimes (habitat shifters). Certain ratios and populations of freshwater eels shift their habitats intermittently between different saline water environments during their continental growth period. Therefore, it is evident that freshwater eels do not necessarily colonize and live in freshwater habitats showing more opportunistic catadromous migration during their continental growth period because diverse migratory histories such as sea eel, estuarine eels and habitat shifters, have commonly been found in many freshwater eel species in addition to the typical catadromous migration of river eels (Arai and Chino 2012; Arai 2020).

Among the three migratory phases throughout their lives, oceanic migration to spawning grounds is still unknown due to the difficulty in tracking silver eels from the continental habitats to offshore spawning grounds. Pop-up satellite archival transmitting tags (PSATs) have been powerful devices to understand the migration route and behaviour for longer periods during their oceanic migrations for spawning in freshwater eels. PSAT studies have been studied in temperate eels A. dieffenbachii (Jellyman and Tsukamoto 2002, 2005, 2010), A. anguilla (Aarestrup et al. 2009; Wahlberg et al. 2014; Westerberg et al. 2014; Wysujack et al. 2015; Righton et al. 2016), A. japonica (Manabe et al. 2011; Chen et al. 2018; Higuchi et al. 2018) and A. rostrata (Béguer-Pon et al. 2012, 2015) and tropical eels A. obscura (Schabetsberger et al. 2013, 2015, 2019; Chen et al. 2018), A. megastoma (Schabetsberger et al. 2013, 2015, 2019; Chang et al. 2020) and A. bicolor pacifica (Chen et al. 2018) and A. marmorata (Chen et al. 2018; Chang et al. 2020). PSAT studies in A. anguilla in the Baltic and Mediterranean seas have been approached for years. The released tagged A. anguilla generally migrated to subtropical regions, and thereafter the eels further travelled to the spawning ground in the Sargasso Sea (Wysujack et al. 2015; Amilhat et al. 2016; Righton et al. 2016). In 2014, one tagged A. rostrata was successfully tracked completely from the continental shelf to the spawning area in the Sargasso Sea crossing the Gulf Stream (Béguer-Pon et al. 2015). One tagged A. marmorata travelled over 800 km towards the possible spawning ground in the South Equatorial Current from Vanuatu (Schabetsberger et al. 2013). All freshwater eels exhibited diel vertical migration (DVM) behaviour during their oceanic migration. All freshwater eels commonly swam deeper water (500–700 m) during the daytime and shallower water (100–300 m) during the nighttime. The migration ecology of freshwater eels is gradually revealed throughout their lives.

This paper examines the migratory ecology of freshwater eels in three phases throughout their lives, i.e., oceanic migration to continental habitats during the leptocephalus and glass eel stages, continental migration during the glass eel, elver, yellow eel and silver eel stages and oceanic migration to spawning grounds during the silver eel stage (Fig. 2). Migration ecology research would also contribute to the management, protection and conservation of freshwater eels, which are currently declining the populations globally.

Oceanic migration to continental habitats

Considerable information on the oceanic migration phase to continental habitats in freshwater eels was revealed by examining otolith microstructure and microchemistry. The initiation of metamorphosis from leptocephalus to glass eel or period of leptocephalus stage can determine the otolith microstructural and microchemical signatures of an abrupt increase in increment width corresponding to a drastic drop in Sr:Ca ratios (Otake et al. 1994; Arai et al. 1997). The completion of metamorphosis (duration of metamorphosis period) can determine the otolith microstructure between the initiation of an abrupt increase and a maximum peak in otolith increment width (Otake et al. 1994; Arai et al. 1997). Based on these criteria, a number of studies regarding the early life history and recruitment such as larval growth, metamorphosis, oceanic migration and colonization of glass eels to continental habitats are conducted for both tropical and temperate freshwater eels. The determination of early life history characteristics depends on the daily growth deposition in otoliths in freshwater eels. The formation of daily growth increments in otoliths has been validated in temperate eels A. japonica (Umezawa et al. 1989) and A. rostrata (Martin 1995) and tropical eels A. celebesensis (Arai et al. 2000b) and A. marmorata (Sugeha et al. 2001b). The formation of daily growth increments in otoliths of freshwater eels supports the determination of various early life history characteristics to understand oceanic migration in freshwater eels. Furthermore, early life history characteristics in combination with the prevailing oceanic currents and population genetic evidence revealed the possible transportation and migration mechanisms in the ocean (Arai et al. 2020; Zan et al. 2020; Arai and Taha 2021; Norarfan et al. 2021).

The timings (ages) of metamorphosis varied among species, and considerable variation was found in A. anguilla (Table 2). Ages at metamorphosis in the temperate freshwater eels are generally older than those of tropical freshwater eels. The mean ages at recruitment also varied considerably among species and were oldest in A. anguilla. Anguilla anguilla was recruited to the estuary at the oldest age, and A. bicolor bicolor was at the youngest age (Table 2). Interestingly, the ages at recruitment of a tropical eel A. bicolor bicolor between the east (148–202 days in Java, Indonesia) and west coasts (68–96 days on Réunion Island) in the Indian Ocean were considerably different (Arai et al. 1999b; Robinet et al. 2003) (Table 2). Such distinctive differences were not found in other freshwater eels (Table 2). Spawning grounds of A. bicolor bicolor are thought to be located in the waters (1) off west Sumatra of the eastern Indian Ocean (Jespersen 1942) and (2) off east Madagascar of the western Indian Ocean (Jubb 1961; Robinet et al. 2003). These two geographically distant spawning grounds are believed to be the occurrence of two divergent populations of A. bicolor bicolor in the Indian Ocean (Arai and Taha 2021; Norarfan 2021). Anguilla bicolor bicolor can be plausibly hypothesized to have different populations on the east and west sides of the Indian Ocean. On the other hand, the ages at recruitment of a tropical eel A. marmorata were less variable and similar in the Indo-Pacific region (Table 2). A multiple-population structure was previously observed in A. marmorata from the Indian and Pacific oceans (Ishikawa et al. 2004; Minegishi et al. 2008; Gagnaire et al. 2009). However, a recent study found that A. marmorata in the Pacific and eastern Indian oceans formed a panmictic population (Arai and Taha 2021). Therefore, A. marmorata might have similar early characteristics in tropical waters.

Table 2 Ages at metamorphosis and recruitment of the freshwater eels of genus Anguilla
Full size table

The mean monthly ages at metamorphosis of tropical eels A. celebesensis, A. marmorata and A. bicolor pacifica ranged from 84 to 95 d, 114 to 158 d and 129 to 171 d, respectively (Arai et al. 2001). Significant differences were found among species; however, no significant difference was found among months in each species (Arai et al. 2001). Furthermore, the mean ages at recruitment of A. celebesensis, A. marmorata and A. bicolor pacifica in each month ranged from 104 to 118 d, 144 to 182 d and 158 to 201 d, respectively (Arai et al. 2001). Significant differences found among species, however no significant difference found among months in each species (Arai et al. 2001). Seasonal variations in the oceanic migration periods (ages at recruitment) were found in temperate eel A. japonica (Kawakami et al. 1999) and tropical eels A. reinhardtii (Shiao et al. 2002) and A. mossambica (Réveillac et al. 2009). These differences would be mainly due to either no seasonal variability or seasonal variation in the velocity of ocean current transportation and directionality from spawning grounds to continental habitats because leptocephali are passively transported by means of oceanic currents.

Anguilla celebesensis, A. marmorata and A. bicolor pacifica leptocephali were found to take approximately 3 to 5 months to metamorphose after hatching (Arai et al. 2001). The earlier-hatched larvae are found to recruit to continental habitats significantly earlier in the recruitment period in a temperate eel A. japonica (Kawakami et al. 1999). In temperate eels, A. anguilla, A. rostrata and A. japonica, recruitment occurs over a limited period (winter to spring) (Tesch 1977; Kawakami et al. 1999), with the spawning period also being limited (Wang and Tzeng 1998, 2000; Arai et al. 1999b, 2000a; Kawakami et al. 1999; Shiao et al. 2001, 2002; Leander et al. 2012). The onset of metamorphosis occurs earlier when eels hatch earlier together with faster growth rates, and thus, the timing of metamorphosis changes following the differences in the spawning period. In tropical eels, however, because of constant larval growth rates and constant age at recruitment throughout the year, spawning is prolonged throughout the year.

Ages at metamorphosis (duration of leptocephalus stage) of tropical eels, A. celebesensis (Arai et al. 1999b, 2001, 2003a), A. marmorata (Arai et al. 1999b, 2001, 2002a, b; Marui et al. 2001; Robinet et al. 2003, 2008; Réveillac et al. 2008; Leander et al. 2012; Hewavitharane et al. 2019), A. bicolor bicolor (Arai et al. 1999b; Robinet et al. 2003, 2008), A. bicolor pacifica (Arai et al. 1999c, 2001; Marui et al. 2001), A. reinhardtii (Shiao et al. 2002), A. mossambica (Robinet et al. 2003, 2008; Réveillac et al. 2009), A. bengalensis labiata (Robinet et al. 2003), A. luzonensis (Han et al. 2016), A. megastoma (Hewavitharane et al. 2019) and A. obscura (Hewavitharane et al. 2019), were 1–11 months younger than those of temperate eels, A. anguilla (7–12 months) (Lecomte-Finiger 1992; Arai et al. 2000a; Wang and Tzeng 2000), A. rostrata (6–8 months) (Wang and Tzeng 1998, 2000; Arai et al. 2000a), A. australis (6 months) (Arai et al. 1999a; Shiao et al. 2001, 2002) and A. dieffenbachii (8 months) (Marui et al. 2001), except for A. japonica (5 months) (Cheng and Tzeng 1996; Arai et al. 1997; Leander et al. 2012) (Table 2). The strong correlation between the timing of metamorphosis and age at recruitment distinctly indicated that earlier metamorphosed glass eels colonize a continental habitat at a younger age in freshwater eels (Arai et al. 1999a, b, c, 2000a, b, 2001, 2002a, b, 2003a; Marui et al. 2001; Shiao et al. 2001, 2002; Robinet et al. 2003, 2008; Réveillac et al. 2009; Leander et al. 2012; Hewavitharane et al. 2019). This evidence suggests that the timing of metamorphosis is a major trigger to initiate active oceanic migration in freshwater eels.

Tropical and temperate eel leptocephali would experience different temperatures during the oceanic migration to continental habitats. Separate temperatures would induce differences in the timing and duration of metamorphosis. General temperatures between the surface and 200 m in open ocean waters where leptocephali were caught are 20 to 25 °C in tropical areas and 10–20 °C in temperate areas (Arai et al. 2001). The spawning grounds of both tropical and temperate eels are located in tropical areas (Arai 2020). Tropical eel leptocephali are transported at higher temperatures in tropical environments; however, temperate eel leptocephali are further transported to lower temperature environments at higher latitudes. Slower growth rates in the temperate eel are predicted because they experienced lower temperatures during the larval migration, leading to an older age at metamorphosis than that of tropical eels. As a result, temperate eels would be further transported to longer distances to temperate regions.

The differences in age at recruitment to continental habitats among the species, e.g. 104–118 d (monthly average range) in A. celebesensis, 144–118 d in A. marmorata and 158–201 d in A. bicolor pacifica recruited to North Sulawesi Island of Indonesia (Arai et al. 2001), 80 d (average) in A. bicolor bicolor, 120 d in A. marmorata and 124 d in A. mossambica recruited to Réunion Island in the Indian Ocean (Robinet et al. 2003), 141 d (average) in A. megastoma, 115–152 d (average range in two recruitment peaks) in A. marmorata and 168 d in A. obscura recruited to Viti Levu, Fiji Islands, in the western South Pacific (Hewavitharane et al. 2019), are believed due to the differences in ages at onset of metamorphosis in each species. Furthermore, ages at recruitment of tropical eels A. bicolor bicolor (Arai et al. 1999b; Robinet et al. 2003, 2008) and A. reinhardtii (Shiao et al. 2002) were 1–12 months less than those of temperate eels, A. anguilla (8 to 15 months) (Lecomte-Finiger 1992; Arai et al. 2000a; Wang and Tzeng 2000), A. rostrata (7–8 months) (Wang and Tzeng 1998, 2000; Arai et al. 2000a), A. australis (7 months) (Arai et al. 1999a; Shiao et al. 2001, 2002) and A. dieffenbachi (10 months) (Marui et al. 2001), respectively, except A. japonica (6 months) (Cheng and Tzeng 1996; Arai et al. 1997; Leander et al. 2012) (Table 2). This may be a result of the ages at metamorphosis in the tropical eels being 1–11 months less than in the temperate eels. Based on the early life history characteristics, scales of oceanic migration to continental habitats would be suggested to be small (local) to middle scales in tropical freshwater eels and middle to large (global) scales in temperate eels (Fig. 2).

The spawning seasons of tropical eels A. celebesensis and A. marmorata (Arai et al. 2001) and A. reinhardtii (Shiao et al. 2002) were found to extend throughout the year. Spawning periods of temperate eels are generally restricted in certain months, i.e., February to April in A. rostrata (McCleave et al. 1987), March to June in A. anguilla (McCleave et al. 1987), July–November in A. japonica (Kawakami et al. 1999), August–December in A. dieffenbachii (Jellyman 1987) and September–February in A. australis (Jellyman 1987). Differences in the seasons of downstream migrations in maturing eels would be the result of the difference in spawning period and timing between tropical and temperate eels. Spawning migrations are generally found to occur from autumn to spring in temperate eels, August to November in A. rostrata (Hain 1975), August–December in A. anguilla (Haraldstad et al. 1985), August–December in A. japonica (Matsui 1952; Kotake et al. 2007), April–May in A. dieffenbachii (Jellyman 1987) and February–April in A. australis (Jellyman 1987). In tropical eels, year-round spawning migration was found in A. bicolor bicolor and A. bengalensis bengalensis in Indonesia and Malaysia (Arai et al. 2016; Arai and Abdul Kadir 2017a).

Tropical freshwater eels were determined to take approximately 2–6 months to migrate from their spawning grounds to continental habitats (Arai et al. 1999a, b, c, 2000a, b, 2001, 2002a, b, 2003a; Marui et al. 2001; Shiao et al. 2001, 2002; Robinet et al. 2003, 2008; Réveillac et al. 2009; Leander et al. 2012; Hewavitharane et al. 2019). The oceanic migration period in temperate eels would correspond to the distance between the spawning grounds and continental habitats integrating the ocean current systems. Some small leptocephali of A. celebesensis (12–20 mm) and A. borneensis (8–13 mm) were collected in the Celebes Sea (Jespersen 1942; Aoyama et al. 2003), which is adjacent to their continental growth habitats in these species. These results suggest that these two species spawn in the vicinity of their continental growth habitats and that spawning migrations would be quite short distances. Interestingly, at least two spawning grounds have been found in Tomini Bay and the Celebes Sea, around Sulawesi Island in A. celebesensis (Aoyama et al. 2003; Arai 2014c). Such short-distance oceanic migration between spawning grounds and continental growth habitats is distinctively different from that of temperate eels (Arai 2014c). The distances between possible spawning grounds and the continental growth habitats in the tropical eels A. celebesensis and A. borneensis were 80–300 km and 480–650 km, respectively (Aoyama et al. 2003; Arai 2014c). However, those in temperate eels A. anguilla (McCleave et al. 1987), A. rostrata (McCleave et al. 1987) and A. japonica (Tsukamoto 1992) were 4000–8000 km, 900–5500 km and 2000–3500 km, respectively. This discrepancy between tropical and temperate eels would relate to the distance between spawning grounds and their continental growth habitats, suggesting that tropical eels establish considerably shorter and local migrations to spawn in areas near their continental growth habitats compared to the long migrations established by temperate eels (Fig. 2). The spawning ground of a tropical eel A. marmorata overlaps with that of a temperate eel A. japonica in the western North Pacific Ocean (Tsukamoto et al. 2011), and both eels are transported and colonized in the same continental growth habitats in East Asian countries, although A. japonica is further transported and colonized to more northern regions. The oceanic migration route and continental distribution in A. marmorata suggest that the migration scale is middle scale similar to that of A. japonica and greater than that of other tropical eels such as A. celebesensis and A. borneensis (Fig. 2). The occurrence of leptocephali of various sizes and stages, including preleptocephalus to metamorphosing stages and oceanic glass eel stages, in waters off Sumatra (Jespersen 1942) and Tomini Bay (Aoyama et al. 2003) supports the possible oceanic migration distance. The occurrence pattern of larvae is completely different from that of temperate eels, in which leptocephali grow gradually along the oceanic migration route. This suggests that larval transportation and migration mechanisms of tropical eels are complicated compared to those of temperate eels. Freshwater eels are believed to originate in the tropics because the spawning grounds are all located in tropical regions (Schmidt 1925; Arai 2020). Leptocephali have a long duration (2–12 months) and are extremely adjusted to an oceanic planktonic life. Passive transport by oceanic currents and several months of oceanic migration periods would lead to global dispersion from tropical regions. Adventitious larval transportation by means of oceanic currents in tropical regions and ranges of larval durations would lead to further drift to temperate regions. Subsequently, new oceanic migration routes and continental habitats are established in temperate eels, leaving their spawning grounds in tropical waters; hence, temperate eels would need to migrate thousands of kilometres between tropical spawning grounds and temperate continental habitats (Arai et al. 2001, 2014c, 2020).

Continental migration

Freshwater eels are widely thought to be a catadromous species migrating between offshore spawning grounds and continental growth habitats (McDowall 1988). However, diverse migratory histories in continental growth habitats are found in all species in temperate eels and several species in tropical eels use otolith Sr:Ca ratio signatures (Fig. 2). A strong positive correlation between otolith Sr:Ca ratios and Sr contents and the ambient salinity level was found in freshwater eels (Tzeng 1996; Lin et al. 2007; Arai and Chino 2017), and other biotic (physiological) and abiotic (water temperatures, food) factors were found to be less affected on their otoliths (Tzeng 1996; Lin et al. 2007). Therefore, migratory history in freshwater eels could be reconstructed using otolith Sr:Ca ratios whether each fish truly recruits and lives in fresh water, brackish water or marine waters during their continental growth period or whether eels move between different salinity environments. Otolith microchemistry elucidated that a number of eels lived their whole lives in brackish (estuarine eels) and marine waters (sea eels) without freshwater lives (river eels) after recruitment to continental growth habitats (e.g., Tsukamoto and Arai 2001; Arai and Chino 2012, 2018; Arai et al. 2020). Furthermore, another unique migration pattern was found in which fishes frequently changed their growth habitats among marine, brackish and fresh water environments (habitat shifters) (e.g., Tsukamoto and Arai 2001; Arai and Chino 2012, 2018; Arai et al. 2020). These studies suggested that all freshwater eels do not necessarily live in freshwater environments; hence, freshwater eels display an opportunistic catadromy in different salinity regimes (e.g., Tsukamoto and Arai 2001; Arai and Chino 2012, 2018; Arai et al. 2020). The results suggest that the definition of freshwater eels as a catadromous fish definitely needs revision because movement into freshwater is not completed through an obligatory passage as is the case for ecophenotypes.

Differences in food resources between freshwater and marine habitats are generally believed to determine the migration in diadromous fish (Gross 1987). According to theory, catadromous fishes such as freshwater eels that colonize tropical waters (lower latitudes) are likely to settle in freshwater habitats because the productivity of continental growth habitats is higher than that of marine waters. Migration patterns and environmental habitat uses would be different along a latitudinal cline, i.e., more river eels would be found in tropical regions, while more sea eels would be found in temperate regions because the primary production of marine habitats in lower latitudes is lower than that in higher latitudes. However, otolith microchemistry studies found a number of sea eels and estuarine eels in tropical continental habitats and tropical eels such as A. marmorata (Shiao et al. 2003; Chino and Arai 2010a; Arai et al. 2013, 2020; Arai and Chino 2018), A. mossambica (Lin et al. 2012), A. bicolor bicolor (Chino and Arai 2010b, c; Arai and Chino 2019; Arai et al. 2020) and A. bicolor pacifica (Briones et al. 2007; Arai et al. 2013; Arai and Chino 2019) showing opportunistic migration ecology. Habitat uses and preferences in tropical eels during continental migration are the same as those found in temperate eels. These results suggest that less latitudinal clines of migration plasticity would be supported in freshwater eels, although latitudinal clines may be found to suppress intraspecies migration (Kotake et al. 2003). In freshwater eels, fishes would live and settle in various salinity environments with less latitudinal clines as the general migration behaviour during the continental growth stages. Migratory ecology and habitat usage would be more influenced by means of ambient environmental factors and habitat environments and/or intra- and/or interspecies competition in each continental growth habitat.

All Anguilliforms except freshwater eels are marine species; hence, freshwater eels would be derived from a marine ancestor (Inoue et al. 2010), suggesting that the offshore spawning ecology would be a conservative trait. Thus, many tropical and temperate eels stay in marine and coastal growth habitats after colonization. However, the mechanisms in the occurrence of sea and estuarine eels are still unknown, whether are because of an ecological plasticity or a remnant genetic trait. The reproductive contribution among migratory types was examined in A. japonica in central eastern of Japan using migrating silver eels in the Pacific Ocean (Chino and Arai 2009). Estuarine and sea eels constituted approximately 81%, while river eels constituted 19% (Chino and Arai 2009). The high proportion of estuarine and sea eels over 80% with a lower proportion of river eels during the spawning migration season suggests that the estuarine eels and sea eels living in coastal habitats and peripheral areas contribute much more reproduction than that of river eels (Chino and Arai 2009). Populations of estuarine and sea eels may be higher than those of river eels. If environmental habitat use is a heritable feature, reproductive isolation and barriers might occur among populations as different migration strategies, leading to genetic differentiation among populations. However, no apparent genetic variation was found in different migration types in A. japonica (Han et al. 2010). Therefore, differences in migration ecology and habitat use might appropriately be behavioural plasticity as result of biotic and abiotic factors such as habitat environments and/or intra- and/or interspecific competition in the habitat. Sea eels, estuarine eels and habitat shifters may be more adaptable and suitable in freshwater eels as they originated from a marine ancestor.

The occurrence of marine resident eels could introduce ecological competition with other Anguilliforms (Moriarty 1978). Anguilla anguilla may experience high competition with a conger eel Conger conger at lower latitudes and hence a lower occurrence of A. anguilla from the middle to southern European and Mediterranean coasts where conger eels are abundant (Moriarty 1978). On the other hand, there are no conger eels in the North and Baltic seas while European eels are plentiful. Thus, such interspecific competition might stimulate migration plasticity (Tsukamoto and Arai 2001).

In temperate continental habitats, there are few plural freshwater eels are distributed except in New Zealand, where A. dieffenbachii and A. australis schmidtii live sympatrically (Glova 1988; Glova et al. 1998). However, plural species live sympatrically in tropical continental habitats; hence, habitat uses might differ depending on the number of freshwater eel species distributed in the habitat (Arai and Chino 2012, 2018; Arai et al. 2013, 2020). In Taiwan and the Philippines, a tropical freshwater eel A. marmorata prefers to live in freshwater habitats (Shiao et al. 2003; Briones et al. 2007). However, sea eels, estuarine eels and habitat shifters were found in A. marmorata in Indonesia, Japan and Vietnam (Chino and Arai 2010a; Arai and Chino 2018). In Taiwan and the Philippines, multiple species are distributed sympatrically: A. japonica and A. marmorata and A. bicolor pacifica and A. marmorata; however, only A. marmorata is distributed on islands of Japan (Chino and Arai 2010a; Arai and Chino 2018). In Japan, A. marmorata can live and move in different habitats from river downstream to upstream after recruitment with no interspecific competition. Indeed, estuarine eels were the most abundant migration pattern (Chino and Arai 2010a; Arai and Chino 2018). Different habitat uses were found in two New Zealand temperate eels A. australis schmidtii and A. dieffenbachii. Anguilla dieffenbachii generally lives upstream, while A. australis schmidtii lives in downstream along the river (McDowall 1990). In Malaysia, a tropical eel A. bicolor bicolor was widely distributed in rivers, but many of the eels were mainly found from downstream to midstream in each river; however, another tropical eel, A. bengalensis bengalensis, occurred from midstream to upstream in each river (Arai and Abdul Kadir 2017b; Arai et al. 2020). Similar habitat segregation was reported in Vietnam (Arai et al. 2013). In A. bicolor pacifica, most fishes (90%) were estuarine eels, and others were marine residents without river eels (Arai et al. 2013). In A. marmorata, the majority were estuarine eels (90%) and others were river eels without sea eels (Arai et al. 2013). Anguilla bicolor pacifica seems to prefer higher saline habitats than A. marmorata in Vietnam. These results suggest that freshwater eels could live and migrate to various salinity habitats during their continental growth phases. However, habitat use and habitat preference in freshwater eels would be different in each habitat depending on the occurrence of sympatric eel species and degrees of habitat use and habitat preference might also differ in each habitat by means of intra- and/or interspecies competition and habitat environments.

In New Zealand, habitat segregation is found by means of physical features in growth habitats (Glova et al. 1998). Current velocities, inclination pitches, bottom materials and carrying capacities are potential factors in different habitat use and migration behaviours (Glova et al. 1998). Anguilla dieffenbachii lives at faster water velocities and larger substrates of riffles; however, the sympatric A. australis schmidtii prefers to settle in slower marginal habitats (Glova et al. 1998). Habitat uses and preferences differed between A. bicolor bicolor and A. bengalensis bengalensis in Malaysia by means of interspecies interactions and intraspecific plasticity to each habitat environment (Arai and Abdul Kadir 2017b; Arai et al. 2020). Anguilla bicolor bicolor lives from downstream to upstream areas in each river with tidal influence; however, A. bengalensis bengalensis occurred from downstream to midstream in freshwater habitats with higher elevation, lower water temperature and no tidal influence. Various environmental factors such as salinity, temperature, water velocity, elevation, river size and carrying capacity, primarily influence habitat preference, habitat use and movement in freshwater eels. The diverse and opportunistic migration ecology and behaviour in freshwater eels found during the continental growth phase display phenotypic plasticity in each continental growth habitat by means of intra- and/or interspecies competition and habitat environments.

Oceanic migration to spawning grounds

Oceanic migration to spawning grounds is the most enigmatic migration phase because it is difficult to track the migratory behaviour and route in the open ocean. After freshwater eels adequately matured for their spawning migration to the ocean, they metamorphosed from the yellow stage to the silver stage. However, the timing and mechanism triggering to the metamorphosis in this stage are still unknown. Numerous morphological and physiological changes are reported upon starting oceanic migration to spawning grounds, i.e., silvering skin in colours (Pankhurst and Lythgoe 1982; Han et al. 2003; Okamura et al. 2007), eyes and pectoral fins enlarge (Pankhurst 1982; Han et al. 2003; Okamura et al. 2007), development of the swim bladder (Kleckner 1980; Yamada et al. 2001), degenerated alimentary tract (Pankhurst and Sorensen 1984; Han et al. 2003), retinal sensitivity change (Andjus et al. 1998; Zhang et al. 2000), musculature development (Egginton 1986) and increases of fat content (Larsson et al. 1990) and chloride cells of the gills (Fontaine et al. 1995). These developments accompanying the onset of metamorphosis into silver eels are believed to enhance the swimming capability and reduce the possible predation in the ocean, leading to a higher chance of reaching spawning grounds.

Freshwater eels have been successfully tracked in the open ocean for several weeks to months by means of the development of PSATs, allowing reconstruction of enigmatic oceanic migration to spawning grounds. To date, the spawning grounds of most freshwater eels are either known with a large degree of uncertainty or not known at all. The massive collection of smaller leptocephali of A. anguilla and A. rostrata could determine their spawning ground in the southwestern Sargasso Sea (Schmidt 1922; Miller et al. 2019), while their eggs and adults have never been caught in spawning grounds over centuries. In A. japonica, the spawning ground was discovered in the North Equatorial Current west of the Mariana Islands in 1991 with collection of massive small leptocephali (Tsukamoto 1992). Furthermore, eggs and fully matured adults of A. japonica and a tropical eel of A. marmorata were collected in the spawning ground (Chow et al. 2009; Tsukamoto et al. 2011). Migration routes and unique vertical migration behaviour during oceanic migration to spawning grounds in several freshwater eel species have recently been revealed by means of PSAT research.

All tracked freshwater eels commonly showed diel vertical migrations (DVMs) during their oceanic migration to spawning grounds swimming deeper in the daytime than in the nighttime (e.g., Jellyman and Tsukamoto 2002, 2005, 2010; Aarestrup et al. 2009; Béguer-Pon et al. 2012, 2015; Schabetsberger et al. 2013, 2015; Westerberg et al. 2014; Wysujack et al. 2015; Amilhat et al. 2016; Righton et al. 2016; Chang et al. 2020). Detailed DVMs have been recorded using PSAT tracking in temperate eels A. dieffenbachii (Jellyman and Tsukamoto 2002, 2005, 2010), A. anguilla (Aarestrup et al. 2009; Westerberg et al. 2014; Wysujack et al. 2015; Amilhat et al. 2016; Righton et al. 2016), A. rostrata (Béguer-Pon et al. 2012, 2015) and A. japonica (Manabe et al. 2011; Chow et al. 2015; Higuchi et al. 2018) and tropical eels A. megastroma (Schabetsberger et al. 2013, 2015, 2019; Chang et al. 2020) and A. marmorata (Schabetsberger et al. 2019; Chang et al. 2020). Interestingly, all freshwater eel species swam similar depth ranges, with depths between 500 and 800 m in the daytime and between 100 and 400 m in the nighttime, with a maximum depth ranging from 1000 to 1200 m. The DVMs are believed to exhibit predation avoidance behaviour (Béguer-Pon et al. 2017). Possible predators during the oceanic migration to spawning grounds have been found in tunas (Blank et al. 2007; Béguer-Pon et al. 2012), sharks (Béguer-Pon et al. 2012) and whales (Schorr et al. 2014; Amilhat et al. 2016). Another possible DVM has been suggested to be thermal regulation to maintain adequately high metabolism and swimming activity (Boëtius and Boëtius 1967; Tesch 1977; Jellyman and Tsukamoto 2010). Freshwater eels ascend to shallower warm water during the nighttime and descend to deeper cold waters during the daytime to maintain their temperature below 11 °C to delay gonad development (Boëtius and Boëtius 1967; Tesch 1977; Jellyman and Tsukamoto 2010). Furthermore, DVMs are found to be related to the lunar cycle in A. japonica (Chow et al. 2015) and A. megastoma (Schabetsberger et al. 2013, 2015), in which these freshwater eels swam deeper layers during the full moon than during the new moon.

Migration routes and possible spawning grounds are reconstructed by means of PSAT research. The possible spawning grounds were indicated by means of the tracking of A. dieffenbachii (Jellyman and Tsukamoto 2002), and A. marmorata and A. megastoma from the Archipelago of Vanuatu were tracked towards their presumed spawning grounds (Chang et al. 2020). One A. rostrata specimen was completely tracked for 2400 km for 52.7 days from the Scotian Shelf (Canada) to the northern limit of the spawning ground of the Sargasso Sea in 2014 (Béguer-Pon et al. 2015). It provided the first direct evidence of oceanic migration to the spawning ground (Béguer-Pon et al. 2015). Further research could track for approximately 7000 km from the Celtic Sea to the Atlantic Ocean (Righton et al. 2016) and for over 2000 km from the Mediterranean Sea to the Atlantic Ocean in A. anguilla, navigating the Strait of Gibraltar (Amilhat et al. 2016). A study showed that A. anguilla from the Mediterranean contributes to the oceanic migration for spawning (Amilhat et al. 2016). The horizontal migration speeds are estimated in tracked freshwater eels: 15 − 31 km d−1 in A. dieffenbachii (Jellyman and Tsukamoto 2002), 2 − 51 km d−1 in A. anguilla (Aarestrup et al. 2009; Westerberg et al. 2014; Wysujack et al. 2015; Amilhat et al. 2016; Righton et al. 2016), 2.5–14 km d−1 (Manabe et al. 2011) in A. japonica, 10 − 54 km d−1 in A. rostrata (Béguer-Pon et al. 2015, 2017) and 21–23 km d−1 in A. marmorata and A. megastoma (Chang et al. 2020). These findings support the estimation of oceanic migration periods to spawning grounds. Estimated trajectories suggest that freshwater eels stay on the continental shelf for a while before migrating across the shelf to offshore spawning grounds (Jellyman and Tsukamoto 2002; Manabe et al. 2011; Westerberg et al. 2014; Béguer-Pon et al. 2015; Amilhatet al. 2016; Righton et al. 2016). Currently, only one specimen of A. rostrata has completely tracked the oceanic migration behaviour and route from the continental habitat to the spawning ground. Compared to other migration phases, apparently more research is needed to understand the migration route and behaviour and discover the spawning ground.

Conclusion

Freshwater eels are extremely important animals that have historical, environmental and cultural significance across a broad spectrum. They are considered a key species in the freshwater and coastal water ecosystems, and if they were to be extirpated, they would be a significant loss and one that would have a knock-on effect on a range of species. Intensive and comprehensive biological and ecological research has been conducted employing the innovative and latest research approaches in six species/subspecies of temperate eels over a century. Compared to those of temperate species, however, there has been considerably less scientific research accomplished on the 13 species/subspecies of tropical eels, even though two-thirds of species are distributed in tropical Indo-Pacific regions. In tropical eels, even fundamental information is at the rudimental level, e.g., the species identification in freshwater eels using only morphological and morphometric analyses is difficult due to the overlapping and similarities of those features (Arai 2014d; Arai et al. 2015; Arai and Wong 2016; Abdul Kadir et al. 2017a). Compared to temperate eels, little fundamental knowledge on the species composition and diversity, biogeography, and life history is available in many tropical eels across the Indo-Pacific region. This issue hinders further biological and ecological research, such as migration ecology in tropical freshwater eels.

In migration ecology (Fig. 2), information on oceanic migration to continental habitats and continental migration gradually accumulates in tropical and temperate eels, while information on oceanic migration to spawning grounds is clearly lacking in tropical and temperate eels. To understand and reconstruct the migration ecology throughout their lives, further research is needed on oceanic migration to spawning grounds using PSATs and innovating devices. Currently, PSATs are the most effective in understanding the behaviour of freshwater eels during their oceanic migration for spawning. The tags can record data for longer periods than acoustic tags used previously. However, to date, PSAT research has been conducted for 7 out of the 19 species only. Tropical eels such as A. celebesensis display local migration less than 100 km from the continental habitat to the spawning ground (Arai 2014c), and fully matured specimens ready for spawning can collect in inland lakes (Arai 2014c). Anguilla celebesensis and A. marmorata were radiotracked by means of the accelerometer tags to understand the diving behaviour of migrating silver eels in an Indonesian lake, Lake Poso (Watanabe et al. 2016), where the downstream of the river mouth is close to a spawning ground of A. celebesensis in Tomini Bay (Arai et al. 2014c). Interestingly, these eels showed the repeated diving behaviour in the lake, such as oceanic migrating silver eels, although these eels did not show clear DVMs in the lake (Watanabe et al. 2016). Furthermore, the silver stage of a tropical eel A. bicolor bicolor is found to occur throughout the year (Arai et al. 2016 and Arai and Abdul Kadir 2017a). Tropical eels, A. bicolor pacifica (Arai et al. 2001), A. reinhardtii (Shiao et al. 2002; Shen and Tzeng 2007) and A. bengalensis bengalensis (Arai and Abdul Kadir 2017a), are also predicted to undergo year-round spawning migration. Therefore, tropical eels would be useful and convenient for tracking research on oceanic migration to spawning grounds.

Populations of freshwater eels are extremely vulnerable to exploitation and global climate change. Drastic declines in glass eel recruitment in temperate eels are critical issues, although such assessments have been less conducted in tropical eels (Arai 2021). The main threats of depletion in eel stocks and coastal recruitment are complex and enigmatic. The exploitation status and stock size in tropical eels are poorly understood. The fecundity of the tropical eels A. bicolor bicolor, A. bengalensis bengalensis and A. marmorata was similar to that of temperate eels, A. anguilla, A. rostrata, A. japonica, A. dieffenbachii and A. australis (Abdul Kadir et al. 2017b). The similarity in fecundity between temperate and tropical anguillid eels and the year-round reproductive characteristics suggests that the population size of the tropical eels is higher than that of the temperate eels (Abdul Kadir et al. 2017a, b). However, the population size of the tropical eels would be different by species due to differences in geographical distribution ranges and population sizes (Arai 2021). Intensive and comprehensive migration ecology research in tropical eels together with temperate eels throughout their lives is further needed to reveal their migration behaviours and routes with consequences for eel stock management, protection and conservation.