Introduction

Ontogenetic habitat shifts between freshwater and marine environments are a common feature of salmonid fishes like charrs Salvelinus spp. Charrs use a wide range of habitats, including small streams to large rivers, lakes, and marine habitats (Dunham et al. 2008). Although many studies have investigated the ecology of charrs in freshwater, little is known about charrs in the ocean (Dunham et al. 2008), especially in offshore waters. It has been suggested that most charrs do not move far offshore and that their movement is generally along the coast (Armstrong and Morrow 1980; Johnson 1980; Power 1980). However, in the case of the Dolly Varden S. malma, there is some empirical evidence indicating that this species is able to use offshore waters (Larkins 1964; Mishima 1975; Volkov et al. 1996). Additionally, a tag-recapture study reported long-distance movement of Dolly Varden between Alaskan and Russian rivers (DeCicco 1992). It is still unknown to what extent Dolly Varden use offshore waters.

The purpose of this study was to examine the distribution, seasonal abundance, and biological features of Dolly Varden in offshore waters. Japanese research vessels have monitored the stock condition of Pacific salmon with extensive fishing efforts (Ishida and Ogura 1992; Azumaya and Ishida 2000). Although the research program was designed to capture Pacific salmon, the by-catch data allowed us to examine the ecology of Dolly Varden in offshore waters.

Materials and methods

Japanese research vessels have monitored the stock condition of Pacific salmon since 1952 using research gillnets, commercial gillnets, longlines and trawls in the Pacific Ocean (Ishida and Ogura 1992). The by-catch data from this offshore salmon monitoring were compiled for the periods 1960–1962 and 1972–2007. The data included fishing location, gear type, sea surface temperature, number of fish caught, fork length, body weight, and sex. Fishing locations are basically at least 22 km away from shore. The total number of fishing operations and the total number of Dolly Varden caught were 17,008 and 17,684, respectively. Fishing stations were scattered throughout the North Pacific Ocean (Fig. 1a).

Fig. 1
figure 1

a Distribution of fishing stations by all fishing gear types (gillnets, longlines, trawls) compiled in this study and b the fishing stations where Dolly Varden were caught

Full size image

Data obtained using research gillnets that consisted of ten mesh sizes ranging between 48 mm and 157 mm (Takagi 1975) were used to evaluate density and fork length distribution. The total number of fishing operations using the research gillnets was 10 073. The catch per unit effort (CPUE), or the number of fish caught per 30 tans (ca. 1.5 km) of research gillnets, were calculated as density index. The CPUEs were stratified by month and 2° latitude × 5° longitudinal grid areas, following Azumaya and Ishida (2000).

Results and discussion

Anadromous Dolly Varden were distributed throughout a wide range in the Pacific Ocean, including the Japan Sea, Bering Sea, Okhotsk Sea, and the Gulf of Alaska (Fig. 1b). In particular, many Dolly Varden were caught in the Bering and Okhotsk Seas. Their distribution in the Bering Sea was biased toward northern areas, which is consistent with a previous report (Larkins 1964). The southern limit of their offshore distribution was in the Japan Sea (38° 46′ N, 135° 02′ E), which is south of the southern limit of fluvial Dolly Varden populations located in Hokkaido Island, Japan (ca. 42°30′N, Nakano et al. 1996). Anadromous Dolly Varden rarely occur in rivers of Hokkaido, Japan (Morita et al. 2005).

The total number and CPUE of offshore Dolly Varden showed a sharp rise from May to August, followed by a sharp decrease in September (Fig. 2). CPUE was comparatively high in the northwestern part of the Bering Sea and western Kamchatka, and the mode of the high CPUE group moved from offshore to coastal areas between July and August (Fig. 3). However, fishing stations were scarce between September and May, so it is unknown whether Dolly Varden occurred in offshore waters during these months. Additionally, Dolly Varden catches from the Japan Sea (ca. 40°N, 135°E) were reported between March and June (Machidori 1986).

Fig. 2
figure 2

Seasonal changes in total number of catch and catch per unit effort (CPUE) for Dolly Varden in offshore areas of the North Pacific obtained by research gillnets

Full size image
Fig. 3
figure 3

Seasonal changes in area specific (2° × 5° grid) CPUE for Dolly Varden in offshore areas of the North Pacific obtained by research gillnets. Crosses indicate grids where fishing by research gillnets were conducted but no Dolly Varden were caught

Full size image

Fechhelm et al. (1997) found that coastal CPUE of Dolly Varden in northern Alaska was typically bimodal, with the highest catches occurring early and late in the summer, and a notable mid-summer decrease. They suggested that this pattern is due to the early-summer outmigration of Dolly Varden from fresh waters (high early-summer catch), followed by dispersal along the coast or out to sea (mid-summer decrease in CPUE), and a late-summer return migration to fresh waters (high late-summer catch). Our results are consistent with this hypothesis, as we found that offshore CPUE of Dolly Varden was unimodal, with the highest catches occurring mid-summer.

The CPUE of Dolly Varden in July in the central Bering Sea (175°E–175°W, 55°N–59°N), where fishing stations were densely distributed between years, did not show significant decadal trends during the period 1972–2007 [linear regression of ln(CPUE + 1) on year, r = −0.153, p = 0.37]. However, large annual variations in Dolly Varden catches in offshore areas would limit an accurate assessment of the population status. Commercial Dolly Varden catches in a Khabarovsk region of Russia were reported to decrease during the same periods (Zolotukhin et al. 2002).

The fork lengths of offshore Dolly Varden varied considerably and ranged from 164 mm to 761 mm (Fig. 4). There were several modal groups, and the lowest modal group (<270 mm) likely represents fish smolted in that year, while the upper modal groups probably represent veteran migrants of post-smolt ages >1 year. This suggests that Dolly Varden used offshore waters after smolting. Upper modal groups appeared offshore first, and the lowest modal group appeared after August. This pattern is consistent with the stage-dependent timing of the seaward migration found in other iteroparous salmonids. That is, veteran migrants descend to the sea earlier than smolts (Jonsson et al. 1990). The sex ratio of the lowest modal group (<270 mm) was not biased, whereas that of the upper modal groups was biased toward females, particularly during early summer. However, most large individuals (>550 mm) were male.

Fig. 4
figure 4

Fork length distributions of Dolly Varden in offshore areas of the North Pacific during the summer obtained by research gillnets

Full size image

Sea surface temperatures at sites where Dolly Varden were captured by research gillnets ranged from 2°C to 14°C, and the frequency of occurrence was highest between 4°C and 12°C (Fig. 5). The average sea surface temperature ± s.d. at the capture sites was 7.7 ± 2.6°C. Thermal habitats of Dolly Varden were similar to the sympatric Pacific salmon Oncorhynchus spp. (Machidori and Ogura 1986; Azumaya et al. 2007).

Fig. 5
figure 5

Sea surface temperature distribution at fishing stations where Dolly Varden were caught (presence) or not caught (absence) in offshore areas of the North Pacific obtained by research gillnets and the occurrence probability [presence/(presence + absence)] in relation to sea surface temperature

Full size image

In conclusion, anadromous Dolly Varden used offshore waters during the summer. Offshore areas are therefore considered an important summer habitat for anadromous Dolly Varden. Common use of offshore waters may be unique among Salvelinus spp. The white-spotted charr S. leucomaenis, which occur sympatrically with Dolly Varden in rivers of the far east of Asia, were rarely caught during the Japanese offshore salmon monitoring (Machidori 1986). Many white-spotted charr were caught in coastal areas instead (Takami et al. 1996), indicating that white-spotted charr do not move far offshore and that their movement is generally along the coast. In addition, migratory behaviour of Dolly Varden would vary among stocks. It has been suggested that most southern form followed the coastline in their migration whereas northern form may undertake oceanic migrations of considerable magnitude (Armstrong and Morrow 1980). Analysis of by-catch data from the Japanese offshore salmon monitoring would provide valuable insight into the ecology of commercially unimportant fishes in offshore waters.