The period of sea entry and the first few months of marine life is a critical period in the life of the Atlantic salmon (Salmo salar) and occasionally characterised by very high mortality (Hvidsten and Møkkelgjerd 1987; Thorpe et al. 1994; Rikardsen et al. 2004; Davidsen et al. 2009). As a preparation for residency in the marine environment, juvenile Atlantic salmon undergo a parr-smolt transformation (termed smolting) before entering the sea (Hoar 1988; Boeuf 1993; McCormick et al. 1998). Among the biochemical, physiological and behavioural changes that comprise the smolting process, development of hypoosmoregulatory ability (seawater tolerance) is considered to be most critical for performance in seawater (McCormick and Saunders 1987). The development of seawater tolerance is associated with an increase in the number and size of the seawater chloride cell subtypes and a several-fold increase in gill Na+, K+-ATPase activity (Hoar 1988; Evans et al. 2005). Fish that have not completed their smolting process are subjected to osmotic stress when they enter seawater which can reduce swimming performance and thus make them more vulnerable to predators (Handeland et al. 1996), while individuals that are completely smoltified have been shown to have higher survival rates after sea entry (Moser et al. 1991).

The timing of the onset and completion of the smolting process is governed by photoperiod and water temperature, respectively (Solbakken et al. 1994; Handeland et al. 2003; Zydlewski et al. 2005), while water temperature and water flow are key factors for the initiation of seaward migration (McCormick and Saunders 1987; McCormick et al. 1998). A successful sea entry presupposes that the completion of the smolt development is synchronised with the timing of migration. Stewart and co-workers (2006) found that smolts from upper tributaries generally begin migration earlier (one or several weeks) than those from lower tributaries, contributing to a more synchronized timing of sea entry of smolts from the entire watershed. It is believed that smolts use environmental cues in the rivers that are predictably correlated with favourable ocean conditions (e.g., temperature and food availability) to control movements in order to arrive at sea when growth and feeding condition is optimal (Hvidsten et al. 2009). However, there is little knowledge about if, and how, the timing of the seawater migration is synchronized with the development of seawater tolerance and to what extent there may be an asynchrony between smolt development and seawater entry within years in salmon populations. The main goal of the present study was therefore to examine the physiological smolt status (using gill Na+, K+-ATPase activity as an indicator of SW readiness) of two groups of smolts migrating at different times within the same year, and their subsequent migration timing to seawater.

In River Alta, northern Norway (70°N 23°E), the main smolt migration occurs during a 2–3 week period in late June to mid-July (Hvidsten et al. 1998). During the smolt migration in this river in 2007, a total of 120 wild smolts were trapped during two periods (2 × 60 smolts) 11 km above the river mouth and tagged with acoustic tags (Thelma AS, Norway, model LP-7.3, diameter of 7.3 mm, length of 18 mm, mass in water/air of 1.2/1.9 g) using methods described in Davidsen et al. (2008). The smolts were released at the capture site 10 min after recovery. Due to an extraordinary high flow in June 2007, it was not possible to operate the trap in the river before 24 June. Based on earlier smolt migration observations in this river (Hvidsten et al. 1998) and the actual daily smolt catches in the present study (Fig. 1), it was assumed that the fish in the first group (26–28 June, n = 60, mean L F 146 mm, S.E. = 0.8; mean mass 28 g, S.E. = 0.5) were tagged and released a few days after the assumed main peak of the smolt migration, while the fish in the second group (2–4 July, n = 60, mean L F 147 mm, S.E. = 1.1, mean mass 30 g, S.E. = 0.7) probably represent the smaller second peak of migrating smolts. River water temperature was 10–11°C during the first period and 13°C during the second period. In order to monitor the time of sea entry, two automatic listening stations (ALS) (Vemco INC, Canada, model VR2) were placed in the river mouth (detecting range 50–100 m). The last record of each smolt on the ALS’s in the river mouth was used as the time of sea entry. In order to examine gill Na+, K+-ATPase activity of the smolts migrating in the two periods, a total of 20 additional smolts from the trap during each tagging period (sampled 26 June and 4 July) were sacrificed by a sharp blow to the head. Samples for measurements of gill Na+, K+-ATPase activity were taken, frozen in liquid nitrogen and later analyzed by the method described in McCormick (1993). Two-way ANCOVA was used to test for differences in gill Na+, K+-ATPase activity between the earlier and later smolt group and Mann-Whitney U-test to test for differences in migratory progress.

Fig. 1
figure 1

Total daily catch of smolts in the trap from 24 June to the smolt run had ended (18 July). Sampling was done consistently during the whole study period. The tagging periods are indicated by black lines at the x-axis and the dates at which gills were sampled for Na+, K+-ATPase activity with an arrow

Full size image

The results showed that smolts captured on 26 June had a four-fold lower gill Na+, K+-ATPase activity (Two-way ANCOVA, n = 40, P < 0.001) than the smolts captured a week later (Fig. 2). Further, the first group spent on average twice the amount of time in the lower part of the river before entering the sea than the second group (96 vs. 48 h, respectively, Mann-Whitney U-test, n = 64, P < 0.001, Fig. 3). In total, a little more than half of the tagged smolts from each group (31 and 33, respectively) were recorded by the ALS’s in the river mouth. The gill Na+, K+-ATPase activity of the smolts captured in the first period, close to the assumed peak of the smolt migration (3.3 ± 2.4 μmol ADP mg protein−1 h−1), was much lower than the activity normally found in fully smoltified Atlantic salmon from the River Alta (≥10 μmol ADP mg protein−1 h−1; Lysfjord and Staurnes 1998; Strand et al. 2007). This indicates that these fish were not fully smolted when passing the smolt trap 11 km upstream from the river mouth. In contrast, the mean activity in the second group (13.8 ± 4.4 μmol ADP mg protein−1 h−1) indicated that these fish had completed smolting. Therefore, there was a negative correlation between time spent by the smolts in the lower part of the river before sea entry and their gill Na+, K+-ATPase activity further up in the river (capture site).

Fig 2
figure 2

Gill Na+, K+-ATPase activity (mean ± SE) of smolt sampled in period 1 (26 June; n = 20) and period 2 (4 July; n = 20). Asterisks denote significant difference between groups (***; p < 0,001)

Full size image
Fig. 3
figure 3

Time from release to sea entry (median ± Quartiles) of smolts tagged in period 1 (26–28 June; n = 31) and period 2 (2–4 July; n = 33). Asterisks denote significant difference between groups (***; p < 0,001)

Full size image

Several studies have shown that post-smolts are exposed to high predation immediately after sea entry (Reitan et al. 1987; Hvidsten and Lund 1988; Dieperink et al. 2002) and that the risk of predation increases if fish suffers from osmotic stress (Järvi 1989; Handeland et al. 1996; Kennedy et al. 2007). It may therefore be suggested that the extended residency of the first smolt group in the lower part of the river was related to their low gill Na+, K+-ATPase activity and a need to improve seawater tolerance before entering seawater. The migration of Atlantic salmon smolts through estuaries is characterized by active swimming and is a continuous movement, usually with no apparent period of acclimatization to adjust to salt-water osmotic and ionic conditions (Lacroix and Knox 2005). Also smolts from the River Alta have a very short residence time in the estuary (Davidsen et al. 2009). The mean time from leaving the river mouth to pass an array of automatic listening stations four km outward in the Alta Fjord was 4 h in the present study. In this area, salinity varied from 10 to 20 at the surface to 30–35 at 4 m depth during the time of the migration (Davidsen et al. 2009), so the smolts entered full strength seawater already few hours after sea entry. It has been shown, however, that a thermal sum of approximately 350 d°C (i.e. the sum of daily mean temperatures) is needed for increasing the gill Na+, K+-ATPase activity from 6 to 12 μmol ADP mg protein−1 h−1 at the temperature (10–13°C) prevailing in the River Alta in late June/early July (Stefansson et al. 1998). The smolts in the first group experienced 40–60 d°C from when they were caught in the trap until entering the sea, which in the Alta Fjord is characterized by high salinity (>30) even during spring (Davidsen et al. 2009). Although the absolute values of gill Na+, K+-ATPase activity may vary between studies, partly due to analytical methods, 40–60 d°C seems to be too little for the first smolt group to complete their smolting process before they actually entered full strength sea water. Hence, if the low gill Na+, K+-ATPase activity of the 20 smolts sampled in the first smolt group is representative for the migrants in this period, a large number of smolts leaving the River Alta in 2007 possibly possessed a suboptimal smolt quality at the time of sea entry. Despite the possible suboptimal smolt quality of the first tagging group, a study on the survival of the corresponding two groups of tagged smolts during their first days as post-smolt in the Alta Fjord did not show any difference in survival (Davidsen et al. 2009). This finding is a paradox since earlier studies have revealed a strong, positive correlation between gill Na+, K+-ATPase activity and hypoosmoregulatory ability in Atlantic salmon (Strand et al. 2007) and further, that hypoosmoregulatory ability is predictive for survival in the sea (Virtanen et al. 1991; Staurnes et al. 1993). On the other hand we did not perform seawater challenge tests with these smolts in order to reveal their actual hypoosmoregulatory ability and the smolts were not tested for gill Na+/K+-ATPase activity at seawater entry. Therefore, categorical statements about smolt quality cannot be made at the current stage.

It is not known whether the difference in gill Na+, K+-ATPase activity found among migrating smolts in the present study represent a general trait among smolts in River Alta, or represents an occasional event. The result indicates that the quality of wild smolts may vary within smolt runs, but further studies are needed to reveal if this is a frequent event and if it is a general trend that earlier migrants are less prepared for seawater than later migrants. Several other factors may also influence the actual survival of early and late migrating smolts. For example, the risk of being severely infected with the parasitic salmon louse Lepeophtheirus salmonis (Krøyer) is shown to be higher for the latest migrating post-smolts in northern fjords (Bjørn et al. 2007), while the food availability often is best for early migrating smolts (Rikardsen et al. 2004; Hvidsten et al. 2009). Such factors may therefore select for early sea entry. The pronounced difference in gill Na+, K+-ATPase activity found between smolts about to enter sea in the River Alta emphasize the need for more knowledge about how physiological and ecological factors influence migration behaviour and survival of this species.

In conclusion, this study showed that smolts with low gill Na+, K+-ATPase activity aggregated in the lower part of the river and delayed their sea entry compared to later migrating smolts that appeared to be completely smoltified. Whether this behaviour is related to a need to improve seawater tolerance before sea entry, or an adaptation to synchronise sea entry in order to reduce the risk of predation, is an interesting question open to future research.