The cost of anadromy: marine and freshwater mortality rates in anadromous Arctic char and brown trout in the Arctic region of Norway

It is hypothesized that in diadromous fish, migrations may occur because of differences in the availability of food in marine and freshwater habitats. The benefits of migration to sea may be increased growth opportunities and reproductive output, while the costs may be increased mortality and increased energy use. Here we examine mortality rates of anadromous Arctic char (Salvelinus alpinus) and brown trout (Salmo trutta) in fresh water and at sea over a 25-year period to test these hypotheses. Daily mortality rates were 5–15 times higher at sea than in fresh water, with highest rates for first-time migrants, inferring a clear trade-off between increased mass gain and mortality risk during the sea migration. Descending smolts were caught in a trap at the outlet of the river, individually tagged, and thereafter recorded each time they passed through the trap on their annual migration between the river and the sea. Brown trout females seemed to benefit to a higher degree from migrating to sea than did female Arctic char, probably because of the higher growth rate at sea, and hence higher reproductive output.

where L is fork length of the fish (in cm). To convert our length measurements to fork length, 157 we used the factor 0.980 (unpublished data). Jonsson and Jonsson (1999) (Table 1). 181 Occurrence of mature individuals in running water was examined by electrofishing.

190
During the 25-year study period, the between-year variation in mean residence time at 191 sea of first-time migrants of anadromous Arctic char was 23.2-44.6 days (with a mean (± SE) 192 duration of 34.0 ± 1.2 days) and of brown trout was 47.3-64.0 days (mean 55.2 ± 0.9 days).

193
Moreover, the mean residence times at the second to fifth sea migrations were 32. 8-35.5 days 194 for Arctic char and 54.5-62.3 days for brown trout (Table 2). All values for brown trout were 195 significantly higher than those for Arctic char (pairwise t-tests, P < 0.001).

196
The mean (± SE) survival rate of Arctic char smolts during the first sea migration was 197 33.6 ± 2.5%, and among those fish that ascended the trap the same summer as they returned to 198 the river, 44.1 ± 3.0% migrated to the sea the subsequent spring (Fig. 2a). Correspondingly,199 28.1 ± 2.0% of the brown trout smolts survived the first summer at sea, and among those 200 ascending again the same summer, 62.3 ± 2.9% descended the subsequent spring (Fig. 2b). In 201 both species, survival was higher for veteran migrants than for first-time migrants, and from D r a f t 10 202 the third sea migration onwards, usually 60-80% of the individuals survived each interval 203 between passing the trap on their way to and from the sea (Fig. 2).

204
Annual survival rates were higher for veteran migrants than for first-and second-time 205 migrants (Fig. 3). In the first three years after smoltification, annual survival rates were 206 roughly similar for the two species (Fig. 3), with mean (± SE) survival rates of 15.5 ± 1.8% 207 and 13.6 ± 1.4% in the first year and 42.6 ± 2.6% and 38.1 ± 2.4% in the second year after 208 smoltification for Arctic char and brown trout, respectively.

209
At a sea age of 3 years, annual survival rates exceeded 50% for both species (52.5 ± 210 3.1% for Arctic char and 54.4 ± 2.1% for brown trout). After that, annual survival rates of

218
Daily mortality rates were considerably higher at sea than in freshwater for both 219 species. In both environments, the daily mortality rates were highest for first-time migrants, 220 decreasing with sea age ( Table 2). The daily mortality of Arctic char was 13 and 10 times 221 higher at sea than in freshwater for first-and second-time migrants, respectively, and 5-7 222 times higher for older individuals. For brown trout, daily mortality rates both at sea and in D r a f t 11 226 brown trout decreased with the number of sea migrations and was 14-15 times higher at sea 227 than in freshwater for first-and second-time migrants, and 6-8 times for older fish.

228
About 6% of Arctic char and 10% of brown trout survived the period from their first 229 migration to sea as smolts until maturity three years later, and between 0.5% and 1% of both 230 species were still alive after seven sea migrations (Fig. 5). Up to 13 sea migrations were 231 observed for Arctic char and 11 for brown trout. Fish densities in the main tributary were low, with brown trout and Arctic char as the 252 predominating species (mean densities 8.7 ± 5.8 (± SD) and 6.1 ± 3.8 individuals per 100 m 2 253 for brown trout and Arctic char, respectively). Altogether, two mature individuals were caught 254 during these 11 years, one brown trout and one Arctic char. Both were males. In the outlet 255 river, brown trout (49.1 ± 12.9 per 100 m 2 ) and Atlantic salmon ( To our knowledge, this is the first study comparing daily mortality rates of 277 anadromous salmonids in marine and freshwater habitats. Gross (1987) tried to get some 278 insight into the cost and benefits of anadromy from existing data by comparing anadromous 279 and nonanadromous forms within seven different species of Salmo and Salvelinus. Egg 280 production by anadromous forms was on average threefold greater than that of their 281 nonanadromous conspecifics. Therefore, he concluded that anadromous fishes may have as 282 much as threefold higher mortality than nonanadromous fishes and still be favoured in 283 evolution. He also pointed out that differences in productivity between marine end freshwater 284 habitats are important, and in cases with high differences in productivity, higher mortality in 285 marine habitats may still favour anadromy. In our study, mortality rates at sea relative to 286 freshwater were considerably higher than that proposed by Gross (1987). High marine relative 287 to freshwater productivity may be the main reason why anadromous populations of Arctic 288 char and brown trout are present in northern Norway, in spite of the high marine mortality 289 rates in that area.

290
Mortality rates in the present study may have been overestimated because of mortality 291 due to handling, anaesthesia and tagging (Hansen 1988). Furthermore, a few individuals 292 stayed more than one year in freshwater before they returned to sea, and others changed from  D r a f t 27 Table 1. Catch of Arctic char and brown trout during gill net surveys in Lake Storvatn in the Hals watershed in 1989 and 1994. For each species, the catch is separated between anadromous individuals (individuals which have been to sea, and were tagged when passing the trap), mature resident individuals (untagged mature individuals), and untagged immature individuals (which may be either anadromous individuals still too young to migrate to sea, or resident individuals). D r a f t D r a f t Table 3. Basic data used to compute Estimated smolt fitness (ESF) of a female smolt of Arctic char and brown trout, which reaches maturity after three sea migrations. EFS was calculated as the summation of a smolt's probability p x of surviving to reproduce at any sea age x multiplied by its fecundity at that sea age. N is number of individuals returning to freshwater. Survival (%) is the probability of surviving from smolt until ascent into freshwater after x sea migrations, mass (g) is the mean mass of individuals at that time, and fecundity is the mean number of eggs produced by females of that mass. D r a f t