212 
Fishery Bulletin 119(4) 
and the conditions that allow for greater growth and larger 
smolt sizes. Smolt outmigration timing has been shown to 
be stock specific and to correlate with optimal local marine 
conditions (Hvidsten et al., 1998). Optimal conditions lead 
to fish experiencing favorable SSTs for growth, increased 
abundance of food resources, or reduced interactions with 
predators (Hvidsten et al., 1998; Mortensen et al., 2000). 
In our study, the end date of outmigration, which could 
only be tested for Chinook salmon in the Chilkat River, 
was positively related to marine survival, indicating that 
smolts of BYs that outmigrated later in the season had 
higher marine survival. Antonsson et al. (2010) similarly 
found that Atlantic salmon (Salmo salar) that migrated 
later had higher survival, but Mortensen et al. (2000) found 
that the earliest and latest emigrating smolt pink salmon 
(O. gorbuscha) had the lowest marine survival rates. Early 
outmigrants experienced colder water temperatures and 
lower prey densities, and that exposure likely resulted 
in lower survival (Mortensen et al., 2000). For Chinook 
salmon in the Chilkat River, migrations that ended later 
likely provided additional time for some smolts to attain 
a larger size before outmigration, given that mean smolt 
length was longer in years when migrations ended later. 
Alternatively, the later migration period may have distrib- 
uted smolts over a wider range of conditions, increasing 
the chance that some individuals encountered favorable 
marine conditions. 
Increased discharge in spring during outmigration can 
influence migration timing in some stocks and may also 
be linked to marine survival. For smolt Chinook salmon 
in the Nechako River, in British Columbia, high discharge 
and river temperatures resulted in shorter migration 
periods (Sykes et al., 2009). Lawson et al. (2004) found 
that high discharge in spring could facilitate outmigra- 
tion by increasing swimming speed and allowing fish to 
be transported downstream more passively, potentially 
conserving energy reserves and improving survival. In 
our study, higher spring discharge resulted in earlier 
outmigrations and lower marine survival for Chinook 
salmon from the Chilkat River. In contrast, discharge was 
not significantly related to the survival of fish from the 
Stikine River. River conditions may be more indicative 
of the environment experienced by smolts that remain 
in coastal waters during the marine entry phase, when 
a large portion of marine mortality occurs. For example, 
smolt Chinook salmon from the Fraser River fared worse 
during years when discharge into the Strait of Georgia 
was high because it caused increased turbulence, reduced 
stability, and low production in estuarine areas (Gargett, 
1997). High seasonal discharge may cause similar pat- 
terns in the inside waters of SEAK where Chinook salmon 
from the Chilkat River reside and, therefore, may limit 
survival. Alternatively, the lack of a relationship between 
discharge and marine survival of Chinook salmon from 
the Stikine River could be evidence of a rapid migration 
upon ocean entry to offshore waters that are less affected 
by freshwater discharge. 
Local SSTs were not related to marine survival of 
Chinook salmon from the Stikine and Chilkat Rivers. 
Mortensen et al. (2000) found that higher SSTs in Auke 
Bay were related to higher growth and survival of pink 
salmon from Auke Creek, but Briscoe et al. (2005) found no 
correlation between local SST and marine survival of jack 
and adult coho salmon (O. kisutch) in the same system. 
Increases in survival related to SST's stem from optimal 
growth conditions because temperature is the primary 
function that controls metabolic and growth rates for ecto- 
therms (Brett et al., 1969; Groot et al., 1995). Sea-surface 
temperatures can also affect ocean conditions, such as 
upwelling, food availability, and predator assemblages, all 
of which may aid or hamper growth and survival of Pacific 
salmon (Cole, 2000; Mortensen et al., 2000; Mueter et al., 
2002b). Mueter et al. (2005) found that warmer coastal 
SSTs were associated with increased survival for chum, 
sockeye, and pink salmon in Alaska. Similar results have 
been reported for Chinook salmon in the Unuk River, 
northeast of Ketchikan, Alaska, with higher marine sur- 
vival occurring with higher coastal SSTs (Graham, 2016). 
The lack of a relationship between nearshore SSTs and 
marine survival of Chinook salmon in the Stikine and 
Chilkat Rivers could be because the majority of individu- 
als grew elsewhere. As mentioned previously, it is possible 
that most of the smolts from the Stikine River migrate 
directly and more rapidly to more open-ocean areas rather 
than occupying more nearshore waters upon marine entry 
as assumed in our study. Marine survival of Chinook 
salmon from the Chilkat River, a stock that is believed 
to grow in the inside waters of northern SEAK, was cor- 
related to migration timing and river conditions but not 
to nearshore SSTs. Conducting trawl surveys during the 
summer in Icy Strait, a known migration corridor for 
salmon in northern SEAK, Orsi et al. (2013) found few 
ocean age-0 Chinook salmon in summer trawl hauls, indi- 
cating that these fish may be nonmigratory upon marine 
entry and remain in more local areas. Alternatively, aver- 
age SST from June and July could have been too broad a 
range to detect a process that occurred over days or weeks, 
rather than over months, for Chinook salmon from the 
2 rivers (Briscoe et al., 2005). 
Overwinter survival and smolt production 
Freshwater indices of survival and production were tested 
only for Chinook salmon from the Chilkat River because 
of constraints in data availability. We found that lower dis- 
charge in fall and higher discharge in early spring were 
associated with increased smolt production in the Chilkat 
River. Results from previous research indicate high vari- 
ability and seasonality in the relationships between the 
freshwater life stage of salmon and river discharge. Low 
discharge levels in the fall and winter have been reported 
to limit survival and smolt production, likely reducing 
migratory pathways in the fall and overwintering habitat 
availability in some systems (Lawson et al., 2004; Crozier 
and Zabel, 2006). High discharge in summer has been 
linked to low production of stream-type juvenile Chinook 
salmon in the Yukon River, in Alaska, possibly because 
high discharge caused velocity barriers that limited 
