Berkman et al.: Effects on survival of Oncorhynchus tshawytscha in rivers of Southeast Alaska 205 
when 95% of smolts had been captured was used to define 
the end of the outmigration period. This index was used in 
this study rather than an estimate of the start or median 
migration date because smolt sampling began after the 
start of migration in some years, making accurate estima- 
tion of passage dates difficult. In addition, this index was 
assumed to be resistant to the biases associated with min- 
now trap catch data, indicating that daily variability in 
catch can be based on location, not on the actual number 
of fish outmigrating on a given day. This is because the 
index used in our study is based on a consistent decline in 
daily number of fish caught over the course of the migra- 
tion period rather than on the number of fish caught on 
any single day. (He and Lodge, 1990). 
Physical data: overwinter survival To determine the local 
factors influencing overwinter survival of Chinook salmon 
in the Chilkat River, environmental data for BYs during 
the fall and spring freshwater periods were examined. 
River discharge data were not available for the Chilkat 
River during 2000-2010 (BYs 1999-2009); therefore, the 
combined standardized mean monthly discharge from the 
Taku, Stikine, and Antler Rivers during 1998-2015 for 
fall (September—November) and spring (March—April) was 
used as a proxy for river discharge in SEAK. These rivers 
were chosen because they were the only glacial systems in 
the region with data for the entire study period. These riv- 
ers represent a range of drainage sizes, with mean annual 
discharges during 2014-2016 of 4 m*/s for the Antler 
River, 401 m°/s for the Taku River, and 1660 m°/s for the 
Stikine River. Discharge data were obtained from the U. S. 
Geological Survey (USGS) National Water Information 
System (available from website, accessed February 2017) 
as follows: Taku River near Juneau (USGS gauging station 
no. 15041200), Stikine River near Wrangell (USGS gaug- 
ing station no. 15024800), and Antler River near Auke Bay, 
Alaska (USGS gauging station no. 15055500). Although 
river discharge during winter may also affect overwinter 
survival (Lawson et al., 2004), the USGS rated discharge 
measurements from ice-covered rivers as poor, and rivers 
in SEAK are typically ice covered from December through 
early March (B. Elliott, personal commun.). 
Long-term monthly temperature data were not avail- 
able for the Chilkat River; however, air temperature data 
were available for a NOAA station close to the Chilkat 
River in Klukwan, Alaska (Haines 40 NW [station ID: 
USC00503504]; Global Summary of the Month data set 
based on the Global Historical Climatology Network daily 
data set available from website, accessed February 2017). 
Because stream temperatures are strongly correlated 
with air temperature above freezing (0°C), daily air tem- 
peratures from the Haines 40 NW station for September— 
November (fall) and March-April (spring) were used as 
proxies for stream water temperatures (Mohseni and 
Stefan, 1999). Although the relationship between air 
and water temperatures may be different for glacially 
fed streams, the air temperature indices used in these 
analyses likely captured the annual variation in water 
temperatures. 
Physical data: marine survival To analyze how regional 
environmental factors that occur during the early marine 
period correlate to marine survival of Chinook salmon 
in the Chilkat River, SSTs from inshore marine waters 
of SEAK (in degrees Celsius) were collected by NOAA 
during the Southeast Alaska Coastal Monitoring survey 
in 1997-2015 (Orsi and Fergusson’). Sea-surface tempera- 
ture was sampled monthly from May through August by 
using a conductivity, temperature, and depth sonde at a 
depth of 3 m at 3 different locations (Orsi and Fergusson’). 
These sampling locations represented inshore (Auke Bay 
Monitor) and strait (Upper Chatham Strait and Icy Strait) 
growing habitats (Orsi and Fergusson”) (Fig. 1). Because 
Chinook salmon in SEAK migrate to the marine environ- 
ment from April through late May, an index of marine 
entry SSTs was developed by averaging the SSTs for 
all stations during June and July (Orsi and Fergusson’; 
Elliott and Peterson”). This index was lagged by 2 years 
(BY+2 years) to reflect conditions during ocean entry. 
River conditions during smolt outmigration were also 
used in the analysis of marine survival. Average river tem- 
perature and river discharge at the time of outmigration 
(April-May) can affect run timing and early ocean condi- 
tions experienced by smolts. River temperature data were 
collected by the ADFG during smolt sampling of the CWT 
program for the period 2001-2011 (Ellott and Peterson’). 
Daily temperatures were averaged for each year to develop 
a variable for overall spring river water temperature. 
Using the previously described proxy for river discharge 
in SEAK, a spring migration period discharge index was 
developed by averaging monthly mean discharge (SEAK 
index on standardized scale) for April-May, the primary 
migration period for smolt Chinook salmon. 
Data analysis The influence of biological and environmen- 
tal factors on overwinter survival and smolt production 
of Chinook salmon in the Chilkat River was determined 
through principal component analysis and principal com- 
ponent regression. During preliminary examination of 
the data, the Pearson product-moment correlation coeffi- 
cient (7) was used to make pairwise comparisons between 
explanatory variables. Results indicate significant pos- 
itive correlation between parr length and fall discharge 
and early spring temperature and discharge (Table 1). 
Because of small sample sizes (n=11), 5 explanatory vari- 
ables and multicollinearity between variables were most 
appropriate for principal component analysis and princi- 
pal component regression. The environmental explanatory 
variables were lagged to represent conditions experienced 
by fish during their first fall prior to wintering in fresh 
water (BY+1) and during the following spring during out- 
migration (BY+2). Parr length, fall and spring discharge, 
and fall and spring air temperature were included in the 
analysis and new linear combinations of the standardized 
° Orsi, J. A., and E. A. Fergusson. 2016. Annual survey of juvenile 
salmon, ecologically-related species, and biophysical factors in the 
marine waters of southeastern Alaska, May—August 2015. North 
Pac. Anadromous Fish Comm. Doc., 71 p. [Available from website.] 
