Goetz and Quinn: Behavioral thermoregulation by adult Oncorhynchus tshawytscha prior to spawning 
259 
stressful for populations migrating through estuaries and 
river corridors that have been modified through channel¬ 
ization, reducing suitable holding habitat, and where struc¬ 
tures such as locks and dams have altered tidal regimes and 
hydrology (Hallock et ah, 1970; Potter, 1988; Russell et al., 
1998; Holbrook et ah, 2009). 
The combination of warming conditions and habitat 
alteration is exemplified by Lake Washington (Fig. 1), in 
Washington (Williamson et ah, 2009), where adult Chi¬ 
nook salmon (O. tshawytscha) enter from July through 
September—the warmest period of the year (Fig. 2A). 
Temperatures in the lake and tributaries have been 
increasing, especially in summer (Quinn et ah, 2002; 
Winder and Schindler, 2004; Fig. 2B). In contrast to large 
systems like the Columbia River, where salmonids use 
tributary junctions as thermal refuges (Keefer et ah, 
2018), and some small rivers apparently lacking thermal 
refuges (Hasler et ah, 2012), the Lake Washington sys¬ 
tem’s primary thermal refuge is stratification in the lake, 
but other small-scale thermal structures (i.e., river con¬ 
fluence and pools in rivers) also are available. The sys¬ 
tem also poses special challenges for salmon, including 
an abrupt change in salinity from Puget Sound to the 
migration corridor above the locks and low dissolved oxy¬ 
gen (DO) levels in parts of the corridor. Increasing 
temperatures may adversely affect migration rate and 
spawning success of Chinook and other salmon species in 
this system, and elsewhere (Connor et ah, 2019), but the 
nature and extent of the effects may depend on the 
behavioral thermoregulation and other tactics used by 
salmon migrating through these altered waterways. 
In this study, we used acoustic transmitters and archi¬ 
val temperature loggers to determine how adult Chinook 
salmon migrate through a complex sequence of struc¬ 
tures and ambient water conditions from a small, marine- 
influenced estuary through a set of locks and a shallow, 
warm navigation canal into a large, stratified lake and 
then into tributaries for spawning. We sought to determine 
whether segments of the migration route have conditions 
that delay migration and expose salmon to thermal stress 
from elevated temperatures that might affect survival 
and reproductive success. We did so by posing the follow¬ 
ing questions: 1) do Chinook salmon move back and forth 
between marine and fresh water prior to upriver migra¬ 
tion? 2) do water temperatures experienced in fresh water 
affect transit time? 3) do the salmon minimize exposure 
to warm water by moving to deeper, cooler water where 
available? and 4) is use of cool water affected by ambient 
DO concentration and salinity? In this context, we use the 
term experience to indicate the thermal and other environ¬ 
mental conditions in the water the fish 
occupied, without making any implica¬ 
tion regarding preference or avoidance. 
Materials and methods 
Study area 
Puget Sound is a large (3700 km 2 ) fjord 
system in western Washington bordered 
by British Columbia, Canada, to the 
north (Fig. 1). The Lake Washington 
watershed covers 1274 km 2 and flows 
from tributaries of Lakes Sammamish 
and Washington through a navigation 
channel (henceforth, the ship canal) 
through the north end of Lake Union, to 
the Hiram M. Chittenden Locks (hence¬ 
forth, the locks), and into Shilshole Bay. 
Salmon Bay, immediately upstream of 
the locks, has a saltwater wedge, and we 
refer to this bay as the upper estuary 
and to the area from the locks down to 
Shilshole Bay as the lower estuary. The 
lower estuary has summer mean tem¬ 
peratures of 12-13°C and salinity of 
15-30. The saltwater wedge intrudes up 
to 6 km into the ship canal above the 
locks, but the highest salinities (1-20) 
and most mixing are within 1 km of the 
locks (Fig. 1, insets). The ship canal is 
12.5 km long from Shilshole Bay to 
Lake Washington (10.8 km from the 
Wsss m 
. 0 * 
j V- Central Puget 
Sound 
) d 
■4 S V \ 
, 
pQUthfiiget r \^ 
_ Sound p uya n up River 
i 
Figure 1 
Map showing the acoustic monitoring network in coastal waters near Puget 
Sound, in western Washington, used to track the migration of tagged adult 
Chinook salmon (Oncorhynchus tshawytscha ) in 2005 and 2006. The shade of 
circles indicates the year of deployment of acoustic receivers. The other maps 
show locations of receivers (circles) and of tagging and release events (stars) in 
the ship canal (top) and locks (bottom) in Lake Washington. 
