266 
Fishery Bulletin 117(3) 
0-10 10-15 15-20 20+ 
Lake depth (m) 
Figure S 
Proportions of temperature records, from tags implanted 
in Chinook salmon (Oncorhynchus tshawytscha), in Lake 
Washington in western Washington during 2005 and 2006 
for each of the depth strata where those temperatures 
were experienced. 
Therefore, there was more variation in arrival dates into 
the system than variation in ascent into spawning streams. 
The total thermal exposure (DD) from tagging to spawning 
stream ascent was higher in the early arriving fish (means 
for the first week: 844.4 in 2005 and 723.8 in 2006) than 
those arriving in the last week (means: 521.1 in 2005 and 
449.3 in 2006)—a consequence of the warmer water experi¬ 
enced by and longer exposure of the earlier fish. 
Discussion 
The Chinook salmon entering the Lake Washington sys¬ 
tem faced environmental challenges that are both unique 
and generic. Few rivers that salmon ascend have locks 
between fresh and salt waters, and such structures pro¬ 
duce unusual combinations of temperature, salinity, and 
DO level with abrupt changes over short temporal and 
spatial scales, both vertically and horizontally. However, 
although the physical structure and these features of the 
environment are unique, the system is experiencing the 
increases in temperature (Fig. 2) that also characterize 
many migratory corridors for salmonids and other fish 
species (Quinn, 2018). In this context, the increasing tem¬ 
peratures in Lake Washington are simply part of a much 
broader pattern of climate-driven change, although each 
river and lake has its own characteristics. These tem¬ 
peratures pose challenges for salmon; therefore, further 
information on how they confront thermal barriers will be 
important to determine their likelihood of survival (Hasler 
et al., 2012; Fenkes et a!., 2016; Connor et a!., 2019). 
We posed a series of questions to help characterize 
the ways in which Chinook salmon address these chal¬ 
lenges, either by making use of refuges and other forms 
of avoidance when confronted with adverse conditions or 
by tolerating the conditions. In addition, given the later 
migration in the season and generally warmer water expe¬ 
rienced by Chinook salmon compared with that experi¬ 
enced by sympatric sockeye salmon (O. nerka ) (Newell and 
Quinn, 2005), we predicted that the former would show 
.greater tolerance of higher temperatures. In the modified 
Lake Washington estuary, Chinook salmon occupied ther¬ 
mal refuges in stratified marine and freshwater areas. Most 
Chinook salmon left the estuary after tagging and spent a 
few days in Puget Sound, often below the thermo-haloeline. 
Upon return to the upper estuary, they either held in the 
saltwater wedge for a week or more or went back to Puget 
Sound again. These movements may have been responses 
to the abrupt change in salinity. If the fish were not fully 
tolerant of fresh water, the scarcity of intermediate salini¬ 
ties may have stimulated them to return to Puget Sound 
and to enter the ship canal again a few days later. To reach 
the upper estuary, most Chinook salmon migrated through 
the locks and were often below the thermo-haloeline in the 
lock, where temperatures were intermediate between those 
in Puget Sound and the upper estuary. A smaller fraction of 
the Chinook salmon ascended the fish ladder. Temperatures 
in the ladder were much warmer than in the locks because 
it contained the surface water from the ship canal. Although 
the tendency to use the locks rather than the ladder was 
consistent with avoidance of warm water, it is certainly not 
proof of it, as the structures differ in many other physical 
aspects besides temperature. 
The warmest and likely most stressful conditions that 
all Chinook salmon had to experience were in the upper 
estuary, which received little marine input and was domi¬ 
nated by the warm surface waters from Lake Washington 
and Lake Union. Fish there occupied the deepest waters 
below the thermo-haloeline for extended periods prior 
to migration to Lake Washington, apparently balancing 
the benefits of low temperature against the drawbacks 
of low DO concentration. Chinook salmon tagged during 
the warmest periods in early August remained longer in 
the estuary (lower and upper) before migrating rapidly 
upstream through the ship canal than did those arriving 
in September, when the water was slightly cooler. Once 
in Lake Washington, most fish resided for some period at 
depths near the thermocline with vertical migrations to 
use cooler water, and they eventually moved toward shal¬ 
lower river deltas where they experienced high tempera¬ 
tures before entering rivers with cooler water. 
Despite avoiding the warmest surface waters, all Chi¬ 
nook salmon experienced temperatures > 18°C and 85% 
of the Chinook salmon experienced temperatures >20°C, 
a threshold associated with adult migration delay, stress, 
and potential mortality (Goniea et al., 2006; Keefer et ah, 
2009). Chinook salmon in the upper estuary and Lake 
Washington occupied vertical positions that minimized 
their exposure to the warmest temperatures, as do Pacific 
salmon in other man-made and natural impoundments 
(Keefer et a!., 2009; Strange, 2012; Keefer et ah, 2015). 
Diel patterns in the upper estuary followed those seen in 
some other salmonids: deeper during the day than at night 
(Newell and Quinn, 2005; Mathes et al., 2010; Roscoe et al., 
