Weitkamp et al Seasonal and interannual variation in juvenile salmonids in the lower Columbia River 
433 
changes in length (collection date as a covariate) and 
length-weight relationships (weight as a covariate). 
Results 
Environmental conditions 
Environmental conditions were extremely variable during 
our study period (mid-April to late June or early July) 
among the 4 years, because of both daily tidal inunda- 
tion and extremely high river-flow events (>14,000 m 3 /s) 
in May 2008 and June 2010 (Table 1); flows during May 
2008 were the highest observed in more than a decade 
(USGS National Water Information System, http://water- 
data.usgs.gov, accessed Aug 2011). Both high flow events 
were characterized by rapid increases in flow, but flows 
remained elevated much longer in 2008 than they did 
in 2010, resulting in 3 cruises in 2008 during which 
mean river flow exceeded 12,500 m 3 /s compared with 
only 1 cruise in 2010 with rates that high (Table 1). By 
contrast, river flow never exceeded 11,600 m 3 /s in either 
2007 or 2009. River flow also was negatively correlated 
with the PDO (r=— 0.48, P<0.05), indicating the influence 
of large-scale climate forces on terrestrial environments 
as expressed by river flow (Lawson et al., 2004). 
River temperature measured both at the Dalles Dam 
and at 1-m depth in the estuary were quite similar 
(Spearman r=0.98, P<0.05), seasonally increasing from 
<10°C in mid-April to 15-17°C by late June (Table 1). 
They also were correlated with coastal SST (r=0.85, 
P<0.05), which displayed similar seasonal increases. 
The snowmelt-driven high flow event in 2008 delayed 
the seasonal increase in temperature, resulting in tem- 
peratures that were somewhat lower in June 2008 than 
in other years. In situ temperatures measured at 7-m 
depth also were positively correlated with upstream river 
temperature (r=0.88, P<0.05) and coastal SST (r=0.78, 
P<0.05), but they were negatively correlated with 7-m 
salinity (r=-0.38, P<0.05), indicating intrusion and mix- 
ing of consistently cold (8-12°C) marine water at depth. 
In contrast to surface temperatures, in situ salinity 
measured before each haul changed dramatically with 
tidal stage, such that salinities ranged from essen- 
tially fresh (salinity<l psu) to largely seawater (>25 
psu) within the 5-7 h of sampling each day. This tidal 
variation was greatest at 7-m depth, and time elapsed 
since low tide was a reasonably good predictor of 7-m 
salinity (linear regression, coefficient of determination, 
r 2 =0.7, P<0.05), but river flow was not (r 2 =0.3, P<0.05). 
However, salinities measured at 1-m depth were influ- 
enced both by tidal cycle and river flow, with maximum 
salinity measured at a depth of 1 m during a cruise 
being negatively related to flow (r 2 = 0.5, P<0.05). In 
particular, maximum salinities recorded at 1-m depth 
when flow exceeded 10,987 m 3 /s were significantly lower 
(mean=3.5 psu) than values recorded when river flow 
was below that level (mean=10.2 psu). This pattern sug- 
gests that high flows result in water column stratifica- 
tion with a surface layer of largely fresh water. 
Fish catches and composition 
We completed a total of 291 purse seine sets (210 round 
and 81 towed) during spring of 2007-10 (Table 1). 
During these 4 years, we caught an estimated 248,822 
fish in round purse seine sets and an additional 58,194 
fish in purse seine tows. These fish represented 27 spe- 
cies (Table 2), including all Pacific salmon species pres- 
ent in the Columbia River: Chinook, coho, chum, and 
sockeye salmon, steelhead, and sea-run cutthroat trout 
(O. clarki). We also documented the presence of 3 spe- 
cies of smelts (whitebait, longfin, and surf), 3 clupeids 
(Pacific herring, Pacific sardine [Sardinops sagax], and 
American shad), and 3 flatfishes (English sole, starry 
flounder [Platichthys stellatus], and sand sole [Pset- 
tichthys melanostictus]). Most fishes were present only 
in juvenile stages, although several species, including 
most of the forage fishes (e.g., northern anchovy, Pacific 
herring, longfin and surf smelts, and American shad), 
threespine stickleback, Chinook salmon, and steelhead, 
were present as both juveniles and adults. 
Catches of individual species-and-age classes were 
highly variable both among species and years, as in- 
dicated by frequencies of occurrence (FO, Table 2) 
and coefficients of variation (CVs, Table 3), and likely 
reflected in part whether species were largely resident 
in the estuary, transitory, or migratory (anadromous). 
For example, 2 species — surf smelt (transitory) and 
threespine stickleback (resident) — were caught in al- 
most every set (FO>80%). Other species caught fairly 
frequently included shiner perch, longfin smelt, Pa- 
cific herring, and American shad. Juvenile salmon 
exhibited varying levels of occurrence, with Chinook 
and coho salmon and steelhead caught in roughly 
half of all sets (FO = 34-65%), whereas both chum 
and sockeye salmon were encountered less frequently 
(FO<24%). By contrast, many species were infrequent- 
ly observed in our catches, including 15 species that 
were completely absent in at least one year (Table 2). 
Mean annual CVs, estimated for the most commonly 
caught species ranged from 47% (surf smelt in 2010) 
to 240% (northern anchovy in 2008), with juvenile 
salmon generally having lower CVs (range: 46-167%) 
than most species except surf smelt and threespine 
stickleback. Accordingly, the abundances of even our 
most commonly caught fishes were highly variable 
among sets. 
Despite this fine-scale variability in abundances, how- 
ever, the catch of fish each year generally followed a 
pattern of increasing abundance and species diversity 
as the spring advanced (Fig. 2). In all years except 2007 
(which had higher overall abundances), mean CPUE 
(number/1000 m 2 ) per cruise was <50 in mid-April, 
increased to roughly 500 in May, and dropped slightly 
in June to 300. The seasonal change in abundance 
was accompanied by a change in species composition 
(Fig. 2B): in all years, the estuarine fish assemblage 
was dominated by threespine stickleback in April, be- 
coming more diverse and variable in May and June 
with sporadically high relative abundances of northern 
