450 
Fishery Bulletin 109(4) 
and 50°N in different years were all strongly and posi- 
tively correlated with shad counts at Bonneville Dam 
in the same years (Table 4). These positive correla- 
Bottom temperature (°C) 
Figure 8 
Cumulative frequency of sampling effort (number 
of tows), log 10 -transformed catch, presence-absence 
(l=present, 0 = absent), and raw catch of American shad 
( Alosa sapidissima) vs. (A) sea-surface temperature (SST 
°C), and (B) bottom temperature during the Alaska 
Fisheries Science Center sampling, and (C) bottom 
temperature during the Northwest Fisheries Science 
Center sampling (both early and late season sampling 
combined). 
tions indicate that the abundance estimates of shad off 
northern Oregon and Washington during these surveys 
were good indicators of the spawning populations of 
Columbia River shad above Bonneville Dam, despite 
the fact that shad were not a target species of these 
demersal surveys. Average weight of shad in ocean sur- 
veys was also strongly negatively correlated with shad 
counts at Bonneville Dam (Table 4), indicating that the 
proportion of younger year classes in ocean sampling 
increased as the abundance of spawners increased. 
The numbers of shad counted are related to ocean 
conditions and the survival of coho salmon as indi- 
cated by the Oregon Production Index or OPI (Fig. 9). 
The OPI is an index of smolt-to-adult survival of coho 
salmon mainly from Columbia River hatcheries. We 
assumed that ocean conditions that affect the survival 
of coho salmon may also affect the survival and re- 
turn of Columbia River shad. During the cool Pacific 
Decadal Oscillation (PDO) regime between 1970 and 
1976 shad counts were comparatively low. At this time 
coho salmon survival was high (Fig. 9, A and B). Dur- 
ing the relatively warm PDO phase from 1977 to 1998 
shad counts increased rapidly, whereas coho salmon 
survival was generally low, especially during the warm 
ocean conditions of the late 1990s. After 2000 shad 
counts increased markedly with warm PDOs, whereas 
coho salmon survival increased to high levels follow- 
ing several earlier years with cool PDOs and then de- 
clined. Shad counts at Bonneville were significantly 
negatively correlated with the OPI index (n=39 years, 
P = -0.45, P=0.004). From these trends we conclude 
that ocean survival of shad and coho salmon off Oregon 
and Washington are inversely related and that warm 
ocean conditions favor increased shad abundances and 
cool, more productive ocean periods favor coho salmon 
survival. In purse seine sets there was also a positive 
correlation between log 10 (caic/i) of shad and SST (n=29 
sets, P = 0.43, P=0.02, SST from about 12.2° to 16.4°C). 
The occurrence of shad off Kamchatka in 1935-1939, 
and again in 1987, all during warm phases of the PDO 
(jisao.washington.edu/pdo, accessed July 2011), indicate 
that shad distributions may increase with predicted 
future climate change and a warmer ocean, just as 
Pacific hake, Pacific sardine, Pacific mackerel (Scomber 
japonicus), and jack mackerel (Trachurus symmetri- 
es) increased off Oregon and Washington after ocean 
warming increased in 1977 (Ware and McFarlane, 1995; 
Emmett and Brodeur, 2000; Emmett et ah, 2006). 
In recent years, numbers of shad counted at Bonnev- 
ille Dam have decreased dramatically. The run in 2010 
was the lowest since 1982 (Columbia Basin Bulletin 13 ). 
Reasons for this decline are unknown, but increased 
incidence of a protozoan parasite, endemic to the Pacific 
Ocean has been suspected (Columbia Basin Bulletin 13 ). 
13 Columbia Basin Bulletin. 2011. American shad: non- 
native to Columbia Basin, runs exceed one million fish, 
peaking at 6.5 million. The Columbia Basin Bulletin, May 
13, 2011. [Available at: http://www.cbbulletin.com, accessed 
July 2011.] 
