Unwin: Survival of Chinook salmon in relation to spring and summer mainstem flows of the Rakaia River, New Zealand 
823 
to a 10 km 2 surface layer that is 10 m deep. Bevolve- 
ment of the resulting halocline would presumably 
require considerable input of wave energy, and un- 
der calm conditions, a surface layer of low salinity 
water may persist for some days until dispersed to 
the north-east under the influence of the Southland 
current (Heath, 1972). 
Implications 
The analysis described in this study relates only to 
ocean-type fry. Because ocean-type fish make up ap- 
proximately two thirds of the Rakaia adult popula- 
tion (Quinn and Unwin, 1993), fluctuations in their 
abundance would have a significant impact on brood 
year survival. However, one third of Rakaia adult 
chinook have a stream-type juvenile life history, and 
there is some evidence that the emergence of stream- 
type behavior in what was originally a fall-run stock 
represents an adaptive response to the lack of es- 
tuarine waters on New Zealand’s salmon-producing 
rivers (Unwin and Glova, 1997). The freshwater habi- 
tat preferences and migration patterns of stream- 
type chinook in the Rakaia River are poorly under- 
stood, and their sensitivity to flow variations is un- 
known. However, the possibility that survival of 
ocean-type fry may increase in seasons of variable 
flow could provide a compensating selective force that 
would help to establish a balance between the incidence 
of the two phenotypes. If so, the ratio of ocean- to 
stream-type fry in any one annual cohort should also 
be positively correlated with the variability of spring 
flows into the Rakaia River during the first year of life. 
To explore this hypothesis further, and hence to 
provide an independent test of the plausibility of the 
mechanisms outlined in the previous section, I ex- 
amined archival material from NIWA’s scale collec- 
tions, aspects of which are summarized in Tables 2, 
3, and 4 of Quinn and Unwin (1993). These records 
include scale samples from salmon taken in the 
Rakaia sports fishery for nine annual cohorts be- 
tween 1965 and 1984. These scale samples permit- 
ted returning fish from each cohort to be classified 
according to juvenile life history. The incidence of 
ocean-type fish among 3-year-old adult chinook was 
positively correlated with mainstem flow variability 
over the period 9 October to 17 November (r 2 =0.54, 
P=0.Q25; Fig. 6) during the year in which they emi- 
grated as juveniles. Of particular interest is the high 
incidence of ocean-type fish in the 1973 and 1982 
cohorts, which were also the broods for which sur- 
vival was highest. 
With regard to the New Zealand salmon fishery, 
which is managed purely for recreational anglers 
90 n 
80- 
JD 
O' 
CO 
i 70- 
03 
<1) 
f 60- 
CD 
O 
O 
50- 
40- 
1. 
Mean flow/median flow 
Figure 6 
The relation between the incidence of ocean-type adults 
among 3-year-old chinook salmon caught by Rakaia River 
anglers over nine seasons from 1965 to 1984, and the ratio 
of mean to median discharge in the Rakaia mainstem for the 
corresponding juvenile cohort, for the same time interval (9 
October to 17 November) as that represented in Figure 4B. 
Sample sizes averaged 127, and ranged from 36 to 373. 
73 
+ 
80 + 
I I I 1 1 1 
,0 1.1 1.2 1.3 1.4 1.5 1.6 
(McDowall, 1994b), the main conclusion to be drawn 
from this study is that despite the observed correla- 
tions between survival and flow variability, interan- 
nual variation in survival remains at best weakly 
predictable. However, spring flow variability is easy 
to analyze on a season by season basis, and the pos- 
sibility of making at least a broad prediction identi- 
fying years of high or low survival up to two seasons 
in advance suggests the effort is worth making. The 
New Zealand Fish and Game Council (which is re- 
sponsible for management of the sports fishery) have 
recently instituted annual monitoring programs for 
key spawning runs in all major salmon rivers, 9 and 
these will eventually enable comparisons between 
survival and flows to be made for other major east 
coast catchments. My results also suggest that any 
reduction in flow variability resulting from develop- 
ment of storage impoundments for hydroelectric or 
irrigation purposes would have a significant nega- 
tive impact on the fishery over and above any losses 
caused by barriers to upstream passage. The high 
bed load of major braided rivers such as the Rakaia 
makes them unattractive candidates for impound- 
ment, but declines in salmon runs in two other riv- 
9 Webb, M. 1997. Central South Island Fish and Game Coun- 
cil, PO Box 150, Temuka, New Zealand. Personal commun. 
