California. Dept. of Fish and Game.

California fish and game (Volume 76, no.3) online

. (page 4 of 7)
Online LibraryCalifornia. Dept. of Fish and GameCalifornia fish and game (Volume 76, no.3) → online text (page 4 of 7)
Font size
QR-code for this ebook






















TABLE 3— Continued

66 — — —

67 — — —

68 — — —

71 — — — 1 32.00

73 — — —

74 — — —

75 — — —

80 — — — 1 35.00 —

82 — — — — — —

84 — — — 1 30.00 —


The results of this study demonstrate that total mortality (Z) is significantly
less when section rather than surface ages are used for slow-growing, long-lived
fishes such as 5. pinniger and 5. diploproa (Table 2). As evident in the larger
reduction in Z for males versus females, the problem is more pronounced in
males, possibly because their slower growth increases the difficulty of assigning
accurate surface ages. This may explain why Z is higher in male 5. diploproa
than in females when surface ages are used and lower when section ages are
used. The absence of older 5. pinniger females also contributed to the smaller
decrease in Z when section ages were used.

This study has documented important differences in estimates of growth and
mortality which result solely from differences in otolith ageing methodology. It
is hoped that subsequent research will incorporate these estimates into various
stock assessment models to evaluate the effect that these different otolith ageing
techniques have on models designed to evaluate yield and production
characteristics of long-lived, slow-growing fish stocks.


This work was supported by a cooperative agreement (number 80-ABH-
00049) and contract (number 81-ABC-00192-PR6) from the Northwest and
Alaska Fisheries Center, Seattle, Washington.


Archibald, C. P., W. Shaw, and B. M. Leaman. 1981. Growth and mortality estimates of rockfishes (Scorpaenidae)
from B.C. coastal waters, 1977-1979. Can. Tech. Rep. Fish. Aquat. Sci. 1048, 57 p.

Beamish, R. ). 1979a. Differences in the age of Pacific hake (Merlucclus productus) using whole otoliths and
sections of otoliths. ). Fish. Res. Board Can., 36; 141-151.

Beamish, R. ). 1979b. New information on the longevity of Pacific ocean perch (Sebastes alutus) . |, Fish. Res.
Board Can., 36: 1395-1400.

Bennett, ). T., C. W. Boehlert, and K. K. Turekian. 1982. Confirmation of longevity in Sebastes diploproa (Pisces:
Scorpaenidae) from ^'° Pb/"'' Ra measurements in otoliths. Mar. Biol. (Berl.), 71: 209-215.

Boehlert, C. W. 1980. Size composition, age composition, and growth of canary rockfish, Sebastes pinniger, and
splitnose rockfish, 5. diploproa, from the 1977 rockfish survey. Mar. Fish. Rev., 42(3^); 57-63.


Boehlert, G. W. 1985. Using objective criteria and multiple regression models foi age determination in fishes. Fish.

Bull., 83: 103-117.
Boehlert, C. W., and R. F. Kappenman. 1980. Variation of growth with latitude in two species of rockfish (Sebastes

pinniger and S. diploproa) from the northeast Pacific Ocean. Mar. Ecol. Prog. Ser., 3: 1-10.
Boehlert, C. W., and M. M. Yoklavich. 1984. Variability in age estimates in Sebastes as a function of methodology,

different readers, and different laboratories. Calif. Fish and Came, 70(4); 210-224.
Campana, S. E., K. C. T. Zwanenburg, and |. N. Smith. In press. -'" Pb/"'' Ra determination of longevity in redfish.

Can. J. Fish. Aquat. Sci.
Chilton, D. E., and R. ). Beamish. 1982. Age determination methods for fishes studied by the Groundfish Program

at the Pacific Biological Station. Can. Spec. Publ. Fish. Aquat. Sci. 60, 102 p.
Dark, T. A. 1975. Age and growth of Pacific hake, Merluccius productus. Fish. Bull., 73: 336-355.
Dark, T. A., M. E. Wilkins, and K. Edwards. 1983. Bottom trawl survey of canary rockfish (Sebastes pinniger),

yellowtail rockfish (5. flavidus) , bocaccio (5. paucispinis) , and chilipepper (5. goodei) off Washington

—California, 1980. U.S. Dep. Commer., NOAA Tech. Memo. NMFS F/NWC-48, 40 p.
Dixon, W. |. 1981. BMDP statistical software. Univ. Calif. Press, Berkeley, 726 p.
Gallucci, V. F., and T. |. Quinn II. 1979. Reparameterizing, fitting, and testing a simple growth model. Trans. Am.

Fish. Soc, 108: 14-25.
Culland, ). A. 1983. Fish stock assessment. A manual of basic methods. John Wiley and Sons, New York, 223 p.
Hirschhorn, C. 1974. The effect of different age ranges on estimated Bertalanffy growth parameters in three fishes

and one mollusk of the northeastern Pacific Ocean. Pages 192-199 in T. B. Bagenal (ed.) The ageing of fish.

Unwin Brothers, Surrey, Engl.
Knight, W. 1968. Asymptotic growth: An example of nonsense disguised as mathematics. |. Fish. Res. Board Can.,

25: 1303-1307.
Leaman, B. M., and D. A. Nagtegaal. 1987. Age validation and revised natural mortality rate for yellow tail

rockfish. Trans. Am. Fish. Soc, 116: 171-175.
Le Cren, E. D. 1974. The effects of errors in ageing in production studies. Pages 221-224 In T. B. Bagenal (ed.)

The ageing in fish. Unwin Brothers, Surrey, Engl.
Neter, |., and W. Wasserman. 1974. Applied linear statistical models. Regression, analysis of variance, and

experimental designs. Richard D. Irwin, Inc., Homewood, III.
Nichy, F. 1977. Thin-sectioning fish ear bones. Fish age determination aids pollution control research. Sea

Technol., Feb., p. 27.
Rao, C. R. 1973. Linear statistical inference and its implications. 2nd ed. )ohn Wiley and Sons, New York, 625 p.
Ricker, W. E. 1975. Computation and interpretation of biological statistics of fish populations. Bull. Fish. Res. Board

Can. 191, 382 p.
Shaw, W., and C. P. Archibald. 1981. Length and age data of rockfishes collected from B.C. coastal waters during

1977, 1978, and 1979. Can. Data Rep. Fish. Aquat. Sci. 289, 119 p.
Six, L. D. and H. F. Horton. 1977. Analysis of age determination methods for yellow tail rockfish, canary rockfish,

and black rockfish off Oregon. Fish. Bull., 75: 405-414.
Vaughan, D. S., and P. Kanciruk. 1982. An empirical comparison of estimation procedures for the von Bertalanffy

growth equation. ). Cons. Cons. Int. Explor. Mer, 40: 211-219.
Wilson, C. D. 1 985. The effects of different otolith ageing techniques on estimates of growth and mortality for two

species of rockfishes, Sebastes pinniger and Sebastes diploproa. M.S. Thesis, Oregon State Univ., Corvallis,

107 p.


Calif. Fish and Game 76 ( 3 ): 1 61 -1 73 1 990





Department of Biological Sciences and

Marine Science Institute

University of California
Santa Barbara, California, U.S.A. 93106

Golden trout, Oncorhynchus aguabonita, from 17 streams in the Kern Plateau
region of the Sierra Nevada, California, were aged using otoliths, and growth rates
were determined using length-age and weight-age relationships. Growth rates,
condition factors, and densities of trout were correlated with site-specific biological
and physical factors using stepwise multiple regression techniques. These stream
populations were highly stunted, and individuals attained quite old ages (9 years).
Densities were usually low and high density had a significant negative effect on
growth ( P < 0.001 ) . In addition, growth was positively affected by amount of aquatic
vegetation, amount of bank vegetation, stream channel stability, and elevation.
While site-specific factors such as trout density may influence trout growth, the low
growth rates throughout the study area were probably due to the low productivity
of these unstable montane streams and the short growing period at high elevations.


Golden trout, Oncorhynchus aguabonita, formerly Salmo aguabonita, have
been widely introduced throughout the United States of America and other
parts of the world, but are endemic to only two watersheds, both in the
southern Sierra Nevada mountains of California (Fisk 1983). There are two
subspecies of the golden trout; O. a. aguabonita is native to the headwaters of
the South Fork Kern River and Golden Trout Creek and O. a. whitei is native to
the Little Kern River, Tulare County. Golden trout were initially thought to be
most closely related to cutthroat trout (Jordan 1892), but are now understood
to belong to the rainbow trout series (Berg 1987).

Although the golden trout is the official California state fish, its natural history
in native habitats is poorly understood. What is known of golden trout natural
history is primarily based on introduced populations of lake-dwelling fish (Fisk
1 983 ) . In these populations, spawning is typically initiated in June and continues
through July. Eggs hatch in 29 to 50 days, and after 2-3 weeks in the gravel the
fry emerge and grow rapidly during the first summer. They reach approximately
4.5 cm by age one year, 12 cm at two, and 19 cm at three years of age
(Needham and Vestal 1938, Fisk 1983). Lake fish reach reproductive maturity
at three years of age and may survive to spawn as many as three times. The
maximum recorded age is six years. Golden trout from streams rarely achieve
lengths greater than 18 cm and ages of such populations are unknown (Fisk
1983). Factors which influence growth rates of stream populations are also

Accepted for publication April 1990.


Growth rates of trout can vary markedly in relation to temperature, food
ration, and density (Elliott 1982). In many cases, the differences within a species
may be greater than those between species. Under good conditions, non-
anadromous trout in streams may achieve a length greater than 40 cm after only
three years (Carlander 1969). Growth is often retarded at higher elevations, due
to lower temperatures and a shorter growing season. Purkett (1951 ) reported
that three-year old rainbow trout, Oncorhynchus mykiss, formerly Salmo
gairdneri, averaged 29 cm at 1,600 m and 24 cm at 1,850 m, and a 24-year old
brook trout, Salvelinus fontinalis, reached only 24 cm at 3,322 m in a low
productivity Sierra Nevada lake (Reimers 1979). If higher elevation habitats
have lower fish densities than lower sites, however, reduced competition at
higher sites may result in faster growth rates (McAfee 1966).

In this paper, we report on ages, growth, and population densities of golden
trout, O. a. aguabonlta, from 12 of its native streams in the Golden Trout
Wilderness, Inyo National Forest. Five additional study streams contained
populations introduced from nearby native populations around the turn of the
century. These included three sites on the eastern edge of the Golden Trout
Wilderness and two sites in Kings Canyon. Several populations in the Golden
Trout Wilderness have been the subject of an intensive management project to
preserve the native habitat and gene pool of the golden trout (Fisk 1983).
Brown trout, Salmo trutta, had been introduced to the South Fork Kern River
drainage, and these predators and competitors were removed by California
Department of Fish and Game biologists from 1976 to 1982. Golden trout were
transplanted extensively during this operation, though trout were not trans-
planted into any of our study streams. Our objective was to provide information
on golden trout ages and growth rates from stream populations and to
determine what biological and physical factors affected trout growth rate,
condition, and density.


The 17 study streams all originate on or near the Kern Plateau of the southern
Sierra Nevada (Inyo and Tulare counties, California; Figure 1). Most are
tributaries of the South Fork Kern River or Golden Trout Creek, both within the
Golden Trout Wilderness. Three streams flow eastward into the Owens Valley
and two are tributaries of the Kings River in the southern portion of Kings
Canyon National Park and drain into the Central Valley.

The southern portion of the Sierra Nevada was largely unaffected by the
Pleistocene glaciation which shaped the valleys north of the Kern Plateau (jahns
1954). Consequently, most of the stream valleys in this region consist of broad
alluvial flats separated by low granitic ridges sparsely vegetated with lodgepole,
Pinus contorta, and foxtail, P. balfouriana, pines. The meadows range in
elevation from 2,300 to 3,200 m, and are composed of relatively unconsolidated
granitic sands and fine sediments. They are more subject to erosion and
degradation than meadows in the glaciated Sierra Nevada (Albert 1982).

Characteristics of the individual streams are provided in Table 1. The streams
of this region are typically of low gradient and stream bottoms consist of
unstable sand and occasional gravels and cobble. Meanders are common in the
unconfined meadow flats, and most streams are relatively wide and shallow.



Discrete riffle-pool sequences are largely absent. The banks are generally
steep-sided, rarely densely vegetated, and active erosion sites are common. True
riparian vegetation is absent, except for mesic herbs and occasional willow
shrubs, Salix spp. In-stream cover which can be used by trout is rare. As a
consequence of stream openness, summer stream temperatures fluctuate
greatly, ranging daily from 3° to 22°C. Golden trout are the only fish species
inhabiting the study streams.


(Cottonwood Cr.


% '2,1


V Korn Pk.^

y e \


MfB. J S

.^ 1 \/> J


J Oloneho Pfc. f
/ * ?

/ s -

J \3oulh Fh. Ktrn RIvtr J

V /^ ^N P-'^

\J V_/

1 •

1 \

i 2

/ \

/ Vi^

\ D

/ ^-^\_

\ o


V c

1 3


V-^ &

5 Blaekreek Mtn.

^ "=


0.5 V


7 __J

FIGURE 1. Map of California showing study sites on the Kern Plateau. Study sites are numbered
and correspond to numbers in Table 1. The two study sites in Kings Canyon
National Park (16, 17) are not shown.


Sampling occurred during the summers of 1983 and 1984. At each site, a 100
m linear transect was established along a representative stream section, and all
samples were taken from this section. The section was mapped three-
dimensionally to assess geomorphological characteristics such as meander
patterns, width /depth relationships, and pool /riffle ratios. Two bottom substrate
cores (15 cm deep x 8 cm diam) were taken and partitioned into five sediment
size classes with sieves (mesh sizes = 2.0, 1.0, 0.5, and 0.1 mm — particles
larger than 2 cm were hand separated from the 2 mm sieve). The three largest
size classes were weighed in the field and the two smallest size classes were
returned to the laboratory for more accurate weighing.

Invertebrates were collected with a 30 cm x 30 cm modified Hess sampler.
The sampler was pushed approximately 10 cm into the substrate and the
substrate within the sampler was vigorously stirred by hand. Any suspended




8 ^

1— '^ v

O ^ *^ i/^ r^
t — no •<r t

^ ^ in rvj

tJ- (N TT ro

C^ 'T O p "^
m rsi ^ ^ ro





— — ^oooooo — o oooo
ooddoddddd do d d d d








*— mrsirvi fNLArs»—

oo ^
T —

— O •— rsi

— t <-M ^












o -

CO 00

rs m in T— rM
in \D I— .— *»■

3 V, 2 3 S S o

o rv o o a^ o
hv r^ CT^ rv ^


in o (^ rn c^
in oo 00 r^ CO



Co t

O fN m a^ P^l
CO O no in hv

r- TT >£ — CO
CO 0> CM 'T hv


(N m (N ON hv.
r*^ fN 0^ v^ lA

5? C!;3

co^DO — 0^ln^^^0^
t— Onnr^i'^rsis^r^Oin


0^ —
in ro

d d

r^< in rsi .— a>
^ rj Tj- ^ .—

d d d d d





v£ CT^ •— in rn

CO nsi d vO -i
r^ •— •—



k. -^







no 00











r^ hx






(N <N





- t


>- uj


^ ~ -

-I >

D "J

Z CQ i


Ln Z

< c

U a: Q

> >-



D 2

3 UJ

> -J „ U i£!

QC LU UJ £ Qi n

.— r^m^inot^ooc^O*—

■^ m o r^


organisms and other organic material was washed into a 330 jitm mesh net
attached to the downstream end of the sampler. Samples were preserved in
70% ethanol for size partitioning and species identification in the laboratory.
Four samples were collected from riffles at each site.

Bank condition was characterized every meter along one randomly chosen
100 m section of stream bank within the study site. At each point, we noted the
type of vegetation, whether or not the stream edge was bare, the presence or
absence of aquatic vegetation (macrophytes and algae), and whether or not the
bank was undercut. If the bank was undercut, the extent of undercutting was

The Pfankuch stream stability rating was calculated for each site (Pfankuch
1975). This index is based on a visual assessment of subjective measures related
to stream channel stability, including bank condition, substrate type, vegetative
cover, width-depth ratio, and pool-riffle ratio. Stream channel stability is
negatively correlated with the index value. For example, narrow, deep streams
with overhanging banks and substrates composed mostly of cobble and gravel
receive a lower score than wide, shallow streams with highly eroded banks and
substrates composed mostly of sand and silt.

Fish were collected by electroshocking (Smith-Root Model 11 Electrofisher).
Three passes were conducted at each site in order to obtain a regression
estimate for population size (Seber and LeCren 1967). However, this method
was not appropriate for the study streams, since the first pass often produced
fewer fish than the second and third passes. Because of the limitations of
population estimates based on the regressions, the data were used to estimate
minimum densities within the stream sections. Fish were retained in buckets
until all passes were completed and then measured.

The standard lengths (sl) of all captured fish were measured to the nearest
mm on a measuring board. We estimated weights by immersing the fish in a
graduated cylinder containing a known volume of water and recording the
volume of water displaced. We assumed, and verified in the laboratory, that the
specific density of fish tissue was similar to that of water ( 1 .0 g/ml ) . Therefore,
a fish that displaced 50 ml of water was recorded as weighing 50 g. Ten fish
from each site, representing a range of sizes, were sacrificed and preserved in
95% ethanol and returned to the laboratory for age determinations. Smaller
samples were preserved from three streams which contained few fish. In
addition, 10 juveniles ( <1 year of age) were collected from each site where
they were present. Since electroshocking did not effectively capture juveniles,
they were instead collected using handnets.

Both scales and otoliths (sagittae) were used for age determination; annuli
and presumed daily growth increments of otoliths were examined at 100-400X
in oil immersion under a light microscope (Campana and Neilson 1985).
Otoliths from fish > 3 years old were first ground with carborundum 600 grit to
enhance transparency, while otoliths from younger fish were viewed directly.

We investigated the importance of 24 biological and physical site character-
istics in determining golden trout growth, condition (K), and density. Site
characteristics included riparian stream cover, substrate size composition, bank
condition, gradient, elevation, suspended particulates, stream surface area,
stream width /depth ratios, pool /riffle ratios, watershed area, abundance of
bank and instream vegetation, fish density, stream channel stability, and aquatic


insect abundance. Data were analyzed using stepwise multiple regression
procedures. The effects of stream channel stability on trout growth, condition,
and density were analyzed separately since the stability index incorporated
many of the other independent variables.

In stepwise multiple regression procedures in which fish growth (SL/age,
weight/age) or condition were included as dependent variables, each fish was
a separate observation. When fish density was entered as the dependent
variable, each stream was a separate observation. Proportional data were
arc-sine square root transformed. The required significance level for inclusion in
the regression model was p < 0.1 5. Only those fish more than one year old were
used in the analyses since young of the year were present in only five of the
study streams when collections were made in 1983.


A total of 376 fish from 17 streams were weighed and measured during 1983.
Of these, 176 fish from 15 streams were preserved for age determinations. In
1984, an additional 20 fish were preserved from two streams from which no
samples had been collected in 1983.

Scales were unsuitable for age determination since annuli were non-existent.
In contrast, annuli were easily counted on nearly all otoliths, possibly due to the
large and discrete differences in seasonal temperature cycles (Campana and
Neilson 1985). Based on annuli counts, golden trout frequently lived more than
five years; the oldest was nine years old. We did not validate annulus formation,
but are confident that our age estimates are accurate, since otoliths accurately
reflected ages of fish up to 23 years old in an extremely stunted brook trout
population (Reimers 1979).

Inter-annular increments were present on all otoliths and were assumed to be
produced daily. Daily increment production has been verified in several fish
species closely related to golden trout including steelhead trout (Campana
1983), Chinook salmon (Neilson and Geen 1982), and sockeye salmon
(Marshall and Parker 1982). The number of increments between successive
annuli ranged from 90-120. Although it is not known for golden trout whether
growth ceases when increment formation stops, the increment counts suggest
a period of rapid growth of between three and four months per year
(June-September). Growth at other times of the year is probably very slow.

Otoliths from young-of-year trout always showed a distinct discontinuity,
after which increments were clearly reduced in width (Figure 2). We believe
this represents the date at which alevins emerged from the gravel after depletion
of the yolk sac. Similar discontinuities were reported from otoliths of sockeye
salmon (Marshall and Parker 1982) and correspond to the date of first feeding.

Lengths of golden trout for all age classes are shown in Figure 3. First year
growth was rapid, after which fish grew at a slower rate, especially after the fifth
year. The length of golden trout for ages 1-9 was fitted to an equation of the
form y,| = 7.30 + 4.02 In x.,^^ (R^ = 0.51, p< 0.0001, n = 138; Table 2).
Weight followed a pattern similar to length, but was more variable (Figure 4).
The weight of trout of ages 1-9 was fitted to an equation of the form
y^, = 2.85 + 19.90 In x^ge (R^ = 0.30, p< 0.0001, n = 138; Table 2). When fit



to a logarithmic equation, the length-weight relationship for all trout was [In
w = -3.59 + 2.70 In I] (R^ = 0.90, p< 0.0001, n = 343).

FIGURE 2. Otolith from a 2.5 cm golden trout. Arrow points to suggested emergence mark. Scale
bar = 100 jam.

Age-length data was also fitted to a Von Bertalanffy growth function
(FISHPARM computer software— Prager et. al. 1987). The Von Bertalanffy
equation for all collected fish was I, = 15.5(1 -exp (-0.42[t + 0.54])). The
asymptotic length for this sample of golden trout was 15.5 cm.

The effects of site characteristics on trout growth (measured as SL/age) were
evaluated for 95 fish from 12 streams using stepwise multiple regression. The
remaining 43 fish were eliminated because their associated stream variables
contained missing values. Fish age accounted for most (62%) of the variation
in trout growth explained by the model (Table 3). The percent of stream
length covered with aquatic vegetation also added significantly to the model and
explained 2% of the variation in trout growth. Fish density (number of
fish/m^)did not vary much between sites (Table 1) and did not explain a

1 2 4 6 7

Online LibraryCalifornia. Dept. of Fish and GameCalifornia fish and game (Volume 76, no.3) → online text (page 4 of 7)