Norcross etal.: Habitat models for juvenile pleuronectids 
505 
(McIntyre and Eleftheriou, 1968; Gibson, 1973; 
Lockwood, 1974; Poxton et al., 1982; Poxton and 
Nasir, 1985; van der Veer and Bergman, 1986), and 
Japan (Tanaka et al., 1989). Abundance and size dis- 
tributions have been related to water depth (Edwards 
and Steele, 1968; McIntyre and Eleftheriou, 1968; 
Lockwood, 1974; Riley et al., 1981; Poxton et al., 1982; 
Wyanski, 1990), sediment size (Poxton et al., 1982; 
Poxton and Nasir, 1985; Wyanski, 1990; Jager et al., 
1993; Keefe and Able, 1994; Moles and Norcross, 
1995), and food availability (McIntyre and Elef- 
theriou, 1968; Allen, 1988; Jager et al., 1993). The 
generally accepted rationale for juvenile recruitment 
to shallow, fine-grained nursery areas includes es- 
cape from predation, increased cover and food avail- 
ability, and decreased intraspecific food competition 
(Toole, 1980; de Ben et al., 1990; Minami and Tanaka, 
1992). The examination of diet diversity among a 
subset of the fishes in the present study showed a 
reduction of both interspecific and intraspecific di- 
etary overlap when flatfishes coexisted in large abun- 
dances (Holladay and Norcross, 1995). 
The coastline of Kodiak Island, Alaska, encom- 
passes a variety of habitats from shallow, fine-grained 
tidal flats to deep and rocky areas. Kodiak Island is 
mountainous and cut by many fjords and open bays 
with shallow waters (<10 m) usually within 0.5 km 
of the beach. The tidal range is 3 to 4 m. The region 
is characterized by deep bays, rough bottom topo- 
graphy, strong currents, and bottom characteristics 
that change rapidly over relatively short distances. 
Around Kodiak Island, juvenile flatfishes occupy fine- 
grained sediments in bays and nearshore waters, as 
do flatfishes in other locations, but waters less than 
10 m in depth are only a minor component of the 
area that is used (Norcross et al., 1995). 
A nursery may be partitioned into areas dominated 
by individual species or intraspecific age groups 
(Edwards and Steele, 1968; Zhang, 1988; Harris and 
Hartt 1 ; Smith et al. 2 ). Habitats occupied by juvenile 
rock sole, flathead sole, Pacific halibut, and yellow- 
fin sole collected on the east and south sides of Kodiak 
Island in August 1991 can be differentiated on the 
basis of depth, substrate, and within-bay distribu- 
tion (Norcross et al., 1995). 
1 Harris, C. K., and A. C. Hartt. 1977. Assessment of pelagic 
and nearshore fish in three bays on the east and south coasts of 
Kodiak Island, Alaska: final report. In Volume 1: Environmen- 
tal assessment of the Alaskan continental shelf, p. 483- 
688. U.S. Dep. Commer., and U.S. Dep. Interior Quarterly 
Reports of Principal Investigators, Anchorage, AK. 
2 Smith, R. L., A. C. Paulson, and J. R. Rose. 1976. Food and 
feeding relationships in the benthic and demersal fishes of the 
Gulf of Alaska and Bering Sea. In Volume 7: Environmental 
assessment of the Alaskan continental shelf, p. 471-508. An- 
nual Report RU 0284.7. U.S. Dep. Commer. and U.S. Dep. Inter- 
ior Environmental Research Laboratories, Boulder, CO. 
We used linear discriminant functions to identify 
tentively the habitat characteristics of juvenile flat- 
fishes with data collected in August 1991 along east 
and south Kodiak Island (Norcross et al., 1995). In 
this study, we repeated the linear discriminant func- 
tion analysis with combined 1991-92 data to include 
observations from a much wider geographic area 
around the entire island of Kodiak collected in Au- 
gust 1992. We refined our previous habitat models 
by using tree-based regression methods (Venables 
and Ripley, 1994) on catch-per-unit-of-effort data. 
Materials and methods 
Sample collections 
Two cruises were conducted in the nearshore waters 
of Kodiak Island, Alaska, during August 1992 (Fig. 
1). These cruises were similar to, but covered more 
area than, two cruises conducted in August 1991 
(Norcross et al., 1995; Norcross et al. 3 ). Cruise KI9201 
consisted of collections taken with a 7.3-m skiff from 
Kalsin, Middle, and Womens Bays near the town of 
Kodiak during 9-14 August 1992. Because these bays 
were sampled with a skiff, extremely shallow collec- 
tions could be made. Collections ranged in depth from 
1 to 60 m. Ten stations were occupied in Kalsin Bay, 
six stations in Middle Bay, and five stations were 
occupied within Womens Bay. Kalsin and Middle 
Bays were also sampled during August 1991. 
Immediately following the sampling from the skiff, 
a counterclockwise circuit of Kodiak Island was com- 
pleted aboard a 24.7-m chartered trawling vessel (FV 
Big Valley , cruise KI9202). Collections during KI9202 
were made from 16 to 29 August 1992 and ranged in 
depth from 5 to 180 m. Areas sampled in 1992, but 
not sampled in 1991, included 52 stations in bays on 
the north and west sides of the island. Collections 
were also made at 41 stations off south Kodiak, 
Sitkalidak Strait, and in Ugak Bay, which were 
sampled during August 1991. 
Sampling gears, vessels, and vessel operators were 
the same in both 1991 and 1992 (Norcross et al., 1995; 
Norcross et al. 3 ; Norcross et al. 4 ). At each station one 
sediment sample was collected with a 0.06-m 3 Ponar 
grab for analysis of grain size, and a portable con- 
ductivity, temperature, and depth (CTD) profiler was 
deployed to measure temperature and salinity. Fishes 
3 Norcross, B. L., B. A. Holladay, and M. Frandsen. 1993. Re- 
cruitment of juvenile flatfish in Alaska, phase 1. Final Con- 
tract Report, NOAA NA-16FD0216-01, 504 p. 
4 Norcross, B. L., B. A. Holladay, F.-J. Muter, and M. Frandsen. 
1994. Recruitment of juvenile flatfish in Alaska, phase 2. Fi- 
nal Contract Report, NOAA NA-26FDQ156-01, 653 p. 
