SAMPLING DESIGN
All sampling was conducted with a bow-mounted
pushnet on a 20-ft deep-v center-console fiberglass boat powered by a 150-hp outboard
engine. The pushnet is a 5.2-m long (body 3.0-m, cod end 2.2-m), four-panel, 1.5 x 1.5-m
Cobb trawl net modified to fit the pushnet frame. The net is constructed of 1.9-cm stretch
mesh, number 110 knotless nylon netting in the body and 1.27-cm stretch mesh, number 126
knotless nylon netting in the cod-end. The pushnet design and its attributes for sampling
juvenile pelagic fishes is described in Kriete and Loesch (Trans. Amer. Fish. Soc.
109:649-652, 1980).
A stratified random sampling plan (SRS) was employed. Each
nursery zone was divided into a series of strata, each 9.3 km (5 nautical miles), and each
stratum further divided into five 1.9 km substrata. Perpendicular to this stratification,
the 9.3 km sections were divided into three nearly equal parts, a center section and two
shoreward sections bounded by the 1.8 m depth contour lines at mean low water (MLW)
indicated on the respective navigation charts. Thus, each 9.3 km stratum was partitioned
into 15 sites. Three sampling sites were randomly chosen from the 15 in each stratum.
Station
map.
The nursery zone in each sampling period was demarcated by
the last upstream and the last downstream stratum in which juvenile Alosa were
captured. A dynamic nursery zone, rather than a static one, and an SRS were chosen because
there is a shift in availability of juvenile Alosa within the nursery zone caused
in part by low summer river flows and the encroachment of saline water. Within the limits
of the nursery zone, juvenile abundance is generally greatest in the central or near
central strata, and this pattern of the distribution of density also shifts as the nursery
zone limits change. The use of a SRS design where there is a shift in availability and/or
the density distribution avoids the inherent possibilities in a completely randomized (CR)
design of expending a large proportion of the sampling effort either in an area where the
fish were previously, but not presently, available, or in a limited area of heavy
concentration.
Sampling was performed on a weekly basis beginning in the
first week of June and continues until the maximal mean CPUE is definitively observed. All
sampling was performed at night because of the negative phototropic behavior of juvenile Alosa
(Loesch et al., 1982). Sampling began 45 minutes after sunset. This is the time at which
availability of juvenile Alosa is maximized to the pushnet (Dixon, 1996). Sampling
began in the upstream strata (i.e., between RM 69 and 65 on the Pamunkey and 59 and 55 on
the Mattaponi rivers) and proceeded downstream (with 3 randomly selected stations within
each strata) until saline water is encountered or until no more juveniles were collected
in a 9.3 km strata. Although stations within strata were selected randomly, the actual
order of sample collection was non-random. Random order collection (e.g., potentially
sampling at RM 69 then 35 then back to 68, etc.) is logistically impossible. For the
Pamunkey River, a typical sampling cruise included 15 to 18 samples while 12 to 15 samples
were typically collected on the Mattaponi River.
To calculate the volume of water sampled, a calibrated
flowmeter was mounted in the mid-point of the net. All samples were collected against the
current. Previous trials with this arrangement, however, indicated that there was no
significant difference in volume of water filtered when samples were taken with or against
the current, and the overall mean volume was 655 m3 (Loesch et al., 1982). In
practice, samples of 5-min duration at 1200 rpms (i.e., speed of 1-m/s relative to the
water column) are taken, and adjusted, as flowmeter values indicate, to the standard of
655 m3 of water filtered (i.e., 1 unit of effort).
Juvenile catch data were also adjusted for a minimum fish
size. Small juvenile Alosa capable of passing through the 12.7 mm stretched mesh of
the pushnet codend are retained to varying degrees by larger fish and debris in the net.
To ascertain escapement, a sleeve of 6.36 mm stretched mesh was loosely fitted over the
codend in a series of 25 samples in 1979 (Loesch and Kriete, 1983). Only 5.4% of the fish <
26 mm were retained in the codend, and a fork length of 27 mm was chosen as a lower limit
for catch-effort considerations. It is believed that this limit increases the reliability
of the estimates, but it is also recognized that the effect of masking (Pope et al., 1975)
could be confounded in the data. However, the effect is believed to be nonsignificant
because the larger counts in the sleeve occurred before the maximal mean CPUE was
attained.
Previous | Next | Home