|
<br />
<br />(260+)
<br />Nonspawning Adults
<br />
<br />I I I I
<br />I I
<br />
<br />Dec(Test Dec/Test Dec(Test Dec/Test
<br />
<br />/1\ /1\ /1\ //I~________
<br />
<br />Y,'l~1 "f 1~11' I 111
<br />
<br />~ ~ /'-. /'-.
<br />GEAR: I EL NET ANG OT ~ EL ~NET EL NET EL NET
<br />
<br />::::~~:. I I ~1 1~6l1l
<br />_;~~.11J_ ___J-L___ __ __ __J. JL J_ _ __ JLL__
<br />
<br />CRITERIA:
<br />
<br />(Length-mm)
<br />LIFE STAGE
<br />
<br />(<2')
<br />Larva
<br />
<br />(21-74)
<br />yay
<br />
<br />(78-'49)
<br />Juvenile 1
<br />
<br />RIVER:
<br />
<br />STRATA:
<br />
<br />YEAR:
<br />
<br />('50-259)
<br />Juvenile 2
<br />
<br />(260+)
<br />Spawning
<br />
<br />FIGURE 1. Dendrogram of data partitions for humpback chub; 'river' is partitioned into Yampa (YA),
<br />Green (GR), and Colorado (CO); 'strata'is partitioned into Blackrocks (BR), Westwater Canyon (WW), and
<br />Cataract Canyon (CA T); 'gear' is partitioned into electrofishing (EL), gill and trammel nets (NET), angling
<br />(ANG), and miscellaneous gear types (OT). The circled numbers represent data partitions used for prelim-
<br />inary 51 curve development.
<br />
<br />then reached on which of the identified
<br />partitions were to be retained. This ap-
<br />proach was used instead of discussing each
<br />individual study, to ensure that standard
<br />selection criteria were applied to all of the
<br />studies included in the database, and to
<br />reduce the amount of time needed to dis-
<br />cuss each study in the 23-year span of data.
<br />Where more than one partition was re-
<br />tained for a particular criterion and the
<br />experts were uncertain about partitioning
<br />or pooling, one of two approaches was used:
<br />The test approach was used to determine
<br />pooling or partitioning strictly on the basis
<br />of statistical results (significant at P < 0.05),
<br />while the decision/test provided the ex-
<br />perts with the statistics but retained the
<br />partitions for a final decision during work-
<br />shop 2.
<br />Based on the data available for each life
<br />stage, following the dendrogram analysis,
<br />the species experts identified the source of
<br />each 51 curve as either category I or II cri-
<br />teria. The category II curves were to be
<br />developed as preliminary 51 curves and
<br />evaluated at workshop 2. The category I
<br />
<br />I~ 34
<br />
<br />curves were to be developed by a Delphi
<br />technique (Linstone and Turoff 1975) in
<br />the interim between workshops, and eval-
<br />uated during workshop 2. The experts
<br />agreed that the microhabitat parameters to
<br />be evaluated with category II criteria were
<br />water depth, velocity, and substrate since
<br />these field observations were available from
<br />the database. Substrate was classified ac-
<br />cording to a modified Wentworth scale
<br />(Bovee and Cochnauer 1977); clay or mud
<br />(CL), silt (51, <0.062 mm diameter), sand
<br />(SA, 0.062-2.0 mm), gravel (GR, 2-64 mm),
<br />rubble (RU, 64-250 mm), boulder (BO, 25-
<br />400 em), bedrock (BE), and other (aT) mis-
<br />cellaneous categories. Water temperature
<br />and cover were also considered; however,
<br />cover data were not available in much of
<br />the database, and the experts felt that water
<br />temperature could not be considered as a
<br />microhabitat parameter in the traditional
<br />sense since it exhibits seasonal differences
<br />and does not, at any given time, provide
<br />the fish with a full range of choices. The
<br />experts also agreed, by consensus, to retain
<br />the English system of measure, since most
<br />
<br />Rivers. Volume 1, Number 1
<br />
<br />J an uary 1990
<br />
<br />of the depth data w
<br />feet, and the velocity
<br />est 0.1 feet per seCOI
<br />
<br />Task 3: Preliminary 5
<br />
<br />Following worksh
<br />tion specified by th
<br />drawn from the dati
<br />priate test or decisioI
<br />partitions for deptl
<br />pooled or retained 0
<br />mogorov /Smirnov (
<br />test (significant at I
<br />1986), while Chi-squ
<br />0.05) was applied to
<br />habitat partitions. (
<br />partitions were det
<br />variances were coml
<br />ues of depth and vel
<br />were presented for
<br />cause the predeterrr
<br />rendered these a d
<br />(i.e., substrate was (
<br />the categories of thE
<br />scale, rather than pr
<br />measurements). The
<br />were then plotted I
<br />gram, and a smooth I
<br />One of three curve D
<br />fit these curves: (1:
<br />mial (Bovee 1986), (
<br />(Haan 1977), or ,(3:
<br />(Bovee 1986). Thes
<br />used because they
<br />mated the distributi,
<br />it was possible to fi
<br />the raw data withou
<br />outlyers. This proc
<br />equation for each Cl
<br />producible, and it .
<br />tion of weighted Ul
<br />of parametric resoh
<br />The result of this
<br />inary curve set for
<br />Each curve set consi:
<br />histograms of raw d
<br />depth, velocity, an,
<br />fitted to depth and'
<br />cy histogram with r
<br />then stored as a sep
<br />viewing and modif
<br />limelite projector d
<br />the curve sets and de
<br />to the species expe
<br />advance of workshl
<br />
<br />I R. A. Valdez et a
<br />
|