7399 - Laserfiche WebLink

Home Browse Search

(260+) Spawning \ of the depth data were to the nearest 0.1 feet, and the velocity data were to the near- est 0.1 feet per second (fps). Task 3: Preliminary 51 Curve Development Following workshop 1, each data parti- tion specified by the species experts was drawn from the database, and the appro- priate test or decision/test conducted. Data partitions for depth and velocity were pooled or retained on the basis of the Kol- mogorov /5mirnov cumulative frequency test (significant at D < 0.05; Press et a1. 1986), while Chi-square (significant at P < 0.05) was applied to test for substrate and habitat partitions. Once the preliminary partitions were determined, means and variances were computed for raw data val- ues of depth and velocity, but no statistics were presented for the substrate data be- cause the predetermined cell distribution rendered these a discontinuous variable (i.e., substrate was classified according to the categories of the modified Wentworth scale, rather than presented as continuous measurements). The data of each partition were then plotted on a frequency histo- gram, and a smooth curve fitted to the data. One of three curve models was used to best fit these curves: (1) exponential polyno- mial (Bovee 1986), (2) exponential density (Haan 1977), or (3) generalized Poisson (Bovee 1986). These curve models were used because they most closely approxi- mated the distribution of the raw data, and it was possible to fit a curve to peak with the raw data without being skewed by data outlyers. This procedure resulted in an equation for each curve which made it re- producible, and it allows for the produc- tion of weighted usable area at any level of parametric resolution. The result of this process was a prelim- inary curve set for each partition of data. Each curve set consisted of three frequency histograms of raw data, one each for water depth, velocity, and substrate with curves fitted to depth and velocity. Each frequen- cy histogram with raw data and curve was then stored as a separate computer file for viewing and modifying with the aid of a limelite projector during workshop 2. All the curve sets and documentation were sent to the species experts and participants in advance of workshop 2 in order for them ----------- co I No Conaonsus Test __----------T''-,,, 1 11 ./'T'--.- /'... /'... · .~~~.-111 _ __1L-- partitioned into Yampa (Y A), Westwater Canyon (WW), and f trammel nets (NET), angling Ita partitions used for prelim- e developed by a Delphi ::me and Turoff 1975) in een workshops, and eval- orkshop 2. The experts Licrohabitat parameters to 1 category II criteria were ocity, and substrate since ations were available from bstrate was classified ac- odified Wentworth scale lnauer 1977); clay or mud 1.062 mm diameter), sand m), gravel (GR, 2-64 mm), ~50 mm), boulder (BO, 25- : (BE), and other (OT) mis- ;ories. Water temperature also considered; however, not available in much of 1 the experts felt that water LId not be considered as a rameter in the traditional hibits seasonal differences t any given time, provide full range of choices. The ~ed, by consensus, to retain em of measure, since most January 1990 J [ R. A. Valdez et a1. to review the preliminary 51 curves and prepare comments. An effort was also made to develop cat- egory I curves for life stages having little or no data using a Delphi technique similar to that applied to other fish species (Crance 1984, 1987; 5tier and Crance 1985). Each species expert was sent a set of blank his- tograms for each life stage and habitat pa- rameter identified in workshop 1 as re- quiring category I criteria development. Each expert was to apply professional judg- ment and expertise, and independently de- velop a category I 51 curve. These curves were to be reviewed, evaluated, modified, and approved during workshop 2. Task 4: Workshop 2 Workshop 2 was held about 3 months after workshop 1. The purpose of this sec- ond workshop was for the species experts to evaluate and approve the preliminary 51 curves, and to develop category I curves for life stages having little or no data. Workshop 2 was attended by 15 people, including 3 workshop organizers, 6 species experts and 6 participants. This workshop was conducted in a manner similar to workshop 1. Each set of preliminary curves was de- scribed by the chairman and data analyst and presented to the species experts on overhead transparencies as well as on hard copies that had been previously sent to the experts. Accompanying statistics were pre- sented for each 51 curve as well as for each test of data partitions that led to its devel- opment. The experts were then asked whether to use the preliminary curve or to pool the data of several partitions to de- velop a new 51 curve. A consensus of ex- perts at workshop 2 was four of six agree- ing on a course of action. Although statistical comparisons were made between all pairs of data partitions, the experts did not rely exclusively on the results of these tests to determine pooling or partitioning. The experts used the sta- tistical tests as a guide to partitioning, and in some cases, overruled a statistical test with biological opinion. For example, al- though microhabitat data collected from fish captured with gill nets sometimes dif- fered from data collected by electrofishing, these conventional gear data were pooled 35 II~