WSP00972 - Laserfiche WebLink

Home Browse Search

5 Relative density of nonnative fishes should nO! be a selection criteria, It will be a response variable. Comments on Evaluation and Refinement of Design Criteria (page 11). The list of response variables is very long and will ultimately complicate analyses of the data. I would select a small number of response variables (i.e., CPUE of native fish, CPUE of nonnative fish I as primary, and additional data as defined by your long list can simply be supportive information around the primary hypothesis testing. In' this manner you will maximize your chance to conclude something concrete from your research and it will allow you to develop an implementation plan based on sound significant differences or none as the case may be. Many response variables will dilute the final conclusions so much that you may never finish the research and get on to management. Comments on Evaluation Design/Methods (pages 11 and 12). \ will re'lterate the importance of a standard survey method. One sampling protocol should be adopted to provide the data to test your null hypotheses. Other sampling can be done as man-power dictates, but unless one standard method is established that can be replicated in each test site type you will have insufficient statistical power to reject or accept the null hypothesis with any precision. I think that the standard survey shculd be designed to sample the test site as a un'it, not individual habitats (j.e., depressions or terraces) within a test site. The amount Of specific habitats in test sites should be quantified and shauld constitute diffe'ences among test sites to be tested by the means differences testing statis~ic. A levee removal site primarliy composed of terrace habitat should not be used as a replicate for a levee removal site primarliy composed of depressions; however, if you make an effort to restore flood function (i.e.. scouring flowsl, after a year or so the river will re-work the site to its own design and then your testing will have to be re-thought. larval fish sampling must not be used as a response variable unless these data can be quantified and qualified. Do larval light traps collect all species in a random fashion in the proportion they exist in the population? To the extent possible, a selection of gear types and fishing effort should be adopted that will provide equalized effort, catch per unit effort between test sites for all target species. In other words, it is not acceptable to use trammel nets in manipulated sites and electro fishing in levee removal sites to provide the response variable for means differences testing. Macroinvertebrates can be a good indicator of differences between test sites. You should adopt a standard protocol such as a series of