Type of Presentation

Poster

Session Title

The Role of Eelgrass Ecosystems in the Salish Sea

Location

2016SSEC

Description

As part of continuing work in Port Gamble, WA a diver-based eelgrass survey was completed to support the application for a Hydraulic Project Approval (HPA), as required by the Washington State Department of Fish and Wildlife (WDFW) for in-water work. The survey was completed following interim guidelines established by WDFW in 2008, although to fit the guidelines to the specific project some methods were modified and approved by WDFW. Collecting statistically robust data proved to be difficult, as shoot density in the eelgrass bed was highly variable. In some areas the variance was so high that power calculations estimated the sample number (n) to be higher than the available number of quadrats. In addition to issues with variance, data were not normally distributed so it is questionable whether parametric statistics should be used on this data. In areas where count data are highly variable and not normally distributed perhaps a more reasonable approach to data analysis is to use a non-parametric resampling method for analysis. This presentation will show the results of a survey completed along an existing wastewater treatment outfall pipe and a reference location, the results of power analysis for the survey, and statistical results comparing the sites using parametric and non-parametric statistics.

Share

COinS
 

A Nonparametric Statistical Approach to Analyzing Eelgrass Density Data

2016SSEC

As part of continuing work in Port Gamble, WA a diver-based eelgrass survey was completed to support the application for a Hydraulic Project Approval (HPA), as required by the Washington State Department of Fish and Wildlife (WDFW) for in-water work. The survey was completed following interim guidelines established by WDFW in 2008, although to fit the guidelines to the specific project some methods were modified and approved by WDFW. Collecting statistically robust data proved to be difficult, as shoot density in the eelgrass bed was highly variable. In some areas the variance was so high that power calculations estimated the sample number (n) to be higher than the available number of quadrats. In addition to issues with variance, data were not normally distributed so it is questionable whether parametric statistics should be used on this data. In areas where count data are highly variable and not normally distributed perhaps a more reasonable approach to data analysis is to use a non-parametric resampling method for analysis. This presentation will show the results of a survey completed along an existing wastewater treatment outfall pipe and a reference location, the results of power analysis for the survey, and statistical results comparing the sites using parametric and non-parametric statistics.