The influence of sampling design on the characterization of in-stream salmonid habitat
Loading...
Date
2019
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Montana State University - Bozeman, College of Letters & Science
Abstract
Pacific salmon have endured widespread population extirpations with some estimates at nearly one third of historical populations. In the western coterminous United States Pacific salmon no longer inhabit upwards of 40% of their historical freshwater range. Reintroducing Pacific salmon has therefore become a common conservation effort. An early step in evaluating potential reintroductions includes quantifying the available habitat, however the quantification, and interpretation of the habitat can be influenced by the sampling design and methods chosen. Reach-based sampling designs have been used extensively to collect fisheries related data; however, few studies have examined how reach-based inferences may be biased, a particular concern given the non-random distribution of factors such as woody debris and the magnitude of site-to-site variability. To address this concern, I collected reach-based habitat data continuously within streams. I then used simulations to resample the streams which were delineated into discrete reaches. During simulations I applied simple random, simple random with unequal probability, and generalized random tessellation stratified sampling designs and chose three habitat attributes that are commonly collected in stream habitat surveys, thought to be important factors for Pacific salmon survival, and expected to be distributed differently across the riverscape. My goal of identifying potential bias and precision under these sampling designs was achieved by summarizing simulations and comparing simulated results across streams, attributes, sampling designs and ultimately the census derived estimate of an attribute. My results indicate the extent of bias and levels of precision varied not only across habitat metrics but also across streams. My analyses suggest the use of reach-based approaches, particularly with low sampling efforts, can result in substantially different estimates of habitat characteristics and erroneous estimates of habitat carrying capacity of fishes.