ESTIMATING FISH HABITAT SELECTION AND MONITORING STREAM HABITAT QUALITY REQUIRES MORE THAN SIMPLY COUNTING FISH

Identifying techniques to more easily monitor, assess and manipulate habitat quality will improve the assessment of habitat restoration and the management of native fish species. We utilized a relatively novel tool the Sand WandTM (Streamside Environmental, Findley,Ohio) to manipulate stream substrates by removing sand We found that this technique can reduce the proportion of sand in the substrate. We observed a 34% reduction in the area of the streambed covered by sand and a decrease from 44% fine sediment within the streambed matrix before the manipulation to 20% post restoration (Chapter 1). In rivers that are heavily aggraded by fine sediments, a large reduction in fine sediments is likely to measurably increase habitat quality for Brook Trout (Salvelinus fontinalis) and sculpin (Cottus spp). We tested how reliably these changes in habitat quality could be inferred using three different metrics that had been used to infer habitat quality in the literature (Chapter 2). We found that immigration most reliable reproduced the assumed changes in habitat quality that resulted from our manipulation. When this metric was applied to other sites in the same river, we observed that in sandy sites native fish appeared to be attracted to exposed cobble, but this relationship did not hold in less sandy sites. We also found that density, a metric commonly used to assess restorations did not reliable infer changes in habitat quality at this spatial scale. We also illustrate that in at least some systems densities vary greatly over short time scales making measuring density more difficult. Abundances varied from 10-39 individuals over the course of 7 days (Chapter 3) This variation can result in low power if not properly accounted for within sampling designs. Finally, we experimentally tested the attractiveness of artificial boulders in different habitat contexts (Chapter 4). Based on previous observations we expected that boulders added to sandy sites to be more often occupied than boulders in rocky sites. This hypothesis was not supported. We observed very few fish in sandy sites and did not observe an increase in occupancy after the boulder addition. As a whole the work detailed here deepens our understanding of how to monitor and assess restoration and how native fish select and use habitats