Fish community assembly in the Forest River, North Dakota, and resolution of Campostoma species presence.

摘要

I was provided the opportunity to investigate the fish community associated with the Red River basin of North Dakota. The first aspect I examined was the presence of the genus Campostoma. In this aspect, I investigated the Campostoma species that were present historically and documented the species that currently inhabit Red River waters in North Dakota. The second aspect of my work was to study the fish community patterns along the longitudinal dimension of the Forest River. I used nonmetric multidimensional scaling to differentiate the patterns at two scales: level III and level IV ecoregions. The final aspect of my study was to investigate habitat preferences of fishes in the Forest River. I sampled fish by macrohabitat and compared the results to a similar study.;In the determination of the Campostoma species present, I investigated the fishes found in the Red River basin to determine presence of central stoneroller, Campostoma anomalum, and largescale stoneroller, C. oligolepis because both species have been reported in the basin historically. Campostoma sp. were found at five sites. These specimens and museum specimens collected during previous studies were C. oligolepis. I developed a novel approach, termed the oblique circumferential scale count, to counting scales around the body to aid in differentiating C. anomalum from C. oligolepis. The sum of oblique and lateral line scale count differentiates C. oligolepis (n = 84 scales) and C. anomalum (n = 84 scales or greater).;The second aspect of the study was focused on the Forest River of North Dakota, which provides a unique situation to assess both environmental and anthropogenic affects on longitudinal fish assemblage because it flows through several ecoregions and is divided by four epilimnetic-release dams (4.6--23.2 m high) impounding over 50 hectares of water each. Sixteen sites were sampled using seining, backpack electrofishing, and traps. Richness ranged among sites from 5--15 species and from 3--6 families. Shannon's Diversity Index (H') ranged from 0.77--1.95 and evenness (J') varied from 0.24--0.81. Nonmetric multidimensional scaling was conducted using percent similarity index values. Site groupings were based on dams but not level IV ecoregions. Sites between two dams were grouped together, excluding the site immediately below dams. The sites directly below dams separated out closer to the most upstream sites than to the other groups. The serial discontinuity concept explains this finding, but the process domains concept does not. When analyzed at the level III ecoregion scale, the site groupings were closely related to ecoregions. Still, the sites that were directly below a dam were closer to the upstream sites than were the sites further downstream of the dams or upstream of the reservoirs. Results from the level III ecoregion scale analysis agree with predictions of both the serial discontinuity concept and process domains concept. The fish assemblage in the Forest River is influenced by the anthropogenic forces of dams at the level IV ecoregion scale. At the level III ecoregion scale, the anthropogenic forces of dams and natural community breaks based on ecoregions were both evident.;The final aspect of the study was to look into fish habitat preferences. Fish data were collected at 16 sites on the Forest River, North Dakota and four macrohabitats (pools, riffles, runs, and backwaters) that were characterized for depth, velocity, substrate, and cover. Habitat preferences for seven species were in general agreement with the habitat preferences in a published report on fishes of Minnesota. Hornyhead chub, common carp, and common shiners were usually found in run macrohabitats that were 55--85 cm deep. Largescale stoneroller, longnose dace, creek chub, and fathead minnows were usually found in riffle macrohabitats that was 15--35 cm deep. Water velocity for the riffle species was generally slower (15--65 cm/s) than for the run species (65 cm/s). Each species has a specific preference with substrate (sand, gravel, boulders) and cover (vegetation, boulders) or no cover. None of these species were found to prefer pool macrohabitats. These findings indicate that the use of stream specific habitat associations may be better than regional habitat association models when trying to conserve certain species.