Habitat restoration and seasonal habitat use by Neosho madtoms (Noturus placidus), and spatio-temporal variation of fish assemblages in the Cottonwood River, Kansas.

Abstract

An artificial riffle was constructed in an area previously dredged of gravel on the Cottonwood River, Chase County, Kansas in February 1992, to restore habitat for the Neosho madtom (Noturus placidus), a federally threatened catfish. The artificial riffle and two natural riffles were sampled for one year to assess use of the artificial structure by Neosho madtoms and determine extent of fish colonization on the structure. Physical characteristics of the artificial riffle and species richness, diversity, and stability of its fish assemblage were assessed and compared to those of natural riffles. Noturus placidus seasonal densities were calculated, and population estimates were made by mark-recapture and depletion removal. Neosho madtoms (n = 255) on each riffle were given different subcutaneous marks in an attempt to document inter-riffle movement by the species. Length-frequency data were compiled, and seasonal habitat use by Neosho madtoms was determined. I used monthly samples to determine amounts of ii spatial and temporal variation in riffle and pool assemblages and identify which species varied in abundance seasonally. Before construction, the artificial riffle site was inhabited by "pool" species, whereas post-construction the structure was rapidly colonized and provided suitable habitat for riffle dwellers, including H. placidus. The artificial riffle assemblage was stable and similar to that of natural riffles. Neosho madtoms exhibited limited seasonal variability in habitat use and appeared to be year-round riffle dwellers. Young-of-year were observed in July, suggesting a June spawn. Only one inter-riffle movement was documented, demonstrating little dispersion among riffles throughout the year. Current speed at the substrate surface, low variance in current speed, shallow water, and loosely embedded substratum were variables most useful in predicting Neosho madtom presence. Cannonical correspondance analysis of the overall fish assemblage (six sites) demonstrated differences between riffle and pool assemblages and indicated seasonal variation, but did not show longitudinal variation in the assemblage. Pool and riffle fish assemblages, independently, did not exhibit spatial variation, but were seasonally variable. Riffles were more temporally variable than pools, with 60% of the total fish captured on riffles showing seasonal variation, compared to 12% for pools.