Fish Assemblages and Small Wood in an Incised Sand-Bed Stream of the Upper Gulf Coastal Plain

Use of woody habitats by stream-dwelling fishes is well documented for salmonids and to a lesser extent warmwater sportfishes (e.g., Micropterus spp.). However, the response of fishes to woody habitats in small, sand-bed streams of the upper Gulf Coastal Plain is poorly known even though wood often provides the only substantial cover in these systems. We were particularly interested in the effects of small instream wood on fish diversity, community evenness, dominance, and assemblage composition and structure (abundance). To evaluate these effects, we installed constructed, woody brush bundles over a 475-m segment of West Fork Cypress Creek (Little Tallahatchie River drainage), northern Mississippi, USA. This small sand-bed stream (max. width ~6.0 m) is deeply incised, flashy, and in most of the study segment essentially lacks any woody or other cover. We installed bundles in two treatment patterns along segments of the stream that lacked cover.  To create patchy woody cover in a reach (Patchy treatment, n=2), we installed three clusters of bundles (6 bundles/cluster) at about 30-m intervals (60-m treatment reach).  To create dense woody cover in a reach (Dense treatment, n=2), we installed three similarly spaced clusters (i.e., 30-m intervals) and also placed groups of 3-4 bundles across the stream at 5-m intervals between the clusters.  As a reference condition (Non-Woody treatment), we selected two reaches that essentially lacked woody cover.  We have sampled for fish abundance and composition and taken physical measurements (depth, width) in the treatment reaches four times (July, November 2009, April, September 2010). Across all samples, we captured 2,635 individuals representing 30 species and nine families.  Rarefied richness and evenness (probability of an interspecific encounter, PIE) were significantly higher in the Dense (13.6 species, 0.83 PIE) and Patchy (13.8, 0.83) treatments than in the Non-Woody (9, 0.65) treatment.  Conversely, dominance (proportion of most abundant species) was significantly higher in the Non-Woody treatment (0.54) than the other two woody treatments (both 0.31).  Abundance (fish/m³) was not significantly different among treatments but was significantly different among sample periods; interaction between treatment and sample period was not significant.  Assemblage differences among reaches were manifested primarily in shifts in abundance of common species rather than wholesale species addition or replacement.  Shifts in individual species abundances appeared to be congruent with habitat affinities and perhaps even predator avoidance.  Addition of woody habitat positively affected diversity and evenness and also was associated with assemblage shifts in this small, highly modified stream.