A 31 month (September 1974 - March 1977) study was conducted on warmwater streams located in the Roanoke Creek watershed of the Piedmont Region of Virginia. The purpose of the study was to determine the effects of stream channelization on the aquatic/riparian wildlife resource and agricultural land-use patterns associated with the altered streams. Three streams, which were channelized 3, 6, and 10 years prior to initiation of the study, and teo unaltered streams, were selected as representative streams for the study.

Recently channelized streams lacked overstory cover but has an abundance of herbaceous and small woody plany cover, Conversely, control streams had significantly larger percentages of trees over 46 m tall. Plant species diversity, foliage height diversity, and evenness diversity increased as age since channelization increased.

No major differences in water quality parameters were found for either channelized or control streams, although channelized streams had greater deposits of sand and lesser amount of rock, rubble, and gravel. These changes in substrate composition did not significantly modify actual stream flow rates.

Fish species composition and species diversity among channelized and unchannelized streams were only slightly different, with most of the differences probably attributable to strays from adjacent habitats, However, evenness diversity for fish communities was lower in channelized streams.

The benthic population showed greater changes than did the fish populations with an increase in Chironominae tolerant of unstable sand substrates in channelized streams. Evenness diversity of benthic populations was also higher and showed more consistency in the control stream than in channelized streams. Evenness diversity of benthic communities in control stream averaged between 0.5 to 0.6 and was quite consistent; whereas, the average in the two youngest channelized streams was 0.3 to 0.4. These data seem to indicate decreased stability of the biota in altered streams.

In general, benthic macroinvertabrate and fish community parameters collected from channelized streams located 1200 m below a reservoir were either comparable to, or intermediate between, upstream (unchannelized) and reservoir tailwater values. The shallow surface discharge impoundments associated with channelized streams appeared to have a highly localized impact on the downstream benthic marcoinvertabrate and fish communities.

During winter, bird species diversity (BSD) among channelized stream sites was not significantly different. During the breeding season, species richness (number of breeding species) and BSD increased with age since channelization. Breeding bird densities were 6.2 pairs/ha in the most recent (3 yr) channelized site and 13.3 pairs/ha on the control streams. Bird diversity and density, particularly for Parulids (warblers), during the breeding season were reduced significantly by removal of tree and shrub layers along channelized streams.

No significant differences were found among study sites for either total number of small mammals or their species diversity indices; although, there was a trend toward increasing diversity as age since channelization increased. Smaller differences in species diversity values for small mammals on channelized sites than for birds suggests that small mammal populations require less time for recovery following channelization than avian communities.

When streams are channelized: 1) vegetation should be removed from only one side of the stream, with minimal disturbance of top-soil; followed by plantings of herbaceous and woody vegetation, 2) hedgrow plantings should be maintained between agricultural fields and the stream for bank stabilization, 3) dead snags and large trees should be left for birds, 4) all channelization projects should be designed according to the most recent guidelines recommended by the SCS and other resources agencies.

In 1958, the Roanoke County Watershed Work Plan projected annual costs of the structured measures (mainly reservoirs and downstream channelization) to be $79,897 and the average annual monetary benefits to be $111,103. With this favorable benefit/cost ration of 1.4, work began in 1960. In 1970, the annual capital cost was 60,780 and operations/maintenance costs were 10,402, or a total annual project cost of $71, 182. High and low values of annual benefits from agricultural income, water supply, recreation, and non-agricultural flood damage were determined for 1970 and compared to annual project cost. The benefit/cost ratio obtained was between 0.25 and 0.58, considerably lower than the 1.4 estimate of the 1958. work plan. This unsatisfactory ratio for the project was due mainly to the failure of the project to encourage large scale cropping of bottomland area. Future projects should be planned with 1) a greater recognition of constraints on farm operator behavior which affect land use change, 2) conservative projection for land use changes in area where agriculture ids in overall decline, 3) increased use of sensitivity analysis to examine the consequences for project economic justification of alternative land use change projections.