be readily identifiable, and in some streams corrective action would be futile or prohibitively expensive. A stream. that is being degraded by poor land use practices benefits little from instream structures unless conditions in the watershed are also improved. Therefore, before stream improvement work is undertaken, factors such as physical and chemical stream and watershed characteristics, flow regime and bank stability must be analyzed.

The effectiveness of various structures is also dependent on proper placement and procedures, as well as on watershed characteristics. Procedures that are functional in certain areas have proved unsatisfactory in others. For example, the removal of the stream canopy recommended for northern trout streams, to promote the production of plankton and aquatic macrophytes, would be detrimental on southeastern trout streams, where removal might result in critically high summer water temperatures. The improper use or design of habitat improvement structures can be more detrimental than no action, if the structures result in bank erosion, increased flooding potential, or formation of migrational blocks.

State and Federal programs designed to increase the abundance of stream fishes through habitat management were initiated in the 1930s. The Civilian Conservation Corps provided the labor for construction, but thorough pre- and post-project evaluations were seldom

conducted. Most of the early work was concentrated on trout streams. Limited analysis of results showed highly variable responses. Improper design and construction and incorrect identification of limiting factors were the primary

causes of failure of many of these structures to improve stream conditions. Log dams were often washed out or undercut, or the stream banks were eroded as a result of the dam presenting too high a profile or being improperly reinforced. In some streams erosion problems were created by the placement of structures in unsuitable locations or in proximity to unstable banks.

Results were successful, however, when structures were

designed within the limitations dictated by the physical characteristics of the individual stream and the limiting factors of the fish population concerned. For example, Shetter, Clark and Hazzard (1946)2 found that log deflectors placed in a Michigan trout stream that had a sand and gravel base increased depth and number of pools, trout standing crop, and angler success. An increase in trout food species was also noted.

The importance of cover to the fish has been evaluated in some recent studies on stream alteration. It has been noted that the territory occupied by trout is often oriented in relation to stream cover such as large rocks, turbulent surface waters and submerged logs. Shaded areas form preferred habitat in water that typically has high light transmittance. Hiding cover is also important in reducing predation and competition. An example of data of this type is presented by Saunders and Smith (1962)3 who found 2Shetter, D. S., O. H. Clark and A. S. Hazzard. 1946. The effects of deflectors in a section of a Michigan trout stream. Trans. Amer. Fish. Soc. 76: 248-278.

3Saunders, J. W. and M. W. Smith, 1962. Physical alteration of stream habitat to improve brook trout production. Trans. Amer. Fish. Soc. 91 (2): 185-188

higher survival rates of brook trout in a Canadian stream after the emplacement of log dams, deflectors, and wired alder and spruce cover devices.

There has been little study of the effects of habitat improvement on warm-water streams. However, several attempts have been made to offset the detrimental impact of channelization by the installation of stream improvement structures. In general, channelization reduces habitat variability, water depth, and velocity and reduces the total amount of habitat by reducing stream length. The loss of environmental variety results in the loss of organisms whose requirements are not met by the restricted conditions. Habitat improvements in channelized streams are usually designed to increase habitat diversity through formation of pools and cover. Such a procedure can simulate prechannelization conditions. For example, the installation of deflectors and check dams in a channelized section of the Weber River, Utah, resulted in the restructuring of the channel to resemble the riffles and pools of an unaltered stream (Barton and Winger, 1973).4 Stream improvement structures are currently widely used in the southeastern U.S. to offset the effects of habitat changes generally resulting from changes of land use in the watershed. Recently we have been involved in a study of these procedures, particularly in Virginia trout streams, which was funded by the Virginia Water Resources Research Center (Maughan, Nelson and Ney, 1978)5. The structures on the four study streams consisted of log dams, rock-filled gabions (rectangular wire cribs of various dimensions) and boulder emplacements. Log dams and gabions are intended to restructure streams to provide additional deep water habitat, and boulder emplacements are designed to increase instream cover and hiding space. Pools are formed below dams through the scouring of substrate materials, and often above the dams from the impounding of the stream. Gabions deflect the current into a restricted channel and, through an increase in velocity, scour a deeper channel. Gabions are used singly or in pairs and also can function to remove silt from stream channels, close alternative channels, and protect eroding banks. The use of log dams is restricted to streams with a maximum discharge below 100 cubic feet per second (cfs). At higher stream flows log structures are washed out and destroyed. On streams subjected to occasional flooding, dams must present a low profile, have their ends well embedded in the banks, and be reinforced with rock.

To ev

evaluate the effectiveness of various structures, we compared stream physical characteristics, aquatic macroinvertebrates and fish populations in improved and unimproved sections of four Virginia streams.

Fish were collected with a 230-volt DC electro-shocker from 50 meter sample stations. Block seines at the station boundaries precluded the movement of fish into or out of the area during sampling. Benthic invertebrate populations were collected using a circular substrate sampler. The

'Barton, J. R. and P. V. Winger. 1973. A study of the channelization of the Weber River, Summit County, Utah. Utah Division of Wild. Resour. and Utah State Dept. of Highways.

5 Maughan, O. E., K. L. Nelson and J. J. Ney. 1978. Evaluation of stream improvement practices in southeastern trout streams. Va. Water Resour. Research Cent. Bull. 115.

physical characteristics examined included stream width, depth, current velocity, substrate composition and riffle-pool ratio.

The study streams were North and Jenning Creeks (Jefferson National Forest, Botetourt Co.), Comers Creek (Mt. Rogers National Recreation Area, Smyth Co.) and Crooked Creek (Carroll Co.). The streams are typical of southeastern trout waters, which flow along mountain slopes and are characterized by steep gradients, rapid runoff and periodic flooding. Three of the streams, Crooked Creek excepted, flow through forested land, are well shaded, and have substrates composed primarily of stone and gravel. Oak, hickory, and yellow poplar dominate the forest community, and rhododendron and mountain laurel are abundant in the understory. In contrast, the Crooked Creek watershed supports extensive livestock grazing, and adverse agricultural practices contribute substantial amounts of sediment to the stream. Streamside vegetation along Crooked Creek is primarily multiflora rose, creek willow and sycamore.

The

he fish community of the streams is dominated by the minnow family (Cyprinidae), the most common representatives being the stoneroller, mountain redbelly dace and blacknose dace. Other abundant fishes are the mottled and slimy sculpin, torrent sucker and the northern hog sucker. The number of species of fish present ranged from five on Comers Creek to 15 on Jennings Creek. The aquatic insects found in the study streams are characteristic of cold, well-oxygenated and high-quality waters. Many species of mayflies, stoneflies and caddisflies, along with the true flies (Diptera), comprise most of the benthic fauna.

On North Creek, pool area in the improved section was approximately seven percent higher in the improved section than in unimproved sections. The greatest area of pool formation occurred above the dams, with smaller pools scoured below each dam. Since the dams on North Creek had been in place more than 35 years, the upstream pools were apparently stable. An increase in pool area was also noted within the structure-improved sections of Comers Creek. However, two dams were washed out in April 1977 when discharge reached approximately 240 cfs. The improper placement of the dams immediately downstream from a meander was probably the primary factor involved in their destruction. The use by trout and other fish species of the pools found in association with the log dams was greatest during summer low flow periods, when other stream sections began to dry. Increasing fishing pressure in the large conspicuous pools was also observed.

In North Creek the total biomass of fish cumulated over the study period was greater within the sections containing the log dams than in unaltered areas. During four of the six fish samples taken on the two streams containing log dams, total weight of trout was greatest in the sections in which the habitat had been improved. However, the dams may not have functioned optimally because the pools formed by the structures were generally devoid of cover, except for areas of turbulent water below the dams and the rock used to reinforce the ends of the structures. No significant differences were found in macroinvertebrate populations between improved and unimproved areas, but

A log dam on Comers Creek in the Mt. Rogers National Recreation Area (Smyth County), Virginia.

slight increases of sand and fine particulate matter were noted in the quiet waters associated with the impounded pools.

Tine single and nine paired gabions were present on Jennings Creek, but the evaluation of the physical effects of the gabions was confounded by the lack of data from before emplacement. Although pools were in close proximity to 11 of 18 gabions, it is not known whether the pools existed prior to or subsequent to the construction. The authors surmise that in the areas where pools began above the gabions, the structures were placed there to enhance and maintain the existing pools. Pools that start at the beginning of the paired gabions likely reflect pool formation subsequent to gabion placement. A number of the single gabions were placed on the outside bend of meanders to direct the stream away from the nearby roadbed. The pools near the paired gabions were sgenerally restricted to a width equal to the diameter of the opening between the gabions. The water behind both single and paired gabions was generally shallow (less than 30 cm) and deposition of sand, silt and detritus was common in this quiet water. No significant differences were found between the study sections in total fish or trout biomass or benthic invertebrate populations over three sampling periods. The general absence of cover associated with the gabions may explain the lack of effect.

The objective of single and multiple boulder placements in Crooked Creek was the development of additional cover through undercutting and pool formation. Five of the 12 placements did result in undercutting and significant depth increases. (Depth, itself, creates cover.) The boulders created some turbulence and tended to accumulate debris, both of which provided additional cover for trout. The depth before placement greatly influenced the effectiveness of the boulders. When the boulders were in the deeper portions of the channel, the current undercut the sides and occasionally the upstream face of the rocks. Conversely, little undercutting occurred when the boulders were placed in the shallower sections of the stream.

The high sand content in Crooked Creek resulted in extensive deposition in the areas of reduced current velocity behind the boulders. Deposition was reduced or eliminated when the boulders were situated close together. The deflection of the current by the boulders during periods of high streamflow presented a problem, in that erosion and damage was noticeable in some areas where large boulders were placed close to the banks. Although no statistically significant differences in total fish or trout biomass were found between the sections containing boulders and unimproved areas, the use of the boulders by trout was apparent during electro-shocking. Significant differences were likewise not found between the aquatic invertebrate populations or physical characteristics of the

stream in the study sections. In summary, while an increase in pool area resulted from the installation of log dams and gabions, no statistically significant differences were found in the fish biomass or benthic invertebrate numbers or kinds among habitat improved areas and unimproved sections of the four streams. However, total fish and trout biomass was more often higher in association with the log dam sections than either gabion or boulder emplacements. As the pools forms by the log dams and gabions were generally lacking in cover, the structures may not be functioning optimally. The increase in fishing pressure observed near the structures further complicates analysis.

he log dams and gabions studied were found to be susceptible to destruction by high flows; however, proper construction design and a low profile could have prolonged their lifespan. The primary study objective of these dams and gabions was to increase pool area, therefore, only in streams where the lack of pools is considered the limiting factor, should their use be considered. Analysis of the depth profiles on unaltered sections of North and Comers

Creeks showed only a very low percentage of deepwater habitat (less than three percent). Therefore, these streams would seem to be reasonable candidates for this type of structure. Boulder emplacements, when situated in deeper sections of the channel, caused undercutting and provided cover for fish. Boulders are therefore applicable in situations where cover is considered to be limiting. When boulders were placed in shallow areas, little undercutting occurred. Boulders can be used in larger streams than those suitable for gabions or dams, and are natural in appearance and can be assumed to function indefinitely. However, the use of boulders may be impractical when they are not available in proximity to the stream.