Conservation Evidence strives to be as useful to conservationists as possible. Please take our survey to help the team improve our resource.

Providing evidence to improve practice

Individual study: Use of bank-cover deflectors to improve brook trout Salvelinus frontalis habitat in Lawrence Creek, Wisconsin, USA

Published source details

Champoux O., Biron P. M. & Roy A. G. (2003) The long-term effectiveness of fish habitat restoration practices: Lawrence Creek, Wisconsin. Annals of the Association of American Geographers, 93, 42-54


Along Lawrence Creek in Wisconsin, central USA, cattle visiting the creek to graze bankside vegetation and drink, were eroding the banks and increasing the sediment load of the channel. This resulted in the loss of resting cover and permanent hiding areas for brook trout Salvelinus fontinalis and also sedimentation of in-channel pools. These habitat features, amongst others, are considered vital in supporting a healthy brook trout population. In the early 1960s, the Wisconsin Department of Natural Resources (WDNR) identified that these habitat features had declined to such an extent that the creek could no longer support brook trout. As salmonid species are important game fish, channel restoration and management is often implemented in such cases in an attempt to improve fish habitat thus hopefully re-instating fish populations. The restoration of part of Lawrence Creek is summarised.

Study area: Lawrence Creek is mainly groundwater fed and drains a catchment of 34.7km². The creek has an annual discharge of approximately 0.5 m²/s which remains fairly constant throughout the year. Channel restoration was undertaken along a 600 m length of the creek. This stretch flows through two types of deposit: terminal moraine composed of heterogeneous glacial material (D50=30 mm) in the upstream section (0-240 m); and sand and fine sediment (D50=1 mm) in the downstream outwash section (240-600 m). (Note: D50 is the median size of the sediment - this is used rather than the mean, so that a boulder on the gravel bed will not disproportionately affect the sediment size measurement).

The study was conducted in 1999 and the results were compared with those of Hunt (1969).

Bank-cover deflectors: In 1963, the WDNR installed bank-cover deflectors along a 1.6 km stretch to narrow the channel and provided artificial underbank refuge for fish. Cover-bank deflectors consist of rock and heavy timbers imbedded into the bank and are generally placed 5-15 m apart (Trimble 1997). They narrow the channel but also provide excellent underbank cover for fish.

It was hoped that by narrowing the channel, flow velocity would increase causing scouring of the bed, thus creating new in-channel pools and deepening existing ones. The deflectors were placed on the outside of bends where there was a long, clear inflection in the bank. Along the morainic section studied, only six deflectors were placed, covering about 40% (96 m) of the bank length. Bank-cover deflectors were placed along the entire length of the outwash section (about 20 meanders) due to its greater sinuosity and hence ameniability, to deflector installation.

Installation of other in-channel features: In addition, boulders and woody debris were placed on the bed of the creek. The aim of this was to cause variable local flow and provided areas of low flow directly downstream of these obstacles as fish rest-up areas.

Overall effects of deflectors on habitat quality: The installation of bank-cover deflectors had a long-term beneficial impact on fish habitat. Habitat quality in 1999 was better than in 1963 (but had deteriorated compared to 1966). Most habitat variables deemed beneficial to brook trout had already improved significantly by 1966 compared to 1963 (Table 1, attached). By 1999, the area of the bed occupied by pools and cover had increased markedly compared to 1963, achieving a significant improvement in fish habitat quality.

Fish habitat deteriorated between 1966 and 1999. There was a marked decrease in cover and a decrease in mean water depth strongly associated with a decrease in pools. However, the area of gravel bed increased from 1966 to 1999. Gravel substrate is important as it provides a spawning area for brook trout.

Narrowing of the channel increased bed shear stress, increasing bed scouring and thus improving the bed habitat but also resulting in degradation of stream banks. Bank-cover deflectors that worked well in the outwash section were poor in the morainic section, exhibeted by greater deterioration - in some areas of high bed shear stress the bank-cover deflectors collapsed.

Morainic section: Bed shear stress in 1963 was highest in the top 180 m of the section due to the gradient of the channel being steepest here. Bank-cover deflectors increased bed shear stress and increased erosion. This caused the deflectors to become undermined and damaged, and by 1999 most were either impaired or had failed. Increased bed and bank erosion also resulted in an increase in the sediment load and aggregation in the outwash section, causing the fragmentation of larger pools. Between 1963 and 1966, the deflectors effectively reduced the channel width and increased bed scouring and subsequently, there was an increase in the number of pools. By 1999, the channel width had increased markedly, the creek became multi-channelled and the physical habitat deteriorated to such an extent that by 1999 it had reverted back to characteristics similar to those of 1963 before the deflectors were installed. Importantly, the number of pools and the area of the bed covered by pools had decreased.

Outwash section: In 1999 several deflectors were still in good condition. The deflectors narrowed the channel effectively and the number and area of pools increased significantly from 1963 to 1999, greatly improving the habitat for fish. From 1966 to 1999, the area of bed occupied by pools remained fairly constant but their size and number changed significantly due to the fragmentation of larger pools. Pool location also shifted. This was caused by aggregation of sediment eroded from the upstream morainic section.

Conclusions: Along the outwash section of Lawrence Creek, bank-cover deflectors were used successfully to improve fish habitat in the long-term. Additionally, an earlier study (Hunt 1971) found that after the installation of bank-cover deflectors, the mean biomass of age 0+ brook trout was 41% greater than that recorded before the bank-cover deflectors were installed.

Long-term habitat improvement in the upstream moranic section was not acheived. It is likely that the coarser moranci substrate was more easily eroded/washed out than that in the outwash substrate and that this, coupled with the greater bed shear stress, particularly in the top 180 m, led to fairly rapid failure of the deflectors.

As well as habitat features, longer-term data on the brook trout population response to bank deflector installation would be desirable but as is the case for many river restoration projects such evaluations are rarley undertaken.

Hunt R. L. (1969) Effects of habitat alteration on production standing crops and yield of brook trout in Lawrence Creek, Wisconsin. In Symposium on Salmon and Trout in Streams, ed. Northeote T. G. 281-312. H. R. MacMillan Lectures in Fisheries. University of British Columbia, Vancouver, Canada.

Hunt R. L. (1971) Response of a brook trout population to habitat development in Lawrence Creek. Technical Bulletin no.48. Madison, WI: Department of Natural Resources.

Trimble S.W. (1997) Streambank fish-shelter structures help stabilize tributary streams in Wisconsin. Environmental Geology, 32, 230-234.

Note: If using or referring to this published study, please read and quote the original paper. Please do not quote as a case as this is for previously unpublished work only.