Stream crossings using culverts can present fish passage problems. Firstly, some species of fish won’t go through small dark pipes, and if the outlet of the culvert is above the level of the stream, while fish travelling downstream may get through, fish travelling upstream won’t make the upward leap into the culvert. They will get "stuck" on the downstream side of the road crossing.

Secondly, culverts are likely to concentrate stream flow and increase stream velocity by reducing the effective cross-section of the stream. Some species of fish will find it difficult to swim through the culvert against the velocity of the on-coming water. This effect is exacerbated by the smooth walls of the culvert, which produce laminar flow with a minimal boundary layer. In a normal stream environment, turbulent flow produces a large boundary layer, and fish use the lower velocity in this boundary layer to swim upstream.

The solution is to use large culverts rather than small ones. The culverts should be positioned marginally BELOW the bed of the stream (see attached Figure). Over time, a lining of river pebbles and other debris will accumulate along the bottom of the pipe. This will create a passage which not only appears more familiar to fish trying to move upstream, but (more importantly) has a rough bottom, a turbulent flow, and a large and negotiable boundary layer. In designing the culvert size, it will be necessary for the design engineer to allow an increase in diameter over the minimum pipe capacity chosen to handle the design rainfall event. It is recognised that this will result in a small increase in the capital cost of the stream crossing.