2) Do Culvert Structures Impede Fish Passage?


In further analysis, stream crossing structures were classified into three different crossing categories; bridges, culverts with no hang height, and culverts with a hang height. Since our data satisfied the assumptions of normality and equal variances (Appendix A), for both weak and strong swimming fish species, a multifactor ANOVA was conducted. This was followed by pairwise comparisons for each of the crossing categories.

Weak Swimmers

Fig. 5. Catch per unit effort (#fish/s) difference (downstream-upstream) of weak swimming fish species belonging to  families Gadidae (burbot), Cyprinidae (northern red belly dace, pearl dace, lake chub, long nose dace), Gasterosteidae (brook stickleback), Catostomidae (longnose sucker, white sucker), Cottidae (spoonhead sculpin), Percopsidiae (trout perch) in three different crossing structure categories, bridges, culverts with no hang height and culvert with a hang height. A positive value indicates that at a site more fish of a particular species were found downstream of a stream crossing, while a positive value indicates higher presence of a species upstream of a stream crossing.  Different letters indicate a significant (p<0.05) differences between the three crossing types.

Fig.6. Median atch per unit effort (#fish/s) values of weak swimming fish species belonging to  families Gadidae (burbot), Cyprinidae (northern red belly dace, pearl dace, lake chub, long nose dace), Gasterosteidae (brook stickleback), Catostomidae (longnose sucker, white sucker), Cottidae (spoonhead sculpin), Percopsidiae (trout perch) downstream and upstream in three different crossing structure categories, bridges, culverts with no hang height and culvert with a hang height.. Different letters indicate a significant (p<0.05) difference between upstream and downstream catch rates.


When comparing catch rates of the three different crossing categories, there is only a significant difference between bridges and culverts with a hang height (p=0.02). There are higher downstream catch rates of weak swimming species at culverts with a hang height than at bridges. In addition, downstream and upstream comparisons were made at each of the three crossing categories. There was no significant difference in weak swimming fish presence above and below culverts with no hang height (p=0.345) and culverts with a hang height (p=0.384). There was, however, a difference at our control bridges (p=0.012). It should be noted, however, that there were several outlier catch rates that may have skewed this result. In the multifactor ANOVA analysis there was no significant interaction effects between the three different crossing categories and upstream/downstream comparisons.

Strong Swimmers

Fig. 7. Catch per unit effort (#fish/s) difference (downstream-upstream) of strong swimming fish species belonging to the family Salmonidae (rainbow trout, arctic grayling, mountain whitefish) in three different crossing structure categories, bridges, culverts with no hang height and culvert with a hang height. A positive value indicates that at a site more fish of a particular species were found downstream of a stream crossing, while a positive value indicates higher presence of a species upstream of a stream crossing.  Different letters indicate a significant (p<0.05) difference.

Fig. 8. Median catch per unit effort (#fish/s) of strong swimming fish species belonging to the family Salmonidae (rainbow trout, arctic grayling, mountain whitefish) upstream and downstream in three different crossing structure categories, bridges, culverts with no hang height and culvert with a hang height.Different letters indicate a significant (p<0.05) difference. 


There was no statistical difference between the catch rates of strong swimming fish species at the three different crossing types (p=0.96). However, when comparing downstream and upstream catch rates at each of the individual crossing types, there was a significant difference at culverts with a hang height (p=0.037). Salmonids were more likely to be found upstream of a culvert with a hang height. There was no significant difference in catch rates at control bridges (p=0.082) and culverts with no hang height (p=0.512). In the multifactor ANOVA analysis there was no significant interaction effects between the three different crossing categories and upstream/downstream comparisons.