Nearly 200 fish were released below Lock and Dam 2 (LD2) in the Upper Mississippi River and tracked to determine both whether and how they passed through this structure, and if passage could be explained using a computational fish passage model (FPM) which combines hydraulics with fish swimming performance. Fish were either captured and released downstream of LD2 in Pool 3 or captured in Pool 2 (upstream of LD2) and displaced below LD2. Tagged fish were tracked using 13 archival receivers located across LD2. Approximately 90% of all fish approached LD2 many times with the displaced species likely attempting to home. Of 112 common carp, 26% passed through LD2 with 15% (most) going through the lock and 6% through the spillway gates. Similar values were seen for bigmouth buffalo. In contrast, although 42% of 31 channel catfish passed through the lock, only 3% went through the gates. Finally, of 22 walleye, only 14% passed through the lock and none through the gates. Ninety percent of all documented passages through the spillway gates occurred when the gates were out of the water and water velocities through these gates were at their lowest levels, an attribute described and predicted by the FPM at LD2. This study strongly suggests that fish passage through spillway gates of LDs is determined by water velocity and can be predicted with a FPM, whereas passage through locks is determined by species‐specific behavioural preferences. Both attributes could be exploited to reduce passage of invasive carp at certain locations.

No abstract

The distribution of river fishes, including invasive species, is influenced by their abilities and tendencies to pass regulatory structures, which in the Upper Mississippi River (UMR) are locks and dams (LDs). However, how and why fish approach and then pass through individual LDs and their structural components is presently unknown. In this study, groups of 25 common carp, a well‐understood model species, were acoustically tagged, transplanted, and released every other week for two summers below a generic UMR LD which has both tainter and roller spillway gates as well as a lock. Common carp distribution, passage routes, and rates were then monitored and analyzed with respect to discharge and gate settings. Although nearly all carp were detected below this LD, weekly passage rates through its spillway gates were negligible until discharge exceeded 85,000 cfs when the gates were ~2/3 open and weekly passage rates increased to ~50% for its tainter gates and ~8% for its rollers. In contrast, carp passed through its lock at a relatively constant and modest rate (~6%), seemingly uninfluenced by discharge. This increase in spillway gate passage, which occurred at discharges shy of open‐river, was predicted by a fish passage model which estimated spillway passage using swimming performance and estimated water velocity (flow‐fields). When observed passage routes were incorporated into another discharge‐based fish passage model, it accurately predicted total seasonal passage. These models could be improved and applied to determine which fish species will pass particular LDs and guide efforts to increase or decrease these rates.