Abstract
Binary fish passage models are considered by many fisheries managers to be the best
21 available practice for culvert inventory assessments and for fishway and barrier design.
22 Misunderstandings between different binary passage modeling approaches often arise,
23 however, due to differences in terminology, application and presentation. In this paper
24 one-dimensional binary fish passage models are reviewed and refined to clarify their
25 origins and applications. For uniform flow, a simple exhaustion-threshold (ET) model
26 equation is derived that predicts the flow speed threshold in a fishway or velocity barrier
27 that causes exhaustion at a given maximum distance of ascent. Flow speeds at or above
28 the threshold predict failure to pass (exclusion). Flow speeds below the threshold predict
29 passage. The binary ET model is therefore intuitive and easily applied to predict passage
30 or exclusion. It is also shown to be consistent with the distance-maximizing model. The
31 ET model’s limitation to uniform flow is addressed by deriving a passage model that
32 accounts for nonuniform flow conditions more commonly found in the field, including
33 backwater profiles and drawdown curves. Comparison of these models with
34 experimental observations of volitional passage for Gambusia affinis in uniform and
35 nonuniform flows indicates reasonable prediction of binary outcomes (passage or
36 exclusion) if the flow speed is not near the threshold flow velocity. More research is
37 needed on fish behavior, passage strategies under nonuniform flow regimes and
38 stochastic methods that account for individual differences in swimming performance at or
39 near the threshold flow speed. Future experiments should track and measure ground
40 speeds of ascending fish to test nonuniform flow passage strategies and to improve model
41 predictions. Stochastic models, such as Monte-Carlo techniques, that account for
42 different passage performance among individuals and allow prediction of the percentage
43 of fish passing would be particularly useful near flow speed thresholds where binary
44 passage models are clearly limited.