Automated Extraction of Dune Crests from Multibeam Sonar Data to Calculate Transport Vectors of Endangered Pallid Sturgeon Larvae on the Lower Missouri River

Engineering of the Lower Missouri River for navigation has resulted in strong downstream advection at the expense of more-complex flows that existed in the pre-channelized river. Channelization has also resulted in substantial loss of low-velocity areas adjacent to the main channel; these areas are thought to be important habitats for the survival of endangered Pallid Sturgeon larvae. Re-engineering actions stemming from efforts to recover the Pallid Sturgeon seek to 1) increase complexity in the flow field to increase interception of drifting larvae into channel-marginal habitats, and 2) to provide food and bioenergetically favorable foraging habitats. Designs for re-engineering require better understanding of the interactions between drifting larvae and channel hydraulics. Interception of drifting larvae may be driven by near-bed velocities, processes that are difficult to measure with hydroacoustic instruments. We present an alternative method for investigating the influence of near-bed velocities on larval interception by analyzing sand dune morphology in a reach of the Lower Missouri River that has been shown to intercept significant numbers of sturgeon larvae. Specifically, we demonstrate an algorithm to extract dune crests from high-resolution multibeam sonar data collected throughout the spawning and dispersal season. The extracted dune crests are then used to define near-bed transport vectors that would otherwise be difficult to observe from hydraulic measurements. We conclude by using the calculated transport vectors and extracted dune morphology features to evaluate the role of near-bed velocities in driving the interception of Pallid Sturgeon larvae.