Fisheries managers, aquatic ecologists, and others interested in studying the habitat quality and availability for fish species in freshwater systems often face an uphill battle. Unlike our esteemed colleagues in the field of terrestrial ecology, we usually cannot rely on widely available satellite imagery and high resolution topographic maps to detail the physical features found in our study organisms’ environment. The information provided in these sources often stops at the river’s edge, and thus the physical habitat in the underwater environment is often left shrouded in mystery.

Imaging equipment that can peer through the water column and shed insight on the physical features in fish habitat is not a new phenomenon. Sonar equipment, which paints a picture of the riverbed by interpreting rebounding soundwaves, has been available for decades. In the recent past, however, this equipment has been prohibitively expensive for many researchers. In the first decade of the 21^st century, manufacturers of fish-finder equipment marketed towards recreational anglers began including side-scan sonar capacity with their top of the line models. This emerging trend prompted researchers to develop ways to utilize this much more affordable technology to augment their studies of aquatic habitat.

The Fisheries Lab at the University of Tennessee, under the leadership of Dr. Brian Alford, has been intimately involved in the reintroduction of the once-extirpated Lake Sturgeon Acipenser fulvescens to the Tennessee River. One of the research needs outlined by the managers overseeing the reintroduction was to assess the availability of habitat critical to supporting the various life history stages of this species. These fish exhibit migrations during the spring spawning season from their lentic habitats in the reservoirs upstream to more lotic habitats where they spawn over clean, coarse rocky substrate with plenty of interstitial spaces. However, the Tennessee River is punctuated by large hydroelectric dams, which are effectively terminals for spawning migrations. Luckily, this species has been found to successfully spawn in the rocky tailwaters below dams in other river systems. Operating under the assumption that as the restoration of the Lake Sturgeon to the Tennessee River continues, these fish will begin forming spawning aggregations below some of the hydroelectric dams along the river, we are using consumer-grade, side scan sonar equipment and GIS to map the substrate in these tailwaters to document the availability of the substrate preferred by these fish for successful spawning.

Using protocols developed by Dr. Adam Kaeser and Thomas Litts (to whom we are very grateful) we collected sonar imagery of the tailwaters below four hydroelectric dams on the Upper Tennessee River with a Humminbird © 1199 ci HD unit.

To ensure that no propeller wash interfered with our sonar imagery, we had a unique aluminum bow-mount manufactured, which allows us to mount the sonar transducer to the bow of the boat, deploying it underwater while we collect data and swinging it up out of the water when we need to run the boat at higher speeds.

After we collected sonar imagery of the substrate in the tailwaters below four hydroelectric dams on the Tennessee River, we randomly selected locations within the tailwaters to revisit. At each of the points selected, we collected real video imagery of the substrate with our underwater video camera system. We will use this information to ground-truth the sonar imagery after processing and interpretation.

After we collected the sonar imagery of the substrate below the dams, we georeferenced the images into spatially-explicit raster layers. With this information, we will be utilizing several methods of image classification to assess the amount of suitable spawning substrate available to Lake Sturgeon in these tailwaters. We will then utilize the ground-truthing imagery to assess the accuracy of the various methods for classifying the substrate, allowing us to evaluate their effectiveness for future aquatic habitat and substrate assessment projects. Already we have noticed that the protocols we used to collect the imagery data have produced more data in less time than other methods of substrate characterization and habitat assessment at this scale. We expect to develop accurate procedures for classifying the sonar imagery collected in this method, and look forward to future applications for these tools and procedures.

Daniel J. Walker^1,2, Todd Amacker¹, J. Brian Alford¹

1 – Department of Forestry, Wildlife, and Fisheries. University of Tennessee –Knoxville. 2431 Joe Johnson Drive, Knoxville, TN 37996

2 – Corresponding author. Contact: dwalke44@vols.utk.edu