USING LOW-COST SIDE-SCAN SONAR FOR BENTHIC MAPPING THROUGHOUT THE LOWER FLINT RIVER, GEORGIA, USA
Article first published online: 10 JAN 2012
Copyright © 2012 John Wiley & Sons, Ltd.
River Research and Applications
How to Cite
KAESER, A. J., LITTS, T. L. and TRACY, T. W. (2012), USING LOW-COST SIDE-SCAN SONAR FOR BENTHIC MAPPING THROUGHOUT THE LOWER FLINT RIVER, GEORGIA, USA. River Res. Applic.. doi: 10.1002/rra.2556
- Article first published online: 10 JAN 2012
- Manuscript Accepted: 6 DEC 2011
- Manuscript Revised: 19 OCT 2011
- Manuscript Received: 6 MAY 2011
- habitat mapping;
- side-scan sonar;
- remote sensing;
An efficient, low-cost approach for mapping habitat features in navigable streams is needed to support the research and management of aquatic ecosystems at the landscape level. We developed a method that uses high-resolution (455 kHz) side-scan sonar imagery obtained with the inexpensive (~$2000) Humminbird® Side Imaging system and ArcGIS to produce sonar image maps (SIMs) used to interpret and map habitat features such as substrates and large woody debris, in addition to continuously recording depth along the survey route. This method was recently demonstrated and evaluated in several small streams in southwestern Georgia (30–50 m width, 40 km mapped). To evaluate the feasibility of this method for mapping substrate and depth in larger rivers and over greater spatial extents, we conducted a sonar survey and generated SIMs for 124 km of the lower Flint River (85–140 m width). We interpreted the SIMs to digitize and classify substrate and bank boundaries. To assess classification accuracy, we visually inspected substrate at randomly assigned reference locations. A comparison of reference and map data revealed an overall classification accuracy of 84%. These results were consistent with previous findings and indicate that low-cost side-scan sonar is also an effective mapping tool for larger rivers. The sonar survey did, however, result in more missing and unsure substrate data and a lower map accuracy for fine-textured substrates than previously achieved when mapping smaller streams. We found a strong, positive relationship (r2 = 0.89) between the sonar range and the proportion of unsure substrate in the map, suggesting that a multi-pass, parallel-transect sonar survey could be used to maintain high-image resolution when stream widths exceed 100 m and/or obstructions, such as islands, are encountered. Applications for sonar-based habitat maps are widespread and numerous. The ability to produce these maps efficiently at low-cost is within the grasp of researchers and managers alike. Copyright © 2012 John Wiley & Sons, Ltd.