Accuracy and Precision of Measurements of Transect Length and Width Made with a Remotely Operated Vehicle

Abstract

Remotely operated vehicles (ROVs) have been used to estimate the density of fish and invertebrates on strip transects. However, there is little published information about the accuracy and precision of measurements of transect length and width, a critical component of the density estimate. In this study, we evaluate straight line constant velocity protocols, the accuracy and precision of estimates of transect length from ultra short baseline acoustic tracking, and compare measurements of transect width based on sonar and lasers. When ROV tracking was compared to distances measured on sonar maps, the difference between linear tracked and mapped distance (|LDt-LDm|) averaged 1.7 ± 0.5 m. The error was not significantly different with distance. Our navigation protocols allowed us to maintain relatively constant heading and speed. Distance computed from velocity exceeded mapped distance by 2-4 m. The error was not significantly different with distance. Measurements of transect width made with lasers and sonar were comparable, particularly when the ROV was within 4 m of the substrate. Based on our data, ROV tracking can be used to measure transect length within 2 m. If tracking fails, distance can be estimated from velocity within 2-4 m. Sonar can be used to measure transect width with considerable cost savings.