Accurate, precise, and efficient underwater visual censuses (UVc) are vital for detecting and monitoring range shifts of reef fishes. The present study compared the utility of time-equivalent belt transects and Global Positioning system (GPs)-tracked roaming surveys for assessing populations of range-expanding tropical fishes off southeastern Australia. roaming surveys were significantly more accurate and precise than belt transects in estimating densities of focal damselfish, Abudefduf vaigiensis (Quoy and Gaimard, 1825), while the accuracy and precision of density estimates did not significantly differ between methods for the rarer focal species, Abudefduf whitleyi (Allen and robertson, 1974). significantly greater species richness and assemblage diversity were detected by roaming surveys than belt transects. roaming surveys were also over twice as efficient, defined as the area searched per unit time, as belt transects; mean efficiency of roaming surveys and belt transects was 33.56 (sE 1.22) and 12.57 (sE 0.66) m2 min-1, respectively. results were consistent among observers with varied experience. reliable density estimates, improved efficiency, and maximized sightings of tropical fishes suggest GPs-tracked roaming surveys are highly suited for detecting and monitoring range shifts of reef fishes.