Butler et al Multibeam sonar mapping techniques for estimates of Haliolis sorenseni 



523 



remotely operated vehicle (ROV) equipped with a for- 

 ward-looking 12:1 zoom color video camera (all surveys) 

 and a high resolution Nikon Coolpix digital still camera 

 (2004 surveys only, Insite Pacific Inc., Solana Beach, 

 CA). We recorded video footage continuously on digital 

 tapes and overlaid footage with temperature, depth, 

 heading, and time information from an on-screen dis- 

 play. Two pairs of lasers were mounted near the front 

 of the ROV in parallel: one pair to measure abalone 

 upon sighting (10.0 cm apart), and one pair to estimate 

 the search field of view (60 cm apart). Upon sighting 

 an abalone, the ROV pilot maneuvered the vehicle into 

 position to zoom in on and photograph the animal for 

 species confirmation, ensuring that the narrow pair of 

 lasers passed over the abalone so that a size estimate 

 could be obtained. Empty abalone shells were also noted 

 and identified. 



The ROV was tracked by using a directional hydro- 

 phone mounted on the ship and a transponder mounted 

 on the ROV. Positional data and all other navigational 

 and physical data associated with the ROV dives (e.g., 

 heading, depth, and water temperature) were recorded 

 every two seconds. Because of the great depth range 

 that white abalone inhabit, it was not logistically fea- 

 sible to compare ROV transect observations with diver 

 transect observations. It is possible that we missed 

 abalone residing on the back sides of rocks as the ROV 

 passed over. 



Sampling design and site selection 



Multibeam and side-scan surveys were conducted the 

 preceding night at sites that were to be surveyed with 

 the ROV during the next day. Transect locations were 

 chosen randomly initially, but after preliminary surveys 

 at Tanner Bank revealed a narrow depth range of aba- 

 lone occurrence from 30 to 60 m and a habitat restriction 



to rock reef or the sand-and-rock reef interface, only 

 areas of suitable habitat (depths ranging from 30 to 60 m 

 [±5 m]) were surveyed. In general, we attempted to run 

 transects for approximately two hours, and to a length 

 of 1 km, although actual lengths were calculated after 

 the dive. Several of the early surveys included transects 

 of greater lengths of time and distances. Georeferenced 

 bathymetry maps were interfaced with the ROV naviga- 

 tional software for all dives, allowing for the ship, and 

 in effect the ROV, to remain within a particular depth 

 stratum and within suitable habitat. 



Postdive processing 



Videotapes and still photos were reviewed to confirm spe- 

 cies identification, measure sizes, search for additional 

 abalone, record search effort, examine microhabitat, and 

 determine the search field of view. Other abalone spe- 

 cies that were encountered were noted (red and pink). 

 We used strip transect techniques to estimate densities, 

 and therefore the field of view was measured for abalone 

 sightings from the first survey (Tanner Bank, 2002), 

 and the average (2 m [±0.85 SD], n=171) was used as 

 the strip width for the entirety of each transect in the 

 study. 



Data analysis 



For habitat classification a statistical approach called 

 textural analysis (described in Cochrane and Lafferty, 

 2002) was used to identify complex rock areas at Tanner 

 Bank as documented by side-scan sonar (Fig. 1, A and 

 B). We estimated the total area of available white aba- 

 lone habitat on each bank by calculating the amount 

 of rocky substrate in depth intervals, as determined 

 by multibeam sonar surveys. A combination of rugos- 

 ity and topographic position analysis (TPI) were used 



