Fnedlander et al.: Sidescan-sonar mapping of benthic trawl marks off Eureka, California 



789 



activities and to compare these results with fishing- 

 effort data from commercial fishery logbook data. 



Materials and methods 



Geological setting 



The study area is the upper continental margin of 

 northern California, north of Cape Mendocino. The 

 shelf and slope in this region are within the Eel River 

 Basin, a forearc basin filled with sediment derived 

 from high sediment loads during the late Neogene 

 (Clarke, 1992). The entire basin is highly faulted and 

 folded from multiple episodes of deformation. Shear- 

 ing and compression of the margin related to under- 

 thrusting of the Gorda plate beneath the North 

 American plate has resulted in thrust faulting, up- 

 lift, and an overall youthful geologic setting, and this 

 has produced varied relief and structure of the sea- 

 floor (Carver, 1987; Clarke and Field, 1989). 



Studies of seafioor stiaicture and sediment history of 

 this area (Field et al., 1999; Gardner et al., 1999; Goff 

 et al., 1999) have identified many features related to 

 submarine landslides and to downslope sediment trans- 

 poi't through channels and gulleys. Sidescan-sonar 

 images of the seafloor along the shelf and slope, as well 

 as multibeam maps, high-resolution seismic I'efiection 

 data, and instrumented tripods from complementaiy 

 investigations, all show abundant evidence of modern 

 processes at work on the seafloor (Field and Jennings, 

 1987; Alexander and Simoneau, 1999; Field et al., 1999; 

 Gardner etal., 1999; Goff etal., 1999;Yunet al., 1999). 



There are three main types of relief on the sea- 

 floor in this region. Along the slope, in the center of 

 the study area, a large growth fold (anticline) has 

 uplifted through the modern seafloor, exhibiting 

 100 m of steep relief. South of the anticline, the slope 

 between 120 and 600 m depths is dominated by the 

 Humboldt slide, an area where slope deposits have 

 undergone deformation and limited downslope move- 

 ment (Field and Barber, 1993; Gardner et al., 1999). 

 North of the anticline, the slope is crossed by a 

 series of shallow gullies that are 100 to 200 m wide 

 and 1 to 2 m deep (Field et al., 1999). Gullies are spaced 

 about 200 to 300 m apart. Gullies also occur south of 

 the anticline where they are generally fewer in num- 

 ber and somewhat deeper (-20 m). The gidlies are 

 thought to represent sediment pathways during ear- 

 lier periods when sea level was at a lower position. 



Sidescan survey 



- Through a collaborative effort of the U.S. Geological 

 Survey (USGS) and the National Marine Fisheries 



Service (NMFS), we examined sidescan-sonar data 

 off Eureka, California, collected by the USGS during 

 the STRATA FORMation on Margins (STRATAFORM) 

 progi-am sponsored by the Office of Naval Research 

 (Nittrouer and Ki-avitz, 1996; Nittrouer, 1999). A grid 

 of 1600 line kilometers of high-resolution, digital, 

 deep-towed (95 kHz) sidescan-sonar data was ob- 

 tained during 1995 and 1996 (Fig. 2). The sidescan- 

 sonar images were obtained by using the USGS 

 Datasonics SIS- 1000 chirp sonar system and an ISIS 

 data-acquisition system. The sidescan sonar operates 

 as a swept FM signal with a frequency band from 90 

 to 110 kHz (port channel sweeps from low to high, 

 starboard channel sweeps high to low). Usual swath 

 width during the cruises was 0.5 s (750-m swath). 

 Sidescan data were recorded on magneto-optical 

 disks and downloaded to a DEC ALPHA workstation. 

 The digital sidescan-sonar data were all processed 

 with USGS MIPS image-processing software 

 (Chavez, 1984; 1986). The raw data were geometri- 

 cally corrected (slant range) by using the position, 

 heading, speed, and nadir depth of the ship. Each of 

 the 8-bit image pixels was geometrically located by 

 assuming a flat seafloor which introduces a certain 

 amount of uncertainty into the actual location of any 

 given pixel. Our estimated location accuracy was 

 about 100 m. After the geometric corrections were 

 made, anamorphic corrections were applied to ac- 

 count for the aspect ratio between along- and across- 

 track directions. Radiometric corrections had to be 

 applied to the 8-bit digital number (DN) that repre- 

 sents backscatter. Each sidescan line was subdivided 

 into segments and, for each segment, the average 

 DN for each binned range was normalized to the 

 mean of all the binned ranges within that segment. 

 Each segment was several hours long, and this cor- 

 rection produced tone-matched adjacent segments. 



Trawl marks 



Ti'awl marks appear on the sidescan sonographs as 

 long, narrow, linear depressions (Fig. 3 ). Trawl marks 

 were traced on mylar overlaid on the sidescan 

 records. Each sidescan track line was divided into 

 10-min time intervals. Vessel speeds during the sur- 

 veys varied between 3 and 4 knots, and therefore 

 each 10-minute interval covered from 0.9 to 1.2 km 

 along the ship's course. The sidescan record covered 

 a 750-m swath so that 10-min segments represent 

 from 675,000 m- to 900,000 m- of area. To quantify 

 the number of trawl marks per time interval, lines 

 were drawn perpendicular to the course of the side- 

 scan track line, and trawl marks that bisected these 

 lines were counted. Mean trawl mark length was 

 0.77 km (SD=0.26) for 20 randomly selected 10-min 



