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



791 



hours per block, was compared to the mean 

 number of trawl marks in each reporting block 

 with a Spearman rank order correlation. Two 

 reporting blocks (129 and 218) were excluded 

 from this analysis owing to a lack of adequate 

 sidescan coverage. To estimate the total area 

 swept by bottom trawls on an annual basis. 

 the total annual number of fishing hours per 

 reporting block was multiplied by a typical 

 vessel speed (5.5 km/h) and a typical door to 

 door width of 85 m. 



Detrended correspondence analysis (DCA) 

 was used to identify clusters of similar report- 

 ing blocks in ordination space on the basis of 

 fish assemblage structure from CDF&G log- 

 book data (Ludwig and Reynolds, 1988). A 

 matrix of reporting blocks by catch was cre- 

 ated for use in this analysis. Habitat types 

 were defined a priori by depth and physical 

 characteristics of the habitat from previous 

 geological surveys and then overlaid on the 

 station clusters created by DCA. A Kruskal- 

 Wallis rank sum test (Hollander and Wolfe, 

 1973) was used to compare the density of trawl 

 marks and total fishing hours per reporting 

 block in each habitat type discerned by DCA. 

 Pairwise comparisons of trawl marks and fish- 

 ing effort between habitats was conducted by 

 using Dunn's multiple comparison procedure 

 at a = 0.05 (Hollander and Wolfe, 1973). 



Results 



We measured the range, density, and orienta- 

 tion of marks on the seafloor caused by trawl- 

 ing activity as resolved by the sidescan-sonar 

 records. Of the 1246 10-min intervals exam- 

 ined, -10% were washed out and could not be 

 analyzed. Densities of trawl marks/km- per 10- 

 min interval ranged from 0.00 to 98.55 and 

 mean density per block ranged from 0.94 to 38.3 with 

 a grand mean of 20.0 marks/km- (Table 1). Trawl 

 marks were most abundant on the slope, particularly 

 the northern portion and least abundant on the con- 

 tinental shelf (Fig. 5). 



Trawl marks commonly were oriented parallel to 

 isobaths and some could be traced for several km (Fig. 

 6). The overall mean trawl direction was 352.6° (95^7^ 

 CI -i-/-5°) (Fig. 7) which agrees with the general ori- 

 entation of isobaths in this area. Because trawl marks 

 were generally orientated in the north— south direc- 

 tion along isobaths, sidescan tracks orientated in the 

 east-west direction (>45° and <135°) might not dis- 

 tinguish as many trawl marks in relation to those 



Figure 3 



Sidescan-sonar image (white is shadow; dark is reflection ) of trawl 

 marks on the seafloor at a water depth of ~ 150 m. Sidescan swath 

 width is 750 m. Tick marks are .5-min time intervals (-0.55 km). 

 Arrows point to examples of trawl marks. Shallow circular depres- 

 sions on the sonograph are gas pockmarks. Ship's course for this 

 sonogram is 345". 



orientated in the north-south direction. Using 25 

 randomly selected time intervals from five east-west 

 sidescan track lines and 25 from five north-south 

 track lines, we found that the proportion of trawl 

 marks sampled with the perpendicular line intercept 

 method was not significantly different from the ac- 

 tual number of trawl marks for the two track line 

 orientations (Mann-Whitney rank sum test=759.5, 

 P=0.775). 



We compared the orientation of trawl mark angles 

 between east-west and north-south orientations by 

 selecting seven locations where east-west and north- 

 south sidescan-sonar track lines intersected during 

 the 1996 survey. The intersecting time interval, along 



