Abstract. Demersal fish surveys 



are used for two purposes: to detect 

 trends in multispecies communities for 

 environmental assessment and to pro- 

 vide fishery independent stock assess- 

 ments for management. We compared 

 remotely operated vehicle (ROV) and 

 swept-area trawl surveys to evaluate 

 their strengths and weaknesses for 

 these two purposes. ROV abundance 

 estimates tended to be higher and have 

 lower coefficients of variation than did 

 trawl abundance estimates. This trend 

 is greatest for benthic species and par- 

 ticularly so for small, cylindrically 

 shaped fishes. For patchily distributed, 

 off-bottom fishes such as rockfish, 

 Sebastes spp., sablefish, Anoplopoma 

 fimbria, and Pacific whiting, Merluccius 

 productus, the results vary between 

 ROV and trawl estimates. For environ- 

 mental assessment, the ROV estimates 

 are superior because, for most species, 

 abundances are higher and smaller 

 changes can be detected. For fisheries 

 management of commercially impor- 

 tant species, the results are divided. 

 Dover sole, Microstomias pacificus, and 

 thornyheads, Sebastolobus spp., have 

 higher ROV abundance estimates and 

 lower coefficients of variation than the 

 trawl. Sablefish, which exhibit more off- 

 bottom behavior, have higher trawl es- 

 timates at two of three depths. The 

 ROV and trawl provide different types 

 of information not available from other 

 gear. Much of the difference between 

 the two types of surveys stems from the 

 nature of the sampling gear and from 

 the behavior, body shape, and size of 

 the fishes. 



Population estimates of Pacific coast 

 groundfishes from video transects 

 and swept-area trawls 



Peter B. Adams* 

 John L. Butler** 

 Charles H. Baxter*** 

 Thomas E. Laidig* 

 Katherine A. Dahlin** 

 W. Waldo Wakefield**** 



* Tiburon Laboratory, Southwest Fisheries Science Center 

 National Marine Fisheries Service, NOAA 

 3 I 50 Paradise Drive, Tiburon, California 94920 



** Southwest Fisheries Science Center 

 National Marine Fisheries Service, NOAA 

 PO. Box 271, La Jolla, California 92038 



*** Monterey Bay Aquarium Research Institute 



1 60 Central Avenue, Pacific Grove, California 93950 



**** National Undersea Research Center 



School of Fisheries and Oceans Sciences 



University of Alaska Fairbanks, Fairbanks, Alaska 99775 

 Present address: Institute of Marine and Coastal Science 

 Rutgers University, New Brunswick, New Jersey 08093 



Manuscript accepted 13 February 1995. 

 Fishery Bulletin 93:446-455 (1995). 



Estimates of population abundance 

 are essential to research for under- 

 standing the impact of human ac- 

 tivities on marine demersal fish 

 populations. Traditionally, swept- 

 area trawl surveys have been used 

 to obtain abundance estimates 

 aimed at fisheries management. 

 Recently, environmental surveys, 

 used to monitor the impacts of pol- 

 lution and coastal development, 

 have become more widespread and 

 important. These two types of sur- 

 veys have different goals. Data from 

 fisheries surveys are used as input 

 for predictive models to forecast the 

 results of alternative fisheries man- 

 agement strategies and are usually 

 directed toward either a single spe- 

 cies or a small species group. Envi- 

 ronmental surveys are used to de- 

 tect trends in populations over time, 

 to distinguish those trends from 

 natural variation, and are usually 



directed at an entire multispecies 

 fish community. 



Both types of survey population 

 estimates are subject to the prob- 

 lems of bias and precision. Bias is a 

 particular problem for fisheries sur- 

 veys. Fishery stock assessment is 

 based on models that integrate fish- 

 ery catch-at-age data with fishery- 

 independent survey estimates of 

 abundance (Deriso et al., 1985). The 

 catch-at-age data, sampled from the 

 commercial fishery, document the 

 trend of population change result- 

 ing from recruitment of young fish 

 into the population and from re- 

 moval of individuals out of the 

 population due to fishing and natu- 

 ral mortality. The fishery-indepen- 

 dent survey data are used as a mea- 

 sure of either relative or absolute 

 abundance (Doubleday and Rivard, 

 1981). These survey data are used 

 to calibrate or "tune" the trend ob- 



446 



