Mullin and Fulling: Abundance of cetaceans in the southern Atlantic Ocean 



611 



and are assumed to have migrated along-shore from the 

 south (Mead, 1975: Kenney, 1990). Aerial surveys of bottle- 

 nose dolphins conducted in the past along the U.S. Atlantic 

 included waters from the shore to 10 m in depth. For waters 

 typically <75 m deep south of Cape Hatteras, the winter 

 1992 abundance from an aerial survey was 12,435 (0.18) 

 (Blaylock and Hoggard, 1994). For waters <25 m deep from 

 Cape Hatteras to northern New Jersey, the abundance from 

 a summer 1994 aerial survey was 26,809 (0.40) (Blaylock, 

 1995). The frequency of bottlenose dolphin sightings during 

 these surveys increased substantially inshore of the 10-m 

 isobath boundary of the ship study area and, compared 

 with the estimate from the ship, may account for the gen- 

 erally larger aerial survey estimates even though they are 

 for smaller study areas. 



There are currently two genetically distinguishable bot- 

 tlenose dolphin stocks designated in the U.S. Atlantic: the 

 coastal stock and the offshore stock (LeDuc and Curry, 1998; 

 Waring et al., 2001). Using mitochondrial DNA from skin 

 biopsy samples obtained during the summer 1998 study 

 and other sampling efforts, Torres et al. (in press) reported 

 no offshore form was sampled within 6 km of shore and no 

 coastal from was sampled beyond 39 km from shore or in 

 waters >34 m deep. Therefore an area of overlap of the two 

 forms occurs within the 1998 study area but the fraction 

 of each stock in our estimate (13,085: 0.40) is unknown 

 because the number of biopsy samples between the two 

 boundaries was very small in the Torres et al. (in press) 

 study. However, 20 of the 38 bottlenose groups we used to 

 estimate abundance were found in waters >50 m deep. 



Abundances were estimated for ten species and three 

 other genera of cetaceans, but other species are known or 

 expected to occur in the study area. Three of these species, 

 right whales iEubalaena glacialis), humpback whales 

 (Megaptera novaeangUae), and harbor porpoises (Phocoena 

 phocoena), occur in the study area seasonally, primarily 

 in months other than summer months, and abundances 

 have been estimated from studies of their primary summer 

 ranges north of the study area (e.g. Knowlton et al., 1994; 

 Palka, 1995; Smith et al., 1999). 



Additional species expected in at least part of the study 

 area include Bryde's whale (B. edeni), Cuvier's beaked 

 whale (Ziphius cauirostris), pygmy killer whale (Feresa at- 

 tenuata), false killer whale (Pseudorca crassidens), melon- 

 headed whale iPeponocephala electro), killer whale (Orci- 

 nus orca), common dolphin (Delphinus delphis), spinner 

 dolphin, and Fraser's dolphin (Lagenodelphis hosei). Each 

 of these species is thought to have a tropical to subtropical 

 or broader distribution worldwide (Jefferson et al., 1993), 

 and except for the common dolphin, an abundance estimate 

 for each species is available for the adjacent northern Gulf 

 of Mexico (Hansen et al.^). However, except for the spin- 

 ner dolphin, each of these species is relatively uncommon 

 in the northern Gulf of Mexico and was not encountered 

 every year during four annual spring surveys with effort 

 similar to that in our survey (Hansen et al.''). Therefore, 

 many of these species may also be uncommon in the At- 

 lantic study area and were simply not encountered during 

 the 1998 survey. During a late summer 1999 ship survey 

 of the inner half of the southern Atlantic study area that 



targeted bottlenose dolphins, a group of Fraser's dolphins 

 and melon-headed whales was sighted in water 3000 m 

 deep east of Cape Hatteras (Roden^). 



Some species may also inhabit the study area seasonally. 

 During the 1992 winter ship survey south of Cape Hatteras 

 (Mullin and Ford'), five groups of balaenopterid whales 

 were recorded, three of which were classified as uniden- 

 tified Bryde's or sei whales. Also during the winter 1992 

 survey, groups of false killer whales and Cuvier's beaked 

 whales were sighted twice, and pygmy killer whales once. 

 Common dolphins were sighted between Cape Hatteras 

 and Maryland in all seasons, except summer, during the 

 CeTAP^ study but were sighted once in this area during the 

 late summer 1999 survey (Roden^). Common dolphins are 

 expected to occur throughout the area surveyed in 1998 but 

 they may not. Although there are stranding records south 

 of Cape Hatteras (Schmidly''), there are no valid stranding 

 or sighting records of common dolphins in the adjacent 

 Gulf of Mexico despite extensive seasonal surveys of the 

 northern Gulf (Jefferson, 1995; Hansen et al."*). 



Precision 



The precision of the abundance estimates was generally 

 poor. For species or genera abundances, only the estimate 

 for bottlenose dolphins (the most commonly sighted spe- 

 cies), Risso's dolphins, and pilot whales had a CV s 0.50 

 (Table 3). The abundance estimate for the Atlantic spotted 

 dolphin, the most abundant species, had a CV = 0.63. In 

 cases where there is human-caused mortality in a cetacean 

 stock, abundance estimates with a CV < 0.50 are gener- 

 ally required to avoid incorrectly classifying a cetacean 

 stock as "strategic" under the U.S. MMPA (i.e. annual 

 human-caused mortality > annual PBR) less than 10*^ of 

 the time (Wade and DeMaster, 1999). For most species, the 

 variance in the encounter rate, varin ), accounted for more 

 than 70% of the var(N). The distribution of most species 

 was not uniform in the study area and precision might 

 be improved by stratifying estimates by water depth (e.g. 

 shelf and nonshelf) and by area (e.g. north and south of 

 Cape Hatteras). 



Biases in abundance estimates 



The survey was designed to meet the assumptions of line- 

 transect theory (Buckland et al., 1993). However, the abun- 

 dance estimates are negatively biased to varying degrees 

 because the central assumption, that cetacean groups on 

 the transect line are detected with certainty (i.e. ^(0)=1), 

 was not met, and data were not available to correct esti- 

 mates for perception and availability bias. By using the 

 CIO methods described by Barlow (1995), we attempted 

 to estimate the fraction of groups missed on the transect 

 line by the primary observers due to perception bias. How- 



' Roden, C. L. 1999. Report of NOAA ship Oregon II cruise 

 99-05 (236) (a cetacean survey of U.S. Atlantic continental shelf 

 and slope waters between New Jersey and central Florida. 

 August-September 1999), 32 p. Southeast Fisheries Science 

 Center, P.O. Drawer 1207, Pascagoula, Mississippi 39568. 



