260 



Fishery Bulletin 93(2), 1995 



estimate or underestimate densities depending on the 

 universality of the pdf. A criticism of strip-transect 

 methods is that observations outside the strip are not 

 included in the analysis and, in the study of rare ani- 

 mals, every observation is important (Eberhardt et al., 

 1979). In this study, however, only 22 of 171 turtles 

 ( 13%) were sighted at distances greater than 0.30 km 

 from the flight line. In the following comparative dis- 

 cussion, the results of strip-transect analysis are cited, 

 but their use does not affect the overall conclusions. 



Estimated densities of sea turtles on the surface 

 of Core Sound were consistently higher than surface 

 densities for Pamlico Sound (Table 1). Where compara- 

 tive data exist, densities in southern Pamlico Sound 

 were greater than in northern Pamlico Sound. Densi- 

 ties ranged from 0-30.5 turtles/100 km 2 in Core Sound 

 to 0-4.5/100 km 2 in northern Pamlico Sound and to 0— 

 6.5/100 km 2 in southern Pamlico Sound. Densities gen- 

 erally were highest during late spring through sum- 

 mer. Densities in northern Pamlico Sound tended to 

 peak at least one month later than in southern Pamlico 

 and Core Sounds. The estimated density of sea turtles 

 on the surface of the sounds was quite different among 

 the study years; turtles were more abundant in 1989. 



The densities reported in this study are surface 

 densities. Sea turtles are estimated to spend 3.8^11% 

 of their time on the surface (Kemmerer et al., 1983; 

 Keinath et al., 1987; Byles, 1989; Byles and Dodd, 

 1989; Musick et al. * ). Thus, the estimated number of 

 sea turtles on the surface represents a small frac- 

 tion of those actually in the sounds. Because of the 

 large range in proportion of time that monitored- 

 turtles spend on the surface, we did not try to ex- 

 trapolate surface estimates to an estimate of the sub- 

 merged population in the sounds. For comparison 

 purposes, density estimates from studies that made 

 the extrapolation were converted to surface densities. 



Comparison of density estimates among aerial 

 survey studies is confounded by differences in plat- 

 forms and altitudes. Aerial surveys utilizing aircraft 

 equipped with bubble observation windows (Fritts 

 et al., 1983; Thompson et al., 1991; Shoop and 

 Kenney, 1992; Thompson 2 ; Lohoefener et al. 3 ) af- 



1 Musick, J. A., R. Byles, and S. Bellmund. 1983. Mortality and 

 behavior of sea turtles in the Chesapeake Bay. Annual report 

 for the year 1982, NEFC/NMFS Contract NA80-FAC-99994, 

 Virginia Institute of Marine Science, Gloucester Point, VA, 41 p. 



2 Thompson, N. B. 1984. Progress report on estimating density and 

 abundance of marine turtles: results of first year pelagic surveys 

 in the southeast U.S. U.S. Natl. Mar. Fish. Serv, Miami, FL, 59 p. 



3 Lohoefener, R., W. Hoggard, K. Mullin, C. Roden, and C. Rogers. 

 1990. Association of sea turtles with petroleum platforms in 

 the North-Central Gulf of Mexico. Report to the U.S. Dep. Inte- 

 rior, Minerals Manage. Serv., Gulf of Mexico Outer Continental 

 Shelf Regional Off., New Orleans, MMS contract 14-12-0001- 

 30398, OCS study MMS 90-0025, Natl. Mar. Fish. Serv., 

 Pascagoula, MS, 90 p. 



forded observers a direct and unobstructed view of 

 the flight line, thus maximizing the area sampled 

 and the number of sea turtles observed per transect. 

 Conversely, our study and other studies utilizing side- 

 viewing aircraft (Keinath et al., 1987; Lohoefener et 

 al., 1988; Keinath, 1993; Epperly et al., in press, b) 

 did not have downward visibility directly beneath 

 the plane, thereby minimizing the area sampled and 

 the number of sea turtles observed per transect. Like- 

 wise, differences in altitude could affect the number 

 of sea turtles sighted. Smaller turtles have a de- 

 creased chance of being sighted at higher altitudes. 

 The altitude used in this study, 152 m, is consistent 

 with that of the 1982-84 study of the offshore wa- 

 ters between Cape Hatteras and Key West, Florida 

 (Schroeder and Thompson, 1987; Thompson 2 ), sur- 

 veys of the Chesapeake Bay and adjacent waters 

 (Keinath et al., 1987; Keinath, 1993) and surveys off 

 the northern North Carolina coast (Epperly et al., in 

 press, b). It differs from the 229 m altitude used in 

 the 1983-86 and the 1988-89 surveys of offshore 

 waters of the Gulf of Mexico (Thompson et al., 1991; 

 Lohoefener et al., 1990) and the 1979-81 surveys of 

 the offshore waters between Nova Scotia and Cape 

 Hatteras (Shoop and Kenney, 1992). Lohoefener et 

 al. (1988), collected turtle data during their 1987 red 

 drum surveys of the Gulf of Mexico using altitudes 

 of 305-457 m. Fritts et al. ( 1983) collected turtle data 

 during marine mammal, bird, and turtle surveys of 

 the Gulf of Mexico and eastern Florida using alti- 

 tudes of 91 m and 228 m. 



Another factor affecting comparability of density 

 estimates is the proportion of suitable habitat sur- 

 veyed in each study. Comparisons of density esti- 

 mates can be made only for surveys with comparable 

 ratios of suitable to unsuitable habitats surveyed. 

 Suitable habitat presumably accounts for all the area 

 surveyed in inshore studies. Offshore studies gener- 

 ally extended well seaward of suitable habitat and 

 in winter included habitat rendered unsuitable by low 

 temperatures nearshore. Because of methodological 

 differences in aerial survey studies, the application of 

 strip- versus line-transect theory, and our inability to 

 reliably correct surface densities for the proportion of 

 the population that was submerged, comparisons of 

 density estimates among studies are nearly impossible. 

 We compare the results of this study only with other 

 studies with comparable methodologies. 



Our density estimates for Pamlico and Core 

 Sounds, respectively, were comparable to those for 

 the mid- (0-8.5 turtles/100 km 2 ) and lower (0-57.4 

 turtles/100 km 2 ) Chesapeake Bay, Virginia (Keinath 

 et al., 1987). Densities in Core Sound and the lower 

 Chesapeake Bay were particularly high, comparable 

 to density estimates of sea turtles in offshore waters 



